Proceedings of the World Congress on Medical Physics and Biomedical Engineering (San Antonio, Texas, August 6-12, 1988)

DOCUMENT RESUME

ED 306 099 SE 050 495

AUTHOR Clark, John W., Ed.; And Others TITLE Proceedings of the WOrld Congress on Medical Physics and Biomedical Engineering (San Antonio, Texas, August 6-12, 1988). INSTITUTION Alliance for Engineering in Medicine and Biology, Washington, DC. PUB DATE 88 NOTE 445p.; Contains small and light print which may not reproduce well. PUB TYPE Collected Works - Conference Proceedings (021) JOURNAL CIT Physics in Medicine & Biology; v33 suppl 1 1988

EDRS PRICE MF01/PC18 Plus Postage. DESCRIPTORS *Biological Sciences; Biomechanics; Biomedical Equipment; Biomedicine; *Biophysics; *Engineering Technology; Global Approach; Human Factors Engineering; *Medicine; Physics; Research Problems; *Science and Society; *Scientific and Technical Information; Technology

ABSTRACT This document contatns the prcceedings of a joint meeting of the International Federation for Medical and Biological Engineering and the International Organization for Medical Physics. Participants from over 50 countries were in attendance. The theme of the program, "Challenges for the Year 2000," was a reminder of the challenges which confront the world community; specifically, better, more widely distributed, and less expensive health care, and a continuing concern for the well being of humanity and of the earth. The proceedings contain abstracts from over 200 papers, speeches, and sessions, and is organized into medical physics and biomedical engineering sections. (CW)

********n*******************************************xt***************** * Reproductions supplied by EDRS are the best that can be made * from the original document. *********************************************************************** U.S. DEPARTMENT OF EDUCATION "PERMISSION TO REPRODUCE THIS Office of Educational Research and Improvenv int MATERIAL HAS BEEN GRANTED BY EDUCATIONAL RESOURCES INFORM?. ITON CENTER (ERIC) XI This document has been reproduced as received from the person or organization originating it. Patricia Horner 0 Minor changes have been made to improve reproduction Quality

Points of view or opinions stated in this docit TO THE EDUCATIONAL RESOURCES rent do not necessanly represent official INFORMATION CENTER (ERIC)." OER1 positan or policy PHYSICS IN MEDICINE AND BIOLOGY Published by 10P Publishing Ltd (the publishing division of The Institute of Physics) for The Institute of Physical Sciences in Medicine

Editor S.C. Lillicrap, Wessex Regional Health Authority, Royal United Hospital, Combe Park, lath BA1 3NG, UK

Publishing Oft zes: IOP Publishing Ltd, 7 Great Western Way, Bristol BS1 6HE, UK, Telephone. 0272 292151, Telex: 449907 STONEB G

U.S.A..Canada, and Mexico: American Institute of Physics, Dept N/M, 335 East 45th Street, New York, NY 10017, U.S.A., Telephone: 516/349-7800. Telex: 960983

Japan: Maruzen Company, Ltd, 3-10 Nihonbashi 2 Chome, Chuo-Ku, Tokyo 103, Japan, Tele- phone: 272 7211, Telex: J26516 MARUZEN

1988 Subscription Rates: UK 205.00, North America US$360.00, Single issue price UK 17.25, US$30.00. Details of special rates for members of the societies listed below are available on request from the Marketing Department, IOP Publishing Ltd

To receive yourfree samply copy and further information please contact: Journals Marketing Manager, IOP Publishing Ltd, Techno House, Redcliffe Way, Bnstol BS1 6NX, UK

An official journal of the: Hospital Physicists' Association, Medical and Biological Physics Division Canadian Association of Physicists. Deutsche Gesellschaft fur Medizinische Physik, European Federation of Organizations for Medical Physics, Japanese Association of Radiological Physicists, and the International Organization for Medical Physics. Published in association with the American Association of Physicists in Medicine and the Amencan Institute of Physics.

Authorization to photocopy Items for internal or personal use, or the internal or personal use of specific clients. is granted by IOP Publishing Ltd for hbranes and other users registered with the Copyright Clearance Center (CCC). Transactional Reporting Sennce, provided that the base fee of $2.50 per copy Is paid directly to CCC, 21 Congress Street. Salem, MA 01970, U S A. This authorization does not extend to other kinds of copying such as copying for general distnbution. for advertising or promotional purposes. for crating new collective works, or for resale. Payment is additionally required fur copying of articles published prior to 1978, Proceedings of the World Congress on Medical Physics

r and Biomedical Engineering

AUGUST 6-12, 1988 SAN ANTONIO, TEXAS

15th International Conference on Medical and Biological Engineering *3* 8th International Conference on Medical Physics 6th International Conference on Mechanics in Medicine and Biology 30th Annual Meeting of the American Association of Physicists in Medicine 0 41st Annual Conference on Engineering in Medicine and Biology 34th Annual Meeting of the Division of Medical and Biological Physics Canadian Association of Physicists 10th Annual Meeting of the Canadian College of Physicists in Medicine

iphriW. Clark 'Pat:tidal Homer IfiectR,Srnith Kathleen Strum ple;nent'jpti hitsliSmIth4he cooperations of the UntonforPhysical-cinct.Enineeriff g'ciences inedicine motional 4 lie, Afton for Medical Physics Proceedings of the World Congress on Medical Physics and Biomedical Engineering August7- 12,1488, San Antonio, Texas

15th International Conference on Medical and Biological Engineering 8th International Conference on Medical Physics 6th International Conference on Mechanics in Medicine and Biology 30th Annual Meeting of the American Association of Physicists in Medicine 41st Annual Conference on Engineering in Medicine and Biology 34th Annual Meeting of the Division of Medical and Biological Physics, Canadian Association of Physicists 10th Annual Meeting of the Canadian College of Physicists in Medicine

Volume 30 ISSN 0589-1019 LC 61-24788

Publisher. Alliance for Engineering in Medicine and Biology for the Organizing Committee of the World Congress on Medical Physics and Biomedical Engineering

Printed in the United States of America Dependable Printing Company, Inc.

Jr"

H Foreword

We welcome participants from fifty nations of the world to the 1988 World Congress on Medical Physics and Biomedical Engineering. This Congress is a joint meeting of the International Federation for Medical and Biological Engineering (IFMBE) and the International Organization for Medical Physics (IOMP), which are joined together in the International Union of Physical and EngineeringSciences in Medicine. It is the first time this joint meeting has been held in the United States, although the IFMBE has had previous international conferences here in 1961 and 1969. The host organizations from the United States are the Alliance for Engineering in Medicine and Biology (AEMB) and the American Association of Physicists in Medicine (AAPM). Meeting with the World Congress is the Interna- tional Conference on Mechanics in Medicine and Biology. We welcome their participation as well as the participation of the Cana- dian Association of Physicists and the Canadian College of Physicists in Medicine. Although the AEMB involves 20 constituent societies, the Biomedical Engineering Society and the IEEE/Engineering in Medicine and Biology Society have been particularly active in organizing scientific sessions. In addition to the scientific program, there is a large commercial exhibits program, and we appreciate the participation of these exhibitors and the support they have provided for the World Congress. We are particularly appreciative of the efforts of the staff of the AAPM and AEMB and for the volunteer help from members of both organizations. The tireless attention of Dr. David Kopp, Secretary General of the Congress, to innumerable details deserves special praise and thanks; the meeting would not have been ab valuable without him. We also are appreciative of the support provided by our own institutions.

Co-Presidents Gary D. Fullerton,, Ph.D Robert M. Nerem, Ph.D University of Texas Health Georgia Institute of Technology Sciences Center at San Antonio

August 1988

6 111 Greetings!

The Alliance for Engineering in Medicine and Biology and the American Association of Physicists in Medicine welcomeyou to San Antonio, to the United States, and to the 1988 World Congresson Medical Physics and Biomedical Engineering. We anticipate that this will be the largest meeting of its kind ever held andwe are pleased to have the support of a very large number of exhibitors. A diverse scientific and educational program is planned whichwe trust will provide a forum for the exchange of information and experiences. An excellent social program is also planned byour experienced Texas hosts. We hope that this meeting will be fruitful andyour stay en- joyable. We are happy to have you asour guest!

Gary T. Barnes, Ph.D Arthur T. Jonnson, Ph.D President, AAPM President, AEMB =.....

Preface

We welcome you to the 1988 World Congress on Medical Physics and Biomedical Engineering. The process of developing the scientific program has been exciting and rewarding for us and we hope that the week of the Congress is equally as exciting and rewarding for you. The 1988 World Congress presents a record-breaking program with respect to the numbers, excellence and diversity of scientific papers, symposia, short courses and exhibits. Over 1950 abstracts were submitted to the scientific program committee from 43 countries. From outside the USA and Canada, there were excellent responses from Austria, Brazil, China, France, West Germany, In- dia, Italy, Japan, The Netherlands, Spain, Sweden, and the United Kingdom. We would like to especially thank the participants in the scientific program and the vendors exhibiting products at this meeting. The theine of the Congress, Challenges for the Year 2000, is a reminder of the challenges, many of a global nature, which confront us. These problems will only be successfully addressed through the efforts of our world community and a recognition of our universal responsibilities toward batter, more widely distributed and less expensive health care, and a continuing concern for the well being of each other and for our mother earth

Alfred R. Smith, Ph.D John W. C!ark, Ph.D Co-Chairman for Co-Chairman for Medical Physics Biomedical Engineering

V 8 Acknowledgments

As these proceedings go to press, all signs point',o a very successful World Congress on Medical Physics and Biomedical Engineering. I received over 1900 inquiries about the Congress from individuals representing 54 countries. Submission of over 1700 regular abstracts and several hundred works-in-progress further demonstrated this great interest. More than 80 commercial exhibitors subscribed toover 180 commercial exhibit booths. Thesuccess of this Congress must be credited to the efforts of a large number of dedicated individuals.I am extremely grateful to my committee chairmen for their untold contributions. They are: Larry Donovan, Nancy Fullerton, Dennis Gilstad, Bill Hyman, and Al Zermeno. Iam also indebted to Elaine Osterman and to Pat Homer and their staffs, in particular TrishIrmen and Susan Leone, for their guidance and tireless efforts. Finally,I gratefully acknowledge the invaluable secretarial support of Barbara Rowe and the understanding, encouragement and support ofmy wife Claire.

David T. Kopp, Ph.D. Secretary General

v1 9 International Union of Physical and Engineering Sciences in Medicine President: VicePresident: Past President Oivind Lorentsen Alexander Kaul John Mallard Norway Federal Republic of Germany United Kingdom Secretan,i.General: Vice Secretary.General: John A. Hoops Robert Clarke Canada Canada Members of the Council: Nandot Richter John R. Cunningham Orest Z. Roy Hungary Canada Canada Lawrence H. Lanzl Robert Nerem Brian Stedeford United States of America United States of America United Kingdom International Organization for Medical Physics

President VicePresident: Past President: Lawrence H. Lanzl John R. Cunningham Alexander Kaul United States of America Canada Federal Republic of Germany Secretan,i.General: Past Secretary. General: Brian Stedeford Rune Walstam United Kingdom Sweden Affiliate Organizations: Australia Greece Norway Australasian College of Physical Greek Association of Norwegian Society for Scientists in Medicine Medical Physicists Medical Physicists Austria Hong Kong People's Republic of China Osterreichische Gesellschaft fur Hong Kong Association of The Chinese Society of Medizinische Physik Medical Physics Medical Physics Belgium Hungary Philippines Societe Beige des Physicians Medical Physics Section, Philippine Organization des Hopitaux Hungarian Biophysical Society of Medical Physicists Brazil India Poland Associcao Brasileira de Association of Medical Physicists Polskie Towarzystwo Fizyki Medyczanej of India Fisicos em Medicina South Africa Canada Ireland South African Association of Division of Medical and Biological insh Hospital Physicists' Assmiation Physicists in Medicine and Biology Physics, Canadian Association Israel Spain of Physicists Israel Society for Medical Physics Sociedad Espanola de Fisica Medica Columbia Sri Lanka Sociedad Colombiana de Fisica Italy Associazione Italiana di Sri Lanka Medical Physicists' Medica y Radioprotection Fisica Biomedica Association Denmark Japan Sweden Danish Association for Japanese Association of Swedish Association of Physicists Medical Physics Medical Physics in Medicine Federal Republic of Germany Switzerland Deutsche Gesellschaft fur Mexico Asocicion de Fisicos de Societe Suisse de Radiobiologie Medizinische Physik e.V. Hospitales, AC et Radiophysique Finland The Netherlands Thailand Finnish Society for Medical Nederlandse Vereniging voor Medical Physics Club of Thailand Physics and Medical Engineering Klinische Fysica United Kingdom France New Zealand UK National Committee for Medical Societe Francaise des New Zealand Branch, Physics, Incorporating British Physiciens d'Hopital Australasian College of Institute of Radiology and German Democratic Republic Physical Scientists in Medicine Hospital Physicists' Association Sektion Medizinische Physik in der Nigeria United States of America Gesselschaft fur Mathematische and Nigerian Association of American Association of Physicists Physikalische Biologie der DDR Medical Physicists in Medicine 1 0 WI International Federation for Medical and BiologicalEngineering

President: Viceresident: PastPresident: Nandor Richter Robert Nerem c)ivind Lorentsen Hungary United States of America Norway

Secretary.General: Treasurer: Orest Z. Roy Uwe Faust Canada Federal Republic of Germany

Members of the Administrative Council:

Marcello Bracale Richard LG. Kirsner Gunter Rau Italy Australia Federal Republic of Germany Astrid GautierLevine Rudolph Miliner United States of America German Democratic Niilo Saranummi Republic Fumihiko Kajiya Finland Japan Jan Persson Sweden

Affiliate Organizatons:

Argentina German Democratic Republic The Netherlands Sociedad Argentina de Gesellschaft fur BioMedizinische Vereniging voor Biofysica Bioingenieria Technik der Deutschen Demokratishen Republik Non.vay Australia Norsk Forening for Australian Federation for Federal Republic oj Germany BioMedisinsk Teknikk Medical and Biological Deutsche Gesellschaft fur Engineering Biomedirinische Technik E.K. Poland Polish Federation of Austria Finland Engineering Associations/NOT Osterreichische Gesellschaft Laaketieteellisen Fysiikan ja fur Biodemizinische Technik Tekniikan Yhdistys R.Y. South Africa Biomedical Engineering Society Belgium France of South Africa Belgian Society for Medical Societe des Eiectriciens des and Biological Engineering Electroniciens et des Korea and Computing Radioelectriciens (S.E.E.) South Korean Society of Medical and Biological Engineering Brazil Greece Sociedade Brasileira de Greek Society for Spain Engeharia Biomedica Biomedical Engineering Agrupacion Espanola de Bioingenier Bulgaria Hungary Bulgarian National Society Merestechnikai es Automatizalasi Sweden of Biomedical Physics Tudomanyos Egyesulet Svenska Foreningen for and Engineering Medicinsk Fysik och Teknik Israel Canada Israel Society for Medkal United Kingdom Canadian Medical & Biological and Biological Engineering Biological Engineering Society Engineering Society Italy United States of America China Associazione Italiana di Alliance for Engineering Chinese Society of Biomedical Ingegneria Medea e Blologica M Medicine and Biology Engineering Japan Yagoslauta Cuba Japan Society of Medical Yugoslav Society for Medical Sociedad Cubana de Electronics and Biological Engineering & Biological Engineering Bioingenieria Mexico Denmark Sociedad Mexicana de Dansk Medikoteknisk Selskab Ingenieria Biomedica

11 VIII World Congress on Medical Physics and Biomedical Engineering Executive Committee

Gary D. Fullerton, Ph.D. Robert M. Nerem, Ph.D. CoPresident, M, Jical Physics CoPresident, Biomedical Engineering University of Texas HSCSA Georgia Institute of Technology San Antonio, Texas Atlanta, Georgia Alfred R. Smith. Ph.D. John W. Clark, Ph.D CoChair, Scientific Program, Medical Physics Co-Choir. Scientific Program. Biomedical Engineering Hospital of the University of Pennsylvania Rice University Philadelphia, Pennsylvania Houston. Texas David T. Kopp, Ph.D. Jack L. Lancaster, Ph.D Secretary General Treasurer University of Texas HSCSA University of Texas HSCSA San Antonio. Texas San Antonio. Texas Participating Organizations International Union of Physical and Engineering Sciences in Medicine International Organization for Medical Physics International Federation for Medical and Biological Engineering International Conference on Mechanics in Medicine and Biology American Association of Physicists in Medicine Alliance for Engineering in Medicine and Biology Canadian Association of Physicists Canadian College of Physicists in Medicine International Advisory Committee Arabinda Bose. Ph.D. Victor M. Tovar Munoz. PhD. Cancer Centre & Welfare Home Inst Nac de Investigaciones Nucleares Calcutta. India Salazar. Mexico Jean Chavaudra. Ph.D. Bertil R. R. Persson, PhD, Institut GustaveRoussy Lund University Hospital Villejuif. France Lund, Sweden Carlos Eduardo de Almeida. Ph.D. Ervin B. Podgorsak. PhD. Inst de Radioprotecao e Dosimetna McGill University/Montreal Gen Hospital Rio de Janeiro. Brasil Montreal. Canada Tapani Jauhiainen, M.D. Masao Saito, Ph.D. Helsinki University Central Hospital University of Tokyo Helsinki. Finland Tokyo, Japan Robert M. Kenedi, Ph.D. Max Schaldach. Ph.D University of Strathclyde Friedrich Alexander i, iniversitat Glasgow, Scotland Erlangen. West Germany George Kossof, D.Sc.Eng. Samuel Sideman. Ph.D. Ultrasonics Institute TechnionIsrael Institute of Technology Sydney, Australia Haifa. Israel John Mallard. Ph.D. Max E. Valentinuzzi, Ph,D University of Aberdeen Universidad Nacional de Tucaman Aberdeen, Scotland San Miguel de. Tucaman. Argentina Anna Benini Mori, Ph.D. Lojze Vodovnik. Ph.D. IAEA Edvard Kardelj University Vienna, Austria Ljubljana. Yugoslavia Waiter Zingg. M.D. University of Toronto Toronto, Canada

12 Organizing Committee

Al Zermeno, Ph.D Arthur Sherwood, Ph.D. CoChair, Scientific Exhibits/ Posters CoChair, Scientific Exhibits/Posters M.D. Anderson Hospital Baylor College of Mediciae Kenneth Vanek, Ph.D. William Hyman, Ph.D. Chair, Commercial Exhibits Chair, Audiovisual Aids Wilford Hall Medical Center Texas A&M University Dennis L Watts, M.S. Donald Johnson, M.S. Chair, Hotel Liaison Chair, Convention Center Liaison Brooke Army Medical Center Moncrief Radiation Center Lany Donovan, M.A. Nancy Fullerton Chair, Social Program Chair, Companions' Program Brooks Air Force Base San Antonio, Texas Dennis Gilstad, M.D. Jack S. Krohmer, Ph.D. Chair, Publicity/Advertising Chair, Registration/ Information Southwest Research Institute J. S. Krohmer Ph.D., Inc.

American Association of Physicists in Medicine

Elaine P. Osterman Trish Irmen Executive Officer Assistant Executive Officer

Alliance for Engineering in Medicine and Biology

Patricia I. Homer Executive Director

Scientific Program Committee

MedicalPhysics Lowell L Anderson James M. Galvin Dennis L. Parker JenyJ. Battista Jan Van de Geijn Michael S. Patterson PeterJ. Biggs Jean St. Germain Nagalingam Suntharalinga,-n Bengt E. Bjarngard John C. Gore Goran K. Svensson Peter H. Bloch James R. Halama Norman Tepley John R. Cameron R. Mark Henkelman Lynn J. Verhey Paul L. Carson Kenneth R. Hogstrom Jeffrey F. Williamson James C. H. Chu Philip F. Judy Joe P. Windham Robert L Chu Faiz M. Khan Michael L. Wood Laurence Clarke Gerald J. Kutcher Kenneth A. Wright David Cohen Paul P. Lin Martin J. Yaffe J. Michael Collier Mark T. Madsen James A. Zagzebski Jake Van Dyk Seong K. Mun Sandra R. Zink Aaron Fenster Bhudatt R. Paliwal

Biomedical Engineering

Kazuhiko Atsumi Roger D. Kamm Helmut Reul James B. lassingthwaighte Fumihiko Kajiya Peter Rolfe Susan Blanchard Hiroshi Kanai Subrata Saha Colin G. Caro Richard I. Kitney Gerald M. Saidel Dudley S. Childress Periklis Y. Ktonas Louis C. Sheppard Yadin Day' ' Jacques Lefevre Samuel Sideman William J. Dorson, Jr. Dieter W. Liepsch D;ck W. Slaaf Morton H. Friedman Marko S. Markov Willis A. Tacker, Jr. Monique Frize Michael R. Neuman Raymond P. Vito Donald P. Giddens TimothyJ. Pedley Dean C. Winter Craig J. Hartley Robert Plonsey Takami Yamaguchi Robert M. Hochmuth Aleksander S. Pope! Ajit P. Yoganathan Michael Y. Jaffrin John M. Radovich Yoji Yoshida 13 Table of Contents

SECTION I. PLENARY SESSION SCIENTIFIC PAPERS xvii

P1. Space Medicine and Biology: Challenges for the Year 2000 xix P3. Medical Physics and Bioengineering in Latin America: Challenges for the Year 2000 xx

P4. Medical Imaging: Challenges for the Year 2000 . . . xx

P5. Artificial Organs: Challenges for the Year 2000 . . xxi

SECTION II. MEDICAL PHYSICS SCIENTIFIC PAPERS 1

MPSC. Medical Physics Short Courses 3 MPS1. Young Investigators Symposium ...... 7

MPS2. Artificial Intelligence: Promises and Reality . . . . 9 MPS3. Issues in Ultrasound Standards and Safety 10 MP1. Therapy: Treatment Planning 1 12 MP2. .1ntitative Imaging 14 MP3. Medical Lasers 15 MP4. Nuclear Medicine 19 MPS. Photon Calculation 21 MP6. General Medical Physics Posters 1 23 MP7. Radiation Protection 30 MP8. Biomagnetism 1 32 MP9. Digital Imaging 1 34 MP10. Therapy Posters: Hyperthermia, Q.A., Beam Modifiers, Accelerators 36 MP11. Therapy: Dosimetry 1 43 MP12. Radiation Therapy Imaging 45 MP13. Therapy Posters: Hyperthermia, Q.A., Beam Modifiers, Accelerators 47 MP14. Therapy: Electron Dosimetry 56 MP15. Therapy: Brachytherapy 1 58 MP16. Biomagnetism 2 60 MP17. Nuclear Medicine Posters: Magnetic Resonance Spectroscopy and General 62 MP18. General Diagnostic Posters 70 MP19. Therapy: Heavy Particles 73 MP20. Ultrasound 75 MP21. General Radiation Therapy 78 MP22. Magnetic Resonance: Spectroscopy 80

MP23. Imaging Posters: CT, Therapy Imaging, General . . . . 82 MP24. Therapy: Treatment Planning 2 90 MP25. Magnetic Resonance: Imaging 92 MP26. Therapy Posters: Dosimeters and Detectors 94

MP27. Diagnostic: General Imaging 4 99 MP28. Nuclear Medicine: Spect and Pet 102 MP29. Diagnostic: Computerized Tomography 104 MP30. Therapy: Electron Calculations 106 MP31. Imaging Posters: MRI, Spect, Digital, PET 108

4 XI MI=j2. Therapy Posters: Dosimetry for Photons, Electrons, Perturbations 118 MP33. Therapy Posters: Brachytherapy 2 125 MP34. Therapy Posters: Particles, Calculations, Computer Applications 131 MP35. Therapy: Dosimetry 2 141 MP36. General Medical Physics Posters 2 143 MP37. Digital Imaging 2 149 MP38. Therapy: Brachytherapy 3 151

SECTION III. BIOMEDICAL ENGINEERING PAPERS 157

BE1A. Electrophysiology 1: Simulation Methodologies in Biomedicine-- Fundamentals of Excitable Membranes 157 BE1B. Electrophysiology 2: Electrophysiology of Skeletal Muscle 158 BE1C. Electrophysiology 3: Electrophysiology of the Peripheral Nervous System 159 BE1D. Electrophysiology 4: Cardiac Electrophysiology . . 160 BE1E. Electrophysiology 5: Functional Neuromuscular Stimulation 161 BE1F. Electrophysiology 6: Brain Motor Control Assessment 162 BE1G. Electrophysiology 7: Cardiac Potential Mapping . . 163 BE1H. Electrophysiology 8: EEG Modeling 164 BE1I. Electrophysiology 9: Electrical Fields for Defibrillation 165 BE1J. Electrophysiology 10: Models in Automated EEG Analysis 166 BE1K, Electrophysiology 11: Cardiac Arrhythmias 167 BE1L. Electrophysiology 12: Clinical Applications of Automated EEG Analysis 168 BE1M. Electrophysiology 13: Inverse Solutions in Electrocardiography 169 BE1O. Electrophysiology 15: The Next Exciting Use of High Intensity Magnetic Field-- Tissue Stimulation for Clinical Monitoring 170 BE1P. Electrophysiology 16: Clinical Aspects of Cardiac Electrophysiology 171 BE1Q. Electrophysiology 17: Studies in Fibrillation and Defibrillation 172 BE1R. Electrophysiology 18: Quantitative Techniques for Evoked Potential Analysis 173 BE2A. Control of Drug Delivery 1: Modeling and Simulation--

The Effects of Drugs on Physiological Variables . . . 175 BE2B. Control of Drug Delivery 2: Research in Auto matic and Adaptive Control, Programmed Infusion

and Controlled Release of Therapeutic Agents . . . . 176 BE2C. Control of Drug Delivery 3: Clinical Trials and Implementation of Strategies for Controlled Drug Administration 177 BE3A. Cardiac Mechanics 1 178 BE3B. Cardiac Mechanics 2 179

XII 1 5 BE3C. Intramyocardial Flow 1 180 181 BE3D. Intramyocardial Flow 2 182 BE3E. Measurements of Cardiac Performance 1 183 BE3F. Measurements of Cardiac Performance 2 CETA. Past, Present, and Future of Biomedical Engineering; Management and Maintenance of Healthcare Equipment 183 CETC. Clinical Engineering: RoleQualification and Performance . 184 184 BE4A. Clinical Engineering: Education andCareer 186 BE4B. Clinical Engineering 2: Standards andRegulations . 188 BE4C. Clinical Engineering 3: Medical Lasers BE4D. Clinical Engineering 4: Cardiopulmonary Parameters, Measurements ane Techniques 190 *BE4E. Clinical Engineering 5: Ethical Issues in Health Care Technology 191 BE4F. Clinical Engineering 6: TechnologyAssessment . . . 193 Programs . . 196 BE4G. Clinical Engineering 7: Practice and BE4H. Clinical Engineering 8: Clinical EngineeringFinance 197 . . . 199 BE4I. Clinical Engineering 9: Computers andNetwo"ks BE4J. Clinical Engineering 10: Physiological Measurements and Instrumentation 201 203 BE4K. Clinical Engineering 12 BE5A. Orthopedics 1: Bone Remodeling- - Mechanical and Metabolic Factors 205 206 BESB. Orthopedics 2: Bone Mechanics 207 BE5C. Orthopedics 3: Bone Tissue 208 BE5D. Orthopedics 4: Orthopedic Implants 210 BE5E. Orthopedics 5: Spine Biomechanics 212 BE5F. Orthopedics 6: Joint Mechanics and DynamicEffects BE6A. Respiration 1: Stability Analysis of Respiration-Apneas and Periodic Breathing 213 BE6B. Respiration 2: Respiration in Sleep-Advanced Methodologies 215 BE6C. Respiration 3: Gas Exchange in the Respiratory System 216 BE6D. Respiration 4: Modeling of the Respiratory System: Analysis and Identification 217 218 BE6E. Respiration 5: Mechanical Ventilation BE6F. Respiration 6: Models of Mechanically-Assisted Ventilation 219 . 220 BE6G. Respiration 7: Lung Parenchymal TissueMechanics BE6H. Respiration 8: Chest Wall Mechanic:, 221 BE6I. Respiration 9: Respiratory Fluid Mechanics 223 BE6J. Respiration 10: Mechanical Models of the Pulmonary Circulation 224 BE6K. Respiration 11: Aerosol Deposition in theLungs . . 225 BETA. Biological Systems Modeling 1: Interaction Between Artificial Intelligence and Simulation . . 226 BEM Biological Systems Modeling 2: Simulation Methodology in Biomedicine 227 BE7C. Biological Systems Modeling 3: Fractals and Chaos in Physiology and Medicine- - of Molecules, Surfaces, andReactions 229

1 6 XIII BE7D. Biological SystemsModeling 4: Fractals and Chaos in Physiology and Medicine-- On Organs and Organisms 229 BETE. Biological Systems Modeling 5: Modeling Aspects of Cardiac Electrophysiology . 230 BE7F. Biological Systems Modeling 6: Modeling and Signal Processing in Electrocardiography . . . . 232 BE7G. Biological Systems Modeling 7: Methods for Modeling in the Time Domain 1 233 BETH. Biological Systems Modeling 8: Methods for Modeling in the Time Domain2 235 BE8A. Advances in Artificial Organs 1 236 BE8B. Advances in Artificial Organs 2 237 BE8C. Artificial Organs 3: Advances in ArtificialOrgans 238 BE8D. Artificial Heart Valves 1 239 BE8E. Artificial Heart Valves 2 240 BE8F. Artificial Heart Valves 3 242 BE8G. Mass Transport in Artificial Organs 243 BE8H. International Aspects of Medical DeviceRegulation 244 BE8I. Artificial Heart 1: Blood Pump Monitoring and Control of the Artificial Heart 245 BE8J. Artificial Heart 2: Energy Converters and Energy Sources 246 BE8K. Cardiovascular Assist Devices 247 BE9A. Atherosclerosis 1 248 BE9B. Atherosclerosis 2 249 BE9C. Atherosclerosis 3 251 BE9D. Atherosclerosis 4 252 BE9E. Atherosclerosis 5 253 BE9F. Atherosclerosis 6 254 BE10A. Heart Rate Variability 1 255 BE10B. Heart Rate Variability 2 257 BE10C. Heart Rate Variability 3 258 BE11A. Flow in Collapsible Tubes 1 259 BE11B. Flow in Collapsible Tubes 2 261 BE12A. Poster Session 1: Flow in Collapsible Tubes 262 BE12B. Poster Session 2: Cardiovascular Mechanics 1 . . . . 263 BE12C. Poster Session3: Renal Systems/Hemodialysis . . . . 268 BE12D. Poster Session4: Clinical Engineering 1-- Administrative Aspects /Equipment and Maintenance . . 270 BE12E. Poster Session5: Clinical Engineering 2-- Computer-BasedAnalysis Systems 274 BE12F. Poster Session6: Artificial Heart Valves 277 BE12G. Poster Session7: Microcirculation 280 BE12H. Poster Session 8: Cardiovascular Mechanics 2 . . . 285 BE12I. Poster Session 9: Ultrasound in Medicine 291 BE12J. Poster Session 10: Cardiac Electrophysiology . . . 292 BE12K. Poster Session 11: Clinical Engineering3 -- Indirect Cardiac Assessment Techniques 295 BE12L. Poster Session 12: Orthopedics 299 BE12M. Poster Session 13: Respiratory System 302 BE12N. Poster Session 14: Mass Transfer 305 BE120. Poster Session 15: Bioheat Transfer 307 BE12P. Poster Session 16: Biomedical Engineering:General 309

XIV 1 7 BE12Q. Poster Session 17: Clinical Engineering: General . . 316 BE12P. Poster Session 18: Cardiac Potential Mapping . . . . 325 BE12S. Poster Session 19: Electroencephalography 327 BE12T. Poster Session 20: Expert and Decision Support Systems 330 BEM. Poster Session 21: Sound Production and Audition . . 331 BE12V. Poster Session 22: Medical Imaging 333 BE12W. Poster Session 23: Biomedical Applications of Spectral Analysis and Signal Processing 335 BE12X. Poster Session 24: Stereometrics and Special Surgical Techniques 337 BE12Y. Poster Session 25: Nerve and Muscle Electrophysiology 339 BE12Z. Poster Session 26: Traditional Chinese Medical Science 341 BE12ZZ. Poster Session 27: Cellular Bioprocessing 342 BE13A. Soft Tissue Mechanics 1 344 BE13B. Soft Tissue Mechanics 2 345 BE13C. Soft Tissue Mechanics 3 346 BE14A. Rheology 347 BE14B. Microcirculation: Mechanics of the Vascular Wall 1 349 BE14C. Microcirculation: Mechanics of the Vascular Wall 2 350 BE14D. Microcirculation 3 350 BE14E. Cellular Biomechanics 1 352 BE14F. Cellular Biomechanics 2 353 BE14G. Cellular Biomechanics 3 354 BE15A. Ultrasound in Medicine: Cardiac Doppler Methods . . 356 BE15B. Ultrasound in Medicine 2: Doppler Velocity and Flow Measurements 357 BE15C. ,ltrasound in Medicine 3: Doppler Signal Processing 358 BE15D. Ultrasound in Medicine 4: Intravascular and Intraoperative Techniques 360 BE15E. Ultrasound in Medicine 5: Improvements in Imaging 361 BE15F. Ultrasound in Medicine 6: Pressure and Tissue Property Measurements 363 BE16A. Auditory System Research 364 BE16B. Vision 1: Eye Movements/Glaucoma 365 BE16C. Vision 2: Normal and Abnormal Visual Function . . . 367 BE17. Analysis and Control of Movement . 369 BE18A. Hemodynamics 1 370 BE18B. Hemodynamics 2 372 BE18C. Hemodynamics 3 373 BE18D. Hemodynamics 4 374 BE18E. Hemodynamics 5 376 BE18F. Hemodynamics 6 377 BE18G. Hemodynamics 7 379 BE19. Biomedical Signal Processing 380 BE20A. Impedance Methods 1: Acquisition of Physiological Data Using Electrical Impedance . . 382 BE20B. Impedance Methods 2: Cardiac Measurements Using Electrical Impedance 383

1 8 BE21A, Rehabilitation Engineering 1:Control of Muscles Through Electrical Stimulation- - Techniques and Systems 385 BE21B. Rehabilitation Engineering 2: Stimulation

of Biological Tissue: Electrodes and Methods . 386 BE21C. Rehabilitation Engineering 3: Ambulation Systems for Persons with Disabilities-- Orthoses and Prostheses 387 BE21D. Rehabilitation Engineering 4: Integration of Disabled Persons with their Environment 389 BE21E. Rehabilitation Engineering 5: Rehabilitation Technology and Functional Restoration 390 BE22A. liage Analysis 1 391 BE22B. Image Analysis 2 393 BE23A. Acupuncture and Related Electrostimulation Techniques 1: Acupuncture-- A Tutorial Overview 394 BE23B. Acupuncture and Related Electrostimulation Techniques 2: Relationship Between Tens and Acupuncture 395 BE23C. Acupuncture and Related Electrostimulation Techniques 3: Tens and Acupuncture-- Controversial Twins 397 BE23D. Acupuncture and Related Electrostimulation Techniques 4: Low Current Transcranial Stimulation 398 BE23E. Acupuncture and Related Electrostimulation Techniques 5: Legal and Regulatory Issues 399 BE24A. Biosensor Research 1: Microelectronics- Based Sensor Systems 400 BE24B. Biosensor Research 2: Oxygen and Glucose Sensors . . 401 BE24C. Biosensor Research 3: Optically-Based Sensors . . . 403 BE24D. Biosensor Research 4: F,'essure Sensors and Measurements 404 BE24E. Biosensor Research 5: Biotelemetry 405 BE25. The Influence of Electromagnetic Fields on Living Tissues 406 BE26A. Bioheat Transfer 1 407 BE26B. Bioheat Transfer 2 409

19 XVI Section I

Plenary Sessions

20 P1.1 P1.2 Medical Researchand theFuture of The German D-Spacelab Missions: Opportunities for Manned Space Flight, C. L. Huntoon, International Life Sciences Research in Space, Karl.E. Klein, Space andLife Sciences Directorate, DFVLR Institute for Aerospace Medicine, D-5000 Kan, FRG NASA/Johnson Space Center, MailCode SA, Houston, Texas 77058, United To utilize the opportunity "microgravity" is of. _ring for studies of States. basic phenomena in fluid physics, material sciences, biology and medicine, the Federal Republic of Germany has initiated a series of microgravity missions with Spacelab. The first one, Spacelab One of the most complex challenges to medical Mission D-1, orbited the earth between 30 October - 6 November science in the twentieth centuryis to make 1985. It had three European Scientists on board, specially trained Truman habitation possible in space and on other to operate a payload compiled from proposals of seven European planets. Over a 30-year period, the physiologic countries and the USA. For the first time, science crew training, response ofhumans to spaceflighthas been experiment integration and payload operation control occured investigated duringshortflights,flights of outside of the USA, i.e. in Germany. more than 300 days, flights to the Moon. Most From a total of 75 experiments one third was devoted to studies in of the systems of the body are affected in some the life sciences with the main objectives being cell functioning, way by conditions of space flight, but research developmental processes, gravity perception, cardiovascular adap- has given us hope that the most serious problems tation and cognitive behaviour. Most of the results were of high can be circumvented. Space medicine and tech- scientific interest, some even unexpected. Among others, they nology have advanced to the point where continu- have ous human habitation of lowEarthorbit is - shown evidence, that - against a long-standing opinion - becoming a reality and a manned mission to Mars gravity effects unicellular systems (like bacteria), is beingdiscussed. Inthissession we are - indicated a synergistic negative effect of microgravity and going to hear about medical findings from long- cosmic radiation on the development of an insect, and and short-term space flights, as well as what it - produced data appropriate to strongly question hypothesis on is like to live and work in microgravity and human vestibular organ responses, valid since many years. some of the experiments planned for future The second mission in this series, the German Spacelab Mission flights. D-2, originally planned to fly in 1988 now is scheduled for 1991. Again, the payload will be microgravity oriented and will contain experiments from many countries, including Japan and the USA. The Life Sciences are represented with three payload elements, including a human physiology research device "ANTHRORACK", and a "BIOTEX"for biotechnology studies.

P1.3 P1.4 Science Operations in Space, Bonnie J. Status of Japanese Spacelab (SL-J) Experiments Dunbar, Mission Specialist Astronaut, Chiaki Mukai, National Space Development Agency NASA-Lyndon B. Johnson Space Center, of Japan, 2-4-1 Hamamatsu-cho, Minato-ku Tokyo Houston, TX 77058. 105 Japan The SLJ mission is the first step of Japanese Throughout the course of human exploration in full-scale space experiments that utilize the space, astronauts have been an integral part of characteristics of the space environment. Thirty- scientific experimentation. During the Apollo and Four experiments (22 on material processing and 12 Skylab programs, astronaut crews conducted scien- on life science) from Japan and several from the tific investigations both as initiators and as ex- U.S. are scheduled to be performed aboard space tensions to ground-based principal investigators. shuttle/spacelab in 1991. Experiments were conducted in a variety of disciplines: human physiology, astrophysics, fluid phys- Material processing experiments will produce new ics, botany, biology, and materials science. The materials in the fields of compound semiconductors advent of the Space Shuttle program and Spacelab and metal-based composits. These experiments will flights has brought more sophisticated experiments, develop space and earth-based processing tech- increased demand of experiment and crew time, and nology and investigate physical phenomena such the need to optimize the human element with auto- as Marangoni convection and gas evaporation. mation. Lessons learned with respect to experiment protocol, hardware design, operations, and "in Life science related experiments are dedicated to flight" maintenance will be discussed. fields of biotechnology (free flow electrophoresis, Recommendations for the development of the Space crystal growth of enzyme, cell culture) and Station eperiments will also be reviewed. biological adaptation to space environment (vestibular function, perceptual-motor function, visual stability, calcium metabolism, endocrino- logy, radiation biology, circadian rhydmi).

SLJ mission status including ground-based control data and hardware for these experiments, especially those focused in the life science field, will be discussed.

1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas XIX 21 P3.1 P3.2 PROBLEMS CONCERNING THE DEVELOPMENT OF MEDICAL PHYSICS IN LAT- MEDICAL PHYSICS AND BIOMEDICAL IN AMERICA, Carlos E. de Almeida, SSDL/Rio de Janeiro-Brazil. ENGINEERING IN ARGENTINA: CHALLENGES FOR THE YEAR 2000. Max E.Valentinuzzi, In the last decades medical physics in Latin America have im- Bioingenierla, Universidad Nacional de proved considerably, nevertheless it is common understanding Tucuman, (UNT), cc 28,suc 2,(4000) that there is still a long way to go in order to reach a de- Tucumlan, Argentina. sirable level of quality. With the advent of linear accelerators the demand for physicists in radiotherapy have incre- Adequate health care on a widespread basis poses a major problem in Argentina. ased but unfortunately in diagnostic radiology and nuclear It faces largo geographical distances, medicine +t has bren Tuit .tationary regardless to the fact poor communication and telecommunication that sophisticated technology has been introduced i.e. CT, systems,and serious insufficiency of SPECT, NNR andULTRASOUND. The creation of the SSDL in se- economic resources. The current situation veral countries ha. been very positive in establishing tra- of our health care system (without considering afew advanced and well ceability to the mf_surements .4 well as to serve as support equiped centers, usually located in big for training. The absence of formal training centers with cities) can be labeled as distressing and the support of Internationa: Organizations contribute to the of serious concern, specially if children increase lack of appropriate education. Several physicists health, prevention of diseases, early sent abroad for training have either not returned of changed diagnosis ant emergencies, taking into account regional pathologies, are jobs or position in their own organization due to lack of consiuered as priorities. This setting scientific motivation or salary.Regional projects i.e. brings us to evaluating a possible high Areal/IAEA oust be explored in order to stimulate the exchan- tech highcost oriented health care ge of experiences well as the Nuclear Information Centers to system versus a lowtech lowcost one. minimize the difficulties in ac ,ss current scientific papers. The first will benefit in all probability a medium to high income minority group in The Latin America situation still deserves more coordinated a relatively narrow band of pathologies, efforts in the areas of training. QA orograms, research and making believe that the system is among a better recognition by the national authorities of the im- one of the bestsystems on an portant role played by medical physics to society. international scale. The second, instead, will probably benefit a low income majority in a wide spectrum of diseases.

P4.1 Expanding the Extraction of Diagnostic P4.2 Information from the Body, Gary D. Fullerton and ManipulatingTissueParameters andDiagnostic Jack L. Lancastrr, Radiology Department, Data to Yield IncreasedDiagnostic Information, University of Texas HSCSA, San Antonio, TX, USA William R. Hendee, Vice President for Science and Technology, American Medical Association, Positron emission tomography (PET), nuclear mag- Chicago, IL, USA netic resonance imaging (MRI), and other computer imaging techniques dramatically expand the range Improved diagnostic capabilities ofmodern of diagnostic information. In spite of thismany imaging methods depends to a large extent onthe physicians and scientists remain unaware of the possibility of influencing contrast properties of broader range of physical parameters available to the body by new receptor specificpharmaceutical provide anatomical and physiological images of a or contrastagents. These materials provide patient. X-rays were limited to images of den- agents to manipulate diagnostic data to reveal sity variation with small overlays of atomic num- subtle details of the patients response. This ber information. MRI images proton density modu- presentation will project this topic into the lated by T1 and T2 parameters determined by mo- future to look at ways to combine information, tional characteristic of biowolecles and asso- match images to observer characteristics, and use ciated water. But it can also be used to create "smart" devices and other computer-level contrast related to flow. Both PET and MR can be techniques as an aid to the diagnosticprocess. used to image the biodistribution of isotopically It will also look at our current understanding of enriched metabolites. We now have the possibil- how we detect and process visual data, and how we ity of imaging in vivo physiological processes. need to improve that understanding to accommodate The location and point of effect of labeled drugs new techniques suchas 3D displays, computer can be followed directly. Thus modern medical visualized images, and other diagnostic aids. imaging has not only application as a provider of anatomical pictures but also as a means of following complex biochemical processes.Such applications, however, demand an understanding of human biochemistry and organ pathologies that pushes the limits of present understanding. Full diagnostic and therapeutic use of these new imaging devices presents a scientific challenge of great import for the year 2000.

XX 1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas 22 P5.1. Clinical Readiness and the Artificial Heart John T. Watson, National Institutes of Health, National Heart, Lung, and Blood Institute, Bethesda, MD, USA

Clinical Readiness and the Artificial /lean

Total mortality from cardiovascular disease continues to decline while the incidence and prevalence of end-stage heart disease continues to increase. Congestive heart fai.ure is now the nation's number one listed Diagnosis Related Group. Presently, biologic or mcaNanical replacement of cardiac function are the only treatment options for ad-stage heart disease. Cardiac transplantation is eminently successful, but its application is limited by donor availability to several huntred procedures per year. It is estimated that additionally 17,000 to 35,000 patients per year below age 70 would benefit from an effective, safe, fully implantable mechanical circulatory support device.

The research, development and reliability testing of electrically powered implantable ventricular assist devices is nearing completion. Powered by electro-chemical batteries, these systems are now being tested for patient safety in laboratory and animal tests. These systems represent a now frontier for technological complexity and reliability never before requirkd for an implanted cardiovascular device.Further, these systems reouire automatic control that mat meet the daily circulatory requirements of the implant recipient for a period of two years of uninterrupted function. The next immediate goal is systems that operate effectively for five years and eventu !ly for ten years or longer. The quality of life and longevity of current cardiac transplant recipients will provide the benchmarks for implantable mechanical circulatory support devices to emulate in the future.

The crgoing Device Readiness Testing Program is designed to provide laboratory reliability and antral test data that will demonstrate that an implantable VAS is ready for clinical evaluation it human patients. Twelve systems are undergoing two year bench testing on a rock-loop. Success is being measured by an 801 reliability at a 70% confidence level. For a sample site of I? this represents to 0 or one system failure. A failure being defined as a Degradation of cardiac output to less than 4 L/min for over 60 seconds. VAS peeformance will be validated in animals during acute studies of transient behavior and twenty-five months of chronic investigations.

Having convincingly shown that the VAS works effectively in the laboratory and safely in animals, a clinical evaluation will be initiated in patients with end-stage heart disease and for whom there is no alternate therapy. The program will have a VAS production center and probably two clinical centers performing 20 implants per center.

1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas XXI 23 Section II

Medical Physics Scientific Papers Medical Physics Scientific Papers

MPSC1. MPSC2. Artificial Intelligence and Expert Medical Laser Concepts and Applications Systems in Medicine. J.A. Siegel, Daniel R. Doiron, Ph.D., Laserguide, Inc. Temple University Hospital, Santa Barbara, CA 93117 Philadelphia, PA 19140 The firs! and still the largest area of application Since the development of digital computers, there is in ophthalmology, but other areas of application has been a growing interest in the creation of are increasing and expanding rapidly. The use of "intelligent" computers. Machines that use com- carbon dioxide lasers for vaporization of tissue is plex information in intelligent ways are said standardly used in various surgical specialties. to exhibit artificial intelligence (AI). AI is The high power Nd:YAG and argon lasers are finding a branch of computer science concerned with the significant applications in endoscopy because of study of: 1) methods for knowledge representation their deliverability using small optical fibers. and 2) techniques for searching that knowledge These lasers are primarily used for coagulation, but in order to derive intelligent conclusions. Its since the introduction of contact delivery systems major areas of application include natural language they are now being used for vaporization and cut- processing, robotics, machine vision, and expert ting. New lasers and medical applications for them systems. An expert system is a rule-based computer are beginning to take advantage of the lasers pGten- algorithm, consisting of a knowledge base and ial of producing very short pulses with extremely an inference engine, designed to emulate the high peak powers. Applications such as laser litho- decision-making behavior of an expert in a specific tripsy and angioplasty are examples cf these new field or domain. Several expert systems respons- applications. In addition, the monochromatic and ible for the birth of AI in medicine (AIM) have possible tuneability of the lasers is now being been developed. Some of these have been so popular used in areas such as Photodynamic Therapy and the that their frameworks or shells (a shell is essent- treatment of birth marks and tattoos. A description ially the expert system without its knowledge of the various lasers, their characteristics and base) are now commercially available. These shells applications, will be presented. are currently being used to develop many new expert systems in medicine. This course will include: 1) a historical perspective of AI, 2) the fundamentals of expert systems, 3) the languages of AI, and 4) the use of AIM.

MPSC3. MPSC4. Hyperthermia Dosimetry Quantitative Techniques Utilizing Computerized L.H. Gerig, Ontario Cancer Foundation Radiologic Imaging Instrumentation. Ottawa, Ontario, Canada K1H 8L6 Stephen R. Thomas, Department of Radiology, University of Cincinnati, Cincinnati, OH 45267 Clinical hyperthermia for cancer therapy Along with the capability for implementation of is the process of elevating rumor bearing well recognized functions (e.g. image tissues to a temperature in the range reconstruction, display manipulation, dir..tal 42° -> 45° for a specified time, archiving, etc.), computerization of radiologic The clinical importance of hyperthermia imaging modalities has offered the promise of as an effective synergistic adjuvant to providing quantitative information. Three cancer radiation therapy or chemotherapy principal categories for quantitation are : (a) is becoming clear. Without adequate spatial, (b) temporal, and (c) feature thermal mapping, hyperthermia will never characterization. Frequently, the defined realize its full potential. parameters of interest cross category boundaries This short course will describe various and incorporate multiple facets. This course will techniques for temperature measurement review the theoretical objectives, procedural and the problems associated with the methodology, limitations and practical clinical techniques. A detailed analysis of the utility for various quantitative imaging clinical use of invasive thermometry will techniques including: 1.) X-ray CT: xenon enhanced be given. The course will discuss the cerebral blood flow, bone mineral densitometry; 2) factors which influence the temperature SPECT: regional quantitation of absolute reported by an invasive thermometry radioactivity, spatial distribution; 3) PET: system and will describe how problems of tracer kinetic models appliedto physiologic, measurement error can be overcome. biochemical and hemodynamic processes (e.g. cerebral blood flow or volume, metabolism, oxygen extraction ratios, receptor density, neurostimulated foci); 4) ultrasound: tissue characterization through back scatter analysis, doppler quantitation of flow velocity; and,5) magnetic resonance imaging: tissue differentiation through calculated relaxation parameters (11, T2), spectroscopic analysis of metabolic components.

19V World Congress on MedicalPhysicsand Biomedical EngineeringSan Antonio, Texas 3 25 Medical PhysicsScientific Papers

MPSC5. MPSC6. Requirements in radioimmunotherapy. Biomagnetism: Principles and Applications in the S-EStrand, Radiation PhysicsDepartment, Lund Study of the Human Brain. Gian Luca Romani: Istitu- University, S-221 85 Lund, Sweden to di Fisica Mace, Universita' di Chieti, and In- stituto di Elettronica dello Stato Solido- CNR, Via The limiting factor in radioimmunotherapy is the Ceneto Romano 42, 00156 Roma: ITALY absorbed dose to normal tissues. Therapy, using The investigation of the physiology and pathology of radiolabelledmonoclonal antibodies, reovires a the human brain by neuromagnotic measurements ispro- proper dose calculation, involving the radlobiolo- viding impressive results and raising increasing gical effects on tumor and normal tissues. The interest from the clinical world.The state of the dosimetric considerations must involve the tumor art of instrumentation permits to detect the spatial geometry i.e. large tumors, small tumors, microme- distribution of magnetic fields over the scalp by tastases or circulating cancer cells. means of 'multichannel' systems featuring a few adjacent magnetic sensors. Each of these measuring Very few published papers on the possibili- units consists of a superconducting magnetometer, ties for radioimmunotherapy, however, have been namely a (dc) SCUM and a superconducting detection focused on the dosimetry, coupled to radiobiologi- coil. This kind of instrumentation is only partially cal considerations. The aim of this course is to satisfying, in that 1 sequential positioning of the review the most important parameters nessesary to sensors over adjacent regions of the scalp is =tir- take into consideration in radioimmunotherapy. ed to detect the whole field distribution associated to a specific source, with significant limitations The review will include suchparameters as to the measurement itself.Nevertheless, important different monoclonal antibodies, selection of findings are being collected in the study oftheor- therapeutic radionuclides, antigen expression on ganization of cerebral primary areas as wellas in tumor cells, biokinetics in vivo, compartment that of some diseases, like focal epilepsy. Several analysis, absorbed dose calculations, dose plan- projects are being carried on in different countries ning, dose rate, radiation biology,etc. to develop large multichannel systems, withas many as 100 adjacent sensors, which in the near future will permit to detect magnetic activityover the whole scalp and, with the use of adequate, devoted computers and of appropriate modeling of neuralac- tivity, will provide real time functional imaging.

MPSC7. MPSCS. Imaging Applicationsin Radiotherapy. Safety Considerations R. Mark Henkelman, for Medical Department of Linear Accelerator, J.A. Purdy, Medical Biophysics,University of Washington UniversitySchool of Toronto, Toronto, CanadaMU 1K9 Medicine, St. Louis, MO,USA.

Medical Imaging is inthe midst of dramatic technological development. Sectional techniques Recent pa;..t events involving medical such as ultrasound, computed tomography, magnetic linear accelerators whichresulted in resonance now pr.:wide detailedthree dimensional overexposures to radiotherapypatients tumour and anatomicaldefinition. have prompted theradiation therapy community to re-evaluatethesafety What is the impact lfthis change for aspects of these systems. radiotherapy? Multimodality Improvements in imagingimpact all treatment units are of particularconcern aspects of oncologicalmanagement: screening, as are those units whichu lize computer diagnosis, staging, response, follow-up, relapse technology in their controlsystems. This and restaging. For radiotherapy planning, refresher course willreview safety developments in imagingpresent the challenges of matters pertaining to medicallinear more precise three dimensional planning. But accelerator acceptance testing,periodic even more important is theprospect of moving qualityassurance testing, operator beyond geographic definitioninto the use of training and incident rep.rting. imaging for tumourcharacterization. Particular attention will bedirected to tests of the operation ofthe entire interlock and computersystems.

4 1988 World Congress on Medical Physics and BiomedicalEngineeringSan Antonio, Texas 26 Medical Physics Scientific Papers

MPSC9. HPSCIO. Computer ManagementSystems in Clinical Facilities Monte Carlo Simulation of Electron- Don P. Ragan, WayneState University Photon TYansport and Applications Detroit, Michigan to Radiation Dosimetry Clinical Management Systems have moved from curiosity to necessity. Physicists and engineers D.W.O. Rogers. aid Alex F. Bielajew, need to take an active part in departmental and Ionizing Radiation Standards, Division of institutional information systems since these Physics, NRCC, Ottawa, Canada K1A OR6 systems are used to measure productivity and profitability and are often the basis for The course begins with a very brief introduction to management decision making. Not only can these the Monte Carlo simulation of electron and pho- systems provide motivation for improved staffing but they can offer substantial quality assurance ton transport followed by a description of the cur- tools. This course will review the characteristics rent state-of-the-art. This will draw heavily on the of departmental information systems. Also International School held last year at the Ettore considered will be acquisition of systems Majorana Centre in Erice, Sicily, on the subject including administrative justification. Computer requirements, interfacing to hospital information of "Monte Carlo Transport of Electrons and systems, cost of operation in terms of personnel, Photons below 50 MeV" (proceedings published time and money and cost benefit will be reviewed. by Plenum Press, 1988, edited by W.R. Nelson et Popular systems in radiology and radiation al.).Special emphasis will be given to contrast- oncology will bc discussed. Limitations and future prospects wi.1 be presented. ing the two major general purpose code systems, ETRAN(ITS) and EGS, and to discussing recent extensions and corrections to these systems. The course finishes with several examples of radiation dosimetry applications which will demonstrate the degree of sophistication and accuracy which is possible with todays software and hardware technology.

MPSC11. MRSC12. Update on Heavy Particle Therapy. An Introduction to Photodynamic Therapy Borje Larsson, Department of Radiation Sciences, M.S. Patterson, Ontario Cancer Foundation Box 535, S-751 Uppsala, Sweden Hamilton, Ontario, Canada, L8V 1C3

Medium-energy ion accelerators are engaged in ra- Photodynamictherapy (PDT) is a method of diotherapy with "heavy particles", i.e. protons, treating localized malignancies in which or heavier ions, neutrons or negative pions. cell-killing is due to chemical reactions Thousands of patients with deep-seated tumors or initiated by the interaction of light with a vascular malformations have now been treated, and photosensitizing drug. The principle has been the prospects for further developments seem to be known for close to a century but clinical use significant. The aim of this course is to con- has grown only recently due mainly to the advent front the participants with the radiological char- of selective photosensitizers, laser light acacteristicsof the various members of the fam- sources, and fiber optics. About a hundred ily of heavy particles. groups around the world are now involved in PDT and, as this number increases, medical The major part will be devoted to physical and physicists are being called on to support these biophysical aspects on the various forms of heavy clinical efforts. particle therapy: interaction of the particles This course will be aimed at such Physicists and with living matter, dosimetry, dose planning, and will stress the "nuts and bolts" required for radiobiology. The accelerator technology will clinical applications. After a brief discussion also be briefly reviewed and problems concerned of the photophysics of PDT, the principles and with the introduction of new radiotherapeutic mo- peculiarities of different lightsourceswill be dalities in clinical practice will be discussed. discussed. This will be followed by a description of various irradiation methods (e.g. To facilitate the understanding of the role of intralumenal, interstitial) and the technology radiotherapy in the treatment of cancer disease, required to achieve them. The important topic today and tomorrow, some relevant themes in the of light dosimetry will be discussed from the evolution of competing or supplementary modalities standpoint of clinical measurement and the will also be touched upon. uncertainties involved. Finally, several areas of research in PDT will be indicated which would benefit from the participation of medical physicists.

1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas 5 27 Medical Physics Scientific Papers

MPSC13. MPSC14. Time Dose Relationships in Radiotherapy.D. Recent Advances in ?RI and MRS Herbert, University of South Alabama, Dept.of John C. Gore Radiology, Mobile, AL. and C. Orton, Harper Y210 University School of Medicine Hospital and Wayne State University, Dept. of New Haven, CT06510 U.S.A. Radiation Oncology, Detroit, MI.

The course consists of two parts. The firstpart In this presention the performance and provides a perspective on the principal clinical potential roles of several advances in NMR dose-time models. In this part thebackground, techniques will be described. Whole body MRI development and clinical uses of each ofthese currently provides excellent images with good models are examined and compared. Modelsof the soft tissue contrast. Recent trends have responses of normal tissues and tumoursto emphasized faster scanning methods to reduce brachytherapy and teletherapy regimensare motion artifacts and provide dynamic considered. A brief introduction tothe information. Extremely fast images can be deficiencies of these models will bepresented._ made using echo planar techniquesthat can The second part of the courseprovidesa resume image the human body in 1/40 second. These of modern statistical methods forthe construc- and other schemescan be used to quantify blood flow. tion, criticism, validation, discriminationand In vivo spectroscopy with good deployment of models of dose-response. spatial localization cannow be achieved to Regression Diagnostics for identificationof detect protonated or phosphorylated observations that are not well explainedby the metabolites in humans, and thesemethods are model or that may dominatesome important being evaluated at fields up to 4.0T with aspects of the fit - or both- are introduced. promising results. NMR images of small samples at high field can be made with Bayesian regression methods for combininga priori and sample informationon the model resolution of under 10 Amso that microscopic parameters are considered. Resampling imaging is also a developing areaThe technological advances needed (jackknife) methods of model validationand to further such estimation of ratios of parametersare examined. endeavours will be addressed. Examples of the use of each of thesemethods on data taken from the literatureare provided.

MPSC15. MPSC16. Radiology Image Management Network. Recent Advances in Ultrasound Imaging Samuel J. Dwyer III and Larry T. Cook University of Kansas MedicalCenter, Kansas City, F.S. Foster, Ontario Cancer Institute KS66103 500 Sherbourne St., Toronto, Ontario, Canada M4X 1K9 Radiology image management networksare designed to acquire, display and manipulate,archive and Ultrasound imaging plays an important role in the generate hardcopy recordings of digitallyformat- detection and management of a wide range of human ted image data. These networks are combinations diseases. Although ultrasound has been used of Local Area Networks and microcomputersystems. predominantly in obstetrics, abdominal imaging Imagin- management networksare currently required and dynamic imaging of the heart, the advent of to acquire 1 GByte of image dataper day and fu- new technologies and scanning approaches has ture requirements will be 50 GBytesper day. Dis- opened up a number of promising areas of play workstations are requiredto display and ma- specialization in diagnostic imaging. Among nipulate 15 GBytes of image dataper day. Archiv- these techniques, neo-natal imaging, ing requirements exceed 10TeraBytes. The exper- intra-operative ultrasound, peripheral vascular ience of implementing Radiologyimage management imaging, realtime blood flow mapping, tissue networks is limited. Only a few departments have characterization and breast imaging are important operational networks in portions of theclinical examples. The purpose of this course is to facility. The benefits of these networksinclude examine the physics behind the newest innovations faster patient throughput andincreased referral in medical ultrasound imaging and to explore physician response. potential future developments in this rapidly expanding field. Topics include. ., e.s....:cronic beamforming with phased linear and annular arrays; 2) effects of beamforming parameters on image quality; 3) the physics and applications of realtime two dimensional Doppler blood flow mapping and 4) progresS in tissue characterization and studies on the interaction of ultrasound with tissue.

28S

6 1988 World Congress on Medical Physics and BiomedicalEngineeringSan Antonio,Texas Medical Physics Scientific Papers

MPS1.2 MPS1.1 DISTRIBUTION AND DOSIMFTRY IN Fast Neutron Dosimetry Using Positron RADIONUCLIDE Emission Tomography, J. Kleck*, J. SOFT TUMOR TISSUE:IMPLICATIONS*TO Smathers, L. Myers, M. Mandelkern and RADICIMMUMOTHERAPY, R.W. Howell , D.V. Rao, W. Digby, University of California, Los University of Medicine & Dentistry of New Angeles U.S.A. Jersey, Newark, NJ;and K.S.R. Sastry, The UCLA Neutron Therapy Facility uses a University of Massachusetts, Atherst, !l7. p(46)Be neutron beam for cancer therapy treatment. Patients undergoing treatment Current methods for estimating the absorbed dose framradionuclideslocalizedin solid tumors were measured for neutron induced assume the activity is uniformly distributed. radioactivity with an Intrinsic Germanium The growing evidence that tumoruptake of gamma spectrometry system. The principal radiolabeledmonoclonal antibodies is Iughlly isotopes produced were identified as C-11 nonuniform requires new methods of calcula and 0-15. Carbon-11 and 0-15 decay by the tumor dose. We have developed a spheri positron emission with half-lives of 20.3 tumor model, in which the radiald,wiericeof the tumor dose is examined as a function of tumor min. and122 sec., respectively. The within the tumor, and radionuclides were determined to be size, activitydistribution radiation type and energy. Turprs of diameters predominantly from C-12(n,2n)C-11and 1.Q cm, containipg spherically 0-16(n,2n)0-15reactions. 0.1 can and- Measurements symmetric activity distributions d performed on similarly irradiated acrylic linearly and exponentially on the!TAM phantoms also indicated the presence of C-11 position, are considered. Dose rate profiles are and 0-15 radionuclides. calculated for thebeta-emittersF-32, Cu-67, The patientswere imaged on theV.A. Y-90, Ag-111,1-131, and Re-188,and for the Wadsworth Positron Emission Tomography conversion electron emitter Pt--.193m. Our results (PET) scanner. Theresultingimages Show that high energybeta-emitters(e.g. Y-90) are effective for large tumors 0:15 1cm) , medium displayed the activation left in the beam for smaller ones paths. Visual inspectionof the PET images emitters (e.g. Cu-67) revealed marked contrast between tissues of eecleln.1 cm), and low energy emitters (Ft193r) for very small tumors. differing elemental composition, due primarily to carbon concentration Work is supporbedbyUSPHS Grant No. CA- 32877. variations. Thus, acrylic phantoms were employed to image activation distributions without the irregularities causedby variations in tissue carbon concentration. The resultingactivationdistributions characterize the high energy component of the neutron kerma. This research is supported by NCI contract CM 973I:.

MPS1.3 MPS1.4 Energy Response of LiF TLD-100 to High Quantitative Magnetic Resonance Imaging of Energy Photon Beams, R. C. Robinson* Diffusion and T. H. Kirby, Radiation Physics I) Chien*, R B Buxton, K K Kwong, T J Brady, B R Dept.,UniversityofTexasSystem Cancer Center Rosen. Massachusetts General Hospital, Boston. Lithium fluoride thermoluminescent dosimeters (LiF- In magnetic r'sonance, the random translational mo- TLD) have been used for a number of years in the tion of water causes attenuation of the spin echo signal measurement of absorbed dose from high energy when magnetic field gradients are applied. This change photon beams used in radiation therapy. To convert from TLD response to dose requires a number of in signal intensity allows the calculation of the diffusion correctionfactors,including: A doselinearity coefficient D. correction factor, a factor to account for the fading We have implemented diffusion imaging on a 0.6 Tesla of the TLD, and a factor to account for the variance whole body imager and have tested the precision of the of TLD response with energy. technique in phantoms. MR imaging of diffusion was This paper presents a definition of TLD energy performed in 18 normal volunteers to study the accuracy response and measurements of that response for a and reproducibility of the quantitation in vivo, and to range of therapeutic energies (Co-60 to 18 MV). The determine the [nobility of water in normal grey matter resultant energy response curveiscompared to and white matter in the human brain. D was calculated theoretical values which are predicted by various cavitytheories(Bragg-Gray,Spencer-Attix,and on a pixel by pixel basis to generate diffusion maps of Bur lin).The practical benefit of this paper is to the brain. The measured D is 1.10 +/- 0.11 cm2/s for present well defined energy correction factors in the grey matter and 0.81 0.13 cm2/s for %%Ike matter clinical radiation therapy energy range for both (measured at a diffusion time of 114 Inset).In con- water and acrylic mediums. trast to unrestricted diffusion, the presence of barriers in tissne causes 1) to vary with diffusion time. This is This work supported by Grant Number CA 10953, awarded by the National Institute of demonstrated in the carrot, and in the human brain, Health, DHHS. showing the effects of restriction on the measured D. 29

1988 World Congress on Medical Physics and BiomedicalEngineeringSan Antonio, Texas 7 Medical Physics Scientific Papers

MPS1.6 A Generalized Method for Constant FeatureSup- STEREOTAXIC LOCALIZATION OF TARGETS IN THE BODY pression in Digital Angiography, DA Reimann*,JP Philip L. Pepmiller; Robert L. Siddon, andNorman Windham, Henry Ford Hospital, Detroit, Michigan48202. Barth, Department of Radiation Therapy, Joint Center for Radiation Therapy, 50 Binney Street, A method is presented which insures thatthe si m of the Boston, MA, 02115 weighting arguments of any temporal linear filteris zero, while preserving the significant aspects ofthe filter. By We have developed a generalized method forstereo- insuring that this stun is zero, constant. featuresas well as taxic localization of targets in the body.The noise which has an expected value of method relies on a two film reconstruction tech- zero is suppressed in nique which places no constraints onsource, tar- the final image. A new sequence is formedby subtracting get, localizer apparatus, or film locations. The the average image from each image in the originalsequence. target and certain parts of the localizer appara- This new sequence is used to define the signaturevectors tus must appear on both films.The method has of features that are to be enhanced and/orsuppressed. been implemented through the construction of two pieces of localizer apparatus, one utilizes These signatures are used in the filter definitionto create a a new linear filter which is then applied to the original large rectangular box into which the patient is placed, the other uses a smaller "U" shapedap- sequence. A mathematical derivation of the convolution of paratus which is placed adjacent to the patient. these two operations shows that thesum of the weighting Relative positions of objects can be determined arguments of the new filter is zero. The method is applied to within lmm. This procedure is useful for det- to the classical matched filter and the eigenimage filter. A exn.ming relative positions of objects suchas radioactive implants and hyperthermia thermometers comparison is made between the technical characteristics of the modified and the original filters.This method is illustrated on a cardiac study anda simulated sequence.

MPS1.7 MPS1.8 NMR Spectrometry of Radiation-Induced Lung Damage Verification of a 3-D Electron Pencil Beam C.H. Newcomb "and J. Van Dyk Algori for Irregu ar Fie s Dose Ontario Cancer Institute, Toronto, ONT. M4X 1K9 Calculations, Harunor Rashid*, Hanne M. Kooy, Roger K. Rice and Lee Chin, Joint Thereare indications thatMagneticResonance Center for Radiation Therapy, Harvard Imaging (MRI) may be sensitive to early physio- Medical School, Boston, MA 02115, USA. logical changes associated with radiation-induced lung damage. In MR images, much of the observed We have implemented a 3D non-coplanar electron contrast is the result of different Tl and T2 pencil beam algorithm based on the Hogstrom relaxation times between various tissues.It is gaussian pencil beam model. We present thechanges of relaxation times associated with measurements of doses and output factors in radiation which will determine thepotential water and special phantoms for a variety of usefulness of MRI for imaging lung damage. irregular fields. The electron fields consist MagneticResonance Spectrometry has been used in of stationary beams and rotational arcs. Arcs an in vitro study of Sprague-Dawley rat lungs. The include rotations performed over field cutouts Ti and T2 relaxation times have been determined fixed relative to the patient anatomy, providing for lung tissue asa function oftime after the ability to shape ,ne dose distribution during the rotation. irradiation. Single component fits have been made Measurements are compared with calculated doses and output factors using to the relaxation data whichcompare favourably our algorithm. The effect of heterogeneities with those determinedfromimaging data. In present in the volume is considered in the addition, multiple and continuousfits have also calculations.Results show good agreement been determined. between measured and calculated dose Significant changes have been observed at times distributions and output factors. The effects which correspond to those times ofknown of heterogeneities and field shape on the output physiological change. factor will be discussed. The results have confirmed the potential role for MRI toimage lung damage and provided the necessary data to optimize imaging parameters.

8 1988 World Congress on Medical Physics and BiomedicalEngineeringSan Antonio, Texas Medical Physics Scientific Papers

MPS1.9 MPS1.10 Computer Algorithms for Scatter Rejection A Ray Tracing Method for Calculating theSpeed with Multiple Fore Scanning Slits, C. Nayak of Sound in Layered Media UsingReflected and D. Plenkovich, Department of Radiology, Ultrasound, G. R. Lockwood ,F. S. Foster and Tulane Medical Center, New Orleans, LA, USA J. W. Hunt, Department of Medical Biophysics, University of Toronto, Toronto, Ontario, Canada. A highly efficient method for rejection of scattered x-ray radiation and veiling glare The speed of sound in a tissue could in many has been developed. Multiple fore slits are cases provide diagnostically useful information scanned across the x-ray beam. For each po- about the state of the tissue. Some success has sition of the slit assembly an image is de- beenachieved calculating the speedof sound tected with an image intensifier-TV camera from reflected signals for a single layered or system, digitized using a frame grabber and parallel layeredmedia, but no satisfactory stored on disk of a MicroVAX II computer. methods havebeen developed for in vivo The resultant image is reconstructed from a measurements of thespeed of sound in series of slit images using an appropriate non-parallel layered media. The signal received computeralgorithm. The peak-detection froma reflectedultrasoundpulse contains algorithmkeeps in each pixel the highest information about thedistribution of the speed value recorded duriag the scan. An advan- of sound along its path. Although the depth of tage of this algorithm is that it does not the reflector is usuallynot known, the average demand precise control of the slitmotion. speed of sound along the path can becalculated However, it requires overlap of the slits using triangulation provided that the source and increasing radiation dose to the patient. receiver are separated. By scanning the source, An alternative method for image reconstruc- a model of the speed of sound at the surface can tion is to threshold each of slit images by be created. Ray tracing through this model cutting off the component due to x-ray allows the calculation of the speed of sound for scatter and veiling glare. The resultant subsequent layers. A computer simulation in are image is obtained by summation of thresh- which images of the speed of sound olded images. This algorithm utilizes all reconstructed from data generated by ray tracing primary radiation, but requires very pre- through two models is presented. Ideas on how cise control of the slitmotion. Neither this methodcould be implementedare also algorithmrequires any synchronization of suggested. slit scanning and video camera readout.

MPS2.2 MPS2.3 EXPERT SYSTEMS IN MEDICINE: REALISTIC EXPECTATIONS Knowledge Based Systems for Oncology Lawrence E. Widman, Dept Med, UTHSC, SanAntonio, TX Nu la Saranummi, Simo Hyodynmaa, Pentti Kolari, Jukka Expert systems are large computer programswhich Rantanen (#) and Antti Ojala (*) Medical Engineering incorporate a well-defined, narrow, butdeep Laboratory, Technical Research Centre of Finland, knowledge of some field of human expertise. Uhen Information Processing Laboratory, Technical Research well constructed, they mimic excellentlythe Centre of Finland *Oncologic Clinic, Tampere University behavior of a human expert in diagnostic and Central Hospital therapeutic tasks. They differ from conventional programs in their capacity forsymbolic reasoning Decision support is required in clinical medicine both in and their explicit expert knowledgedatabase. making the diagnosis, planning and prescribing therapy and in managing the patient during therapy. It can be provided with This overview talk is intended for listeners new knowledge based on systems that are integrated with patient to the field of expert orknowledge-based systems. data. in oncology e.g. with the patient database, diagnostic It will review the basic principles ofthis image database and clinical register. In addition, a decision technology. the advantages and disadvantagesof the e.upport system must cooperate with the user when giving most commonly used methods forrepresentation and advice and suggestions not compete with him. Data input application of symbolic knowledge, and examplesof required from the users must also be minimized. successful expert systems developed in medical fields. Building on this introduction,.:he short- Our solution to these requirements, CARTES, builds on the and medium-range potential of thistechnology in results of the Nordic Computer Aided Radiotherapy medical applications will be discussed. programme (CART), where a data set standard and a network concept was developed to facilitate the exchange of data between radiotherapy modules. The CARTES workstation acts as one module in the CART network and comprises three knowledge based modules integrated with each other and the clinical register and protocol database. These modules support therapy selection, protocol design and outcome analysis. A prototype has been built of the 1st module.Its domain is therapy selection for the unoperable non-small cell lung carcinoma. This was done in Xerox-1186 AI workstation using Loops and Inter Lisp. Presently we are working with the outcome analysis module and the clinical register.

1988 World Congress on Medical Physics and Bicmedical EngineeringSan Antonio, Texas 9 Medical Physics Scientific Papers

MPS2.4 MPS2.5 The Role of Virtual World Systems in Medical The Diagnostic Radiology Workstation in Imaging in the 21st Century. the Year 2000 Henry A. Swett, Dept Diagnostic Prof. Henry Fuchs, University of'orth Carolina at Imaging, Yale Univ Sch Med, New Haven, CT Chapel Hill Evolving techniques in artifical intelligence, picture archiving and communication systems, and Inspired by Ivan Sutherland's 1965 vision of an "ul- medical information management systems are con- timate display", many groups are developing real- verging in a way that will lead to a powerful time 3D graphics systems with tiny head-mounted intelligent radiologic workstation. By the turn displays whose images are continuously modified to of the century, radiologists will no longer sit account for head position so that the user feels as in front of x-ray view boxes and they will no if he is moving inside the simulated environment. longer be limited to making diagnoses of entities These advanced systems allow immediate comprehension that they have seen before. Instead, they will of the data and direct hand manipulation of the sim- view images on an intelligent radiologic work- ulated objects.Medical imaging may be the great- station. This workstation will not only display est beneficiary of such systems which allow viewing and manipulate images, but it will provide of complex 3D anatomical data together with direct highly specific information that will enable the ealt-ing, such as the elimination of obscuring re- radiologist to function with a much higher and gion,. b., "erasing" them by hand, and allow easier broader level of expertise than previously understanding of the spatial relationship between possible. This information will include selected multiple objects. In radiation oncology, the user and annotated excerpts from patients electronic may view patient anatomy together with treatment medical record, artifical intelligence decision beam sources, and isodose contours. The systems support and analysis tools, and automatic retrie- may provide direct manipulation of design parame- val of pertinent selections from the medical ters such as location and orientation of radiation literature. sources and blocks. The realization of such systems presents daunting technical challenges. The system's three components, a real-time computer image generation system, a small head-mounted display unit, and a real-time head tracking system, each challenge the state of the art in engineering and computer science. Recent developments about each will be discussed.

MPS3.1 MPS3.2 Acoustic Output Levels and Performance Technical Factois and Clinical Uses of Standards for Modern Ultrasound Instruments, Pulsed Doppler for Studying the Fetus, J. Zagzebski *, University of Wisconsin- P. Burns*, Department of Radiology, Madison, Madison, WI, 53706, USA. Thomas Jefferson University Hospital, Philadelphia, PA, USA. The purpose of this symposium is to discuss and elucidate current issues regarding safety, perfor- Pulsed Doppler instruments are receiving consider- mance and standards in diagnostic ultrasound. This able attention at the present time, both in re- background should be relevant to the practicing gards to the rapid technological developments and medical physicist and bioengineer in evaluatingthe their expanding clinical uses.However, these several standards which are extant and to bere- devices currently produce the highest time average leased in the near future. acoustic output levels among diagnostic ultrasonic instruments, with time average intensities in some One of the difficulties facing the user when making systems exceeding 1-2 W/cm2.The purpose of this decisions regarding the choice of ultrasound in- presentation is to review recent technical develop- struments for specific applications is that of ments of pulsed Doppler devices and report on new obtaining data on the acoustic output and quantita- uses of these devices for studying the fetus. tive performance levels of these devices. One user Tradeoffs between pulsed Doppler performance and organization, the American Institute of Ultrasound acoustic output levels will also be addressed. in Medicine, has accumulated acoustic output data through a manufacturer's commendation process; and these cumulative data will be summarized. This presentation will also outline a standard being developed for measuring imaging performance of pulse echo scanners.

10 1988 World Congress on Medical Physics and Biomedical Engineering G San Antonio, Texas Medical Physics Scientific Papers

MPS3.3 MPS3.4 Problems in Characterizing Acoustic Concerns With Acoustic Exposures: Alterna- Emissions From Diagnostic Instruments: tive Safety Proposals for Medical Ultrasound The Manufacturer's Perspective, J. Abbott*, Equipment, P. Carson*, University of Michigan Diasonics, Inc., Milpitas, CA, 95035, USA. Hospital, Department of Radiology, Ann Arbor, MI, 48109-0030, USA. The multiplicity of operating modes and sophisticated electronic control of beam focusing patterns The concept of "derated intensity" was adopted to in modern ultrasound instruments presents a major allow meaningful comparisons between acoustic out- challenge when measuring acoustic output levels of put levels produced by modern ultrasound instru- ultrasound instruments as well as when reporting ments, which usually utilize high frequency trans- data to the user. Important features of a new ducers, and levels of older, lower frequency sys- standard, being developed by the Americ-i Institute tems. of Ultrasound in Medicine and the National Electri- The adequacy of "derating models" will be discussed cal Manufacturers Association, for specifying and an alternative, which may be particularly rele- acoustic output levels of diagnostic ultrasound vant for fetal exposures, will bq presented. This equipment will be described.The standard requires presentation will also address the topic of thermal measurements and intensity calculations in water, mechanisms for specifying potentially hazardous but also allows calculation of "derated" intensity exposures for diagnostic ultrasound. In addition, values assuming a uniform sound beam attenuation an ultrasound power control feature fordiagnostic path. The merits of both reporting systems will instruments will be recommended. be addressed.

MPS3.5 MPS3.6 Possibility of Cavitation Induced Bioeffects Second WFUMB Symposium on Safety and from Medical Ultrasound and Shock Wave Standardization of Ultrasound in Devices, B. Fowlkes*, University of Mississ- Obstetrics, G. Kossoff*,Ultrasonics ippi, Department of Physics, University, MS, Institute, Sydney, NSN 2000, 38677, USA. Australia.

One of the mechanisms by which the large pressure An overview of the ultrasound safety issues, par- amplitudes associated with diagnostic ultrasound ticularly as seen from the ultrasound community pulses could induce a biological effect is through outside the United States will be presented. cavitation. Recent experimental and theoretical These issues will be discussed in greater detail Medicine data for cavitation thresholds for diagnostic at a World Federation of Ultrasound in ultrasound will be presented. and Biology (WFUMB) meeting on Ultrasound Safety and Standards, scheduled for October 22-23, 1988, in Washington, D.C. The goals and anticipated outcome of this meeting will be discussed.

3 3

1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas 11 Medical Physics Scientific Papers

MP1.1 MP1.2 A Quality Assurance Protocol For The A Comparison of Treatment Techniques for Prostatic Testing Of Treatment Planning Systems, Cancer Irradiation.M. Sontag*, R. Whittington and E. U.F. Rosenow*, M.K. Islam, G. Mitev, Cheng. Radiation Oncology, University of Pennsylvania S. Rajaratnam, O. Cintron, T. Urrutia, 3400 Spruce St., Philadelphia, PA. 19104 Albert Einstein College Of Medicine and Montefiore Medical Center,Bronx, An analysis has been made of 30 treatment plans for 5 patients NY 10467 to determine the optimal treatment technique for prostatic cancer irradiation.15 MV x-rays were used. Techniques Quality Assurance of treatment planningsystems is examined included 4 field box, bi-lateral 110° arcs, 300° structured, in three different levels ofperfor- anterior arc and 360° rotation, each without and with blocks mance: (1) initial system tests, aimed ata most Simulation of blocked arc treatments with multi-leaf comprehensive test of total system performance; collimation were also performed. CT scans were taken at 1 (2) constancy checks whichare to be done at re- cm intervals throughout the irradiated volume and plans were gular intervals and comprise ofa subset of the generated using a'3-D radiation therapy treatment planning initial system checks; (3) intermediatesystem system. Isodose distributions, dose-volume histograms and tests, which are performed followingany hardware, dose statistics were made available to the physician for software or data base modifications.These con- treatment plan evaluation.Each plan was optimized to sist of any number of checks fromconstancy check maximize target dose homogeneity for a minimum target dose up to the full set of initial system tests depend- of 70 Gy. Target volume size varied from 87 to 163 cm3. ing on the nature of the modification. No correlation among target volume, bladder and rectal For reasons of practicability theset of test volume was found.Analysis of the dose to the femoral cases has to be rither limited. It consists of heads,bladder, and rectum indicates that a clinically acceptable various single beam:: in a water phantomincluding femoral head dose is obtained regardless of the treatment oblique incidence, wedges, blocks and simplein- approach. All blocked beam techniques result in a clinically homogeneities where comparisons betweencalcula- significant reduction of the bladder and rectal dose, tions and measurements can be made. Further in- particularly for smaller target volumes. The best results were cluded are opposing and matching beams,and a obtained with the 4 field box and bi-lateral 110° arc selection of patient plans. The latter are mainly techniques, each having comparable bladder and rectal doses. used for constancy checks.Results obtained with An error analysis indicated that both blocked and unblocked the quality assurance protocol withdifferent ver- bi- lateral arc techniques is somewhat more forgiving of patient sions of a treatment planning system willbe preset-up uncertainty and motion than the blocked 4 field box sented. technique.

MP1.3 MP1.4 Constraints in 3-D Treatment Planning Dynamic Wedge Field Techniques through SystemDesign, B.A. Fraass*, D.L. Computer Controlled Collimator Motion and McShan, and W.A. Roberts, University of Dose Rate,DennisD. Leavitt*, Melissa Michigan Medical Center, Ann Arbor, MI Martin+, and John H. Moeller, Univ. of Utah, 48109. Salt Lake City, Utah 84132, and St. Joseph Hospital+, Orange, CA 92667 A numberofgroups havereported on various features and uses of their 3-D treatment planning systems. Clinical treatment planning situations arise which Althoughdesign philosophy and aims require different wedge angles within segments of a have a large impact on each of these systems, the single field. way in whichvarious Solutions to these cases can be attained design constraints are through dynamic wedge field generation handled is often of crucial importance. We describe a number of basic constraints and design Dynamic wedge field generation is achieved by 1) considerations for specifying the desired doseat a seriesof points in the 3-D planningsystems and the plane of the wedge, under the constraint that the points variousways in which those issues have been addressed during the four years of development of may be at different depths but must be on different the 3-D planning system U-HPlan. fanlines (cannot specify dose at more than one depth on In a general 3- same fanline); 2) calculating the dose contribution to D system, a three-way trade-offdecision must often be each point due to collimator-defined asymmetric strip made betweeninteractivity, disk space fields having one contiguous edge while increasing in and accessspeeds, and computer resource usage. A second basic type of constraint is that imposed width across the field; 3) determining optimized field weightsforeachfield,whileaccountingforthe by systemhardware (whichcan besolved, or at least changed, dependence of output factor of each strip field on width by purchase of new equipment) and and radial distance. more difficult limits caused by operating system Calculation and measurement of dynamic wedge fields software design. A number of these limits have been reached and verification by film dosimetry are reported for the during thedevelopment ofour Varian Clinac 1800 and Clinac 2100-C. software, includinglimits on disk space, real and virtual memory, CPU usage, and multiple user environments. Majordesign changes, operating system reconfiguration, and new hardware have all been used to address the various limits. 34

12 1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas Medical Physics Scientific Papers

MP1.5 MP1.6 OSCAR - An objective approach to treat- Evaluation of Three-Dimensional Photon mentplanevaluation, s. Shalev*, Treatment Planning of the Intact Breast, D. Viggars, P. Therrien, P. Hahn, J. Chut L. Solin, M. Sontag, L. Brewster, Manitoba Cancer Treatment and Research E. Cheng, K. Doppke, R. Drzymala, M. Hunt, Foundation, Winniplg, Canada R3E OV9. R. Kuske, J. Manolis, B. McCormick, J. Munzenrider, Univ. of PA & FCCC, Phila., PA, Mem. Sloan-Kettering Cancer Center, NY, NY, An objective approach to treatment planning is Mass. Gen. Hosp., Boston, MA, Mallinckrodt described in which aplanned dose distribution is Inst. of Radiology, Wash. Univ., St. Louis, MO evaluated in terms of prescribed limits of acceptability, and any discrepancies (referred to as Early breast cancer can be effectively treated with "regions of regret") are displayed in the form of lumpectomy followed by radiation therapy. Uniform a parametric image inwhich colorsare used to dose distribution within the target volume is represent different types and degrees of regret. difficult to achieve due to the significant breast The program OSCAR (Objective Scoring with Coloured contour variations and the large dose purturbations Areas of Regret) is used in conjunction with the produced by the lung tissues inside the irradiated AECL Theraplan treatment planning system to dis- volume. Two intact breast patients have been play images of regret :n addition to conventional extensively planned and 39 three-dimensional treat- isodose plots, and is used for the comparison of ment plans were evaluated as part of a multi- alternative plans in terms of adequate target cov- institutional treatment planning project sponsored erage and minimal irradiation of sensitive organs. by NCI. Dose distributions, dose volume histo- Quantitative score functions are derived for tar- grams, and dose statistics were used to quantify get coverageand over-irradiation of normal tis- the dose to the target volume and eight normal sues, which provide a clear indication of the tissues. The physical parameters studied include: acceptabilityof the dose distribution in each effect of contour variation, effect of tissue organ. Required tumor dose levels and organ-spe- inhomogeneity correction, effect of photon energy, cific tolerancedoses are prepared inadvance by effect of dose modifiers, and the use of field the clinician for each site. The method is objec- shaping blocks. Our preliminary data indicate tive and reproducible, and provides clear documen- that the dose heterogeneity can be as much as 23% tation of the selection process. Examples are within the treated volume and it is important to shown for selection of onefrom several rival consider the entire irradiated volume when evaluat- plans for treatment of the prostate, esophagus and ing treatment plans for intact breasts. lung.

MP1.7 MP1.8 Dose Surface DisplayforThree THE EVALUATION OF A NEW 3-D SCATTER CORRECTION Dimensional TreatmentPlanning, ALGORITHM WITH ALREADY EXISTING METHODS. A. T. J.A. Purdy*, J.W. Wong,W.B. Harms, Redpath.and D. I. Thwiates, Dept. of Medical Physics, Western General Hospital, Edinburgh, B.Emami, R.E. Drzymala, and J.W. Matthews, WashingtonUniversity Scotland. School of Medicine, St.Louis, MO, USA. A number of methods have been developed to correct for the presence of inhomogenous structures when calculating absorbed dose distributions for radio- The evaluation of the volumetric three therapy. The two algorithms that take into dimensional (3D) dose distribution has account changes in the scattered component of proven to be a difficult and time consum- absorbed dose are the power law tissue air ratio ing task in 3D treatment planning. We have method and the equivalent tissue air ratio method developed code for our prototype 3D treat- (ETAR). Both algorithms have been implemented on ment planning system which allows the gen- a treatment planning system running on a SUN 3/180 eration of user specified isodose sur- microprocessor. A new algorithm based on the faces. Surfaces are represented as cross- principles of ETAR has been developed and has hatched contours, and allow one to view also been implemented on the treatment planning target volumes and/or internal structures system. which are enclosed or protruding from the isodose surface. While dose surface dis- All three algorithms have been assessed in a play for 3D treatment planning is not new, variety of situations designed to simulate those its implementation on a real time display occuring in radiotherapy practice, and their system is. Dose surface display in combi- performance compared to experimental results. nation with high speed interactivity has proven to be an extremely effective tool in plan evaluation. Examples will be shown demonstrating dose surface displays versus multi-planar 2D isodose distribution displays as a more effective way of evaluating 3D treatment plans. (Supported in part by NCI CM-47696)

1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas 13 Medical Physics Scientific Papers

MP1.9 MP2.1 Horizontal Tangential Technique and Dose Calculating AN ANALYSIS OF LEFT VENTRICULAR STRUCTURE BASED Methods for Breast Irradiation, Yi-min Hu*, Chun -ii ONREGIONAL WALLTHICKNESS AND Zhang, Zheng-gong Luan, and Shin Zhang, Radiation CURVATUREBY GATED CT, G. Tomonaga*, K. Iga, Oncology Dept., Cancer Institute (Hospital), Chinese T. Tamamura, H.Geh,T. Matsumura, K. Academy of Medical Sciences, Beijing, China. Hori, Tenri Hospital, Tenri, Nara 632, Japan A horizon Al-tangential three-field technique has Many cardiologists have assumed that wall been developed to solve effectively the technical stress remains within normal range in compensated hyper- problems and difficulties encountered by the con- trophic or normally growing hearts. To investigate ventional three-field technique used to treat breast this hypothesis, we attempted to measure regional cancer by radiation. A mathematical model is built up for Cobalt-60 and 8 MV X-ray. wall thickness and curvature of the left ventricle And from this in6 normal and model, one can easily choose a proper wedge filter 3 hypertensive subjects using ECG-gated computed tomography. (as tissue compensator) for individual shape of The longitudinal imageof theleft ventricle was target volume for a better dose distribution with reconstructed at end-diastole. From the tracings of outer an ununiformity of less than 10%.By using a gra- and inneredges ofthe ventricular wall,mid- vity-oriented trapezoid tangential field block es- wall line (MWL) was determined by measuring wall pecially designed, a proper matching of the fields thickness. Meridional radius (Rm) was measured between tangential and supraclavicularones can be sequentially by fitting circles on MWL. Minimum obtained. A special treatment couch for the tan- distance from each point on MWL to longitudinal gential fields and a water level buried-in slide- axis was measured as circumferential radius (Rc). pointer is made to improve theaccuracy locali- Mean value of the two reciprocals [(1/Rm + 1/Rc)/ zation and the repetition of settingup of patients. 2] was defined as mean mid-wall curvature. The relationship between regional wall thickness and curvatureshowedhyperbolic correlation in every patients, that was statistically significant (p<0.01 in8cases, p<0.05in 1case). This result suggested that the product et the regional wall thickness and curvature is constant. Accord- ingly, regional wall stress may be constant through:Jot the left ventricle.

MP2.2 MP2.3 RELIABILITY OFT AND To AS QUANTITATIVE PARAMETERS A Comparison of Selective Cancellation Excitation WITHIN LONGTERMCLINICALRESEARCH PROTOCOLS, Methods for Flow Measurement by NMR, R.C.Samaratunga*, S.R.Thomas, Dept. of Radiology Eduardo Benitez-Read,* Arvind Caprihan, and andW.Ball, Childrens HospitalMedical Center. Eiichi Fukushima, Lovelace Medical Foundation, University of Cincinnati, Cincinnati, OH 45267-0579 2425 Ridgecrest Drive SE, Albuquerque, NM, USA.

The recognized potential value of reproducible NMR flow measurementsinthepresenceof an quantitative NMR parameters such as calculated T 1 abundance of stationaryspinsrequirethatthesignal and T includes their practical clinical utility as 2 fromthestaticspinsbesuppressed. Cancellation indicators of tissue changes secondary to disease excitation either drives the static spinsto equilibrium or in relationship to treatment protocols. or inverts the magnetization to obtain the signal only Investtgations of whitematterabnormalities in from the moving spins, while the subtractive methods pediatric brains following bonemarrow transplant are based on subtraction of two data sets,in which are being conducted with the objective of defining the signals from thestaticspins arethe same while progressive or regressive -euroresponse on MRI. A thesignal from the moving spinsisdifferent. The major component of this st dy involves establishing cancellationexcitationmethodshaveanadvantage and maintaining a clinical calibration standard for over subtractive methods in that the dynamic range of longterm monitoring of (apparent) tissue T, and the A/D converter is more adequately utilized. T Preliminary data acquired over a period of two 2 . The method tobe analyzed consists of a pair of months on a 1.5 Tesla GE Siena unitutilizing a selective rf pulses with opposite phases and appropriate compartmental plexiglass phantomwith different sliceselectiongradients. Othervariationsofthe concentrations of CrC11(0.1 to 1.0 mM) indicate: a) pulsesequenceincludetheuseof oneor more r Long term reproducibilEy of 8% to 15%, b) T1 and To pulses for reducing the sensitivity to Bo and rf inhomo- position dependence of 8% and 23%. A newly designed geneities, phase cycling therfpulses, and alternating phantom consisting ofa matrixof sealedglass the polarity of gradient pulses.These variations will vials (of sterilized CrC1 solutions) in a loading 3 be compared for their sensitivity to instrument imper- medium is being used to evh-uate: a) Long term fections (Bo and rf inhomogeneities, and eddycurrent reproducibility, b) Optimum pulse protocol and c) effects), flowcharacteristics (mean velocity, and Measurment precision in conjunction with periodic laminar or plug flow), and pulse sequenceparameters verification of phantom T1 and T2 stability using (theratio of repetition time Tr to Ti, slicewidth, spectroscopic mesurements.Temperature dependence pulse angle, and frequency offset). of the calibration standard is also monitored,

14 1988 World Congress on Medical Physics and E::--)medical EngineeringSan Antonio, Texas Medical Physics Scientific Papers

MP2.4 MP2.5 X-ray Spectroscopic Imaging, JM Boone*, New Appoach to Quantiliying Cenebtat Blood JA Seibert+ and GS Shaber, Departments Ftow Mea4olomontA hy tho Xenon/CT Technique of Radiology, Thomas Jefferson University, Abund 0. Wistl Panos P. Fatouros, Philadelphia, PA and +University of Richard L. Keenan, and P.R.S. Kishore California-Davis, Sacramento, CA Medical College of Virginia, Richmond, VA

Dual energy image pairs are conventionally used With the recently developed Xe/CT method for to render bone or tissue subtracted images. measuring cerebral blood flow, high resolution X-ray Spectroscopic Imaging (XSI) is a technique blood flow maps are possible which contain a in which the spectral properties of the trans- large amount of information not readily access- mitted x-ray beam are overlaid using subtle ible. By using pattern recognition techniques color mapping, onto the original (raw) grey we areproducing composite imagescombining scale radiograph. Dual energyimages are ac- anatomical and hemodynamic data. Specifically, quired, and are numerically processed to gener- information from the partition coefficient ate the relative hardening index (RHI) image, ("lambda") mapsiscombined with blood flow which is a parametric map qualitatively related distribution in composite maps indicating nor- to the relative beam hardness.The RHI image is mal, ischemic or hyperemic areas. Further by colorized and overlaid on the raw grey scale drawingisoflow lines intothese maps the image. Therefore, the transmitted x-ray spec- physician can determine quickly the actual tral properties are essentially displayed using blood flow at a given point. To speed up the the visible spectrum. XSI can be used to dis- calculation of the blood flow maps (for each criminate between calcified and non-calcified pixel a two parametric non-linear regression nodules, and thus could be clinically useful in analysis has to be carried out), we developed a mammography, chest and abdominal radiology. A new high speed algorithm which can solve the brief theoretical introduction, computer simu- appropriateequations by about a magnitude lation results, and spectroscopic images will faster. We have incorporated several other im- be presented. provements designed to improve the accuracy and readibility of the information presented. The high diagnostic capability of this method will be demonstrated with a series of blood flow maps obtained from recent head injury patients.

HP2.6 MP2.7 In Vivo Flow Measurement by NMR Imaging. Individual and local lung density M.A. Smith; M. Tarnawskt, D.J. West, G.M. Taylor, variations observed by computed tomography V.T. Ayton, R.G. Gosling and M.N. Maisey. in rodent lungs after irradiation Division of Radiological Sciences, United Medical and E. El-Khatib*, S. Lehnertt Dental Schools of Guy's and St. Thomas's Hospitals, KFSH&RC, Riyadh 11211, Saudi Arabia Guy's Hospital, London. tMcGill University, Montreal, Canada Lung toxicity is recognized as a major limiting In vivo flow measurements have been performed using factor in radiotherapy of the thorax region. a 1.5T MR system (Philips Gyroscan) with a flow The factors associated with radiation damage to adjusted gradient (FLAG) pulse sequence.The lung are expected to produce an increase in the technique utilises gradient echoes and pairs of bipolar density of lung. Densitometry using computed gradient pulreF to encode velocity information in the tomography (CT) is a non-invasive method to phase of th t NMR signal.Two types of flow images measure tissue densities in-vivo and has been can be ac;, red:MR angiograms of thick (150mm) used to monitor radiation damage to lung. sections to remonstrate the presence of flow and Measuring an average lung density, however, is MR phasesaps of thin sections (5mm-10mm) to not a sensitive index of radiation damage since give quanti.ative measurements of velocity. radiation pneumonitis is a focal disease where Calibration was performed using a rotating flow areas of increased density are adjacent to areas phantom with a maximum velocity at the edge of of decreased density. 4m/s.The accuracy of velocity measurements was In the present study, the average lung density 5.6%. The maximum velocity detected through the was measured repeatedly in rats irradiated with sec' on was 2.6m/s and within the section was various single doses to the thorax. In addition, >4m/s. the density of small elements along the outer Results will be presented of flow measurements in parenchyma of the lung was recorded. Abnormal the aorta, head and legs in 30 normal subjects. MR regional densities and density gradients were angiograms of the leg have also been performed and observed at lower doses and earlier times than compared with X-ray angiography in 15 patients. changes in average lung density. Moreover, areas of very low density were observed close to denser areas. The regional density analysis thus proved to be a more sensitive index of radiation damage to the lung. 3 7 198.: World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas 15 Medicrii Physics Scientific Papers

MP2.8 MP2.9 A Method of Quantifying and Cor-Lecting for Sensitivities for Elemental Imaging thr.rffe-4--9 ofScatter, Veiling Glare and in Computerised X-ray Tomography, Beam Hardening in Quantitative Arteriographic K.J. McCarthy'and J. Fryar Imaging.I.A. Cunningham* and A. Fenster, National Institute for Higher Victoria Hospital and imacrina Research Education, Dublin, Ireland Laboratories, Roberts Research Inst, Unix of Western Ontario, CAMDA. There are two basic ways to determineanalyte concentration by using monochromaticX-ray Obtaining quantitative estimates of arterial lumen beams in computerised tomography. Thefirst narrowings (stenoses) From image data is made method is to scan the matrix, add theanalyte difficult by the effects of scatter, veiling glare, and scan again. The second method isto scan beam hardening and image noise. We describe a at two or more energies which straddlean ab- method of quantifying the influence each of these sorption edge of the analyte. The resulting effects has on the accuracy of stenosis severity reconstructions show the analyte concentration. measurements from a direct measure of the system The sensitivities of both methods have been modulation transfer function (Mi) . The Mil' is theoretically analysed for the situationwhere obtained for a particular set of conditions using the analyte is homogenously mixed witha cyli- an edge technique, which produces an accurate ndrical matrix. For the firstscan method, it measure of the low frequency drop, both with and is shown that the maximum sensitivityoccurs without a phantom to produce scatter. The effective for a matrix diameter of 2.5 mean freepaths. attenuation coefficient of the iodine contrast The sensitivity at any other diametercan be media, as would be estimated from attenuation readily determined by using a universalcurve measurements using image data, is modeled theoreti- of relative sensitivity versus thickness in cally as a function of beam path-length throughthe mean free paths. Sensitivities are shown to iodine.This model is verified experimentally, and be approximately 100 times better thanwhen is used to determine the raative importance of the analyte is added to the equivalentof one scatter, glare and beam hardening in determining voxel. In the second experimental method itis stenosis severity for a range of conditions. It is shown that the same universalcurve as for the shown this approach can be used to correct image first method also applies to the determination data for these effects with images of iodinated of the sensitivity at any diameter when that step wedge phantoms. at one diameter is known. The theoreticalpre- dictions are compared with experimentalmeasur- ements.

MP3.1 MP3.2 Holopraphic Displayof X -rayAbsorptioninMedical RELATION BETWEEN REFLECTANCE IMAGE Specimens, M, J.Flynn*. Roebuct_t. Lacey. R. Block, AND LESION FORMATION ON SIMULATED Henry FordHospital and Holo-source. Inc.,Detroit, EYE MODEL _Michigan, Y.Yang*, A.J.Welch, H.G.Rylander III,W.S.Weinberg A method based on radiographic measurements has been University of Texas at Austin developed to depict the Interior structures of a specimen witha white light hologram. The method has been applied to display One of the goals of our laser robotics work at UT Austin is the complex topography of blood vessels within the brain. to control the size of a lesion using reflectance feedback for Specimens are placed In a cylindrical chamber ani rotated In .5 ophthalmic treatment. A simulated eye model has been degree increments. Digital radiographs (1024') with water created that is composed of three layers. A thin absorptive equalization are obtained at each angular position. Each layer to simulate the pigment epithelium is placed betweer. radiograph Isrecorded with 2x magnification using a .3 two thick transparent layers containing protein.The millimeter focal spot to achieve a resolution of 5Ip/mm. The pigmented layer is heated by Argon laser irradiation and a digitalradiographs are computer processedtoremove video camera with a 32*32 pixel array operating at 420 nonuniformities and enhance structures of Interest. A total of frames per second records the reflectance images (in ital 120 radiographs covering a 60 degree angular field of view are time) owing to coagulation of the protein. Simultaneously transferred to a highesolution, black/white cine film. The cross section temperatures of some lesions are recorded with sequence of cine views Is reprojected through an Image a thermal camera. Different protel. and absorptive materials enlarger onto a master plate using an argon laser mountedon are chosen to indicate the effects of absorption and an Interferometrically stable optical bench. The hologram is coagulation rate constants. Formed lesions are examined produced from a single laser projection of the master plate onto under a microscope and the extent of damage is compared to a second photographic plate. Stereoscopic perception of the predictions with a computer model. specimenIndifferentorientationsIsachieved when the holographic plate Is viewed using a white fight point source. This work was supported in part by the Free Electron Laser Biomedical/Materials Science Program: ONR Contract number N00014-86-K-0875.

16 1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas Medical Physics Scientific Papers

14'3.3 MP3.4 Comparative Analysis of Laser Ablation of Plaque Using Direct Radiation Safety and Quality Control Lam fr--r: and rt Me.tal Conrac.r Probe Proaram for Medical Lasers, Suresh M. Brahmavar*, David Pleet, Cathy Miller GL LeCarpentiert, S Rastegar+, AJ Welch, and Marilyn Riddle. Baystate Medical SA Prahl, and H Hussein++ Center, Inc., Springfield, MA01199

Biomedical Engineering Program, The Radiation Safety and Quality Control (RS/QC) University of Texas at Austin program for medical lasers was developed at Bay- +Bioengineering Program, Texas A&M University s,ate Medical Center in compliance with DHHS/CDRH/ ++Trimedyne Inc., Santa Ana, CA FDA and DPH requirements. The RS /QC procedures are based on the recommendations of ANSI Z136.1, The purpose of this research was to establish the theoretical ANSI 2136.3 and AMA Council report. The RS /QC temperature distribution and ablationcharacteristicsin program manual includes five (5) distinct groups atherosclerotic plaque resulting from both direct Argon laser of procedures for use of medical lasers. The pro- irradiation and the application of a metal contact probe. The plaque cedures include regulatory compliance, laser safe- was assumed to be a single-layered homogeneous medium, and the laser light distribution within the the tissue was calculated using a ty, quality control standards for approval of delta-Eddington approximation. One dimensional temperature laser user, service and r,intenance.The details distributions and ablation depths were calculated using an included in above procedures developed for clini- immobilized finite element method. Thermal and optical properties cal lasers relate to laser environment, patient of the tissue were assumed to remain constant. Ablation front and personnel safety, electrical safety, smoke positions and temperature profiles at times prior to ablation, at the plume, fire safe*y, institutional safety policies, onset of ablation, and during the ablation process werecalculated compressed gases, medal surveillance of staff, for both modalities. Additionally, ablation velocities were LSO responsibilities, MPE for skin/eye, nominal calculated as a function of both tissue water content and heat flux hazard zones, pre-op/post-op check lists, daily transmitted through the metal probe tip. For the typical operating start-up and user log, administrative and engi- conditions used, calculated values agreed qualitatively with experimental results. Temperature calculations after the onset of neering controls, educational and practical experi- ablation indicated high sub-surface temperature rises beyond the ence standards, periodic QC tests, output measure- surface ablation threshold temperature for the direct irradiation caw ments, testing of protective devices, records and butmonotonically decreasing profiles for the metal probe. audits. The clinical lasers at Saystate Medical However, deeper within the tissue, calculated temperatures Center include CO2, Argon, Nd-Yag in OR, Ob-Gyn appeared to be higher for the probe than for direct irradiation Service, Immunology Labs and in Eye Surgery.

MP3.5 MP3.6 Infrared Laser Heating of Thermoluminescent Measurement of the concentration of Plates for the Purpose of Spatially and hematoporphyrin in different tissues Dosimetricallv Characterizing Radiotherapy by opto-acoustic spectroscopy, Beams, M. Grupen*and K. Kearfott, Arizona U. Bernini, A. Andreoni, E. Quarto, State University, Tempe, AZ 85287-6006. P. Russo*, A. Mancini, M. Mastrocinque, of Naples, Naples, Italy. Laser heating of a Thermoluminescent (TL) plate may University provide a method of extracting spatially encoded Hematoporphyrin (Hp) is a known photosensitizing dosimetric data. Two-dimensional collocation of agent, used in the treatment of tumors by the digital data extracted will provide a dose activation with lasers, since it absorbs in the distribution map or image. This information may be red and accumulates into tumors more than into valuable for quality assurance of radiation- normal surrounding tissues. We measured the Hp producing equipment or for obtaining beam isodose concentration in different organs of lab micas, curves which are prerequisites for radiotherapy patient treatment planning. To develop and that received the drug 24 h in advance, without optimize a laser heating device, time variant extracting it, by opto-acoustic spectroscopy. As spatial temperature distributions within the TL the technique is less noisy with powders, detector must be known. A numerical model (solved samples of known weight were lyophilized and the using the Alternating Direction Implicit Scheme) signal from the powder were compared in has been developed to determine radial and axial temperature profiles in a 71. layer when a 4 Watt amplitude to the signal obtained from a mixture carbon dioxide Gaussian laser beam is used to heat of Hp with BaSO4 (having a uniform absorption in a single or a series of spots. The results for LiF the spectral range of interest). Although the plates reveal heating rates on the order of 350 K/s absorption of Hp almost overlaps that of to 2000 K/s for plates 0.5 mm to 0.1 mm thick, hemoglobin, we proved itis possible to measure respectively. Laser heating of thicker TL plates the drug levels in different tissues following becomes impractical due to the limited depth of infrared radiation penetration within the material I.V.injection with drug doses of the order of layer. Furthermore, it is revealed that spatial the therapeutic dose used for phototherapy with resolution of such a system is a stronger function lasers on tumor models implantableinto the of laser beam diameter than radial conduction. mice. 3 9 1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas 17 Medical Physics Scientific Papers

MP3.7 MP3.8 HUMAN ERYTHROCYTE'S MEMBRANE DAMAGE DUE TO LASER AND PHOTOSENSITIZERS. S. Ganesan*, V. Velmurugan, Comparison of surface temperatures and damage volume incanine bladders duringinvivoirradiation at 1.06 gm wavelength. V. Masilamani and C. Prasad nppt of Physics, Anna University, Madras, India. W.F. Cheong*, AJ. Welch, S. Thomsen*, A.C.von Eschcnbach*, J.N. Torres, R. Swanson* The use of lasers in Photo Radiation Therapy (PRT) has assumed great importance because of its speci- Biomedical Engineering, The University of Texasat Austin; ficity and its capability of regressing tumors +The University of Texas System Cancer Center, even in internal organs. The characterisation of Houston, Texas. this PRT is still incomplete. In this connection photohemolysis of human erythrocytes are studied The efficacy of cw light at 1.06 run in creating thermaldamage inin as a simple model system using five different vivobladder tissue were assessed basedon peak mucosal and lasers and three photosensitizers - Acridine serosal temperatures, and histology of denaturationpatterns. Two Orange, Fluorescein and Rose Bengal. The percent physical conditions were examined: (1) air-tissue-air boundaries Hemolysis (pH) in the presence of laser and sensi- where exposed canine bladders, suspended with minimum tizer was found to be higher when compared to hemorrhaging, were irradiated; and (2) water-tissue-air interfacesin laser alone. The pH was found to increase a water-filled bladder under cystoscopic irradiation. Orly the serosal towards shorter wavelengths. It was also found spatial and temporal temperatures were monitored incase 2. Thermal that the pH depends on post incubation time, in- images were recorded with 8-12 p.m and 3-5gm thermal cameras. In cystoscopic irradiations at 7 s and 4 s using dicating the photochemical nature of the damage powers greater than 30 W and 40 W respectively, serosal charring aid collapseof ulood due to delayed photo oxidation via the cross vessels were observed with only simultaneous irrense linking between biomolecules. cpagulation at Also, of the three the mucosal surface. Subsequent correlation analyseswill determine sensitizers, Acridine Orange has the higher effect whether surface temperature can be used for prediction ofthe in photo-modification of cells. volume of thermal denaturation.

lIP3. 9 MP3.10 Tissue Effects of Neodymium-YAG Laser Fiberoptic Neuroendoscopy and Nd-Yag Irradiation of Pig Ureter, R.A. Siwek*, Laser Interfaced to Imaging-Guided R. Kodama, J. Srigley and L.Klotz, Stereotaxis-Henry Ford Hospital Det., III Toronto-Bayview Regional Cancer Centre and Zamorano Chavantes C, Dujovny M, Sunnybrook Medical Centre, Toronto, Canada Malik, G, Ausman J I.

Neodymium -YAG las'nx radiation at 1060 nin produces Imaging-guided stereotaxis constitute the most ac- a deep homogeneous coagulation in tissue.This curate method to approach intracranial lesionsMith is due to its low absorption and high scattering 3D/Multiplanar Imaging, size, shape, axis of lesions properties. The Nd-YAG laser has been can be considered to select the best approach for successfully applied endoscopically to treat diagnostic and therapeutic purposes. This informa- bladder neoplasms. In select patients, the tion can be transposed into stereotactic space with treatment of well differentiated tumours in the high accuracy and reliability. Nevertheless major ureter would lend themselves to laser therapy. critics are lack of direct visualization and intra- However, the wall of the ureter is much thinner operative monitoring,changes in intracranialcoor- than the bladder and a reduction of laser energy Jinates after aspiration of cystic cavitiesor in- would be necessary. No safe dosage guidelines traventricular lesions and difficulties on aspir- have yet been established. ation of high dense contents like hematoma, tumor We report our preliminary data of Nd-YAG therapy etc. New adaptation for direct visualization of on the normal pig ureter. This was an acute stereotaxic procedures has been designed that in- study on laparotomised pig ureter. Each ureter clude:rigid and flexible endoscope with a 2mm workwas opened longitudinally.Using a specially ing channel each, specially designed instrumenta- designed laser fibre holding device, the spot tion and Nd-Yag laser fiber.Methodology, experimen- size was kept constant. Various doses of laser tal and clinical experience will be presented. In- radiation were delivered.Heat measurements at terfacing of Imaging-guided stereotaxis withneuro- the ped-adventitial surface of the ureter were endoscopy not only changes enintraoperative blind measured with microthermocouples.Ureteral procedure into a direct visual one, but also in- sections were examined histogically for the creases the scope of use of these procedures in the effect of the laser radiation. Our results management of intracranial lesions. It allows to correlating radiant exposure and tissue explore ventricles, cysts, cavities extensively temperature are reported, together with heat with minimal damage to surrounding brain tissue: distribution and histology. drainage? biopsies, and resection of lesions under direct visual monitoring became possible. 40 18 1988 World Congress on Medical Physics and BiomedicalEngineeringSan Antonio, Texas Medical Physics Scientific Papers

MP4.1 MP4.2 MEASUREMENT OF THE COUNT RATE PERFORMANCE Scatter from Distributed Sources Within a OF THE MULTIWIRE GAMMA CAMERA BY DECAYING Uniform Scattering Media. MS Rosenthal*, SOURCE WITH 9.3 MINUTE Ta-178. R. Adams,' and LA Henry, University of Pittsburgh, J.L. Lacy, M.E. Ball, and L.J. Martin, Pittsburgh, PA, 15261, USA Loma Linda University, Loma Linda, CA and Baylor College of Medicine, Houston, TX. Scattering is a major source of quantitation loss in nuclear medicine. It is also difficult to cor- The multiwire gamma camera has unusually rect for. We have investigated scattering in uni- good performance at high count rates. This form media with either distributed or multiple study was designed to evaluate this pro- point sources. We performed Monte Carlo simula- perty by a decaying source method. Data from a Ta-178 source in the NEMA tions of this scattering problem using a 22 cm diameter cylindrical acrylic shell phantom filled scatter phantom were acquired in 1 minute with water. The code modeled the transport of frames. A FORTRAN program corrects the net cPs values for deadtime loss by a photons into a large field of view gamma camera trial deadtime CT) and converts to their (Tc-99m, 3/8" NaI crystal, LEAP collimator). Tested were various sizes of ring sources, as well natural logarithms. A lnd order polynomial as multiple point sources plus a combination of is fitted to the data: In R = A + BT + CT2 the two. In addition, experimental measu,ements where R is the net corrected cps, and T is were made of the same configurations as calculated the elapsed time.''(' is adjusted and data by the transport code. We took into account corrections and curve fittings are re- Compton scatter, coherent scatter, diffraction, peated until C approaches zero, to yield photoelectric effect, and production of lead x- the deadtime ("r), measured half life, and rays. Statistics generated were 500 to 1000K the input versus observed count rate curve. events detected by the camera, both my calculation The paralyzing deadtime measured 0.41 and experiment. Psec, and the half life, 9.19 minutes, only Results of this modeling lemonstrate that the 1.40 low. The validity is also suPPorted Monte Carlo simulations leproduce the experimental by close agree lent between fitted and data relatively well. For the most part, the sim- corrected cPs values. The camera can be ulations agree within statistics with the experi- operated up to 840,000 cPs with Ta-178 in ment. first-pass studies without exceeding 50% data loss, 2-3 times greater cps than any other commercial NaI device.

MP4.3 MP4.4 The Use of a Restoration Filter to Improve the NUCLEAR DECAY BY INNER ATOMIC SHELL ;ONIZATTON: Quantification of 1-131 Scintigraphs, M.T. Madsen, BIOLOGICAL EFFECTS IN VIVO, D.I.7..Rao , R.W. Howell, University of Medicine & Dentistry Thomas Jefferson University, Philadelphia PA. of New 4ersey, Newark, NJ; and K.S.R. Sastry, University ofMassachusetts,Amherst, MA. Most of the radionuclides used in Nuclear The quantification of activity from scintigraphs Medicine decaybyinner atomic shell ionization of 1-131 labeled pharmaceuticals presents special processes (electron capture and/or internal problems because of scatter and the large amount cormersion). The numerous low energy Auger electrons time follow the decay event result in a of septal penetration associated with the collima- highly localizeddepositionof . When tors routinely used for this radioisotope.Be- these Auger emitters are incorpora into cause of tais, the counts associated with 1-131 mammalian cells,the biological effects may be severely underestimated by currently adopted sources are very dependent on both the location dosimetric methods, which considertheam energy and size of the selected regions (ROIs) drawn electrons as trivial. Our experiments over the past several years about the source images. The efficacy of a rest- have repeatedly Shown that -Auger cascades do oration filter applied to the 1-131 images to cause severe biological damagein vivo. The circumvent this problem was evaluated for a vari- targets in these studies are the highly y ety of source distributions. The restoration radiosepsitive spermatogonia in mouse testis nd the primary oocytes in the mouse ovary. The filter was derived from an image of an 1-131 point biological end points are survival of source located on 10 cm of polystyrene. Test (primary oocytes) and induction of abnormalabnarmal sperm-Head shapes. Our conclusions are: images with both point and disk sources of 1-131 (1) Auger-emitters (F1=201, In-111, Fe-55, etc.) were acquired with 0,5, 10 and 15 cm of poly- are much more effective than their beta-emitting styrene. The counts in two different sized analogs, when injectedinthesamechemical form. (2) The sane, Auger-emitter In-111 as oxine concentric ROIs about the 1-131 source images were is far more effective than as a citrate because compared for both the raw and restored images. very different subcellular distributions in the The counts in the concentric ROIs varied in the two chemical forms. Hence, subcellular distribution depending on the chemical forms of raw images by 20-40% while those in the restored the Auger emitter is an important determinant of images varied by less than 5%. These results the effects. (3) conventional dosimetry is suggest that the the application of a restoration inadequate, and cellular dosimetry is essential to predict the effects of internal Auger filter to 1-131 images will significantly improve emitters. the estimates of activity. Work is supported by USPHS Grant No. CA-32877. 41

1988 World Congress on Medical Physics and Biomedical Engineering 0 San Antonio, Texas 19 Medical Physics Scientific Papers

MP4.5 MP4.6 Eigenimage Filtering of Nuclear Image Sensitivity and Applicability of Quality Sequences of the Liver. W.J. Potvinx, Control Indices for Gamma Cameras, A.B. Abche, J.B. Farison, L.R. Low, A. Hughes* and P. F. Sharp, Universityof J.P. Windham, Medical College of Ohio, Aberdeen, Aberdeen, Scotland, AB9 22D, Department of Radiology, Toledo, Ohio 43699 USA. Much of the work carried outon quality control of gamma cameras has concentrated on definingways in The effectiveness of the eigenimage filtering which measurementscan be made. Very little work technique as applied to human liver c namic has been done into the practicalvalue of these nuclear studies has been investigated. iiocess- indices particularly into two majorproblems; ing was performed on sequential digital images firstly, how sensitive the indicesare both to obtained during first transit of technetium-99m changes in device performance andthe conditions labeled sulfur colloid. Temporal signature under which they are measured and,secondly, how vectors were obtained from regions of interest different indices interact in determiningimage established on the aorta, vena cava and parenchy- quality. mal liver tissue.These vectors were used to Studies of the sensitivity of variousmeasures of generate signature matrices corresponding to gamma camera uniformity were carried out. It was image features designated as desired or interfer- shown that sensitivity of uniformityindices to ing. For each choice of a desired of interfering the presence of a localisedarea of non-uniformity process, an associated weighting vector is used increased with count density, beingmost sensitive to produce a composite eigenimage in which a for images containing10 000 counts/pixel.Images feature of interest is enhanced while suppressing acquired according to the NEMASpecifications were undesired processes. less sensitive to changes in detectorresponse. The uniformity indices whichwere the most sensi- The results of this processing provide images in tive were those calculated usingall the pixels in which features demonstrating arterial blood flow the image contrary to the NEMA recommendations. can be clearly separated from structures asso- The interaction between the spatialresolution and ciated with the venous system. Applications to sensitivity of a collimator has beenrecognised as the liver's arterial and portal blood supplies a problem for many years. Our work demonstrates and vascular and avascular pathological struc- how the perceived sharpness ofan image depends tures are presented. upon image count density.. The effectiveness of varxous figures.of merit in.combining.the effects ot both resoiution and sensitivity willbediscussed.

MP4.7 MP4.8 ON THE CODED APERTURE IMAGERS WITH PAS- Pediatric Total Body Potassium Counter SIVE AND ACTIVE ELEMENTS, K V Ettinger, Kenneth J. Ellis*, USDA/ARS Children's Nutrition II University of Rome and J Brondo*, Research Center, Department of Pediatrics, Baylor Scientific Innovations Inc.,Wainscott College of Medicine, Houston, Texas 77030. NY, USA The direct in vivo measurement of total body potassium Coded aperture systems of imaging are capable of (TBK) can provide information concerning the body's lean imaT.ng gamma rays with energies which are un- tissue mass. Although many nutritional studies have been suitable for conventional multi-hole collimators performed in children, few have measured lean tissue An efficient system of coded aperture imaging, mass directly. The noninvasive nature of the TBK using uniformly redundant arrays is described as measurement makes it ideal for pediatric use. a basic building block for assembling a scanner, Longitudinal studies in infants, however, are difficult be- a gamma camera or a tomographic imager. The cause of changes in the counting geometry caused by the principle of redundant arrays is to ensure that rapid growth of the infant. To overcome this restriction, the obscuration of detectors from radiation we have designed a whole-body counter that is invariant sources being mapped, is such that the signals to body size from the premature infant (36 cm, <2 kg) to can be uniquely unscrambled. Modulation is per- the 3-yr-old toddler (100 cm, <17 kg). The counting formed by switching between two masks,one being system consists of eight NaI(T1) detectors (7.6 cm x a complement of another. The system can be im- 7.6 cm x 100 cm) positioned lengthwise along the child. plemented either with scintillation detectors The cross-sectional geometry of the counter can be varied (BGO or NaI) or with germanium solid state de- to match that of the child. The signals from the two tectors. Discussion includes comparisons between photomultiplier tubes at each end of the 100-cm length circular and rectangular masks as well as incor- are operated in multi-parameter mode. Custom software poration of coincidence technique into the coded has been written for an ND9900 series analyzer thatcom- aperture design. It is worthwhile to note that bines the two signals to give energy and spatial informa- coded aperture systems can be used to image fast tion per measurement. The average resolutionsare neutron distribution,albeit with different de- ±8.5% full width half maximum at 662 keV for energy tectors than for gamma rays. and ±2.5 cm for position. For a 1000-sec counting time, the average error is less than ±4% for TBK measurements in :dents.

42 20 1988 World Congress on Medical Physics and Biomedical Engineeringo San Antonio, Texas Medical Physics Scientific Papers

MP4.9 MP5.1 Development of a Knowledge Based System for the Monte Carlo calculations of tissue dose Radionuclide Renogram, to determine stochastic and nonstochastic P.J.Vassilakos*, Ch.Pierrakeas. N. Pallikarakis. risks in Radiology. G. Williamsw, Dpt. ofNuclear Medicineand Dpt.of Medical W.J. Potvin, G. Drexlert, R.A. Brinker, Physics, University of Patras,Greece. Medical College of Ohio, Department of Radiology, Toledo, Ohio 43699 USA. The Radionuclide Renogram (RR) using 95mTc- DTPA, has been establishedduring the last decade as a basic diagnostic tool inNephrology In certain tissues such as bone marrow and and Urology. Today,the efforts are oriented periosteum, realistic dose calculations and towards a quantitative analysis of the measurements are difficult to make. The parameters expressed by the Renogram curve and determination of the risks of leukemia and their relation to the pathophysiologyof the osteosarcoma induction for children and young kidneys. Such parameters are, the time and slope adults is important. Also there are many other to peak activity, which expresses the ability of radio-sensitive tissues in the body at risk. In the filtrationofrenal parenchyma as well as diagnostic and therapeutic radiology significant the time for half peak activity which is related doses can be deposited in such tissues. A Monte to the excretory ability of the kidneys. Carlo method coupled to realistic 3D CT models The analysis and processing of these has been used to calculate tissue doses. The parameters may lead to the development of a results of the method to determine doses in Knowledge Based System (KBS) for computer as- radiology will be given. Also the latest data sisted diagnosis of renal diseases. from this work, published in ICRP 44 and ICRP So far 14 such parameters have been identified 51, will be discussed and the work of a new ICRU as quantitative indicators for the characteriza- Report Committee which is studying experimental tion of the renogram curve using data from 20 and theoretical phantoms in radiology, radio- cases from our electronic archive. According to therapy and radiation safety, will be described. the results obtained we can conclude that we are able to built a KBS for diagnosis of renal func- tISS, GSF, Ingolstaedter Landstrasse 1, tional abnormalities. D8042 Neuherberg, West Germany

MP5.2 MP5.3 Convolution Monte Carlo Dose Calculation in a Is a Strict Computation of Three-Dimensional CT Image Applicable to the forMegavoltage C. Chui* and R. Mohan Dose Distributions Field and J.J. Memorial Sloan-Kettering Cancer Center Photon Beams ? G.C. Institute New York City, New York, U.S.A. Battista*, CrossCancer Edmonton, CANADA The Monte Carlo code EGS4 has been employed to simulate radiation transport through a three dimensional CT image to Two kernel-based algorithms are calculate dose distributions for both electrons and photons. usedtoproduce dose distributions Since dose measurements in arbitrary geometries are often for Cobalt-60, and 15 MV X radiation difficult if not impossible, especially in regions of electronic beams incident on a water medium. disequilibrium, Monte Carlo simulation remains the only We first review the convolutionand as viable alternative to produce accurate results. In principle, all superposition integrals physical processes can be included in such simulations. descendants of the general dose Calculations were first performed in a water phantom and equation, and identify major other simple heterogeneous geometries for which reliable assumptions adopted by various measurements can be made to ensure the validity of this investigators. Dose distributions method. Calculations were then performed on Rando are then computed with a VAX-11/780 phantom and actual patient CT images employing arbitrarily computer in both the real and oriented and shaped beams in 3D treatment plans. Since Fourier domains and are compared Monte Carlo simulations require aerge amount of computing withdoses measured experimentally. time (up to weeks on a VAX-8550), it cannot be used The impact of different routinely. However, they may serve as benchmarks against implementations on computational which other approximate methods may be compared to speed is measured while the effects estimate uncertainty distributions of those methods. We on dose accuracy are visualized with present the simulation method, results and describe some colour maps of dose differences. We variance reduction techniques that may be used to increase conclude with recommendations on the the efficiency of simulations. practical usefulness of the strict (fast) convolution and the less restrictive (slower) superposition algorithms. 4 3

1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas 21 Medical Physics Scientific Papers

MP5.4 MP5.5 Dose Perturbations at Interfaces in Photon Convolution Isodose Distributions of X-Ray Fields Beams: the partial fluence model. Derived from MonteCarlo Spectra. A.L. Boyer*, Y.P. Zhu, F. Pascal, L. Wang, Department of Rad- Barry L. Werner, Ph.D!' IndraJ. Das, M.S. iation Physics, The University ofTexas System Department of Therapeutic Radiology Box 494 CancerCenter, M.D. Anderson Hospital and Tumor University of Minnesota Hospital and Clinic Institute, Houston, Texas 77030. Harvard Street at East River Road Minneapolis, MN 55455 Isodosedistributions of unwedged and wedged x-ray fields weremodeledusing convolutions An improved version of the partial fluence employingthe fast Fourier transform. The dist- model (Med Phys 14,585, 1987) for photon beam ributionswerecalculated on a 64x16x64 array dose distributions at medium interfaces,is pre- representing voxels 0.5cm x 2.0cm x 0.5cm. The sented. The original model explains how photon convolution kernels were based onMackie's Monte beam dose distributions in soft tissue surrounded Carlocalculations. TheEGS-4 Monte Carlo code by bone depends on the beam energy, thickness of was also used to simulate the production ofx-rays the bone, and orientation of the interfaces with in a Siemens linac and to transport the radiation. respect to the beam. Although the original model through a complex beam modifyinggeometry. Wedged predicts approximate limits between which dose fields were modeled by ray-tracingthrough the perturbation can be expected to be found, these physical wedge. 6 MV and 18 MV beam spectrawere limits are not narrow enough to calculate dose calculated as the weightedsumof 2,5 and 10 perturbations for clinical treatment planning componentsand comparedwith measureddistribu- purposes. The present, improved version of the tions. Small changes in the spectra attributable model describes how calculations of the actual to scattering from the collimatorswere determined dose perturbation can be made for photon beams of by fitting to measured depth dosedata. When five convolutions any energy at interfaces between materials of any were used, the accuracy was on the composition. Dose perturbation predictions made order of 1% and thecalculation time was on the with the improved version of model are compared order of 20 min usinga VAX/750 computer. A technique with measurements made in layered phantoms con- was expl,red to reduce the number of sisting of polystyrene and materials of atomic convolutions while maintaining goodaccuracy. number ranging from boneto lead, forphoton beams ranging in energy from Co-60 to 24 MV. This investigation ws supported by PHS Grant CA43840 awarded by the 0.C.I., DHHS.

MP5.6 MP5.' Comparison of Algorithms forSolving Angular Distributionof Thick Target Bremsstrahlung Inverse Problems in Radiotherapy Photons, R. Mohan; G. Miklos, C. Chui, Memorial Sloan Bengt K. Lind, Department of Radiation Kettering Cancer Center, New York, USA and A. Bielajew, Physics Karolinska Institute, P.O. National Research Council of Canada, Ottawa, Canada. Box 60204 S-104 01STOCKHOLM, SWEDEN Radiotherapy treatment planning Angular distribution of bremsstrahlung photons emitted of today from the target of a medical linear accelerator affects the is mainly focused around what iscalled theforwardproblem inradiotherapy, photon energy fluence and spectrum as a function of radial i.e. calculation of the resultant dose diFt.cince from the central ray as well as the magnitude of distribution in the patient froma given photon angular spread resulting from scattering from the treatmentsetup, including beam modali- flattening filter and the collimating system. The variation ty, energy, beamangles etc. Asthe of energy fluence and spectrum with radial distance goal of radiationtreatment is to determines the magnitude of "horns". The angular spread deliver a specificdose to a target determinesthevariation of output with field size and volume, withoutdamaging surrounding affects the sharpness of the beam boundary. Energy spectra healthy tissue or organs atrisk, some and angular spread may be calculated by Monte Carlo sort of optimization has to bedone. simulation of treatment machine heads and the EGS Monte Carlo code has been employed widely in medical physics for An alternative approach is to solve the this and other purposes. We have not been able to predict inverseproblemofradiotherapy, this accurately the phenomena cited above using the default has the benefit of includingthe optimi- zationpart in a natural way. We have version of the EGS code.In an attempt to calculate more developed an iterative inversionalgo- accurately, we have incorporated improved rithm for solving the inverseproblem in bremsstrahlung angular dependence in EGS and employed radiotherapy. The propertiesof this al- "parameter reduced electron-step transport algorithm" gorithmwill be discussed, and compared (PRESTA) of Bielajew and Rogers to calculate energy with other methods. fluence and spectra and angular distribution of photons as a function radial distance from the central ray. The effect of these improvements on "horns" and output factors will be demonstrated and a comparison of results of calculations with experiments will be shown. 44 22 1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas Medical Physics Scientific Papers

MP5.8 MP5.9 Beam Data Generator for Asymmetrically Inclusionof Electron Multiple Collimated Fields, D.D. Loshek Scattering in Photon Dose Calcula- Marshfield Clinic, Marshfield WI, USA tion, C.X.Yu*and J.W.Wong, Washington University School of X-ray fields that are shifted away from the Medicine, St. Louis, MO, USA. central axis by independent collimators exhibit alterations of output, beam profile and average energy. These perturbations are not well Changes in atomic number can result in described by treatment planning systems that rely large dose changes by perturbing electron on discrete fan-line beam data. Two software multiple scattering. To incorporate atomic modules capable of generating appropriate fan- number effects in inhomogeneity correc- line data for use in these systems are described. tion, explicit electron transport based on Both employ an off-axis output function end a Ferme-Eyges theory was made for electrons profile generator. The profiles are generated as generated by primary photon interactions. a product of three field-size-independent Monte Carlo simulations of broad photon functions. These are a function describing the beam incident on thin slabs of different x-ray intensity in the absence of the collimator materials were used to obtain the energy assembly, a function describing the scattering spectra of secondary electrons. The angu- phenomena and a function describing the effects lar and spatial distributions at each of the collimators. The variation in beam energy energy increment were also scored to de- is described by an explicit correction for the rive initial scattering parameters. Elec- variation in zero-field-size tissue-maximum-ratio trons were then transported layer by layer as a function of off-axis distance. One module using a mini-pencil summation method. utilizes the fan-line data for a symmetric field Electron louses during transport were cal- of equal size to calculate the values of scatter- culated with an empirical function. Calcu- maximum-ratio while the other module uses a lations are being made to predict the sector integration over the field boundaries of small inhomogeneity data. The model forms the scatter-maximum-ratio data. The execution the basis of a photon dose calculation time of these modules is sufi,ciently fast to algorithm which include electron multiple allow in-line use in the beam definition process. scattering.(Supported by NCI grant CA- 41574). Supported in part by NCI grant 1 ROl CA40253

MP6.1 MP6.2 Where Does the LQ Model "Fit"?, D. Herbert, Hazard Analysis and Radiation Survey of UV University of South Alabama, College of Medicine, Lamps in Operating Rooms, Srxesh M. Brah- Dept. of Radiology, 269 CSAB, Mobile, AL. 36688. mavar, Cathy Miller*, Helen Bollea and Marilyn Riddle. Baystate Medical Center, The LQ hypothesis is now regularly deployed Inc., Springfield, MA01199 to explain, with apodictic assurance, a wide variety of so-called stochastic radiation The adverse effects of exposure to germicidal UV responses at both high and low dose. For lamps (254 nm) used in destruction of bacteria in example, mutagenesis in plants (pink mutants in air are erythema, eye burn, skin irritation, ex- Tradescantia 02); carcinogenesis in humans cessive dryness and incidents of conjunctivitis. (leukemia incidence and cancer mortality at The beneficial or adverse effects depend on the Hiroshima and Nagasaki); cell survival (bone product (uwsec/cm2) of time and intensity of UV ex- marrow stem cell survival in the mouse and the posure in operating rooms. A hazard analysis of rat); radiation "complications" of normal UV lamps was undertaken to establish reasonable tissues (hind leg paresis in rats). For each of output levels at various locations in operating these endpoints we have compared the appropriate room. The intensity measurements were made at generalized linear model (either Poisson or seventeen (17) different locations in two ORs for Bernoulli response) of the LQ hypothesis a set of seven (7) UV ceiling lamps per room. The with the cognate model of a leading rival intensity controls for inner and outer group of UV hypothesis (eg., Target theory, NSD, etc.) on lamps were set independently to read 20 uw/cm2 di- published experimental data. Our analyses rectly below each lamp at 6' from the floor. This include the respective regression diagnostics results in output of 30 uw/cm2 at 3' from lamp for both models of each set of data as well as surface. The locations included are: 6 spots at some use of semi-Bayesian regression methods a height of 6' near OR table and 4 spots on the where appropriate. The paper includes OR table at a height of 3'. The questionable ef- discussions of - and empirical evidence for - fectiveness of UV lights in OR is further compli- several current heterodoxies. One conclusion is cated by the need to observe short time of expo- that the choices of the respective published sure to avoid erythema or eye burns. Therefore, versions of the LQ hypothesis cannot be easily a RS/QC program for UV lamps has been established defended on the basis of these data - except as, to overcome observed adverse health effects among in M. V. Berry's locution, "an invocation." OR staff. 45 1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas 23 Medical Physics Scientific Papers

MP6.3 MP6.4 Some Aspects of Radiation Environment for Hazards to the Earth's Ozone Environment over India Space Radiobiology: A Selected Review, B.L. Acharya*, A.A. Chougule and G.R. Agarwal, Dr. R. Datta and Dr. S. Datta*, Department of R.N.T. Medical College, Udaipur, INDIA Radiology, LSUMC-S, Shreveport,Louisiana,USA. It is increasing concern of artificial production Space radiobiology has just born as a consequence of nitric oxide and fluromethanes, nitrogenous fer- of manned missions in space. Previous missions in tilizers and nuclear explosions reduce the ozone space involved short trips in the relatively less through reactions which are now fairly well under- hazardous radiation environment of equatorial low- stood and for which reaction rates are now avail- earth orbits. Future planned missions, however,are able. In the present investigation suitable model expected to involve increasing human activities in for ozone and recent available reaction rates con- space, flights to Mars, extended excursions to geo- stant have been used to find out the concentration synchronous and polar orbits for extended periods. of ozone. The ozone data obtained from baloon The earth's magnetic field is populated with sonde recently for various Indian stations .ike trapped electrons, protons, and low energy heavy Trivandrum, Poona, Delhi, Ahmedabad, Mount Abu have ions.The radiation environment in free space, how- been compared with that obtained from model calcula- ever, are the geomagnetically trapped, galactic tion. It is found that the ozone layer concentra- cosmic, and solar flare-associated particulatera- tion is reduced day by day and the flux of ultra- diations. Radiobiological factors in a rathercom- violet radiation reaching the earth have been in- plex radiation exposure to humans in space explore- creased. This can result in a variety of environ- tionmay have synergistic effect. Virtually nothing mental consequences including possible increase in is known asto their effects on biological systems, skin cancer, stoppage of tissue growth, albumin, especially at low doses for long term exposure and coagulation, change in global temperature, rain under microgravity conditions. Scenariosas to ra- fall conditions and introduction of ecological dis- diati induced-effects need to be quantified. Rel- turbances. Biological, climatic and ecological evant topics such as spaceflight dosimetry,biologi- effect of ozone depletion have been discussed in cal effects of the so-called HZE particles, inter- detail. actions of dynamic spaceflight factors and radiation, and finally the necessary trade-offs ofsys- tem performance to address sytems safety concerns as to radiation effects on both electronic components and population at risk are highlighted.

MP6.5 MP6.6 In vitro measurements of 5-iminodaunomycin ROLE OF LASERS ANDDIHEMATO laser induced toxicity, A. Andreoni, PORPHYRINE ETHER (PHOTOFRIN-II) IN A. Colasanti, U. Giani, A. Kisslinger, DAMAGING THE MEMBRANES OF HUMAN A. Mancini, M. Mastrocinque, G. Roberti, ERYTHROCYTES. S. Ganesan*, S.R.Manoharan,V.Masilamani and P. Russo*, University of Naples, 1-80125 G.N.S. Prasad, Dept. ofPhysics, Naples, Italy. Anna University, Madras-25, INDIA.

In order to evaluate the effectivenessof Photoradiation therapy for local treatment antracyclines in tumoral phototherapy, we of tumours utilizing hematoporphyrin deri- investigated the effects of laser light on vative as photosensitizing drug is under- cellular cultures of human liposarcoma incubated going clinical trials in several countries. But the in a medium containing 5-iminodaunomycin. Cells action mechanism in the killing of cells is still unclear. The were irradiated, with different light doses, at present study was under taken to explore the absorbance peak of the drug (590 nm), witha the posible role in membrane damage of cw dye laser pumped by an Argonlaser. Cell human erythrocytes due to lasers at diffe- viability wasevaluated by measuring their rent wavelengths and the photosensitizer, uptake of 04 thymidine. Our results show that Photofrin-II. It was found that, i) the the treatment with percent hemolysis increases in increasing drug (at a given the concentration of the drug, ii) increa- concentration) and laser light affects the cell ses with pre-incubation time,indicating mortality morethan drug alone (atthesame the uptake of drugs by the cells, Ai) de- concentration). The results are discussed taking pendson the post incubation time indicating :ato account some possible interaction the photochemical nature of the damage and complete mechanisms between 5-iminodaunomycin and DNA hemolysis was found within 24 hours after irradiation and iv) the percent hemolysis purely depends on the wavelengths of laser used, indicates that, the nature of the effect was purely chemical one and not thermalone. 46 24 1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas Medical Physics Scientific Papers

MP6.7 MP6.8 Ultrasound Irradiation of Chicken Embryos - BI-PYRAMIDAL MULTIMODE RESONATORS FOR Technique. MS Rosenthal*, JW Yip, YPL Yip, IN-VIVOELECTROMAGNETIC SPECTROSCOPY, and C Capriotti, Universit7 of Pittsburgh, Brondo, Scientific Innovations Inc., Pittsburgh, PA, 15261, USA Wainscott NY, USA.

Documentation of bio-effects from non-ionizing A large bi-pyramidal cavity resonator is descri- radiation requires precise knowledge of the expo- bed, intended for the measurement of thermal sure to the animal system and as natural environ- noise emission in a broad range of microwave ment as possible. We report here an exposure sys- frequencies from a human or animal subject pla- tem that allows good quantitation of the embryos' ced inside the cavity. This is a more sensitive exposure to ultrasound irradiation. arrangement than an anechoic chamber with an an- The embryos are prepared J3r making two openings in tenna array and it is a matter of interest to the egg shell, one at the air sac, the other at establish not only the shape of thermal noise the opposite end. Over the second hole a gas per- emission spectra but also to register any non- meable, ultrasound transparent membrane is applied. thermal components of emission, which have been For irradiation, the egg is placed vertically with postulated by others. The relationships between the air sac towards the bottom in a large water the dimensions of the source, the cavity and the bath maintained at 37°C. the embryo is positioned wavelengths in question are such that the cavity in the middle of the second hole. The ultrasound operates as a multi-mode resonator. The bi-pyra- energy is transported from the transducer to the mid shape permits an introduction of pick-up egg using an acoustic coupler. Below the egg is a couplers into the side walls with a smaller microprobe whose output is fed into an A/D con- amount of perturbation than in other designs. verter on board a PC/AT computer providing online The system can be calibrated with a white noise monitoring of the exposure. generator and the cavity can be also used to Results indicate that the SWR is approximately 1.1 measure absorption by the body in the microwave with this configuration, and that there is 80% region. The absorption measurements require a transmission through the egg. broadband noise generator and a spectrum analyzer. In the absence of the body, a swept frequen cy source is needed to measure the intrinsic response of the cavity.

MP6.9 MP6.10 BACTERICIDAL EFFECTS OF LASERS AND Study of the Metabolism of Nuclear Fallout ACRIDINE ORANGE ON E-COLI:IN VITRO Products by means of Stable Isotopes as STUDIES, S.Ganesan*, V.Jayapal+, Tracers. V.Masilamani, G.N.S.Prasad, M.C.Cantonexo,C.Cevicchiolixo,G.Gambarini+0, S.P.Thiagarajan+. and D.Subramniam+, N.Molhoxo,L.Pirolaxo,Ch.Hansen*,E.Werner* rtaioartimento di Fisica dell'Universita, Dept. of Physics, Anna University, Milano, Italy Madras-25 +Dept. of Medical Micro- o I.N.F.N. Sezione di Milano, Milano, Italy biology, Institute of Basic Medical + Istituto di Fisica del Politecnico, Science, Univ. of Madras, Taramani, Milano, Italy * Ges. fUr Strahlen-und Umweltforschung, Madras-113, INDIA. Frankfurt/Main, FRG Laser play a vital role in orthopedics The evaluation of the dose due to the intake of as an adjunct treatmentfor chronic osteo- radioisotopes from fallout is based on specific myelitis. The paper presents the study of metabolicmodels. These models in somecases bactericidal effect of lasers in the are quite detailed and may be supported by much Orange experimental evidence fromobservation on presence ofa sensitizer Acridine humans. In other cases, data may be limited to in on E-Coli, theorganism encountered a few observations on a particulargroup of chronic osteomyelitis. The resultsshhows populationor on species of experimental ani- effect han mals. that Nitrogen laser has better Webeliveof particular significance to study other laser has better effectthan other the humanratabolism of elements, likefor lasers and the killing effectenhanced in instanceMo, whose radioactiveisotopes are So the included in nuclear fallout. Considering that the presence of Acridine Orange. radioactive and stable isotopes follow the same Bactericidal effect was essentiallyphoto- metabolism, stable isotopes are used as tracers chemical in nature and not a thermal one. to prevent radiation damage. Proton NuclearActivation technique, wich allowed us to study iron metabolism by means of stable isotopes, was applied to the study of Mo metabolism. The optimization of the relevant parameterswas previously tested in orderto nval.iate the sensitivity limits for an instrumental analysis.

1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas 25 Medical Physics Scientific Papers

MP6.11 MP6.12 Study of the Tannin Influence on Iron PHOTON ACTIVATION ANALYSIS IN-VIVO FOR Metabolism studied by means of Stable Isotopes as Tracers. MAJOR BODY ELEMENTS, K V Ettinger, II University of Rome, Italy and U J Miola* M.C.CantonexoG.Gambarini+0,M.Molhoxo, Royal Armed Forces Hospital, Riyadh, L.Pirolaxo,P.Salaxo,Ch.HanseEw7E-Warner* x Dipartimento di Fisica dell'UE,versita, Saudi Arabia . Milano, Italy o I.N.F.N. Sezione di Milano, Milano, Italy Photon activation analysis has got a number of + Istituto di Fisica del Politecnico, Milano, Italy advantages in comparison with the technique of * Ges. fUr Strahlen-und Umweltforschung, neutron activation analysis in present use. The Frankfurt/Main, FRG uniformity of detection is very good and the effect of the body shape can be easily accounted The possibility ofstudying in humans the for or even ignored. It is possible to measure correlation between intestinal iron absorption oxygen, carbon and nitrogen using bremsstrahlung and the assumption ofparticularsubstances of sufficiently high energy, in order to form includodin alimentary diet, is related tothe methodused for investigating the intestinal positron emitters which could be distinguished absorption itself. from each other by the differences in their half By considering that it is not free of danger to lives. One of the most common measurements in- carry outsuch a study on humans by means of radioactive tracers, we belive that amethod vivo is the determination of nitrogen, which is using stable isotopes as tracers finds agood most easily measured by photonuclear method, field of application. having the lowest formation threshold (10.6 MeV). We developeda method for the studyof iron metabolism by means of stable isotopes measured Potehtially most accurate are measurements which by proton nuclear activation. This method was utilize as aas a source of photon fluence,X- first tested on rabbits and then on humans. rays produced in reverse Compton scattering of The present study is aimed to verifythe laser light from electrcn beams, a facility hypothesisthat tannininfluences the iron intestinal absorption. which is now becoming available in most of the For this purpose healthy people is investigated synchrotron light sources. Two examples are before and after assumption of a tanninrich food. given of the installations for photonuclear activation analysis in-vivo, The radiation dose imparted to the patient ,in order to attain a result with accuracy and precision adequate for clinical use, is about 1 - 2 cGy.

MP6.13 MP6.14 Influence of Rotating Magnetic Field on the IMPROVED DIAGNOSTIC PERFORMANCE ON THE SEVERITY Activity of Alkaline Phosphatase in Iaflam- OF LEFT VENTRICULAR HYPERTROPHY USING BODY matory Exudate, Hu Jixiang* , Wang Mibin, SURFACEMAPPING. M. Yamaki: K. Ikeda,I Kubota, Chen Dezhong, Yang Zhenbzng, Henan Medical K Nakamura, K Hanashima, andS Yasui. Yamagata University, Changsha, Hunan, The People's University School of Medicine, Yamagata, Japan Republic of China To improve the ECG diagnosis on the severity of left ventricular hype trophy (LVH), 87 uni- The purpose of this experiment investigation was polar ECGs were recorded from 87 pts with LVH trying to search the mechanism of antiedema effect (34 hypertension,23 HCM,30 AR). The ECG of rotating magnetic field (antiedema was one of changes were evaluated by two following points: the effects of magnetic field). The rat paw was 1)increasing in QRS voltage, 2) delaying local immersed into hot water of 60°C for 10 seconds to activation. We measured for each lead, a)R induce local edema, one of the rat paw exposed to voltage, b) AQRS (the net area of QRS), c)VAT rotating magnetic field of 1000 Oe. for six hours. (ventricular activation time), and d) DIof At this time an exudate sample of 2-4 microlitre AQRS and VAT, where DI=(X-mean)/SD. From those was taken by microlitre injector. The rapid method measurements, we calculated 7ECG parameters for determining alkaline phosphatase activity in for each patient: 1) R max, 2) AQRSmax, 3) serum with thymolphthalein monophosphate as a new AQRS-max DI, 4) AQRS +2SD area (area size where substrate was used. DI>2),5) VAT max,6) VAT-max DI,and 7) VAT The activity of alkaline phosphatase of the expo- +2SD area. Among these, we selected themost sure group and the control was employed for stati- effective parameters for the LVH diagnosis stical estimation. Significant difference (n=33, employing a stepwise multiple regression P 0.01) between two groups was noted. It indicated analysis.In the hypertension and HCM group, that the activity of alkaline phosphatase in ex- the combination of VAT +2SD area and Rmax was udate of tie treated group was higher than that of the best for estimating the wall thickness the control. (multi-ple R =0.73). In the AR group, the AQRS max was the best for estimating LV Internal dimension(r= 0.73). Developed regression equation correctly diagnosed 87% of subjectson the presence of the LVH. These weresuperior to 48 the 12-lead ECG. We can improve the ECG diagnosis for LVH by ECG mapping.

26 1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas Medical Physics Scientific Papers

MP6.15 MP6.16 A Simulation Study of 3D Vascular Bed Study of Biomechanical Mechanism Reconstruction. J. Wu*and D. L. Parker, LDS for Occupational Exposures to Vib- EfospftEdAjniversity Utah, SLC, Utah 84143. ration, J. Xie*, S. Ma, M. Xiong, and X.Liu, Hunan Medical Univercity This paper presents a simulation study of the effects Changsha, Hunan, P. R. China of various imaging parameters and physical image degradation sources on the accuracy and precision This paper describes the study of mecha- of 3D artery reconstruction. 3D coordinates of the nism on secondary Raynaud's phenomenon vessel centerline are determined from the according to biomechanical principles and centerline coordinates in two views.Orientation methods. We finished the researches of corrected densitometric information is averaged 8 mine areas and 1067 pneumatic drillers. over the entire, visible, arterial bed, to determine a We made some animal experiments as well. densitometric conversion factor for each image. The results follows: a). Irradiating with Absolute lumen cross sectional area is determined vibration can make micrangiums resistance densitonatrically for all vessel positions. to increase up to 4.5 times. b). Irradi- For the simulation study,a3D representation of a ating with vibration can cause blood human left coronary artery is used to generate two vessel collapse. The collapse grade is projection images at arbitrary orientations.The determined by the vibrational accelera- effects of beam hardening, x-ray scatter and system tion and the stress due to the power of gripping. Usually, the effect of stress blurring are incorporated in the forward projection, is bigger than acceleration's. c). As to match the imaging physics. The resulting 3D far as vibrational harm is concerned, in representation is compared with the original "true" the part of high frequency, Raynaud's representation. In the absense of beam hardening phenomenon is the main damage. In the and when the vessel is clearly seen in both views, part of low frequency, the whole body an average error of 0.08 41- 0.13 mm in vessel vibrational damage is main. radius is obtained. While the system is moderately insensitive to scatter, beam hardening causes an large and small x-ray paths through iodine to be under- and overestimated, respectively.

MP6.17 MP6.18 llicroccroputer-basedNuclear Probe: an ATPase Enzyme Structure Extraction From experience in Cardiac Surgery, R.M.V. Its Blurred Image, Liuda Chu, ...nstitute Piva*, J.M.S.Pascual, J.O.C.Auler Jr., J.C.Meneguetti, C.P.Melo and C.C.Robi- Of Electronics, Academia Sinica, China lotta, Heart Institute and Physics In titute of University of Sao Paulo, C.P. 8091, Sao Paulo - BRAZIL. The structure of ATPase enzyme has been studied by using different methods The interfacing of a scintillation detector (a 2" x 2" NaC1(11) crystal) with an appropriate col - including protein sequencing which has limator to a microcomputer can provide a simple produced its general structure. The method for beat-to-beat ventricular function evaluation. The technique, which includes an injection shape and density taken under microscope of Tc -99m labelling red blood cells, wasprevicusly show high background noise and low reso- validated with cardiac gated blood pool imaging using a conventional game camera system. lution. Image processing algorithms have At the Heart Institute, the potential of this de- been developed and implemented for redu- vice as a bedside monitoring system has been tested in patients submitted to various cardiac surgical ing noise, enhancing and detecting edges procedures. Ejection fraction, heart rate,systolic to extract contours. Gauss-Laplacian and diastolic intervals were measured at pre-operatory period as basal parameters to becompared filter, histogram equalization, linear with the ones obtained at post-operatory period transform and non-linear statistical mainly in high risk patients. It was noted a good correlation with clinical evolution. transform have been used. In addition The method is presented as a low cost and non-inva geometrical transform has been combined sive alternative forcontinuous monitoring of cardiac function in.patientssubmitted-to-cardiac with a digitizer to achieve rearrange- surgery. ment of those enzymes for comparasion. It is proved practical and useful. 49

1988 World Congress on Medical Physics and Biomedical Engineering c San Antonio, Texas 27 Medical Physics Scientific Papers

MP6.19 MP6.20 Predicting decay in radical concentration CARDIAC OUTPUT MEASUREMENT USING DOPPLERTECHNIQUE. in an amino acid following heavy ion ex- S. Sakthivel, S. Suresh, I. Suresh, andG. Prasad% posures. J.W. Hanseeand K.J. Olsen, Rise Dept. of Physics, Anna University, Madras-2S, National Laboratory, DK-4000 Roskilde, INDIA Denmark. Herlev Hospital, 2730-Herlev, Denmark. Cardiac output is the most important descriptor of ventricular function in clinicaluse today. The decay rate of the CH COON radical inthe ami- Doppler principle allows accurate assessment of no acid L-a-alanine depelids on the depositeden- cardiac output noninvasively. To determine car- ergy 9.nd is less than 1% per year after exposure diac output we need the heart rate and thestroke with 00Co y-rays to doses an order of magnitude volume. The stroke volume can be calculated from below the dose for maximum obtainable radical the flow velocity integrated andcross sectional production, saturation dose. A decay of 13% in area. The cardiac output was calculated from the 2000 hours has been measured afterexposures to Doppler in formation measured from the apical saturation doses with fast electrons of highdose window. The cardiac output can be measured either rate. Increasing decay rates of radicalconcen- at mitre' valve or at aortic root at pulmonary tration have been measured at low average doses, valve. In this study, the cardiac outputwas single tracks, from heavy charged particles. measured from LVOT close to aorticroot. In case Exposure with penetrating 16 MeV protons results of pulmonary flow, it isvery difficult to get in a small but significant decay, while an ap- the proper window. The cardiac output is seen preciably larger decay has been measured for to increase with advancing age from this study stopping particles of higher average LET in the (childhood to adulthood). It reaches a plateau sample. We have compared measured radical decay which adult stages and itwas also found that the after exposures to low average doses from heavy stroke volume and heart rateare inter-related. charged particles with an expected decay obtained If the stroke volume is high, the heartrate is from calculations on track energy density distri- less and vice versa. The compensates the final butions and measured decay after fast electron output. exposures to saturation doses. We see a good agreement between experimental data and results obtained from this model.

t1P6.2i MP6.22 T3,T4 &TSH levels estimation for testing the thyroid function in Adivasiarea, G.R. Agarwal, A Preliminary Experimental Study on Reflection, Transmission and B.L. Acharya* and M. Tyagi, R.N.T. Medical College, Waipur-313001(India) Absorption of Light By Human Skin, Tang Jianming*, Fu Changyu, Third The thyroid hormones estimation is an important Military Medical College, Chongqing, investigation in assessing the differentthyroid P.R.China diseases. The tote:1 1050 Cases have been considered in the present investigation. In the heart disease Cases (35) serum T3 is foundto be low. It was found in this study that light irradiat- the stress generally canses a fall inserum T3 ed on the human body is reflected not only by ravels in heart disease patients. In diabetes the superficial skinbut also by the tissue patients (../7) T3 is found to be low inmost of layers underneaththe skin. The absorption cases where as T4 and TSH levels are normal in peaks of the reflectance curves for the light diabetes cases as well as in heart diseasecases. wi+ he wavelength between 400-800nm by the The hormones levels estimation in malignanthead hum. body are similar to those of the transmit- and neck cases (45) show after externalirradia- tance curves by the humanblood; and in the tion induces cuthyroid sick. The female cases in same range of light wavelenght, the reflectance the age group of 16-30 years are ma..imum. The and transmittance curves of theskinor skin laboratory normal range has been 9!'ven usingcon- plus hypodermic tissues possess the similar ten- trol charts for T3, Tel&TA. The results have dency of alterations. Thus this study provides been discussed. certain useful parameters for the treatment and protection against 1:ght and laser.

5©

28 1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas Medical Physics Scientific Papers

MP6.23 MP6.24 EPR and Dielectric Relaxation applied to in vitro Soft X-ray Contact Micrographs of Various Myogenesis: Effectsof Ions,M.T. Santini, Lab. Biological. Cells: A Test of Resolution, G. D. di Fisica, ISS, Rome, Italy, A. Bonincontro, C. Guttmann* and M. R. Howells, Center for X-ray Cametti, Dip. di Fisica, U. di Roma, Rome, Italy, Optics, and E. L. Alpen, Life Sciences, Lawrence R.E. Hausman, Biology Dept., Boston U., Boston, Berkeley Laboratory, Berkeley, CA 94720, U. S. A. MA and P.L.Indovina, Dip.diFisica, U. di Napoli, Naples, Italy Numerous biological samples have been examined by soft x-ray contact microscopy at LBL. An effort has been made to establish a protocol for Electron Paramagnetic Resonance (EPR) anddielectric preparing biological samples for examination in relaxation (in the radiofrequency range) allow information the soft x-ray contact microscopy station at LBL. of membrane order and the conductivity and permittivity Development and electron microscopy examination of of membranes tobe obtained. We appliedthese two the exposed x-ray sensitive resists have also techniquestothestudy of membrane changes occurring followed an established procedure. The images during myogenic differentiation.With EPR, we have shown presented will test the resolution of soft x-ray thatmembrane order varies during myogcnesis and that contact microscopy as applied to biological thisvariau isaltered by the addition of monovalent samples. Typical samples being examined are ions (Li+, Cs').An analysis using a "single-shell" model premature condelt. d chromatin, 9L glialsarcoma oftheconductivitydispersionducto interfacial cells, and vesicles made from the "brush border" polarization (Maxwell-Wagner Effect) secninmyoblasts of the LLC-PK (pig kidney tubule) cells. This during development dcmonstratedthatconductivity and research is supported by the U. S. Department of permittivity of the mcmbranc goes down abruptly during Energy under Contract No. DE-AC03-76-SF00098. the period of known fusion.Moreover, the addition of toe above-mentioned ionsshiftedtheabruptdropin conductivityabdpermittivitytolaterculturetimes, indicationg that fusion had also been delayed. A possible model for the effects causcd by these ions on myogcnisis issuggested.

MP6.25 MP6.26 INTELLIGENCE, COGNITION AND FUZZY LOGIC The Short-Range Acoustical Field Calculation IntelligentSystems Research and Resolution Estimate of Linear Array Madan M. Gupta: Laboratory, College of Engineering, University in Medical Ultrasonic Tomograph of Saskatchewan, Saskatoon, S7N OWO, Canada. R.C.Cheneand Y.Z.Jing Shanghai Jiao Tong University. Shanghai. China This paper is concerned with themodeling aspects of the process of human mentationand cognition using fuzzy logic for the emulation of intelligence and other associated attributes on amachine. The humanmentationacts upon the acoustical fieldradiated by a The calculation of cognitive information. The naturalsensors, definite aperature is based on Huygeus-Fresnel principle. eyes, ears, nose etc. acquire the fresh The mathematical operationmay be simplified, if Fresnel information from our environment on a continuous the or Fraunhpfer approximation is adopted .But the condition basis. This information is relayed to the brain, through sensory for approximation are usually unsatisfied for the most of cognitive processor, The cognitive processor recallsthe detected regions in medical ultrasonic apparatus. In this neurons. previouslystoredinformationin the memory, article, the caculation is mainly accomplished by meansof (experience, knowledge-base) andconverts the for all regions angular spectrum method which is suitable freshly acquired information into relative in acoustical field.With the applicationof FFT algori- gtades. Then the process of mentation does an thum in this method, the time of the numerial calculation appropriate processing yielding the phenomenon known as 'perception'. This process of mentation may be saved. In this paper, the choiceof 'nterval and thus deals with the graded-uncertainty and its amount of sampling points is discussed with an example of cognate graded-information. We call such an The result of calculationprovides a basis linear array. informationas thecognitiveinformation and for researching lateral resolution of linear array andfor model it using the notion of graded membership dynamic focussing segment by segment. embedded in fuzzy logic. This notion of graded membership leads towards the natural mentation, cognition and perception phenomena. Presently our studies are directed towardsthe emulation of natural vision-perception phenomenon and cognitive process.

29 1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas Medical Physics Scientific Papers

MP6.27 MP7.1 Meeting the Demand of Medical Physicists for THE RADIOLOGICAL ACCIDENT IN GOIANIA: AN OVERVIEW the year 2000, U. Madhvanath, DRP, Bhabha C. E. de Almeida; J. J. Rozenthal and R. N. Alves Atomic Research Centre, Bombay-400 085, India. Comissio Nacional de Energia Nuclear - BRAZIL Human resources development is required to be A radiation accident involving a radiotherapy Cs-137 source,with consciously planned if we are to effectivelyman the diagnostic and therapeutic facilities of popul- an activity of 1375 Ci, has occurred last september in Goiiinia ations in the emerging countries-all looking into (1.100.000 inh.). A radiotherapy source head in disuse, was

the year 2000. The unique experiment carried out removed by a group of people inv,...1.:ed with the junkyard by the Bhabha Atomic Research Centre ina period of business. Later on, it was broker., releasing its radioactive 25 years stands testimony to this development pro- content to the environment in the chemical form of Cesium gramme. Sensing the advent of cobalt-60 teittherapy cloride. Due to the high solubility of this the machines to India in mid-fifties, stx physicists of compound this centre were deputed to TorontJ, London, Stock- dispersion into the environment was favoured by the winds and holm and other places of radiotherapy centres and rainfalls. However, the major contamination vectorwas indeed following their return, a one year trainingcourse the people directly involved, since they have transfered parts on Hospital Physics and Radiological Physics was of the source with the lead shielding to three junkyards in the started in 1961 to train 15-20 physics graduates city and distributed small parts of the source to members of drawn from all over the country each year and provi- their families. "The blue glow emmitted by the compound,led them de them to hospitals, industries and research insti- to believe that they could be facing a precious stone".As result tutions handling radiation sources. Country's demand is abundantly, met. They man todat 150 telethe- of this, seven major radiation spots was found inthe city rapy facilities, 90 each of brachyth'rapy and nucl- within the same neighborhood, with radiation levels up to 110R/h ear medicine centres. The success this programme at one meter, immediately evacuated and isolated. Externaland rested in the generous financial support by the internal contamination was found in 244 people out of 118 220 Government of meeting the living expenses of the monitored; 129 people required medical care, including some with candidates as it was realized that a new discpline and profession is not viable at the universities. severe internal contamination and external radiation injuries. Problems of developing countries, when advanced Four people died of acute radiation effects and one had his right technologies of the West were borrowed, can only be forearm amputated. This paper will present an overview of the overcome Ly organised training programmes of this accident, the scenario, the emergency procedures, the dosimetry type. and its environmental impact.

MP7.2 MP7.3 THE CHINESE HUMANOID PHANTOM AND ITS APPLICATIONS A Survey of Radiation Exposure in Diagnostic IN MONITORING FOR RADIATION DOSAGE IN CHINA. Lin Radiology in Latin America,G.P.Hanson ,World Da-quan* and W. Da-ke, Chengdu University of Science and Technology, Sichuan, P.R.C. Health Organization,1211 Geneva 27 Switzerland and G.G6mez Crespo,PAHONHO Consultant. The humanoid phantom has been widely used in medi- In 1984 the Pan American Health Organization cal radiation fields. For the first time in China arranged visits by the authors to Argentina, since 1984, we have developed a new humanoid phantom which has successfully applied in monitoring Brazil,Colombia,Costa Rica,Ecuador,Mexico,Nether- lands Antilles and Nicaragua in order for radiation dosages. This phantom, which is to collect 165 cm in height and 57.5 kg in weight, is bALqd information concerning the availability and utilization of radiodiagnostic services in Latin on the average adult man of china. The thermo- Ameri- plastic coupled materials, which radiation equiv- ca and the Caribbean.That study included a survey of the radiation exposure received by patients alence is assured among 12-1250 KeV and its devi- during the most frequently requested ation is within + 5%, are adopted as the tissue x-ray examinations. substitute for all parts of the phantom except the To our knowledge this survey is unique lungs. Our phantom consists of 11 parts and 20 because a cross sections and each section has 3x3cm rectang- standard methodology was used and all themeasure- ments were made with instruments provided and/or ular hole grid and organ hole, and the TLD holes calibrated by a single reference dosimetry (3x2Omm) are over 500. Assemble devices have been laboratory whose secondary standard calibration instru- developed which provide a new ease of use in pat- ments are compared directly with a primary radia- ient-like positioning. Internal nylon rods pass through holes in each section. tion standards laboratory,the U.S.NationalBureau of Standards. The type of phantom has been produced 10 sets since 3984 in China ane thoy are the important experimental tool in radiation protection, medical diagnosis, radiotherapy, medical education and the radiation research field, etc. F2

30 1988 World Congress on Medical Physics andBiomedical EngineeringSan Antonio, Texas Medical Physics Scientific Papers

MP7.4 MP7.5 CT useage in Manitoba and the correspond- A Comparison of Several Mammography Films Using ing patient doses (1977-1987).W. Huda* a Contrast-Detail Imaging Phantom and a 152-ACR and G.A. Sandison, Medical Physics, MCTRF, Phantom. R.C. Fleischman*, L.N. Rothenberg, 100 Olivia Street, Winnipeg, Canada, R3E OV9. Memorial Sloan-Kettering Cancer Center, New York, NY 10021. The use of CT in Manitoba has increased from an 3 initial rate of 4.2 patients per year per 10 poR- Reading a mammogram is one of the most diffi- ulation in 1978 to 18.2 patients per year per 10 cult tasks required of a radiologist, the film/ in 1987. Values of effective dose equivalents, screen system greatly influences the ability of H's have been evaluated for the EMI 5005, GE 9800 the radiologist to detect clinically important E ' and Siemens DRH CT scanners used in this province information. The dose required to obtain an image over the 11 year period under study. Modern CT is also an extremely important characteristic scanners result in an HE value of 200 mrem for of the film. typical head examinations and HE values in the We have compared Dupont sR-329, range 600-1,700 mrem for typical body examinations. Dupont SP-150, Kodak 0M-1, Kodak TM and Fuji MI-NH The per caput population effective dose equivalent mammography films using a Kodak Min-R cassette in Manitoba increased from 0.52 mrem in 1978 to for all exposures. A 152-ACR phantom approved 8.1 mrem in 1987. for use in The American College of Radiology Mammography Accreditation Program was radiographed to obtain images of clinically significant targets.A contrast-detail phantom was also radiographed to evaluate the films. For each type of film, three radiographs were taken of each phantom. The radiographs were ranked in order of the detectability of the various targets. The dosimetry characteristics of the films were measured using a Radcal 10X5-3M chamber. The film speed as well as the H&D characteristics of the films were obtained using BR-12 phantom material.

MP7.6 MP7.7 Image Analyzing System to Measure Etched Biological Half-life of 1-125 in Nuclear Tracks on CR-39 Detectors to Neutron Japanese Thyroid :Kunihide Nishizawa, Dosimetry and LET Spectroscopy of Heavy Nagoya University, 55 Tsuruma-cho, Charged Particles Showa-ku, Nagoya 466 Japan

it§ Julius J. Aimasi and Cornelius A. Tobias Lawrence Berkeley Laboratory In vivo thyroidal monitoring was University of California, Berkeley, CA 94720 executed for 51 workers doing iodization and for 38 workers doing radioimmunoassay Neutron spectrometry and dosimetry are important from 1984 to 1987. The monitoring subjects of radiation safety at medical system was composed of a multichannel accelerators. CR-39 track etch detectors are pulse height analyzer(MCA) and two the most sensitive low cost passive neutron NaI(T1) scintilation detector or an dosimeters. CR-39s are also used to three NaI(T1) survey meter having a connector dimensional LET spectroscopy of heavy charged to the MCA. The counting efficiency was particle beams and can be used to heavy ion determined with an anthoropomorphic neck radiography as well as to several other phantom. Contamination was detected in applications. All the above-mentioned 17 persons doing iodization and the applications of the CR-39s require a fast measurements of the biological half-lives and reliable image analyzing system which can were completed for 6 iersons of them. discriminate between tracks end background No contamination was found in the workers spots and can count and measure a large number doing radioimmunoassay. Biological halfof etched nuclear tracks. We are reporting life of iodine-125 ranged from 20 to 90 about the development of such a fully automatic days, the average being 48.3 days. These low cost and high performance system. We will values are considerably shorter than the present examples of the application of our system biological half-life of iodine of 120 at different dosimetry and spectroscopy days for reference man of ICRP. measurements.

NIH postdoctoral fellow, Present address Siemens Medical Labs., 2404 N. Main Street Walnut Creek, CA 94596

1988World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas 31 Medical Physics Scientific Papers

MP7.8 MP7.9 Radiation Protection in the Treatment. Estimated Doses in riasues Resulting Room during Intraoperative Radiation From Radionuclides in Dental Porcelain. Therapy P. Tortal*,Radiation Health Service J. Rao Nibhanupudy*, G. King, E. Ashayeri, Department of Health, Sta. Cruz AL. Goldson - Howard University Hospital Manila, Philippines 1003

During intraoperative radiation therapy in certain cases, it may become a require- U-238, T1-208, K-40 and Cs-137 were found to be ment for medical personnel to be in the present in dental porcelain through a gamma snec- treatment room near the patient. In tral scan performed using high purity GeLi scin- order to provide radiation protection for tillation detector coupled to an ND66 multi-chan- medical personnel during treatmet,measure- nel analyzer. ments were made behind lead shielding for Annual doses to the tongue, gums and the inner 6 MeV,9 MeV, 12 MeV, 15 MeV, and 18 MeV cheeks and lips were computed assuming continu- electron energies at different locations. ous contact with the exposed tissues. Calcula- Results are reported. ted doses from alpha and beta were derived from the definition of absorbed dose. Doses due to gamma were calculated using the MILD method. The resulting dose to the tongue was estimated to be 1.57 rads per year for male and 1.50 rads per year for female. The resulting dose at the gums and the inner cheeks and lips was estimated to be 0.80 rad per year.

MP8.1 MP8.2 Noninvasive Functional Localization by Biomagnetic Methods to Improve Neuromagnetic Imaging of Brain Methods, T. Kati la, Laboratory of Biomedical Engi- Functionality in Evoked Response Studies. E. R. neering, Helsinki University of Technology, 02150 Es- Flynnw. D. L. Arthur. C. J. Arne. J. S. George. P. poo, Finland A. Medvick. P. Lewis. W. Overton. G. S. Sullivan. and D. van Mulsteyn. Los Alamos National Various imaging methods, CT, MRI, ultrasound and Laboratory. Los Alamos. New Mexico. USA. radioimaging are today indispensable in clinical work. Neuromagnetic images of functional neural sources However, usually these methods do not give informa- may be obtained by using evoked responses from tion on the functional level, or in practice only long visual, auditory, and somatosensory stimuli.The time averages can be obtained. weak magnetic fields arc detected by an array of Electric surface potential measurements can be used SQUID coupled gradiometers and neural generators for noninvasive studies on the bioelectric functioning deduced by solving the inverse electrortragnetic of e.g. heart, brain, eye and muscles. The time scale problem. The results may be overlayed onto nuclear magnetic resonance anatomical images to present of these studies ranges from milliseconds to sustained three-dimensional images of functional structures phenomena. within the human brain.The limitations of this Lately, biomagnetic measurements hp e often turne technique are signs] -to -noise (STN) characteristics out to be superior to electric surrace studies in func- of the sensor output, localization accuracy of the tional localization of the bioelectriactivity. Sources of sensor-head system. and parameters of the soum.e abnormal cardiac electric activity have been localized itself. The first of these (STN) way be with an accuracy of about 1- 2 cm. In brain studies, substantially improved by a proper magnetic shielded environent, the application of singular- cortical sources of evoked potentials have been localized value decomposition methods and the use of with a reproducibility of about 2 mm and source shifts correlations between mcr",-ers of the sensor array. even smaller than that are detectable. For more accu- The second, localization of sensors with eesnot.t to rate results, the source and volume conductor models landmarks on the head, may be addressed by the use as well as the measurement policy are important. An of multiple sets of transmitting orthognal dipoles overview on the factors affecting the localization accu- in the form of a headband with the neuromagnetic racy is given. sensors acting as receivers. Finally, the source problem is approached by theoretical modeling of increasingly complex source descriptions. 4

32 1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio. Texas Medical Physics Scientific Papers

MP8.3 MP8.4 Equivalent Localization in the Brain Pathological Foci in AC- and DC-Magnetoen - to Sumatosensory Evoked Magnetic Fields cephalogram (MEG) Compared to Electroen- Using the ECD Technique, Y.Uchikawa*, cephalogram (EEG), J. Vieth: P. Schueler, C. A.Adachi,T.Hasegawa, and M.Kotani, Meyer, S. Malliarou, S. v.Harsdorf, University Tokyo Denki University, Hatoyama,Saitama, Erlangen, D-8520 Erlangen, Fed. Rep. Germany. 350-03 Japan.

MEG (1-channel D.C.-SQUID) and EEG (reference: contralateral ear or average) was recorded in 16 We have performed the magnetic measurement at patients with a focalEEG-activity caused by more than 25 different positins on the surfaceof epilepsyand/or tumor or iselemia. The recording the head in a nonmagnetically shielded room with grid was narrow (point distanc: MEG 2-3 cm: EEG a SQUID connected to a second order gradiometer In epilepticpatients averaues of 2-6 cm). with a baseline of 3.2 cm and a coil of 2.48 cm spike activity vere performed. The spikes ofthe diameter. In the measurement of somatosensory and EEG averages were selected and triggered MEG evoked magnetic fields(SEFs) the electric current by using a suitable EEG electrode (in case: sphe- was applied to stimulate the midian nerve ofthe noidal electrode). In all patients with a right hand with the duration of 0.2 msec and the pathological slow wave activity (delta or theta) interval of 0.5 sec. All ragnetic recordings were the "relative covariance" method was used. In all averaged with more than 500 samples. cases isocontour maps were made. 3D- localizing In order to estimate the location of current sou- of the focal source only was performed if the map rce in the brain generating the magnetic field showed a clear dipolar pattern (12 cases). distribution the 3-D ECD(equivalent current dipo- The EEG never showed a dipolar pattern. TheMEG le) technique was introduced in the meaning of did it and was able to localize the anatomically the best fit between experimental and theoretical lesions in all these patients. The verified distribution. Using this technique we could esti- activity during the start of an epileptic seizure mate the equivalent localization in the brain of was similar in the AC-MEG and theAC-EEG. The the SEFs in latencies of around 80 msec to 150 directly coupled (DC) MEG recording showed a sim- msec. The calculated ECD was located at the Rola- ilar shift preceding the start of an epileptic ndic fissure of the left hemisphere of the head as the DC-EEG recording via a foramen ovate fit where is the motor region related to the finger. electrode (STODIEK and WIESER, 1987).

MP8.5 MP8.6 Improvement of Biomagnetic Field Representa- Volume Current Effects in Magnetoenceph- tion by Ysing the Current Multipole Expansion, alography, E. Ducla-Soares*, D.F. Rose and S.N. Erne, L. Trahms and Z. Trontelj-, P.T.B., S. Sato, Neurophysiology Unit, MNB, NINCDS, Institut Berlin, Abbestr.2-12,D-1000 Berlin 10 National Institutes of Health, Bethesda, MD and* Phys.Dept. University E.Kardelj of Ljublja- na, Jadranska 19, 61000-Ljubljana, Yugoslavia. Magnetoencephalographic data have usually been analysed in terms of Biot-Savart's Law for the magnetic field of an impressed current. This method neglects the effects of volume currents. The inverse problem in bioelectricity and in bio - However, for he multi-channel magnetometers now magnetism is usually studied due to its nonunique available and for numerous clinical problems like solution in a model approach. Already the simplest temporal lobe epilepsy, that approach is a poor equivalent current source in a form of a current di- approximation. We present results on the magnetic oole brought success and allowed to localize some field of a current dipole implanted in the sources in the heart and in the brain. On the other temporal zone of a skull filled with a saline agar hand, magnetic field distributions which cannot be solution. They show a 5n departure from Biot- are also explained by a singledipolar source Savart's predictions for the relative amplitude known. of the extrema and also a considerable difference More complicated equivalent sources have to be con- in the position of the zero line. We interpret sidered: We will take into account the next higher these results in terms of a spherical model of order term in the current multipole expansion- the concentric layers and also with a homogeneous current quadrupole. In the present work the current model that matches the digitized internal surface sources are supposed to be active in a homogeneous of the skull. These models, which include effects half space. The analysis of the equivalent current of volume currents, show considerable improvement quadrupolar term contribution to the magnetic field relative to Biot-Savart's approach and are capable is presented.The attempt to localize the source of of localizing the implanted dipole within several bioelectric activity from the measured biomagnetic millimeters. field (MCG and MEG) will be also shown.

t/P'

1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas 33 Medical Physics Scientific Papers

MP8.7 MP8.8 Traveling Source Model of ABR Electric Distribution of the magneticfieldassociated and Magnetic Fields, S. Ueno*, T. Tanabe, with P"300, Z. Dunajski1,6C. Massen, E.v.d. K. Iramina, K. Harada, S. Wada, and S. Voorden, P. Elias and C. Brunie Matsuoka, Dept. of Electronics, Kyushu University of Technology Eindhoven, University, Fukuoka, 812, Japan, and Postbox 513, 5600 MB EINDHOVEN, the Dept. of Neurosurgery, Univ. Occupational Netherlands. and Environmental Health, Japan The P-300 wave present in elctrically measured The origins of auditory evoked brainstem responses evoked brain responses is of great interest in (ABRs) have not been fully understood. In the psychological research. Until now present study we investigate a possible mechanism very limited knowledge excistrabout the magnetic field elicited for ABR. A source model of ABR is proposed, which from task relevant events and about the possibility .simulates, in part, ABR waveforms. It is assumed of applying a magnetic method for investigation. that ABR signals are generated by actionpotentials Measuring possibilities and conditions ofthe which propagate along auditory nerves.Pathways of magnetic investigation of P-300are studied in auditory nerves are represented by a combination of six curved lines which involve both ipsilateral some details. The distribution of the magnetic and field around the head, evoked by auditory contralateral pathway... stimuli The sources of ABR signals in the oddball paradigm is measured are modeled by a twin-dipole which travels along using a rf- SQUID with a second-order gradiometer the curved lines. in an un- The twin-dipole consists of two shielded environment. The frequentl oppositely oriented dipoles which kHz and infre- are located quent (target)2 kHz tonebursts within a distance apart from each other. were presented at The 80 dB SPL with probability ratio effect of the distance between two dipoles 5:1. Reliable on ABR recordings of the magnetic P-300 waveforms is also focused in the present wave have been study. obtained. The results show that magnetic Based upon our source model, the waveforms measure- and ments of P-300 provide further information spatial distributions of electric potentials about and sources of underlying processes. The magnetic magnetic fields of ABRs are calculated.The P-300 may be either positive periodic patterns in ABR waveforms or negative depending are simulated on the measuring position above the head. well by the model. Its The component waves in ABR are shape and latency are also explained by the change of dipole orientation. position-dependent and reflect the involvement of active The spatial pattern of wave V is also simulated sources. well by the model.

MP4.1 MP9.2 Digital Tomo-synthesis By Inverse DIGITALLY ENHANCED MAMMOGRAPHY: Filtering, A.Iwata*, H.Matsuo, TV I.Horiba, ,H.Wei, N.Suzumura,and SCANNED FILM VERSUS T.Sawada, Nagoya Inst. of Tech., LASER ACTIVATED IMAGING PLATE Showa-ku, Nagoya, 466 JAPAN C. Kimme-Smithe, L. Bassett L. Gormley, R. Gold Dept. of Radiological Sciences, UCLA Conventional X-ray tomographic technique School of Med. Los Angeles, CA90024 withX-ray film provides a sagittal imagebyone scanning of measurement. OM/Min R film-screen images oftest objects Digital Tomosynthetic technique, however, containing simulated calcificationsand images of can produce sagittal images at breasts of 40 women (25 with pathology any arbitary slice and 15 of the body. normal) were scanned by This paper proposesa new reconstructing a Damon Mammography methodfor digital Enhancement System. The same test objects and Como- synthesis. It breasts were also imaged with FUJI high utilizes a 3-dimensional convolution resolution plates and laser activated bya Philips Computed process with the inverse filter function, which is derived analytically Radiography (PCR) system (.100mm pixel size). The images produced by the two methods from the point spread were function compared. The PCR system was found to be superior determinedby the projectiongeometry. for imaging dense breasts becausethe H and D Unstabilityof inverse filter hasbeen eliminated considering the curve and overall film density could be localization manipulated, while the Damon system of frequency component ofprojectional was found to images. By be superior for examining thesimulated micro- the phantom experiment, it calcifications. has been shown that The Damon system, however, was the proposed unable to compensate for improper algorithm is able to film exposure, reproduce the so that dense breast parenchyma below .45 expectedgray level corresponding with O.D. and X-rayattenuation photon-poor on film remained noisy whenenhanced. coefficiets on the Imaging with the FUJI plates required slice. Itcan also reduce the blurred half the exposure of the film-screen images. Out of 10 images from outside of focusedslice. cases with calcifications, threewere not well represented on the FUJI plates. 56

34 1988 World Congress on Medical Physics andBiomedical EngineeringSan Antonio, Texas Medical Physics Scientific Papers

MP9.3 MP9.4 A Novel Digital Radiographic System, Description of a Three-dimensional Motion UsingSelenium Platesand Laser Correction Algorithm for DigitalSubtraction Scanning Readout, J.A.Rowlands* and Angiography. D.R. Pickens*, J.M. Fitzpatrick, V. D.M.Hunter, R.B.Holmes Radiological Mandava and J. Grefenstette, VanderbiltDept of ResearchLaboratories, University Radiology, Nashville, Tennessee 37232 of Toronto, Toronto, Ontario, M5S 1A8Canada andN.Araj, Noranda Patient motionwhich occursbetween a mask image Research Centre, Pointe Claire, and a contrast image is often a source of arti- Quebec, Canada. facts in digital subtraction angiography (DSA). We have been evaluating a method forthe automatic registration of selected mask and contrast sub- A new imaging system for radiography basedon imagespriorto subtraction which isable to charged photoconductive plates with read-out by remove orreduce the effects of three-dimensional a scanning laser beam is described. It is motion. The technique uses an eight term poly - shown theoretically that this system should nomial, whose coefficients are determined automati- have an exceptionally good spatial resolution cally to stretch a mask subimage, resulting in a as well as an adequate signal-to-noise ratio reductionofartifactsaftersubtraction. The (SNR). These expectations are confirmed with a polynomial coefficients are determined by a heu- prototype systemdesigned in our laboratory ristic search method whichseeks to minimize the which has shown a modulation transfer function absolute value of the difference between mask and of 50% at 5 1p/mm; 20% at 10 1p/mm and a SNR of contrast subregions. A factor, the Jacobian, is 1000:1.A clinical system is beingdeveloped used to alter the gray level values ofthe mask which willbe suitable for the most demanding subregiontocompensatefor stretching or con- radiological imaging tasks currently undertaken tracting. There is improved visualization of ar- withscreen/film combinations. Because it is terial segments in subtraction images of phantoms digital new tasks impossible with current and in clinical DSA images dueto artifact reduc- screen/film combinations will be feasible. tion whencompared tothe results of subtraction Evaluations ofthe system using the Wiener without motion correction. spectrumwill be used to compare this method with stimulable phosphor systems. Supported inpart by NIH 1-R01-HL33951, by NSF #ECS-8508588, and the Whitaker Foundation Equipment Grant.

MP9.5 MP9.6 Noise reduction in dual energy imaging, Post Imaging Contrast Enhancement in Digital Mammography, P.J. Bjorkholm* and M. Annis, W. Kalender, Siemens Medical Systems, American Science and Engineering. Inc., D-8520 Erlangen, West Germany Cambridge, MA 02139, D.B. Kopans and J.M. Taveras, Massachusetts General Hospital. Boston, MA 02114 An algorithm for noise reduction in dual energy radiographic imaging using A high spatial resolution detector with good X- ray absorption characteristics over a wide the basis material decomposition energy range has been used to obtain mammograms technique is presented; it is based on on a number of volunteers. Because the digital images can be contrast enhanced after acquisi- exploiting the negative correlation tion, the X-ray spectrum utilized in imaging is between the noise in the soft tissue not constrained as in conventional imaging which requires a low voltage to achieve an acceptable image and the noise in the bone or calcium contrast. Low voltage usually results iodine image. The algorithm will be in high contrast for micro calcifications and a limited dynamic range. The new digital system, derived by physical considerations and Microdose Mammography. utilizes a 70 kVp tung- be demonstrated using simulated data. sten anode spectrum. While the transmitted image has a lower intrinsic contrast for the Typically, pixel noise is reduced by a calcifications. the contrast enhanced images are factor of 2-5 without an increase in presented with nearly identical contrast for micro calcifications with excellent dynamic dose. Results of clinical studies are range. Because of the higher voltage spectrum presented both for scanned projection on the peak dose to the breast is of the order of 30 times less with the Microdose system as digital radiography and for hybrid DSA. compared to standard film screen techniques, while achieving comparable spatial resolution and superior contrast resolution. 57 1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas 35 Medical Physics Scientific Papers

MP9.7 MP9.8 Non-recursive Parametric Filtration of Incorporation of Various Physical Factors in Medical Images for Scatter Rejection, Statistical Image Reconstruction Methods,C.-T. J.A. Seibert* and J.M. Boone, University Chen, The University of Chicago, Chicago, of California, Davis, California 95616 Illinois, U.S.A.

Recent developments in statistical theory andassociated X-ray scatter is a major degradationin 2-D computational techniques have openeda new avenue for transmission medical images. Using a parametric solving problems In image reconstruction.Mathematical technique developed in our laboratories, a models based on statistical theorycan be built to incorporate scatter point spread function (PSF) distribution every physical factor involved in image formation. Various is measured, based upon imaging geometry, object estimation methods, e.g., maximum likelihood,Bayesian, thickness, and presence of anti-scatter devices etc., can then be employed to derive iterative algorithmsfor (grids). The PSF is parameterized as a modified computing estimate of the original object.In principle, symmetric Gaussian function in terms of a improved image qualitycan be achieved by using these scatter fraction and distribution width. statistical methods becausemore faithful and more complete Transformation of the PSFinto the frequency mathematical models can be employed. Inaddition, since domainpermitsinverse filteringofFourier stationarity is not a necessary conditionin the derivation of transformed digital imagesto non-arbitrarily the formulas employed in these algorithms,some common remove the scatter component. Inverse shift-variant characteristics can be compensatedfor as well. transformation provides the scatter corrected Various implementation strategies have beeninvestigated for image. incorporating physical factors suchas photon attenuation, scattered radiation, system spatialresolution, etc., in We describe methods that allow theuse of such a emission computed tomography. Variousmathematical technique in a clinical environment, basedupon models that describe these physicalprocesses have been input variables such as field size,air gap, formulated for generation of relatedprobability density photon energy and patient thickness. Improved functions. Computer simulationstudies have been image contrast and quantitative accuracy for performed to evaluate these strategies interms of resulting densitometric measurements, with minimal image quality and computational requirement.In general, negative effects such as increased image noise images reconstructed by these statisticalmethods have or artifacts are obtained. Implementation as an superior quality at the expense ofmore processing time, optional processing scheme has been acomplished when compared to those obtainedby conventional in the clinic on our digital angiography unit. convolution methods.

MP9.9 MP10.1 Focused collimator for scanned projection Total Body Irradiation with 10MVxrays, digital radiography, IL Sones, Mt. Sinai S.N. Rustgi*, G. Popescu, J. Rodgers and J. Medical Center, Cleveland, OH 44106. Torrisi, Department of Radiation Medicine, Georgetown University Hospital, The fan beam collimator typicallyused in Washington, D.C. 20007 scanned projection digital radiography consists ofa single slit positionedroughly midway We are participating in a protocol using total between thex-raytubefocal spot and image body irradiation (TBI) for immunosuppression receptor. To produce a thin fan beam with this prior to bone marrow transplantation for adult type ofcollimator, a small focal spot mustbe leukemia. This paper details the dosimetric used, and focal track loading becomesa dominant considerations of using a 10MV xray beamat an system constraint. Furthermore, the slit extended distance of 5 m. We have febri%.ated a collimatorgenerates a fan beam whose thickness patient support and a lung shielding block increases with distance from the slit. For most support system, which minimizes patient move applications this divergence is undesirable. ment during AP/PA treatment and helps position the lung blocks accurately.The patient stands 30 cm behind a 1 cm thick lucite sheet to A focused collimator design isproposed which, enhance the skin dose to 93% of the dmax dose. for given fan team thickness, permits theuseof a significantly larger focal spot and provides Nondivergent lung blocks, 1 HVL in thickness, are placed at a distance of 30 cm from the 2-5times greaterx-rayfluencethan a slit patient midline and their position checked collimator. The focused collimatorconsistsof daily by port filming. The chest wall under a set of thin, closely spaced, radio-opaque vanes oriented nearly parallel the lung blocks is boosted with electron beams to the central of appropriate energy. planeof the fan team, but slightly angled to The measured TMR values at 5 m were found to be within 2% of the focus at a point within the central ray. This values obtained at the nominal SAD of 100 geometry allows non-diverging or even converging cm. We have also performed in vivo measurements of fan beams to be generated. A mathematical model of a focused collimator entrance and exit doses on a number of TBI isdeveloped, and patients using a calibrated silicon diode potential implementations are discussed. detector. Aspects of this treatment approach and dosimetry will be presented. 58 36 1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas Medical Physics Scientific Papers

MP10.2 MP10.3 Dosimetry of lung and bone for total A Technique for Hemibody Skin body irradiation:Comparison of 18 MV to Co" Skin Irradiation, D.L. Wilson, Dept. of and 8 MV photon irradiations Radiation Oncology, University of Louisville Brown Cancer Center, Louisville, KY 40292 Ellen El-Khatib., Jurek Rymel, F. Mahyoub Biomedical Physics Dept. A new technique for electron beam skin KFSH&RC, Riyadh 11211, Saudi Arabia irradiation of mycosis fungoides or Raposi's sarcoma patients is presented- A In most centers total body irradiation (TBI) is uniform skin dose is delivered to either performed with either Co" or X-rays of energy the upper or lower half of the body, while below 10 MV. At the King Faisal Specialist sparing the opposite half. The treatment is Hospital And Research Centre (KFSH&RC) an 18 MV done with a Therac 20 linear accelerator. X-ray beam is used for TBI. This higher energy The patient lies on the floor and assumes beam produces a more homogenous dose distribution six different orientations which correspond to the body for parallel opposed beams while to the six standing positions normally used giving a lower radiation dose to the lung. Lung in the standard total skin technique as toxicity is a limiting factor in TBI. Most of described by the AAPM TG 30 protocol. A the data available on lung and bone radiation 0.6 cm lucite sheet is used to lower the dose is for photon energies in the range 1-10 nominal 6 Melt beam energy to 3.9 Mev. The In the present work radiation absorbed dose in untreated portion of the body is blocked by and beycnd bone and lung in tissue equivalent lead sheets. Two adjacent fields are used phantoms and also in a humanoid phantom was to achieve beam uniformity. The separation measured and calculated. Results were compared of the fields was determined by fitting the for 18 MV, 8 MV, and Co" large field irradiat- beam to a gaussian curve and looking at the ions. The TBI technique and dosimetry are two field combination that gives the most evaluated in terms of dose homogeneity and a uniform dose over one meter. Calibration comparison of bone and lung dose for various factors were derived from ion chamber and photon energy irradiations is presented. TLD measurements made 'according to recommendations from the AAPM TG 30. All field parameters,calibration and phantom measurements are presented.

MP10.4 MP10.5 Sweeping, an universal technique in The design and evaluation of Jung compen.,ators Total Body Irradiation, R.G. MUller, for total body irradiation treatment, C. Simms*, Institut fUr Radiologie, Universitat, L. Doyle and J. Conway, Department of Medical D-8520 Erlangen, F.R.G. Physics and Clinical Engineering, Weston Park Hospital, Whitham Road, Sheffield, S10 2SJ., Englana. Sweeping is a technique without the need of a spe- A total body irradiation (TBI) technique, for patients requiring bone marrow transplant, has cial room design. The patient lies in supine and recently been developed to allow a fractionated prone position on a low couch on the floor. The regime rather than the previously employed single beam direction is along the shortest diameter, dose treatment. No bolusing of the patient requires the design of customized tissue compen- ventral to dorsal. The required accessories as sators for dose uniformity to the body. This moving flatness filter defined lung shielding paper describes the methods used to provide a thorax compensator to correct for increased lung blocks, and beam attenuation will be demonstrated. transmission and tissue obliquity. A 'beams-eye- The midplane dose is calculated to 12 Gy (6 frac- view' facility developed for the Picker Treatment Planning System allows the precise identification tions, 3 days). The lung dose is reduced to 8 Gy. of compensator shape required with respect to the patient's arm, which acts as the compensator for The underdosage of the chest wall is compensated the posterior section of the lung. Analysis of by electron beams with matched energy and bolus the resulting 3D dose distribution is obtained using 'dose volume histograms'. material. The maximum dose rate is reduced to 0.2 Gy/min Measurements taken in a thorax 'Rando' phantom, using thermoluminescent dosimetry (LiF powder) will at midplane, and the effective dose rate in the be presented. lung is about 0.07 Gy/min. Patient positioning and compensator alignment is discussed with regard to dose uniformity for this TBI method. t

1988 World Congress on Medical Physics and Biomedical Engineering 0 San Antonio, Texas 37 Medical Physics Scientific Papers

MP10.6 MP10.7 Hyperfractionated T.B.I.:results and dosimetric The Physical Aspects of Hyperfractionated considerations after three years experience. Total Body Photon and ElectronIrrediation G.Gobbi.,C.Raymonde,M.P.Leogrande,S.Beneventi, for Bone Marrow Transplantation.Azam Niroo- S.Ricci and P.Latini. Fisica Sanitaria and Radio- mand-Rad*, Michael T. Gillin.Dept. of Radi- ation Oncology, Medical College of terapia*.Policlinico Hospital - Perugia- Italy. Wisconsin 8700 W. Wisconsin Ave. Milwaukee,WI 53226 USA 48 patients received hyperfractionated T.B.I.(1.2Gy Total body photon and electron 13 fractions 3 times a day, total dose 15.6 Gy). irrediation(TBI) involves delivering an adequate immunosuppressive do- A modified 60-Co ,snit with a horizontal beam anda 2 se before transplantation to acheivea successful 190x125 cm field is used,the dose-rate being about marrow graft and to eradicate the mallignentcells 70 mGy/min at 226.5 cm source-midline distance.The while avoiding paeumonities andtoxicity to organs patients lie supine for the lateral fields andon such as lung and liver. To acheivethese goals a hy- their sides for AP fields, the legs slightly bent. perfractionated regimen for TBI withpartial lung and liver blocks have been adopted. In the supine position a compensation bolus is Chest wall areas which have been shielded by lung applied to the neck and the arms are alternatively blocks are boosted with electron beams. CTscans are used to select el- raised. The most uniform dose distribution is ectron beam energy and to calculate lungdoses. A achieved with 9 AP fractions and 4 lateral fract- convetional linac in a standardtreatment room is ions (inhomogeneity indexes: AP=3.2, lat.=6.9, utilized to deliver 18 MV X-Raybeam at a dose rate 9 AP+4 lat.=2.6). In the treatment planning the of 20 cGy/min at 350cm treatment distance. A special chair is used to treat the patient AP/PA TPR for very large fields is used taking themean in an upright sitting position with the legs bent thickness of 3 transverse sections: pelvis,umbili- backward. The dose to lower extremitiesare monitored and raised cus and nipples. During the first 2 fractions TLD's to the total body dose. Following theguidlines re- are placed in 7 body sections and midline dose dis- commanded by AAPM Task Group No.29(1),the physical tribution is evaluated from combined entrance and aspects of TBI is described. The dosedistribution and the dose uniformity exit doses, thus allowing treatment time correction are presented. Details of The mean ratios of measured/calculated doses this scheme, in-vivo and in-vitrodosimetry, boosting are: techniques and lung dose calculationsare 0.919 in AP and 0.970 in lateral irradiation. presented. (1)AAPM Repot No. 17: The PhysicalAspects of Total and Half BodyPhoton Irrediation(1986).

MP10.8 MP10.9 Analysis of Short FSD Techniques for Total Skin Energy Degradation of Low-Energy Electron Irradiation, D.I. T.hwaites, Medical Physics, Electron Beam for Treatment of the *stern General Hospital, Edinburgh, Scotland, U.K. Superficial Malignancies,W.Sewchand, M. Sarfaraz *, N. Bautro, Rad. Therapy Total skin irradiation is the treatment of choice for certain Dept., Univ.of Maryland Hospital, conditions. Long fsd, high-output techniques are generallyus- Baltimore, MD 21201 ed, with multiple beans or patient rotation for uniformity. Problems arise if this cannot be achieved due to machine or Energydegradation of conventional low- roan limitations, or if the patient cannoteAd for the tre- energy electron beams, by delibrate atment tine involved. Short fsd scanning or tracking has been placement of degraders in the path of used, but the moor four fields generally employed inevita- the beams is often used in Total Skin bly reduce uniformity. Electron Beam (TSEB) therapy atextended Cur technique uses a short fsd (.150 cm) and large fields Source-Surface Distance (SSD). When aided along the patient. Using a specially designed applica- local areas require booster doses,it is tor, distributions added side-by-side can be made essenti- desireable to adminster it at standard ally flat. A four-position technique is used with energies of SSD with a degrader that produces 2- 4 Mr, as required. Depth doses and distributionssere similar dose characteristicsas TSEB. measured in detail in phantoms and checked on patients. They Extensive measurements were carried out depend heavily on cross section and number of fields. For a to study the effect of various basic four-field approach under these conditions, distribu- thicknesses of polystyrene degraders on tions around a trunk section typically average 130- 140Z of a conventional low-energy (6MeV) given dose on equally ueighted beams, with a 120% variation. electron beam at a standard (100Cm) SSD However the technique allows gantry angulation, to increase for five different fieldsizes. The study the effective number of fields, achieving spreads within 107.. evaluated the effect of degraderson: 1) To correct for variations with cross-section, longitudinal energyand depth dose characteristics, field edge positions can be selected and different weighting 2) surface doses, 3) field flatness, 4) applied at each level depending on average cross-section. bremstrahlung component, and 5) the The technique is more complicated and time-consuming to set output for certain monitor setting. The up than typical long fsd treatments, but its flexibility results were analyzed and significance allows very acceptable distributions to be achieved in of the study will be discussed. reasonable machine tines. 60

38 1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas Medical Physics Scientific Papers

MP10.10 MP10.11 Total skin electron arc technique Modification of an Intraoperative Radiation for disabled patients TherapyApplicator System, N. Samsami*, F.R. Bagne, Medical College of Ohio, Toledo, Ohio. B.J. Gerbi*, F.M. Khan,F.C.Deibel,and Earlier studies have shown that the intraoperative T.H. Kim. University of Minnesota, Dept. of radiation therapy applicator system developed Therapeutic Rad., Minneapolis, MN 55455. at theMedical Collegeof Ohio can produce a clinically significant dose nonuniformity across Mostof thecurrent techniquesused for theradiationfield by as much as 18%. This total skin irradiation using electron beams pre- dose nonuniformity is caused by radiation suppose that the patient is capable of contamination from different components of the maintaining a standi, position for the duration system. We have developed a technique for of the treatment. For patients either weakened eliminatingthisundesired dose ron-uniformity. bydisease orthose suffering from aloss of In this study, the scattercontributionsfrom limbs, this is often an unrealistic expectation. different components of the system are Wewilldescribe atechniquethat allows the differentiated. Energy-specificringsarethen patient to remain in a reclined position without designedwhich fitat the gantryend ofthe sacrificingdoseuniformity. This technique applicators. The materials usedare aluminum, employs two,symmetric (+48 from the vertical stainless steel and copper, all of whichcan axis) pendulum-arc electron beams to provide a be cold-gas sterilized. To determine the optimum field uniformity of +5% over a range of -220 cm thickness, a series ofmeasurements are done along the length of the patient. The standard using rings of varying heights and thicknesses. six patient positions [1] are employed to provide All data aretaken inwaterusing 6, 9, 12, a uniform dose to all surfaces of the patient. A 15 and 18 MeV electrons. The applicators employed description of tne *reatment technique along with are circular in cross-section with diameters details of the dosimetry will be given. ranging from 1" to 4" and have eitherflat, 15°, or 30°beveled end. This paper presents 1. V. Page et al., Radiology 94:635 (1970). the resultsofourstudy and recommends the ring height, thickness and material for various electron energies and applicator sizes necessary to provide a uniformdosewith less than5% variation.

MP10.12 MP10.13 A Method for Shaping Arbitrary Fields for A Study of the Factors Affecting the Perform- Intraoperative Radiotherapy(IORT) , P. ance and Use of Intra-Operative Radiation Biggs*, W.Shipley and C.Willett, Mass. Therapy Applicators, P. Stafford*, J. Chu, Genl.Hosp.Cancer Centre, Boston, Ma USA G. Hanks, Univ. of PA and Fox Chase Cancer Center, Philadelphia, PA. A technique has been devised for arbitrarily shaping electron fields in IORT. This has been achieved by cutting a 15mm thick lead disc, Intra-operative radiation therapy is now in which fits inside the base of an 8cm diam. routine use at many facilities, both dedicated circular cone and which is attached through a and non - dedicated in nature. The applicator series of holes around itsperimeter.The generally consi...s of a holder/periscope and desired shape of the blocked area is determined a series of electron cones of various diameters by the therapist at surgery, using a dummy cone. apd angulations. This paper investigates one After cutting the disc, the cutout is sterilized such system (Radiation Products Design, Inc.) and taken to the therapy room. This lead piece for the effects of: a) less than perfect is attached to the cone by one or more screws alignment of the viewer-cone assembly; b) the before the cone is inserted into the patient. presence of a suction tube or other surgical The dosimetry for two shielding arrangements has apparatus in the cone; c) field size setting been tested:(i) a single straight edged piece, of the photon jaws; and d) the behaviour of varying in width up to 4cm and (ii) a circular outrut factors with distance from the nomlnal cone down, varying in diameter from 7 to 4cm. 100 cm treatment distance. Both radial and These sizes represent the extreme limits of any depth profiles will be shown illustrating likely blocked fields. As expected, the %depth these effects, as well as graphs for field dose decreases with field size (independent of size and output factor behavior. field shape) and increasing energy (up to 3mm at 9 MeV, 8mm at 15 MeV and 17mm at 23 MeV) as does the location of dmax, while the surface dose also increases (up by an average of 5% for (ii) and -1.5% for (i)). The 'cone ratios' show an increase of -3% for 9 MeV, 3-5% for 15 MeV and 4- 9% for 23 MeV for (ii) and less than 3% overall for (i) 0-iri 4

1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas 39 Medical Physics Scientific Papers I

MP10.14 MP10.15 MRI Stereotaxis:An Indirect Method of Recurrence vs radionecrosis:anindication for 30/ Transposition of MRI-Generated Datainto Multiplanar Imaging-guided stereotaxis. CT Multiplanar Images Zamorano L,* Dujovny Zamorano Lic,Kataaik D., DujovnyM.,Yakar D.,Malik G M, Malik G, Mehta B - Heetnryu., rdEaEard Hospital Henry Ford Hospital, Detroit,MI48202 and nonlinearities induced by eddy E "ternal RT has proved to be currents during the pulsesequence distort images an important adjuvant in the treatment of malignant and produces displacement ofcoordinates in both gliomas. Its effect on survival is dose-dependent. Intents the X-Y plane and Z axis. Correctionof geometric to localize the field of high dose RT to the distortion requires special algorithms,phantom tumor area have been performed with the aim to measurements, shim adjustment, etc. Amethodology spare damage of normal brain tissue. Between 8/83 of indirect transposition of MRIgenerated data in- to 1/87, 33 patients with malignant astrocytoma to CT Multiplanar stereotaxis ispresented, which underwent high dose louses available software from MRI (GE Signa 1.5T) calized hyperfracctionated externalRT after surgery. Patients received 57.6 G./ to the and MRI laser beam to establishthe plane of place- tumor+edema area and a localized boost of 7.4, 14.4 ment of stereotactic frame. Axial, coronalandsag- or 24 Gy. In the follow-up, 16 patients died ittal 11, T2 weighted imagesare displayed. Xcoor- with evidence of CT/MRI increase of size of lesion. dinates are selected from axial andcoronal cuts.Y Since 7/87 11 patients have undergone from axial and sagittal and Z from allthree cuts. multitrajectory serial stereotactic biopsies: 4 patients Trajectories are determined and plottedin similar were "real" recurrence of tumor and 7 were radiation necrosis. ways. Areas and volumes of region of interestare Recu- rrences were treated with 1-125 interstitial determined. Stereotactic frame is placedon the RT. Radiation necrosis were treated predetermined region. Frame is isocentricto the CT with corticoides and diuretics, obteining partial gantry and parallel to the scanningplane. 1.5mm or complete remi- ssion of symptoms and decrease thickness slices of the region ofinterest are or disappearence of CT lesions. Should these localized performed. Error factors for MRIgenerated data are external boost be replaced by 3D stereotactic calculated for each cartesian plane. Thisfactors interstitial boost can be introduced on computer program that auto- after surgery and conventional externalRT? Advan- tages would be: higher intratumoral matically will correct any MRIcoordinates. These dose, less da- can be plotted on axial, coronal, sagittal and mage of sourrounding brain tissue andmore favorable radiobiological conditions. paraxial reformatted images and AP andlateral scoutviews. Methodology detailswill be presented.

MP10.16 MP10.17 A high energy photon collimatorfor stereotactic irradiation with Some Physical Problems Associated with a linear accelerator. D.A.Mazal, J.C.Rosenwald*, J.J.Baron. Combined Electron and Photon Fields in Institut Curie- 26,rue d'Ulm. 75005.Paris. France. the Treatment of Mesothelioma, M. Soubra*, P. Dunscombe, D. Hodson Small size, high energy X-ray beamsfrom a linear accelerator and G. Wong, are used to perform stereotacticconvergent irradiations of Manitoba Cancer Foundation, small volumes in the brain. Additionalcollimators have been 100 Olivia Street, Winnipeg, Manitoba developed to produce small circularbeams, well adapted to CANADA this technique, and to ensure theaccuracy of their dimensions and positioning. The collimatordesign is based on the following criteria : precise beams diameters, 5 to 40mm at Combined fieldsofelectrons and photons can be isocenter, without introducing modificationson the accelerator used effectively to treat the pleural membrane of and maintaining a unique apertL. a of the main collimator; patientswithmalignant mesothelioma. Various adjustable direction and position,with the possibility of problems may be identified in the implementation changing collimators without modifying these two parameters ofsuchatreatment strategy: and without moving the patient; the non-uniform dose transmission factor and the distribution in the join-upregionof the two primary collimator opening (where thebeam is attenuated only fields; by the additional collimator) should increased electron penetration into lung; be minimized, as well as the increased lungdose from photon side-scatter and collimator weight and beam contamination, for all the energies the effectof the ribs on both the electron and under consideration; strong, easy to manufacture from cheap thephotonfields, are among the major problems and available materials, easy to manipulate. encountered. The modular design we have chosen To investigate the significance of gives independent theseeffects functions (fastening, micrometric a phantom simulating, under simple deplacements); 10 cm lead geometrical ensure less than 1% transmission for conditions, bone, lung andsoft a 5X5 cm2 field, tissues has been constructed. measured in the range 6 to 25 MV; Dose distributions a brass cylinder provides under mechanical protection and allows various simulated clinical conditions, have strong and precise beer measured and compared with calculated data. attachements. The distance from the collimator base to the The significance of the effects studied for the isocenter (28 cm) ensures enoughfree space around the treatment of pleuralmesnthelioma is discussed. isocenter and avoids beam contamination.Total weight of the system is 14.3 kg, with 7.8 kgcorresponding to the exchangeable collimator. Two removablecross hairs are used for precise centering. 62

49 1988World Congress on Medical Physics and BiomedicalEngineeringSan Antonio, Texas Medical Physics Scientific Papers

MP10.18 MP10.19 A Modified Technique of NeckIrradiation in Radiation Therapy Techniques ID Head and Neck Carcinoma: Its Effectiveness, Breast Cancer, J.S.B.Onnnseknr, R. Datta*, F. Ampil and S.Datta, LSU Medical V.Dilip Kumar, S.J.Supe, J.B.Sesane Center, Dept. ofRadiology,P.O.Box 33932, Dept. of Radiation Oncology, Jaslok Shreveport, Louisiana 71130 Hoepitalaombey 400 026, India,

In the treatment of head and nee carcinoma, the In 183 breast cancer patients, various radio- junctionalareabetweenthe parallelopposed therapeutic techniques were used, with the aim lateral upper neck fields and the anterior lower of prophylactic, radical and palliative mea- neck/supraclavicular neck field poses a chal- sures depending upon the stage of disease. lenging problem. The major concern involves the For the treatment of chest wall irradiation, desire to adequately treat the entire area at tangential portals were used with open fields, risk of recurrence without exceeding 'tolerance' breast appticator and wedges and the patients of the spinal cord. To overcome this diffi- were grouped accordingly. Comparison of these culty, a modified technique of irradiation has techniques are done by (1)plotting of !modes* been utilized at the Louisiana State University distribution (done .anually because of non Medical Center since 1980. This report retro- availability of computer) of the techniques in spectively evaluates the effectiveness of this two chest wall contours (one with chest wall technique in 47 patients given adjuvant post- alone and the other Breast and Chest wall and operative external beam irradiation during 1980- (ii) Observing the radiation sequels° in these groups of patients during and after the comp.. 1984. Actuarial four-year disease-free survival was 51 percent. Recurrentdisease was not letion of treatment. observed at the junctional blocked (and under- With the advent of various concepts like CRC, dosed) area of irradiation in any of the treated TDF and T3D,attempts have been made in various patients. Neitherwas therean example of centres to compare the treatment plane based radiation-induced myelitis. We feel that this on isoeffect lines using these concepts.Tn modified partial block radiotherapy technique is the present study isoresponse Wee using useful, provided evident disease is not being these concepts for radical treatment techniques blocked. Description of this technique, its were obtained and compared. dosimetric aspects and its effectiveness will be highlighted.

MP10.20 MP10.21 Matching of Orthogonal Fields in Cranio- Orthovoltage Therapy with X-ray Contrast

spinal Irradiation, M. Tatcher,* Quality Media for Brain Tumors, K.S. Iwamoto , Assurance Review Center, Rhode Island M. Wollin , A. Olch , J. Bellotti , A.R.

Hospital, Providence, RI 02903. Kagan , and A. Norman*, Departments of Radiation Oncology, UCLA, and Radiation Patients with malignancies that require cranio- Therapy, Kaiser Permanente, Los Angeles, spinal irradiation (CSI) are usually treated with California. parallel-opposed lateral cranial fields coupled to a long posterior spinal field. To improve dose uniformity at the junction, many radiotherapy pro- Orthovoltage x-rays have two potential advantages tocols specify compensation for beam divergence by over conventional high energy photons for the rotating the cranial fields so that their inferior treatment of brain tumors. One, if the tumor is edges follow the projection of the superior edge of first loaded with x-ray contrast media, the high the spinal field. In this method, when the borders photoelectric absorption in iodine makes it of the orthogonal fields are matched in the mid- possible to enhance the tumor dose (Med. Phys., sagittal plane, there is no gap at the junction of 10,75-78, 1983; Radiother. Oncol., 8,161-170, the fields on the posterior surface of the body. A 1987). Two, the low exit dose spares normal review of treatment charts from institutions that brain tissue on the far side of the tumor. The participate in Jinical trials involving CSI shows disadvantage of large entrance and bone doses can that some departments separate the cranial and spi- be reduced by using multiple fields. We have nal portals by up to 1-2 cm when treating by a di- measured the dose distributions from orthovoltage vergence compensating technique. To study the x-ray beams in phantoms and used these to con- dosimetric effects of such gaps, we have calculated struct treatment plans. The results show that dose distributions in the junction region for typi- multipl, orthovoltage fields, when used to treat cal CSI treatments by cobalt and 4 MV photons. selected tumors loaded with contrast media, have Stationary and moving junction methods were con- potential advantages over conventional opposed sidered. The results demonstrate how the extent field cobalt-60 gamma rays.They also indicate and magnitude of overdosage and underdosage depend a potential use for this technique to boost the on the width of the gap. tumor dose after conventional therapy by treating either through the skull or through a bone flap. Supported by NCI Grant CA29511. 63

41 1988World Congress on Medical Physics and Biomedical Engineering San Antonio, Texas Medical Physics Scientific Papers

MP10.22 MP10.23 Radiation Dose To Radiosensitive Organs In Treat, Reduction of Gonad dose during treatment ment Of Medulloblastoma, F.R. Bagne , D.G. Bronn, of Testicular =burs. T.M.Kehoe and S. Hoke, Medical College of Ohio, Toledo, Ohio D.I.Thwaites *,Western General Treatment of medulloblastoma requires irradiation Edinburgh, Scotland. of the entire cranio-spinal axis. The alternating heat; positiontechnique is a methodby which The large field treatment of testiculartumours presents the the entire CNS of a child is treated in a single problem of reducing the dose to the regainingtesticle to an field with the patient in the prone position. acceptable level. Unfortunatelyan acceptable dose has never The head is turned to the opposite lateral been clearly defined and in general thephilosophy is to ke- position on alternative treatment days. A ep the dose as low as pracUcal. We have designeda new shi- separate cast block is made for each alternating eld to protect the testicle fran scatteredradiation on a head position which excludes the faceandthe 9MV linear accelerator. eyes. To determine the dose to radiosensitive The design had to allow the patientto rotate from supine to organs, TL dosimetry was done on patients. prone for parallel opposed fields. An analysis of theroutes Itwas foundthat withthestandard Cerrobend by which scattered radiation enters the testicleis presen- blocking, atotaldose of 3600cGy to the spine ted. The majority of the patientscatter enters via the tru- results in44cGy tothe gonads,380cGy tothe nk. To attenuate this component thesuperior face of the sh- thyroid, and 648cGy to the lens. A fractionated ield is placed between the testicle and the trunkand is 1cm dose above 400cGy is considered cataractogenic. thick. The anterior and posterior faces of theshield need To reducethe lens dose,supplementaryblocks to attenuate high energy scatter frcrn the field defining bl- were placed directly over the proximal eye. ocks and the collimators, they are 2cm thick. Ilelateral The total lens dose was then reduced to 308cGy. sides are 0.5cm thick for rigidity and the inferiorface is Gonadal doses above 250cGy are associated with left open. Ile anterior lid is =tile and lockable,and is temporary sterility. Our data indicate that positioned by the patient between the testicle andpenis the gonadal dose is not a concern with this effectively trapping the testicle beneath it. Thisreduces technique. The absolute threshold for the patient scatter contribution toa minimum The shield is hypothyroidism or secondary cancer of the tnyroid made of Pb and weighs 51b 1Cbz. 111) measurenentsreveal dos- has not been established. However, it is es of 50cGy to the testicle for a course of 3000cGy to mid- generally accepted that the thyroid dose should line is obtainable. be kept to a minimum. Our study indicates that the thyroid dose can be further reduced by 50% by means of supplementary blocks.

MP10.24 MP10.25 MINIBEABS AND THEIR APPLICATION IN CEREBRALS IRRADIATION. Maria Cristina Plazas A Mono-Isocentric Breast and Supraclavicular Irradiation Technique Using AsymmetricBeam In the last years, the development of Computer To- Collimation, U.F. Rosenow, AlbertEinstein mography (CT) and Molter Magnetic Resonance (I t) College of Medicine and MontefioreMedical imaging, have considerably changed the possibility Center, Bronx, NZ10467 to give treatment to cerebral tumors. These methods give the possibility to an early diagnosis of small A ney breast and supraclavicularirradiation lesions and to a move precise treatment of the lesi- technique is described whichuses a single iso- on with stereotactic methods. center for the tangential and supraclavicular fields. The therapeutic problem is to distinguish between The technique is especiallysuited for the tumor that "surgically accessible and theme that treatment machines with independentlymovable are posisioned in a non accessible topography. The collimators but can also be performedusing half- minibeams with high energy X- rays were proposed by beam blocking on conventionalmachines. The ad- different groups and the goal of this work is the vantages of the techniques are: dosimetry of minibeams of different diameters (5mm -accurate matching of the supraclavicularand to 20 mm diameters) and with energy of 60Co,6 MV, tangential beams 10 MV, 18MV, 23 MV, 25 MV x -rays. -use of a well-defined set-up point The following caracteristic have been considered: -simplified set-up 1. The dose measurement along beam axis (depth dose, -insensitivity to day-to-day set-upvariations TMR) with the different dosimetric methods. -coplanarity of beams, which simplifiesthe 2. The dose measurement off beams axis (profile, calculations of dose distributions scans, isodose curves). A dosimetric evaluation includingthe sensitivity 3. The evaluation of the different dosimetric methods: to day-to-day set-up variations isgiven. Some film, thermoluminescence, ionisation chamber and possible methodical variations ofthe technique diodes. will be discussed. The optimization of the dose 4. The collimating system. distribution, within the technique,to achieve a 5. The problem linked to lack of lateral electronic better dose homogeneity in thetarget volume and equilibrium and the existence of a possible virtual better lung sparing will also headdressed. source. 6. Monitor calibration. 7. Comparison with others methods of irradiation. 64 42 1988 World Congress on Medical Physics and BiomedicalEngineeringSan Antonio, Texas Medical Physics Scientific Papers

MP11.1 MP11.2 Energy Spectra an Angular Distributions The Effect of Charged-Particle Equilibrium on the of Electron Beams Scattered by Different Response of an Aluminum-Walled Argon/llopane Allan F. deGuzman and Phantom Materials, L. Andre* and Ionization Chamber. H. Neuenschwander, Dept. of MedicalRadiation R. J. Schulz, Yale University Physics, Inselspital, University of Berne The evaluation of the dose from neutrons in a CH-3010 Berne, Switzerland mixed neutron/low-energy-gamma field using an argon/propane ionization chamber requires a The geometrical acceptance angle of our spectro- careful analysis of energy deposition within the meter is 4.2 degres. Within this angular resolu- argon or propane from the low-energy gamma rays. tion we measure the differential fluenceof elec- Using a plane-parallel, guard23-field ionization cm plate trons from medical accelerators. The energy reso- chamber (10.2 cm in diameter, spacing), measurements made with 1-125 (effective lution (NE/E with 997 valleys) depends onthe width energy n 27 keV) revealed an argon/propane ratio of the variable entrance slit and isgenerally of of 15. As the charged-particle equilibrium doses the order of 3%. to argon and propane, which are obtained at the Td obtain angular distribution, twodifferent midplane of the present chamber, are in the ratio methods are used. For small angles we subtract the of 36.2, there is a clear suggestion ttt dose spectra measured with a small entranceslit from buildup (in argon) and builddown (in propane) those measured with a large one in order toelimi- occurs near the aluminum plates. To verify this, a computer program was written which tracks the nate the scattered electrons.For larger angles we asymmetric blinds. paths and energy deposition of secondary electrons use an entrance slit with The electrons are assumed to linear acceler- in the chamber gas. We have measured the spectra of the travel in straight lines with a range equal to A 45 (BBC) ator SL75/20 (Philips) and the betatron 0.70 times the CSDA range. The computational behind different slabs of water-,lung- and bone prediction of the argon/propane ionization ratio equivalent materials as well as without any is 39.9 which is in reasonable agreement with absorbing Laterial. The increase of spread in the experiment considering assumptions neccessarily angular and energy distribution of the spectrumand included in the program. the decrease of the most probable energywith This work supported by NCI grant CA-09259. increasing thickness of phantom materialwill be discussed.

MP11.3 MP11.4 The Use of Single Field Superpositions to An Experimental Determination of the Penumbra for Design Uniform TSET Dose Distributions. Multi leaf Collimation James M. Galvin*, Alfred Smith. D.P. Fontenla*, H. Weiss, Dept. of Medical Hospital of the University of Pennsylvania, 3400 Spruce St., Physics, Memorial Sloan-Kettering Cancer Philadelphia, PA. Ronald Moeller. Varian Radiation Division, Center, 1275 York Avenue, New York, NY 10021 611 Hansen Way, Palo Alto, CA. In Total Skin Electron Beam Therapy (TSET) the patient, located behind a 1 cm Multi leaf collimators (MLC) offer the interesting possibility of thick polystyrene screen, is irradiated directly shaping static treatment fields, thus eliminating the need at various angles of rotation by low lergy for cerrobend blocks. This study compares the penumbra for electrons. The purpose of the procedure fields shaped with MLC with the edges for standard blocks. The when exposures at each angle of rotation MLC was double-focused, following the divergence of the x-ray are weighted appropriately is to provide as beam, and projected a width of 1 cm at 100 cm SAD. The dose near as possible a uniform surface anddepth changes rapidly in1,12_,Q, dimensions in the penumbra region for dose over the whole skin. To determine the tl'e "stepped" MLC. We have used film to monitor this dose dis- effect of different weighting factors, on tri,Aimand will present a technique which corrects for the inac- the distribution of the dose around the curw.ies usually seen when this dosimeter is used for measure- phantom, a humanoid phantom has been wrapped ments in the penumbra region. An Optronix rnicrodensitometer at chest height with film at various depths. with a 400 micron spot size was used to determine dose distri- Dose distributions obtained from anterior, butions perpendicular to the beam center axis. The penumbra for posterior and lateral films from single MLC was scalloped at the beam edge as a result of approxi- angle exposures were mathematically combined mating the block as a step function. The measure of penumbra in different ratios by computer. The com- used was the 80 to 20% (and 90 to 10%) width of the beam bined distributions obtained in this way are profile at the edge. For the scalloped MLC edge, an "effective" compared with the dose distribution obtained penumbra was determined as the distance between the outermost when irradiations from multiple angles were peaks of the 20% line and the innermost peaks of the 80% line. made on a single film, by rotation of the Scalloping varied according to the exact shaping situation, but humanoid phantom, as is done in the actual was most severe when the block cut diagonally across the field treatment. This is a report on the appli- (the situation investigated here). The change in the penumbra cation and advantage of combining exposures width for the two field shaping techniques was determined at 5 to explore and facilitate the design of and 10 cm depths for a 15x15 field and a 6 MV accelerator. treatment. For 5 cm depth, the edge increases from 5 mm to 10 mm (22 mm to 24 mm) wide when the MLC was used. For 10 cm depth, the edge increased from 8 mm to 12 mm (31 mm to 34 mm) wide. 6J 43 1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas Medical Physics Scientific Papers

MP11.5 MP11.6 PERTURBATION CORRECTION OF A CYLINDRICAL THIMPLE-TYPE The Influence of calculation gridsize on the CHAMBER IN A GRAPHITE PHANTON FOR Co-60 GAMMA RAYS , C. E. de Almeida: A. M. Perroche Roux,and M. Boutillon, error In dose estimation, A. hliernierke, M. Goitein, BIPH F - 92310 Sevres - France Massachusetts General Hospital, Boston, MA02114

The determination of absorbed dose in graphite for 60Co Gamma The standard technique in which the dosedistribution is rays by means of an ionization chamber located in a graphite presented as an isodose map is basedon interpolation phanton requires the knowledge of the perturbationon the photon between doses calculated at a matrix ofequally spaced points. We explored the question of how the grid and electron fluence made by the chamber cavity and its wall if used for the dose matrix affects the the material or the density of the wall differs from that of the error in the interpolated value of the dose at any point and hencein the position of phanton. The necessary perturbation correctioncan be determined a calculated iscxhose contour. We defined thedose theoretically for a flat chamber(Boutillon 1983 and Niatel 1983). calculation error as a difference between theinterpolated However, for a cylindrical chamber(thimble type)the theoretical and true dose for a given point, anda position error as a calculation is difficult, due mainly to the geometry of the distance between that point and thenearest point which chamber. has, in fact, the same dose vaiue. We An experimental determination has been made of the examined the problem using an analytical and measuredbeam profiles perturbation correction to be applied to a graphite cylindrical of Co -60, X -rays and protons. Our analysis ionization chamber. showed that The results indicate that the magnitude of the interpolation errors areproportional to the correction increases with depth and that is also closely related curvature of the dose distribution, andare relatively to the size of the cavity. This makes possible the high in use of sgch a either side of, but not including, thesteepest chamber as an excellent standard of absorbed dose in graphitefor part of the penumbra. Our results showhow big an 60 Co gamma rays, as well as an appropriate transfer instrument. interpolation error one should expect fora given size of the calculation grid. The specification of Boutillon, M. Phys. Ned. Biol 28, 1983 pp 375-388, dose accuracy should be cast in the form of a pair ofrequirements, one Niatel, H - T. Phys. Med. Biol. 28, 1983, pp 407 - 410. as dose and the other as position. At a given point, only This work was undertaken at the Bureau 'International des one requirement need be satisfied. Theposition Poids et Measures during the tenure of an American Cancer requirement is almost always the lessdemanding in clinical practice and, if accepted, permits Society - Eleonor Roosevelt - Felloship awarded by the the use of larger grid spacing than if only International Union Against Cancer. a dose requirement is applied. As a rule of thumb. we recommendeda odd size . ; 01 10 , when that error is in the rant.1 of from 1to a few millimeters.

MP11.7 MP11.8 in vivo dosimetry during external photon beam Experimental treatment using diodes, S. Heukelom, H. Evaluation of the Hogstrom Lanson Pencil Beam Algorithm and B.J. Mijnheer*, The Netherlands in theVicinity of Cancer. Insti- Cylindrical Bones with 2D and 3D tute, 1066 CXAmsterdam, The Netherlands geometry. J. Antolak*, E. Mah, J.J. Battista,J.W. Scrimger Cross Cancer Institute, Edmonton,CANADA Silicon diodes are well suited for in vivodosi- metry if a number of characteristics, such as We have studied the influence of energy and temperature dependence of the dosime- small bone substitutes of' rious density and cylindrical ter response, are taken into account.The energy geometry on the dose distribution for and dose rate dependence is reflectedin a depth- incident electronbeams of 10 and18 MeV (1). Dose field size correction during externalphoton beam distributions were measured witha scanned p-type treatment. The variation in the valueof this diode and compared with 2D and3D dose correction factor and the calibrationfactor was computations based on the work ofHogstrom (2). determined as a function of time and totalamount The cylindrical rods havea fixed diameter (1 cm) of radiation received by differenttypes of dio- but variablelengths (1-10 cm), separations des. After establishing accuratelythese correc- (1.2-5.0 cm) andcompositions tionand calibration factors, in vivo dose (CTNumber of 495-2224 HU). Measurements measurements were performed, mainly during for increasing pelvic density andlength show the progressive changes treatment. Relatively large series of measurements in "off-axis" electron scattering.The gain on the same patient showed that theaverage in doseaccuracy achieved with a 3D implementation standard deviations of the entranceand exit dose appears significant only for thesmaller higher determinations are about 1% and 1.5%,respectively. density bones and requires Slightly larger values were observed during wedged a computation time of 5 min. on a VAX-11/780 processor for the central beam irradiations. Actual dose deliveryto the dose plane (4096 pts). patient was compared with stated dose The clinical implications values ob- of these results will be discussed in tained from treatment planning. conclusion. (1) Phys. Med. Biol. 32, 1073 (1987) (2) Phys. Med. Biol. 26, 445 (1981)

44 1988 World Congress on Medical Physics and Biomedical Engineering San Antonio, Texas Medical Physics Scientific Papers

MP11.9 MP12.1 BOSIHETRICAPPLICATIONSOF FIVE RECONSTRUCTED First clinical experience with an ionisation BREHSSTRAHLUNG SPECTRA BETWEEN 4HV AND 10HV chamber device for portal imaging. H. Meertens*, A.Piermattei*,G.Arcovito,L.Azario, Catholic Uni- M. van Herk and J. Bijhold, Netherlands Cancer versity, Rome,Italy - C.Bacci, G.Giacco La Institute, Amsterdam, The Netherlands Sapienza University,Rome, Italy- Laianciardi, H.Benassi. Nat.Cancer Inst.,Rome,Italy - E.De A portal imaging device based on the measurement Sapia,S.Camillo Hospital - Rome,Italy. of ionisation currents in a matrix ionisation

There is a lack of experimental bremsstrahlung x- chamber became recently available for clinical ray spectra produced by radiation-therapy accele- application. The field of view of the system is rators due to several difficulties. We report relative trasmission data obtained by measurement 32 x 32 cm or 16 x 16 cm, depending on the pixel along the central axis andoff-axis of five size: 2.5 mm or 1.25 mm respectively. The noise photon beamsgenerated by two 4HV, and 6,9,10 HV linear accelerators. A numerical analysis of in the image is less than 0.3% for a scantime of trasmissiondata has been applied to determine 2 s. During dose delivery, images can be dis- the fractional energy fluences on the central axis and off-axis of each beam.The central axis played every 2 s. The qtality of the images ob- spectrahave been used to find the Spencer-Attix tained with this device is sufficient for the stopping power ratios and a comparison with the dataobtained by methods reported on protocols purpose of patient set-up verification. Portal are shown. A correlation between the mean energy images of various tumor sites will be shown and of thespectrum and the quality indexTPRI:was found. The off-axisSpectra have been used to the following clinical potentialities of the account for the softening effect on the flatening device will be discussed: a) the use of instant filter. Corrections of clinical computer programs using conventional zero field size This were portal information, i.e. obtained within 10% of made.With these results and those in phantom the fraction dose; b) the patient motion during measurements, off-axis quality indexeswere obtained. dose delivery; c) the variation of fieldedges during dose delivers.

MP12.2 MP12.3 Digital Image Processing to Enhance the Detail of Anatomical A digital fluoroscopic imaging device for Structures within a Portal Film Image Obtained with High radiotherapy localizationP. Munro , J.A. Rawlin- Energy Photons.P. Bloch*, W. Powlis and L. Buhle. son, University of Toronto, Toronto, CanadaM4X- University of Pennsylvania School of Medicine, Department of 1K9, and, A. Fenster, John P. Robarts Research Radiation Oncology, Philadelphia, PA. Institute, London, Canada, NBA 5K8. Portal films are used to ensure that the Anatomical details seen on the portal film are viewed correct region of the patient is treated during together with higher contrast films obtained during treatment radiation therapy. Unfortunately, portal films simulation with lower diagnostic x-ray energies, to identify suffer from low contrast and often lack sufficient relevant discrepancies in beam alignment between simulation detail of the bony anatomy to determine what and treatment due to shielding block or patient positioning. region of the patient has been irradiated. While The treatment portal image, however, has low tissue contrast, the poor image quality is primarily due to the primarily due to the minimal differences in attenuation high energy x-ray beam, which reduces subject properties of high energy photons in bone and soft tissue. To contrast, film granularity and the poor display improve the quality of the portal image, digital processing is contrast of film contribute to the image degrad- used to (1) increase the display contrast between tissue types ation. and (2) compensate for degradation in the portal image due to We have developed a digital fluoroscopic imaging the finite size of the x-ray source target and the metal-screen device as an alternative to portal films. The film cassette. The portal image was digitally copied using a system consists of a metal plate/phosphor screen laser scanner microdensitometer, into 4096 grey-scale levels detector, a 450 mirror, and a high quality T.V. and 1684x2048 pixels each 200 microns2. A 2D-Fourier camera. While other investigators have demons- Transform of the image was obtained, and filtered in the trated systems of similar design, we have made frequency domain. The filter function used the measured modifications to the detector and the operation of inverse MTF for both the focal spot of the high energy photon the T.V camera in order to improve the signal-to- source and the metal - screen film combination. The noise noise ratio of the images.The phosphor has been power of the imaging system is also incorporated into the filter bonded directly to the metal plate, the light is function. The method permits a rapid reconstruction of the port integrated on the target of the T.V. camera for image that can be interactively adjusted to emphasize desired lengthy (1-10sec) periods, and the T.V. camera is particular anatomical features. operated in a slow-scan mode. The effects of these modifications on camera performance and on image quality will be presented. ')7

1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas Medical Physics Scientific Papers

MP12.4 tf12.5 Contrast enhancementfor on-line portal On-Line Radiotherapy Imaging Using imaging, K. Leszczynsky*, S. Shalev, Fiber -Optic Image Reducers, Y.A. Manitoba Cancer Treatment and Research Cheng*, J.W. Wong, W.R. Binns, J. Foundation, Winnipeg, Canada R3E OV9 Klarmann and J.W. Epstein, School of Medicine and Departmentof Physics,Washington University, Portal verification images (from film or on-line St. Louis, MO, USA. imaging systems) suffer from poor contrast and digital processing is used to enhance the visual perception of anatomical landmarks. Local histo- On-line digital imaging devices gram manipulationroutines arewidely used for are being this purpose, in which grey level transformations developed for radiotherapy verification. are computed for Optical imaging is one promising approach predefined neighborhoods around and relies on capturing the light each pixel, and various schemes have been pro- output posed to reduce from an irradiated fluorescencescreen the computational burden by with a digital camera interfaced interpolatingbetween a sparse grid of such to a transformations. A new algorithm is proposed, small computer. There are no movingparts Moving Histogram Equalization and the components are readily available. (MHE), which dif- However, fers from the widely accepted technique of Adap- the system requires alarge tive Histogram Equalization (AHE) in that linear mirror to direct the light tothe camera located rather than twodimensional interpolation of come distance awayfrom the transformation functions isemployed. MHE is radiation and is inconvenient. We havere- compared to AHE and othermethods of contrast placed the mirror in suchan optical enhancement ,global histogram modification,win- System with an array of optical fiber dowing, filtering, etc), and its advantages are image reducers to "pipe" the image.Each demonstrated in terms ofcomputational speed and reducer itself is made up of a bundleof superior image quality. The ripple artifacts 15x 15 plastic fibers with an input commonly associated with AHEare greatly reduced dimension of 1.6mm x 1.6mm. Theoutput is by MHE, and examples are presented for the reduced to lam x lmm withina height of enhancement of portal images using the new algor- 12cm. The chaxacterLstics of the reducers ithm. for image transfer are studied. Imagescan be acquired in the order of seconds and will be pre-ented.

MP12.6 MP12.7 Lott Energy Imaging WithHigh Energy Brems - The Use of Storage Phosphors for Portal Strahlung Beams,Duncan M. Galbraith*, Imaging in Radiation Therapy: Therapists' Princess Margaret Hospital, Toronto, Perception of Image Quality. D Gur*, Ontario, Canada M Deutsch, PA Clayton, MS Rosenthal, AG Bukovitz, ..1C Weiser. University of A method isdescribedfor verification imaging Pittsburgh, Pittsburgh, PA, 15261, USA withlowenergy (keV) photonsproduced by a radiotherapy accelerator inthe bremsstrahlung A major drawback in the interpretation of portal process. These diagnosticenergy x-raysare images in radiation therapy is their low contrast. normally filtered out by thick high Z target This is due to the low differential absorption of materials. Using a berylliumtarget with no body tissues at therapeutic energies andto a rel- flattening filter and a low(4-6 MeV) operating atively large focal spot. We used an experimental energy, low energy photon images can 'beproduced storage phosphor imaging system (EasLman Kodak Co.) on film in contact with a thin high 2 fluorescen to produce portal images. The system consists of screen. Thescreen selectively absorbs the low a storage phosphor detector, a laser scanner (12 energy component of the broad spectrum of photon bit), am image processing module, anda laser energies. Theimaging performance of this beam printer (12 bit). Ninety-three patients undergo- relative toconventional diagnosticand thera- ing radiation therapy treatments ha' botha conven- peutic beams is assessed by comparing, a) measured tional portal image and a s.orage phosphor image object contrast, b) images of the RMI contrast tam The storage phosphor image was scanned, im- phantom and c) images of theAlderson Rando age processed, and displayed on a laser-printed phantom. film. The conventional image and the storage phosphor images were then displayed side-by-sideand were evaluated independently by three radiotherapists for information content. Each of the three radiotherapists found the storage phosphor images to be superior (p < 0.001) to the conventional images.

This work is supported in part by the Eastman Kodak Company. 68 46 1988 World Congress on Medical Physics and BiomedicalEngineeringSan Antonio, Texas Medicai Physics Scientific Papers

MP12.8 MP12.9 Image Manipulation and Processing Digital Enhancement of Treatment Verifica- in Radiation Oncology, M.S. tion Films using a low cost video digitizer with a personal computer. Edward Weinhous,* Washington University Mok*, Cancer Therapy & Research Center, San School of Medicine, St. Louis, MC, Antonio, Texas and John Feldmeier,D.O., USA. The UniversityofTexas HealthScience Center at San Antonio.

Radiation Oncologists use a variety of Digital image processing canbe used to dissimilarimaging modalities (plane improve the image quality ofradiotherapy films, CT,MRI, portal films, etc.)in treatment portal films. However most of the existing digital enhancement systems require prescribing and verifying treatment. We expensive video digitizers and computers, have implemented digital image manipula- which prevents them from being broadly used tion and processing techniques to provide by any treatment center. Recently low cost the physicians with a consistent, corre- video digitizers with high resolution have lated, and appropriately enhanced set of become readily available for use with personal computers. A video digitizer manu- images. Consistent in that the rendered factured by Matrox Electronic Systems Limited images have the same magnification fac- is used here.It can produce up to 1024 x tors, points of view, and similar con- 1024 pixels resolution with 256 gray levels, trasts. These consistently rendered images with "real time" digitizing capabilities. are examined for both spatial and temporal Software was developed to perform edge enhancement, histogram equalization, low and correlation (e.g. a similar film with a high frequency filtering and other methods of portal film with a calculated portal image enhancement. The images could also be stored [from CT or MR]; or serial portal films so that anatomical landmarks as well as field withoneanother).When appropriate, outlines can be compared with different enhancement is accomplished by contrast simulation and treatment days.Examples of modification, high-pass filtering, deblur- enhanced images are presented. ring, etc. Ultimately,physicians are provided with an improved data set from which to make clinical judgments.

MP13.1 MP13.2 Development of 430MHz Microwave Heating System APPLICATIONOF CALORIMETRYTOEVALUATION OF with Lens Applicator for the deep seated tumor MICROWAVE ANTENNAE FOR INTERSTITIAL treatment HYPEETHERMIA, D.A. Wilkinson*, T.K. Saylor, M.K. i* 3 M.Kikuchi, ,Y.Nikawa2,, T.Arioka , Mizicko and p.N. Shrivastava, Allegheny-Singer T.Yoshida', T.Terakawa' and T.Matsude, Research Institute, Pittsburgh, pA 15212 1. National Defense Medical College, Saitama, Japan. We have constructed a simple "adiabatic" 2. National Defense Academy, Kanagawa, Japan. calorimeter for the purpose of measuring total 3. Tokyo Keiki Co., Ltd, Tokyo Japan. energy delivery from microwave antennae. This 4. Tokyo Metropolitan Komagome Hospital, Japan. device has a volume of 270 c.c. and is filled with 0.9% saline. Temperatures are measured For microwave hyperthermia, it is difficult to with a calibrated thermistor before and after deposit electromagnetic(EM) energy deep in the the application of power, equilibration tissues. In this paper, we have newly developed occurring within 2-3 seconds. Several types of the lens applicator based on our previously antennae (BSD, Clini-Therm and Cheung) have been reported lens technique which can deposit EM evaluated at different depths of insertion.Our energy in the lossy medium with the convergent data show that this technique is a simple, quick effort of radiated EM field, and the microwave means of testing the usefulness of an antenna heating system(HTS-100) operates at 430 MHz. This prior to hyperthermia treatment. system consists of the compact high power unit (solid state type, 500W), the lens applicator, the Supported in part by NCI contract #N01-04-37512. multipoints temperature measurement unit, the cooling unit and the control unit which installes the calculations of thermal dose, abnormality detections and automatic safety stop. We have also carried out the experiments on heating both an Agar phantom and the liniature pigs to evaluate the temperature distribution and control performance. These results show that the heating depth of approximately 60mm was obtained, which implies great capability in clinical use. c 9 1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas 47 Medical Physic; Scientific Papers

MP13.3 MP13.4 REGIONAL HEATING APPLICATORS:CHARA- The Use of Double Node Microwave Antennae CTERIZATION IN PHANTOMS, A.M. Di for Simultaneous Intraoperative Interstitial Nallot L.Begnozzi, G. D'Ermo andM. Hyperthermia (IOHT) and Intraoperative Benassi, Regina Elena NationalCan- cer Radiation Therapy (TORT) with Electrons. Institute, Rome,Italy. G.C.King,!.Sc.*, R.Nibhanupudy, M.SC. J.Smyles, M.D.,A.L.Goldson, M.D.,E.Ashayeri, M.D. The citotoxic effect ofheating and the enhancemert of the radiotherapyeffecti- veness The dosimetry problems, because of thepre- combined with hyperthermiahas sence of double node antennae and nylon favouredthe development oftechnology flexiguide in the radiation field for loco-regional are treatments. discussed. The solution will help to For abetter choose ofhyperthermic eliminate the use of numerous single node treatment technique, measurements have antennae for uniform heat distribution. been It carried out, in order to is essential to investigate these problems characterize two different units: a to insure an uniform heat and radiation couple of capacitiveelectrodes working distribution during che.use of simultaneous at 18.56 MHz (M. M. E.) and a waveguide TEM Intraoperative Hyperthermia (IOHT) and working at E'T MHz (Aeritalia). Intraoperative Radiotherapy (IORT) for The phantoms used were builtusing com- primary and recurrent tumors. positions suitable to simulate tissues dielectric constant andelectric con- The findings presented are basedon ductivity at the frequencies of inte- observations of a 915 MHz microwave field rest. Measurements were carriedout to used to maintain a temperature of 44°C for investigatethe heating omogeneityand 30 minutes while 6,9,12,15,or 18 MeV the possible existence of disperse electron beam. fields and hot spots. These two different heating techniques are used in our Institutionfor pelvic tumors treatments.

NP13.5 NP13.6 NEW INTRACAVITARY ANTENNAARRAY FIXTURE, Low frequency I 'Applicators for localized Thomas R. Canfield, Ph.D.* and Anne Hyperthermia, G.Banci, C.Franconi*, McCall, M.D., Loyola University of C.A.Tiberio and L.Raganella, II Univ. of Chicago, Stritch School of Medicine, Rome, Via O. Raimondo, 00173- Rome, Italy Maywood, IL and Edward Hines Jr. Hospital, Hines, IL. A family of 27 MHz hyperthermia applicators A new fixture has been designed for of innovative design have beendeveloped, introducing a hexagonal array of 915 Mhz that provide almost uniform heating in coaxial microwave antennae into intracavitary space. The antennae are insert- pseudo axial symmetry body segmentswith ed into closed end catheters that are the highest symmetry congruence. Theyhave mounted ingrooves parallel to the axis have a of a 1 cm diameter lucite red. In one potential use in regional configuration all six antennae junctions hyperthermia that can be circumscribedwith are in a planeorthogonal to the rod. The rod has a telescoping mz.chanismthat ring electrodes.Two implementationswill be permits three alternating antennae junc- presented, i.e. the active ring electrode tions to be repositioned toelongate the energy pair and the Toroidal Transformer deposition pattern. Normally systems, pairs of antennae would be activated also working at 95 MHz, at which together. frequency The device has been testedin vivo in the dogLarynx. These results axial focusing is exhibited. Resultswill and phantom data will be presented. be shown in which these devices provide localized heating at intermediatedepths also in combination with inductivedevices such as the Twin Dipoles and theopen-mode Toroidal Transformer to forman hybrid type of applicator. 7 0 48 1988 World Congress on Medical Physics and BiomedicalEngineering San Antonio, Texas Medical Physics Scientific Papers

MP13.7 MP13.8 RF Inductive Hyperthermia for Deep Seated Tumor, TARGETING OF CANCER CHEMOTHERAPY BY MICROWAVE IRRADIATION, Naoki NEGISHI*, Y s,.hiko TOMITA, H.Ikeda, M.Fujii, K.Sakamoto, H.Kanai*,Sophia University, Chiyoda-ku Tokyo, Japan 102 Makoto KIKUCHI, Yoshio NIKAWA ,and FUmiaki OKADA1, Dept. of Med. Eng., Nat. Defense When a cylindrical body is heated by a surrounded Med. College, Saitama359, JAPAN., ' Dept. of coil, temperature rise in the superficial partof Elec. Eng., Natl. Defense Academy, JAPAN. the model is very large,but is very small in central part of it. We have already reported about a new coil system for induction heating. As the In order to develop a temporal and spatial drug- anywhere administration method using physical actions, a results, it was shown that tumors seated at basic reseach was made on the anticancer drug in body can be heated by this new method. delivery system using microwave in the range of Temperature rise distribution in head cross-section are calculated, when this new type coil is A physical-targeting method which 430-MHz band. Both ultrasound heating and RF achievesa localized drug-administration by applied to brain. capacitive heating methods can not be applied to selective irradiation of microwave is proposed in However, new our meth- lieu of conventional biochemical-targeting method. brain because of .he skull. Experimental results obtained from When irradiated with 433-MHz microwave(MW, to the od be applied. polymeric prodrug of methotrexates coupled with phantom experiment are discussed togather with biodegradable poly(hydroxypropyl)- L- glutamine calculated results. Phantom is homogeneous the (PHPG),its temperature rise was higher than that cylinder made of ager solution which has almost of human muscle. of neutral PHPG. Polyelectrolytes which showed same electrical properties as that large values of the relative loss factor(e") were We applied various type coiles to the phantom. Frequency and power are 13.56 MHz and about 50Jw, easy to generate heat by 433-MHz MW irradiation. respectively.After heating, temperatur,, rise are When a strong dolyelectrolyte(PE) was injected is into the phantom simulated to muscle tissue in measured by a small thermistor probe. Hot spot value of e",it was confirmed that the PE was more obser'oed in the central part of phantom. heated than the phantom by external MW Experimental results agree well with calculated If we cool the superficial tissues, deep irradiation.A localized drug-administration results. induced by selective microwave irradiatfinn,that seated tumor can be heated well by this method. is,a physical targeting can be expected if the The effect of inhomogeneity and blood flow change by this method are macromolecular prodrug in polyelectrolyte system on temperature rise distribution also discussed in this paper. is utilized.

MP13.9 MP13.10 CLINICAL INTEREST OF RIDGED-WAVEGUIDES OPERATING A Study of QA program of High Energy AT 27 Mhz FOR DEEP HEATING. C. Marchal, P. Bey, Therapy Machines in The Republic of China, and P. Aletti*, Dept. of Radiotherapy, Centre Mann Tchao Wang* and Shiaw Chyan Chang, Alexis Vautrin, Vandoeuvre les Nancy, FRANCE. National Yang Ming Medical College, Shih-Pai, Taipei, Taiwan, R.O.C. A complete hypertlermia device named HPRL 27 has been developed in collaboration with SAIREM Ets (Lyon, France). The second generation of material High energy therapy machines are widely spreading is proposed for the treatment of pelvic tumors. in this country and the tread of use are rising A 27.12 Mhz generator of 1400 watts drives a rapidly. Since very few medical physicists on the bended water-filled waveguide of 150 Kg. This island, there are some problems in maintainance applicator is held by a motor-powered mechanical and QA works. These may lead to potential harms arm. The effective aperture of this waveguide to the patient, employee and general public. A is greater than 400 cm2. The penetration depth task group was formed by the Atomic Energy Coun- was found about 8 cm but using a strongcooling cil to conduct a survey to review the situation of the skin the maximum heating can be dropped and tried to find a solution. The over-all rezult between 3 .ind 6 c..1 deep. Technical, physical shows that the most serious problems were the and experimental results on "tissue-equivalent" dose rate and the field size congruences but the phantoms and on anima's are given. Our five years radiation safety survey reveal to be highly clinical experience on 54 tumors treated is pre- satisfactory. sented. A postal method for monitoring the QA programs by using LiF and films was established to improve the problems. The dose rate and beam profile can be monitored by LiF powder. Cross hairs, field congruence, beam flatness and symmetry were measured by one film. The mechnical isocenter and radiation center measured by a second film. A special phantom was made to carry out the task. It was proved a highly satisfactory way to monitor the hospital in remote area. 7 1

1988 World Congress on Medical Physics a nd Biomedical EngineeringSan Antonio, Texas 49 Medical Physics Scientific Papers

MP13.11 MP13.22 AN EVALUATION OF A QUALITY ASSURANCE A Computerized Quality Assurance PROGRAM FOR MONITOR UNIT CALCULATORS, System, Vdananen A.*, Puurunen H., R. A. Dahl, E. C. McCullough, D. E. Kuopio University Central Hospital, Mellenberg*, Mayo Clinic and Mayo SF-70210 Kuopio, Finland. Foundation, Rochester, Minnesota. Acomputer aideddosimetrysystem is Computerized methods are frequeltly being used for constructed for routinemeasurements, computing monitor units or timer sets.These documentation and analysis of theresults programs can be developed by the local institution, of quality assurance (QA) in radiotherapy. be "imported" from outside or be part of acommer- A multimode electronic unit, controlledby cial treatment planning package. An evaluation a desktop computer, consists oftwo and quality assurance protocol for monitor unit precision electrometers, programmable calculators was developed to systematically voltage supply, adisc station fc data compare and check results from MU calculators. communication and steppermotor controller The results from this testing revealed that errors for waterphantom and densitometer. in MU calculators are possible when they are tested Software consistsofabsorbeddoseand over a wide range of field sizes (including blocked dosedistributionmeasurementprograms, fields), SSDs and depths. Our protocol presents a QA-analysis and data documentations. minimal program to assure proper calculator Using the electrometer readings and stored accuracy and software integrity. A key element of calibrationandcorrection factors the any protocol is the designation of acceptability absorbe( dose can becalculated. The criteria since meeting these criteria can have system controlstherealization of the medical-legal implications. Our results suggest QA-programsand messages about unaccepted that calculators be checked against hand calcula- values. All the data of QA measurements tions, and this be permanently documented. are saved bycomputer where they are Monitor unit calculation program maintenence will easilyobtained. Written instructions be outlined in the light of liability and software about various measurementsare not needed. corruption concerns. Because the time-consuming analysis, documentations etc. arenot needed, the system will help the dosimetrist indaily routine.

MP13.13 MP13.14 Criteria for acceptance testing and Experience in the Acceptance Test- commissioning of modernmaltimodality ing and Use of aComputer Con- linear accelerators, Jatinder R. ?ate* trolled Therapy Linear Accelerator, and N. Suntharalingan, repartment of M.S. Weinhous*, J.A. Purdy and C.O. Radiation Therapy and Nuclear Madimine, Granda,Washinyton University School Thomas Jefferson University of Medicine, St. Louis, MO, USA. Philadelphia, PA 19107

Theacceptance testing, carmissioningand quality In August,1987, a Varian Clinac-1800 assurance of modern aultiamdality linear accelerators linear accelerator was retrofittedwith have beccce more complex. 'Nese =whims offer dual the vendor's computer controlledconsole. photon and multiple electron energies, independent or In light of recent past events involving asymmetric motion of jaws, electron and photon arc computercontrolled therapy linear therapy and multiple sedzent therapy. Digital logics accelerators, extraordinarymeasures were and microprocessors are also incorporated into taken to assure the safe operationof the accelerator control and mmitoring to handle these entire interlock and computersystems. complex functions. The 'Code of Practice for x-ray During acceptance testing, the traditional therapy linear accelerator:" in current use is not machine acceptance parameterswere measur- adequate for these accelerators. the inadequacies are ed. In addition, however, each interlock in the following areas: (1) determination of radiation was tested as close to the fault/signal isocenter with independent jaws, (2) field size source as was safe and practical. Finally, dependence for asymmetric fields,(3) beam flatness several days were spent tryingto force requirements formultiplephoton beams, (4) beam the console computers out of their defined calibration, (5) measurement of anomalous radiation modes and into indeterminateor dangerous leaks, (6) software and hardware interlock checks, and modes. oetails of the successfulaccep- (7) acceptance andquality assurance of computer tance testing and the results ofone year software. Criteria and methodology for acceptance of operation t.ill be presented. testing, commissioning and quality assurance of modern accelerators arc presented. Ehys, 2, 110-121 (1975) 72

50 1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas Medical Physics Scientific Papers

MP13.15 MP13.16 Longterm Variation in Beam Symmetry as a Function A Compute.ized Tissue Compensator System. H.I. Amols; Columbia Univ.,L.E. Reinstein of Gantry Angle for a Computer-Controlled Linear Accelerator, M. Loyd*, J. Laxton and R. Lane, The and B.C. Baldwin, S.U.N.Y. at Stony University of Texas Medical Branch, Department of Brook, New York, U.S.A. Radiation Therapy, Galveston, Texas 77550. 1 A teletherapy tissue compensator system Longterm beam symmetry measurements have been per- is described. Patient contours are ob- formed on a Philips SL20 dual energy computer-contained by sequential tracings with a hand tiolled linear accelerator. Symmetry in both the held 3D magnetic sensing wand. Contours radial and transverse axis of each x-ray beam was are analyzed and displayed on-line by the monitored for eight gantry positions. computer, with automatic 3D surface In the radial or gun-target axis of the beam, both smoothing. Menu driven options allow for computer look-up tables and servo systems are re- contour editing, isometric display, gan- quired to maintain x-ray beam flatness and symmetry try, and collimator rotations. Compensa- with gantry angle. In the transverse or A-B axis tor molds are cut into styrofoam tem- only the look-up tables are used. plates via a computer controlled milling The look-up tables recurd beam steering current machine, with the user specifying scaling settings used to produce a flat, symmetrical beam. factors for compensator size. Molds can These settings are recorded every 3 degrees for 366 be filled with high density cerrobend degrees of gantry rotation. Recorded settings are alloy, or a special medium density gypsum used for each of the adjacent degrees of gantry stainless steel mixture (useful for tis- angle positions as well as for the tabulated degree sue deficits of up to '5cm). Mounting of gantry angle. insets are milled into :ale styrofoam These measurements were undertaken to evaluate the template to permit accurate mounting on effectiveness of this beam steering system. Eval- standard blocking trays. Patient contours uation of the data shows that careful set-up proce- .:an be ootained in a matter of minutes dures result in x-ray beams in which fluctuations with this system, with compensator fabri- in symmetry as a function of gantry angle are cation performed off line. A detailed withal 4. 1.5%. However, day to day instabilities system description, plus sample isodose produce a total overall variation in beam symmetry profiles will be presented. Comparable data which is on the order of 4- 2.0%. 1 obtained from a Clinac 6/100 and a Clinac 4/100 Huestis Corp., Bristol, R.I., U.S.A. linear accelerator will also be presented.

MP13.17 MP13.18 A Computerized Digital Video System CT-Based ConformationalBolusinq Technique For Contour Acquisition and for Electron Beam Therapy. F. Bagne*, Compensator Design, J.W. Andrew*, J. L. Low, L.A. Fadell, Medical College E. Aldrich, M. E. Hale and J. A. of Ohio, Toledo, Ohio Berry, Cancer Treatment & Research Foundation of Nova Scotia and Dept. In radiation therapy with electron beams a bolus of Radiation Oncology, Dalhousie is often used to compensate for surface University, Halifax, N.S., Canada. irregularities or to conform the isodose B3H 1V7 distribution to the shape of the treatment volume. Because of the three dimensionalnature of the Recent advances in computer and video problem, a clinically satisfactory bolus has technology have enabled significant not as yet been fully developed. We have advances in the field of surface developed a highly accurate patient-specific topography measurement. This panc: will computerized bolus for this purpose. The system describe F system for the design utilizes CT slices of the pertinent region under radiother-gy photon beam compensators. treatment position. Initially, the area to The system obtains contours using a fan- be treated is marked on each slice. These data shaped light beam and a video camera along with other anatomical information are which are mounted on a tray which fits then entered into a VAX-780 computer. A computer into a radiotherapy simulator. program converts thesaggitaldata intobolus Transverse contours are obtained from one information inaxial, or coronal planes. The digitized video image or, in the case of number of bolus planes is determined by the opposed fields, from two or more image thickness of the bolus slices and by the bolus frames taken at different gantry angles. material. Thispaper presents the design of Entire surfaces are measured by scanning the bolus system; describes the computer program the bed longitudinally and analyzing a used for the process andillustrates the steps number of images. Posterior surfaces are necessary for computation and fabrication of obtained by scanning a mold. A printer the conformational bolus. Finally, the clinical produces a template for making the application of the system is discussed by compensator from sheet lead. The cost of presenting actual cases which have utilized the hardware needed to build the system. the system over a one-year period. was approximately $10,000 (U.S.). 7 3

1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas 51 Medical Physics Scientific Papers

MP13.19 MP13.20 Dose Conformational Transmission Blocking Comprehensive Study of Effective Attenuation Teanique, F.R. Bagnil, D.G. Bronn and L.A. Coefficient for Use in High-Energy Fadell, Medical College of Ohio, Toledo, Ohio Photon Treatment Planning, F.R. Bagne*, N. Samsami, Standard cord blocks are generally used in lung S. Hoke, Medical College of Ohio, Toledo,Ohio cancer treatment after the spinalcord reaches its maximum tolerance dose. This technique Precision radiation therapy with highmegavoltage has the disadvantage of providing a lower-than photons requires a uniform dose throughout the biologically-adequate daily dose to the treatment volume. In compensating for the missing tissue, mediastinal strip. Alternatively, posterior oneareawhichhas beenneglectedin cordtransmissionblocks of a fixed thickness the past is the region of buildup. Clinically, have been used to maintain a higher fractionated this region is ofparticularimportancewhen dose over the course of treatment. The latter the involved tissues,such as lymph nodes, are technique neither corrects for the geometry located oithin a few centimetersfrom the skin. of the cord nor for the patient's surface contour. We have undertaken a comprehensivestudyto This paper presents a conformational transmission determine theadequacyof a single effective blocking technique to compute the thickness attenuation coefficient, p ef, for a given ofthe transmissionblock alongray-lines and compensating material for use in calculating to provide a uniform fractionated doseacross the missing tissue. Measurements were made the entire spinal cord. Initially, a series in a tissue-equivalent phantom using 10 MVphotons of CT scans are taken over the region of interest. (Varian Clinac-18). The compensatingmaterial Thelocation of the spinal cord on each slice used was polystyrene. Our results indicate is entered into a VAX-780 computer via a that a single ueff does not provide the desired digitizer. A computer program combines the accuracy over the clinically-used range of field anatomical information and the measured sizes, depths and missing-tissue thicknesses transmission data for megavoltage photons to and that u eff can vary Ly as much as 70% over generate a demagnified blocking pattern. A the region of buildup. This paper presents standard block-cutter is modified o produce the specific variation ofueff with depth, field the necessary conformational transmission block. ;ize and missing-tissue thickness andrecommends The final Cerrohend block is attached to the a techniquefor minimizing the number ofpeff trey along with the field blocks. This paper values required for treatment planningpurposes. pre..ents examples of theresults and discusses Application of the resultsto computer-assisted its clinical significance. tissue compensation techniques is alsodiscussed.

MP13.21 MP13.22 Comparison of Penumbra betweenFocussed PENUMBRA GENERATORS FOR CL1NICALEIBETRON BEAMS, and Straight Edged Blocks - Implications R. G. Kurop*, S. Wang, and G. P. GlargOw, Loyola- for Multileaf Collimators (MLC), P.Biggs*, Hines Radiotherapy, Loyola University Medical J.Capalucci and M.Russell, Mass.Genl.Hosp., Center, Maywood, Illinois 60153 Boston Ma USA Clinical electron beams used in radiotherapy have The penumbra for straight edged(SE) blocks has been compared with that forthe collimator jaws narrow penumbra. This makesabutmentof these (focussed blocks) of a 10MV linac for square fields extremely difficult. It is desirable to fieldsat three depthsby measuring the distances between the 50% andthe 80%, 90% and increasethe width of the penumbra at the region 95% doses respectively. Beam profiles were taken of abutment of the fields.One way of achieving with a water phantom scanningsystem using 1.5mm diam. diodes. The results showthat at a depth this is to introduce narrow, thin plasticstrips of 2.5cm, dmax, the penumbra is worse for SE or wedges along the edge where the penumbra is to blocks when field sizes exceed15cm at the 80% and 90% dose and 20cmat 95%. At llcm depth, be enhanced. The effect of such penumbra this 'break even' field size remains the same generators were studied. The influence of the for 83% whereas at 90% thepenumbra differs only for fields greater than 30cmwhile at 95% the width, thickness, and shape of such penumbra data overlap. The max. differences in penumbra crenerators on the beam penumbra and their role in amount to2 and i.5mm for the 2.5and 11cm depths respectively. At 15cmdepth the two sets field matching for electron beam energies form of penumbra are indistinguishable. Since most 6 MeV to 18 MeVwillbe presented clinical treatments prescribeto the 95% isodose or greater, SE and focussed blocksare equally effective e:.cept closeto dmax. Thus the mechanical complexity ofan MLC could be significantly reduced fora small sacrifice in penumbra. Further studies showthat SE block*' with either a 1mm ora 0.5mm offset at mid-block height do not naticeably degradethe penumbra.

52 1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas Medical Physics Scientific Papers

MP13.24 Standard Thorax Compensators for AP-PA Treatments Limitation of "Clear-Pb" as a filter to Dept. of Rad. reduce electron contamination produced Wendell Lutz* and JeffWilliamson, by hig13, energy x-ray beams., S. Guru Oncology, University of Arizona, Tucson, Arizona Prasad ,K.Parthasaradhi,YoungLee, Rajesh Ruperal, Rafael Garces,Evanston Dose delivered byfirst courseAP-PA portals to Hospitn1,Evanston,I1.60201. lung and esophagus tumors is generally limited by cord tolerance. Because of the long fields involved The dose in the build-up region of separation differences between superior and high energy accelerator x-raysin- inferior borders may be as largeas 14 cm. For a 6 creases whenever blocking trays are MV photon beam this results in cord doses 13% above used. The amount of electron contamina- and 7% below the dose at midline on the central tion depends not only on the collimator axis. Wehavedesigned a setof four, anterior and the flattening filter but also on field, standard 1 dimensional compensators for our the field size. It is also dependent 6 MV and 10 MV linacs. For over 80% of the cases, on the x-ray energy. The use of one of thesecompensators fits sufficiently well electron filters are among the sug- that corddoseinhomogeneity is limited to 3%, gested methods to reducethis con- eliminating theneed for customized compensation. tamination. Several filter materials The compensators, which accommodate 30x 30 cm2 have been previously studied. It ap- fields, are made from 0.031" thick copper sheets pears that the previous studies have mounted to aluminum trays that slide into the wedge not examined the field size in limiting slots. Each compensator is designed to compensate the usefulness of the filter. a differently oriented sagittal contour each with a We have studied "Clear-Ph" as a filter 41 cm radius of curvature. The contour orientations to reduce the electron contamination are defined bythe angle a line, tangent to the produced by 6and 10 MV x-rays gen- contour at the central axis, makeswith the erated by Clinac-1800 with special em- horizontal plane. The angles used are 15', 20°, phasis on the field sizes over which it 25', and 30'. To quicklyandquantitatively can he used. This trey material can be evaluate the fit of a standardcompensator to a used up to 25x25 cm' for 6 MV x-rays particular patient, transparentoverlays of the and up to 20x20 cm2 for 10 MV x-ray lateral viewofthese "standardcontours" are beam. compared with a sagittal contour drawn from measured patient separations.

MP13.25 MP13.26 Clinac 4 Treatment Accessories Revisited Dosimetry of Eye Shielding Block Wendell Lutz* andJeffWilliamson, Dept. of Rad. F.J.Lin *, H.H.Chen,T.C.Chen,J.H.Hong, Oncology, Univ. ofArizona, Tucson, Arizona. Department of Radiation Oncology, Chang Gung Memorial Hospital, 199, Tung Hwa Because of theshort80 cm SADof the Clinac 4 North Road, Taipei, Taiwan, R.O.C. accelerator, treatments such as tangential breast and certain head and neck cases involving oblique angles coupled with the use of cerrobend blocks, Lens is very sensitive to radiation and cataract wedges, and/or compensators are spacially difficult can be induced even by relatively low dose of to treat.We have designed a compactwedge- radiation. In treatment of head and neck tumors, compensator-block holder that maximizes the shielding eye block _13 used frequently and vari- distancefrom isocenter to the most distal ation in distance between the eye block and the accessor" Most notable is themounting of the eye surface can affect the dose to lens remark- blocks ?side-down". This removes lateral ably. Different divergent blocks were used to protrusions in addition to saving considerable find out the optimal distnace of eye blocks for space when the gantry is rotated from one field to Co-60, 6HV and 10HV'xrays from linear ::::celer- another with no blocks in place. For head and neck ators and to evaluate the dose distribution under cases e rigid, long, narrow extension was designed eye block. PTW thin window parallel-plate ioni- that attaches to the end ofthe couch. The head zation chamber was used in the measurement and rests on this extension and is immobilized with a all data were acquired in polystyrene solid thermoplasticmaterial molded to the patient's Phantom. The eye blocks have a surface area with surface and attached to the extension. This diameter of 2 cm at the distance of irradiation. extension is constructed of 0.050" thick aluminum Eye block at optimal distance can reduce the dose rolled into a half-moon cross-section so that to lens from 17% to 7% for 10MV and 6MV linear posterior and posterior oblique treatment beams are accelerator, from 14% to 8% for Co-60 comparing attenuatedby less than 5%bythe extension. to block on the regular tray. The optimal Remounting the wedges, rotated 90' fromtheir distance is 0.9 cm, 1.4 an and 6 cm for 10MV, 6MV original position,proved to be more convenient and Ca-60 respectively.When block diameter is bothfor treatments and blockmounting. In larger than 2 cm, the dose to lens is independent addition, when blocks areused with wedges, the of block size. blocks no longer havethe possibility of "falling outof their slots at certain gantry angles.

1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas 53 Medical PhysicsScientific Papers

MP13.27 MP13.28 Commissioning the NEC Linear Accelerators In House Repair of Co-60 Teletherapy Units. F.J.Lin*, C.S.Chen, R.H.Lan, C.P.Tu, H.H. Amendment Requirements, Peter R. Almond, Ph.D., Chen. Department of Radiation Oncology, Brown Cancer Center, University of Louisville, Chang Gung Memorial Hospital, 199, Tung Louisville, Kentucky 40202 and Kenneth R. lima North Road, Taipei, Taiwan, R.O.C. Hogstrom, Ph.D. The University of Texas M.D. Anderson Hospital, Houston, Texas 77030

Three NEC linear accelerators, two in Lit. In April of last yeas the Nuclear Regulatory Medical Center and one in Kaohsiung Medic.. Commission issued information notice number 87-18 Center, have been installed recently installed in "Unauthorized Service on Teletherapy Units by Chang Gung Memorial Hospital. Commissioning Non-Licensed Maintenance Personnel."The out- procedures included the acceptance test, radi- come of this notice is that only very limited ation protection survey, the collection of service can now he done by "in-house" staffon dosimetry data, and acquisition and assimilation these units. Although not a new regulation by of patient treatment aids.When the machine the NRC, it does represent a stricter interpre- installation was completed, the acceptancetest tation of the existing standard licensecondi- were performed by the NEC engineers first and tions. Although the notice did not address the must be performed again by hospital physicists to question c,f how licensees were to get their verify the speficinations and characteristies of units repaired, it did state that "Licenseeswho the facilities. It included safety, mechanical, desire to have these operations performedby digital, optical, dosimetry and lasersystems, qualified "in-house" staff members whoare not then measurements of dosimetry datawere made authorized by the NEC or an AgreementState must with thin wall parallel plate chamber inwater request an amendment to their license." In phantom and cylindric chamber in solid phantom. order to test this provision, a license amendment Dose calculation is by Task Group 21 method in was submitted to the appropriate authorities in our department. After commissioning procedure, Kentucky. The progress of this amendment and quality assurance program were initiated and its content will be discussed and comparedto performed for all facilities. Specified commi- guidelines for repair on Cobalt-60 unitsprepared ssioning procedures and qualityassurance program by the staff at M.D. Anderson Hospital inTexas are essential for maintaining high quality of by request of the Texas Bureau of Radiation radiation therapy. Control.

MP13.29 MP13.30 Manual For the Use High Energy X-Ray TI-erapy Single Board Computer Upgrade toa Linear Accelerator: Clinical and Physical Simulator Field Size Display, Aspects, E. Gaona*, V.M. Tovar, Depto. de Fi- J.J. Szabe& M.K. Sanford, Saint John Regional Hospital, sica, Instituto Nal. de Cancerologia-Universi- Saint John, N.B., Canada. dad Autonoma Metropolitans Xochimilco, San Shen a Treatment Simulatorwas given a major Fernando 22, 14080 Mexico, D.F., Mexico. overhaul rather than purchasea new one, a certain component could not be purchased The purpose of this manual is to present appropiate because the parts were no longer manufactured. material for the physicist and radiotherapist in Lack of this component rendered the clinical and physical aspects, incluying historical existing analog display system useless. summary, criteria for selection of linear accelera- This display indicates the treatment field parameters tors for radiation therapy, acceptance and periodi- for subsequent patient radiation therapy ca] test, high energy X-ray dosimetry, monitor unit - a vital loss. A completely new digital display calibration and monitor unit calculation in treat- system was therefore designed to replace ment planning. Futhemore other topics related to the old analog system. radiation therapy. This system consists of a 20 digit vacuum fluorescent This Manual covets the basic needs of physics ra- digital display controlled by a central single-hoard diation therapy and clinical applications in tea- computer. ching. Features of this system include: low cost of parts, quick and This Manual was published in spanish by National easy calibration, transparent to Simulator Operator, and Cancer Institute of Mexico, incluying a brief a self checking routine that notifies the summary in english. Operator of calibration errors. In addition, the computer can be optimized in processing speed to achieve minimum response timeto changing inputs while also minimizing the inputsen- sitivity to random noise.

54 198E World Congress on Medical Physics andBiomedical EngineeringSan Antonio, Texas Medical Physics Scientific Papers

MP13.31 MP13.32 Electron Beam Enerccy Monitoring Photon Energy Spectra of X-ray Beams from Medical Electron Accelerators, E.J. Born, using a Staircase Filter, Dept. of Medical Radiation Physics, Inselspital, P. O'Brien*, R. McVittie, B.A. Gillies, University of Berne, CH-3010 Berne, Switzerland Toronto-Bayview Regional Cancer Centre, Toronto, Ontario The knowledge of photon energy spectra of X-ray beams from medical accelerators is essential for certain dose cacuP:Lion algorythms, for investi- The matrix dosemeter (1) is a computer driven gations of treatment machine head design, for the square array of 49 ionization chambers spaced at calculations of stopping power ratios etc.. approximately 3 cm. A polystyrenc staircase A Monte C?rio program using the EGS4 code system filter in conjunction with the matrix dosemeter has been developped, which allows the simulation of is used to monitor electron beam energy in the radiation transport from the electron target addition to the normally monitored parameters through Cle treatment machine head to the patient output, flatness and symmetry.The staircase surface, where particle charge, position, energy filter is constructed as a series of 25 varying and angle. of incidence are sampled. Currently, thicknesses of polystyrene each of which shadows photon beams hlve been simulated from 8 to 45 MV a different pair of ionization chambers. The produced by two different accelerators (SL75-20, ionization readings are divided by the readings Philips; Asklepitron 45, BBC). Energy, angular and from an unshielded exposure on a one-to-one basis spatial distributions of the photons were deter- to correct for field non - uniformity. Integration mined and the electron contamination of the beams of the results yields a number which is was evaluated. For comparison to experimental data independent of electron beam flatness or symmetry depth dose curves were calculated resulting from but directly proportional to mean beam energy. the simulated photon spectra and from the electron This technique can reliably detect energy shifts contamination of the beam. Problems of treatment of less than 2% and is being used for nominal machine head design can be investigated e.g. by beam energies from 7 to 25 MeV. Data presented keeping track of the last interaction point for will demonstrate the software used and the each photon arriving at the patient surface.Work precision and accuracy of the technique. is in progress to use Ele photon energy spectra as input to a convolution algorythm for photon beam CO MARTELL et AlInt J Rad Onc Biol Phys 12, 2 dose calculations and to experimentally determine pg 271, 1986 these spectra by detecting the Compton electrons with a magnetic spectrometer.

MP13. 33 MP13. 34 Product ion of Photoneut rons in a Incorporation of Emerging Technologies into the Lead Shield*by High Energy X-Rays , Design of a Radiation Oncology Facility: Problems

P. McGinley , K. Long and R. Kaplan, Encountered, J.D. Bourland*, M.A. VE.ria, G.W. Emory CI inic ,1365 Clifton Road, Sherouse, P.E. Stancil, L.D. Stanley, E.L. Chaney, H.L. Atlanta, GA 30322, USA McMutry, J.E. Tepper. Radiation Oncology, University of North Carolina, Chapel Hill, NC 27599-7510. Lead and other high Z metals are occasionally used in place of concrete to We have recently specified an academic radiation oncology fabricate primary barriers for mflical facility which includes 3 megavoltage treatment rooms, a accelerators. In this work measurements combined high dose rate remote afterloading and minor of photoneutron levels behind lead shields procedures suite, 2 hyperthermia rooms, an isotope storage irradiated with high energy x-ray (20 to room, a simulator, a CT room (used for simulation and 45 MV) beams were made. Neutron fluence treatment planning), and, in anticipation of a highly values of the order of 7x 106 n/cm2 per computerized department, an expansive infrastructure for u -lit absorbed dose (Gy) of x-rays data communications. Merging this collection of existing delivered to the point where the beam and emerging technologies imposed unique demands on the enters the shield were observed behind design team. Special problems included optimization of a the lead barrier. By measuring the change floor plan comprised of multiplediagnostic and treatment of neutron fluence as a function of the functions and desir. and integration of a coherent, facility- lead thickness it was found that the wide data communications system. Considerable effort was majority of the photoneutrons were requiredtoeffectivelycommunic&..thesevaried Froduced in the first 2.5 cm of the shield. requirements to the designers. We found careful pre- It was observed that when 7.6 cm of planning and early education of hospital plant engineering borated polyethylene was added to the and the architectural firm to be essential.Our design front of the lead shield no reduction of philosophy, physical layout, data communications network, the neutron fluence behind the lead and practical suggestions for the physicist involved in the barrier was detected. Based on this work design of a treatment facility incorporating new technologies it was concluded that a radiological will be presented. health hazard can be produced by the replacement of a concrete barrier with lead.

1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas 55 Medical Physics Scientific Papers

MP14.1 HP14.2 Comparison of kr and Rcor Values with A The Influenceof beam parameters on percentage Measured Energy SpectraofElectron Beams depth dose Inelectron arc therapy,M. Pla, C. Pla ', and from Medical Accelerators, L. Andre* and E.B. Podgorsak, Department of Radiation Oncology, McGill H. Neuenschwander, Dept. of Medical Radiation University, Montreal, Canada. Physics, Inselspital, University of Berne, CH-3010 Berne, Switzerland Rotational or arc electron beam therapy is the modality of choice in treatment of certain large superficial tumor volumes which follow The most probable enerEs E and the mean energy curved surfaces. For a given treatment geometry an angle 13can be E at surface are paramaterh of the energy spectra defined which is uniquely determined by the two treatment parameters: the depth of isocenter di and field width w. The o? medicalaccelerators.In clinical routine these characteristic angle (3 is always smaller than the treatmentarc angle. parameters are evaluated using 1150- and R- values We found experimentally that electron beams with combinations P of from depth dose curves or depth-ionisationcurves. di and w, which give the same characteristic angle (3, actually exhibit We have measured the depth dosecurves for all very similar radial percentage depth dose characteristics even energies of the linear accelerator SL7S -20 though they may differ considerably in individual diand w. (Philips) as well as the depth dose and the depth- Moreover, there is a gradual change in the properties of radial ionisation curves for all energies of the betatron percentage depth doss curves with an increase in angle Q. For example, for a very small 13 the percentage depth dose distribution Asklepitron 45 (BBC). The energy spectra are given of the rotational beam is similar to that of the stationary beam; withan by an integration over all angles of the electrons increase in(3,however, the depth of dose maximum moves towards fluence in the beam. We have measured these energy the surface, i.e., the beam becomes less penetrating,and the spectra with a magnetic spectrometer.The rela- surface dose increases. tionship between electron energy and field strength We will discuss the dependence of the rotationalor arc electron B in the homogeneous part of the magnet include, beam percentage depth doses on the characteristic anglepand beyond universal constants, only the product R * B show that the (3 concept can be used in clinical situationsto predict the shape of the percentage depth dose curve when (R is the virtual radius of 1...3 trajectory). As w and di are known, or more importantly, it can be used to determine long as the measurement of the magnetic field is the appropriate w when di and the desired percentage depthdose linear we can use for all energies a constant for characteristics are known. the virtual radius multiplied by the absolute calibration factor of the field measurement for the evaluation of the electron energy.

MP14.3 MP14.4 TE Sf"ntillation Detector for Dosimetric Properties of an TORT High-Energy Photon and Electron Applicator System (SOFTDOC) for the Beam Dosimetry, A.S. Beddar*, P. D. Higgins and C.H. Sibata Mevatron-80 Accelerator, C.E. Nelson* S.M. Rakfal, R. Cooke, ECU School of A minimally perturbing Tissue-Equivalent Scintil- Medicine, Greenville, North Carolina. lation detector has been developed for high-energy beam dosimetry. The detector is composed of a Acustom system for intraoperative miniature plastic scintillator probe (1.4 mm di- electron radiation has been designed that ameter and 3 mm long) connected to a radiation-re- does not require physical 'docking' between the patient sistant fiber optic bundle optically coupled toa and the accelera.ot gantry structure. The 10 mm photomultiplier tube in series witha current systemuses clear plexiglas for most key or charge integrating instrument. The collected components and two accelerator current or charge is directly proportional to the mounted laser beams for coaxial alignment of the dose that would be deposited in the surrounding patient-cone axis with the accelerator tissue. beamaxis. This Noise levels associated with the radiation provides both excellent target visualization response of the selected optical fibers was min- and patient-accelerator alignment. A dosimetric imal compared with the resnvnse of the scintillator. study of the sensitivity of this system to Radiation damage was fot:ild to be negligible and slight (1-2 ) non-coaxial alignment and non- recovery rapid. The reproducibility of the de- standard SSD between the patient-fixed elec- tector for Co-60 was found to be better than 0.27.. tron coneand the accelerator-fixed electron and its response to be dose-rate independent. The collimator has been completed fora varietyof comparison of ionization chamber and scintillation electron beamenergies, primarycollimator detector measurements for photon and electron bews settings, and patient-fixed cones. were in good agreement. Furthermore, the physical The results indicate that within clinically dimensions of this detector offer a high spatial acceptable dose variations (+ 10%), this system resolution, especially when operating in the build- provides dose distributions whichare comparable up or penumbra region. to fixed docking systems with substantial improvements in clinical set-up time and vastly improved target visualization. 78

56 1988 World Congress on Medical Physics andBiomedical EngineeringSan Antonio, Texas Medical Physics Scientific Papers

MP14.5 MP14.6 On the Reliabilityof In Vivo Parameterization of Total Skin Electron Electron Dosimetry with Semiconductor Beams. J.C. Dennett*, L. Antonuk, and Detectors,E.Vartiainen,T.Lahtinen*, R.K. Ten Haken, University of Michigan A.Vddndnen,H.Puurunen, University Medical Center, Ann Arbor, MI 48109. Central Hospital and University of Kuopio, SF-70210 Kuopio, Finland. The calibration of a total skinelectron beam (TSEB) for a linear accelerator located in a Direct measurements of entrance dose from large treatment room (up to 6 m SSDfor TSEB clinical electron beams were tested. The treatments) and equippedwith a high dose rate detectors were commercially available option for 6 MeV electrons (-100 cGy/min at 4.5 m semiconductor diodes for photon and SAD) has recently been completed according to the electronbeams. Ascomparedwith the spirit of the proposed AAPM TG-30 protocol. This reference geometrythe response of some set of conditions coupled with the use of a 2.2 m detectors washighlydependent on the long linear scannerallowed the measurement of directionof irradiation. Futhermore the TSEB treatment parameters under a variety of detectors had a complicated field size conditions using ionization chambers, films and dependence. The highest deviations were TLD's. The changes in beamuniformity in air, found for very small and very largefield skin dose uniformity on phantom, doserate at sizes. For non-standard SSD, the inverse d-max in phantom, depth of 80%, 50% and 10% dose, square law correction does not hold for and level of x-ray background were parameterized tested diodes. The SSD correction factors as a function ofSSD for combinations of 6 dual were different for various detectortypes (up plus down) or single (normal Incidence) and for scattered and scanned electron fields both with and without the inclusion of a 1 beams. The study suggests thataccurate cm thicklucite diffuser in front of the patient in vivo semiconductordosimetrywith treatment plane. Ultimately, two beams were electron beams is not possible at present selected for clinical use (one withand one without many correction factors. without diffuser for shallow and deeper penetration). A summary of the measurements, a parameterization of the results, and the process used to select and clinically implement those beams will be presented.

MP14.7 MP14.8 Relative Outnut Factors of Irreaulariv Shand Electron beam dosimetry measurements on Electron Fields, Brian T. McParland, Division of the Philips SL-25, Jatinder R. Palta*, Radiation Physics, C.C.A.B.C., victoria, Canada K. Ayyangar, and N. Suntharalingam, Department of Radiation Therapy and The close outnut of an electron beam is high1Y Nuclear Medicine, Tnomas Jefferson University Hospital, Philadelphia, denendent unon the collimation sYstem design, the PA 19107 be energy and field dimensions. For irregularly Clinically useful uniform electron beams are shared fields, the output is also a function of generated on the Philips SL-25 using optimal the snatial variation in fluence caused by the settings of the primary collimators,a dual scattering foil system and new design shielding and of the degree to which ]ateral applicators. The open-sided applicators are electronic enuilibrium is present ez the central intended to properly collimate the beam while the second vtattering foil serves as a axis. A method of calculating the relative out-nilz flattening filter thus not requiring any wall factor of an irregular electron field has been scatter. A comprehensive set of electron dosimetry measurements for the nine available develoned and is m:esented.This relative nutnnt electron energies(4-22MeV) fordifferent factor is derived as the nroduct of a weightinn size applicators was made.These measurements include thepercentagedepth dose, surface factor (calculated from a discrete nencil beam dose, isodose curves, field size dependence of model) and the known relative outnut factor of the machine output at the nominal SSD, output at extended distances, cerrobend shielding regular field from which the irregular field is effects on depth dose, and leakage radiation The impact of formed. The rc-sults of calculation.: are nresented from the electron applicators. this type of applicator design on irregularly and are cannared to data measured for a veriet,r of shaped electron fields is described along with irregular fields at various beam energies. measurement data.

1988 World Congress on Medical Physics and Biomedical EngineeringSan Antor.'), Texas 57 Medical Physics Scientific Papers

MP/4.9 MP15.1

Isodose Shift Factors for Obliquely Dose Distribution Patterns of some Afterloading Devices. Incident Electron Beams, K. Ulin*, and H.-K. Leetiand G. Wiesen, Institut fUr Radiologische E.S. Sternick, Tufts-New England Medical Physik der UniversitUtskliniken im Landeskrankenhaus, Center, Boston, MA 02111 0-6650 Homburg, FRG.

It is well known that when electron beams are incident obliquely on the surface of a phantom, Formeasurements of dose distribution patterns around the the depth dose curve measurednormal to the applicatorsof afterloadingsystems a very precise detector surface is shifted toward the surface. Based on support forup to 36 TL- Dosimeters (HarShaw TLD 100 rods) was geometrical arguments alone, the depth of the constructed. The TL-dosimeters are placed in thin polyethylen nth isodose line for an electron beam incident tubes which are circularly or helically arranged. Fittings for at an angle e should be equal to the product of the afterloading catheter are also provided,allowing mea- cos 0 and the depth of the nth isodose line at surements in planes perpendicular and parallel to the appli- normal incidence. This method, however, can cator axis. cause significant errors in isodose placement Dose distribution patterns of some commercially available af- for beams at large angles of incidence. terloading systemswhith 137-Cs and 192-Ir sources were mea- A table of isodoseshift factors has been sured in air and in water.The experimental results are dis- developed fromwhichone may easily calculate cussed in order to establish a relation between dose pattern, the depth of any isodose line for beams at structure of applicator, structure of source and incident angles of 0° to 60°. The isodose shift arrangement of sources and spacers.By comparison of dose in air and dose factors are tabulated in terms of beam energy in water build-up factors were evaluated. (6 MeV - 22 MeV) and isodose line (10%- 90%) Based on theexperimental results modifications of a voxel- and are shown to be relatively independent of oriented dose calculation model were studied in order to in- beamsizeand incident angle for angles less corporate a suitablecorrection method. As a first approach than 60°. Extensive measurements have been made theapplicator structure, source structure and source spacer in a polystyrene phantom with a Capintec PS-033 arrangement was transformedinto thevoxel-array.Physical parallel plate chamber. The dependence of the propertiesand geometricalaspectsare taken into account chamber response on the incident angle of the by an appropriate modification of the radiological path length electron beam has been checked, and the scaling which then canbe applied factorof the to the primary component of ra- polystyrene phantomhas been diation. determined to be equal to 1.00. Results of dose measureaentand dose computation are demonstrated and discussed.

MP15.2 tiP15.3 A Convolution Method for Brachytherapy Dose A Dose Computation Model forAm-241 Computation in Heterogeneous Geometries. Vaginal Applicators Including Source- J.F. Williamson* andR.S. Baker Dept. of Rad. to- Source Shielding Effects, Chang Oncology, Univ. of Arizona, Tucson, AZ 85724. H. Park, Ravinder Nath* and Christopher King, Yale University School of Medicine, Currently-available brachytherapy dose computation Dept. of Therapeutic Radiology,333 algorithms ignore heterogeneities such as tissue- Cedar St., New Haven, CT, 06510, U.S.A. air interfaces, shielded gynecological colpostats, and tissue-composition variations in 1251 implants Recently, vaginal applicators containing despite dose computation errors as large as 40%. encapsulated sources of Am-241 have been To calculate dose in the presence of tissue and developed for intracavitary irradiation of applicator heterogeneities, we have developed a gynecological cancers (1). A vaginal applicator computer code that describes scatter dose as a 3-D may contain up to six sources, each source having spatial integral which convolves primary photon an activity of 2, 5 or8 Ci. Since Am-241 fluence with a dose-spread array. The dose-spread sources themselves presenta significant amount array describes the distribution of dose due to of shielding,radiation from one source is multiple scattering about a single primary shielded by the others. A dose computation model interaction site and is precomputed by our Monte for the determination of dose distributions Carlo code. To correct for heterogeneities, the around these vaginal applicators has been primay fluence at eachvolume element is adjusted developed and tested by comparing the model with for heterogeneities traversed by the primary dose measurements using LiF TLD in a solid photons, the dose-spread array is renormalized to water phantom. The dose computation model takes into reflect the density and composition of the account self-shielding and source-to-source element, and the distance to the point of lAterest shielGing effects on both primary and scattered is scalc4 by the pathlength of the int'rvening photons. This model has been incorporated into medium. Convolutioncalculations for 1251 and the Theraplan treatment planning system. 137Cs pointsources in the presence of finite phantoms, air voids and high-density regions have (1) R. Nath, R. Peschel, C. Park and J. Fischer, been compared to thecorrespondingMonte Carlo Int. J. Radiat. Oncol. Biol., 1988, inpress. calculations. Our results showexcellent agreement with the MonteCarlo calculations and,a gain in efficiency of 50. 0

58 1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas Medical Physics Scientific Papers

MP15.4 MP15.5 Dose Homogeneity in Multiplanar Implants A Brachytherapy Programme to Sort the in the Manchester, Quimby and Paris Seed Strands in an Ir-192 Implant, P.Biggs, Systems. J. . Paul*, R. F. Koch, Mess. Genl. Hosp. Boston, Ma USA P. C. Philip and F. R. Khan. NWCH, 800 W Central, Arlington Heights, Il Recent trends in brachytherapy suggest that 60005 differential seed loading techniques, either by changing the activity of different seed strands Idealized volume implants consisting of three or their duration in the implant are becoming parallel planes have been analysed for dose increasingly important. Although many computer homogeneity and overall quality for the programmes exist that determine seed locations Manchester, Quimby and Paris systems for a set based on random seed entry, none of them sort of given targets using all the applicable rules out the individual strands, as encountered, for of source distribution of each system. Since example,in the case of Ir-192 implants. The there are lo rigid rules for source plane present programme uses an algorithm based on spacing in the Manchester and Quimby systems, the interseed separation and the angle between except those conventionally followed, optimal three successive seeds to identify individual spacing had to be determined first before strands from the total list of seeds,thus comparing then with the Paris system of implants. permitting individual weighting .)f these The parameters calculated for comparison are strands. This programme has been successfully treatment volume, total activity, integral dose tested on a series of implants containing from expressed in terms of uniformity index, integral 32 to 142 seeds and varying in complexity. The volume that receives a dose rate more than twice results show that for plane parallel implants, the stated value, which have a bearing on the the interseed spacing is found to be 0.8-1.1cm quality of any implant. These values indicate for anominal lcm spacing and96% of the that the Paris system of implants is as good as measured angles are >170° and all are >160°. other systems and generally has an improved value For implants where the strands are considerably of uniformity index, a parameter that quantifies morecurved, whilethe interseed spacing the quality of interstitial implants. remains the same, only 47% of the angles are >170° and 93% >160°. Moreover, based on these tests, several checks have been incorporated into the algorithm to prevent misidentification of seeds in the wrong strand.

HP15.6 MP15.7 The Calibration of Sr-90 Beta Fye 3-D Correlation of CT Eye Anatomy with Applicators with Radiochromic 125 -I Plaque DoseDistributions Using Radiation Detectors, J.A. Sayeg'and Scleral Markers. R.K. Ten Haken*, D.L. R.C. Gregory, Department of McShan, A.K. Vine and A.S. Lichter, Radiation Medicine/Health Radiation Univ. of Michigan, Ann Arbor, MI 48109. Sciences, University of Kentucky, Lexington, Kentucky, 40536. Using our custom-designed, thin, heavy-alloy plaques, individual 125 -I seeds may be localized radiochromic material which is a on plane radiographs taken with those plaques in solid state solution of place. Previous attempts to correlatedose hexahydroxyethyl pararosanaline nitrile distributions computed from those films with eye in a nylon polymer (i.e., thin foil], anatomy onCT scans (obtained without the plaque has been used to determine the surface due to artifacts) weremade by immobilizing the dose rate and dose distribution of patient and aligning boney landmarks on the films Sr-90 eye applicators.The detectors with matching ones on digitally reconstructed are rugged, easily handled, have an radiographs from the CT dataset. That technique equivalent response to photons and has been greatly improved by suturing 4-6 small electrons, and produce high resolution (1-1.5 mm) barium spheres to the scleral surface images.They have been found useful of theeye prior to CT. Plaque placement and for this application due to the high filmlocalization follow. Dose distributions surface dose rates (10-100 rads from the films are now correlated with CT anatomy H20/secondl and their low sensitivity by minimizing differences in the relative (approximately 106 rads for an optical coordinates of those markers as they appear density of 1.0].The foils have been between datasets. ARMS deviation of 1.1 mm in evaluated on a He-Ne scanning laser position (5.5% in dose) for 32 markers used in 7 densitometer with a resolution of 0.3 patients has been achieved to date. In contrast, micron.Comparison with NBS doses calculated to the optic nerve and center of extrapolation ionization chamber the lens using this method differ on the average measurements indicated surface dose by 10-60%and 5-40%, respectively, from those rate agreement to within 5%. The computed using a standard 22 mm eye model; dosimetric properties and clinicaluse pointingout the importance of more accurate of these sources will be discusr.id. planning for these treatments.

19,...8 World Congress on Medical Physics andBioradicalEngineeringSan Antonio, Texas 59 Medical Physics Scientific Papers

MP15.8 MP15.9 Improvements in Directional Iodine-125 Dosimetry Analyses of 1-125 Plaque Plaques in the Treatment of Tumours of Using FilmDosimetry, the Eye, D Shackleton*, R Sealy, Andrew Wu and Frank Ernsin,Tufts-New England Medical C Stannard, J Hill, E Hering, J Korrubel, Center,Boston, MA, USA Depts of Medical Physics, Radiotherapy and Ophthalmology, Groote Schuur Hospital Nationwide studies of usingan episcleral and University of Cape Town, Observatory plaque containing 1-125 radioactive seeds 7925, RSA for treatment of choroidal melanoma have been undergone for two years now.However, Gold foil backed iodine-125 plaques have comprehensive dosimetry studies been used by us for 13 years to treat are scarce and controversial. Since the intra-ocular malignant melanoma and changes of energy spectrum of 1-125 in retinoblastoma. Several improvements tissue are negligible, filmdosimetry now enhance this technique. A scribed seems to be a logical choice, especially globe permits geometrical details to be fot high spatial resolution. In this readily determined in spherical coordi- paper, we first established an H&D curve nates. The plaques now extend to a with doses expressed in a unit of reference line 5mm from the limbus, specific dose rate constant. This avoids increasing positioning accuracy and absolute dose measurements. All film reducing time in theatre and radiation measurements are made and normalized to exposure to staff. Angled gold foil the density of dose exposed with 1 mCi at edges permit tailoring of the radiation 1 cm. The isodose rate distribution of field to the lesion, allowing a tumouri 1-125 seed (model 6711) plaque with and cidal dose to be delivered within lmm without gold shield are compared of the optic nerve while sparing that and discussed. A:so, the effects on dose structure. Gold foil ridges effectively distributions due to the presence of gold focus the radiation onto the target. shield along the central axis or around These improvements lead to better geo- the lip of the plaque are presented. metrical precision than photon or electron or proton external beam therapy appear to offer.

MP16.1 MP16.2 Neuromagnetic Studies of Evoked Pain,J. Joseph*, CARDIAC POTENTIAL AND MAGNETIC MAPS 9F R.T. Wakai, E. Howland, H. Backonja, C. Cleeland and VT- PATIENTS, R. Merritt, G. Stroink J.R. Cameron, University of Wisconsin, Madison, Wisconsin. M. J. Gardner" and B.M. Horacek, Dalhousie University. Halifax, Canada, B3H-3J5. We have measured evoked neuromagneticfields and electricpotentials usingpainful electrical High resolution body surface potential stimulation of tooth pulp and intracutaneous tissue. Toothpulpstimulation is ideal for maps (BSPM) and magnetic field maps these studies because it is predominantlyinner- :MFM) have been obtained of 14 patients vated by pain fibers. In contrast, intracuta- with ventricular tachycardia (VT). The neousstimulation excites a mixture of sensory potential maps are constructed from 117 and nociceptivE nerve fibers whichresults in a leads distributed over the body surface more complex subjective sensation. The strongest and sampled simultaneously. The MFMare components of theevoked responseoccurs at obtained by sequentially measuring the latencies of 150 and 250msec. These signals are magnetic field at 56 points ona 4 cm used to provide anobjectivemeasure of pain grid in a plane j..rallel to and just sensation under various experimentalconditions. above the anterior surface. The measurements ore obtainer' in a shielded room. Inspection of the bidirectional filtered Frank leads reveal late potentials in 10 of the 14 peients.The BSPM and MFM, plotted each 2 ms near the time instances that the late potentials occur, were analysed using, single current-dipole, inverse solutions. The succes of this analysis will be discussed in terms of the location of such dipole and the S/N ratio in the two sets of measurements. 82 60 1988 World Congress on Medical Physics and Biomedical Engineering San Antonio, Texas Medical Physics Scientific Papers

MP16.3 MP16.4 MOVING SINGLE DIPOLE FIT TO MCG AND BSP Measurement of the Magnetic Fields produced MAPS USING A HOMOGEUOUS TORSO MODEL, C. by the ON-OFF Response of the Humaa Retina,

Purcell, G. Stroink , B.M. Horacek and Y.Uchikawa, A.Adachi,T.Hasegaw4N.Okuyama, T.J. Montague, Dalhousie University, and M.Kotani,Tokyo Denki University, Hatoyama, Halifax, Canada, B3H-3J5. Saitama 350-03 Japan.

We have measured the 117 lead Body Surface Potential Maps (BSPM's) and 56 The MRGs have been measured in a nonmagnetically lead Magnetocardiographic (MCG) maps of shielded room with a SQUID connected to a second a group of 10 normal male subjects (mean order gradiometer with a baseline of 3.2 cm and age 55 y.) and 15 subjects (mean a e 57 a coil of 2.48 cm diameter. To induce the magne- y.) recorded on average 77 days tic fields from the retina we used two kinds of following their first documented light stimulations. One is by a flash(white light) myocardial infarction (MI). The BSPM and and the other is by the red light of LED transmi- MCG maps, plotted _ 10 ms apart during tted to the eye through optical fibers. The magn- the T wave interval, were analysed with etic recordings were averaged with more than 100 an inverse solution for a single current samples. dipole that uses a homogeneous torso It was found as followings; the amplitude between model with a realistic outer boundary. the a-wave and the b-wave was the range of 200 to We found that, using this model, 93% of 300 fT. The a-wave of the MRG did not observe as the data content of the first 80% of the seen in the ERG when the intensity of the light T wave magnetic maps and 98% of the ita was decreased. Especially we will present the content of the electric maps durin, .his magnetic recordings produced by the OFF response interval can be attributed to a single of the human retina which is the first measurement. dipole. A classification of abnormality The fesability to analysis for transient response in the results of the BSPM data of the of the retina was confirmed by introducing this MI's in terms of the values for dipole ntw approach of noninvasive biomagnetic measureme- components and location are in agreement nt. with the diagnosis based on the 12 lead ECG. The analysis of the MCG maps classified 9 of the 15 MI patients as abnormal.

MP16.5 MP16.6 SQUID-based Magnetic Sensor cooled by High- Resolution SQUID Magnetometers for Blomagnetie Studies, John P. Wikswo, Jr., Vanderbilt Univer- a Refrigerator,S.J. Williamson*, D.S. Buchanan, D. Paulson, and G. Klemic, sity, Nashville, TN 37235 USA Department of Physics, New York Univer- in studies of the human magnetocardiogram and sity, New York, NY, 10003, and Biomag- magnetoencephalogram, the separation between the netic Technologies, Inc., San Diego, CA, bioelectric source and the SQUID pickup coils is determined by both the thickness of the intervening 92121. bone and soft tissues, and by the distance between Neuromagnetic measurements relying on the SQUID sen- the pickup coils and the outside of the Dewar sor have up to now required liquidhelium for cooling. To containing the SQUI% Tradeoffs between signal-to- avoid this and gain flexibility in operation, a sensorincor- noise ratio and resolution are optimized when the porating a dc-SQUID and second-order gradiometerhas coil diameter is approximately equal to the coil- been developed that is cooled by a two-stage refrigeratorin source separation, typically 1 to 2 cm. While the dewar, coupled by flexible lines to an external compres- this size of instrument may allow localization of sor. A Gifford- McMahon cycleprovides a stage at 15 K, effective dipole sources in the cortex to within 1- and a Joule-Thomson expansion chamber supportedfrom 2 mm, detailed mapping of the sources is precluded this sustains a temperature of about 4 K for the SQUIDand by the severe low-pass spatial filtering from the gradimeter. The sensor's white noise level isabout 20 1-2 cm source-coil spacing and coil diameter. Me'', and residual noise introduced by the Gifford- Intervening bone and tissue can be removed for McMahon displzzer is virtually eliminated by real-time experiments on in vivo and in vitro animal prepara- computer filtering. In application the dewarcontaining the tions, so that reducing the coil-Dewar spacing is sensor can be pointed in anydirection, including upside paramount for high-resolution mapping of current down, and the intrinsic noise level is insensitive to orienta- :_Jurces. We present a design study and progress tion. A pair of such "CryoSQUIDs" has been installedin a report on SQUID systems with 3 mm diameter pickup magnetically shielded room, independently mounted so that coils, 1-2 mm coil-sample spacing, and a potential they can monitor activity in widely separated areas of a sensitivity of 50 fT/VHF. The decreased spacing subject's bi Ai. Their performance used in conjunction may lead to an order of magnitude increase in with a conventional 5-sensor system will be described. signal-to-noise ratio for hi,:hly-localized sources. Such instruments should be able to detect as-yet- unohserved current distributions associated with differences between the intracellular and extracellular anisotropies of cardiac muscle. 83 1988 World Congress on Medical Physics and Biomedical EngineeringSan rintonio, Texas 61 Medical Physics Scientific Papers

MP16.7 MP17.1 Two-dimensional Inverse Problems inBiomagnetism, Use of Magnetic Resonance Spectroscopy B.J. Roth, S. Tai: and J.P. Wikswo, Jr.*,Vanderbilt (MRS) to AssessMetabolic Effects of University, Nashville, TN 37235 Therapeutic Ultrasoundonthe Heart. There ins a scant but growing literature applying 0.Z. Roy*, G. V. Forester, and R. Best, Fourier transform techniques to thebiomagnetic National Research Council of Canada, inverse problem. We have solved the inverse Ottawa,Ontario,Canada KIA OR6. problem analytically for a two-dimensional current We have established distribution. If the normal component of the that therapeutic ultrasound applied to acutelyischemic myocardium magnetic field, Bz(x,y),is measured over a plane in the at a height z above a two-dimensional reperfusionstage increased recovery. We used current 31P MRS to measure highenergy phosphate density distribution, J(x,y), then fromthf law of compoundreservesand intracellular pHwhile Blot and Savart J can be related uniquelyto Bz by expressing both as functions of the spatial monitoring the mechanical response of Langendorf perfused working rat frequencies, kz and ky. The filter function heart preparations. In non-protected heartsarrested for 60 minutes, relating Bz(kx,ky) and J(k ,ky) is proportionalto the function, exp(- kz2+ky2 z). ultrasound applied during reperfusionenhanced Thus, the magnetic recovery. field provides a low-pass filtered version Withcardioplegicarrest,hearts also of the recovered current distribution, with the amount of completelyafter 60 minutes of arrest filtering with and without ultrasound, but only depending on the distance z. Other field compon- ultrasound improved recovery ents, different gradiometer configurations, withcardioplegic arrested and the hearts at 85 minutes and 105 minutes. effects of finite pickup coil size By 155 can be included minutes of arrest, analytically. We use fast Fourier transform (FFT) application of ultrasound produced noimprovement in recovery. algorithms to transform the magnetic fieldinto the MRS metabolic profile indicatesthat recovery spatial frequency domain, filter thedata with the is preceded by an overshoot in the Phosphocreatine appropriate function, and use the inverseFFT to obtain the current density distribution. concentration; without the overshoot the heart The does not recover. The intracellular pH returns filter that represents the solution to theInverse to normal whether or notthe problem is unique, but the stability ofthe results heart recovers contractile function. These data are depend upon geometric parameters and thesignal-to- important inelucidating noise ratio. The extension of this approach to 2-D the effects of ultrasound incardiac improvement and in and 3-D bidomain systems and to 3-Dmono-domain helpingto determine optimum systems may encounter non-uniqueness. conditions for reperfusion ofthe ischemic myocardium encountered in the clinical situation.

MP17.2 MP17.3 ApplicationofMagnetic Resonance 1 Spectroscopy to Study Preservation of II NMR Characterization of Human Blood Plasma, the Heart for Transplantation Ann E. Wright, University of Texas Medical G. V. Forester; 0. Z. Roy*, K. Butler, Branch, Galveston, Texas 77550. R.Deslauriers,andI.Smith, NMR characterization of human blood plasma is of National Research Council of Canada, great interest because of recent evidence that oc- Ottawa, Ontario, Canada K1A OR6. cult cancers may be detected by NMR measurement of the spectral peak width of lipoproteins (1).At Acorrect immunologicalmatch is animportant this institution we have used a JEOL 270 MHz spec- consideration inthe use of the donor hearts for trometer to perform spectral analyses of plasma cardiac transplantation. A match may be found samples from both normal subjects andcancer that cannot be used when the timeto remove, patients using a Hahn spin-echo technique.After transport andimplantthe heart falls outside the a TE of 80 msec, a long T2 component not normally preservation window of 4-5hours. Wt. have present was discovered in sanples from the cancer developed an isolated perfusedpigheart model that patients. High performance liquid chromatography gait beplacedin thehorizontalbore of a 4.7 testa (HPLC) is being used to analyze the spectral peak Bruker magnet with Medspec spectrometer, 31P containing the long T2 moiety.HPLC has been spectroscopy has allowed us to follow highenergy uhown ubeful for analysis of molecular species of phosphate metabolism and intracellularpHas hearts phospholipids and promises to provide particularly are stored for 5 hours at5Cand 12.C.Wefollow novel information with respect to changes in molec- the recovery of the hearts with reperfusion at '7.0 ular composition brought about by pathologiccon- while measuringthe metabolic and mechanic,1 heart ditions (2). Isolated intact molecules Muted from correlates.This allows us to determine optimal the HPLC column are vacuum dried, res. ended in conditions for preservation and recovery of donor deuterium and reanalyzed on the NMR spectrometer. hearts,and will leadto awider geographical Preliminary evidence indicates the long T2 peak harvest of suitable donor hearts. arises from a molecular species most abundant in plasma chylomicrons. 1. New Eng J of Med (1986) V.315, 22:1369-1376. 2. J Lipid Res (1986) V.27:131-139. (Note: The NMR spectrometer was purchased with 4funds donated by the M.D. Anderson Foundation.)

62 1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas Medical Physics Scientific Papers

MP17.4 MP17.5 Early post mortem changes of 1114N MR- THE ENIGMA OF RELAXOMETRY IN BIOLOGICAL SYSTEMS relaxation times in mouse liver necropsies SYED FAROOQ AKBER*, Dept. of Biomedical Engineering, Worcester Polytechnic Institute, Worcester, MA E. Moser* ,H. Ethsel5 and J. Schuster The present theories of proton magnetic resonance Institut fur Medizinische Physik and relaxation times fall short of explaining satisfac- .,Jtitut fur allgemeine and experimntelle torily the phys1.-al and biological mechanism re- Pathologie, Universitat Wien, A-1C Wien, sponsible for thedissimilaritiesof relaxation AUSTRIA behavior in normal and pathological tissues. An alternate approach to understand the mechanism Systematic investigations of1H-NMR relaxation is needed. times in mouse liver, depending on temperature and time-after-excision,have been done to be It is well known that when water is degassed (dis- able to control storage and measurement condi- solved oxygen removed), the relaxation time of water tions. Measurements have been performed with a increases. The spin-lattice (T1) relaxation time of BRUKER pc20 spectrometer. Data were fitted using protons in water completely saturated with oxygen a mono- and biexponental modelfor relaxation is 1.4 ± 0.1 sec, andin water completely free times.Extensive test measurements as well as ofoxygenis 3.6 ± 0.2 sec at one atmospheric computer simulations have been undertaken to pressure at 20°C (1). This provides a clue that check the validity of our results. a similarprocess maytake place in normal and pathological tissues, therefore it is hypothesized We found significant differences in normal mouse rat the dissimilarities ofrelaxationbehavior liver T,(depending on time-after-excision) due in normal and pathological tissues are most likely to sex hnd feeding status. Degenerative changes due to the concentration of paramagnetic molecular due to different glycogen content/decrease may oxygen dissolved in cell-associated water (2,3). responsible and time course of relaxatior. times 11- s paper will discuss the physical and biological (in vitro) could be used for diagnostic purpose. me_uanism of oxygen availability and oxygen consumption and its influence on proton spin-lattice i present address:Dept. Biochmistry and Bio- (T1) relaxation times in normal and pathological physics, School of Medicine, University of Penn- tissues. sylvania, Philadelphia PA 19104 1. G. Chiarotti et al.: Phys Rev 93:1241, 1954

2. S.F. Akber : Med Phys 14:910, 1987

3. S.F. Akber : Nuklearmedizin 28:37, 1988

MP17.6 MP17.7 Detection of the Differencein Functions Tl/T2 Contrast in Fast MRI as a Function of Field of CerebralHemispheresby Differential Strength, S. Vinitski, R. Griffey**, B. Griffey**, NMR,H.Kamei*,Y.Katayama and H.Yokoyama, J. Boone*, A. D'Amato, Thomas Jefferson Univ. Electrotechnical Laboratory,1-1-4 Umezoro Hospital, Philadelphia, PA, **Univ. of New Mexico, Tsukuba, Ibaraki 305, Japan. Albuquerque, New Mexico. Many pathological changes are associated with We developed a novel NMR technique to detect the increased water content which prolongs T2 to a difference in functions of human cerebral hemi- greater extent than Tl. Soft tissue Tl increases spheres. The method is based on the fact that the with magnetic field strength, suggesting that T1 blood flow throgh the brain tissues varies in- contrast may diminish in h..gh field spin echo homogeneously according to their levels of mo- imaging. However, differences between Tl/T2 tabolism and functional activity, and that tt.e ratios of edematous lesions and normal tissue inhomogeneous blood flow enhancement results in should incre ,se with the field strength. Using the changes of proton NMR signals gathered -om different solutions to the Block equations and region-selective detectors located at se'-. -"al taking into con,:id,.ration differences in suscep- parts of the head surface. The experiments were tibility of biological tissues, we calculated Tl, made on a homemade whole-body MRI system. Percep- T2 and T1 /T2 grey/white matter and grey matter/ tion of music, response to taste and activation edema contrast in both fast gradient refocusing of visuo-sensory areas were examined on normal (GRASS or FISP) and spin echo imaging. Cat and and right-handed volunteers. It was found that rabbit brains were then imaged in vivo at 1.5 and the left hemisphere played dominant roles in 4.7 Tesla. perception of muzical rhythms, and the right Experimental results support the theoretical prediction that while grey matter/ hemisphere in the perception of melody. When a edema contrast increases moderately with field droplet of the extracted solution of red pepper strength, the grey/white contrast in spin echo as dropped on the left side of tongue, the right imaging is maximum at 1-2 Tesla and then grad- somato-sensory area (not taste-sensory) was acti- ually declines. On the contrary, both grey/white vated. Linear relationship was observed between matter and grey matter/edema contrasts in gra- logarithmic illumination intensity of light spot dienc tnfocusing imaging increases drastically in the right visual field and the intensity of with fie strength. Thus, the ratio TI/T2 may the signal from left visuo-sensorv. The rr ults becone a strsitive indicator of disease in high on time-resolved measurements will be reported. field fast imaging. 85 1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas 63 Medical Physics Scientific Papers

MP17.8 MP17.9 NMR STUDIES IN BONE MARROW AND Is a Faraday shield ever useful in MRI coils? TISSUES, A.Kennedy,G.N.S.Prasad* S.V.Baskaran# and A.V.Lakshmanan,# Dept. of Physics, Anna University D. K. De and P. R. Moran Madras and # Cancer Institute, Radiology Department Adyar, Madras-25, INDIA. The Bowman. Gray School of Medicine, 300 S. ':awthorne road, Winston-Salem, The aim of this work was to evaluate the NC 27103 use of Nuclear Magnetic Resonance(NMR) for descriminating the normal bone marrow and tissues from the We calculate the relative MR coil malignant marrow and dielectric-losses, Pc and Pf, respectively tissues. The parameters T1 (soin-lattice due to the Coulomb's-law electric-field relaxation)and T2(spin-spin x.?.laxation) contribution (electrostatic field) andto the were measured for various bone marrows and Faraday's-law electric-field contribution tissues. There was a cleardifference (inductive field), within a conducting subject, between T1 values for normal marrowand inside an MR/ coil. We show that Pc is malignant narrow. There is no clear demar- insignificantly smaller than Pf. This result creation between the values of T2 obtained quantitatively depends stronglyupon the for diffe-ent marrows. However the Ti subject's dielectric constant and characteristic values for malignantmarrow is consider- dimensions; and this therefore distinguishesMRI ably higher than that of the normalmarrow coil-behavior dramatically from the experience In general the Tlvalues, for all kinds of of small-sample MR spectroscopy. The relativeeffect normal marrows, varies between 1000 to of Pf cannot be eliminated by 1600 m.sec and for malignant marrows the a slotted Faraday shield, since the relative strengthof Bl must be range varies from 1900 to 2500 m.sec at maintained to keep optimum NMR signalsensitivity. room temperature. The NMR parameters Tiand Thus the real usefulness of T a Faraday shield for were measured for arious kinds of tis- MRI subjects cannot be to reduce subject/coil sues also. The T1 values for normal dissipation effects, e.g., MR thermalnoise, and tissues are lesser than that of malignant Faraday shields are otherwise highly tissues. Thus NMR can be used as a diag- questionable. nostic tool in detecting the maligancy.

MP17.10 HP17.11

The PossibleDiscoveryof aReagentfor Quantitation of regional bone metabolism by Cancer Diagnosisby Urine N M R analyses, dynamic bone scintigraphy and its age-dependence Yong Jin Kim* , Department of Physics, of skull,S,Wada$,IL Hayama,lidilaeda, The Nippon Hanyang University and Central Lab., Samil Dental University,school of Dentistry at Niigata, Pharm. Co., Seoul, Korea. 1-8 Hamaura-cho Niitaga city, Niigata, Japan

Urine N M R analyses show that the N M R signalscor- Bone scintigraphy with seaTc-phosphate compound is said responding to four aromatic protons very distinctive- to be reflected bone mineral turnover, Therefore it is ly appear in normal and various diseased urine. How- expected teat quantitative parameter of regional bone ever the results of many repeated N M R experiments metabolism can be measured by quantitation of bone scin- on the urine i- Icate that the proton signals are tigraphy, We mathematically analyzed time activitycurve much more frequently observed in cancer urine. Area- of dynamic bone scintigraphy, applying compartment model gent made by mixing a dissolved metal with the well- analys:s of tracer kinetics, Investigating to abiod the known Millon indicator is used to identify them by effect the main artery and soft tissue we made computer observing the urine color reaction. Red precipitates program which calculate the quantitative parameters caused by the, reactionof the phenolic compound of fitting obtained solution with time activity curve of tyrosine are formed in the urine. Like the N M Rre- patient, sults it is fond that the urine reaction frequency The analysis was carryed out in every normal part of is particularly higher in the cancer urine. Moreover. cranial and facial bones using irlage data of patents from the reaction it is observed that thecancer who have been performed dynamic Lune scintigraphy, (63 urine is very sensitive to the reagent in comparison men :34 women), with other urine. With distinction of sex and location quantitative para- An attempt is made to determine the reagent sensitiv- meters were statistically investigated with regression ity and specificity of the urine collected bya ran- analysis,The age-ralated increase of a parameterwas dom sampling for a differential diagnosis between the observed in cranial and part of facihl bones in both cancer and non-cancer urine. The results of theat- sexes, And the tendancy was more rem? ''able in women of tempt show that its sensitivity and specificity are over 50 years old,suggesting corres: at least above 75 %. fence to the postmenopausal bone change, 83 64 1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas Medical Physics Scientific Papers

MP17.12 MP17.13 Use of Concentration Instead of Activity for A General Theory for the Transport and Exch- Absorbed Dose Estimation; E.A. Silverstein*, A.M. ange of Inert Diffusible Tracer in the Brain, Zimmer, and S.M. Spies. Northwestern Memorial T.Y. Lee* and I.W.T. Yeung, Radiology, St. Hospital, Chicago, Illinois. Joseph's Hospital ana Medical Biophysics, University of Western Ontario, London, Absorbed doses from radionuclides are usually Ontario, Canada N6A 4V2 calculated using the MIRD formalism. In the S value formulation, the self-dose to an organ is expressed as a product of the cumulated activity Cerebral blood flow (CBF) and capillary and an S value, tabulated for the organ in permeability (PS) are two important neurophy- question, and the radioisotope used. In this work siological parameters of the brain. Until now, the advantages of using the cumulated concentra- they have to be measured in separate experim- tion insteadofthe cumulated activity are ents. CBF is measured using inert diffusible presented. The absorbed dose is then expressed as tracer with the Ket- whereas PS is obt- a product of the concentration and a factur mS, ained by men. extraction efficiency where m is the "standard man" mass of the organ. or transfer _onstant. We propose a new theory It is shown here that mS varies only slightly for which treats the transport and exchange of different organs and radioisotopes, being inert diffusible tracer more exactly such that essentially constant as expected for pure beta both CBF and PS are determined simultaneously emitters. By use of the concentration approach it in a single experiment which uses quantitative is then possible to estimate doses for organs of computed tomographic techniques to measure the unknown size and for activity-containing regions tissue concentration of tracer noninvasively of non-standard shape. The use of these in vivo. It will be shown that under differ- techniques can give more accurate dose estimates, ent simplifying approximations, the proposed especially in conjunction with quantitative SPECT general theory reduces to the Kety model, the which is well suited to measuring values of Crone and Renkin model, and others which have concentration. been used by previous authors to determine CBF and PS separately. One application of the proposed theory will be the measurement of CBF and PS in patients using either H2015 and PET or D20 and magnetic resonance imaging.

MP17 .1' MP17.15 In Vivo Neutron Activation Laboratory Scattering Within Uniform Media. LA Henry K.J. Ellis* and R. Shypailo, USDA/ARS Children's and MS Rosenthal*, Uaiversity of Pittsburgh, Nutrition Research Center, Department of Pediatrics, Pittsburgh, PA, 15261, USA Baylor College of Medicine, Houston, Texas 77030. Compton scatter along with other interactions cause The quantitative assessment cf body composition, espe- image degradation and loss of quantitation in nu- cially protein, water, fat, and skeletal mass, can provide clear medicine studies.As a first step in charac- data that may increase our understanding of basic physiologic or pathologic mechanisms and aid in a patient's terization of scatter, we have modeled scattering within uniform scattering media via use of a Monte diagnosis and clinical management, Thus, a body composi- Carlo transport code. This code includes Compton tion laboratory has been set up at the Children's Nutrition scattering, coherent scattering, photoelectric ef- Research Center for in vivo neutron activation analysis fect, diffractions and production of lead x-rays (IVNAA). The four major facilities include infant- and adult- size whole body counters, a variable geometry neutron irra- From the collimator. diator for delayer; activation analysis, and a scanning The code models a 22 cm acrylic shell phantom ir- system for prompt-gamma activation analysis. The delayed radiating a 3/8" thick Nei crystal through a LEAP Within the phantom a single source is activation system uses 56 Am,Be sources, 15 Ci each, collimator. positioned along the longitudinal axis using either which can be arranged into four irradiation geometries air or water as the scattering media. For both ex- Ismail animal, infant, normal adult, and obese adult). The perimental data and calculations, 500 to 1000K adult counter consists of 32 Na! detectors (10.2 cm x events were detected by the gamma camera. 10.2 cm x 45.7 cm) arranged above and below the supine The Monte Carlo transport code was run on a DEC VAX subject. For a 2-min activation time (<3 mSv), 1-min delay, and 15-min count, the precision for total body Ca, 11/750. Both experimental measurements and calculations were performed as a function of position Cl. P, and K is <2%. The prompt-gamma IVNAA Within technique uses four collimated Am,Be sources and four along the longitudinal axis of the phantom. 15.3 cm x 10.2 cm Na! detectors for a single scan of the statistics, the spectra produced by the Monte Carlo subject. These measurements require

1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas 65 Medical Physics Scientific Papers

MA17.16 MP17.17 A new approach to the measurementof Attenuation Correction in Rectilinear High- glomerular filtration rate and blood Sensitivity Scanning, D. Gvozdanovic *, Dep. flow of each kidney using 99mTc-DTPA,G. ofBio-Medical Physics &Bio-Engineering, IZZO*& A. MAGRINI,Fisica Medica - Univ. of Aberdeen, Aberdeen, Scotland, U.K. Universita Tor Vergata, A. FAVELLA,F. FRINGUELLI & M. BANCI, Medicina Nucleare - Clinica Ars Medica, F. FRANCHI &L. A transmission scanning procedure, using AGATENZI, Ospedale S. Andrea- USL RM12, appropriate energy photons, was developed in Rome, Italy conjunction with high sensitivity (10 min. scan of 0.1 MBq in w. b.1 rectilinear scans in an attempt to assess the Absolute quantity of the Model analysis of the early part (3min) of the tracerin varies organs and sites of thebody, renogram was performed using alarge-field following long-term element turnover studies. computerized gamma-camera (90x 1 s4- 15 x 6 s images) and 99mTc-DTPA as a tracer. The Thetransmission scan is taken with a collimated Cme-activity curve calculatedover the left beam of the radionuclide used in the study, both ventricle was used as an input functionto a this andthesubsequent tracer scanshave an compartmental kidney and extrarenal tissue image matrix of16x52 pixels, anterior and model. Fitting of therenogram by taking into posterior images being taken at the same time. By account the real input enabled the glomerular using an IBM compatible computer, the imagesare filtration rate and blood flow of eachkidney to interpolated to a 46x126 matrix, filtered by a ")e determined independently from background. two-dimensionaldeconvolution process applying 32 patients undergoing routine DTPAtest were theimage transfer function (15x15 pixels). A studied with the proposed approach,as well as geometric meanis derived for each pixel from with two classic nuclear-medicine methods(Gates the anterior and posterior scans. This is and Constablemethods). Compared with the corrected for attenuation us4.ng appropriate pixel creatinine clearance values, resultsappear to in thetransmissionscan. Integration over demonstrate that the proposed method more selected area gives the fraction of tracer when accurate and/or simpler than the above-mentioned compared with the integral of the whole image. methods, also providing more completeclinical Phantom studies gave an estimationaccuracy of information. < 5% when clearly defined areas are involved.

MP17.18 MP17.19 Assessment of hepatic and splenic Halif-Value Layer (HVL) of Radionuclides hemodynamics with a single radiotrdcer Emitting Polyenergetic Gamma Photons, test, A. MAGRINI* &G. IZZO, Fisica Ibrahim B. Syed*, Muzaffar Ali, and Mark Medica - Universita Tor Vergata - Roma, H. Crosthwaite, V. A. Medical Center, Bell- A. FAVELLA, F. FRINGUELLI & M. BANCI, ermine College and University of Louisville Medicina Nucleare - Clinica Ars Medi - Health Sciences Center, 800 Zorn Avenue, Roma, L. VALERI, Ospedale S. Giacomo - Louisville, Kentucky - 40202, U.S.A. Roma, S. SPOSITO, Ospedale Civile- Velletri,F. FRANCHI & L. AGATENZI, Ospedale S. Andrea - USL RM12 - Roma, A knowledge of half-yalue layer (HVL) is impor- Italy tant for shielding purposes in radiation safety and also for clinical interpretation of photon- deficient areas in images in nuclear medicine. The MARK method (Model Analysis of Radiocolloid For monoenergetic photons and for a given Kinetics) for non-invasive study of hepato absorber(attenuator), HVL is calculated divid- -splenic hemodynamics is described. ing 0.693 byp (the linear attenuation Data collected by means ofa computerized coefficient). But in nuclear medicine, radio- gamma-camera for 4 min after i.v. administration pharmaceuticals are used with radionuclides of 99mTc-colloids (4 mCi) were analyzed usinga having multiple principle gamma emissions. An model formulated on a physiopathological basis. equation is derived for calculating HVL of such Several parameters of clinical interest were ragionuclides. determined, namely,liver arterial and venous blood flows, intLahepatie`:::::n..ed flow, spleen A ..omputer progra. has been developed to deter- flow, hepatic and splenic vascular spaces of nine the im of radionuclides which emit poly- distribution of thetracer, transit times, energetic gamma photons by Newton's method and extraction efficiencies and clearances. On the iteration. basis of 4 years experience and 8,000 tests performed, it is concluded that the MARK method is sufficiently accurate, reproducible, safe and practical, and may thus be considered suitable for routine clinical use. 8 8

1988 World Congress on Medical Physics and BiomedicalEngineeringSan Antonio, Texas Medical Physics Scientific Papers

MP17.20 MP17.21 The Use of Electworation for Labeling 13E117)1:T11:w= MiloaaCtteilv-et=ne, Leukocytes with Tc Radiopharmaceuticals, M 0 Shackleton*, I D Werner, C S Slater, William Phillips, M.D., University of Depts of Medical Physics and Radiotherapy Texas Health Science Center at San Antonio, Groote Schuur Hospital and University of San Antonio, TX 78284-7800 Cape Town, Observatory 7925, RSA Electroporation has been shown to create reversi- Radioactive iodine was used by us for ble pores in cell membranes without loss of cell thyroid ablation in 133 patients during viability. 9beukocytes electroporated in plasma 1983-1987. Only one patient developed containing mTc-MDP demonstrated a significant acute oedema of the neck. This patient increase in labeling compared to a control group had had one thyroid lobe excised of non-electroporated leukocytes. The labeling aviously. This very rare condition efficiencies ranged between 1.6%-5%. This new appears to be reported only once by technique for radionuclide cell labeling has another group. The oedema developed potential for applications in diagnostic imaging. within 48 hours of receiving 3 GBq cl radioactive I-131. The patient was treated with corticosteroids and responded well to the treatment. Five. days later the patient developed symptoms of vocal cord paralysis. Nuclear Medicine scans were performed on the patient to establish the location of the residual thyroid tissue. Sensitivity to 1-131 had been ruled out by using appropriate tests. From dose calculations it is concluded that the oedema was caused by acute radiation effects. Further tests are being done to investigate this condition.

MP17.22 MP17.23 MINIMUM PREDICTED GAMMA CAMERAINTEGRAL AtiD DOSIMETRY OF AUGER EIECTRON EMITTERS: IMPLICATIONS 4ND POTEREMANTICATIONS, UNIFORMITIES J.R. Halama , DIFFERENTIAL FLOOD K.S.R. SAstty , University of Massachusetts, and R.E Henkin, Loyola University Medical Center, Amherst, MA; R.W. He cell, DX, Rao, Maywood, IL. University of Medicine & Dentistry of,NJ, Newark NY; A.I. Kassis, S.J. Adelstein, Harvard Medical School, Boston, MA; and Gammacamera performance is routinely specified C. Haydodk, Mayo Foundation, Rochester, MN. by computing flood field uniformity defined by MAX and Several biomedical radionuclides (I-125, In -111, the ratio u.(MAX-MIN)i(MAX+MIN), where TC-99m, etc.) are prolific emitters of low energy MIN are the maximum and minimum counts observed Auger electrons with highly subcellular es. in a digitized flood image. For integral uni- Biophysical implications of sudnAuger emir searched, and for are not predictab]e by opnventional.dosimetry formity the entire image is (e.g MU D Schema), which ignores the importance differential uniformity a 6 pixel range is used. of these electrons. Our studies with mammalian Due to counting statistics, these indices do not, cells in culture, andonexperimentalmouse models show that the biological-effects of Auger however, converge to zero for a perfectly cascade electrons may be highly disproportionate uniform gamma camera response. This minimum can t9 the energy carried by tnem. Sur results con- be predicted by determining the probability firm the inadequacy, o f conventional dosimetry, distributionof u. Based on order statistics, and reveal several important parameters ignored hithert9. In terms,of these determinants, pheno- the joint probability distribution function for mem+ogical expressions are presented for a bio- ,,:seruing MAX and MIN is described as follows: physically meaningful dosimetry of internal Anger g(MAX,MIN)=n(n-1)[F(MAX)-F:MIN)rf(MAX)f(MIN) emitters. Possible ,approaches to exploit the highly, localized action of the low energy elec- where F and f are the distribution and density trons in cancer therapy,areexaminedudAlfSb-119 functions governing each pixel value, and n is and Pt-193m as illustrative examples. the number of pixels in the search. The dis- Work supportedinpart by USPBSGrant Nos. tribution function for u can Then be obtained by CA-32877 and CA- 15523. in'ec;rating g(MAX,MIN) over the region defined by u. Assumingr is Gaussian, the expected value and variance of u were obtained numerically versus n and average pixel counts. Results compared favorably to computer simulations and actual gamma caa:nra flood images. With this information, we can better ascertain the true uniformity of the prom.: camera by u.

1988 World Corgress on Medical Physics and Biomedical EngineeringSan Antonio, Texas 67 Medical Physics Scientific Papers

MP17.24 MP17.25 Dual-isotope Measurement of Lung Water in The Applications of Radionuclide Phase Dogs, R. Y. L. Chu*, P. V. Carlile, Jr., Image Analysis in Coronarj and Noncoro- G. Basmadjian, University of Oklahoma nary Disease, Zhao Huiyang* and Zhu Ju- Yealth Sciences Center and Veteran:. Admin- nren, Zhong Shan Hospital, Shanghai Me- istration Medical Center, Oklahoma Ci*y, dical University, Shanghai, China OklaLtma, U.S.A. 73104.

To develop a noninvasive method for measurement of This paper reported the results of 349 lung water, we have beet. studying the use of iodo- cases radionuclide phase image analysis, antipyrine. 1-131-labeled iodo-antipyrine and Tc- including 21 normal subjects, 169 pati- 99m-labeled erythrocytes were used to measure ents with coronary disease, ) patients water content in lungs. These radioactive tracers with LBBB, RBBB and WPW syndrome, 101 were injected into 11 dogs with injured 1ngs. hypertrophic, virus and dilated myocar- Blood samples were drawn and the animals were sac- ditis. 19 cases of coronary patient rificed. The lungs were removed, weighed and homo- were performed by coronary angiograph,; genized. Samples of blood and lung homogenate were 27 cases were performed at rest and assayed for I-131 and Tc-99m. Samples also were with stress. The results of phase image weighed before and after drying to a constant analysis at rest and with stress were weight at 70-75°C. Extravascular lung water was analyzed. determined by the dual-isotope technique and again The sensitivity was 92 %, the specifi- by gravimetric analysis. The average ratio of the city was 28.6 %, but phase image analy- results f-o.,1 the two different methods is 1.14 ± sis with stress was more sensitive and 0.20. The two methods were compared also by better than LVEF in the early diagnosis regression analysis and the correlation coefficient of CAD. Abnormal phase angle was visua- is 0.97 ± 0.09. This investigation suggests the :i7:d in noncoronary patients. possibility of in vivo measurement of lung water with the equilibrium distribution of iodo-antipyrine.

This work was supported by a research grant from the Vete..ans Administration.

HP17.26 MP17.27 Brain Scanning with 99m-Technetium Glucohep- The 160(a,pn)18F and the 18Uip,n)18F channel:4 az tonate and quantitation of specific uptakes in attennativez On the production o6 "F- toed .en tumors * tadiochemicat &betting. A. Hermanne M. Gysemans, R.Ravichandran,S.Devaru,M.Ananthal,B.S.Das2,V.K.J N. Walravens, P. Van den Winkel, W. Vanryckeghem ain3 1) KIMIO,Bangalore.29,2) NIM- and L. De Vis. Vrije Universiteit Brussel, HANS,Bangalorc29,3PIMAS,Dclhi.7,India. B-1090 Brussels, Belgium. Our earlier At the VUB Cyclotron (Ea 47MeV ; work hasshowed that conbination of 2-ieoxy-D- max = Ep,max 45 MeV) two nuclear reactions areused for carrier glucose (2DG) with ionising radiationsis ex- free 18F- production : pected to improve therapeutic efficacy by inhibi- - the rarely cited 160(a,pn)18F tion of DNA and cellular repairprocess in cells with Etreshold = 25 MeV. with high glycolic activity. Clinical trialsare - the now common 180(p,n)18 18 planned in brain tumor radiotherapy with 2DG. For F on 0 enriched targets Real competitivity of these channels for let label- selection of patients and monitoring themwe proposed to use 99m-Technitiqm glucoheptonate ling experiments depends on both the isotope production yield and the ultimate incorporation effi- (TcGH) a3 glucose analogue in the place of ciency. Data on the following physical and chemical positron emission tomography. The specific uptakes of ToGH were quantitated in human brain tumors characteristics for both reactions will he presented * Excitation curves (a from 25 to 47 MeV ; using a conventional gonna camera and measured p from i0 to 25 MeV). build up factors. * Relation between the incident beam intensity and The specific TcGH uptakes were expressed as ratios saturation yield ; correlation with loss of tar- against normal brainto account for variations. get material. Theratios of specific TcGli uptakes correlated * Levels of short (150) and longer lived radioiso- well with the grades of brain tumors, 4.4+0.7(n=8) sler, seGa, 67Ga, topic contaminants (46V ue). for malignant,2.40.9(n=6) for metastatic and * Labelling yields for a standard 18F exchange 1.6+0.6(n=8)for benign tumors. Concentration of reaction and influence of foil breakdown resul- TcGH in viable parts of recurrent high grade ting in depassivation of the target. tumor, increased ratios for TcGH against similar ratios for Tc-RUC blood pnol scans suggest that the concentration of TcGH in brain tumors may be related to rotabolism. 90

68 1988 World Congress on Medical Physics and Biomedical Engincering San Antonio, Texas

I ,IIIIMPRPLOLVIMENPROMRITIM 1111111111101101110r

Medical Physics Scientific Papers

MP17.28 MP17.29 Validity of the assumption of tracer equili- Internal Dose Estimation from Skin Dose Measurements brium with native Vitamin B1, pool,C.M.Pathakt using TLD in 18-FDG Positron Emission Tomography R.R.Sharma and K.C.Das,BiopfiYsics Department, (PET). J.F.D.Chubaci* & S.Watanabe (Univ.Sao Paulo, CP.20516, 01498-Sao Paulo, Brasil); H.Fukuda, H. PGIMER,Chandigarh,India. Seino, Y. Ota, T.Matsuzawa, M.Ito & J. Hatazawa (To- Serum Vit B level does not reflect native Vit B12 hoku University, Japan). 12 pool, since it gets altered in various pathological conditions.Tbtal body vit B12 can beestimated A method has been developed to estimate the from specific activity measurement of a biopsy sam- dose on target organs of PET patients by ple provided tracer equilibrates with native pool. Thermoluminescence dosimetry. TL dosimeters were held on the skin, close to the target organs, of It is generally assumed that whole body monoexpon- ten Japanese subjects with surgical tape, a few mi- ential fall of tracer reflects equilibrium state. nutes before the intravenous administration of This study Lritically examines the validityof 2- (F- 18)Fluoro- 2- Dcoxy -D- Glucose The skin doses this controvertial assumption. due to the tget organs were obtained by reading Whole body kinetic parameters in rats using57 Co- Vit Bi,(s.c)showed monoexponential decline of tra- out the TL dosimeters and extrapolating the values for 37.0 MBq of activity and for 658.8 minutes of cer atter about 32 days,whereas inter organdistri- exposure time, whie.1 is equal to six half lives bution of tracer did not stabilise even upto64 of Fluorine-18. The internal dose, on various days,reculting in continuous change of specific organs, were estimated using the organ/skin radioactivity in organs.Thus whole body Vit B12 values nuclide concentration ratio. Thi. ratio was calculated from different organs at 64 days ranged calculated from printed PET matrix data for each between 3.9 pg(kidney)to 89.3 pg(muscle)with mean scanned organ, by a method developed through tissue of 37.5 pg. Whole body turn over rate of1.47% per equivalent gelatine phantom studied. The largest day probably represents an average of wide spect- internal dose obtained was 41.3 pGy/MBq in the organs.Variation rum of turnover rates in different bladder, tnen 34.5 pGy/MBq for the spleen; other of T1 biol. of tracer in various organs may be values are for liver, heart, lung, brain, pancreas, related to interconversion of Vit B 12 into other kidneys and testes. The value3 found in the present forms and their metabolic functions. This study work are in close accordance .7ith previously the shows that assumption of equilibrium between reported works, showing that our method may be of tracer and the endogenous pool is not valid even atter tracer is excreted- monoexponentially. great usefulness to evaluate absorbed dose with good accuracy for emission tomography.

MP17.30 MP17.31 Bone Mineral DensilomeIty With X-Ray Sources, Dual Photon Absorptiome try of Forearm M. Mori', D. P. Chakraborty M. V. Pester, and G. T. Barnes, Using Monochromatic X-Ray. T.Tam eg a i,K.Taura,S.Ki mura,K.ka w arn ura, Dept. of Radiology, University of Alabama at S.Ninomiya, ALM Co.,Ltd.. Tokyo, JAPAN Birmingham, 619 19th Street South, Birmingham, AL 35233 The newly developed dual photon absorptiorn- etry(DPA)bone mineralcontent measuring instrument is reported. The instrument con- Two methods of oual photon absorbtiometry utilizing an x-ray oi area scanningmechanismforforearm, tube instead of a Gd-153 source have recently been X-ray generator, diffraction unit, two Nal(T1) introduced. In one method the kVp is switched at each pixel, in scintillation detectors and a smallcomputer. Thecollimated X-raybeam fromconventional the other, a tingle kVp spectra is used, but photons in two X-ray tube irradiate the difIraction gratings energy windows are counted. We will present a theoretical and a part of the diffracted ray are collimat- study of the two methods for a typical patient (20gicm2 of soft- ed through pinhole collimators and extracted beams. The tissue and 1 g/cm2 of bone), with respect to the achievable as the monochromatic X-r system uses two energy beams 27 -411(-1 53 keV,so works as coefficient of variance of thebonemineral count (CV) and the dual photon absorptiometry. Sufficientbeam dose. Our study indicates that while the single kVp method is intensity forabsorptiometryare obtained at more dose efficient, the kVp switching method is not count rate the conditions of 70 kV and5mAon the X-ray Basic studies on theinstrument are limited and can therefore achieve a smaller CV. Our model is tube. performed including photon energy identifi- based on published x-ray spectra and attenuation coefficients cation,beamintensity and stability,linearit> and is in good agreement with data.It is of interest that a point of bone mineral Lontent, long and short term source of Gd-153 would be even more dose efficient. However, precision andcomparison withconventional single photonabsorptiometry,Results of those in practice the source is distributed and of limited strength, and studies suggest that the uistrurlent can be use consequently the radionuclide method cannot achieve a CV as asDPAof forearm. Also the instrument eliminate small as either x-ray method in similar scan times. theexpenseof frequent repla-ing radioisotope source and decrease scanning cost. Thesystem Supported in part by NCI Grant No. CA43615-02. is expected as theuseful instrument. f) 4 ai

1988 World Congress on Medical Phy tics and Biomedical EngineeringSan Antonio, Texas 69 Medical Physics Scientific Papers

MP17.32 MP18.1 Is detailed biodistribution data necessary Logistic for nuclear medicine dosimetry? Parameters CompAefLatith_Comtional W. Huda* Sensitometry for Radiograoi* and G.A. Sandison, Manitoba Cancer Foundation, Film-Screens, C. E. Willis* (KRUG) and J. A. 100 Olivia Street., Winnipeg, Canada R3E OV9. Bencomo (UTSCC), KRUG International, 1290 Hercules, Houston,TX 77058

Using the concept of effective dose equivalent, HE, Analysis of CDRH sensitometric it is shown that a knowledge of detailed biodistrI- data confirms empirical observations that the Ntion data in most organs and tissues taking up film characteristic curve is a logistic function of exposure. "mTc labelled radiopharmaceuticals isunnecessary. Data for seven commercial radiographic emulsions using Dosimetric precision of about + 25% can be readily four intensifying screens were normalized to obtained from urine excretion data alone providing obtain fractional response between there is no significant uptake within the gonads fog and saturation. Saturation was determined from the or thyroid. Special attention should only be paid graph of absorbance versus the inverse ofexposure. to absorbed dose measurements in red bone marrow, Norma- lized dataweretransformed into logistic units skin, lungs, gonads and thyroid, with the great-st (Logits) and plotted versus the log of exposure. attention directed toward the retention and dosi- These graphs are linear from <2%to >98% fraction- metric aspects of radioactivity in the gonads and al absorbance, or a factor of 100 in exposure. thyroid glands. An examrle of the calculation of Regression analysiswas conducted to determine H for 99mTc labelled d,l-HM-PAO is presented dem- E logistic sensitometric parametersbased onstrating the importance of these specific organs on the explicit statistical weights oflogistic data. The to radiation risk. Using biodistribution data logistic slope, or contrast analog,had an average obtained in rats, the HE value for a 20 mCi of 1.95 +/- 0.19. The logistic intercept (value administered activity was computed to be 880 mrem, of exposure required to produce 50% of saturation of which 62% was due to the risk of genetic absorbance) or speedanalogvaried effects. from 11.2 to 36.3 units of relative exposure. Previous reports show contrast analogis unaffected by development conditions, however, speed analog, saturation absorbance, and fog dependon degree of development. Tabulation of logistic parameters f"r photographic emulsions couldsimplify catalo9,as of sensitometric data. The linear logistic representation facilitates quantitativemeasurements.

MP18.2 MP18.3 Simple QC Test Tools for DiagnosticRadiology, Computation of Off-Axis X-Ray Spectra,D.M. P. Mora*, Escuela de Fisica, Universidadde Costa Tucker', D. P. Chakraborty, and Rica, San Jose, Costa Rica and J. Cameron & G. T. Barnes, J. Knoche, Univ. of Wisconsin, Madison, WI Dept. of Radiology, University of Alabamaat 53706, U.S.A. Birmingham, 619 19th Street South, Birmingham, AL 35233 Hospitals in developing countriesc not usually affordsophisticated qualitycontrol (QC) instruments and do nothave the trained pro- Birch and Marshall's1 method of fessionals to use them. computing x-ray spectra has been extended to include off-axis(i.e., "heel") effects. Both Over the last five years variou. simple QCtest bremsstrahlung and characteristicspectral components are tools for useby radiographers have been included. With knowledge of the developed, primarily at x-ray tube anode material and the University of angle, kV and mA waveforms, Wisconsin. The total cost of the materials fora and inherent and added filtration, set of the test tools is anout $100 U.S. About the model allows ore to predictthe primary x-ray beam intensity ten hours of semi-skilled labor is needed for and photon energy distribution their at any point in the radiation field. construction. TheQC tools measure: Predicted and experimental 1) quantitative image quality for general x-ray beam intensities and qualities radio- versus off-axis angle are compared graphy, fluoroscopy, mammography, angiography, for different x-ray tube and tomography; 2) sizeof the focal spot; voltages, filtrations and targetmaterials and target angles. The 3) accuracy of the timer; 4) accuracy of theset agreement between the two is typically kV; 5) collimator function; 6) alignment of within 1-2%. Also the discussed am potential apHications grid; 7) film-screen contact; and 8)consivercy of the model in dual-energy of the output for standard exposure conditivas. imaging and patient dosimetry. The devices were designed to be used by 1R. Birch and M.Marshall, Phys. Med. Biol. 29, 505 radiographers. A set of the test toolsare being (1979) evaluated by radiographers in several major Supported in part by NCI GrantNo. CA43615-02. hospitals inCostaRica following a two day courseon their use. The test tools will be described and their useunder field conditions will be discussed. 2

70 1988 World Congress on Medical Physics and BiomedicalEngineeringSan Antonio, Texas Medical Physics Scientific Papers

MP18.4 MP18.5 Optimization of Cine Angiography Analyses of Different Diagnostic Using Variaable Input Exposures, Xray Procedures Used in Switzerland J.S. Whitingand J.M. Pfaff, R. Mini, Abt.f. med. Strahlenphysik, Cedars-Sinai Medical Center, Inselspital, CH-3010 Bern Los Angeles, California, USA. In many countries thediagnostic X-ray We performed an analysis to investig,,,te the use of procedures gives the largest man-made a variable camera aperture to optim'ze signal contribution t., the population dose. detection in coronary angiography. Signal/noise It is therefore important to analyse (S/N), contrast, image intensifier exposure and skin the different medical procedures in entrance exposure (D) were calculated for detec- order to see which one is the most tion of a 0.5 mm cube of 370 mg/nil iodine in significant with regard to the radiation water thickness ranging from 14 to 28 cm and risks. The individual organ doses, x-ray tube potentials from 50 to 120 kV.Assum- resulting from a specific procedure ing 34 kW iso-power operation, S/N reached a depend very much upon the radiation maximum at 95 kV for all tissue thicknesses, but technique which may vary from doctor varied by less than 10% from 65 kV to 120 kV, to doctor. Therefore our Departement while maximum (S/N)2/D occurred between 60-70 of Radiation Physics has made an extensive kV. Skin exposure increased linearly and contrast survey in order to assess the variability decreased exponentially with increasing tube of the technical parameters used for potential for fixed tissue thickness.When patient the different diagnostic procedures. thickness drives x-ray tube potential above 85 kV The data of 200 doctors where analysed. skin exposure may be reduced 70% with less than As expected we found that even in a 10% penalty in S/N by reducing image intensifier small country like Switzerland there input exposure 50%.Film exposure can be main- is a big variation in the used procedure tained by compensatory opening of the camera techniques. This makes it difficult aperture one f-stop.We conclude that incorpora- to establish mean dose values which tion of aperture control into automatic exposure could be representative for the Swiss control systems can significantly minimize patient population. dosc and/or tube loading without diminishing image quality.

MP18.6 MP18.7 Reject Analysis in Diagnostic Radiology Complexity i Complexity Contrast in Radiological 1.ges. R. Reed: W.M. Gentles, W. Ho, U. Giugni,R.Gallini,S.Belletti:V.Berna V.V. Velocci, Sunnybrook Medical Centre, Spedali Civili,Universita' degli Studi, University of Toronto, Toronto, Ontario, M4N 3M5, Canada. 1-25100 Brescia, Italy Radiological images can be generalized as Reject analysis is an important part of any Quali- containingstructures of interest their surround ty Assurance ;QA) programme in diagnostic radiolo- and noise. The surroundingand overlaying anatomical structures have a masking effect, gy;this analysis provide', a means of early detec- reducing the conspicuity of a known target in a tion of changes in reject rates,determines the ba- scene. If the characteristics of the surrounding anatomical structures are varied, the target seline for a QA programme and monitors the effecti- conspicuity will vary, becoming small as the veness of that programme. target and surround become similar. A model is presented which provides an To this end we collected 8061 films (2303 examina- analytical method of measuringthe amount of tions),coming from 7 radiological depts. (2 belon- anatomical structure or anatomical noise in a or scene. This leads to the definition of median size ho- target ging to regional hospitals,3 to complexity. The theory also provides a method of spitals,1 to a pediatric hospital,1 to a private determining if frequency dependent Image processingcan improve the conspicuity of a known institution).We choose the dimension of each sample target in a surround of known complexity. according to the annual workload ID': the dept. The model is based on two assumptions: 1) The We considered 588 spoilt radiographs,determining 17 complexityof a structure or surround is related to its spatial frequency content, and 2) a figure different categories of "reason".Some categories representing complexity is derived from a give less than 1% of msted films,whereas other five linearly-weighted function of power spectrum. Asecond figure of merit named complexity reach a percentage of 10% or more. contrast is introduced. It gives ameasure of This investigation represents the starting point complexity differeuces between a target and its surround. It is shown that in the limit of of a QA programme,which also involves a patient do- frequency approaching zero, complexity contrast se survey and an imaging quality assessment. approaches the definition of optical contrast. Complexity contrast is related to a form of SNR between a target and the anatomical structure (ie noise) of the surround. Research supported by Regione Lombardia (Italy), graat N° 482 93 1988 World Corgress on Medical Physics and Biomedical EngineeringSan Antonio, Texas 71

I- Medical Physics Scientific Papers

1P18.8 MP18.9 Comparison of different types of collimator design used in Performance Evaluation of Gamma dental radiography - R. Badiellol T. Bernardi and C. Passeri- Cameras in the Philippines, E. ni - C.N.R. Servizio di Sicurezza del Lavoro e Protezione Sa- Valdezco, Phil. Nuclear Research nitaria; Ni'versitb, Servizio di Fisica Sanitaria, Bologna, Inst., Diliman, Quezon City and A. Lobriguito,Radiation Health Service Italy Sta. Cruz, 1003 manila,Philippines

The continuous increase in the use of X-rays in dental practi The gamma cameras in the Philippines vary in ce calls for an accurate analysis of the radiation protection ages. This paper presents the results obtained of both the patients and the workers. In the present comtuni- from the quality control tests done on these cameras, adopting the International Atomic cation a comparison iz made of different types of collimator Energy Agency standards. The physical parameters design used in dental X-rays apparatus (cone, cylinder, paral- measured are the intrinsic uniformity, intrin - lelepiped)in view to reduce the dose to the patient andto in- sic resolution, intrinsic linearity, count rate crease the quality of the imaging. It is well known that the loss, energyresolution, maximum count rate open ended collimators should be preferred to those that have and sensitivity. Several measurements are done plastic ends because these act as a source of scattered radia- for every parameter and the performance of the gamma camera is evaluated using the mean value, tion. The paralletepiped shape offers further advantage with to determine the clinical acceptability of the respect the cylinder one for what concerns the radiation pro- camera in obtaining maximum diagnostic infor - tection of O. patient. In fact the irradiated volume is stron mation. The current mean standard (CMS) is der- gly reduced (by a factor two) with a consequent reduction of ived and the performance of the camera is com- the absorbed energy at the same exposition. The skin dose is pared with the CMS. It is hoped that, with this paper, minimum also reduced since the retrodiffusion factor is smaller in performance standards will be established and will the case of the rectangular area. Moreover such a type of co' be used nationwide. limator contributes to a better definition of the radiological image. The results are discussed and applied to the estima to of the collective dose equivalent for oral diagnostic radio logy.

MP18.10 MP18.11 EVALUATION OF SCREEN-FILM The Scatter Point Spread Functionin COMBINATIONS FOR MAMMOGRAPHY Diagnostic Radiology, JM Boone* and Marlene H. McKetty, Ph.D. JA Seibert"-, Departments ofRadiology, Howard University Hospital Thomas Jefferson piversity, Phila- The purpose of this study was to evaluate films delphia, PA and University of Cali- for use in screen-film mammography from twomanu- fornia-Davis, Sacramento, CA. facturers and to determine if the use of a medium -spend screen, that is now availelle, provides Monte Carlo techniques were used toevaluate the images that would be acceptable to radiologists point spread function of scatteredradiation in in our departent. Breast imaging is difficult diagnostic radiology.A homogeneous model of because extremely small structures must be infinite lateral extent and ofvarious thick- visualized and because much of the tissues have nesses was analyzed using an ideal detector; similar attenuation coefficients and produce this simplified model adheres to therequisites little subject contrast. As the variety of of linearity and stationarity. 80, 100 and 120 mammography films and screens increases, mammo- kVp team spectra were used in thisstudy, with graphy facilities must choose carefully among air gaps of 0, 2 and 4 cm. An analytical equa- them to ensure that they obtain the best image tion was derived using four physicalparameters quality at low doses. as input--object thickness, air gap and the attenuation coefficients of bothprimary and Several images were made of two mammography phan- scattered radiation. Though the analytical de- toms using different types of film and differing rivation assumes only single scattering,reducing speed Kodak Min-R Screens. Techniques were P(scatter) by about 10% results in varied to produce the highest quality images excellent correlation with the Monte Carlo results. at reasonable 4'es. The images were evaluated by three observers for parameters suchas density, contrast and visibility of test objects in the phantoms. The data to be presented show that the single emulsion films from bothmanu- facturers were equally acceptable to the observers and that the double emulsion film was not acceptable for routine use.

72 1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas Medical Physics Scientific Papers

MP18.12 14P18.13 An Image Data Compression System for X-Ray Tissue-Equivalent Phantoms

Medical Image, Wen Gaeand Masao Saito, P. Dvorak, G.R. Symonds; G.R.J. Robertson Saito Lab Inst of Ned ElecUniversity of and B.M. Keller Bureau of Radiation and Medical Devices Tokyo Japan Department of National Health and Welfare Ottawa, Canada, KIA OL2

In PAC:3 image data compression is one of Phantome designed to simulate the patient in important fccnktions because of t..ices from trans quality control tests,n diagnostic X-ray equipment should ideally match the tissue in energy mission and storage of medical image,, An design depend^nce of X-ray absorption and scattering of medical X-ray image data compression system properties, and be sufficiently light and smell to be handled easily. These two requirements are named ADCTVQ (Adaptive Discrete Cosine Transform difficult to meet simultaneously. A computer Vector Quantization) 's presented in this paper model to evaluate the best combinations of materials for the suitability as tissue- ADCTVQ uses adaptive bit allocation scheme for equivalent phantoms has been developed and OCT efficients matrix and vector quantization solutions suitable for narrow beam measurements using an ionization chamber, TLDs, films or technique By the adaptive bit allocation scheme similar detectors have been found.A simple a result of average 1dB up can be gellen andby experimental verification of this approach has been performed with encouraging results. vector quantization an effect of minimal quantiza-

tion distortion can be made . Simulation results

show thatfor X-ray chest images the AUCTVQ can give receptabte image with a compression radio

not less than 12. Consideration of the real-time requirement ADCTVQ is constructed into parallel

structure

MP19.1 ttP19.2 OPTIMIZATION OF PROTON AND HEAVY ION A Test of DraggGray Cavity Theory for Protons THERAPY USING AN ADAPTIVE INVERSION Verhey, Dept. of Radiation Medicine, ALGORITHM, A. Brahme*, P. KM1man and Mass. General Hospital, Boston, MA 02114 B. Lind, Dept. of Radiation Physics, The Karolinska Institute, Box 60204, An ideal Bragg-Gray cavity is defined as one in which S-I04 01 Stockholm, Sweden all energy loss is local, the cavity Is much smaller than the range of secondaries and the production of secondaries In the cavity Is negligible.For a thimble To fully utilize the dose distributionaladvan- ion chamber in aphoton beam In charged particle tages ofpions,protons and heavy ions it is equilibrium, these criteria are well satisfied.in the necessary to cover the targetvolume as well as case of a proton beam, however, the ranges of electron possibleby Bragg peaksand avoid irradiating secondaries are typically less than 1 mm of air and the surrounding organs at risk. The most general way proton interaction probability per mm of gas is very to achieve thisgoal is by three-dimensional high, so Bragg-Gray cavity theory cannot be used scanning of the beam so that its Bragg peakfall without modification. exclusively in the target volume. We have investigated the significance of this affect by comparing the responses In proton and Co60 beams of We have recently deve'oped a very powerful iter- ion chambers with a varietyof geometries and ative inversion algorithm capable of calculating dimensions, including 0.1cc, 1.0=, and 80cc spherical the optimal scanning pattern of the proton beam chambers, and a thin-window parallel-plate chamber. required to get the desired dose distribution in The ratio of responses for the same gar In the chambers the target volume. From this point of viewthe of widely different surfaceto volume rutro-, can present technique implies a reversal comparedto quantify the departure from classic Bragg-Gray theory. conventional treatment planning, where treatment The ratios of 1.4-2tnn responses with different gases in fields generally must be chosen by a trialand the same chamber gives us a way of investigating W error-like approach to maximize tumor dose and values for protons which are needed for routine minimize normal tissue irradiation. Typical dose dosimetry. The results of these measurements and their plans will he presented. significance for proton dosimetry will be discussed.

1988 World Congress on Medical Physics and Biomedic &1 EngineeringSan Antonio, Texas 73 Medical Physics Scientific Papers

MP19.3 MP19.4 Computed microdosimetry of epithermal Investigation of Alternative Dose Calculation Alyo- neutrons in selected biological materials. rithms for Charged- Particle The -apy. P.L. Petti, Lawrence R.E. Wallace* and P.Bloch,Univ of Penna, Dept of Radiation Onculogy,Phila., Pa. Berkeley Lab, Berkeley, CA 94720 In tissue-equivalent materials,epithermal neutrons generallyimpart dose via Dose calculations for charged-particle radiotherapy typ- elastic nuclear collisions. These reac- ically use ray-tracing techniques to calculate thewater- tions produce lowenergy charged recoil equivalent pathlength along the beam direction through nuclei which are cnaracterized by their a matrix of CT data. Although these calculationspro- low lineal energy, typically-10keV/pm for lkeV neutrrns. At low neutron energies, vide accurate dose distributions in horngeneousor sim- reactions producing high-LETparticles can ple heterogeneous systems,as Urie et al.'have dis- occur in tissue materials. For example, cussed, errors due to multiple scatteringarise in more the N(np) reaction produces a 600keV pro- complex heterogeneous geometries. Thepurpose of ton with a dose average linealenergy of 85keV/pm in muscle. For 100ev neutrons, investigation is to explore the possibility of using meth- this reaction accounts for5.7% of KERMA ods similar to the "differential- pencil -bear technique while the 2.2MeV y-raysfrom neutron cap- described by Mohan etal.,2which rely on p.calculated ture and elastic recoils inhydrogen con- Monte Carlo data, to calculate charged-particle dose tribute nearly 94%. The high-LET N(n,p) distributions. Calculations based protons, have higher RBE than they-rays or on such an algorithm the recoils. For example, using single will be compared to Monte Caro simulationsfcr se- fraction survivalcurves for human kidney lected heterogeneous systems. Since the distanceover cells at the 1% levelas a dose response which adjacent particle tracks contributeto the dose archetype(1), -28% of the totaleffective at any given point is small for charged particles, such dose for neutrons below 200keVis delivered by the N(np) protons. Computed micro- techniques may not be prohibitively time consuming. dosimetric results in brain,muscle, fat, This work is supported by NCI grant 1CA19138. and bone will be presented. The microdosimetry of 5eutron capture inboron and lUrie M, M Goitein, WR Holley, GTY Chen, Phys. Med. lithium infused tissues will bediscussed. 31, 1-15, 1986 (1) Barendsen, G.W., Los Alamos Laboratory 2Mohan, It, C Chui, L Lidofeky, Med. Phys.13, 64-73, 1986 report ALP- 5180C, p. 120, 1972.

MP19.5 MP19.6 Dosimetric Characterization of Helium Ion Beams OptimumThickness of Li and Be Targets Used in Radiotherapy at the Bevatron, J.M. Collier: T. for the Production of Therapy Neutrons fal Renner, S. Henderson, P.L. Petti, S. Pit luck, Research Protons and Deuterons of Plfferent Energies, M.A.Chaudhri*, Austin Hospital, Medicine and Radiation Biophysics Div., Lawrence Berke- and University of Melbourne, Melbourne, ley Lab CA Berkeley 94720 Australia.

Because of the dismantling of the 184-inch cyclotron It was suggested by us in 1970's that, at the Lawrence Berkeley Laboratory, the heliumion from light elements (1,1,Be,etc.),protons from a modern cyclotron would produce radiotherapy has been transferred to the Bevatron.Ini- -ore tially, the treatments are being carried penetrant neutron beams for therapy out using the -ban deuterons from the same machine,and "Wobbler" dipole magnets to spread the beam laterally that moderately thick targets would in circles up to 30 cm diameter and plastic "propellers" produce neutrons with higher moan to spread the particle stopping region in depth. This is energies than thick targets. In this paper in the room currently used for Neon ion radiotherapy we are presenting results of our calculations on neutron production also. However, a new treatment from room is being built. In Li and Be targets anddifferent backup this room will be placed the unique patientpositioner materials. Cyclotrons of different from the cyclotron. A double scattering system will be sizes, and backing matenals of Cuand Ta employed to create the lateral spread. Dosimetric char- , have been considered. On tho basis of our results, it has been suggested that acterization of the beam in the first room has been done moderately thick targets of Li and Be using film and ionization chamber, includinga 144 el- would not only produce morepenetrant ement linear array chamber in which each element is 2 neutrons, but also enough in.ensity tc mm wide by 1 cm long. Results will be shown which carry out neutron therapy. Moreover, as the give the slope of the spread out Bragg peak, he distal energy absorbed in such a target is smaller than in a thick target, the heat gradient (3-4 mm for 90% to 10% drop off), and lateral dissipation problems would bemore easily penumbra for different treatment situations. Supported manageable. by NCI Grant 1CA19138. 96 74 1988 World Congress on Medical Physics r.nd Biomedical EngineeringSan Antonio, Texas Medical Physics Scientific Peners

MP19.7 MP19.8 Heavy Ion Medical Accelerator in Chiba, MicrOdosimetrie Measurements of a K.Kawachi*, H.Clgawa, T.Kanai, M.Endo and High-Energy Neutron Therapy Beam, Y.Hirao,National Institute of Radio- S.H. Benedict *, J.B. Smathers, L.T. logical Sciences, Chiba-shi 260, JAPAN. Myers, R.E. Wallace, University of California, Los Angeles U.S.A.

Microdosimetric measurements were made in a A project of Heavy Ion Medical Accelerator in neutron therapy beam using a Rossi propor- Chiba (HIMAC) has been started since 1984 in tionalcounter witha 1.50-cm wallof Japan. The final design studies of the HIMAC A-150 T.E. plastic. Previous microdosimet- including the buildings have been oampleted. ric studies have been performed on neutron The HIMAC consists of ion sources, RFQ linac, beams of14 and 15MeV; howeer, few Alvarez linac, synchrotron, high energy beam studieshave been performed at higher transport system and irradiation facilities. energies. The CP-45 isochronous cyclotron The main accelerator synchrotron consists of at the UCLA NeutronTherapy Facility double rings which are operated with the generates a neutron beam with 46 MeV pro- repetition rate of 0.5Hz, and with the phase tons that bombard a thin Be and C target. Both the target design and collimation diffeYence of a half period each other.The system are unique to the facility. maximum output energy of synchrotron is 800 The charged-particle energy loss spectra MeV/u for the ions with the charge to mass were measured and compared atvarious ratio of 0.5. depths in phantom, in the penumbra reaion, The HIMAC has both horizontal arrl vertical andoutside thebeamusing standard beam delivery courses. The sophisticated ir- techniques. The calculation of lineal radiation facilities cNonsist of beam spreading energy distributions and gamma components system, range modulating system, field shaping from the spectrums was performed tsing system, beam monitoring sysban and patient's established calculational techniques. positioning system. A comparison of the microdosimetric data of The HIMAC will be completed in 1993 and be able the neutron therapy beam properties of the UCLA facility to that of other fast neutron provide the beams for the clinical trials. to therapyfacilities is presented. This research is supported by NCI contract CM 97315.

MD19.9 MP20.1 SPATIAL DISTRIBUTION OF FREE RADICALS Quantitative imaging of acoustic backscatter coefficients, E. Boote; J. Zagzebski, E. PRODUCED BY NEUTRONS AHD ?SR DOSIMETRY, Madsen, T.F.all and M. Pozniak, Univ. of C. Franconi*,K. V. Ettinger, 3. Holowacz#, Wisconsin, Madison, WI, 53706 1I Universita di Roma, Italy Current clinical ultrasound scanners provide only qualitative inforoation on acoustical properties # On leave from Politechnika Gdadska, of tissue becauseoperator, instrument and Gdansk, Poland transmission path dependencies are inadequately accounted for during image formation. we have Determination of absorbed dose in ESR dosimetry developed a technique for producing quantitative, is based on measurement of the concentration of instrument independent images of the acoustic free radicals in the solid used as a dosemeter, backscatter coefficient and are implementing this on a clinical scanner.Our imaging algorithms are e.g. a polycrystalline amino acid.At the same derived from a previously reported method of data time it appears possible to estimate in the same reduction whichrigorously accountsfor all sample the spatial distribution of radicals. This experimental factors in a pulse-echo experiment, estimation is based on the known dependence of including the transducer pressure field, pulser- the microwave powersaturation effect inESR receiver electronics and attenuation, yielding a spectrometry on the value of spin-spin re.axation measure of the intrinsic scattering properties of constant. This relaxation constant is,in turn, a a medium. A spatial mapping of the reduced data simple function of the spin density. The results in a grey scale image, where the experimental ratios of freeradicaldensities brightness of each pixel is associated with a produced by 7.6 MeV neutrons and 1.25 MeV gamma unique small range of oackscatter coefficients. rays are within 1.5 - 4.0 .If reactor neutrons To implement this techniqueon a real-time sector scanner, a high speed transient recorder are comparedwith rays, the 1.25 MeV gamma and a memory buffer digitizes and stores echo rf ratios of radical densities are within 1.4 - 3.3 signal for the entire image during a sweep of the Inorder toobtain theabsolute values of transducer. The digitized data are sent to a radical densities from the saturation curves it computer for off-line analysia and ioage display. is necessary to know the intensity of microwave Initial tests of the system produced accurate magnetic field .n the measured sample.This is backscatter coefficient images of tissue-mimicking accomplished by means of perturbation technique phantoms containing regions of varying scatter in which a minuscule scatterer is introduced into levels. Quantitative images have been formed of the cavity. The ensuing shiftof the cavity in vitro samples of canine kidney (normal and resonance is a measure of the field intensity. transplant rejection) and human liver as well as of abdominal organs in vivo. 5 7 1988 World Congress on Medical Physics and Biomedical Engineering o San Antonio, Texas 75 Medical Physics Scientific Papers

MP20.2 MP20.3 High resolution imaging through imagereconstruction Correlation of Ultrasonic Attenuation with of axially deconvolved B-Scans. Pathologic Fat and Fibrosis in Liver Disease, A. Herment, P. Pironneau: G. Demoment T. Lin, J. Ophir*, G. Potter, Department of INSERM U. 256, Hopital Broussais, Paris, France. Radiology, University of Texas Medical School, 6431 Fannin St., Houston, Texas, USA In ultrasound imaging, where reflection images of a same structure can often be acquiredunder The attenuation coefficient in two groups (n1 = 70; various incidences, a methodhasbeen developed n2 = 59) of pathologically graded in vitro human for acquisition and processing of data. Ii.consists liver specimens was measured over a frequency in : range from 1.25-8.0 MHz and fitted to a power law . Acquisition of R.F. B-S..ans model. The slopes of the power law curves at 5MHz . Line by line 1D deconvolution of these scans were correlated with pathological score for fat in . Video detection Group I, which consisted of normal and fatty . Image reconstruction by a filtered back projection livers (no fibrosis); they were also correlated process. with the pathological score for fibrosis in Group A study on simulated data and phantoms forS/N II, which consisted of normal and fibrotic livers ratio ranging from 50 to 5 dB will be discussed. (no intracellular fat). Significant (p<0.002) As compared to2D deconvolution, for a typical differences were found between the two groups. The (256 x 256) image reconstructed from 16 scans fatty liver group exhibited approximately a 0.23 + of 32 lines each distributed on 120°, theprocess 0.06 dB cm-1 MHz-1 (fat grade)-1 behavior (mean + shows : s.d.), while the fibrotic liver group exhibited . A gain of 10 on processing time approximately a 0.11 ÷ 0.01 dB cm1 MHz-I (fi- . A gain of 4 on S/N ratio brosis grade) -1 behavior. These results may ex- . A p.in of 5 on image definition. plain some of the conflicting literature in this Works are now in progress to assess the influence area, and suggest that attenuation may in princi- ofthe variable propagationspeed in tissues on ple be used for screening for liver disease, but the image quality. not for differentiation between fatty and fibrotic conditions. Supported by National Institutes of Health grants #R01 CA38515 and ROI CA44389.

MP20.4 MP20.5 Comparison of a Visual and A Digital BONE F'OPERTIES AND ACOUSTO-ULTRASONICBEHAVIOR Measure of Relative Echogenicity Applied A. Sivas, I. Leichter*, A. Mittelman, and to Fetal Lung and Liver, P. Carson*, A. Weinrb, Racah Institue of Physicsand School of M. DuPuy, C. Meyer, R. Bowerman, E. Chiang, AppliedSciences Hebrew University ofJerusalem, R. Hayashi, University of Michigan, and the Jerusalem Osteoporosis Center Samples of Ann Arbor, MI48109-0553 equal size of three groups offemoral heads were investigated: 32 from normalcadaveric bones without known bonepathology; Relative echogenicity is a key source of 39 from osteoporotics (OP) (5 of them diabetic), and7 from information in ultrasound images.An algorithm for osteoarthritic patients (OA). quantitative measurement of relative echogenicity Foreach sample three parameters weremeasured: was implemented, which corrects for digitally 1) The bone density (BD) ingrams /cm3was determined by the Compton scattering technique. estimated differences in attenuation coefficients 2) The bone mineralcontent (BMCv) of two tissues with presumably identical overlying was determined by the NorlandBoneMineral Analayzer. The tissues. resultsin This measure V.F. relative echogenicity was gr/omzwer., divided by the bode heightto obtain the BMCv studied on fetal lungs and livers in a sample of ingr/cml 3) The peak amplitude (PKA) ofa 172 digital images from 8 third trimester fetuses. transmitted ultrasonic pulse and the pulseduration Two independent digital data sorters agreed well (PD) were measured by a AET 5000system. From these data the with each other; the 95% confidence limits for relaxation constant (RC) of the one exponentially decaying pulse was calculated. sorter predicting the other's next data point on The PKA and RC values were the same for theOA and individual images and on average values for a fetus OP bones, although the BD and BMCv valuesof the OA were 2.8 and 1.3 dB, respectively. Corresponding bones were significantly higher than those ofthe OP predictions of either sorter's echogenicity from bones andthose of thediabetic bones in particular. ThePKA visual readings by our most experienced readerwere and RC values of thenormal bones, nowever were significantly within 4.4 and 2.2 dB, with 95% confidence. smaller than those for the diseased bones, although the BD and The somewhat higher precision and expected higher BMCvvalues for the normals la; between thoseof objectivity of the digital measure suggests theOP and OA. The acousto-ultrasonic, parameters po5sible value in repeating the (generally for normal males and females didnotdizfer while negative) earlier studies of relative lung to liver the BD and BMCv values were significantly different. echogenicity as a possible indicator of fetal lung These results show that there is no simple relation maturity. The digital measure is a better, but not shipbetween the acousto-ultrasonic parametersand essential, method of incorporating relative the BD or BMCv ialues. While the latter represent echogenicity in a multiparameter study of fetal measuresof bone mass, the formermay strongly lung maturity. depend on the structural elements of bone.

r 76 1988 World Congress or. Medical Physics and Biomedical EngineeringSan Antonio, Texas Medical Physics Scientific Papas

MP20.6 MP20.7 Artifacts in Sonography: Physical The Use Of Ultrasoundto Assure Correct Mechanisms. F.W. Kremkau, Center for Plamoent of =an' aye Plaques, B. Japp* Medical Ultrasound, Bowman Gray School C.J. Paviin, H.D. M -Gowan, D.G. Payne, of Medicine, Winston-Salem, NC E.R. Simpson, P.J. Fitzpatrick, Princess Margaret Hospital, Toronto, Ontario, The occurrence and manifestation of approximately Canada. eighteen artifacts of acoustic origin in sonography are reviewed. Physical mechanisms involved are A popular brachytherapy method for treatment of described. They include pulse length, pulse width, choroidalmelanoma uses an applicatorconsisting interference, reflection, refraction, side lobe, of 3-25I seeds in a goldshell. The plaque edge grating lobe, attenuation, focusing, reverberation, results in a very sharp dose gradient so it is resonance, speed error, and high pulse repetition importantto assure thatthe tumor base is within frequency. Unexplained cases are included to theperiphery of the plaque. Ophthalmicultra- stimulate thought and discussion. These artifacts sound is used routinely for imaging of choroidal are manifested on images as added or missing melanoma. Imaging at the time of plaqu, _lacement objects and objects with incorrect brightness, and post-operatively may beused to show plaque location, size, or shape. Some artifacts are position relative tothe tumor and sensitive useful in interpretation and diagnosis (e.g., structures. Scans of the plaque are °brined in shadowint, and enhancement), while some can cause vitroinorder to familiarize the operates_ with confusion and error (e.g., section thickness and the imagesand/oroptimize the materialsof a reverberation). A proper understanding of dummy plaque in order to get a better image of the artifacts and how to deal with them when plaque center or edge. At the time of operation, encountered enables sonographers and sonologists to the plaqueis Secured over the tumourand its use them in diagnosis while avoiding the pitfalls position is monitored with ultrasound or the dummy which they can cause. is positioned and scanned before exchanging it for the loaded plaque.We have usedthis method of assurance with 3 plaqus designs over the past7 years.Goodlocalizations and some misplacements will be demonstrated.

MP20.8 MP20.9 Effects of Ultrasound Irradiation on Axonal Ultrasonic Computed Tomography used Outgrowth in Chick Embryos at 1.1 MHz. in Noninvasive Temperature Measuring JW Yip, MS Rosenthal*, YPL Yip, C Capriotti, Yi-hong Yang & Hong-zhang Wang *, Dept.of Precision Instrument,Shanghai BQ He, University of Pittsburgh, Pittsburgh, Jiao Tong Univ.,Shanghai,P.R.China PA, 15261, USA

Studying the effects of ultrasound irradiation on In the ultrasonic hyperthermia,for the noninvasive temperature measuring and mo- the developing central nervous system is important nitoring a method of ultrasonic computed but difficult because of the immense structural complexity of the vertebrate brain and spinal cord. tomography with a ring transducer array model and a suitable algorithm have been Using the less complex chick sympathetic system as developed. In order to get high resolu- a model allows us to investigate the effectsof tion with shorter calculation time sparse ultrasound irradiation on one aspect of neuronal scan and sparse reconstruction have been development, axonal outgrowth. Two-day-old chick embryos (stage 15) had two open- proposed for measuring the increment dis- ings made in their shells to allow for ultrasound tribution of sound velocity and then the increment temperature distribution. irradiation in ovo. The embryos were irradiated pulse mode (2000 Hz) at 1.1 MHz with a pulse width Since the sound beam can be partly hind- Average power was 1 W/cm2. The em- ered by the bone and air cavity in human of 75 psec. body,resulting in a lose of projection bryos were irradiated for 5 minutes, on three con- data,the method of projection space ite- secutive days starting on the third day of incuba- Following each irradiation, em- ration and the method of truncated pro- tion (stage 17). jection have been used to make a good re- bryos were stages and returned to an incubator (37° construction from a limited data. C). One day post final irradiation, embryos were examined for the degree of symrathetic pregang- The method of ultrasonic diffraction to- lionic axonal outgrowth using the marker horse- mography and the fast convolution-back- radish peroxidase (HRP). propagation algorithm are used for image Results to date have shown no evidence of any ef- reconstruction. Also,a new approach of fect on sympathetic preganglionic axonal outgrowth diffraction reconstruction algorithms with this scheme following expotre of 75 embryos from limited data is proposed. and 75 controls. 13 9

1988 World Consitas on Medical Physics and Biomedical EngineeringSan Antonio, Texas 77 Medical Physics Scientific Papers

MP21.1 MP21.2 Weighted Moving Beam Therapy(WMBT) NONINVASIVE THERMOMETRY: for Fast Neutrons (with special THE POTENTIAL OF MRI IN CLINICAL HYPERTHERMIA,N.A. Detorie*, consideration of the neutron D. Jenkins, F. Tkacik, specific inhomogeneity problem), Richland Memorial Hospital, U.S.C. Medical School,Columbia, K.H.Haver*, B.Hesse,W.J.Lorenz, South Carolina Deutsches Krebsforschungszentrum, D-6900 Heidelberg, FRG Temperature was used asa contrast agent to alter the relaxation times of muscletissue imaged with A significant benefit for fastneutrons a 1.5 Tesla MRI system (Philips Gyroscan).Spin as reported recently is related to high echo images prior to and energy P(65)+Be or - using monoenergetic post heating were T1- weighted: TR=800msec. and TE=30msec.These images 14 MeV d-t neutrons tosophisticated demonstrate an expected andobservable loss of MR treatment techniques. For the fastneu- signal with increasingtemperature. tron therapy facility KARIN at Temperature the Ger- sensitivity can be dramaticallyenhanced via image man Cancer Research Center the WMBT- subtraction of the "prior" from technique and the corresponding "post heating" treat- image. Furthermore, selective point ment planning system have been temperature devloped measurements in tissue "prior" and"post imaging" with the view towards optimizationdose potentially provide a calibration distribution. for quantitatively assessing thetemperature distribution This computer controlled treatmenttech- throughout the entire imaged nique allows to match isodoses volume. Twelve image close to slices (128x128) utilizinga single MR measurement the target volume and helpsto overcome can be acquired in apprlximately some problems followed from the two minutes. The neutron subtraction images provideinformation whicL may specific inhomogeneity problemwhere be applicable not only energy deposition is highly depending to tnermal modeling but on also treatment planning. Our experience to date atomic composition of tissues. regarding heating and imaging The results of CT-based computerized both in vitro and !n patients will be discussed. radiotherapy treatment planningverified by phantom measurements willbe demonstrated and discussed in thepresentation.

MP21.3 MP21.4 Radiation Dose Estimation for Radioactive The Impact of Dose Response Curves on she Isotopes of Lead Labeled with Antibodies, Accuracy in Clinical Dosimetry, P. Kartha* F. Hosainl W.W. Layne and E.M. Billings, L. Lanzl, S. Reddy and F. Hendrickson, Department of Nuclear Medicine, Univ. of Rush-Presbyterian-St. Luke's Medical Center, Connecticut Health Center, CT 06032, USA. 1753 West Congress Parkway, Chicago, IL 60612 It has been possible to label a monoclonal anti- There is enough evidence that the dose body (B72.3) with lead-203 in response our laboratories. curve is rather steep and that the accuracy of Cyclotron produced Pb-203 (Tp of 52 h with 80% of dose is important clinically. The purpose of this 279 KeV photons), and Ra-224 generated Pb-212 (Tp paper is to present dose response curves derived of 1C.6 h with alpha emitting daughter products) from clinical data obtained when treating certain are potentially useful for diagnosis and therapy, types of tumors e.g. head and neck and respectively. cancer of Radiation dose estimates have been the cervix and their implications in the carried out for anticipated clinical application. accuracy of clinical dosimetry. The discrepancies observed Calculations are based on preliminary animal data in physical dosimetry, dose prescription, obtained with the Pb-203 labeled antibody, treat- and ment planning, clinical dosimetry and uaily patient kinetic modeling. Information based on a kinetic set-ups will be presented. model using human plasma clearance of indium-111 These inaccuracies may lead to errors in delivered target absorbeddose labeled F(ab')2 predicted a maximum of 12% uptake of such magnitude that might change tLe of Pb-203 radioactivity in tumors of the size probability of of local control appreciably. Our experience and kidneys (usual size of hepatomas treated) Pb-203, results in treating over 15,000 patients during which reduced to 3.6% in the case of Pb-212 due the period 1971 to 1967 on the megavoltage to its short half-life. units: Further, a rapid uptake Theratron 780, varian Clinac 4 and an AECL Therac of 30% in the liver (clearing with a biological 20, will be discussed.We have used dedicated half-life of 1 day), and 5% uptake in trabecular computers to monitor and verify the set uppara- bone without any biological clearance were taken meters while treating the patients with radiation. for dosimetry. The S-absorbed dose factors were An analysis of our data indicates that derived for 24 x 25 source-target combinations. a good quality assurance program in Radiation Oncology Estimated doses were found to be 2.2, 0.25, 0.03 Physics alone can prevent the errors from mGy/MBq with Pb-203 for tumor, liver and marrow occurring. compared to corresponding values of 23, 9.2, 3.6 for Pb-212 in equilibrium with daughter products. C; 0

78 1988 Work! Congress on Medical Physics andBiomedical EngineeringSan Antonio, Texas Medical Physics Scientific Papers

MP21.5 MP21.6 Virtual Simulationina PhysicalWorld: Some Dosimetry of 1-131 Labeled A6H Antibody,*TK Practical Considerations, George W. Sherouse', J. Johnson, RL Vessella, SE Mittelstadt, ML Daniel Hour land, Kevin L. Reynolds, Harris L. Mathisen, BW Wessels, MK Elson, FM Grund, DJ Leightner, RK Chiou, PH Lange, VA Mc Murry, ThomasP.Mitchellt,JulianG. Medical Center, Minneapolis, MN. Rosenman, Edward L. Chaney, Radiation Oncology, University of NorthCarolina, ChapelHill, NC A6H is a highly specific monoclonal anti- 27514, and tRadiation Oncology, Universityof body targeted against Renal Cell Carcinoma. Florida, Gainesville, FL 32610 In phase I-II clinical trials, 10-20 mgrars of antibody, labeled with 5 mCi of 1-131, Clinical implementation of virtual simulation (the use of CT was administered to patients with primary images and computer software in place of a conventional and/ur metastatic disease to ascertain in- treatment simulator) requires the ability to transfer the vivo pharmacokinetics. A higher dose (50 geometry of the virtual patient and treatment unit to the mCi, 20 mg) was administered if a tumor clinical setting in a way which is both reproducible and effi- site imagedas part of a therapeutic dose cient enough for routine use. We have separated this pro- evaluation series.Dosimetry estimates were made using the MABDOS software of cess into 1) repositioning of the patient with respect to the Johnson. Data were modeled as four treatment couch, 2) localization of an absolute coordinate compartments comprising whole body (blood), origin in the patient/couch system, and 3) absolute position- tumor, liver and spleen. The MABDOS ing "by the numbers" of the treatment couch, gantry, and environment allowed interactive model collimator. The first problem has becn addressed by use of specification, integration of the resulting full- or hemi-body polyurethane foam molds formed within differential equations, calculation of alignment jigs, the second by use of a clinically-locatable cumulative activities, interactive coordinate reference, and the third with the help of three- definition of tumor site(s), Monte Carlo axis coordinate scales.Our early experience with use of simulation of prominent photons for each source compartment, and dosimetry theserepositioning andlocalizationtechniqueswillbe presentation in a tabular format by organ. presented. Estimates of tumor dose ranged from 100 to 1300 rads, and were highly dependent on estimates of tumor mass. Extrapolations indicate that much higher doses (100 200 mCi 1-131) might be tolerated by patients.

MP21.7 MP21.8 Evaluation of the new 50 MeV race-track micro- IMPLEMENTING AN IN-HOUSE QUALITY CONTROL PROG tron with multi-leaf collimator for patient AT HYPERTFIERMIA FACILITIES P. Shrivastaa,* 1 treatmelt, M. Karlsson*, H. Nystrom, J. Stenberg T. Sayl9r , D. Wilki pson , B. Paliwal , D. and H. Svensson, Radiation Physics Department, Beuchler , R. Steeves ,"Hyperthermia Physics University of UmeS, S-901 85 UmeS, Sweden. Center, Allegheny-Singer Research Institute, Pittsburgh, PA 15212-9986, and University of Clinical electron and photon beams for patient Wisconsin, Madison, treatment have been ,:eveloped on a new type of 50 MeV race-track microts-on in UmeS, Sweden. The A strong commitment to quality control measures collimating system on this unit is a new type of is necessary at every hyperthermia facility to high resolution multi-leaf collimator. achieve safe, effective and reproducible Electron beams at energies between 5 and 50 MeV reatments. Without these measures, the can be used and the beam flattening is obtained by variability in heat distributions can be so a computer controlled scanning system. A minimum .arge that the benefit of treatment may become of scattering material is introduced in the beam inconsistent. The recommendations for an and the measured depth dose data will therefore be in-house QA program developed by the very close to what is theoretically achievable. Hyperthermia Physics Center (HPC), under Photcn beams can be extracted at the same energies contract #N01-CM-37512 will be presented in this as for electrons. Also these beams are scanned and paper. These recommendations are Lased on the the use of very sharp flattening filters are there- HPC experience in reviewing a minter of by avoided. This will decrease the neutron conta- hyperthermiafacilities in the USA using a mination at high energies and also make the beam variety of equipment and participating in flattening more stable. The collimating system is national cooperative hyperthermia trials. The a newly designed multi-leaf collimator with 64 recommendations include QA procedures and leaves. To reduce the electron scattering in the criteria for ensuring the proper operation of treatment head the air has been substituted with equipment and guidelines(1) foroptimal.'patient helium gas. It is therefore possible to use the setup, treatment delivery and documentation. same collimating system for both electrons and photons. (1)HyperthermiaQualityAssurance Guidelines, Except for a technical description of this new HPC Report No. 8803 (Radiation Therapy Oncology system, the presentation will include a full set Group, 1101 Market Street, Philadelphia, PA). of beam data, dosimetric data and some experience from the first patient treatments. 1 (/ 1 1988 World Congress on Medical Physics and Biomedical Engineering a San Antonio, Texas 79 Medical Physics Scientific Papers

MP21.9 MP22.1. "Throidal Hormonal Changes During In Vivo MR Spectroscopy of the Brain. And After HeadAnd Neck Irradia- M.A. Smith; D. Porter, V.T. Ayton, M. Lowry, C.J. tion" A . K . Shukla , D . K . Hazra , T . N . Twelves, M.A. Richards, P.B. Garlick and M.N. Edoliya , Y C Asarw al V. K . Sachan, Maisey.Divisions of Radiological Sciences and J.Chandra, J.K. Cancer Institute, Clinical Oncology, United Medical and Dental Schools Kanpur (India). of Guy's and St. Thomas's Hospitals, Guy's Hospital, London. The functional disorders of thyroid gland induced by its incidental inclusion in the radiation field during head and neck irradiation may In vivo human 31P spectra have been obtained using be encountered. The onset of thyroidal hormlnal a 1.5T MR imaging and spectroscopy system (Philips changes may bepredicted to be of measurable Gyroscan):A closely fitting phosphorus head coilis order even during the progression of radiotherapy placed round the subject's head and both are as well as after its completion. Thepresent positioned within the proton imaging head .coil. study included 29 patients undergoing irradiation Images are obtained to identify the volume tobe of head Andneck region. The serum T3, T4 studied.The magnet is shimmed to 0.25ppm on the and TSh levels were measured using radioinnuno- volume of interest using SPARS (spatially resolved assay technique and the results were statistically spectroscopy).The phospnorus spectrum is then analysed. The changesin mean T3 levels were acquired using ISIS (image selective in-vivo observed to be statistically insignificant both spectroscopy).Methods of spectral processing have during and immediately after the course of been investigated to give a reliable estimation of radiotherapy, whereas the alterations noticed peak area ratios. in mean T4levels were mere pronounced and Localised brain spectra have been obtained from 16 sound to be statistically significant at the normal subjects of which 10 have been repeated after end and one month after the course of radiothe- a delay of at least 1month.The mean pH was 7.00 rapy.The TSH levels indicated similar changes with a standard deviation of 0.07; the percentage andit was concluded that the radiation related difference between repeated measurements was 0.6%. lormalitiet may be more common than speculat- Results will be presented on twenty patients with ed in the p,It. brain tumours, in some cases both before and after therapy.

MP22.2 MP22.3 In vitro Dosimetry by the Determinadon of AnAdiabaticHalf-PassagePulsethHuman the ATP-Decay in Devitalized L1210-Cells Forearm31P NMRSpectroscopy. S. A. with the Help of 31P-NMR-Spectroscopy. Altobelli*, M. V. Icenoglef, and E Fukushima, W. Ulmer, St. Marienkrankenhaus, Radiology, Lovelace Medical Foundation, and VA Hospital 5900 Siegen, FRG Ambulatory Care Unitf, Albuquerque,NM,USA Radiobiological dose-effect-relationships are usually determined by an observation of the colony- This study evaluated an amplitude and phase modulated forming ability of clonogenic cell cultures. 31P- pulse'in forearm 31P surfacecoilspectroscopy at1.9 NMR-Fourierspectroscopy provides a tool to deter- T. This pulse performs an adiabatic rapid half passage mine the fraction S of devitalized cells in a more toreducetheadverseeffectsof RF magneticfield direct manner, because it is possible to measure inhomogeneity. Spectra generated using thepulse have the ATP-cortents via the chemical shift after ir- 15% or more S/N thanspectra obtained using a single radiation. Devitalized cells are not able to syn- hardpulsewhich was chosenempiricallytoproduce thesize new ATP, and the remaining ATP must under- maximal signal and which is approximately a a pulse at go hydrolytic decay processes. The experiments thesurface. Increased S/Nwillimprovethetime with L1210-mouse leukemia cells, which have been resolutionof 31P spectroscopy during studies of irradiated by 200 kV X-rays, show that the inte- forearm exercise.The imager/spectrometer(NALORAC gral ATP-concentration represents and adequate Cryogenics)allowsspecificationof shaped RF excita- surrogate for the survival probability, and the tionintermsofatextfiletableof RF relative applicability of the linear-quadratic survival amplitudeandphase. Theactualamplitude,pulse function duration (to),transmit phase and offset frequency (fs) S = exp(-ct D -pD2) can then be specifiedat run time.In these experiments, N-pointRF shape tables of the form:amplitude has been tested by the determination of the ATP- = sin(mw), phase = Ksin(mw), with m = 0,1,...,N-1 were contents via 31P-11MR-spectroscopy and by the con- used. N was either 256 or 512, w was a/(2(N-1)), and ventional method of the observation of the survi- K was constant. The amplitude of thepulse follows val fraction. one-quartercycleofsineandthephasevariation causesafrequency sweepwhichfollowsone-quarter cycle of cosine. The sweep ends at fo+fs and begins atfo+fs+NKw/(360tp), where f0 is the NMR frequency. 1021 MR Benda!! and DTPegg, J. Magn. Res., 67, 376 (1986).

80 1988 World Congress on Medical Physics and Biomedical engineeringSan Antonio, Texas Medical Physics Scientific Papers

MP22.4 MP22.5 Me DIX Pulse Sequence: Depth and Design of Notch Band Magnetic Resonance Inversim withSurface Coils,G.D. PulseSequencesUsingOptimal Control Clarice* and R.B. Rehr, Medical College Techniques, V. Smith*. Department of of Virginia, Richmond, Virginia, 23298 Radiation Oncology, University of California, San Francisco, CA 94143. Inversionof nuclear magnetization is difficult when attempting selective excite- Optimal control theoryhas recently been ticn with surface coils becmase the inhamo- used to design selective pulses, both for geneous magnetic fields produced by these magnetization inversion and fortime probes will nutate the magnetization through reversal. This approach hasbeen a large range of angles. We have developed extended to design pulse sequences and tested amiss sequence which can invert providinguniform excitation over a magnetization ataselected chemical shift specifl.ed frequency range except for a while also allowing a degree of spatial narrow band having essentiallyzero localizatiw for in vivo spPatroseopy. excitation. Such sequences ccal be used, Ashaped, complex sech pulse is used for for example, for water-suppressed proton the irnr=sianofselected peaks within the spectroscopy bycenteringthe notch of entire sensitive volume of the sample.A two zeroexcitation atthewater resonant cycle, 90 msec each pulse will produce near- frequency. The Polak-Ribiere conjugate ly uniform invenrion of a 100 Hz region gradient method wasusedto minimize an inaspectrum. After a variable delay, four objective functionbasedon the product consecutive pulses are cycled through 16 of theactual and desired magnetization phases which producesa"depth pulse ". An vectors. This product was calculated, FID from a volume undergoing a nutation typically, over apulse seque.nce of 128 close to 900 is observed. The Ti's of vari- pointsfor each of 21 frequencies ous metabolites cmn be =seared using the coveringthe range of interest. The inversion recovery experiment. The DISC excitation profilcs produced closely sequence canalso produce reliable values matched the ideal and did not exhibit the forthe flux through the creatine kinase errors seenwhen the inverse Fourier reaction as demonstrated by both in vitro transformsmall-angle approximation is and in viva inversion transfer measurements. used.

MP22.6 MP22.7 HYDRATION OF NITROGENOUS COMPOUNDS A NOVEL PROBE FOR A 7.6 CM WIDE BORE NUCLEAR IN THE URINE MEASURED BY H-NMR MAGNETIC RESONANCE SPECTROMETER D.R. Harrison, SPECTROSCOPY, S.Nakaya*, Y.Tanaka, S.A. O'Connor', D. Reid, P.J. England & P.G. Morris Y.Yoshioka, and N.Yasuda, Dept. of Smith Kline & French Research Ltd, The Frythe, Physiol. Sch. of Med., Iwate Welwyn, Hertfordshire, AL6 9AR U.K. and Medical University, Morioka, Japan. Biochemistry Dept., University ofCambridge)

Critical data on how medicinal compounds affect the Hit,- resolution H-NMR spectroscope (80MHz, 1.9T.) function of whole animal organs can be obtained by has been used to analyze human urine. The peaks means of a wide bore 7.6 cm nuclear magnetic reson- of many components, especially nitrogenous ance (NMR) spectrometer, operating at a frequency compounds such as creatinine, free amino acids, for protons of 360 MHz and phosphorus of 145 MHz, etc. have been identified in tne FT-speccra of which can perform detailed biochemical analyses of normal urine. In the present study, we examined the whole working organ. To maximise the data the interaction between several such compounds; which may be obtained we have designed, developed and water by comparing T1 and T. of water protons and tested, with a variety of organs, a novel re- (NMR relaxation tunes) at several different producible probe for the NMR spectrometer, which concentrations (0.1 - 10 M for urea, 0.01 - 0.8 M allows: the temperature in the rek,ion of the coil for croatinine,1 - 10 mM for uric acid, 0.01 - 3 M to be monitored and kept constant; the temperature for glycine). With the increase of the of the perfusate to be kept at 37C; the organ to concentration, Tl decreased linearly (2 - 3 sec) be stimulated electrically if necessary; the cap- and Tdecreased exponentially (0.1 - 2 sec) with 2 acitors on the RF coil to be fine tuned whilst the all compounds examined. The hydration number probe is within the spectrometer. The probe is (mol H0/mol) calculated based on the T/T ratio 2 1 2 versatile in that different sizes of coil for diff- (Kumosinski and Pess.,n) was 8 for glycine, 4 for erent whole organs may be accommodated up to the urea, and 0.5 for creatinine. For uric acid, maximum size of the probe. the degree of hydration seemed to be smaller in The excellent signal to noise characteristics of acidic than in alkaline solutionswhich sugTested the probe have enabled us thus far to study rapid that the urate crystals might be formed more changes (10-20s) in the concentration of high- easily in aciduria. energy phosphate compounds in perfused rodent hearts induced by established as well as novel cardiotonic agents. 13

1988World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas fil Medical Physics Scientific Papers

HP22.8 MP22.9 PMR SPECTROSCOPIC ANALYSIS OF BLOOD PLASMAFROM LimitsforFastExchange:Practical HEALTHY AND CANCEROUS PATIENTS, D.A. Wilkinson*, Considerationsfor NMR Relaxation H.R. Giles, H.F. Acevedo, P.N. Shrivastava, B. Measurements, L. John Schreiner,McGill Jarmillo, Allegheny-Singer Research Institute, University, Department of RadiationOncology, Pittsburgh, PA 15212 and J. Dadok, A.A. Montreal, Canada. Bothner-By, P. Mishra, Carnegie-Mellon University, Pittsburgh, pA 15213. The apparent nuclear magneticrelaxation rates measured for heterogeneous systems containingprotons in a number of Blood plasma from fasted cancerous patients was different environments, or spingroups, are functions of compared with plasma fran fasted healthy males exchange rates between the spingroups and of the true and females using 620 MHz PMR spectroscopy with inherent relaxation rates and spin densitiesof each group. If various degrees of resolution enhancement. the exchange is 'slow', the apparentrelaxation rates are the Significant variations in the methylene and inherent rates; if exchange is sufficiently'fast', the whole methyl resonances, arising mainly fran the system is observed to relax with a singlerate. The exchange lipoproteins (VLDL, LDL, HDL) were observed rate conditions for the slow and fast regimesare determined along with changes in the levels of with respect to the inherent relaxationrates of the spin groups. -hydroxybutyrate. These changes are Therefore, for example, the apparent protonTi's of tissue are accentuated by using difference spectroscopy, a well explained by the fast exchangetwo-state model, while the technique which may provide more andbetter apparent T1p' s and T2's of tissueare not. information about blood plasma than other The details of the two-site exchangomodel will be reviewed to techniques currently employed (e.g.line-width clarify the behaviour of the apparentrelaxation in the transition measurements). Our data also show that factors from the slow to fast exchangeregimes. It will be shown that such as sex and age are important inassessing throughout the transition the spinrelaxation is characterized by the spectra. two magnetization components. As thefast exchange regime is approached, the magnetizationfraction of the high relaxation rate component becomes vanishingly small.A practical limit for the onset of fast exchange, in termsof the inherent relaxation rates and the spin densities of the spingroups, will be presented. This limit for apparent fast exchangeis dependent on the ability of the spin relaxationmeasurement to resolve the small fast relaxation component. Theerrors associated with the failure to resolve this component willbe discussed.

MP23.1 MP23.2 Measurement of Three DimensionalBone Architecture by Micro-computed Tomography. L. A. Feldkamp, S. A. A QCT Method for Analyzin$ theSpatial Goldstein, M. J. Flynn*, Ford Motor Co., Univ.of Michigan, Distribution of the Bone Mineral within and Henry Ford Hospital, Detroit, Michigan. the Femur/', M.M. Goodsittl*, D.C. Kushner2, S.C. Orphanoudakis3, R.F. Kilcoynel, A system for high-resolution, three-dimensional,x-ray computed M.L. Richardson', S.M. Ott', W.P. Shuman', tomography has been constructed and is 'University of Washington, Seattle,WA, being applied to 2Massachusetts General Hospital, studies of the architecture of in yin'g specimensof trabecular Boston, and cortical bone. The system is comprisedof a microfocus x- MA, 3Yale University, New Haven, CT ray source, rotational and translational stages,x-ray image intensifier, vidicon camera, video digitizer, and Presently, nearly all QCT bone mineralanalysis computer with techniques determine the array processor. Images obtained at a sequence ofrotational mean mineral positions are converted to maps of 2D projection content within the trabecular region. This integrals and mean value, however, may not be processed with a cone-beam reconstructionalgorithm(1). very Typical recoqtruction sets for specimens inthe 1 cm3 range representative of the distribution ofbone consist of 200 density values. Resolution for this mineral content, and it doesnot yield specific sample size is information about where the boneis being rouyhiy isotropic with FWHM of 50 micrometers. Better resorbed and remodeled. resolution can be achieved over a smaller volumeby increasing We have developed the a new method which automatically outlines geometricmagnification. inapplicationstobone the borders of the cortical and architecture, reconstructions are subjectedto morphological trabecular analysis. regions in cross-sectional CT imagesof bones, and determines the bone mineralcontent 1. Feldkamp, Davies, Kress, J. Opt. Soc. AM (BMC) within multiple concentriczones in each A2, 612 (1984) region.The BMC is referenced to that ina set of 6 calibration rods (97- 1460 mg/cc CaCO3) work supported in part by NIH-SIO-RR03348,AR34399, and AR31793. which is scanned separatelyfrom the patient. Histograms of the BMC within the corticaland trabecular regions are generatedas are tables of the BMC within eachzone.The technique has initially been applied to CT imagesof the midshaft of the femur, and will in thefuture be applied to other bones. 'Work supported by NM BRS X04 grant RR05432 82 1988 World Congress on Medical Physics and BiomedicalEngineeringSan Antonio, Texas Medical Physics Scientific Papers

MP23.3 MP23.4 CT Brain Volumetrics for Alzheimer's Disease, Improvements on Locally Optimal Run- Nai-Chuen Yang*, Peter K. Leichner, Godfrey Length Compression on CT Images, J. D. Pearlson, Jose A. Mendizabal, Joseph B. Liu and M.W. Engelhorn, Zhejiang Uni- Bryer, Upma Dhingra, Alan Tien and William G. versity,P.R. China Hawkins, Department of Radiation Oncology and Department of Psychiatry, The Johns Hopkins In locally optimal run-length compres- Hospital, Baltimore, Maryland. sion on CT images differential entropy is selected as a criterion for compres- A new algorithm is developed to calculate the cere- sion. The reported compressed results brospinal fluid (CSF) and brain volumes from CT for 12-bit CT images are 8-9 bits/pixel scans of the brain. Computer software is used to which are far from tne criterion of the- contour the brain automatically and strip off the se images which are from 5.5 to 6 bits/ skull. The CT number distribution for each slice pixel. is generated and saved in a histogram. Beam-har- Analysed result has given one of reason dening effects are corrected in two steps. First, why the compression can not reach the a linear regression analysis is carried out for upper limit is that the strategy is not mean CT numbers versus brain sizes for the brain perfect, option of a optimal compression slices. The mean CT numbers for all slices are format in every run is not always ensur- then shifted to a standard value to take care of ed with that method. The compressed re- the "apical" artifact. "Shifted" histograms are sul.a after the improvements are gene- then generated for all slices and summed together rally closed to the differential entro- to get a global histogram. A global CT number py,about 5.5 bits/pixel for 12-bit CT threshold is then calculated with a two-gaussian images. Thus the results have challenged model and it is accurate enough to separate and to that the differential entropy is se- highlight the CSF's and the brain tissues in the lected as a upper limit for compression. central regions of all the "shifted" images. The "cupping" artifact causes the peripheral CSF's too high in CT numbers and an empirical formula is then used to correct for it.Preliminary studies show that the CSF versus brain ratios range from 12-22% in Alzheimer's Disease and 3-7% in normal controls.

MP23.5 MP23.6 Locally Cptimal Run-Leng'h Compres- Practical applications of the use of sion on CT Images witrout Restriction re-constructed CT images in radiotherapy to Pixel Velue,J.Liu *and M.W.engelhorn planning, C. Sims*, I. H. Manifold and Zhejiang University, P.R. China J. Conway, Weston Park Hospital, Sheffield, S10 2SJ, England. Locally optimal run-length compression Some tumours, notably those of the head and neck onCT images reported by other contri- and the oesophagus, require to be treated in a butor is only applicable to CT images non-transverse plane. They may be displaced whose pixel value is 12-bit. However, horizontally or vertically or both. Conventional pixel va_ues of CT images may be va- planning of such tumours is difficult due to the riable frcm 8-bit to 16-bit. difficulty in interpreting simulator films in such Single byte coding(SBC) has been puted planes. CT gives improvements in the localisa- forward to solve this problem. No mat- tion of the tumour and possibly, more importantly, ter howmany word-length of pixel value critical structures such as the spinal chord and is, the single byte codini. based on the brain atom. We have developed a method of byte has no trouble with it alwys.But reconstructing transverse CT images into the one of most important problem in SBC appropriate oblique plane. Software has been is how a run - length size of noncon?res- written on the Picker Independent Treatment siblepixels would be decided. So coiled "noncompressible pixels" are not certa- Planning System (ITPS) to allow planning to take place utilising central axis and off-axis oblique inly in fact noncompressible, only n CT images Fain from compression of these pixels is too small, so they are regarded as The results of phantom si-udies will be presented noncompressible. In order to search to show the accuracy of the method in addition to out a length of noncompressible pixels detailed examplesor patients with tumours of the a search method and a discriminator ere parotid, middle ear and oesophogus. Finally, the defined. Then run - length of noncompres- problems of verification will be addressed, sible pixels crn be determined reasor- together with a general discussion on the problems ably with the strategy, r.nd experiments and pitfalls of planning in oblique planes. have shown SEC are successful. I 3 5

i988 World Congress on Medical Physics and Biomedical Engineering San Antonio, Texas 83 Medical Physics Scientific Papers

MP23.7 MP23.8 Errors in CBF Determinations by Xenon- Ionization Cher:her Imaging of MegavVtage Enhanced CT. WF Good, D Gur*, H Yonas, X-ray Treatment Portal p, U. Luxton `, 2 JM Herron. University of Pittsburgh, , comez M.A. AstrahanA and T.T. Pharr , Pittsburgh, PA, 15261, USA "University of Southern California School Rf Medicine, Los Angeles CA 90033, and The contribution of errors in arterial xenon con- California Institute of Technology, centration estimations, CT noise and xenon arrival Pasadena CA 91109. time to errors in the determination of regional cerebral blood flow by the xenon-enhanced CT tech- Imaging of megavoltage x-ray beamswas nique was evaluated by computer simulation and investigated using a 1 atm. gas ionization confirmed thTough a series of in vivo and phantom chamber, filled either with air or fluorinated experiments. The results of this study indicate hydrocarbon (freon), used for its higher that approximating the arterial data by a single density. The prototype detector consisted ofa exponential build-up introduces relatively small parallel array of 64 linear anodes separated 1 errors. On the other hand, significant sources of mmcenter to center. Anode to cathode gap was error were found to be caused by CT noise and in- 4 mm. Images were formed by computer accuracies in estimating the xenon starting time reconstruction from multiple projections, using in relationship to sequential scanning times. The the technique of filtered back projection,as overall system's errors in estimating CBF using in computerized _tomography (CT). Reconstruction this method depend on many factors and can be sig- time on an IBM TM AT-compatible computerwas 2 nificant. Using current technology, it is not un- min. A 10 mm diameter cavity of 25 mm depth in common that estimated floe value in a single pixel a 50 mm thick acrylic phantom was imaged with have errors associated with it which exceed 100%. an air filled chamber at 4 MV. With freon fill, the one-dimensional full width half maximum (FWHM) was 2.0 mm for a slit ina 4 MV x-ray beam, almost independent of thickness of lead r.creen used to enhance contr.ast. Chamber resolution for a slit in a 21 MV beam decreased from 2.0 mm FON for no screen to 3.2mm FWHM with a 1.5 mm thick lead screen.

hP23.9 MP23.10 A portal verification system for external Solid-view 3D-imaging for Brain Implants, beam radiotherapy using a two-dimensional K. Ayyangar*, N. Suntharalingam and array of silicon transit dose detectors, Leslie Tupchong, Department of Radiation I.Dchidal Y.Ozai, T.Tomaru, T.Yamashita*f Therapy and Nuclear Medicine, Thomas M.Nakazawa ** and K.Kanota*fCancer Institute, Jefferson University Hospital,Philadelphia, Tokyo,**Cancer Institute Hospital, Tokyo, PA 19107 and K.T. Daniel Yeung, Keh-Shih Japan. Chuang, Medical Image Processing Group, Dept. of Radiology, University o! A portal verification system using a two-dimen- Pennsylvania, Philadelphia, PA 19026 sional array of 289 silicon diodes has been developed to verify patient setup and the correct Improved visualization of brain implants and the delivery of dose to the patient. Using thi3 associateddosedistributions, through either device, instantaneous information about the solid-view modelling or wire frame display has actual treatment beam passing through the patient recently been incorporated intotheplanning can be collected simultaneously with the commence- procedure atTJUH. Thiswas accomplishedby m.nt of irradiation. Tie collected data are auopyass5cLv.LCLls.al...... LC CLV1I5Wile" Wt... stored on a floppy disk for analysis at a later Vax 8300 and Lexidata Image processor hardware. time. By doing quantitative dosimetric analysis, It is now possible to display tumor volumes, size and shape of each treatment field and water- radioactive segments of catheters and isodoses equivalent thickness map within a treatment all as a 3D volume display. The resultant 3D portal are displayed on a CRT and printed out. object is interactively rotatedto visually the comparison of the data obtained during analyze the adequacy of the proposed implant. different treatment sessions, field placement The display time for each solid view is in the errors and changes of the patient thickness can range of 1-2 min per image. Wire frame data can be estimated and also from instant data obtained berotated much faster at20 sec per image. during a treatment, patient movement and motion Currently, the display timeis a limitation. of internal anatomy can be detected. Work is progressing however, to improve interactive time using hardware capabilities of the Lexidata. The clinical usefulness of this new feature will be discussed with 106representative implant cases.

84 1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas Medical Physics Scientific Papers

MP23.11 MP23.12 Computer Modeling of Fast Heavy Ion Radiography Reduction of scatter in portal imaging. S. Shalev*, T. Lee *§, and L. Reinstein, Julius J. Almasi and Cornelius A. Tobias Manitoba Cancer Treatment and Research Lawrence Berkeley Laboratory Foundation, Winnipeg, Canada R3E OV9, University of California, Berkeley, CA 94720 and Stony Brook University Hospital, NY 11794. Heavy ion radiography (HIR) images are usually reconstructed from the residual range distributions of the primary heavy ion beam Portal films and on-line portal verification behind the object. In the past plastic images obtained withhigh energy radiation detector stacks were used to determine the exhibit poor contrast resolution and are notori- location and residual range of the primary ously difficult to use. Image qualitycan be heavy ions. we amend the possibility of improved with theuse of a thick metal plate to using a single plastic detector sheet. The preferentiallyabsorb incident electronswhich location and residual range of the primary havebeen scattered in the patient. We have heavy ions behind the object can be determined examined thescatter to primary ratio for films from the coordinates and the parameters of the exposed under copper plates of different thick- etched tracks on a single detector sheet. A nesses, using radiation from linacs with 6, 15 series of calculations have been carried out to and 25 MV maximum photon energy. A 20 cm thick determine the possible resolution and contrast tissue equivalent phantom simulated patient scatter. It was found of the (HIR) images using different heavy ions that the scattered radiation can be reduced to about 10% of the primary beam with different energies. We have calculated the for small fields and sufficiently thick plates. absorbed dose in the patient as well as other The experimental technique is described, and the benefits and limitations of this new method. implications in terms of improved portal image Results of this calculation will be presented. quality arediscussed. The thickness of front We found that the new technique of analysis is screens in most commercial portal film cassettes faster than analyzing whole stacks of plastic is shown to be inadequate for optimal contrast and it can also give higher density resolution. resolution.

NIH postdoctoral fellow, Present address Siemens Medical Labs., 2404 N. Main Street Walnut Creek, CA 94596

MP23.13 MP23.14 On-line portal verification, Patient Dose Estimation Using S. Shalev*, K. Leszczynsky, T. Lee, Measured Portal Dose Distribution, Manitoba CancerTreatment and Research X. Xing* and J.W. Wong, Washington Foundation, Winnipeg, Canada R3E OV9, University Schocl of Medicine, St. and L. Reinstein, A. Meek, Stony Brook Louis, MO, USA. University Hospital, NY 11794.

Treatment planning calculations are only A digital imaging system has been constructed for nominal representations of the dose deliv- the acquisition of on-lineportal images and the ered due to many treatment uncertainties. verification of beam locations during treatment. Portal dose distribution, or image, in- The detector consists of a metal/phosphor screen clude the treatment uncertainty informa- combination which is viewed by a high sensitivity tion. With accurate dose calculations; CT video camera via an angled mirror. The video digital pro- data that adequately described the pa- images are processed byan IT-151 tient; Software an approximate model of the CT cessing system with a host IBM PC-AT. "error", the measured entrance and portal has been written tc permit rA?I-time frame summa- dose images can be used as constraints to tion and non-uniformity corrections, and contrast adaptively improve the estimation of the enhancement is attained by thresholding, filtration, real-time convolution and histogram modifi- actual dose delivered. The CT data are first used to predict a portal image. The cation techniques. Spatial resolution is better and contrast resolution is greatly ratiosof the measured and predicted than 1 1p/mm, portal doses are used for CT corrections. superior to film. Examplesof on-line portal images are shown for the head and pelvis, using a The corrections can be made in terms of cobalt therapy unit and a 4 MV linac. A "m.vie thickness and/or density and are confined mode" permits the visualizationof patient move- to the region based on the "error" model. ment during treatment. Portal images are com- A recalculated image is then compared with pared with the prescribedportal boundaries from measurement to begin another iteration simulator films, and any significant discreprz- until the images agree to satisfaction. In cies arA! displayed for immediate operator inter- cobalt-GO simulations, excellent correc- vention. tions were made for both density and/or thickness changes. (Supported by NCI grant CA-41574). 1 C 7

1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio,Texas 85 Medical Physics Scientific Papers

MP23.15 MP23.16 Accuracy in matching of portal and simulator The precision of determining compliance images for patient set-up verification. H. Meer- with prescribed fields from conventional tens*, J. Bijhold and M. van Herk, The Nether- portal films, P.B. Dunscombe* and K. Fox, lands Cancer Institute, Amsterdam, The Nether- Manitoba Cancer Foundation, Winnipeg, lands Manitd,a, Canada R3E OV9.

In order to determine discrepancies t,,..tweer in- Increasing emphasis is being placed on ensuring tended fieldedges as documented on the simulator that the treated portal in radiotherapy is identical to that prescribed by the physician. image and actual fieldedges as determined from With a variety of portal imaging techniques either portal images, corresponding anatomical structures in use or under development it Is timely to consider the precision with which Field Placement in both images have to be matched. We developed Errors may be quantified and the factors which a method for this purpose. Alter matching the influence this precision. 370 measurements o- -eld Placement Errors (FPEs) have been made by digital images, field rotations and translations 16 observers on 20 prescription-treatment film can be determined. The effect of the accuracy in pairs taken during routine radiotherapy for Ca prostate. Analysis of the distribution of FPEs the delineation of anatomical landmarks used for measured under several different conditions has matching on the accuracy of the measirement of shed light on the significance, for the precision of the measurement, of the following factors: displacements of fieldedges was studied. Results (a) relative magnification of prescription- showed that precise matching is extremely treatment films; (b) the experience of the observer; (c) the contrast on the treatment important if fieldedge discrepancies less than a film and (d) a double exposure technique. tew millimeters have to be determined, The in-

fluence of the choice of anatomical structureson the measurement will be shown for various tumor sites.

MP23.17 MP23.18 iethodology and Computer Program forTrans- EVALUATION Or A COMMERCIAL MAKMOORAPHY IMAGE position of Image-defined Trajectories into ENHAN-MENT SYSTEM, R.H. Behrman*, P.O. 2rmenhof, Arch-Quadrant Stereotactic Devices- Henry and M.J. Homer, Tufts-New England Medical Ford Hospital, Det., MI Zamorano L,* Dujovny Center, Boston, MA 02111 M, Martinez-Coll A, Mali:, G.

Arch-quadrant stereotactic frames work basedon the principle of aiming a target point fromany entry A commercial mammographic image enhancement of a formed sphere of known radius. Advancesin system which implements a real-time local Image-technology (CT,MRI,PET) intracraniallesions adaptive video processing algorithm basedon can be geometrically characterized with 3D or multi the Wallis equation (1) is evaluated. Measure- planar software and the best trajectoriesto aim ments of the conspicuity of microcalcifications these lesions can selected and plottedon the image, in mammograms made with a mammographic phantom The transposition of these trajectories (angular will be presented. These measurementswere made setting)into stereotactic systems ismandatory in with and without the enhancement system forcon- many stereotactic procedure like multi-trajectory tact as well as magnification mammograms, under serial biopsy and to correlate Image-datawith and over exposed films, and "realistic"mammo- histology, interstitial intracavitary radiotherapy grisuns iron actual breasts superimposed on the and image-guided laser resection. We developeda phantom. Distortion, high and low contrast methodology and computer program (Fortran, Basic)to spatial resolution, and variations of thecon- calculate angular setting of arch-quadrant frames. spicuity across the field of view are also characterized. Frames need to be adapted isocentricallyto the CT MRY and parallel to the plane of scanning.Atra- jectory will be defined by two points coordinates (1) Wallis, R.N., "An approachto space variant that are measured directly from an axial,a refor- restoration and enhancement of images," Proc. matted or scoutview images or the 3D display.The Symp. on Current Math. Problems in Image Science, mathwatical fundaments to calculate Angle A inthe Pavel Postgraduate School, Monterey, CA, Nov. 1976. coronal plane and angle B in the sagittal planewill be represented. Experimental and clinical examples assessing the accuracy of the methodolgy andpro- gram are presented. 1 0 .8

86 1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas Medical Physics Scientific Papers

M323.19 MP23.20 3-D Reconstruction of Left Ven- Noise Properties of Cineradiographic Films, tricle(LV)appliee to Coronary M.Kiguchi, Y.Fujimura, T.Furukawa, Artery Disease(CAD),LU PING and R.Masuya, H.Yuasa, S.Nakagaki, Y.Ishimi MA ;SONG FU,MAILBOX 163,Zhejiang and Y.Kodera*, Hiroshima University, 1-2-3 University,Hangzhou,PRC Kasumi, Minami-ku, Hiroshima, Japan A microcomvuterized image processing sjs- Noise properties of two cineradiographic films tem for 3-Dreconstruction of IN has been (Kodak CFT and Fuji MI-CF) were evaluated by developed and applied clinically for dia- Wiener spectral method. Films were processed by daosis of CAD.Using image saquence analy- two different chemicals with several temperatures sis technique,it will be fast to detect of developer (20-26°C) and different processing automatically endocardial and epicardial speeds (12-20 feet/min). At 24°C and 14 feet/min boundary of echocardiograms and reconstr- with Fuji chemical, relative speed and average uct 3-D LV.LV will be divided into 20 3-D gradient were 100 and 1.67 for CFT film, and 174 regional walls and 3-D images can be dis- and 1.35 for MI-CF film, respectively. Three noise played dynamically in CRT with 4 windows samples for each film which hadthesame gradients which indiviivally and rotatabV display and sensitivities were prepared with different 3-D1;shortening of wall motion,3-1114wall thickening,3-D wall thickness and normal re- processing conditions. lation of coronary artery purfusior to 20 For all processing conditions, both quantum mottle regional walls in pseudo color quantita- and film graininess were higher in MI-CF film than tively.The loGation of coronary artery ob- CFT film, Although MI-CF film showed no evident structions can be predicted 1):; observing differences among Wiener spectral values at any the1.bnoraulity of wall motion, thickening temperatures of developer, CFT film showed and :hickness.The location,sizes and ex- reduction of about 25-40% of Wiener spectral values tent of myocardial infarct and ischemic at the low temperature. can be dctected.A group of patients have been examined with this technique.Contra- stint& with coronary angiography,the result has made good coincidence.

M>23.21 MP23.22 A Critical Look At Video-Based Imaging Three Parameter Equivalent Spectra as a Workstations. V. V. Velocci: W. M. Gentles, W. Ho, Measure of Beam Quality, John M. Boone, R. Reed. Sunnybrook Medical Centre. University of Ph.D., Thomas Jefferson University Toronto, Toronto, Ontario, M4N 3M5, Canada. Hospital, Philadelphia, PA19107 Theapplication of video-based imageprocessing workstations to medical and other imaging Radiological imaging equipment has undergone con- disciplines is becoming commonplace. Most of these siderable technological improvements in recent sye.tPms incorporate commercially available hardwarewith custom software to perform the years, and as a result there is an increased need fundamental functions of imageacquisition and to precisely monitor the quality of x-rays for processing. The power available in todays personal specific examinations. This is especially true computers allows rapid analysis and processing of in low dose, quantitative imaging and dual energy images but the performance of components responsible for displayandespecially image subtraction applications. In diagnostic radio- acquisition is far from optimal (even with high- logy, the half value layer remains the "gold priced components). The inherent image quality of standard" index of beam quality. A technique is the original films can be significantly degraded presented which, based on theoretical and empir- as a result of the non-linearcombination of effects at everystage of quantization and ical models, is capable of accurately describing display. the x-ray energy spectrum using three parameters; Four categories of effects may be the equivalent kilovoltage (3Veq), equivalent demonstrated: 1) variations in transillumination aluminum filtration (Aleq), ad the anode angle. intensity, and video camera response, 2) geometric distortion introduced by the video camera, thy; Aluminum attenuation data are used to determine frame acquisition board ani the final display the values of kVeq and Aleq, while the anode monitor, 3) noise from the camera and acquisition angle is generally known a priori.The concept electronics and, 4) a compromised dynamic range I. of the equivalent spectra is easily understand- the overall system due to signal amplitude mismatching and quantization limits. able, and provides an accurate yet concise way All these factors, when not corrected, of reporting x-ray beam properties.The 3 para- combine to compromise systeN spatial resolution, meter equivalent spectra mudel includes tungsten dynamic range and geometric representation of the target attenuation and characteristic x-ray image. Thisstudy will detail these effects and show that even when reasonable corrective measures production. The equivalent spectrum may be are taken satisfactory image quality is achievable preferred over the HVL for applications where only under some conditions. spectral quality is important. 109

1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas 87 Medical Physics Scientific Papers

MP23.23 MP23.24 Variances in Exposures Due to Patient Dose Monitoring in Diagnostic Energy and Intensity Responses of Radiology using Therapy Verification Contemporary Diagnostic Exposure Film,R.A. Geise* and H.J. Ansel, Meters, Louis K. Wagner*, University of Minnesota Hospital and University of Texas Medical Clinic and V.A. Medical Center, School, Houston, Texas, Frank Minneapolis, MN Cerra, Thomas R. Fewell, Thomas R. Ohlhaber, and Burton Conway, Food and Drug Administration, Rockville. Maryland. The measurement of patient skin entrance exposures during fluoroscopic procedures is compli- Variances in exposure measurements using cated by the frequent movement of the patient a variety of contemporary diagnostic and the use of automatic brightness controls exposure meters may result from syste- for spot filming, cine and fluoroscopy. Small matic errors due to calibration, energy- dosimeters such as TLDs can not be properly dependence and intensity-dependence. placed on the patient beforehand and kVp and The majority of generalpurpose diag- mAs data may be inadequate for determining nostic exposure meters tested agreed to doses due to changing focus-skin distances. within 10% of eachother if exposure Radiotherapy verification film has been found rates were below 5 R per second and to have a response in the useful exposure range beam-qualities were typical for general for GI exams and cardiovascular cine. The de- purpose radiology, excluding mamm- pendence of sensitivity on kVp and variations ography.Variances up to 40% were ob- in backscatter are small (less than 5%). Skin served for general purpose exposure entrance exposures determined using film agreed meters at mammographic beam -qalities. with TLD measurements within a maximum range of In the mammographic range, main. -aphic 21%. Film dosimetry also provides information exposuremeters varied less th-. 3%. about the field size which is needed for deter- For exposure rates comparabletothose mining scatter doses. in barium enema radiography the variances ranges up to 25%.

MP23.25 MP23.26 Effects of X-ray Tube Voltage Waveforms Mean Organ Radiation Doses in Cardiac and Carotid on Image Quality and Patient Dose, ESA,A. Mustafa; and J. Janczek, Faculty of M. Matsumoto*, H. Kubota and H. Kanamori, Medicine, Kuwait university. Osaka Prefectural College of Technology, Osaka 572, Japan Patient skin doses as well as radiation sensitive mean organs doses from cardiac and carotid digital The measurement of photon spectra has been subtraction angiography procedures were measured carried out for various tube-voltage ripple-rates using a randc phantan, TLD calibrated dosimeters, at 70 kV using objects of various thicknesses made and a Digitron 2 system. of aluminium and acrylic resin.Photon spectra In cardiac DSA, a 30 degrees RAO projection was were transformed to energy spectra and exposure performed using the continuous mode at a rate of spectra, and then exposure attenuation curves were 25 frames/second over a period of 10 seconds. In obtained. By using the radiographic sensitivity carotid DSA, the high load pulse mode technique spectra of five screen-film systems and x-ray was used at the rate of 2 frames/sec. to obtain energy spectra, their radiographic sensitivity RAO, LAO, and 'Down Voews. A total of 20 frames/ attenuation relations were introduced.From these View were acquired. The phantan positioning, radiographic sensitivity attenuation and exposure preparation, screening and exposure factors attenuation relations, mAs values and patient doses were made as close as possible to those done on for obtaining the same radiographic density levels patients. were compared for various tube-voltage ripple- rates. In cardiac DSA a mean entrance dose of 121 mGy The blurring sizes in the radiographs due to object was measured. The highest mean organ dose was velocities were discussed. The results show for the lungs with a value of 14.4 mGy, and a q-antitatively that the smaller ripple-rate yields range of 1-60 mGy. In carotid DSA, mean entrance less mAs values, hence there is less blurring in doses of 177 mGy (RAO), 142 (mGy) (LAO) and size due to object motion, as well as being able to 185 mGy (Towns view) were measu e . The highest use a lower patient dose. The advantages of,three- mean organ dose from the three projections was phase units compared with single-phase units were to the thyroid with a value of 21 mGy followed quantitatively demonstrated. by the eye lens with a mean dose of 4.7 mGy. All other organs received mean doses below 1 mGy from either procedure. 110

88 1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas Medical Physics Scientific Papers

MP23.27 MP23. 28 Non-Invasive kVp Measurement on High Frequency A Prsiotype of Multiple-slit X-ray Beam Imaging Generators, M. High*, N.Y. Medical College, System for Clinical Applications. Y. Kume*, K. Doi, Valhalla, NY 10595, A. Olson, SUNY Medical Center, X.-W. Xu, P. S. Prieto, R. A. Schmidt, L. B. Dixon, Brooklyn, NY, and J. Eisenbarth, Beth Israel Med. Dept. of Radiology, The Univ. of Chicago, Chicago, Center, NY, NY IL 60637

Seven different commercial non-invasive kVp Multiple-slit beam imaging technique has been measuring devices have been evaluated on several proposed to obtain high quality radiologic images with high frequency x-ray generators, both fixed and I.1. -TV digital system.This imaging technique can variable frequency types. The kVp monitors were provide high signal-to-noise ratio and low patient dose all calibrated against the same invasive voltage by removal of scatter and veiling glare, if the physical divider on the same three phase generator to parameters involved with this technique are properly assure they worked properly with generators they selected. were designed for. Based on our previous studies, we have designed The accuracy and usefulness of these kVp a prototype of multiple slit beam imaging system for monitors was then determined when used on clinical applications.We employ a linear stepping several of the newly popular high frequency motor for accurate control of driving a multiple-slit generators. There was quite wide disagreement assembly (MSA) and a high frame rate for image acqui- among the non-invasive kVp monitors, implying sition(60frames/sec)inorderto achieve short that the user of a non-invasive kVp meter should exposure time (less than 0.3 sec).We shall also be aware of just how suitable his individual present methods of reducing line artifacts which may device is for use with a particular high-frequency be caused by the inaccuracies in the slit width and slit generator. separation of the MSA, and by the variation in x-ray exposures among different original frames.Prelimi- nary results on clinical cases will be presented.

Supported by USPHS Grant CA24806.

MP23.29 MP23.30 Problems and Solutions in ,Planning the Anthropomorphic Phantoms for X-Ray Imaging System Acceptance Testing of the Radiodiagnostic Evaluation, C. B. Caldwell andM. J. Yaffe, Equipment of a New University Hospital, University of Toronto, Toronto, Canada M4X 1K9 N. Pallikarakis*, Z. Kolitsi, G. Panayiotakis x-ray imaging and B. Proimos, Medical Physics Opt, Univ. Conventional methods of evaluating of Patras, 260 01 Patras, Greece. systems generallyinvolve assessment of the detectability of simpleobjectsagainst uniform backgrounds. It is not clear how well results of The new University Hospital of Patras with a capacity of these tests relate toclinical radiography where radiologists search for specific anatomic details 730 beds has recently been completed. It is the first of We have been three hospitals to be handed over to medical service fully against highly complex backgrounds. developinganthropomorphic phantoms, radiographs furnished and equipped by the contracting enterprise. images, similar to X-ray diagnostic units include radiography, fluoroscopy, of whichpresent realistic those of patients, to the observer. By providing anglography, conventional and computed tomography, more realistic detection tasks, such phantoms may mobile units and peripheral accessories. The acceptance testing procedure is designed to: (a) ensure that the be moreappropriate than conventionalphantoms performance meets thecontract specifications,(b) for assessing and optimizing imaging systems. establish a reference performance level and (c) comprise a pilot guide for the acceptance testing of the two other We describe the development of anthropomorphic using data derived from actualpatient hospitals. phantoms images. Phantoms are designedin layers, which A work plan and protocols have been developed to cover combination of the required tests. The main checks apply to mechanical are fabricated by a numerically-controlled millingand photochemical and electrical safety, beam geometry, beam quality and Methods of designing both high spa- intensity and radiation safety features. Specific tests apply techniques. tial resolutionphantoms and tissue- equivalent to biplane angiography, fluoroscopy,tomography and willbe the peripheral accessories, i.e.grids, cassettesand phantoms fromdual energy radiographs presented. Finally, a study willbe described processors. The above tests will be carried out with the aid abilityto of commercially available as well as 'borne made" whichseeks to assess our phantoms' evaluate imaging systems by correlating equipment. The experience gained during this procedure radiologists' ratings of image quality of patient will be discussed. and phantom radiographs. 11 1 1988 World Congress on Medical Physics and BiomedicalEngineeringSan Antonio, Texas 89 Medical Physics Scientific Papers

MP23.31 MP24.1 Horphorogical Reference system ofHuman Structure using A Treatment Planning Techniquefor 3-D Dynamic Computer Graphics,Naoki Itou2,Tetsuo Okamura2 Stereotactic Radiosurgery, D.Lefkopoulos*, P. ME Research unit Jikei Univ.Schoolof lied.2Nihon KogakuK.K.2 Nouet, M.Rouzaud, J.N.Foulquier, M.Schlienger, 105 Minato-ku Tokyo,Japan A.Laugier. Services de Radiotherapie, Weipital Tenon, 4 rue de la Chine, 75020 Paris, FRANCE.

The goal of stereotactic radiosurgery (SR) We have developed a system of is the morphological database todeliver a large radiationsingle dose with of the human body structure using a computer graphics a high degree of spatial accuracy intoa small technique.We designed this systemas a general purpose target volume (1 to 3 cm), within the brain. referense ..ystem of body structureby reviewing the three We propose a treatment planning techniquefor calculation of 3-D dose distribution with dimensional image of bodies accordingone s purpose.This a15 MV Linac (Saturne 43 CGR MeV). Weuse eight addi- system is able to use withan ordinaly 16 bit personal tional cylindrical collimators (6 to 20mm diame- computer and possible to carry thesystem only by diskets ter). For target localization andpatient immobi- that install the program . lisation during treatment, the TalairachStereo- The system has 6 subprograms. They are sale wholebody, tactic frame is used. Anatomical data (3f)are female wholebody,cranial part,thoracicpart,heart and obtained from stereotactic arteriography,GT-scan andmultiple abdominal part. Users are possibleto rotate or enlarge stereotactic biopsies (ifindica- ted). Calculation algorithms the three dimensional imageon CRT and also possible to are bused on coordinate transformations and use the measuredelemen- eliminate a disused part to clearthe aimed organs. tary beam centralaxis percentage depth doses, Dissection techinique is available to dissect the body dose profiles and TMR data to computethe 3-D surface or the organ to observe the inner structure of the dose distribution. The computing system is a model. microVAX-II computer and an FPS 120Barray pro- This system is supposed to be usefulas a diagnosis sup- cessor. We already have performe,4 56 SR (50 MAV - 6 tumors). port system for various medicalimagings and as a tool for the education of medical students. Optimization is mandatory in SR because ofthe high dose level. We have developped sucha method imposing a 3-D previsional dose matrix. A"best possiole" plan of irradiationcan be found.

MP24.2 MP24.3 Dose Distributions In Dynamic Stereotactic A Radlosurgery, B. Pike', E.B. Podgcrsak, T. Peters, Computer-Controlled Radiotherapy Technique C. Pia, and A. Olivier, McGill University, Departments of for Carcinoma of the Esophagus, C.W. Cheng; Neurosurgery and Radiation Oncology, Montreal, Canada R.L. Siddon, B.E. Bjarngard, JointCenter for Radiation Therapy, 50 Binneystreet, Boston MA 02115, U.S.A. A treatment planning technique for calculation of dose distributions The curvature of the esophagus in dynamic stereotactic radiosurgery with and its close a 10 MV isocentric linear proximity to the spinal cord render accelerator will be presented. The treatment planning for dynamic major difficulties in the treatmentplanning for radiosurgery is a three dimensional problem, since during treatment esophageal cancer.A both the gantry and the couch rotate simultaneously, the dose inhomogeneity of gantry 125% and a cord dose at thetolerance limit from 30° to 330° and the couch from 75° to -75°. Thedose are considered acceptable. calculation accounts for the three dimensional We have studied a nature of the computer controlled radiotherapytechnique problem and is based on the two dimensional Milan and Bentley using irradiation at a numberof fixed gantry algorithm for the calculation of tissue-maximum-ratio distributions. angles. At each gantry angle,a number of The patient surface and anatomical information is obtainedfrom a static fields are applied. Along the gantry family of CT or MR scans, and a stereotactic frame is used fortarget rotational axis, they abut tocover the localization, treatment set-up, and patient immobilization duringthe desired volume, while each fieldis adjusted treatment. The dose calculational algorithm follows thegantry and to conform to the target in therelevant couch rotation In an incremental fashion, and relieson measured transverse plane. This is made possible by stationary beam central axis percentage depth doses anddose independent collimator jaws. With dose rate profiles to calculate the normalized tissue-maximum-ratio control, the technique producesthe same distributions over a matrix of points defined onone of three effect as if tissue compensatorswere used orthogonal planes (transverse, sagittal, or coronet). The algorithm for each gantry angle. Dose distributions has been implemented within a general purpose stereotacticimage are compared with a four-field techniqueand analysis software system, allowing the user to analyzestereotactic conformation rotation without thesubdivision into fields along the rotational CT or MR images, localize target points, as wellas to calculate and axis. The superimpose the dose distrioutions directly ontoany CT or MR proposed technique givesa dose inhomogeneity of 107% throughout the entire image. The dose calculation algorithm will be discussed In detailand volume and a the calculated dose distributions for the dynamic radiosurgery cord dose well below tolerance. The beam-on compared with measured data. time is 1.7 times that of thefour-field technique.

1 1 2 90 1988 WorldCongress onMedical Physicsand Biomedical EngineeringSan Antonio,Texas Medical Physics Scientific Papers

MP24.4 MP24.5 EVALUATION OF CORRECTION FACTOR DISTRIBUTIONS IN TREATMENT PLAN OPTIMIZATION FROM QUANTITATIVE THORACIC TREATNENT PLANNING BEAM'S EYE VIEW VOLUMETRICCONSIDERATIONS, L.C. Katherine Maetand Jake Van Dyk Myrianthopoulos*, G.T.Y. Chen and D.R. Spelbring, Ontario Cancer Institute, Toronto, CANADA M4X 1K9 University of Chicago, Chicago, IL, USA

Dose computational algorithms which account for The Beam's Eye View (BEV) method has many clinical the effects of inhomogeneities are becoming increa- applications. We haveexpanded BEV's potential by singly sophisticated. However, the implementation calculating thevolume of structures intersected of inhomogeneity corrections in routine treatment by b beam which isblocked so as to encompass the planning remains controversial. Certainly in radio- target, viewed froman arbitrary angle. Plots of therapy ofthe thorax, differential transmission such volumes, as afunction of gantry angle, show through lungtissues canleadto errors in both distinct minima and maxima. Inthe case of a tar- target and lung doses. Consequently, this results get in the right lung, occupying approximately 20% ininaccurate radiotherapy, major inconsistencies of itsvolume, thefraction of the Jeff: lung in dose data collectedfrom multicenter clinical volume irradiated was minimum (0) at 300, reached trials, and inaccurate dose-response data. 328 at 105 0, became 5% at 200 0 and 408 at 275 0. To assess the impact oftissue inhomogeneities (AP=1800, left ;900). The minima for the cord were onroutine treatments, theequivalenttissue-air centered at 1350 and 3150and were separated by ratio method was used to derive correction factors. two broad maxima. Treatmentplansare best com- CT-based distributions consisting of inhomogeneity pared dusimetrically usingdose volume histograms correction factors were generatedfor 100 patients (DVH). However, current computational power ren- who were entered into a prospective, clinical study ders the calculation of DVH too time consuming for on lung damage. The correction factors were analys- clinical use. This work introduces the extended ed within the target volume, at the point of target BEV technique, which relies on geometry, as a much comparing treatment plans. Absorbed dose, within the irradiatedlung volume faster alternative to are correlated with the and at the spinal cord.Resultswill be presented Results for singlebeams technique is then extended According to photon energy and treatment technique. corresponding DVH. The keeping track of the number The effects of ignoring inhomogeneities on our dose to multiple fields by pixel. The volumetric response data for lung damage will be discussed. of beamsthrough a given results of the combinations chosen by BEV are compared wiTh the correspondingdosimetric ones, as well as with the results for the actual treatment.

MP24.6 MP24.7 Accuracy of Beam's-Eye View Localization.Robert 3-D Non-Coplanar Treatment Planning, G.J. Kutcherl: Smith*, James Galvin.Hospital of the University of and C. Burman, Memorial Sloan-Kettering Cancer Pennsylvania. 3400 Spruce Street. Philadelphia, PA. Center, New York, NY, USA

Beam's-Eye View (BEV) treatment planning has been routinely In conventional treatment planning the central employed in our clinic over the last two years using a G.E. axis' of the treatment fields are usually parallel 9800 CT scanner and G.E. Target planning system. This to axial planes of the patient. With the study examines the accuracy obtainable from BEV planning availability of the new 3-D dose planning systems when the patients are scanned in the treatment position by it is possible to design, calculate, and display personnel trained as both CT Technologists and Simulation treatment plans with non-coplanar fields. Technologists. In order to test the accuracy of BEV planning, Strictly, these would be plans in which the simulator films were obtained by setting up to the marks placed central axis' of three or more radiation fields as part of the CT procedure and used as reference for the BEV do not form a plane. However, plans in which procedure. Transparent hardcopy overlays of the BEV output the radiation fields are coplanar, but are in a which includes the target volume, treatment field and block plane inclined with respect to an axial patient placement were used to test the agreement of the two plane comprise an important class of treatment procedures. In order to facilitate the comparison, the BEV plans that also can be designed with 3-D dose images were presented as "shadow" images instead of the more planning systems. typical "wire frame" approach. The comparison included an Data will be presented on target and normal tissue examination of the position of the reference marks on the dose distributions for elliptically shaped tumors simulation film and measured differences in the position of that are irradiated with non-coplanar fields. The reference structures drawl on the simulator film and also use of planar compensators to produce homogeneous identified on the BEV image. In general, the two procedures dose distributions will also be discussed. The were found to agree to within 5 mm for a wide range of use of non-coplanar fields to reduce doses to treatment sites. Possible explanation of the differences is the normal critical structures will be illustrated effect of volume averaging of the reference structures when with treatment plans for patients with para-nasal thick CT cuts (0.5 or 1.0 cm) are used. The magnitude of this and pelvic tumors. difference is important if the BEV is to be used to generate shielding blocks for those cases where it is difficult to determine block placement using a simulator radiograph. This situation occurs, for example, when overlying structures hide information on the simulator film. Problems associated with this treatment planning technique will also be discussed. 113

1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas 91 Medical Physics Scientific Papers

MP24.8 MP24. 9 AnInteractive Method for Optimization of AutoContouring of RiskOrgans in Electron and Photon Beam Mixing Through the Use of RadiotherapyTreatment Planning, J. N. a Feasibility Search Algorithm, G. Starkschall'l Roy*, P. Steiger C. and and C. Ling, G.A. Henkelmann2, Dept. of Radiation Physics' Department of Radiation and Dept. of Clinical Oncology, Radiotherapy', Univ. of University of California, SanFrancisco, Texas M.D. Anderson Hospital and Tumor Institute, CA 94143. Houston, TX 77030. It is desireable to automatically locate and contour riskorgans When such as kidney, combining electron andphoton beams for liver and spinal canalson CT images for radiation therapy treatment planning, the provi radiotherapy planning. This is often sional determination of optimal beam weights can difficult, with simple techniques beachieved based on by specifying doseconstraints at threshold criterion because of overlap various points and selecting any combination of in CT numbers withthe surrounding beam weights that yields a dosedistribution tissues. Thus, contouring satisfying these techniques constraints. Thisprocedure require substantial operator differs from the more conventional optimization intervention. Our aim in this workwas to schemes in which dose values are specified and the minimize operator dependence andoptimize optimization algorithm calculates that beam weight the efficiency of the contour tracking distribution that yields a dose distribution that process for these organs. We have adopted comes closest to the optimal distribution. In the the quantitative image evaluation event that there is no combination of beam weights technique (QUIET) which takes into that satisfies the choice of treatment modalities account a prior knowledge and of the organ doseconstraints, this method allows for anatomy and offersamong many options, interactive user modification of treatment contour tracking, contour modalities and dose constraints. smoothing and contourshare analyses. Our contour tracking algorithm allows defining a contour in extremely noisyandlow contrast environments and contour smoothingtechniqueprovidesselective elimination of convex andconcave contour segments. Further detailsabout this technique will be discussed.

MP25.1 MP25.2 MRI-PET CORRELATION USING A COMPUTERIZED Minimizing Brain Motion Contribution in MR REGION-OF-INTEREST ATLAS AC Evans; C Beil, S Mar- Diffusion Imaging rett, CJ Thompson, AM Hakim. McConnell Brain Imaging K K Kwong*, D Chien, R B Buxton, T J Brady, Unit, Montreal Neurological Institute, Montreal, Quebec. B R Rosen. Massachusetts General Hospital, Boston. A procedure is described for combining anatomical information from MRI (or CT) and functional information from PET In magnetic resonarim imaging of diffusion, thediffu- hi a rapid fashion. MRI data is combined witha procedure for sion weighted image can have significant signal changes the definition, storage and recall of anatomically-based regions- when the diffusion encoding gradient is inone direc- of-interset (ROIs). An atlas of standard ROIs, defined fora set tion but not when the gradient is rotated by 90 degrees. of 18 parallel planes spaced at 6mm intervals, providesa starting ROI template for each patient slice. Global adjustments Our interpretation for this is that signal attenuationis to produced only when the applied gradient is along the scale, orientation and position are applied to obtainan initial match. Individual ROIs may then be moved, deletedor re-drawn motion direction, therefore areas of anisotropic brain as needed to refine the match. motion will or will not have signal attenuation, depend- The ability to store the ROI templatesensures reproducibil- ing on the direction of the diffusion encoding gradient. ity of analysis over long periods and introducesa standardization We observe a shift in position of signal attenuationin of analysis technique. In 25 brain structures, themean coefficent the image when the diffusion encoding gradientis ri.,- of variation in glucose utilization measurementsobtained by 5 tated by 90 degrees. neuroanatomically-trained observers was reduced from 8.1% for A subtraction of the two images with diffusion sensi- manual ROI definition to 4.0% using the template approachwith tization in orthogonal directions gives MR.I. a map of 'flow' or `motion' signals. Adding the absolute value of Template analysis for space-occupying lesions such this dif- as tumors ference image to the original image enables the calcula- or infarcts is illustrated with PET data from a stroke studyem- phasizing the facility for rapid, reproducible analysis tion of the diffusion coefficient minimiz;ng brain motion of multi- or flow contribution. functional studies. MRI/PET matching fora structurally intact caudate nucleus having reduced glucose utilization inHunting- ton's disease emphasizes the accuracy of anatomical localization required to quantify small structures. 114 92 1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas Medical Physics Scientific Papers

HP25.3 MP25.4 Comparison of Presaturation and Surface Coil Excitation with Different Selective Excitation for Slab Transmit Pulses for High-Quality M' DefiRition in 3D FLASH MRI, Imaging,R. Bader', J. Gehrig, H.-J. Wood& V.R. Runge, Tufts University and W.J. Lorenz, German Cancer Research and New England Medical Center, Center, D-6900 Heidelberg, West Germany. Boston, MA 02111 High-quality MR imaging is performed by using surface coils as RF antenna systems. However, the in- Selective excitation and presaturation homogeneous RF transmitting field of surface coils have been evaluated as two techniques for reduces the quality of MR images in cases were no defining a slab from a more extensive homogeneous excitation with body or head resona- volume in three-dimensional (3D) FLASH tors is possible. A method leading toareasof magnetic resonance imaging (MRI). nearly homogeneous sensitivity or constant signal Selective excitation was achieved using a intensitiesfor uniform phantom materialsis de- 2.6 millisecond radio-frequency (rf) scribed. Thenewtechnique bases on a surface coil pulse. The presaturation technique used excitation with different transmit pulses and the a 2.6 ms selective rf pulse to excite superposition of the resultant image matrices to an tissue on both sides of a volume of averaged image. The RF pulse intensities are deter- interest. A non-selective rf pulse for mined by computer simulations of the signal yield imaging followed 10 ms later and it shapes versus the axial distance to the center of a excited the slab more uniformly so that flat coil or loop. The signal intensity at a distance typically the slices deviated in of half the coil diameter exceeds that obtained intensity only one-half as much as with with a single surface coilexcitation by a factor selective excitation. The reduced image of 4 and is even 1.5 times that with a homogene- intensity within presaturated regions ous body or head resonator excitation, if the signal minimized the consequences of aliasing. intensities are ntrmalized to the respective maxi- The presaturation technique also allowed mum signal intensities at the center. The evidence a 3 ms shorter TE and avoided incomplete in Image quality improvement will bedemonstrated. rephasing in the slice selection direction.

MP25.5 MP25.6 MagneticResonance Imaging Studies Two-Dimensional MR Flow Imaging. R.R. Price*, D.R. o6 Mid Velocity Pickens, T. Conturo, G. Smith, W. Riddle, C.L. Partain and A.E. James, Jr., Vanderbilt Kenneth A. Kraft, Panos P. Fatouros University, Nashville, Tennessee and Ding-Yu Fei, Medical College of Virginia, Richmond, VA 23298 Constant flowalong , linearmagneticfield gradient produces a shifL in the phase angle of Velocity profiles in flow models simulating a the transverse magnetization indirect proportion variety of normal and pathologic vascular tothevelocitycomponentalong thegradient conditions have been investigated using mag- direction. Several investigatorshave utilized netic resonance imaging (MRI). Validation "balancedgradients" and paired RF pulses to experiments under identical conditions were compensate for phase shifts resulting from flowing carried out using laser Doppler velocimetry spinstoproduce MR angiograms.MR angiograms Fluid having a Ti relaxation time and (LDV). derivedfrom one-dimensional flowcompensation viscosity similar to that of blood was circu- sequences often suffer from artifactual flow voids lated at a constant rate through the imager's when the flow direction deviates from the flow- From MR images of flow pro- 2.35 T magnet. encoding direction. We have developed imaging files, quantitative velocity information, as sequences which in essenceinterleave two imaging well as estimates of wall shear forces can be sequenceswithin a singleacquisition in which Images of flow through axisym- extracted. flow-encoding is achieved forboth orthogonal (x clearly reveal jet metric constrictions and y) directions. Following acquisition, special phenomena, flow separation, and regions of reconstructionsoftware derives the component Likewise, flow profiles through turbulence. velocity images as well asthe direction- symmetric bulbs demonstrate flow reversals in insensitive magnitude image. Measurements have consistent with vortical separation zores, been carried out in flow phantoms, prepared animal The correlation of MRI- secondary motion. models andvolunteers. Flow images illustrating derived velocities over a wide range (5 - 180 reduceddirection-sensitivityachieved by two- cm/s) with identical LDV measurements was dimensional flow encodingascomparedto one- excellent. dimensional encoding sequences are presented. 115

1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas 93 Medical Physics Scientific Papers

MP25.7 111325.8 Optimizing of Exciting and Saturating Selective Pulses for MR Imaging Slice Separation in Simultaneous Multislice MRI by J.B. Weaver, P. K. Spiegel and T.R. DeRoo, U. Klose, Radiologische Klinik, Dartmouth-Hitchcock Medical Center, Hanover, D-74C0 TObingen, FRG New Hampshire, USA. MR imaging requires a excitationwith frequency-selective pulses to achieve an image ofa well defined slice. To Simultaneous multislice MRI acquires images of realize thin slices, theuse of small frequency band widths multiple parallel body sections with the same number of echoes that a conventional sequence is necessary, requiringgreater pulse lengths in the time requires for a single slice. The technique requires domain. As signal intensitydecreases exponentially with time,it the select gradient to be on during the echo to isdesiredto shorten echo time as muchas possible. Therefore the choice of separate the sections in the bandwidth sampled by RF pulse length has to be the system. As many as sixteen sections have been a compromise between accurate sliceprofiles realized by long RF pulses and short echo time. imaged simultaneously. While there is potentialto increase the number of sections imaged per unit time rc: a selective pulse several pulseforms and filter by a factor of sixteen or more, a factor of two to four fur.ctions can be used.To adjust the pulse form to the is more reasonable for routine imaging with the demand of a choosensequence, the profile of the excited present hardware. slice can:le measured by a specialsequence. In a similar The limitation on the number of sections obtain- way the optimal pulses for the saturationof well defined able with the simultaneous multislice technique slices can be determined, whichis desirable for some appli- is that the images become more blurred as the cations iaL MR imaging likezooming free of fold-over effects. parallel sections are brought closer together and For the design of selective pulses for as the field of view is made larger. We identified excitation and an ii'dex of blurring. saturation a computerprogramme has been developed to generate arbitrary pulse forms, choose We will describe the simultaneous multislice the needed frequency technique and two ways of limiting the field of band width and shift the centralfrequency of the RF-pulse. The properties ofa variety of pulse forms has been view. The first method of limiting the field of determined experimently and will view involves limiting the extent of the 180 pulse be demonstrated.For in the read direction. The second method uses example it can be shown, thatthe slice profile aftera excitation with a half sinc pulse presaturation pulses. We will compare the is sufficient for imaging, methods, describe the effects on image quality leading to a shorter time forrefocusing the slice selection gradient and a shorter echo time. and show Images obtained using both methods.

MP25.9 MP26.1 APPLICATIONS OF IMAGE REFORMATTING FOR Response characteristics ofplane-parallel MAGNETIC RESONANCE IMAGING, M.E. Brqmmer, ionization chambers for obliquely MSc*; R.L. Eisner, PhD; W.A. Fajman, MD; incident photon beams M.C. Brunner, MD; D.G. Schwartzberg, MD; Emory University Dept. of Radiology, B.J. Gerbi* and P.M. Atlanta, GA 30322. Khan, University of Minnesota, Dept. of TherapeuticRadiology, Minneapolis, MN 55455. Multi-Planar Reformatting (MPR) is the generation of planar images in any orientation froma three It has previously beenshown that fixed- dimensional (30) data set. We have developed a separation plane-parallel ionization chambers dedicated MPR software package foruse with magnet- overestimate the dose in the buildupregion of ic resonance (MR) images acquired withthree dimen- normally incident high-energyphoton beams [1]. sionally symmetric resolution using 3DFourier Th,2se chambers also tendto overestimate the dose acquisitions. Bi-linear interpolation is used to in the buildupregion of obliquely incident generate single-oblique reformats, and tri-linear photon beams. This overresponse at obliqueinci- interpolation for double oblique cuts. Typical dence is greater than that data set sizes vary from 256*128*12 to 256*256*64 observed for normally incident beams and is a function ofchamber pixels with approximately 1 mm resolution inall design, angle of beam incidence,and energy of directions. the beam. Data will be presented thatillustra- 3D MR knee images required interactive MPR,allow- tes this overresponse of ing multiple oblique views for optimal visudliza- fixed-separation plane parallel chambers versus resultsobtained using tion of meniscus, cruciate ligaments andother both an extrapolation chamber irregularly shaped structures. and thin layers of LiF thermolumineacentpowder. A method to cor- Also, using a 3D acquisition followed by MPR, the rect for this overresponse will also be scan protocol for knee imaging is substantially presented. simplified, compared to conventional multi-slice techniques. MPR of 3D acquisitions of the pituitary 1. D. E. Velkley et al., Med.Phys. 2, 14 gland provided excellent visualization of small (1975). lesions. The extremely large datasets (typically 3-10 MBytes per scan) will make specific demands on hardware architecture and software for future MR image processing stations. 116

94 1988 World Congress on Medical Physics andBiomedical EngineeringSan Antonio, Texas Medical Physics Scientific Papers

MP26.2 MP26.3 THE EFFECT OF THE BUILD UP CAP MATERIAL IN THE IONIZATION VALIDITY OF DOSE MEASUREMENTS AT HIGH ATOMIC NUMB- CHAMBER RESPONSE FOR Co-60 GAMMA RAYS. M.P.O. Rocha, Dep. de ER INTERFACES IN NEGAVOLTAGE PHOTON BEAM; Indra J. Das*, Kenneth R. Kase, and Faiz M. Khan+, Dept. of Protecyio e Seguranya RadiolOgica/LNETI, Lisboa, Portugal, Radiation Oncology, Univ. of Mass. Med. Center, L.N. Rodrigues, E.R. Cecatti, and C.E. de Almeida*, Instituto Worcester, MA 01605, and +Dept. of Therapeutic Rad- de Radioprotey'io e Dosimetria/CNEH, Rio de Janeiro, Brazil. iology, Univ. of Minnesota Hospital, Minneapolis, MN 55455, USA. In order to determine the absorbed dose in a medium through the Measurement of dose or dose perturbation factors calibration factor based on exposure or kerma in air, the knowl- at high Z interfaces are usually performed with a edge of the influence ofthe wall and build up cap materials on thin window parallel plate ion chamber. In the the ionization chamber response is required. The effect of the past, the use of ion chamber readings for the dose attenuation and scattering of 60Co gamma rays in the wall and measurements has been questioned. This presentat- build up cap of the ionization chamber as well as their no-air ion describes the various measurement problems and studies in detial the associated problem of the equivalence were studied for TKO' ionization chamber(Delrin validity of dose calculation. Out of three factors wall). For that, build up caps with different thicknesses were (stopping power ratio, charge collection at the constructed from Delrin, Lucite, graphite, C-552, A-150, and interface and W/e) needed for the determination of aluminum. It was performed three series of 25 measurements each dose, stopping power ratio and charge collection for each cap which were individually corrected for pressure and were critically investigated. Using several types temperature. The experimental set-up was theraically insulated of commercially available parallel plate chambers, extrapolation chamber and TLD powder, the validity therefore variations less than 0.1°C have been observed during of dose measurement was studied for photon beams each serie of measurements. The original chamber cap were meas- in the range of Co-60 gamma rays to 24 MV X-rays. ured through out the experiment and it was observed an agreement It was observed that at high Z interfaces, most better than 0.1%. The response results for a special material parallel plate ion chambers accurate2y measure dose were plotted as a function of the wall and cap total thickness for higher energy beams (). 10 MV), but for lower energies, there is a significant difference between and extrapolated to the thickness corresponding to the mean cen- measured doses with different detectors. It is sug- ter of electron production. The correction factors for a specif- gested that at high Z interfaces and lower energ- ic thickness were analysed in relation to the cap material and ies, dose measurements should be performed with the relative responses compared to the values calculated by the either an extrapolation chamber or with a very thin Spencer-Attix formalism for the cavity theory. window parallel plate chamber.

MP26.4 MP26.5 COMPARATION STUDY OF TWO PLANE DOSIMETRY WITH IN AIR CALIBRATED IONIZATION PARALLEL IONIZATION CHAMBERS. CHAMBERS FOR MEDIUM ENERGY X-RAYS. C. Cinos and M.C. Lizuain4: Hospital J. Seuntjens, H. Thierens*, A. Van der Plaetsen de Bellvitge. Barcelona (Spain). and O. Segaert, Standard Dosimetry Laboratory, Rijksuniversiteit Gent, Ghent, Belgium.

The use of plane parallel ionization chambers is According to current codes of practice the absor- recommended for accuracy dosimetry of the high bed dose to water, Dw, for medium energy X-ray energy electronbeams and they must be used in qualities is determined by in phantom measurements electron beams of energiesbelow Eo = 15 MeV with ionization chambers calibrated in air under (Spanish Protocol of Radiotherapy Dosimetry). reference conditions. The absorbed dose to water

Dw , is found using the equation A study of two plane parallel chambers has been Dw = Mu Nk (171 /p) puku made for obtaining the correction factors. The en w,air polarity effect, the recombination correction and In present work the (Nen/P)w conversion fac- the absorbed dose to air calibration factors have air been measured. The stability of both chambers and tors, averaged over the energy fluence spectra at the Ice'age current have also been determinated. the measuring point inside the phantom, are determined for X-ray qualities specified by kV HVL of the primary beam. Ionization chambers: PTW 0.4 cm3, M 23343 and These spectra are c5lculated NACP - graphite from measured primary spectra by the Monte Carlo method. Electrometer: N.E. 2500/3 Ionex Perturbation factors pu for a 0,6 cm3 graphite cylindrical chamberare determined following the same procedure. Irradiation equipment: Linac Therac 20-Saturno. Quality correction factors ku are deduced from an experimental determination of the energy response. For X-ray qualities with 2 mm Al < HVL < 5 mm Cu the pu and ku corrections are respectively smaller than 2 % an 1 %. The results show also that accurate determination of the air kerma-dose to water conversion factors require beam specification by kVp and HVL. 1 1 7

1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas 95 Medics: Physics Scientific Papers

MF26.6 MP26.7 ArallaktforLACI,IF thermoluminescentdosimeters in The orientation of TLD's- effects on canical neut urn beams with energimrangingtroro, dose assessment. T.M.Kehoe and d(14)+Be to 12(65)±119 D.I.Thuaites*, Western General Hospital, S. Vynckier B. Hociniand A. Wambersie, Edinburgh, Scotland. UCL-Cliniques Universitaires St-Luc, 1200 Bruxelles, Belgium. The main advantage of TLD materials is their capabilityof dose estimates at isolated sites without the necessity of Application of thermoluminescent dosimeters for fast direct connection to cunberscne ancilliary equipment. Their neutron dosimetry isrestricted due to theirlow size permits many TLD's to be postioned in close proximity, neutron sensitivity. In this Apo- we report the giving good spatial resolution. Although sore rrethois have variation ofthe relative sensitivity ltu (neutrons been devised to introduce 11D's into internal cavities, most are placed on the patients skin. compared to 60-Co photons) of Lithium-7 Fluoride At sites deep within the body which are inside the primary (TLD700) as a function of differs a neutronenergies. bean and well away fran the penunbra region, the scattered The neutron irradiations were performed at the radiation is reasonably uniform. Under these conditionswe neutrontherapy facility of Louvain- la -Neuve usinga would not expect the orientation of the ILD to be important. d(20)+Be, a d(50)+Be, a p(34)+Be,a p(45)+3e, a p(65)+Be In areas of non-uniform scatter such as outside tin primary and a p(75)+Be neutron beam, and at the Gent facilty bean and at the surface, orientation is important. using a d(14.5)+Be neutron beam. The orientation effects on LiF chips,discs and rods placed The ku-value increases with energy. For example, on the surface of a phantcm have been investigated for both 10.55 and 5756 was found at the maximum dose depth photon and electron beans. In addition the orientation effects for scatter photon fields outside the primary ben have forthe d(14.5)+Be and the p(65)+Be neutrons been investigated. The results are presented with anphasis respectively. The low ku-valueat low neutron on radiotherapy applications. energies restricts their applicability to the determination of the y-component, as reported in the literature. The increase at high neutron energies allows the measurement of the neutron dose,since the y-component for these neutron beams is small. The influence of total dose (supralinearity), depth in phantom and field size will also be reported.

MP26.8 MP26.9 THE ENERGY RESPONSE OF LiF TLD-100 to Low and High Experience witha computerizedTLD Energy Photons andtoHighEnergy Electron:, for Reader, H. Gfirtneriv,T. Gerster Varying DosimeterThickness. C. Reft, Michael and H. Kett, Staedt. Reese/University of Krankenhaus, ChicagoCenter for Radiation Med. Physik,D-839 Therapy, Chicago, IL, 60637 Passau, FRG.

A number of papers have reported differences We workonTLdosimetrysince many in sensitivity of LiFTLD-100to x-rays and high yearsand reached twoyears ago a re- energy electrons in comparisonto that for Co-62 producibility of betterthan 2%.(single gamma rays in waterandpolystyrene media. standard deviation). Inorder to ameli- Cavit7theoriesand atomicnumber difference;; orate precision furtherand perhaps fa- between the dosimeter andmedia have been used in cilitate the samplehandling we inter- attempts to explain thesevariations. In view of faced last year the difficulties our Harshaw model 2000 of calculatingthe relative TLDanalyzer to aPDP 11/35minicompu- responses of TLD's over a range of energies, it is ter. recommended to calibratethemusing radiation of TLsignal, temperature signaland clock the same energy as that to be measured.This is are acquired during each readoutand not always practical orconvenient. In this work stored on floppy disk.The terminal's the energy response of pre-selected LiP TLD-100 bellisusedto inform the operator chips was measured in water for photon energies of about the readout status.(e.g. announ- 0.010. 0.024, 0.045, 0.25, 4, 6, 10, 18, and 24 MV ces that the nextTLDhas to be put on as well as CO-60, and in polystyrene for electron the planchet) energies of 6, 9, 12, 15, 18 and 22 MeV. The Theevaluationoftheglow curves, results for three different TLD thicknesses (0.25, which is performed afterthe readout, 0.4 and 0.9 mm) were compared with the runs either interactivelyor automati- theoretically calculated energy responsefrom cally. Burlin cavity theory which is a function of cavity Using this computerizedTLDreading we thickness. Adiscussionof the experimental and achieved a reproducibility theoretical results will be presented. than 1%. of better 1 A. McKinley, 1981 Thermoluminescence (Bristol, U.K.: Adam Bigler) 66. 1 1 8 96 1988 World Congress nt Medical Physics and Biomedical Engineering S San Antonio,Texas Medical Physics Scientific Papers

MP26.10 MP26.11 Measurement of Planar Dose Distributions Using A Pit Fall of Film Dosimetry for Combined Modali- Filmin TherapeuticElectron Beams, A.SShiul ties of Treatment Beams, M. Kao* and W.L. Hsu, V.A. Otte and K.R. Hogstrom, Dept. ofRadiation Rush-Presbyterian-St. Luke's Medical Center, Physics, University of Texas M.D. Andetson Chicago, IL, and Tri-Service General. Hospital, Hospital and Tumor Institute, Houston, TX77030. Taipei, Taiwan.

In this paper, two techniques were used to Since the constant sensitiiity of film to higher obtain the isodosedistributions in a plane energies photon beams and the constant mass stop- containing the beam e.xis. The first technique ping power in the medium of the therapeutic energy reconstructs isodose distributions from the range electron beams, the film has been used for central-axis percentage depth dose (CAXXDD) and megavoltage photon beam and electron beam dosime- the off-axis profiles. These profiles were try studies in medical physics.The data obtained determined from the film optical density profiles from a single exposed film is equivalent to hours corrected for filmnon-linearity. The fluence machine time with ionization chamber and phantom factor and scaling factor were applied to the CAX measurements. Film has been used for routine XDD measured in the solid film phantom to convert clinical dosimetry evaluation in the area such as to CAX XDD in water. Thesecond technique the scattering from foreign bodies in a homogene- combines the ion-chamber CAXXDD measured by an ous phantom, the beam penumbra and uniformity ionchamber in waterwith the off-axis profiles studies, and at the junction of more than one beam measuredusing film in the solid phantom to of a patient set-up, etc. This film dosimetry has construct the isodose distributions. Comparison been used for combined modalities of beams and of the resulting isodose distributions from both energies together. In the study we find a pit fall techniqueswith ion-chamber measurements for 7, of dose interpretation of film dosimetry in apply- 12, 15 and 18 MeV electrons will be presented. ing to the junction of electron beam with , .oton The agreement L.:tween theresults of the second beam. Some reasons are due to the film response technique and the ion-chamber measurement is difference to the incident direction of the elec- within 1% from 95% to 10% isodose lines. tron and the photon beams, and the dose calibration to the media with respect to the phantom used. These investigations were supported 4n part by Without proper precaution, this film dosimetry NCI grant CA-06294. leads to more than 7% deviation and even more from the actual dose expected. In some caseS,it underestimates the hot spot att.:II:junction considered.

MP26.12 MP26.13 STUDY OF THE CHARACTERISTICS OF SEMICONDUCTORS iOR Dose estimates with a water calorimeter. IN VIVO DOSIMETRY DURING TOTAL BODY IRRADIATION AT T.M.Kehoe, G.Galloway,J.R.Cheming and LOW DOSE RATE, J. Van Dam* and A. Rijndens, Univer- D.I.Ihmaites *, Western General Hospital, sity Hospital St. Rafadl, Dept. of Radiotherapy. Edinburgh, Scotland. 3000 Leuven (Belgium). A water calorimeter is an absolute instrument for direct The characteristics of Thetados EDE semiconductors measurement of absorbed dose in the medium of clinical have been studied in view of their use for in vivo Interest. Recent and on -going work is establishing the dosimetry during TBI with 18 MV X-rays at low dose technique as a practical method of obtaining absorbed dose. rate. The water calorimeter built at Edinburgh has been used in Six semiconductors have been exposed simultaneously both high and low LET therapy radiation beans. Dose esti- with a Nuclear Enterprises0.6cc thimble ionization mates have been compared to ionisation chamber measurements chamber at depths of 3, 19 and 34.5cm in Perspex, converted to dose for the low LET radiation using the lat- corresponding to dose rates of 0.07, 0.05 and est HPA protocols (1983,1985). The calorimeter overestim- 0.03Gy.min.-1 respectively. The mean ratio of the ates by for 10M0 electrons and "4% for 4 and 9MV semiconductor over the chamber signal increases from photons. Results are presented for a wider range of energies 1.009 to 3.015 at 19 and 34.5cm depth respectively, over a range of dose rates and field sizes and different with reference to the value observed at 3cm depth. water conditions. The influence of dose, dose rate and degradation of The over-response of the water calorimeter has been attri- photon energy with depth will be analysed selective- buted to the heat defect in water i.e. the disparity betw- ly. This study in TBI conditions will be extended een the absorbed energy :.am the radiation field and that to higher dose rates of max. about 4 Gy.min. -l. appearing as heat via the observed temperature rise. An Moreover the influence of temperature increase from independent method of measuring the .cat defect by a comp- room - to skin temperature has been studied. The arative absorption technique is described. dosemeters were applied on the outer surface of a Perspex water tank with heating element. The surface tempe lture was measured by a thermistor. Marked differences were observed between the six dosemeters: when increasing surface temperature from 22°C to 31 °C (average skin temperature) the increase in responses varied between 0.5 and 3.5%. -1-19

1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas 97 MedicalPhysicsScientific Papers

MP26.14 MP26.15 Telescopic Method to Measure Backscattered A METHOD OF INCREASI.jG SENSITIVITY OF Radiation from Collimator to BeamMonitor PIEZOELECTRIC QUARTZ DOSEMETERS, Chamber, Hideo Xubo, University of Rochester A Ghcsh, Royal Armed Forces Hospital, Cancer, Rochester, New York 14642 Riyadh, Saudi Arabia and K V Ettinger* II University of Rome, Italy. Charge collected by the monitor chamberis produced primarily by the primary photon beam. Piezoelectric quartz dosimetry is basedon chan- The controversy is whether or notan additional ges in the resonant frequency of resonators fol- charge is created by the bet! tattered radia- lowing-an irradiation. The frequency changes tion (BSR) from the secondary are collimator jaws. cased by alteration in the elastic A pair of 10cm thick cerrobend blocks constants of with a quartz caused by changes in the charge state of 6.3mm hole in the middle was used to measure impurities, which act as electron traps. The BSR. One block was placed against the secon- radiation response of quartz resonatorswas dary collimator and the other 200cm from the modified by drifting alkali :1,tal icnsinto the source. A Farmer chamber with proper build-up bulk of the resonator, using Doherty method.It was positioned behind the second hole to was also possible to introduce very small measure charge as a function of the collimator amounts of Al by means of drifting from Al-Hg opening. Because the two holes were alignedto amalgam applied to the surface. The resonators see a fixed area of the source and flattening were fabricated from originally pure material,of filter, but not the secondary collimatorjaws, so called 'swept' variety. After the impurities the chamber reading should neither beinflu- were introduced into 0.3 mm thick resonator,the enced by flattening filter scattering nor continuous application of an electric field re- collimator scattering and should reflect the distributed these impurities throughout thema- correct amount of BSR. Results for 18 MV terial. The measured semAtivity to gamma radia- x-rays from a Therac-2G and a Clinac-1800 tion has increased by a factor of 8 - 25, brin- indicate that approtimately 8% andlt, ling the useful range of quartz dosimetry very respectively, of fitad sire dependentoutput is close to the radiotherapy dose range. For mass attributed o BSR when tht, field size was use the doping may be provided at the stage of changed from 2m x 2..mn to 40cmx 40cm. The crystal growing from the hot solution. Therac-20 result is in agreement withthat by Patterson and Schragge (Med. Phys. 8, 312, 1981).

MP26.16 HP26.17 Clinical Experiences of Patient Dosimetry. ESR READ-OUT OF INORGANIC SOLID STATE G. Rikner* and S. Halldn, Department of Hospi- DOSEMETERS, K VEttinger, II Univer- tal Physics, University Hospital S-751 85 Upp- sity of Rome, Italy . sala, Sweden There are several dosimetric calculations which Many of the solid state radiation dosemeters de- have to be performed before the prescribed dose pend in their action on formation of paramagne- can be delivered to the patient in external radiatic species, like free radicals, colour centres, tion therapy. To ensure correct calculations, electron and hole traps. ESR read-out is very direct patient dose measurements have been per- accurate and non-destructive. In pure alkali ha- formed for many years with semiconductor detectors lides colour centres formed on irradiation give connected .o DPD 5 electrometers. rise to an ESR signal, with amplitude depending To achieve field size independent readings build parabolically on the absorbed dose. Zero-dose up caps have covered the detectors.However, this signal depends on the degree of purity andcon- at the expense of perturbation of the radiation centration of point defects. Introduction ofim- field, which is depending on the radiation quality. purities may result in appearance of paramagne- A maximum perturbation of about 5% has been tic centres, even without irradiation, depending noticed. upon the value of ground state spin of incorpo- In a retrospective study which contains more than rated ions. Material known as TLD-100, which is 500 measurements the arreement between prescribed essentially Lir with a few controlled impurities and measured doses hev3 been investigated. It was gives a strong ESR zero-dose signal,owingto the found that 95% of all measurements showed agree- presence of transition metal impurities, but ment within 5% for the SSD and isocentric treat- LiF:Mg, LiF:Ca, LIF:Ba, NaCl:Mg and KC1:Mg, all ments. This figure increased to about 8% when prepared by interdiffusion and coprecipitation wedged fields were used. No reasons for these devi- give only a weak zero-dose signal, whilepreser- ations could be noticed. For larger deviations ving strong response to radiation, almost linear these could, however, be explained by 1) not with dose. Some of the paramagnetic centres in correct dose per monitor unit setting, especially bivalently doped alkali halides have been iden- in complicated irregular fields, where it is diffi- tified and are described in more detail. cult to estimate an equivalent field size, 2) not stable dose monitor of the accelerator and 3) other systematical errors. 120

98 1988World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas Medical Physics Scientific Papers

MP26.18 MP26.19 Environmental Effects of CR-39 Track Postal Dosimetry Comparisons with Detectors Alanine Dosimeters, K.J. Olsen"and

Julius J. Almasi , Siemens Medical Labs. J.W. Hansen2 ,1 Dept. Rad. Phys. 2404 N. Main Street, Walnut Creek,CA 9456 Herlev Hosp., 2730 Herlev, Denmark. 2 Acc. Dept. Risc,c Nat. Lab. DK 4000 CR-39 detectors are more frequently used Roskilde, Denmark. in various fields cf Medical Physics and CR-39s have many Health Physics. When alanine is exposed to radiation stable free advantages over other low cost passive radicals are produced which may be measured by detectors. However, in experiments someone Electron Spin Resonance (ESR). Our dosimeters must be very careful with the application consist of alanine mixed with 5% polyvidone in of these detectors because of the the shape of 2 mm thick cylinders of 4.5 mm dia- environmental effect of CR-39s We are meter. Alanine dosimeters have been exposed to reporting about an environment effect of various beam qualities in the dose range 10-100 CR-39s which can cause orders of magnitudes Gy at radiation therapy centers in USA, Finland, A of errors in dose measurements. Poland and Denmark and measured at Risoe Nat. mathematical model will be presented to Lab. Doses were determined by comparison with describe the geometry of etched tracks on alanine dosimeters exposed to a "Co beam cali- detectors which suffered from this This model can also brated with Fricke Dosimetry. At doses above 10 environmental effect. Gy the standard deviation for dose measurement be used to estimate the errors of a was 1.1%. Dosimetry comparisons better than 2% measurement caused by this effect. We will at the 95% confidence level are possible. The present several different methods to avoid dosimeters are easy to handle, robust and cheap this environmental effect. We will present developed to and may be read repeatedly. Fading after 100 Gy a computer program :IhIch we of "Co gammarays is less than 2% in 4 years. estimate the sensitivity and the resolu- Alanine dosimeters are useful for dosimetry com- tion of CR-39s at a particular measurement. parisons both for photons and electrons and the This program takes into account not only negligible fading make them ideal for documenta- the environmental effects but also the tion of patient doses in radiation therapy. errors introduced by the etchingprocedure and the image analysis system used to count and measure the etched tracks.

MP26.20 MP27.1 System for THE COMMISSIONING OF AUTOMATED SCANNING Large Field K-cdgc Subtraction Coronary Anglography using Synchrotron DOSIMETRY SYSTEMS, C. R. Blackwell, Radlation.I:K-edge Subtraction System in JAPAN, R. A. Dahl, D. E. Mellenberg*, Mayo M.AkIsada*,S.Watanabe,I.Anno,T.Takeda, Clinic and Mayo Foundation, Rochester, M.Kaklhana,K.NIshImura.S.Hasegawa.K.Konls1, Minnesota. K.Hyodo,M.Ando,Univ. of Tsukuba,Saltamt Medical School,Unlv. of Electro-comm.,Unlv.of Kyushu, The complexity of modern commercial scanning Photon FactoryKEK dosimetry systems requires a systematic approach The K-edge subtraction method is suitable for to assuring their proper operation. A protocol the reduction of the mIsregIsttatIon in the for testing these systems was developed to subtracted Image of the Intrar.nous coronary Radiation(SR) assess correct performance. The following areas angiography. Synchrotron 1) Accuracy and reproducibility available at the Photon Factoo+,National were included: Laboratory for High Energy Physlcs.provIdes a of detector positioning; 2) Precision and new X-ray source which Is highly suitable for stability of system detectors and electronics; K-edge subtraction. This is due to Its high 3) Method of data acquisition; 4) Data processing intensIty,Its parallelsm and Its and 5) Data output consistency.This protocol monochromaticity available in a monochromator was applied to two commerciallyavailable systems, system.Since the SR beam size is relatively including the PS/2 Model 60 controlled Wellhofer sma:1 for In-vivo imaging, a special diagnostic Imaging system, were Results of the protocol measurements system:tvo dimensional WP600C. The potential use and value of This protocol represents an constructed. will be presented. energy subtraction were successfully important component of an overall radiotherapy demonstrated. equipment quality assurance program. And also an attempt at enhancing theboundary obtained from IVDSA coronaryanglograohy was performed using "Minimum Difference of Density Method", because the boundary of the°pacified coronary artery was difficult to be differentiated from the overlapping°pacified ventricles.

121

1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio. Texas 99 Medical Physics Scientific Papers

MF27.2 MP27.3 Expanded visualization capacity Evaluationof a Slot in the display of medical images. Scanned X-ray Imaging System incorporatinga TDI CCD, D.W. Holdsworth Paul R. Horan, PhD and R.K. Gerson, Univer- sity of Toronto, Toronto, CANADA, Wake Forest University, The Bowman Cray h4Y. 1K9 School and A. Fenster, John P. Roberts Research of Medicine, Department of Radiology Institute, London, CANADA, NO 5K8. Winston-Salem, NC 27103 Radiographic imagequality Human visualization of medical image is degraded by noise due to scattered photons. Scannedprojection display is severely limited, becauseeye-brain radiography witha 1 to 2 cm wide slot beam of physiology restricts to contrasts of18 , and radiation willremove mostof this unwanted less in non-ideal viewing conditions. The scatter without introducing excessive x-ray tube visual dynamic range, therefore, issomething heat loading. We have developedsuch a slot less than 7-bits of gray-scale. Over the FOV in medical images, scanned system in whichacharge-coupled device however, the originally operating in acquired image-array pixel values he time-delay-integration mode is can easily used to detect the optical span dynamic ranges of 13 to 15 bits. outpu, from an x-ray Thus image intensifier. S6 image lines are the intensity-scale of image information collected is simultaneously to reduce often 50 to 250 times greater than what image acquisition time. the The systemhas been designed to improve human observer canvisualize lesion in asin'le detectability in digital angiography conventional image renditior. and digital Steven Pia -r mammography. and colleagues at UNC showed thatimage-array values may be mapped,in principle, so that We have constructed aprototypeimagingsystem the essential detail information can be which acquires 512 X 256 images inabout 2 s. The well-displayed and visualized in 7-8-bitsof limiting spatialresolutionof the displayable data. The system is methods for that determined primarily scientific demonstration, by the x-ray image adaptive histogram intensifier MTF. equalization Measurements of the CCD dark (AHE), ct.n be vastly improved noise (270 noise equivalent electrons)and full through a recently discovered alternative well capacity(580,000 electrons) indicate adaptive image-intensity mapping, a dynamic camera dynamic range of 2150 when cooled to 5° adaptive contrast-windowing methodologies. C.

MP27.4 MP27.5 Reconst-uction of Dynamic Left Ventricu- QUANTIFICATION AND VISUALIZATION OF 3DCARDIAC lar Cavity Shape from DI-directional IMAGERY.N.F. Ezquerra*,E.V. Garcia, J.W. Projection Images, H.Matsao*, I.Horiba, Peifer, C.D. Cooke, J.L. Klein, J.P. Skelton, and A.Iwata, N.Suzumura. ('.Takao and W.S. Briggs, Georgia Tech and Emory University K.ilayasaki. Nagoya Inst. of Tech. Nagoya.466,Japan Io order to aid physicians in the visualization and determination oF the extent and severity of coronary artery disease,a technique has been 3-dimensional imaging of :nft ventricle developed to unify the physiological information is required to improve diagnostic eva- obtained from thallium-201 (T1 -201) perfusion luation of cardiac wall motion andcar- distribution, and the anatomical diacfunction. We propose an algorithm information obtained from coronary arteriography. The to reconstruct dynamic 3-D shape of left perfusion information is obtained Pmm tomograms ventricle from bi-directional DF(digital of the myocardium after patient infection of fluorographical system) images. It 201 at pe,' exercise, and the informxtion is then utilizescontrast density information quantifies o obtain maximal count profiles that and the conic beamco-ordinate. The are mapped ito a 3D ellipsoid that approximates evaluation ofthe algorithm hasbeen the myocr. um. The patient-specific arterial performed by the experiments using tree, which s superimposed on this ellipsoid, is plasticphantom modeled on a real left obtained f i a 3D geometric reconstruction ventricle of a hog. Fairly good recon- performed E simultaneous biplane aniographic structing results have been achieved. images. ThE (constructed arterial structure is Wehave applied the method to recon- approximat, I by successive segments of structa human left ventricular cavity generalized elliptical cone frusta, and along shape using sequential images of left with the color-coded myocardium are modeled using ventricle taken from RA030 and LA060 by asoftware package that we developed. conventional The OF system. Theresulting methodology has the capability of displaying recontructed images arc displayed on TV- models that can be viewed and illuminated from monitor as a movie ofheart-beating any orientation in space. This methodology has iAlage, which can clearly express local been tested with a phantom and is currently being wall motion disorder. tested with patient studies.

4 c% 4. Z, 100 1988 World Congress en Medical Physics and BiomedicalEngineeringSan Antonio, Texas Medical Physics Scientific Papers

MP27.6 MP27.7 A New Approach .to Light Imaging oj .the Bteast Evaluation of Radiographic film MTF bymeans of a Linear array of Pholadiades Abund O. Wistl Panos P. Fatouros and P.R.S. G.E. Kishore, Medical College of Virginia, A. Castellano+, G. Borasi °, E.M. Staderini Richmond, Virginia 23298 Gigante# + Dip. Fisica Universita' di Lecce, ° Ospedale Civile di Reggio E., # Centro per New, non-invasive imaging techniques utilizing ingegneria Biomedica dell'Universita' di Roma "La non-ionizing radiation for detecting malignan- Sapienza", Corso Vittorio E. 244, 00186 Roma. ITALY. cies of the breast are continually being sought of as an alternative to mammography. We have The method to evaluate the performances systems is actually that to designed, constructed and currently are testing radiographic determine the MTFof a film, or film-screen a light imaging system which includes apencil combinations. To do this is necessary to measure light beam, a scanning mechanism, provision for the LSF using tet samples and a significant scatter reduction, a variable wave- microdensitometer (m.d ),the procedure is quite length light source, and a low-noise sensitive long, because of the reading of the film pattern with the m.d.. For many radiographic detector. This system configuration was used examinations the spatial frequency upper limit is to quantifyscatter reduction and produce below that of the m.d. The availability of linear images of breast phantoms having low, medium arrayof photodiodes and CCD has opened an and high scattering properties. The use of a attractive way to perform densitometric of these tightly collimated beam improves dramatically measurements.The spatial resolution dIstance of twoneighbor of detectors, i.e. the contrast and resolution at the expense sensitive elements, are in the range 7-25 microns. detected signal strength. Similar measurements Therefore, the Hyquist frequetw of these with aconventional wide-angle configuration detectors are in the range 70-20 Lp/mm, without These illustrate a severe loss of imagequality any optical magnification of the pattern Work in performances are lower then that attainable with a especially in high-scatter phantoms. but areenough to measure1.he LSF of a and use of m d., progress includes the development radiographic film, used for typical clinical more realistic breast-like phantoms, more examinations. Ve have tested this using a linear powerful light sources, and faster and more array of photodiodes (Reticon) to measurethe MTF comparison with a sensitive detection systems. of radiographic films, a traditional procedure has been done. The results obtained by means ofthe prototype system are

satisfactory . Nov we will develop a dedicated densitometric system, to be used in the evaluation of MFT.

MP27.8 MP27.9 Information Spectra of Radiographs Including Specifications and Acceptance Testing for a E2fects of Scatter and Object Velocity, of Diagnostic X-ray Equipment Developing Country, M.E. Tecson*, I. Talusan H. Kanamori*, M. Matsumoto and H. Kubota, Kyoto Institute of Technology, Kyoto 606, and E. Estoesta, Radiation Health Service, Department of Health, Manila 1003, Japan Philippines. The information spectrum is offered as an The Radiation Health Service of the Department overall expression of image quality and system of Health is the regulatory agency for the con- performance. Its formula is trol of electronically producedradiation in Oftentimes it I(v) = log2 + p(v)/n(v) bits the Republic of the Philippines. is consulted for purchase specifications and where v, p(v) and n(v) are spatial frequency, acceptance testing of diagnostic x-ray equip- signal Wiener spectrum and noise Wiener spectrum, ment. respectively. The information spectrum combines The report discusses the difficulties encoun- synthetically the contrast, the latitude, the tered in preparing purchase specifications and sharpness by p(v) and the granularity by n(v), and in doing acceptance testing for newly installed is expressed as a function of spatial frequency. 100-mA, 125 kVp x-ray units for primary care The information spectrum can be readily calculated hospitals. Several factors normally taken for by using the dynamic range, and the MTF and noise granted for developed countriesareof major Wiener spectrum at medium dynamic range. The significance in the Philippines. Examples are detcctability of a low-contrast object and the problems in the availability, quality and sta- resolution of an object can be explained by infor- bility of power supply. Once equipment has mation spectral values in low- and high-frequency been delivered and installed, a medical physics ranges respectively. team fromour office has to travel to remote The contrast degradation effect of scatter, places to do the acceptance testing. Factors the blur due to object velocity, and the detecta- like difficulties in travel, security risks bility-enhancement effect of poor MTF are demon- and thelack of suitable instruments for rug- strated by the information spectra. ged travel have to be taken into account. Oth- er administrative factors must also be considered. 123

1988 World Congress on Medical Physics and BiomedicalEngineeringSan Antonio, Texas 101 Medical Physics Scientific Papers

MP28.1 MP28.2 A Review of Futuristic Detector Design Concepts Correction for the Effects of Spatial for Achieving Ultra High Resolution in PET Cameras. Resolution in Emission Tomography, Wai-Hoi Wong, University of Texas Graduate School R.W. Rowe, University of Texas of Biomedical Sciences, Houston, Texas, USA. Health Science Center, Houston, TX 77030. The resolution of PET cameras has been improved steadily in the last 10 years. Current commercial The quantitative accuracy of emission tomographic cameras have 'intrinsic' detection resolutions in images (PET and SPECT) is limited by the spatial the 5-6mm range which is still two times worse resolution of current camera systems. Activity than the theoretical limit.Current efforts are and size measurements in small objects are in improving the intrinsic resolution to the 2-3mm subject tosevere errors because part ofthe range and simultaneously decreasing the expense of activity at any point is detected at neighboring using thousands of photomultipliers with their locations. This so-called spillover causes affiliated gamma-ray counting electronics. blurring of edges, cross-contamination of Decreasing the hardware complexity is desirable adjacent activity spaces and inabilityto recover from the inventory, production, reliability and true activity (so-called partial volume loss). service considerations. Interesting concepts with Based on computer simulations,we have developed various degrees of compromises between resolution a pixel by pixel image correction procedure for and complexity have been proposed recently. These cardiac PET which removes spilloverfrom the compromises also have significant impactson the myocardium into the ventricular bloodpool, usable dynamic ranges in the count-rates of the liver, stomach and lungs, and removes spillover fromthe cameras. Furthermore, some designs focus at high activity myocardium into the bloodpooland improving the central resolution while others low activity regions of the myocardium. As a stress resolutions in a larger area. Esoteric result, normal myocardial activity isrestored to scintillation light sensors from silicon avalanche within a few percent of its true value photodiodes to specially designed multi-anode and the accuracy of defect activity measurement is (or cross-wire-read-out) proximity imaging photo- improved. Size restoration also results. multipliers have been employed. The Various interest- same principles can be applied to other imaging ing 1 and 2 dimensional analog decoding schemes applications in both PET and SPECT and promise using common photomultipliers have also been to enhance the clinical utility proposed. of emission This pap.r reviews these interesting tomography. design concepts and compromises.

MP28.3 MP28.4 Evaluationof aMaskedOrbiting Effects of Axial Resolution and Sampling Transmission Source (MOT3) for PET on Image Contrast for Multiplanar Emission Attenuation Correction. Tomographic Systems, E.M. Kluksdahl* and K.J. Kearfott, Electrical and Computer N. T. Ranger' and C. J. Thompson Engineering Department, Arizona State McConnell Brain Imaging Centre, University, Tempe, AZ85287-5706. Montreal Neurological Institute: and Medical Physics Unit, McGill University Because of partial volume effects, axialspatial resolution may greatly affect object visibilityand Transmissionscans are routinely performed to data integrity for multiplanar emission tomographic correct for gamma ray attenuation in positron imaging systems. Simulations have determined the emission tomography (PET). The transmission scan relative importance of axial resolution and must be done before administration of the positron sampling on object contrast. Gaussiancurves labeled compound.A new techinque is proposed and represent the point spread function of a evaluated in phantoms and human subjects which multiplanaremission tomographic system. Ring-to- allows the transmission scan to be done after, and ringseparationand axial full-width-at-half- in some cases simultaneouslywiththe emission maximum (FWHM) are varied to determine thecontrast scan. The transmission source, which orbits around of objects at differing axial positions. the inside of tpg patient aperture, consists of a Whenthe FWHM is increased from 6 mm to 12mm, the 3 mm disk of Ge enclosed in a lead collimator. wider FWHM has a 50% decrease in contrast compared Only events fromdetector pairs which are to thenarrower FWHM. For a 3 mm object centered collinear with this source are selected for the in a detector ring, 25% of the truecontrast is transmission scan, significantly improving obtained for a FWHM of 6 mm, in comparison to 13% contrast.Most remaining coincident events can be withaFWHM of 12 mm. When the object is used to reconstruct an emission scan with reduced positioned at the axial midpoint of the detector efficiency. The lead surrounding thesource ring and cross plane for a ringseparation of 13 prevents contamination of the transmission scan by mm, 21% of the true contrast is obtained fora FWHM the emission scan. An attenuatiopocoefficient of of 6 mm compared to 12% witha FWHM of 12 mm. The .096 cm. for both water and ,1 Ge solution with decreased object contrast of widerFWHM point specific activity of over 5 kBq/cc, demonstrated spreads, in the absence of specialized 3D excellent rejection of scattered radiation and reconstructive algorithms, could thus produce data radiation from the object being scanned. losses or misleading artifacts more serious than ose due to small inter-ring gaps.

102 1988 World Congress on Medical Physics andBiomedical EngineeringSan Antonio, Texas Medical Physics Scientific Papers

MP28.5 MP28.6 Relative Performance of Simulated Ann- Ectomography using a Rotating ealing and Traditional Iterative Method Slant Hole Collimator, for a Miniaturized SPECT, S.E. Hill*, S. Daleand P. Edholm, K.J. Kearfott, Arizona State University, Karolinska Institute, S-10401 Tempe, Arizona. Stockholm, LinkOping University, S-58185 Linkoping, Sweden. An image simulation and reconstruction system for a novel miniaturized :aulti-detector SPECT design has been used to compare the relative Ectomography is an alternative tomogra- performance,of simulated annealing and a trad- phic method in gamma camera imaging. itional iterative reconstruction technique. A With a rotating slant hole collimator 64 variety of simulated phantoms were used with a projections are registered and the de- range of photon counts between 50,000 and 700,000. sired sections are reconstructed by a The SPECT design ',ad 128 detectors arranged in two-dimensional filtered back projection groups of 16 surrounding each of 8 pinholes. technique. When imaging sections that The 5 cm. radius sample was divided into 1961 are parallel and close to the body sur- 2 mm. square reconstruction bins. Results show face ectomography is advantageous to comparable reconstruction quality between the SPECT: the distance from the camera to two methods for all noise levels. Acceptable the patient will be shorter and constant reconstructions were obtained for about 10C,000 during the whole registration. This re- counts and up on the basis of a picture distance sults in a high and uniform resolution measure of 0.55 using a normalized root mean and a 1..z and uniform attenuation. So squared distance measure. The simulated ann- far we have studied 1) coronar 99m-Tc ealing method is also shown to be consistent sections of healthy and fractured hip in reconstruction quality and time, although joints for prediction of the healing its reconstruction time is yet excessive for course, 2) short axis 201-T1 sections of a microcomputer. the heart for detection of coronary heart disease or infarction and 3) coronar 99m-Tc sections of the thyreoidea. The hip joint and the heart were imaged with both SPECT and ectomography.

MP28.7 MP28.b The Hartley Transform: A High Speed, Real Left Ventricular Mass Estimation from SPECT TI-201 Uomatn courieriranstorm substitute, M. Myocardial Perfusion Short Axis Slices Using Bulls- btecknerI U. Urost, I. Prato, St. Joseph's MJ Tames; Health Centre of London and University of eye Analysis and Edge Detection in Dogs. Western Ontario, London, Canada. RL Eisner, JC Cedarholm, AS Boyers, D Dunn, SE Mar- tin, R Greene, MT Jones, A Gbber, RE Patterson. Car- The Hartley Transform (HT), a realdomain Fourier Transform (FT), was applied to Nuclear lyle Fraser Heart Ctr/Crawford Long Hosp of EMory Medicine image processing. The HT is 40% faster University, Dpts of Medicine (Cardiology) & Radio- and requires 50% less memory than an FT. logy-EUSM; Georgia Inst of Technology, Atlanta, GA The discrete representation of the HT is:

N-1 N-1 The bullseye protocol (BP) synthesizes SPECT T1-201 NH(f)=Eh(t)cas(2uft/N) h(t)=EH(f)cas(2uft/N) myocardial perfusion three dimensional (3D) data in- f=0 t=0 to a 2D display. When this data is compared against where cas(2uft)=cos(2uft)+sin(2uft) and N is the a normal (nl) reference file it accurately detects number of data points. The HT is equivalent to coronary artery disease. We have used BP and an the FT if and only if the ,input data is real. .The Fast FT algorithm was modified for the edge detection algorithm (EDA) to reasure total left HT'. Since no complex operations are required, ventricular (LV) mass and the %LV that is underper- computation time was reduced by 40% and memory fused. Twenty anesthetized dogs were injected with requirements were nalved. Clinical and bar phantom isoproterenol as a stress agent and imaged: 9 were Nuclear Medicine images were filtered (Henning filter) by both FT and HT ,and the output images nl references, 7 others were also nl, and 4 had par- compared. Total counts (3x100) differed by less than tial coronary stenosis. The %LV supplied by the .001% and no one pixel (mean 500 counts) differed by more than one count. A difference map did not stenosed artery was measured by post-mortem stains. suggest any systematic pattern. The HT can replace SPECT images were reconstructed, and BP was per- the FT in all signal processing applications formed on the stress images.The EDA extracts endo- where the input data is real, e.g. SPECT imaging cardial and epicardaal boundary points on each short where filtered back projection image reconstruction requires extensive filtering. axis slice through a radial search for the zero crossing of the 2nd derivative of the count intensi- * Cooley JW, Tukey JW: An allorithmforthemachine ty profile. Ischemic areas on each slice were de- calculation of complex Fourier series. Math Computation APR/65, v19, p297-30I termined by BP. Estimated LV mass vs. measured LV mass correlated reliably (r=0.88). Preliminary Bracewell RN: The Hartley Transform. data suggest that BP can predict the $LV that is New York: Oxford, 19db. ischemic during stress (r=0.90, SEE=3%, n=4). 125 1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas 103 Medical Physics Scientific Papers

MP28.9 MP29.1 A ProbabilisticAnalysis of Thallium Myocardial A Modified Anger Camera System for SPET Images 3 Coherent Scatter Tomography ICJamzad, A.Uchiyama, H.Ttlyama, and H.Murata *&1 Dept.ofElec.&Clannl. Sch.of Sci.&Eng. Waseda U. J. Kosanetzky; G. Harding Tbkyo. 2) Inst.of Clin. Med. T8UkUba U. Tsukuba. Philips GmbH Forschungslaboratorium Hamburg, 3) Div.of Nucl.Med. Tbranombn Hospital, Tbkyo. Vogt-Koelln-Str. 30, D-2000 Hamburg 54, FRG

A probabilistic method is developed for analyzing the extend and severity of isehemia bycorparing When x-rays impinge on an object, part of the the early and delayed SPELT images of Thallium quanta will be scattered elastically into a small after the phisical exercise. For both, early and cone of about 10 deg.aperture. Due to inter- delayed stages, 10 short axial images(slices) are ference effects the spatial distribution of the extracted from apex to base. For each image the scattered radiation fora certain material is center of left ventricular cavity, the inner and characterized by an individual scatter-pattern. outer edges of myocardium are autematicaly deted- One can use these pattern to differentiate be- ted. Each slice is divided into 8 equi-angle sec- tween materials. Use of a conventional x-ray tube tors. Also the heart is divided into 8 equi-angle allows a CT-like imaging of a slice through an vertical sectors from apex to base. For each extended object within a short measuring time. sliceof earlyand delayed,adefect probability Image contrast reduced by the polychromatic na- matrix P is created such that0 < P(i,j).41 ture of the incident radiation can be enhanced by shaves the severity of ischemia of pixel(i,j) of use of a modified 'Anger Camera' system as a de- myocardium of that slice. tection unit. A short overview of the physical Using these matrices, the defect probability for background will be presented together with our each slice, for each sector of each slice and for new detection system. Results will be given which each vertical sector are calculated and display- demonstrate the improvements in image contrast ed on CRP by means of images, curves and histogr- and the feasibility of our method for material ams. They show what percentage of the area of characterization. each slice and that of lateral, anterior, septum andinferiorhas how much defect. These curves, histogramsand imagesgive acomprehensive and easilyundrestoodinformation about thecondition of the heart in early and delayed stages.

MP2c.2 MP29.3 A Method for Multi- element Synchrotron Radiation Computed Tomography Imaging in X-ray CT. with large area monochromated beams, J. Fryaeand K.J. McCarthy F.Toyofuku*,K.Konishi,K.Nishimura,K.Hyodo, National Institute for Higher M.Ando,M.Akisada,A.Maruhashi,A.Suwa, Education, Dublin, Ireland S.Hasegawa,and E.Takenaka,Kyushu University, 3-1-1,Higashi-ku,Fukuoka 812,Japan. When analytes are already present within an object (matrix) they can be imaged in a single X- ray CT scan by passing a continuum of X-rays Synchrotron radiation has a lot of advantages as a through the object and measuring transmitted source of X-rays for computed tomography. Several fluxes within narrow energy bands on either systems have already been developed using a slit side of the absorption edges of the analytes. beam geometry, aiming to achieve the highest spa- It is shown that by extrapolation of results to tial resolution of 10 pm or less. We have develop- the absorption edges, matrix effects are ellim- ed a synchrotron radiation CT system with an area inated. In the process, dead time effects are beam geometry using an image intensifier and a minimised. The projection data is analyte equi- high S/N TV camera as a two-dimensional detector. valent thickness along the beam direction. A Large area Jochromatic X-rays were obtained from reconstructed image shows analyte concentration. a wiggler beam using a silicon crystal with (311) A small apparatus was built to test the theory. asymmetric reflection. A rat phantom injected with In previous work the analytes were discretely iodine on a rotating table was projected onto the located within the matrix. In the present exp- image intensifier. A total of 100 projection video erimental results the analytes are intermixed images covering 180 degrees were acquired in a si- within matrix. When reconstructing an im- ngle scan. Axial sections were obtained by using age for any one analyte, the other analytes the filtered backprojection method. By taking two are treated as part of the matrix. The recons- images above and below an iodine K-absorption edge tructed images are compared with images produ- (33.17keV),it was possible to obtain an iodine im- ced by raster scanning a thin slice of the ob- age. Three dimensional images were also obtained ject taken from the CT image plane. In a water from all of the reconstructed cross sections. like matrix a few centimeters in diameter the It was suggested that this system has the possibi- sensitivity is of the order of 10-3-10-5g/cc lity of rapid three-dimensional volume imaging of for 107 incident photons. a specific element an a human body. I 2:

104 1988 World Congress oa Medical Physics and Biomedical EngineeringSan Antonio, Texas Medical Physics Scientific Papers

MP29.4 MP29.5 Potential medical applications of differential An Isotope Computed Tomography Scanner for the tomography, R. Cesareo; Centro per l'Ingegneria Precise Measurement of Bone,T. N. Hangartner, Biomedica,Universitil di Roma, Italy, S.Q.G. Ma- WrightState University, Dayton,Ohio, U.S.A. htaboally°, Dept. of Physics, University of Raj- shahi, Bangladesh, L.J. Pedraza°, Centro de Es- A special computed tomography (CT) scanner has been tudios Nucleares, Universidad N.A. de Mexico. developed for the precise measurement of bone. The scanner, which employs an 125i source and counting Attenuation coefficient of each element exhibits sharp electronics,is a translaterotate system and allows the discontinuities (K,L....) in correspondence with the mini- measurement ofcrosssections up to 200 mmdiameter. mum energy values for the ejection of an electron from The machine is optimized for the quantitative evaluation its different shells. These discontinuities can be employed of trabecular bone at thq distal end of arms and legs. for detecting and isolating the presence of an element in These measurementsitesare particularlyimportant any matrix when a pair of monoenergetic radiations are because fractures of these bones represent good indicators used which closely bracket the energy of discontinuity of forosteoporosis. the element. The difference between two tomographies Theprecisionof the measurement of trabecular bone den carried out with a selected pair of radiations is only sensi- sityisabout 0.5%. This is an improvement of close to an tive to the presence of a given element. Ity employing an order of magnitude compared with commercial scanners in X-ray tube with external interchangeable secondary tar- average clinical settings.Responsible for this high pre gets, quasi monoenergetic X -rays can be produced. A mi- cision are the mechanical design of the scanner; the choice ni CAT-scanner has been employed, using NaI(Tl) /HPGe of an optimal energy for the photonsource;the stability detectors connected to a multichannel analyzer and a PC- and efficiency of the detectionsystem;thevarious cor- Apple H E for obtaining differential tomographs. The pre- rectionsfor deadtime, background, beamhardening and sence of Ag, Pb and I in different samples have been stu- scattering; the system of evaluating the images. died by using secondary targets of Sn/Sb (Ag-isolation), This scanner will be used primarily in a research setting. Y/Zr (PB-isolation) and LaI/Lall (I-isolation). This method The high precision of the measurements allows the evalu may be used for studying the I-flow in blood vessels and to ation of changes in bone density through disease or treat enhance the presence of a given tracer element used in ment over weeks and months rather than years.Smaller medical diagnostic applications. patient groups will suffice for-obtaining significant data as the measured changes can be evaluated for each individual ° Intern. Centre for th oretical Physics (Trieste-Italy) for patient, and no need for averaging over large groups is Training and Research in Italian Laboratories. necessarytoreduce thenoiseofthe measurement method.

MP29.6 MP29.7 Precise Delineation of Radiotherapy Targets Geometrical Aspects of Computed by Quantitative Computed Tomography, T. Lei Tomography: Sensitivity Profile and W. Sewchand *, Dept. of Rad. Oncology, and Exposure Profile, R.M.Gagne, Ph.D., Univ. of Maryland Hosp., Baltimore, MD 21201 Office of Science and Technology, CDRH,FDA,Rockville,Maryland A major disadvantage of current CT-based techniques for delineating radiotherapy targets is their reli- A simple model has been developed for the genera- ance on visual perception and subjective selection tion of theoretical sensitivity and exposure of CT-value thresholds. A new quantitative automatic profiles perpendicular to the tomographic plane method has been developed which accurately identi- in computed x-ray tomography. The model incor- fies and >recisely quantifies a target-of-interest porates the functional dependence on scanner (TOI) on the CT image. The basis of this approach geometry, focal spot size and shape, and detector is that the pixel CT image of normal tissues has an sensitivity. The sensitivity and exposure profi- asymptotic Gaussian distribution and isergodic in les are best depicted as the convolution of mean and variance. Thus, estimates (no/ ) of the functions, when appropriately scaled, describing mean and variance are obtained by spatial sampling the focal spot distribution and the transmittance and spatial averaging. A CT-value filter with pass- of the pre- and post-patient collimators. Pre- band (m-kv,m+kv) is designed to create an accurate dictions of the model agree well with experimental isolated image of a specific normal tissue. When results on both sensitivity and exposure profiles multiple normal tissues are enclosed in a region -of for small and large nominal slice thicknesses -interest (ROI), unique boundaries between images from five different CT x-ray systems. The CT of these tissues are obtained by a suitable choice x-ray systems selected represent both state of of k-values that ensures no overlap between tissue - the art and older scanner models. Comparisons specific filter passbands. When a TOI (e.g. tumor) are also made on the degree of matching of the is included in the ROI, specific normal tissues are sensitivity and exposure profiles for each scanner subtracted from the ROI to yield an isolated true which can be used as a measure of geometrical image of this TOI. efficiency in the direction perpendicular to the This approach provides an accurate 2D/3D display of tomographic plane. TOI's and generates significant necessary data (area /volume, density distribution, relative location, etc.) for quantitative evaluation of response to treatment. Examples of this application will be presented. 127 1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas 105 Medical Physics Scientific Papers

MP29.8 MP29.9 Observer Efficiency in Size-Discrimination on CT Images, Geometrical, Sampling, and Algorithmic A.R. Bleier, P.F. Judy*, R.G. Aspects of Computed Tomography Swensson, and I. Kazda,Harvard Medical School and Brigham and Women's Hospital, R.M. Gagne: Ph.D., Mary Pastel Anderson,M.S. Boston, Massachusetts, 02115 Office of Science and Technology, CORR, FDA, Rockviile, Maryland This experiment compared the efficiency of observers relativetophysical image calculations,in A complete, self consistent, analyticmodel has size-discrimination tasks of singlefeatures or feature been used to generate the theoreticalpoint spread pairs on CT images.The stimuli were images of two function (PSF) and the associated modulation disks added to CT scans of water.Eight sets of images transfer function (MTF) for computedtomography were studied at two levels of disk contrast.Observers (CT) using filtered backprojection.The model performed three rating tasks with each set of images: incorporates the functional dependence ofthe 1) the sizes of single features (smaller vs larger), 2) reconstructed PSF on scannergeometry, focal spot the size-differences of feature-pairs and 3) the av- size and shape, and detector sensitivity. The erage sizes of pairs. model explicitly includes aliasing effectsdue Within each specified task and disk contrast, the not only to sampling within the projections and mean da (an index of observer performance, derived angular sampling but also to thetwodimensional from the rating ROC curve) was prorvortional to Task sampling inherent in the reconstructedimage. SNR (a cross-correlation calculation that measures Unlike previously reported models, specialemphasis how accurately the physical image information could is placed on the impact of offsettingthe detector be used), with a slope (observer efficiency) less than sampling points in opposing views for unity. For low-contrast disks these slopes were similar scanners with detectors rotating through 360 degrees. In in all three tasks (0.386±0.011). For high-contrast addition, the effects of interpolationbetween disks these slopes were lower, only slightly in the sampling points and of the reconstructionalgorithm size-difference task (P.3521-0.004) but considerably so are also considered in the development.Predictions in the single,-feature task (0.267±.006) and the of the model for the PSF and MTFusing the known average-size task (0.260±0.014). geometrical characteristics of thescanner and The higher level of observer efficiency in the size- estimates on the reconstruction algorithmfrom differencerating task suggests that it is a more empirical noise power spectraagree well with "natural" task for observers to judge differences in experimental results for the PSF andMTF from four the size than the sizes of single features or the different CT xray systems. For these comparisons, averagesize of feature-pairs. the significance of the aliasingerrors is enumera ted 'n the final results.

MP30.1 MP30.2 On a numericalcalculation of energy A Monte Carlo Method for Electron deposition by broad high energy electron beams, Beam Dose Calculations, H. Neuenschwander, E.J. Born* H. Huizenga * andP.R.M. Storchi, Rotterdam Dept. of Medical Radiation Physics, Inselspital: Radiotherapeutic Institute, 3075 EA Rotterdam, The Netherlands. University of Berne, CH-3010 Berne, Switzerland A macroscopic Monte Carlo (MMC) methodbased on Dose deposition by broad high energy electron traditional Monte Carlo precalculationsas sug beams as calculated by Monte Carlo codes have gested by Mackie and Battista (1)has been deve been reported by several authors. These lopped. Precalculations were done using the EGS4 calculations take a lot of computer time and the code system to simulate the results transport of perpendi feature statistical uncertainty. Up to cularly incid9t electrons through now Monte Carlo macroscopic calculations are not suited for (typ. 0.5 gcm ) spheres. radiation therapy planning. The symmetry of Therefore, the this geometry minimizes storage feasibility of a requirements for numerical calculation method the resulting probability distributions has been investigated.The method is based only of emerging electrons at the cost of on the basic physicalinteraction processes of a rotational transforma tion that has to be performed after each high energy electrons and matter.The method is electron a drastically improved version of the phase step during the MMC calculation. In order to get space time evolution method as described by physically reasonable results, severalessential Cardoroand Zucker (Nucl.Sc.Eng 45,107,1971). modifications to the original concept (1)are ne The calculated depth dose curves, energy spectra cessary and will be discussed. MMC results for and angular distributions agree very wellwith calculations in water are in goadagreement with the Monte Carloresultsof Amdreo and Erahme full EGS4 simulations andwere obtained 1 to 2 (Rad.Res. 100,19,1984).For the general case of orders of magnitude faster (dependingon parameter radiation Thirapy planning, combination of this values) than EGS4 results for slabgeometries. method withthe "redefinition pencil beam This time gain is expected to besignificantly method" of Storchi and Huizenga higher for simulations for heavierelements and for (FM13,30,467,1985) andShiu and Hogstrom (9th dose calculations in closemeshedscoring grids. ICCR,1387) seems promising. Current work concerns the generalizationof the MMC code to accept measuredenergy spectra as input data and to consider inhomogeneitiesby taking into account pixel based density information. (1) Mackie, Battista; Proc. 8th Int.Conf. on the 128 Use of Computers in Rad. Therapy, 123,1984 106 1988 World Congress on Medical Physics and BiomedicalEngineeringSan Antonio, Texas Medical Physics Scientific Papers

MP30.3 MP30.4 A Model of Fast Electron Penetration LOW ENERGY ELECTRON DIFUSION IN TISSUE LIKE MEDIA D.J. Perry*, S.K. Raisis, F.C. Deibel, and C. Malamut*, Int.Rad.Dos./BRAZIL F.M. Khan. Univ. of MN., Dept. of Therapeu- P.J.Paes-Lece and A.S.Paschoa,Pont.Oniv.Cat./RJ.,BRAZIL tic Rad., Minneapolis, U. 55455

We introduce a multiple scattering model of The spatial - energetic distribuction of lou(E<50 key) charged particle penetration basedon previous electrons was studied for a source located in a medium analys3s ofYangand Perry. Ourdevelopment simulating biological tissue. The time independent Boltzmann removes the main limitations of the closely related Fermi-Eyges picture but retains its equation was used to model mathematically this distribution. considerable pot.ntial for successful therapeutic Ionization was assumed to be a pertubatioa to a quasi - elastic applications. collision process between the electron and the medium. A

diffusion limit was obtained, by using a scale parameter , In the time dependent picturediscussed leading to a sequence of recursive partial differential first, particles penetrating a material are found on thin curved surfaces which expand and deform . equations, whosesolutionswere obtained by numerical as they move away from their point of origin. approximations. Electron ranges were estimated based on this These surfaces are an exact feature of the multi- solutions and then compared with values reported in the open ple scattering transport equation and are defined literature based on experimental and Monte Carlo aethods.Local by three moments of the particle density: the dosimetry of low energy electrons from internal emitters can average depth of penetration and the lateral and benefit by the knowledge of these estimated ranges in longitudinal widths. The model is indifferent to biological tissue. Auger electron emitters, for exemple, have the use of small or large angl '3s. been object of a number of investigations, not only due to We use this approach to construct spatial their toxicity in general, but also due to their potential distributions under time independent conditions, appplication in the selective irradiation of neoplanstic cells. giving results of practical interest. We compute the energy absorbed in thinplanes located at various distances from the source and the lateral extent of these distributions. Results are compared against a previous Monte Carlo computation for 5, 10, 20 MeV electrons incident on water.

MP30.5 MP30.6 ELECTRON TRANSPORT IN NON-HOMOGENEOUS MEDIA Monte Carlo evaluation of methods for Rospana C. Falcao+, C. Malamut*, P. J. Paes-Leme§ deriving electron pencil beam spread. +Inst.Fisica-UFRJ ,*Inst.Rad.Dos.-CNEN,Dept2.Katemat.PUC/R4 G.A. Sandison*l, W. Hudal, J. Battistaand Due to the importance of obtaining, with high degree of accuracy, D. Savoie'. 'Manitoba Cancer Foundation, 100 Olivia Street, the dose variation at the vicinity of the interfacebetween two Winnipeg, Canada, R3E OV9. media, we propose a mathematicaly rigorous method to calculate it. 2Cross Cancer Institute, 11560 University Ave., The Singular Eigenfunction Method is used to solve a one speed Edmonton, AB, T6G 1Z2. one dimensional Boltzmanitransport equation P 60k,11) +0(x,11).411.(x,1)41 The EGS4 Monte Carlo code has been used to gen- ( 1 ) erate "benchmark" dose distributions resulting .(x,0)is the particle flux at poit x, with direction is thethe from electron pencil beams of 16 and 22 MeV cosine between the x axes and the particle's velocity). energy incident upon homogeneous muscle, bone and For an infinite medium eq.(1) has the general solution lung equivalent media. A repeat set of Monte +x/v ;x/v -x/v 1 AP)+11A()4' .(x,11).a e o +aO4, o +1 e dv ) Carlo data was also generated in witch brems- 0--o strahlung is ignored. These data are compared where to three methods of deriving pencil beam dose ).vc .(11). cv + A(v) 6(v- 11) 2(v..0) distributions from therapeutic electron beams. 'Vco+21vp) 2Ivp) and v The results indicate that systematic errors The coefficients a a and A( v)are to be determined by the 0+' 0- associated with each experimental method can lead boundary conditions and ortogonality relations along the 4,s. to substantial errors toward the end of the elec- Suppose we have two adjacet media, a source at x -te and that tron range. The largest contribution to these Assuming that the flux might be continuum at the systematic errors is the bremsstrahlung back- limp(x0.0. ground present in collimated broad beams.A new interrace, we can obtain the fuxes at medium 1and at medium 2 empirical formula based on the Monte Carlo data +x/v x/ . (x,p) ap) e 0 e-4% '+f 1A (v).(p)e vdv to modify the pencil beam spread parameter at o- + o 1,1 depth in tissue predicted using Fermi-Eyges theory -x/v 1 x/v .2 (x,10. ap) e 0 + f.A (v).( d v is suggested. Presently we aA clICulating i;`(21), '0+(2), Al(v ) and A2(v ) trough numerical methods since the expressions which give these coefficients cannot be integrated analliticaly.

1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas 107 Medical Physics &dentine Papers

MP30.7 1P30.8 Effect of Dimensionality ofHeterogeneity 3-D Dose Calculations forTotal Correctionson Three Dimensional Electron Pencil Skin Electron Therapy, L.D. Beam Calculations, G. Starkschalll K.R. Hogstrom, Simpson*, Washington University S.W. Bujnowski, L. Wang, and A.S. Shiu, Dept. of School of Medicine, St. Louis, MO, RadiationPhysics,'Univ. of TexasM.D. Anderson USA. Hospital and Tumor Institute, Houston, TX 77030.

Electron beam dose distributions have been cal- Total skin electron beam (TSEB)therapy culated on a three dimensional grid using a pencil for mycosis fungoides may requirethe use beamalgorithm taking into account effects of het- of 6-8 different patient positionsat 3-5m erogeneities in either one, two, or threedimen- SSD. Each position may require 2-3 low sions. Execution times for each of the algorithms energy electron fields at split gantry on a VAX-based treatment planning system are compositions. A 3-D dose computationalmodel pared. Execution times are comparable for both the will be presented which permitsanalysis oneand two dimensional heterogeneity corrections, of alternative TSEB treatmentplans. in spite of the fact that the two dimensional het- Single-field, experimental input dataare erogeneity correction requires an additional calcu- required: off-axis profiles for normalin- lational loop over fan lines.Execution times for cidence;depth dose asa function of the three dimensional heterogeneity correction are incident angle; and the dependence of the approximately a factor of ten longer than that for dose at dmax on distance from thesource. the two dimensional correction. The logic and The patient surface (or a segment of approximations the employed to optimize the algorithms patient'ssurface) is represented as for speed will be presented. Finally, the potent- points on ellipsoidal surfacesegments. ial applications of thesealgorithms in electron The total dose, at each surface point, 'beam treatment planning are discussed. at depths normal to the surface is givenas the sum over all clinical positionsfor Thiswork supported in part byNCI grant CA06294, all physical fields. The model produces NCI contract CM67914, and a research contract with results consistent with the phenomena International General Electric of New York, Medical experimentally and clinically observedin Systems Division. these treatments. Several examples will be discussed.

MP30.9 MP31.1 Application of Measured Electron Beam An Instrument for Producing Constant Bremsstrahlung Distributions to a Laminar Flow in an MR Imaging Phantom, Pencil Beam Algorithm, J.E. Rodgers , R. J. Kriz*, R. Stortzum, G. Dobben and G. F. Popescu and S.N. Rustgi, Georgetown A. Horowitz University Hospital, Department of Radiation Medicine, Washington, D.C. An imaging phantom has been designed that 20007 will produce constant laminar flow at variousaverage velocities through rigid tubes of various diameters. Accurate assessment of pencil beam algorithms The phantom consists of six circulartubes 20 cm requires a proper representation of the brems- in length, having diameters of 0.102cm, 0.207 strahlung component, especially at high ener- cm, 0.318 cm, 0.397 cm, 0.476 cm and 0.794cm. gies. Previous measurements (1) of bremsstrah- A large reservoir supplies the tubes. lung distributions in 6-18 MeV electron beams Constant flow of 0 to 10 cc/s is provided bya from a Varian Clinac-18, which demonstrated that separate reservoir specifically designed to bremsstrahlung is produced primarily in the maintain a constant head pressure. accelerator head, have been parametrized by the The difference between the fluid levelin the product of the reciprocal of a cubic function reservoir and exit overflow is adjustablefrom (beam envelope) and a Cauchy function near and 0 to 15.3 cm and kept constantto within 1 mm. beyond the cone edge. This model was incorpora- The pressure produced by the differentialfluid ted into the MDAH(Hogstrom) computer code and level is reduced by the fluid passing compared to measured ionization distributions. through an orifice in the exit tubing from the phantom. Pencil bekim parameters yielding best agreement Thus the pressure drop across the phantomis will be presented along with effects these ,ontrolled and constant. improvements have on ability to account for simple slab inhomogeneities and air gaps.

(1) Rustgi, S.N. and Rodgers, J.E. Med. Phys. 14(5), 884 (1987).

108 1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas Medical Physics Scientific Papers

MP31.2 MP31.3 3D Magnetic Field Mapping and Confirmation A Chronicle Of Philips Gyroscan S5 of Hazard Zones for Clinical MRI, Suresh M. MRI System: The Evaluation Results Brahmavar, Susan Majka, Martha LaFrance* and Of Image Qualicy Parameters For The J. Robert Kirkwood, Baystate Medical Center, Past 3Years, Pei-Jan Paul Lin, Inc., Springfield, MA 01199 Northwestern University Medical School, Chicago, Illinois 60611 The 3D magnetic field mapping of a clinical MRI U.S.A. (1.0 Tesla) :!acility was done to meet compliance requirements of DPH and DON. The customary exten- Various MRI image quality parametersand sive planning support provided by the manufacturer descriptors havebeen evaluated and was augmented by out own efforts to confirm regu- recordedbased on a set ofmeasurement latory magnetic hazard zones of 15, 5, 3 and 1 protocols for a Philips Gyroscan S5 MRI gauss lines. Such confirmation wqs of utmost im- system. The performance of the MRI unit portance because of unique location of MRI fazili- has been recordedstarting from the ty at Baystate Medical Center. The third flJor prestaging/acceptance at the factory MRI facility is surrounded by other areas such as [Time: 0], attheinitial acceptance pathology offices, clinical laboratories, lab com- testingconducted at the time of "turn- puters, electronic cables directly below and adja- over" to the Hospital [4th Month], at the cent patient rooms, gift shop, etc. The 3D magne- time of system re-tuning (14th Month], at tic field measurements were carried out by two the timeofsystemup-grading (26th gauss meters and axial probes calibrated using Month], and at the end of two and ahalf 100G, 200G, 1000G standard magnets. Several mea- year of clinical services [34th Month]. surements were taken at different locations on The results, the improvements and/orthe long axis as well as on the short axis of 1.0 Tes- lack of improvements over the 2-1/2year la MRI magnet. The ten (10) magnetic field zones time span after being placed for clinical ranging from 500G to 1G were established. The re- imagingcan be employed toestablish a sults were utilized to implement radiation safety basic quality assurance (QA] programand and quality control (RS/QC) program for MRI facili- to optimize the QA procedure. Abrief ty. The working guidelines include safety precau- discussionona few spec.ific problems tions for magnetic field lines less than the FDA encountered suchas low "signal-to- caution zone of 5G. noise", RF shielding, andshimmingof magnet will be included.

MP31.4 MP31.5 Utilization of Inversion Recovery Pulse A Quantitative Study of Magnetic Sequences with Short TI at High Field, Resonance Imaging, W.K. Chu, S. Albert* and N.E. Leeds, Department University of Nebraska Medical of Radiology, Beth Israel Medical Center, Omaha, Nebraska, USA Center, New York, N.Y. The objective of our study was to Inversion recovery (IR) pulse sequences with sho- invest'gate, through quantitative analyses rt inversion time (TI) were developed and applied under theframework of pattern chiefly at low (0.15T) and medium (0.5T) magnetic recognition, the possibility of enchancing field strengths. The objective of the study was d et ec ty and/or specificity of to develop useful IR pulse sequences with short lesions shown in Magnetic Resonance TI times for high field imaging and to discuss Imaging ( MRI ). Sixty sets of MRI liver and demonstrate the analytical and practical con- images were studied. Thirty were obtained siderations imvolved. from patients with hemangioma, a common Our studies have been carried out at a field of benign liver tumor, and thirty were from 1.5T with a whole body superconducting imager (GE patients with other types of localized SIGMA). Spin-echo data collection is used in our lesions. A 1.5 Tesla MRI scanner was used experiments and proper interleaving is required for data collection. Quantitative features in the multi slice sequencestooptimize acquisi- measured included grey scale intensity tion time. An additive contribution to contrast values and variations of normal liver from both T1 and T2 and fat suppression were tissues, intensity values and variations obtained with a short TI time of 100ms. Signal within lesions, calculated T2 relaxation suppression of any tissue can be achieved by app- and backgrounds.More features were ropriate selection of TI value. developed based upon these measured Comparison of the IR images with spin echo (SE) values. After feature selection process, T2-weighted images demonstrate in most cases a was found that a normalized intensity better contrast to noise ratio, a shorter acquisi- ratio between the lesion andit's tion time and reduced chemical shift artifacts of surrounding normal liver tissues provides the IR images. the highest separation between hemangioma and all other localized liver lesions. With a threshold ratio of 16.8 an accuracy of 95 percent was achieved. I 3 1 1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas 109 Medical Physics Scientific Papers

MP3 1.6 MP31.7 Dispelling Some Doubts About Back Projection ASystemfor Tissue Characterization Reconstruction in MR Imaging, Studies by NMR, K. Homma* and E. A. V. Lakshminarayanan% Dept.of Radiology, Takenakal , Mechanical Engineering Laboratory, MITI and National University of Alabama at Birmingham, 619 South DefenceMedical College' , 1-2, 19th Street, Birminy im, AL 35233 Namiki, Tsukuba Science City, Ibaraki 305 JAPAN Itis widely believed that the(back) projection reconstruction technique of acquiringMR images is more susceptible to A system for tissue characterization studies in Medicine and Biology is proposed. The system magnetic field non-uniformities that 2DFT methods and results is constructed by a NMR (Nuclear Magnetic in images with poorer signal-to-noise characteristics. This belief Resonance) and two subsystems an image is due to misimplementation of the projection reconstruction processing system and an expert system. The image processing system installed in a computer scheme and a failure to compare equivalent scan strategies. extracts features from three dimensional NMR Images are presented that demonstrate the robustness of the images reconstructed in accordance with back projection method in the presence of inhomogeneities imaging methods (MRI). Tissue functions are and the relative advantages and disadvantages of the two also analyzed by the chemical shiftanalysis (MRS). On the other hand, the expert system approaches. Compared are signal-to-noise ratios, field and RF evaluates and characterizes biological tissues uniformity requirements, gradient requirements, the ability to based on the extracted features (spin density, obtainhigh-resolutionimages, effectofmotion, and relaxation times and three dimensional computational complexity. structures). The system is equipped with a knowledge base concerning these features in normal and pathological human, rat and rabbit tissues. The LISP language is used for the tissue characterization. The system communicates directly with a PACS (PictureArchiving and CommunicationSystem).

MP31. 8 MP31. 9 PlIOT0t,.1T.,-DEPENDENT CALCULATED Biexponential transverse relaxation VOLUW.i.S FOR MAGNETIC RESONANCE IMAGES IN CANINE GLIOMA.Robert L. accuracy of an 0.5 T MR imaging unit. Galloway, Jr., Robert J. Maciunas, Anne A Fransson*, Karolinska Institute, S-10401 L. Failinger. Vanderbilt University, Stockholm, A. Ericsson, B. Jung, University Nashville, TN. Hospital, S-75185 Uppsala, Sweden. Primary CNStumors have a structure which includes asolid tumorcore surroundedbya zoneof In magnetic resonance imagingbiexponential infiltratingtumorcells. Theplanningofsafe transverse relaxation processes in biological tissues resections andefficacioustherapiesdependsona have been reported. To investigate the capabilities clear understandingofthe tumormarginsand volume. Inthe presented study, cloned gliosarcoma of biexponential resolution of our imaging unit, cellswereimplanted inadult,mongreldogsand measurements were performed using a binary grown for two weeks.Each animal was imaged using phantom as a model. Mono- and biexponential four common magnetic resonance (MR)timing sequences and protocols: T1-weighted, T2-weighted, relaxations were assessed with software developed protondensity and gadolinium-enhanced T1-weighted fordiscretemultiexponentialfunctions.Only images. The volumes were determined by marking transverse relaxation was investigated by applying tumormarginsonspatially-sequential,parallel,slice a spin-echo-sequence with 32 echoes. images. Theimages werethen computationally The results indicate a relatively accurate and stable stacked and the volume was calculated basedon the imagepixelspacing and theslicethickness. Each resolution of the slow component in different sequence provided a different,distinct volume. The environments. Theratioofrelaxationrates possiblesourcesfor these differencesandtheir relationship between the two components demands a factor of totheunderlyingphysiology and "true" two for a satisfactory resolution probability. tumormargins are discussed. Thisworkwas supported in part by a grant from Sun Microsystems, A system specific component systematically added Inc., Mountain View, California. to the signal in one direction has been found, resultinginbiexponentialrelaxationfrom monoexponential signal-input. 132 110 1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas Medical Physics Scientific Papers

MP31.10 MP31.11 Development of a protonNMR imaging A Comptex Phantomfor a Fast and Complete image system based on industrial standard QUaltty naluationto MRI B. Aubert* N. Aubert and VME-bus for Bio-Medicaland Bio- F. Aubry, InstitutGustave-Roussy, 94805 Villejuif technical applications,L. Malmgren, Cedex, France.

B.R.R. Persson , B. Blad, T. Pole, M. Frank, Lund University, Dept of In order to perform a fast, easy but complete ima- MRI, a complex phantom Radiation Physics, S-221 85 LUND, ge qualityevaluation in Sweden. has been designed. It consists of a perspex cylinder (20 cm diameter, 30 cm length) containing five The aim of the present project is to deve- identical modules surrounded by a copper sulfate lop a semiportable general NMR imaging resolution. Threeof these modules are located at search system for both in vitro measure- the isocenter of the phantom in order to study the ments and in vivo imaging. This system will axial, coronal and sagittml plmees, the IAA, alien: fill the great need for low-field NMR-equip- at a distance of t 11 cm (away from theisocenter) ment to be used for measurements of tissue to study the multi-slice mode in the axial direc- samples and imaging of experimental animals tion. Moreover, oblique planes crossing the cylin- in parallel with the use of costly high- der axis can also be analysed. Each module allows aproton ficld clinical NMR-systems. Thus to evaluate fourteen parameters including unifor- NMR imaging system for experimental research mity parameters, slice parameters, spatial resolu- have been developed based on a broad band tion, geometric distortion, T1 and T2 measurement, spectrometer, two high resolution (16 bit) ADC's, and a four channel programable gradi- ghosting, ... This phantom is very useful for daily qualitycontrol (10 minutes acquisition time) ent generator all compatible to a VME-bus. after We have an air-winded whole body Oxford mag- or for a more complete quantitative analysis net shimmed at 0.07 tesla (possible to rise preventive or curativemaintenance, or modifica- 0.15 tesla) which is specially dedicated to tion of the machine. It has been used with various machines. A special software giving an research and technological development. The MR imaging NMR imaging system is built as a stand alo- automatic imageanalysis has been developped on a ne system which can be used withvarious ty- Signe System(1) (General Electric). pes of magnets. The very flexiblegradient controler and pulse programmer will select (1) CIERM - H8pita: de Biatre - all desired pulse sequences required for 94290 le Kremlin -Bicatre, France. different applications.

MP31.12 MP31.13 A Coil Desensitizing Technique to Suppress QA/QC Phantom for NMR Imagers. MS Rosenthal*, University of Pittsburgh, Pittsburgh, PA, Respiratory Artifacts in MR Imaging of the 15261, USA Spine, A. V. Lakshminarayanan*, Dept. of Radiology, University of Alabama at Birmingham, We have developed a QA/QC phantom for NMR imagers. 619 19th Street South, Birmingham, AL 35233 This phantom, constructed out of acrylic, provides the ability to measure slice thickness, slicespac- ing,SNR, CNR, linearity, and intrinsic sensitiv- Motion of the abdomen (in particular, the subcutaneous fat) ity. The phantom consists of a cylindrical tube introduces artifacts in 2DFT MR imaging which seriously with two inserts. The first one is another smaller interferes with the diagnostic utility and quality of lumbar spine cylindrical tube with an acrylic rod at an angle examinations. Flat (License Plate) surface coils tend to reduce for slice measurement. The second insert consists of a 'plus' sign arrangement of acrylic with holes these artifacts in spine images but result in nonuniform signal for 13 tubes containing H2O and D20. For each brightness in the PA direction.Described is a receiveonly tube the concentration of water varies from ap- "surface coil" based on the well-known Helmholtz type of proximately 1 mmol/mL to 55 mmol/mL. The rest of design. The geometry of the coil is such that its response is the phantom is filled with a copper sulfate solution and propanol which is used to inhibit growth In addition a mechanism has been added to uniform. in the solution. desensitize the imaging system to signals from moving We have used this phantom to test an NMR system as structures both during RF transmission and reception. well as compare the chAracteristics of different Images are presented that demonstrate a high degree of pulse sequences. This phantom provides an easy method to test the Also uniformity and the suppression of motion artifacts. intrinsic sensitivity of the NMR system which will presented are images that demonstrate that the same indicate changes in machine function. techniques can be used to eliminate fold-over artifacts.

`s' 3.

1988 World Congress on Medical Physics and Biomedical Engineering c SanAntonio, Texas 111 Medical Physics Scientific Papers

MP31.14 MP31.15 NMR Local Imaging by the FresnelTransform Reconstruction of intracranial vascular systemusing Technique using a Nonlinear FieldGradient, 3-planes magnetic resonanse tomograms,Naoki Y.Yamada*, S.Ito, K.Tanaka andz.Abe, Suzuki",Kouji Utsunomiya Univ., Utsunomiya 320,Japan. Simizu2,Tetsuo Oamura' ME Research Unit JikeiUniv. School of Med.' Shimazu K.K.2

NMR Fresnel transform imagingtechnique proposed by authors uses a nonlinearfield gradient with It is very difficult to reconstruct the threedimensinal a quadratic field intensity profile. The center (3D)image of intracranial vascular systemwith only a set position of the nonlinear field gradient, where of transverse CT image of the cranium.Becausethe intra- the differential gradient iszero, is moved at a cranial vascular system shows frequentlycircling and fixed amount in each successiveimaging pulse meandering. We tried to reconstruct the 3Dimage of whole sequence to produce nonlinear phaseencoding in intracranial vascular system using 3-planes that moving direction..A spindensity image is tomograms. reconstructed from collected signalsby the com- The CT examination was performed bya super conductive MR! puter processing. The imagingregion in this scanner(SNT-50,Shimadzu) and transverse,sagital and cora- technique is limited to themoving range of the nal images are utilized as 3-planes tomograms. nonlinear field gradient. When suitable conditions Only transverse sectioned vascular images are selected, the aliasing artifact in each tomogram is not pro- are picked up and processed for the 3D image.Vertically duced in the reconstructedimage. The conditions running vessels along Z axis for producing local image withno artifacts are are reconstructed by a set of discussed and experimental resultsare presented. transverse tosograms.ln the same way,vessels runningalong Unlike the case of conventionalFourier trans- X and Y axis are reconstructed bya set ofsagital and form technique, no selective excitation procedure coronal tomograms, respectively.Thesethree images are or no use of surface coil is needed to oktain a combined in 3D coodinates toa shale vascular local image. A spatial resolution system in better than 1 the cranium including carotid arteries and jugular mm can be easily obtained. Moreover, veins. the imaging Anterior,middle and posterior cerebral arteriesand time in this technique is comparableto that of basilar artery were also able to display the conventional Fouriertransform technique. on the 3D image. In this method, it was possible to display notonly arteries but also veins such as superior sagital,transverseand sigmoid sinus in one image.

MP31.16 MP31.17 141 Cases of Brain SPECT Imagingwith 99mTc-(d,3.)HM-PAO and Quantitative Quality Control Studios in Estimationof Analysis, Yuan Aina and Zhao Huiyang*, Lesion Size in SPECT,M.M.Rehani* Zhong Shan Hospital, Shanghai Medical Department of Medical Physics and Bio-Engg., University, Shanghai, China. Sher-I-Kashmir Institute of Medical Sciences, Srinagar-190011,INDIA.

This paper describesa recent study of The accurate estimation of lesionsize in SPECT 141 cases with SPECT ofthe brain using requires stringent quality control workup and 99mTc-(d,l)HM-PAO. Among 141cases, 93 evaluation. Besides flood acquisition and off- cases were examined by CT and 6 confir- set correction, a number of parameters like med by surgical operation. Themethod matrix size,count rate, choice of tilter,atten- used in this study allows theproper uation etc.affect the results. For out study and quantitative diagnosis of the regi- we used ZLC 3700 gamma camera with Array Proce- on and the degree of regional ischemic. ssor and Data Spectrum Phantom with six plexi- Comparisons with CT were performed.Re- glass cold spheres of dia 31.8,25.4,19.1.,15.9, sults show that ror stroke andtumor, 12.7 and 9.5 mm. Tc-99m activity ininjection the ischemic regions obtainedare con- vial was used for simulatinga hot source in sistent with those obtained usingCT, cold surrounding. The number of pixelsalong while the method used is muchbetter the diamet,r and voxel estimation in different than CT for the diagnosis of transient silices was resorted to for estimatingthe ischemic attack, migraine, epilepsy, size. The results show that,for the countrates dementia and slight trauma. In addition, normally encountered in clinical practice,the this method is used to indicatethe de- choice of 64x64 matrix, use of Shepp- Logan- gree of these brain diseases and itcan hanning reconstruction filter,appropriate off- also be used to detect the locationof set correction and attenuation correction result the epilepsy focus, to predict the in estimation of lesion size within + 7.5X stroke and to measure the blood flow between 16 mm to 32mm. This study has great of tumor. utility in quantitative imaging.

112 1988 World Congress on Medical Physicsand Biomedical Eng:rxering ' San Antonio, Texas Medical Physics Scientific Papers

MP31.18 MP31.19 ASSESSMENT OF MYOCARDIALSPECT RECONSTRUCTIONS Synthesis of Stereo Pairs From Reconstructed SPECT USING A HEART PHANTOM UNDER CLINICAL CONDITIONS Data for 341 Viewing, E.A. Silverstein*, S.M. *1 41 2 Spies, A.M. Zimmer and W.G. Spies. Northwestern C.C.Robilotta , M.F.S.Rebelo , R.Abe Memorial Hospital, Chicago, Illinois. 1-Instituto de Universidade de Sao Paulo 2-Instituto do Corn-go, Hospital das Clinicas In this work, an attempt is made to utilize the Sgo Paulo, S.P., Brasil human stereopsis capability to obtain a more effective presentation of SPECT data. Planar In order to assess the effects of 360° and parallel projected views were synthesized from 180 samplings and those from 3 different' 64x64 reconstructed SPECT data and viewed as phantom reconstruction methods, cardiac Rather than displaying opaque plastic stereo pairs. studies had been developed. The surface contours, our displays have the form of a heart phantom, containing 2 small infarct- We believe that this style in an elliptical translucent image. ed regions, is placed makes better use of the available data, since the phantom, simulating the human chest. Both whole data set is always in view. Viewing was 360° and 180 data had been collected with done by various means: Pairs were shown side by the whole phantom, with and without scatter side on a CRT or 8x10 film and fused with a stereo medium, positioned under typical clinical Alternating blue left and red right conditions, using a single-head rotating viewer. images in the same location on the CRT were viewed camera. A second positioning with the heart A centralized along the rotation axis has with blue left and red right eye filters. commercial liquid crystal polarizing display was also been used. The kinds of images vied included the also tried. Backprojection reconstructions using Tc-99m PYP cardiac images, T1-201 myocardial ramp, Hann and Parzen filters had been ap- images, Tc-99m PYP bone images, and labeled plied to produce the tomographic images. monoclonal antibody images with both 1-131 and In- Visual inspections, contrast-ratios and 111. Far less smoothing than usual was needed to count profiles analyses had been carried Synthesized single out and provide means for evaluating the 2 make the images acceptable. controversial sampling modes as well as to views suggest the possibility of eliminating planar imagingin some situations. Projected identify the scatter/attenuation effects enhancements include the use of motion by panning, and the reconstruction artifacts. tilting and rotation, the use of real perspective, § Financial supports from FAPESP and FINEP. and increase of resolution to 128x128.

MP31.20 MP31.21 Evaluation of Parallel Hole Collimators Used for The Effect of Different Flood Field SPECT Imaging. E.A. Silverstein* and S.M. Spies, Correction Methods on SPECT Images. A van Herbst, MG Northwestern Memorial Hospital, Chicago, Illinois. Aswegen , PH Pretorius, CP Letter and PC Minnaar, University of the Consistency of septal angulation of a parallel Orange Free State, Bloemfontein, South hole collimator is more important for spocr than Africa. for planar imaging. In this work two types of measurements, one local and one global, are Effective quality control is of utmost developed for evaluating this consistency. The importance in performing reliable single local measurement involves placing a point source photon emission computed tomography of radioactivity (Tc-99m) about 15 cm from the (SPECT) studies. SPECT images have to be non- collimator under test. A gamma camera faces the corrected for flood Field (FF) other side of the collimator at a distance of 50 uniformities prior to reconstruction in Three The sou-ce is fixed with respect to the order to eliminate ring artifacts. cm. i.e. camera. Thus, the same portion of the camera types of FF images were evaluated field is used for all measurements. Suitable (i) a Tc-99m filled sheet source (TSS), (both setups permit the collimator to be moved sideways, (ii) a Co-57 sheet source (CSS) or the camera-source system can be moved. acquired with a low energy high Variations in septal angulation cause the image of resolution collimator) and (iii) a Tc-99m without the point source to move. The long 50 cm lever point source (TPS) imaged (IU) arm amplifies the angular variations. A precision collimation. Integral uniformity all FF. The of 0.5 mm in measuring the image position gives a measurements were done on the FF on tomographic sensitivity of 1 milliradian, with greater effect of sensitivity realizable. The variation in septal resolution and uniformity was evaluated. angulation can thus be assessed. The global IU of TPS was superior to CSS and TSS for measurement uses images of point sources placed at central field of view. No significant for either a largo distance from the camera-collimator system difference could be found being tested. Visual and computer analysis of resolution or tomographic uniformity for these large paint source images reveal defects in the different FF. TPS is therefore septal angulation not apparentin the flood recommended for non-uniformity correction SPECT performed with a sheet source. provided collimators used for acquistron are intact. ..kc4.

1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas 113 Medical Physics Scientific Papers

MP31.22 MP31.23 A Comparison of Radiation Dose from A Quantitative Theory for the External Sources of Gd-153 and Eu-155, Measurement of Cerebral Blood Flow with 99mTc-HMPAO C.D. Frey*, C. Starr, C. Lam and and SPECT T.Y. Lee* and K.M. Spicer, Medical University of Yeung, Radiology, St. Joseph's ,spital and Medical Biophysics, South Carolina, 171 Ashley Avenue, University of Western Ontario, Charleston, South Carolina 29425. London, Ontario, Canada N6A 4V2

When single photon emission computed tomography (SPECT) is done simultaneously with transmission The radiopharmaceutical, 990Tc-HMPAO,has been developed with the computed tomography (TCT) expectation that it done with a can be used to measured cerebral radionuclidesource; blood flow the TCT data is used to (CBF) with SPECT. produce attenuation measurements However, because of its and scatter limited extraction efficiem%/(EE) corrections used to improve the fidelity of the and unstable binding to the brain parenchyma, SPECT images. 'This study calculated the quantif- the ication of CBF with it has radiation doses from transmission sources proved difficult of uaing simple models. We have Gd-153 and Eu-155. Doses were calculated developed a for theory in which the transport, slabs of tissue from 5 cm to 20 cm thick. transcapillary The exchange and reversible binding results are presented in the form of curves that of 99mTc- HMPAO are modelled. It is shown that the show the average dose to the tissue perprimary total tissue concentration of thematerial, At, transmitted photon. This data can be used to is related to the arterialconcentration, CBF select appropriate nuclide sources for TCT. and EE. At can be measured in-vivononinvas- ively with SPECf, and thearterial concentration of the lipophilic form ofHMPAO is measured by octanol extractionsuch that CBF can be estimated by non-linearcurve fitting using the operating equation ofour proposed mdoel.

MP31.24 MP31.25 SPECT Imaging by Using a Multislice QUALITY CONTROL Of SPECT SYSTEMS : Fan Beam Collimator,Y. Akiyama* S. Sakata and N. Yui, Chiba Cancer EFFECTS FROM SOME BASIC MISADJUSTMENTS 61 Center Hospital, 666-2, Nitona-cho, C.C.R2bilotta*1,M.A.Olivelra M.F.S.Rebelo61 Chiba-shi 280, Japan. R.Abe .1-Instituto de Fisica, Universidade de Sao Paulo, 2- Instituto do Corn-6o, Hospital das Clini- A multislice fan beam type casting collimator for cos, Sao Paulo, S.P., Brasil. SPECT was introduced in our institute, (hole size: It is wideiy known that quality control 1.8mm, septum: 0.3cm, hole length: 40mm, energy: of 160KeV, focus: 750mm, F.O.V.: 350mm). An inter- equipment conditions, data acquisitionand polation method developed for a fan beam type reconstruction methods are extremely critical in order to produce reliable SPECTimages. X-ray CT was modified and utilized to reconstruct Effects due to: a)detector the SPECT images.This algorithm was confirmed by gantry misalign- ment, b)unsatisfactory detector tuning, c) means of computer simulation. Primary photopeak data were obtained using a 20% energy window cen- wrong computer centre of rotation, d)inhomogeneity of collimator septa/holes tered at 140KeV for Tc-99m and 159KeV for 1-123. and others are often mentioned but seldomshown Beforehand we studied various ploblems concerning in the literature. the fan beam collimator by computer simulation method and by the use of a DIV/CON collimator, Point- and line-sources, cold- and hot-rods phantoms studies had been developed whose conversing side produces similar data to a for a fan beam collimator. wide field of view rotatingcamera coupled We comparied the fan beam to a low energy high resolution collimator. collimator and a low enegy high resolution parallel hole collimator. Reconstructions were accomplished using Concerning the resolution ramp, Hann and Parzen filters. and sensitivity the fan beam collimator is supe- The aim of this contribution is ior to the parallel hole collimator. to show the This was effects due to the factors a),c),d) confirmed by various phantom studies and clinical as well examinations. as source-detector distance, scatter medium and counting statistics on the sinograms The authors thank F. Kinoshita for expert techni- and the reconstructed images. cal assistance. This work was supported in part by a Grant-in-Aid for Cancer Reseach (60-39) from the Ministry of Health and Weliare, Japan. 6 Financial supports from FAPESP andFINEP. .136 114 1988 World Congress on Medical Physics andBiomedical EngineeringSan Antonio, Texas Medical Physics Scientific Papers

MP31.26 MP31.27 Analysis of reconstruction filters on Effects of collimators design on image tomographic noise in SPECT imaging,R. characteristics in tomographic gamma - Moretti, A. De Agostini, S. Belletti* camera system, A. De Agostini, R. Mo-

L. Mascaro,Health Physics Dept., Spe- retti, S. Belletti* , L. Mascara, Hea- dali Civili, 25100 - Brescia, Italy lth Physics Dept., Spedali Civili, 25100 - Brescia, Italy. Noise level of a reconstructed image modifies lesion detection capabilities of a tomographic sys- Selection of corrected collimator system for tem. Selection of the appropriated back-projection qualitative and quantitative gamma-camera SPECT filter function is of greatest importance in opti- imaging is of the utmost importance in optimi- mizing image characteristics. zing Nuclear Medicine tomographic procedures. We have analized the influence of various window Collimator design ( number, diameters and len- functions (Bt, Hm, Pz, Wn, Mz) on the noise level ght of holes ) modifies the physical parameters of images obtained with a rotating SPECT gamma-ca- of reconstructed images, such as spatial reso- mera system. lution, contrast, noise level, lesion detection Computer simulation of cold lesions (diameters and so on. ranging from 10 mm to 35 mm) embedded in a uniform Three types of parallel holes low and medium e- background activity with a Gaussian noise function nergy collimators were used for quantifying ima- distribution was compared with theoretical and ex- ge parameters. perimental data. Results show the importance of collimator choice Analysis covered usefull matrix size, angular step in modifying image physical characteristics, and and was extended to low, medium and high image co- were used for optimizing acquisition p.rameters unt densities. of clinical emission tomography. Results were used for obtaining practical indications to clinical implementation of automatic filter selection.

MP31.28 MP31.29 SPECT Reconstruction of Symmetric Image Properties of Digitized Radioisotope Distributions From a Radiographs, Y.Kodera*, M.Ogawa, Very Limited Number of Views. T. M.Fujita and T.Wada, Hiroshima Lowinger*, D.S. Marsden, Department University, 1-2-3 Kasumi, Minami-ku, of Radiology, St. Luke's/Roosevelt Hiroshima, Japan Hospital Center, New York, N.Y. 10025 A digital image processing system for a medical Objects with symmetry properties may be recon- x-ray image was developed. The system is made structed in a tomographic study from a limited up of a minicomputer, a high resolution laser number.of views. We studied a number of recon- scanner and a laser printer. A digital image was struction algorithms that have previously been obtained from a film radiograph by the laser developed for astronomy and electron microscopy scanner. The pixel size was variable from 56.6 and used them to reconstruct 3-dimensional iso- to 198.1 pm with a pitch of 28.3 pm and the tope distributions from a single projection. density resolution was 10 bit. An output image An algorithm invented by astronomers to deter- of the laser printer was produced with a pixel mine mass distribution in globular clusters has size of 85 pm at a 8 bit density level. been applied to tumors.Often a tumor will ex- As sigmoidal relationship was found between hibit radial symmetry in its uptake of isotope input and output densities, linear transformation thus, 3-D isotope distribution may be calculated was then made by use of a look-up table. from a single view. In bone scanning, cylindri- Resolution of this system was mainly determined cal symmetry is exhibited by parts of the humer- by the pixel sizes of the laser scanner and us and other bones. Such structures can be re- depended on the scanning directions. In a digi- constructed from a single planar view. We have tized image of screen film systems the radio- obtained tomographic reconstructions from a graphic mottle was reduced in high spatial fre- single view and from 2, 3, 4 views of phantom quency range, but a little increase in low fre- and patient data. The 3-D isotope distributions quency range. Some clinical radiographs were obtained compared well with theoretical predic- processed and would be presented. tions-and-actual SPECS studies. This method of tomography is available for any scintillation camera interfaced with a computer. 137 1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas 115 Medical Physics Scientific Papers

MP31.30 MP31. 31 Quantitative Uriformity Analysis in Radiological Image IntensifierTV Computerized Analysis of the Heart in DigitalChest Systems Using a Digital Image Processor. Radiographs.N. Nakamori*, K. Doi, V. Sabeti, H.-P. Chan, S. Katsuragawa, H. Mae Mahon. Kurt Rossmann P. Cooney; K.P. Maher andJF Malone, Lab. for Rad. Image Res., Univ. of Chicago,Chicago, F.D.V.H., Meath Hospital, )ublin 8. IL 60637 Uniformity is a well defined term used to describe the response of gamma cameras to a "uniform" Some of the most useful diagnostic information radiation input. It is not routinely included in in chest radiographs is related to the size andshape of the performance specification of xray image the heart. We are developing an automatedmethod to intensifierTV (XIITV) systems, where it can determine the contour of the projected cardiac image influence such applications as vascular, cardiac in PA chest radiographs and to obtaina number of and quantitative digital studies. Its measures related to the size and shape of the heart specification poses a number of problems such as including the long and transverse dimensions, thepro- the absence of any agreed measurement methods and jected heart area and the cardiothoracicratio. In tolerances, and the compound effect of the order to obtain "gold standards", four radiologists individual component uniformities. traced their best estimates of the entire contour of the heart on 12 radiographs, including the largely A quantitative study of XIITV system uniformity invisible inferior margin. These contours expressedin was performed, based on a refined version of the terms of radial distances from the center of the heart NEMA Definition for gamma cameras, with the use were analyzed by using Fourier analysis.Results of of a realtime digital image processor (Quantel, Fourier coefficients were used as a guide to obtainan I-100). Uniformities of 50% and higher, were approximate pattern of the heart image by fittinga observed. The influence of image noise and matrix shift-variant sinusoidal function to detected edges of size has been investigated. Measurements and the heart.The heart and diaphragm boundarieswere definitions were also developed for the individual determined from the detected edges in horizontaland system component uniformities and applied to a vertical profiles, respectively.We found that the number of systems in routine hospital use.A contours of heart images obtained withour automated considerable variation between systems was evident. method agreed well with those obtained byradio- A detailed analysis of the results will be logists. We shall present preliminary resultson presented. clinical usefulness of our method. Supported by USPHS Grants CA24806, HL36238,and The Whitaker Foundation.

MP31.32 MP31.33 The Use of Rare Earth Filtration in Image r ornzats: rive rears "vier Digitally Enhanced High Milleamperage me Anna ztanaara Format far Digital Image Interchange,G.Q. Maguire Jr:: Fluoroscopy, G.D. Frey*, R.A. Miller and M.E. Noz+, Department of and R,C. Hoigate, Medical University Computer Science, Columbia University, New York, NY of South Carolina, 171 Ashley Avenue, 10027 +Department of Radiology, New York University,New York, NY Charleston, South Carolina 29425. 10016 The publication of AAPM Report No. 10 was the firstattempt to The use of high mA digitally enhanced fluoroscopy standardize image formats in the medical imaging community. Since has made it possible to do operative argiography then, two other groups have formed (CART- the Scandinavian with mobile Carm devices. However whe conven collaboration for Computer Assisted Radiation Therapy threatment tional aluminum filtration (1 mm AL) is used the planning, and ACR-NEMA, a collaboration whose purposeis to skin entrance exposures can be very high. Expo formulate a standard digital interface to medical imaging equipment). sures greater than 30 R/min occur at 120 kV and The AAPM format uses key-value pairs in plain text to keep track of 20 mA. The use of a rare earth filter(*) reduces all informaticn associated with a particular image. The radiation the skin entrance exposure by a factor of a 2 at oncology community in the U.S. has been defining key-value pairs for 120 kV and a factor of 4 at 60 kV. However image use with CT, nuclear medicine and NMRI images. Additionally, both quality deteriorates if the rare earth filter is ACR-NEMA and CART have been defining fields for use with the used without adjusting the kV to lower values. same types of images. The time has come to unify the naming of If the kV is reduced image quality is maintained. these fields. The CART collaboration has introduced a data base This study determined the appropriate kV's touse which is available electronically, but is maintained by agroup of with the rate earth filter. Image quality individuals. ACR-NEMA operates through committeemeetings. To studies with phantoms and patient studies show insure a consistent set of field names in such a rapidly developing that substantial exposure reductions are possible arena requires the use of a server rather than a committee. Via a without loss of, image quality. server a person would inquire if a particular field had been defined. If so, the defined name would be returned. If not, the person would be *Rfilter RSP Inc., San Carlos, CA given the opportunity to define the field. The next inquiry would return the previously defined field. A number of computerized services already operate in this manner. As new modalitiesare adds l to the imaging repetoire, it would be easier and faster to ensure the consistency and adequacy of the data base; e.g., in the present version of its standard, the ACR-NEMA fields are adequate for CT but there are almost no fields suitable for describing the parameters associated 138 with nuclear medicine and NMRI images. 116 1988 World Congress on Medical Physics and BiomedicalEngineeringSan Antonio, Texas Medical Physics Scientific Papers

MP31.34 MP31.35 A CCD Photoreceptor System for Measure- performance of Antiscatter Grid in Digital ment of Image Quality Descriptors, F.J. Radiography: Effect on Signal-to-Noise Ratio. McCraw*, J.A. Bencomo, and L.M. Marsh, Chan*, K. L. Lam, Y. Wu, H. Fujita, K. Doi, Kurt University of Texas System Cancer Center, Rossmann Lab. for Radiologic Image Research, Dept. 1515 Holcombe Blvd, Houston, TX. 77030 of Radiology, Univ. of Chicago, Chicago, IL60637

A 256 pixel CCD line detector designed We havestudied theeffect of antiscatter grid for optical character recognition has on the signal-to-noise ratio (SNR) of low-contrast, been adapted for capture of its video quantum-noise-limited objects in digital radiography. output signal. This signal was condi- In atheoreticalanalysisof the SNR,we found that the tioned and syncronized for digitization effectiveness of a grid is determined by a performance by an IBM Data Acquisition and Control index, which is proportionalto the contrast improve- Adaptor in an IBM XT computer. A lens mentfactor and inversely proportionalto the square and mirror system is used to both focus rootof the amity factor of the grid. Results of Monte an image on the receptor and to remove Carlo simulation studies and experimental measure- the sensitive CCD device from the direct ments with an I.I.-TV digital system indicate that, for x-ray beam. Software has been developed a given patient exposure, the improvement in SNR to provide for acquisition control, achieved by the useof a conventional grid islimited. storage and processing of the data. An The reduction of scattered radiation, however, reduces initial evaluation of the device sensi- the relative effect of system noise on the SNR of a tivity, dynamic range and accuracy has low-contrast object, thereby extending the useful been performed. It is expected that this dynamic range of the imaging system. device can be used to facilitate measurement of image quality descriptors nec- Supported by USPHS Grant (MIRA) HL36238, The essary to optimize clinical image Whitaker Foundation, and USPHS Grant CA 24806. techniques.

MP31.36 MP31.37 /wage PC-Lady of the architecture of a wank- Use of the effective dose equivalent, HE, station for image processingut a medical in PET studies. W. Huda* and G.A. Sandison, environment E.M. Stmlehm and G.E. Grgantel*, Manitoba Cancer Foundation, 100 Olivia Street, Centro Ingegneria Biomedica deliginiversita. di Winnipeg, Canada, R3E OV9. Roma. Corso Vittorio E. 244 00178 ROMA The effective dose equivalent, HE, can be used as The personal processing of images is standard radiation dose parameter in all imaging goinv to be aprimary needin such modalities which utilize ionizing radiations fields of medical research in which including Positron Emission Tomography (PET). images are a medium to get transfer and A simplified method for evaluatiqg HE,yalues deliver diagnostic knowledge. The for the positron emitters 11C, 'IN, I.)0 and I°F of an image processing is presented. Hvalues for a range of PET studies architecture E system is now well established, have been computed based on biodistribution data consistng of an image capturing block, a available in the scientific literature. Low dose processing unit and an output display. PET studipinclude a bolus administratoni of The availability of fastarithmetic 28 mCi CO 40 (HE = 100 mrem) and 2 mCi8F L dopa processorsallowedusto develop a (HE = 119 mrem). High dose PET studies include versatile system with an hardware 20 mCi C DM0 (HE = 560 mrem) and a lAour (steady convolver capable of doing convolutions state) inhalation of a total 250 mCi C00 (H, = with programmable size and shape 890 mrem). The mean HE value of 13 diverse PET kernels. The system is completed with studies was computed to be 390 mrem. high capacity storage units (hard disk, optical disk) configuring acomplete personal workstation. The system is controlled by an IBM-PC and uses a high resolution CCD camera toacquire images. The firstresults obtained with such system, regarding its performances and limitations, are discussed. The workstation is going to be used in a digital radiographic system whose preliminary results are shown. 1 3 9

1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas 117 Medical Physics Scientific Papers

MP32.1 MP32.2 Radiotherapy Rosirretry Intercanparison in the U.K. Comparison of protocols for photon dosimetry, D.I. Ihreites and J.R. William+, * Medical B.J. Mijnheer*, The Netherlands Cancer Institute, Physics, Western General. Hospital, Edinburgh, Amsterdam, The Netherlands, F.W. WittkSmper and Scotland; + Medical Physics, Belvidere Hospital A.H.L. Aaibers, National Institute of Public Glasgow, Scotland. Health and Environmental Hygiene, Bilthoven, The Netherlands, and L.M. Chin, Harvard Medical School, The IPSI (PA) Radiotherapy Topic Group has initiated a study, Boston, USA involving all UK radiotherapy centres, aiming to establish the degree of variation between measured and stated dose for all Recently a number of national and international MV photon beans under reference conditions and to comparemeas- organizations revised their protocols formega- ured doses in three-field treatments to those calculated by the voltage x-ray and 7-ray beam dosimetry,to incor- local planning procedures. A phantom has been designed andcon- porate new concepts and refined data. The diffe- structed, having relatively simple gecaetry, but enabling a rences in recommendations in these reports with range of factors to be tested. Doses measured in this, using respect to type of reference instrument, calibra- an ionisation carter, are compared to those stated or calcul- tion, measurement procedure and referencedata for ated fran clinical data in each centre. About half the centres physical parameters, will be reviewed.Despite have now been included. these differences, it will be shown that the A preliminary analysis of the earlier results shows a mean disagreement in absorbed dose determinations isvery crepency under reference conditions of 0.1% (standard devia- good. In addition, the values for absorbeddose tion 1.7%, maxima discrepancy 4%). Consistency with field derived according to one of thesenew protocols ql-Me variation was generally within 1% and always within 2%. shc;ed agreement within 1.0% with valuesobtained For planned three-field dose distributions, the average dose with a graphite calorimeter. ratio (rteasured to calculated) was 1.020 with no inhcmogen- The absorbed dose to a medium irradiatedby ortho- eity present (s.d. 1.2%, range 1.007 to 1.033). With a lung- voltage x-ray beams is usually determinedaccor- equivalent insert present, these results changed only margin- ding to the formalism recommended bythe ICRU. A-I:tendedupdated analysis is presented. As sufficient data Experiments using different types of ionization is obtained analysis will consider the effect of different Var chambers and analysed according to the ICRU and iables, each as bean energy, treatment planning system factors, a new formalism, will be described. Results will etc. be compared with those obtained recently by Sponsors of the project include NE Technology, Negretti Aviat- other groups using different calibrationor ion, ItE Medical Systems. measurement procedures.

MP32.3 MP32.4 EET:MATION OF OUTPUT OF 12 MeV LINEAR Dosimetry of the Photon Beam of ACCELERATOR (MEVATRON-74) FROM WIT-PUT the CGR Orion 5MV Linear FACTORS,Surinder K.KaulI Wajahat H. Accelerator, C.W. Coffey*, R.E. Andrabi, RakeshK. Kaul George+, B. Stokes, B. Berner and M. Sanders, Radiation Department of Medical Physics and Medicine Department, University Dio-engineering,S.K.Institute of Med. of Kentucky Medical Center, Sciences,Soura,Srinagar,Kashmir-India. Lexington, Kentucky, 40536. +Radiation Oncology Department, A simple emprical method has been School of Medicine, Indiana suggested for the output measurements University, Indianapolis, IN, of the arbitrary fields from known output 46223. factors for 10 MV Photons from Mevatron- 74 Linear Accelerator Beam characteristics and dosimetry of OF (X,Y)=OF (X,10) OF (10,Y) -(1) t h e C G R Orion 5MV, 100-cm SAD OF (X,Y)=MORX,10) OF (10,Y) and (2) isocentric,linear acceleratorare for rectangular and square fields presented. The relevent physics respectively. Where 'OF' stands for parameters of central axis depthdose, output factor and 'X,Y' denote the tissue-maximum ratios, surface dose, and field axis.The results obtained have relative field size correction factors been compared with experimental are presented for two independent Orion measurements of 14)4V photons. The 5MVaccelerator installations. percentage variation in individual Additionally, the percentage depth dose cases is less than 1% .It has been data is compared with data fromother noticed that the output of fields commercial low-energy acceleratorphoton X x Yi Y x X. A possible explanation beams.Other clinical beam parameters for this discrepancy has been given. investigated and reported include: field flatness, symmetry, isodose curves, and wedge attenuation. assessment of dosimetry for theAn independent-jaw collimator assemblyalso 140is presented.

118 1988 World Congress on Medical Physics and Biomedical Engineering* San Antonio, Texas Medical Physics Scientific Papers

MP32.5 MP32.6 The Characteristics of Small X-ra Beams Explanation for the Degree of Back Scattered rom Linear Acce era ors, J-S Tsai, GK Radiation from Collimator into Monitor Chamber, Svensson and BE Bjtrngard,Joint Center for Hideo Kubo* and Pocheng Cheng, Department of Radiation Therapy, Harvard Medical School, Radiation Oncology, University of Rochester Boston, MA 02115, U.S.A. Cancer Center, 601 Elmwood Avenue, Rochester, New York14642, Department of Radiation We use 6 MV'X-ray beams of diameter > 12.5 mm Oncology, Albany Medical College and VA for stereotaxic intracranial irradiations. Hospital, New Scotland Avenue, Albany, New York Smaller fields are of interest for treatment of 12208 AVMs and functional disorders. This study aims to characterize such beams < 12.5 mm. Most By simulating a flattening filter, beam monitor measurements have been made with radiographic chamber and collimator jaws of a linear film. When the beam size is reduced, the dose at accelerator, we have measured the relative a given depth goes-down and is about 50% at 2 in amount of back scattered radiation (BSR) from of that at 8 mm diameter. The TMR is not the collimator jaws into the monitor chamber. affected, indicating that the dose/kerma ratio BSR was found to exist in photon beams of all changes with beam size but not with depth. three linear accelerators we studied, i.e., a Geometric unsharpness dominates the penumbra. Clinac-6, Cllnac-20 and Therac-20.The Clinac- The influence of electron spread is limited to 2 6 6MV photon beam showed little BSR dependence mlaterally. Results with a beam-spot camera on collimator jaw opening. This was attributed show that an accelerator with a 90° bent beam to the presence of a fixed mirror in a beam has a larger and more elongated beam -spot than path which absorbed most of the BSR.The one with a straight beam. In the latter machine, Clinac-20 18 MV photon beam showed a smaller the source position was found to be more stable BSR dependence than the Therac-20 18 MV photon for different gantry angles. The isocentric beam due primarily to 1) a larger chamber-to- accuracy of our straight-beam linear accelerator is better than ±1 mm, as is the mechanical collimator distance, and 2) a thick exit window accuracy of the stereotaxic positioning aids. It on a monitor chamber. The Therac-20 18 MV is concluded that for 6 MV x-rays fields down to photon beam showed the strongest BSR dependence 8 mm diameter can be utilized. For smaller on collimator opening due to 1) a small fields, penumbra effects gradually contribute to chamber -to- collimator distance, and 2) a thin the spread of dose outside the target. exit window on a monitor chamber.

MP32,7 MP32.8 THE EFFECTSOF SECONDARY BLOCKINGONOUTPUT Dosimetric Evaluation of a D-shaped Field Used for FACTORS FOR 4, 6 AND 24 MV PHOTON BEANS. C. Reft, Treating Retinoblastoma With 4MV X rays. MichaelReese/University of ChicagoCenter for Radiation Therapy, Chicago, IL 60637 Shirish K. Jani*, Edward C. Pennington, and David H. Hussey, University of Iowa Hospitals & Most dose calculationsystems account for the Clinics, Iowa City, Iowa change in machine output as a function of rectangular collimator field sizes through the use Retinoblastoma is a common intraocular child- of output factors. However, most treatment fields hood disease treatable with radiation. The most are irregularly shapeddue to secondary blocks, widely used treatment technique includes a small intended to shield vitalorgans or normal tissue. single D-shaped lateral field using megavoltage x Insomeclinicalsituations, suchas extended rays. The aim is to treat the entire retinal mantle andhalf-field head and neck treatments, surface and spare the lens. In spite of it's more than half of thearea defined by the machine widespread use, only a limited amount of dosimet- collimators maybe blockedby secondary blocks. ric evaluation of this field is r4orted in the The effects of secondary blocking can be literature. Here, we report on the dosimetric significant as shown byMeli fora 6 MV photon aspects of a 2.5 cm x 2.5 cm D-shaped field with 4 beam. This work assesses the effects of secondary MV x rays from a Varian Clinac 4/80. A 12.5 cm blocking as a function of collimator opening, SSD, thick lead block was used to obtain sharp beam blocked field area and arrangnmentfor 4, 6, and edges. Surface dose and buildup regions were 24 MV photon beams. A simple but accurate method measured with a parallel plate ionization chamber. is described for calculatingoutput factors for The output, percent depth dose, beam flatness and blocked fields from a limited number of in-air and the penumbra were obtained. Results showed that in-phantom dose measurements as a function of the beam output was approximately 7% higher than collimator setting. Comparisons of calculations an equivalent open field. Dosg to the temporal withmeasurements for some typicalpatient retina at a depth of 0.5 gm /cm was 97% of the treatment fields will be presented. maximum dose requiring no added beam spoiler. It is concluded that special dosimetric evaluation of 1 J. Meli, Med. Phys. 13:405, 1986. this D-shaped field is needed for clinical use.

141

1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas 119 Medical Physics Scientific Papers

MP32.9 MP32.10 Dosimetry of Photon Beams From A Philips Characteristics of photon beams from SL25 Dual Energy Linear Accelerator, G. Philips SL-25, Jatinder R. Palta*, Mitev*, S. Rajaratnam, M.K. Islam, U.F. K. Ayyangar, Inder Daftari and Rosenow, Albert Einstein College of Me- N. Suntharalingam, Department of dicine and Montefiore Medical Center, Radiation Therapy and Nuclear Medicine, Bronx, NY 10467 Thomas Jefferson University Hospital, Characteristics of 6 MV and 25MV photon beams Philadelphia, PA 19107 from a Philips SL25 Linear Acceleratorare des- Philips SL-25 is cribed and compared to publisheddata. Included a multimodality linear are measurements of percentage depthdose, out- accelerator with new type of beamtransport put factors, crossbeam profiles, andthe result- system and independent collimator jaws. ing isodose distributions.Special attention is A comprehensive dosimetric studyof 6MV and 25Y given to the following SL25capabilities: Dose x-ray beams produced by this machineare presented. distributions for effective photonenergies re- This includes: (1) the beam sulting from mixed 6 MV and 25 MVbeams; effective flatness changes with depth, fieldsize and energy, wedge angles from 0° to 60°,generated with the (2) surface and depth dose motorized 60° wedge, by mixing wedgedand open characteristics of individual beams (3)mixed beams of beams; off-centered fieldsas defined by the indifferent fractions, (4) output factors for dependently movable collimators. The feasibility openand wedged fields, of generating effective photonenergies in be- (5) isodose distributions foropen and wedged fields, tween 6 and 25 MV with the SL25 mixedbeam mode and (6) lead and cerrobend is investigated, and an exampleof matching a transmissionforfieldshaping. The beam 15 MV photon beam from the SiemensMevatron 20 is characteristics at off-axis points foropen and wedged presented. The usefulness of a continuouslyvar- fields are alsodescribed to evaluate its iable versus a discrete set ofeffective wedge impacton the asymmetrically angles is discussed. Characteristics of asymme- placed fields. tric beams are demonstrated.The consequences of these special features fortreatment planning are briefly outlined.

MP32.11 MP32..12 Dosimetric Measurements ofSuperficial X-ray Therapy Machine, F.J.Lin*, C.S.Chen, Analysis of the penumbra components in high energy photon C.P.Tu, N.M.Tsang, Department of Radiation !ems, M. Simonian and J.C. Rosenwald*, Institut Curie, Oncology, Chang Gung Memorial Hospital, 26, rue d'Ulm 75005 Paris, France. 199, Tung Hwa North Road, Taipei, Taiwan, R.O.C. At the edge of higher energy photon beams, several factors contribute to the penumbra : Superficial x-ray treatment facilityhas been used for many decades.It is still used in some hos- - The finite size of the source (geometrical penumbra) pitals for the treatment oflocalized superficial Theleakagethroughthecollimator(transmission leasions such as skincancer. Dose and relative penumbra) dose measurements were made withcylindric chamber - The radiation scattered from the head and normalized at phantom surface in many - The photons scattered from the medium institutes. The measurement results of relative - The finite range of the secondary electrons spread out dose for cylindric chamber and parallel plate laterally. chamber varied about 30% at the surface but agree These components are analysed using on one hand a within 0.5% at 0.5an depth when data were hieoretical approach based on the 3D superposition of point- normalized at 1 cm depth.The surface dose is spread functions, and on the other hand experimental dependent on the chamber wall thickness. Measure- ments of percentage depth dosewere performed with results obtained in conditions where itis possible to parallel plate chamber inwater and polystyrene. quantify each of these components. The difference of the values inwater and poly- It Is shown how the penumbra slope and shape are modified styrene was about 10% from a depth of 1an to 3 as a function of the energy, field size and depth for each an for data normalized at 1 mm depth.We also component. studied the output, relativeoutput factor, field We thank R. Mackie for providing us with software and data flatness, uniformity index withcalibration depth used for theoretical calculations. at 1 cm depth to minimize the uncertaintyof surface dose.

142 120 1988 World Congress on Medical Physics and BiomedicalEngineeringSan Antonio, Texas Medical Physics Scientific Papers

MP32.13 MP32.14 DOSE PERTURBATIONS AT HIGH ATOMIC NUMBER INTERFACES Comparison ofAir Gap Buildup IN PARALLEL OPPOSED MEGAVOLTAGE PHOTON BEAM IPRAJI- Factors for 18 MV 6 MV, and Cobalt 60 ATION, Indra J. Das*, Faiz M. Khan+, and Kerneth R. Photons. D.L. Wilson, Dept. of Radiation Kase, Dept. of Radiation Oncology, Univ. of Mass. Oncology, University of Louisville Brown Med. Center, Worcester, MA 01605, and +Dept. of Cancer Center_k_Louisville., KY 40292 Therapeutic Radiology, Univ. of Minnesota Hospital Minneapolis, MN 55455, USA. A thin window ion chamber (Markus) placed Uniform dose distribution is required in a target inside a polystyrene phantom was used to volume for curative radiation therapy. The dose measure dose buildup factors behind various uniformity depends on photon energy, thickness of air spaces for 18 MV, 6 MV, and Cobalt 60 irradiating volume and presence of inhomogeneity. photon beams. A large dose buildup effect Dose perturbations in the vicinity of high Z inter- is observed for smaller field sizes. This faces cause dose non-uniformity due to the loss of effect becomes more pronounced at the electronic equilibrium. In a single photon beam, higher energies, and may cave clinical there is a dose enhancement on the backscattered signifigance when treating small fields side of the inhomogeneity for all photon energies, with 18 MV photons. The dose buildup effect but on the transmission side dose enhancement occ- is of much less importance at larger field u:s for the energies greater than 10 MV and a dose sizes. There is no signifigant reduction of reduction occurs for the energies lower than 10 MV. dose behind air cavities when very large For parallel opposed beam, the dose enhancement is fields are used such as in whole body or always observed on either side of the inhomogenei- hemibody treatments, even at 18 MV. When ty. Dose perturbation is seen up to _he range of high energy photons are used for whole body secondary electrons generated by the photon beams. treatment, superficial doses can be Dose enhancement was studied for the commonly used effectively increased by adding a thin inhomogeneities such as bone, steel (in the form of plastic sheet well in front of the patient. prosthesis), and lead, as a function of photon ene- Buildup data for various field sizes and rgy, thickness, width and depth of inhomogeneity, air cavity dimensions are presented from and radiation field size. These data will be valu- measurements made on the Therac 20, Therac able in predicting dose at high Z interfaces in 6, and Theratron 780 therapy machines. parallel opposed beams used clinically.

MP32.15 MP32.16 Direct Measurement and Use of Tissue-air On the Necessity of Scatter-air Ratios, Ratios for Megavoltage Radiotherapy Beams, B. Thomadsen*, L. Asp**, T.R. Mackie, S. B. Thomadsen*, T.R. Mackie, S. Shahabi, Shahabi, B. Paliwal, University of Wisconsin, B. Paliwal, University of Wisconsin, Madison, Madison, Wisconsin, United States of America. Wisconsin, United States of America. The usual practice for calculating dose rates in irregularly-shaped radiotherapeutic Uncoupling the z-drive on a commercial photon beams entails the summation of scatter-air scanning water phantom, fastening the z-position ratio (SAR) contributions from partitioned sectors read-out to a float, and fixing an ionization of the field (referred to as a Clarkson procedure) chamber in the water tank, modifies the system to (1,2). This procedure yields an effective SAR directly measure tissue-air ratios (TAR). As per which, when added to the tissue-air ratio (TAR) Suntharalingam, Shalek, and Lanzl (1972), by cal- for a "zero area field" (TARo), gives the TAR for culating the dOse in-air as the dose in-phantom the point of calculation. divided by the TAR (as the ratio of readings, not Shrivastava (3) has pointed out the insensitivity of the resultant TAR doses), in use, inappropriate factors cancel. to the accuracy of the TARo, because the SAR table For high energy beams (e.g., 24 MV) graphite yields more compact build-up caps than lucite, but comes from subtracting the same value of TARo from TARs that is added again in the Clarkson procedure. without the pair-production problems of aluminum. In fact, the Clarkson procedure requires The use of in-air measurements with a build- neither the TARo nor SAR, as shown below: up cap to characterize the intensity of the radi- TAR(irg) = TARo + (1/n) ESAR(radius) ation incident on the phantom simplifies the separation of collimator-dependent from patient- = (1/n) E [TARo + SAR(radius)] = (1/n) ETAR(radius) geometry-dependent variables, as compared to Thus, only the TAR values need be summated. methods often used with tissue-phantom or tissue- **Current affiliation Bowling Green Medical Center, maximum ratio systems. Bowling Green, KY, USA. Sample tables will be presented. 1.Clarkson, J.R. Brit. J. Radiol. 14 (1941) 265. 2.Cunningham, J. M.TallaTtaiTTIT7tinkinson, J. Com uter Programs in Biomedicine 2 (1972) 192. 3. rivastava, P., Summers, R., Samulski, T., Suntharalingam, N., Shalek, R.J., Lanzl, L., Baird, L., Phys. Med. Biol. 24 (1979) 818. AAPM QuarterlyBulletin, June 1972, 61-62 143 1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas 121 Medical Physics Scientific Papers

MP32.17 MP32.18 Dosimetry Measurements in Presence of Dental DosePerturbationsAround Stainless Materials for Therapeutic Radiation Beams, Steel Implants, E.C. Scarbrough*, N.C. Subhash C. Sharma*, Ph.D., Maritza 0. Jenkins**, Ikoro,P.P. Antich, U.T Southwestern Niranjan Bhandare and Allan G. Farman**, MedicalCenterat Dallas, Dept. of Ph.D., Dept. of Radiation Oncology, School of Radiology, 5323Harry Hines Blvd., Medicine and School of Dentistry, University of Dallas, TX75235-9071 Louisville, Louisville, Kentucky 40292. Interfaces are created in tissues and bones by Metals (13 x 16 x 1.5 mm approximately) of Stern implantable stainless steel trays used in the hard type II gold, medium type III gold, Amalgam, surgical management of oropharyngeal cancer; Tirconium, and Lead placed in a polystyrene dose perturbations around such interfaces are phantom, forming a symmetric field of 5.0 x 5.0 measured with ultrathin (nominally 50 microns) an were exposed to different therapeutic beams TL dosimeters. The ultrathin dosimeters are of radiation (Cobalt-60 photons, 6 MeV, 9 MeV, of dimension and sensitivity appropriate for 13 MeV, 17 MeV, and 20 MeV electrons). A thin- precise and accurate measurement of dose window parallel plate small volume ionization perturbationsat thesesites. Resultsare chxdoerwas used to measure transmission and consistent with previous work on Cobalt-60 backscatter radiation. As expected, the results gamma beams. New data on 6 MV photon beams is showed dose enhancement at the entrance side of presented. Implications for therapy and the metals. In addition, we will reportresults techniques to reduce the dose enhancement due of using film and thermolundnescent dosimetry to backscattered electrons are discussed. thin chips to measure dose distribution around the metals. We intend to use this experimental data to evaluate the computer algorithms used in the computerized treatment planning, specifically for head and neck cancer patients containing the dental restorative materials and treated with radiation beams.

MP32.19 MP32.20 Loss of Beam Edge Sharpness for High Z...igy Influence of Aluminum and Bone on Dose Photons in Low Density Materials, Kenneth E. Distributions for Photon Beams Ekstrand*, Bowman Gray School of Medicine, *1 Winston-Salem, NC27103 and Walter E.E. Klein R.K. Rice, B.J. Mijnheer, H. Barnes, Moses H. Cone Memorial Hospital, L.M. Chin, Joint Center for Radiation Greensboro, NC27401. Therapy, Harvard Medical School, Boston, MA 02115 The problem of central axis dose reduction for high energy photon beams traversing normal lung The influence of bone and aluminum simulating bone on dose distributions for tissue is well k.own. An additional problem, 4 and 15 MV photon beams has been studied which may be not as well appreciated, is the loss in detail. The data is being used to of electronic equilibrium on the periphery of high evaluate existing heterogeneity correction energy photon beams, resulting in an increase in algorithms and to guide in the development the penumbra occurring in lung. We have compared of new ones. Various phantom geometries, profiles of x-ray beams ranging inenergy from 4 including bilateral femur geometries were MV to 18 MV. The profiles were measured at 10 cm studied. Measurements were made with a depth in unit density and lung density (.26) phan- thin window parallel-plate chamber, a toms. At the highest energy the 20X to 80X small diode, and film. The vertical (both penumbra width was measured to be 7.5mm In the proximal and distal) and lateral unit density material, whereas in the lung phantom interfaces were studied for various field the width was 18 mm. At 4 MV the situation was sizes and bone sizes. Significant reversed, that is, the penumbra was slightly underdosing was observed for the 4 MV smaller in the lung phantom. photons distal to the bone and aluminum. This is followed by a second buildup Most radiation therapy treatment planning computer region. Overdoses were observed both programs do not take into account this change in proximal and lateral to both bone and beam profile when calculating dose in the lung. aluminum of about 15 %. The bone and This can result in an unanticipated underdosing aluminum have similar influences on photon dose distributions. when high energy x-rays are used. We recommend that lung treatments be carried out atx-ray 1 energies of 6 MV or less. Present address: Cleveland Clinic Found. Cleveland, Ohio 44106

122 1988 World Congress on Medical Physics144 and Biomedical EngineeringSan Antonio, Texas Medical Physics Scientific Papers

MF32.21 MP32.22 INFLUENCE OF HIP PROSTHESIS IN HIGH Modification of the Dose in Radiation ENERGY PHOTON DOSE DISTRIBUTIONS Fields Due to the Presence of a Mandibular Prosthetic Implant,K.P.Doppke*,S.Sailer, C.H. Sibata, J. Mastandrea, H.C. Mote.** and Massachhusetts General Rospital,M2,USA P.D. Higgins. Cleveland Clinic Foundation. 9500 Euclid Ave, Cloveland, OH, 44106,USA. Patients with head and neck cancers are often treated after major reconstructive surgery and they occasionally have had replacement of their Radiotherapy treatment of patients having mandible with a stainless steel prosthesis. This hip prostheses are a common problem facing prosthesis is 8mm wide,3mm thick and with dosimetrists and physicists when a holes spaced every 8mm center to center. The treatment plan is being done which resulting modification of the dose distribution requires irradiation of the pelvic area. was evaluated for low and medium energy x-rays Normally this problem is avoided by and 6,9, and 12 MeV electron beams normally used changing the 4-field"box" technique to in the treatment of these patients. Dosimetry one which does not irradiate the measurementswere performed usingfilm to prosthesis. To quantify the perturbation evaluate the dose distribution and a parallel of these devices, studies of attenuation plate chamber was used to study thelocal and transmission were made on 6 and 18 MV changes due to the backscattered radiation.The photon beams using two different hip results demonstrate that the dose beyond the prosthesis compositions. These studies stainless steel is significantly modified in the have shown that an attenuation of as much electron fields, with a maximum increase in dose as 40 % can be found in a single beam of 17% due to backscattered electrons. There profile under the prosthesis. In are areas in depth where the dose increase was principle, the perturbationcaused by 24% higher then the open field for the 12 MeV these devices couldbe calculatedand beam and somewhat less for the lower energy accounted for by using compensators. We beams. A dose reduction of greater than 50% is have attempted to study the capability of seen behind the stainless steel. The a dose planningsystem topredictthe modification of dose in depth for the photon attenuation andtransmission of these beams was less, but a dose increase of 26% wzs devices as compared with measurements. measured due to backscattered radiati:*n in the cobalt-60 beam.

MP32.23 MP32.24 Radiation Dose Non-Uniformity Caused by DOSE PERTURBATIONS ASSOCIATED WITH Permanent Vascular-Access Injection RETRACTEDMISSING TISSUECOMPENSATORS, Ports. F.R. Bayne*,H.W. Merrick, N. D. Robinson*, J.W. Scrimger, Department STOW, R.R. Dobelbower, Jr. Medical of Medical Physics, Cross Cancer College of Ohio, Toledo, Ohio. Institute, Edmonton, Alberta, Canada

Injection ports are frequently used to maintain permanent vascular access in patients with malig- Wehavestudiedexperimentallythedegree to nant disease. The injection ports allow blood which retracted missing tissue compensators in sampling as well as infusion or injectionof the presence of irregular surfaces are able to chemotherapeutic agents directly into the circula- restore the dose distribution obtained with a tion. The access catheters are usually placed in normally irradiatedregular surface. Initial the subclavian vein with the injection port being results were obtained for a flat slab geometry', implanted in the infraclavicular area. These andrecentinvestigations have concentrated on injection ports are entirely self-contained under- conical geometries which are amenable to neath the skin and may have been placed in an area mathematical analysis. The experimental which subsequently requires radiation treatment. arrangement involves a hollow cone immersed in A comprehensive study of the perturbation effects water, with the associated water deficiency of the injection ports when placed in the path of compensated by a geometrically reduced retracted the beam was carried out. All dose measurements unit density cone. were performed either inwater orpolystyrene Results of measurements made in60Co, 6MV and using stainless steel and titanium injection ports 15MV photonbeams will bepresented, andthe manufactured by Davol. The radiation beams were 6 extent to which a simple analytical approach is MV and 10 MV X-rays and 6, 9, 12, 15, and 18 MeV able to explain the data will be discussed. electrons.The data indicate that the presence of 'Megavoltage photon beam dose reduction with injection ports significantly alters the radiation retracted tissue compensators. DM Robinson, JW dose end dose uniformity. For example, the dose Scrimger. Phys Med Biol1987; Vol. 32, No. 8: underlying a stainless steel port is reduced by 1031-1037. 47% when 18 MeV electrons are used. This paper presents the dosimetric data, discusses the clinical significance of the results, and provides recommendations for design modification of he ports. 145 1988 World Congress on Medical Physics and Biomedical Engine 2ringSan Antonio, Texas 123 Medical Physics Scientific Papers

MP32.25 MP32.26 Comparative Analysis of Dosimetry Protocols for Dosimetry Of Open And Shaped Electron Beans - sore Physical and Clinical Electron Beams From A Philips SL25Accelerator, Implications, D.I. "[hakes* and A.E. Hahn+, S. Rajaratnam*, G. Mitev, * Medical Physics, Western GeneralHospital, M.K. Islam, and U.F. Rosenow, Albert EinsteinCol- Edinburgh, Scotland; + Medical Physics, The lege of Medicine and Royal Marsden Hospital, Sutton, England. Montefiore Medical Center, Bronx, NY 10467

Various national and international dosinetry protocols have Output factors, depth dose appeared between 1980 and 87. Different approaches have been curves and isodose distributions for electronbeams from a Philips SL25 followed and different values used for variousparameters, due will be presented. A systematic investigationwas to advancing knowledge during this period. Thus international also carried out on the effect consistency has been lost. National consistency is also reduc- of using cutouts. Centrally and asymmetricallyplaced cutouts of ed, as inevitably all centres do not adoptnew recamendations differing field sizes, uniformly at the sane tine. were used on each of the cones. Attempts were made to determine The various electron protocols are canparel to each other and a functional relationship betweenoutput or dosedis- to older recamendatices and the inconsistencies discussed. tribution and cutout/cone combination. The pos- The clinical implications of the changes in dose (up to..5%) sibility of using equivalent fran older to more recent schemes are considered. Inparticu- fields for electrons was also investigated. In addition, a strong de- lar, the influence of using different phantom materialsunder pendence of virtual older and recent protocols is detailed. The non-recognition of source position on beamenergy, and collimator sizewas observed. The implica- fluence ratios in pre-1980 protocols and their generalrecomm- tions of these results fortreatment planning will endation of polystyrene or water as interchangeable phantoms also be discussed. has significant implications for consistency then, butalso for the present situation on changing tonewer protocols. Finally, an analysis of the revised HPA Code of electron dosinetry is presented. An expression is derived for Ce as used in that Code and this is evaluated for the NE 2571 cylind 1- cal chamber, using the most recent physical data. ICRU 37 stopping pokers are used consistently throughout. HPA Ce increase too slowly with depth at all energies, but there is good agreement with the present work at depths close to Dim.

MR32. 27 MP32.28 Depth Dose Flattening of Low Energy (=10 The Disance Dependence of Electron Dose from MeV) Electron Seams, H. Alasti*, D.M. a Varian Clinac 2500 Linear Accelerator, P.H. Galbraith, Princess Margaret Hospital, Huang* and R. Hagg, Division of Radiation Toronto, Ontario, Canada Oncology, Scripps Clinic & Research Foundation, La Jolla, CA 92037, U.S.A. Low energy electron beams commonly have large dose gradients on both sides ofd....., limiting the Dose variation with distancewas measured at dmax depth of tumor which can safely be contained in a polystyrene phantom witha parallel-plate ion within a ±5% dose specification. Clinically it is chamber furnished with appropriate buildupfor six preferable to minimize the dosegradient while electron energies of 6-22 MeV froma Varian Clinac maintainins the sharp gradient beyond 2500 linear accelerator. Relative dose normalized Several methLds are dv,cribed which satisfactorily to 100 cm source (assumed to be at 100cm from the flatten the depth dose in this manner for cone and isocenter) to detector distance (SDD)was deter- trimmercollimated beams at small and large field mined from 10C 130 cm SDD for five cone sizes sizes. ranging from 6xo to 25x25 cm2. Significantdevia- tion from inverse-square divergence with respect to the assumed source positionwas noted for most of the electron energies andcone sizes. Relative dose versus SDD was plotted in log-logscale to calculate the slope of thecurve. The slope varied fr:m 3.17 for 12 MeV electronenergy and 6x6 cm2 cone to 1.9 for 22 MeV electron ?.nergy and 20x20 cm2 cone, resulting ina virtual source at 60.5 and 106 cm to the isocenter, respectively.For the Caine cone, the variation of the slopewas relatively insensitive to different electronenergies except for 6 MeV. However, the slope changeddras- tically with cone size for each electron energy. These indicate that dose measurement ina phantom should be made for treatment distancesother than calibration distance for electron irradiation. 146

124 1988 World Congress on Medical Physics and BiomedicalEngineeringSan Antonio, Texas Medical Physics Scientific Papers

MP32. 29 MP33.1 Fluence Ratios for Electron Beans for Various Uncertainty in Point A Definition: A Plastics canpared to %ter, D.I. Theites, Comparison of Different Methods, Medical Physics, Western General Hospital, T. Ganeshl N.R. Datta, R.C. Joshi, Edinburgh, Scotland. S.Narayanan, Rajendrakumar and G.K.Rath, IRCH, AIIMS, New Delhi, INDIA. There is a gradmil move away fran plastic phantoms for electron beam dosimetry, due to charge storage effects and uncertainties Despite the fact that point A has been exten- in fluence ratios to water. However most protocols still reco- sively used for dosimetry in the intracavitary mend plastics and many centres world-wide still employ them. irradiation of carcinoma uterine cervix, there Previous work on fluence ratios, canparing clear polystyrene is still considerable uncertainty in its exact and water, has been extended in energy. This shows a linear var- definition. According to the Manchester sys- iation inMp fran -4.025 at effective energy of 3 bbV to tem, in which point A nomenclature wasfirst -.1.035 at 15 %V, using a graphite-walled Farrier chamber. Thus introduced, it was defined as a point situated differences are observed of up to 1% to the values used in the 2cm laterally from the uterine canal and2cm MEM dosinetry protocol, but throe is reasonable agreerP.nt with up from the level of the mucous membraneof values extracted fran the HPA Code. Results are presented and the lateral fornices. In clinical practice, canpered to other published values. Discrepancies reflect diff- however, point A is often taken 2cm up from erences in polystyrene samples, in the characteristics of the the flangeof the intrauteri catheter or chambers used and in angular and energy spread of the electron from the last active source in it and 2cm beams. Sao of these factors are investigated. lateral from the central canal. Since point A In addition fluence ratios have been measured for other plas- plays a crucial role in the treatment of cer- tics, including sale camercial solid -water equivalent phantom vix carcinoma, a study has been carried out on materials. For RNA, values are unity for effective energies patients treated with Selectron to assess how above 10 be, falling to-.0.99 at 3 ItV (MA:ft). these different practical definitions produced water (older formulation) shows differences to actual water of variations in point A dose rate, ICRU refere- around half those observed for polystyrene and is thus a better nce volume and doses at other refernce points. water substitute than polystyrene for electrons. More important- Computerised dosimetry has be done for dif- ly, it shows negligible charge storage effects. Results for oth- ferent point A definitions and the variations er materials are presented and discussed. have been analyzed.

MP33.2 MP33.3 Choice of LDR vs HDR for Ca-cervix Dose-volume Study for Combined External Treatment:Radiobiological considerations. and Selectron Intracavitary Irradiation, U. Madhvanath, DRP, Bhabha Atomic Research T. Ganeshl N.R. Datta, R.C. Joshi, S.Narayanan, Rajendrakumar and G.K.Rath, Centre, Bombay-400 085, India. IRCH, AIIMS, New Delhi, INDIA. Shouldered survival curve of HDR (1.5-2.5 Gy/min.) changes to exponential at LDR of 1.5-2.3Gy/h. This implies change in the repair capacity of irr- The ICRU brought out its report # 38 on 'Dose adiated cells whether that of cervix carcinoma andVolume Specification for Reporting Intr- (tumour) or normal tissue (of rectum, bladder or acavitary Therapy in Gynecology' in 1985. colon). Both these dose rates either at Point 'A' Since then its theoretical and practical adv- or at critical normal tissues have relevance beca- antages have been largely acknowledged by many use of the differences in the Tc between carcinoma radiotherapycenters around the world. Based and normal tissue cells. Because of shorter Tc of on these ICRU guidelines, a dose-volumestudy Carcinoma cells, 15h as against 45h of colon cells, has beenperformed on patients treated by a differential cancer cell killing advantage is combined external (15MV X-rays) and intracavi- obtained in the case of LDR technique. Further, tary radiation (Selectron). The total refere- resistant S phase is overcome as irradiation times nce air kerma (TRAK), reference volume, doses are always more than 15h and perhaps the anoxic at referencepoints in organs at risk and cells as well. In the HDR technique as both the those related to bony structures for a typical tumour and normal critical structures are in iden- Selectronapplicationhave been calculated. tical high fields, no repair differential is avai- The relationship between TRAK and reverence lable and as irradiation times are only a few minu- volume for different doses of external radiates per session, multiple treatments (atleast 5 or tion hasalso been studied. The reference 6) are required that too at weekly intervals to volumehas been calculated by constructing overcome the stage resistance. Thus, radiobiolo- voxels around the applicator instead of taking gical considerations advocate LDR with better cure the product of its three dimensions. This ratewhereas administrative patient load consi- dose-volume study was done using the Nucletron derations led oncology centres to go for HDR Treatment Planning System. technique. 147

1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas 125 Medical Physics Scientific Papers

MP33.4 MP33.5 Use of Scout Views for Source Localization in CT Compatible Gynecological Implants Version of the Fletcher E. Yorke*, George System for Washington University, Washington, D.C. Intracavitary Treatment, K. 20037& C. J. Weeks* and Clifton Ling, University of California, S. L. Schoeppels, San University of Francisco, San Francisco, CA 94143 Michigan Medical. Center, Ann Arbor, HI 48109.

When CT scans are used for evaluation ofdose Theuseof CTscans to delineate the anatomic distributions in gynecological implants thesource relationships of intracavitary positions are determined from orthogonal films. cervix applicators, tumor, rectum and It would be desirable to use AP and lateral bladder has been scout limited by streak artiftcts produced views taken concurrently with the scans for by the metal source applicator and/or tungsten shields. localization. To eliminate The distortion which is known to this artifact, a new variation exist in the scout views has prevented such of the Fletcher use. systemfor treatment Much of the distortion is caused by the marked of cervix cancer has been developed. The applicator beam divergence due to the short FAD of CT has been constructed of acrylic. The tandem and colpolstats have scanners. It can be corrected by geometrical been designed along the lines of the Fletcher-Suit- transformations provided that the patient position ;:elcos applicator used previously relative to the scanner axis is not changed in our department. Theprincipal between scout views and the axis position is design modification has been to afterloadboth sources and tungsten marked on both scouts. A phantom with wires shields. The choice of shield approximating gynecological implant sources was positon is flexible and is chosen used to compare endpoint localization to conform to Fletcher's accuracy originalapplicator. This from orthogonal scout views and from simulator lightweight system enables the radioactive films. Simulator films are most accurate (end- source positions to be clearly defined in relationship to the tumor as points localized with standard deviation of well as critical structure volumes, such +0.7 mm). Scout views corrected by simple magni- as rectum and bladder, through the use of fication are least accurate (+3.1 mm with some a routine CT scanafter placement. This work presents the endpoints in error by more than 5 mm). Correction rationale, designand procedural aspects of the for beam divergence improves localizationto new systemas well as examples of simulator and +1.3 mm with all endpoints localized to better CT imaging of the system after placement. than 4 mm. Volume dose calculations, shielding design, and future development work will be discussed.

MP33.6 MP33.7 Dosimetric Evaluation of a New Gynecol- Recovery of Arbitrary 3-D Radiograph Geometry in ogical Brachytherapy Device, E.D. Brachytherapy from a Minimal Set of Reference Slessinger*, P.W. Grigsby and C.A. Markers, Bruce A. Lulu, Radiation Oncology, Perez, Washington University School of University of Arizona, Tucsor, AZ 85724, USA. Nk-dicine, St. Louis, MD, USA. in radiation therapy, verification of A new device brachytherapy plans ideally involves the forgynecologicalbrachytherapy calculation of true source positions relative to application has recently beendeveloped.It is some fiducial landmarks. This method permits the called MIRALVA, an acronym for the Mallinckrodt use of arbitrary unknown film geometry, as might Institute of Radiology AfterloadableVaginal Ap- occur in surgery with a portable X-ray unit, as plicator. The development of MIRALVAwas prompted well as display of source positions (and by the need to treat the wide range of vaginal therefore isodose lines) relative to patient and cervical cancer target volumes withan opti- anatomy seen on CT. This technique is also used mal dose distribution, in a singleapplication, to locate artereovenous malformations relative utilizing either manual or remoteafterloading to a stereotaxic frame. technique. The plexiglass device is comprisedof With two radiographs, one can reconstruct a cylindricalportionthat enlarges at the source positions in 3-0 by finding the vaginal apex. It contains threechannelsfor the intersection of the two lines passing through placement of Cesium-137 tubesor spheres.The the same arbitrary source. We consider the central channel allows for a tandem thatextends situation where the fiducial markers are either to the vaginal apexor higheras an coplanar but not collinear, and that the 3-D intrauterine device. The remaining twn channels coordinates of these markers are known insome allow source placement laterally in the rightand arbitrary coordinate system. Absolutelyno other left aspects of the vaginal aIlex. The outer knowledge of the geometry is given. We then dimensions can be varka to p; widean optimal derive the coordinates of a line passing through fit while the surface contours are designed to the target and the brachytherapysource, as well provide a relatively uniform dose distributionto as the target coordinates relative to the the vaginal mucosa.Thermoluminescent and film previously mentioned arbitrary coordinate dosimetry studies in a special solidpolystyrene system. We show that for coplanar, non-collinear phantom provided the necessary data toevaluate points, four points are necessary and sufficient the resultant dose distributions. 148 for this reconstruction.

126 1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas Medical Physics Scientific Papers

MP33.8 MP33.9 Use of a Pass-throughIonization Dosimetry of Large Interstitial Templates Chamber for Brachytherapy Quality J. M. Paul*, R. F. Koch, P.C. Philip and Assurance. H.A. Earle*, K.H. Hutchins F. R. Khan, NWCH, Arlington Heights, Il, and R.K. Ten Haken, Universityof 60005 Michigan, Ann Arbor, HI 48109. Although the Paris system has been used in inter- An integral part of hrachytherapy quality stitial brachytherapy using single or two plane assurance programs should be confirmation of implants, no attempts have been made so far to use source strengths stated by manufacturers prior to this system for large multiplanar implants, as use in patients, as pointed out in recent reports encountered in the Syed-Neblett type of inter- by AAPHTask Groups 24 and 32. Even if pending stitial gynecological templates. We have national regulations make these checks manditory. investigated this possibility in order to quantify they willonly be consistent with existing ALAAA the dose delivered with such templates, involving programs if they canbe accomplished quickly, many hundreds of seeds with as many as 44 needles, easily andwith minimumexposure to staff. To instead of arbitrarily quantifying the dose that end, wehavedeveloped an in-house assa;# delivered on the basis of isodose curves generated program based around a pass-through, re-entrant with treatment planning computers. Based on our well air ionization chamberenclosed in a lead analysis, using three-plane Paris implants, we brick cave. Different inserts have been designed find that the overall quality of these implants and fabricated for use with individual 125 -I could be improved in most cases while using only seeds, strings of 192-1r seeds, 137-CS tubes and one-half or less as many needles; which 32-P suspensions. For example, a thin glass tube considerably simplifies implant procedures suspended bylow mass supports along the central and subsequent localization and dosimetry. axis of the chamber is used for assaying strings These results are presented in this paper using of 192-1rseeds. With this arrangement, strings integral volume vs. dose rate curves, and are threadedfrom their lead shipping container, tabulated values such as: integral dose, through the calibrator, and into another lead variation of the basal dose rate, volume contains for transfer to the patient. Using the irradiated to dose rate greater than twice the respons function for a single seed vs. position, stated rate and an ob)ective parameter called appropriate conversionfactors have been derived uniformity index. for any arbitrary number of seeds per string.

MP33.10 HP33.11 Multiplanar-CT Interactive Treatment 1-125 Dose Distributions in Various Planning and Stereotactic 1-125 Tissues, D. Huang*, M. Schell, K. Weaver, BrachytherapyZamorano L,* Dujovny M, C. Ling. Department of Radiation Yakar D, Malik G, Sheehan M.-trii/Ford Hosp. Oncology, University of California, San Computer close calculations for interstitial im- Francisco, CA 94143. plants of radioactive seeds within intracranial tumors requires a knowledge of the spatial coor- The low energyphotons of 1-125 interact dinates of each see. A method is presented by with tissues primarily via the which Multiplanar-CT under stereotactic conditions photoelectric effect whichis strongly Z is used for computer dose calculations. Multi- dependent. Thus, the dose distributions planar-CT is used to determine size, shape and of I-125 sources inmedia of different main axis of tumors; this information is integrat- effective Z may not be the same. We used ed with a computerized planning system.The aim is LiF andCaF2 thermoluminescentrods to to achieve a distribution of a selected isodose measure dose distribution intissue curve adjusted to the defined tumoral borders. equivalentmaterials correspondingto After preplanning selection, CT scan with the fix- muscle, adipose, and bone. The measured ed stereotactic frame is performed. Multiplanar data wereparameterizedas polynomikls CT images and scoutviews (AP and lateral) are used and compared toMonte Carlo calculations to plot the trajectory of selected planes of treat- andtothe findings of Dale*. The calculated ment and coordinates are calculated.Catheters are and measured results, for placed stereotactically loaded with codified distances between 0.5 and 7 cm., were in good agreement. dummy seeds. Postoperative CT with stereotactic These data provide basic data fordose frame is performed; Multiplanar and scoutviews distribution calculations when 1-125 seeds are implanted in ornear images are used to determine the location of dummy inhomogeneous seeds and the relationship of catheters to tumor tissues, e.g. breastor bone. borders. Measurements are compared with preplanned and minor discrepancies can be corrected by altering distance, number, activity of seeds or dose/ * Medical Physics 10(2) 1983, p. 178-183, and 13(6) 1088, p. 983-984. rate. Simulation with conventional orthogonal film technique and final isodose distribution curves in all 3 planes are generated and superimposed. 449 1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas 127 Medical Physics Scientific Papers

MP33.12 MP33.13 TLD Dosimetry for /-125 Eye Plaque Dosimetry Considerations for 1-125 Seeds S.T. Chiu-Tsao*, L.L. Anderson, L. Stabile Implanted on a Curved Surface, S.C. Prasad*, Memorial Hospital, New York, NY 10021 D.A. Hassan°, P.I. Fear, G.A. King and J.A. Winfield, SUNY, Health Science Center, Extensive measurements using LiF TLDs were made Syracuse, N.Y., U.S.A. for COM* 20 mm eye plaque both containing a single 6711 seed at the center and containing a 1-125 seeds are commonly used as permanent inter- complete loading of twenty-one seeds. TLD cubes stitial implants for various malignant tumors. of1 ma on an edge were used in a solid water eye For volume and planar implants, currently available phantom. Measurements were performed at 2 mm or tables can be used to determine the number of closer intervals up to 14 mm from the "sclera" seeds needed, the activity of the seeds and the throughout a central plane within the eye.The seed separation. In our institution, a new im- single-seed measurements were made with and plant technique is being used to treat highly without the gold backing. The results for single malignant brain tumors using I-125 seeds. Tumors seed configuration were compared with the dose of the brain, with a histology of high grade astro- distrihxtion in a homogeneous solid water phantom. cytoma or gnome, are first surgically removed. We found that the presence of silastic (Zeffa11.2) 1-125 seeds are next implanted, as a permanent between the seed and the measurement points implant, along the wall of the cavity formed by introduced dose reduction in the eye by about 10% the removal of the tumor. No simple nomogram (central axis) to 15% (off-axis).Adding the gold exists to preplan an implant in which the seeds are backing over the silastic did not appear to modify distributed on the surface of a hollow cavity. We the dose significantly. These data with silastic have developed a computational model to determine and gold agree well with our Monte Carlo results. the number of seeds needed and the activity of The TLD results for the multi-seed configuration seeds once the size of the cavity and the dose are were found to agree with calculations which treats specified. The model is based on spherical the seeds as isotropic point sources with dose as geometry and has been implemented on an LS/-11/23 a function of distance derived from the transverse- computer. The details of our technique and results axis data in homogeneous phantom and incorporating dn. be presented. a correction factor for filtration by silastic. +Collaborative Ocular Melanoma Study Work supported in part by NCI contract NO1CM57776

MP33.14 MP33.15 Substitution of Pd-103 for I-125 in Angular Dependence of the Dose Rate for the Treatment of Choroidal Melanovis, High Activity Ir-192 Sources Used in E.C. Scarbrough*, N.C. Ikoro, P.P. Remote Afterloaders, G. A. Ezzell* Ant: ±ch, U.T. Southwestern Medical Gershenson Radiation Oncology Center, Cewcar at Dallas, Dept. of Radiology, Harper-Grace Hospitals/Wayne State 5323 Harry Hines Blvd., Dallas, TX University, Detroit, MI 48201 75235-9071

Pd-103 has recently become commercially High activity (typically 370 MBq) Ir-192sources available for use in bracytherapy as a are frequently used in high dose rate remote potential substitute for 1-125. A computer afterloading applications. These sources are programhas been written to compare the usually calibrated by measuring the exposureor treatment of choroidal melanomas with air kerma at a reference distance on the perpen- eyeplaques containing each of the two dicular bisector of the source axis. Patient isotopes. The program takes into account the dose calculations usually assume an isotropic increased tissue absorption of Pd-103 gammas, point source model. The dose rate anisotropy anisotropies of the radiation field, reduced was measured in water for distances represen- coherent scatter due to the presence of the tative of patient treatments. DI es at 1 cm gold eyeplaque, and the cut-off of 'he from an endobronchial treatment catheter were treatment volume due to the goldplaqt . also measured using TLD and a 0.03 cc ion Results of this study indicate that the chamber. The results show that the simple point anisotropies and increased tissue absorption source model underestimates the dose by 2 -6X do not place serious limitations on the use of depending on the active length of thesource Pd-103 for this purpose. In addition, the train, and that this underdosage can be increased tissue absorption reduces the explained by applying the anisotropy uAta. radiological hazard to personnel in the Appropriate parameters of a line source model vicinity of a patient with such an implant. are presented.

128 1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas Medical Physics Scientific Papers

MP33.16 MP33-17 INTRABRONCHIAL IRRADIATION WITH MODERATE Dose Correction Factors for Ir-192 in the DOSE RATE USING A SIMPLE ECONOMICAL AFTER Presence of Inhomogeneities, S.N. Rustgi*, LOADING DEVICE TO HANDLE THE HIGH ACTIVITY G. Popescu and J. Rodgers, Department of Ir-192 SEEDS, WILLIAM C. KAN*, STEVEN H. Radiation Medicine, Georgetown University STOKES, RADIATION ONCOLOGY DEPARTMENT, Hospital, Washington, D.C. 20007 SOUTHEAST ALABAMA MEDICAL CENTER, DOTHAN, ALABAMA 36302-698' The influence of inhomogeneities on dose in water around a 10 Ci Ir-192 source has been assessed. The inhomogeneities studied include An inexpensive manual after-loading device aluminum, cork, and styrofoam which simulate used for intrabronchial irracLation with bone, lung, and a low density medium, respec- high activity Ir-192 seeds vat, constructed. tively. The measurements were made possible by Twenty patients have been treated with one the high activity of the Ir-192 source and were to four sessions receiving 600-700 cGy at performed with a 0.1cc PTW ionization chamber 1 centimeter radius per treatment. Ninety which has good spatial resolution. Since dose percent of patients experienced relief of in the vicinity of small sources has steep symptoms associated with an obstructing gradients, positional accuracy is extremely intrabronchial carcinoma. Review of the important. The measurements were made in a film and ring badge reports on both the Therados RFA3 phantom which has a positional Radiation Therapist and the Radiation accuracy of + .5 mm and a positional reproduci- Physicist performing these procedures did bility of + .2 mm. Dose correction factors, not show an excessive radiation exposure. which are ratios of ionization in the presence Our after loading device and method of of inhomogeneity to ionization at the same point treatment has been proven to be simple, in a homogeneous water phantom, were measured economical and safe for the treatment of for distances up to 20 cm from the inhomogeneity. intrabronchial carcinoma. Construction of A 0.7 cm aluminum slab reduces the dose beyond our device and method of treatment will inhomogeneity by 5% at distances from 3 to 20 cm. be discussed. For a 2.5 cm thick cork inhomogeneity the correction factor increases from 1.04 at 1 cm to 1.12 at 20 cm. The dose correction factor for a 6 cm slab of styrofoam increases from 1.12 at 1 cm to 1.39 at 19 cm.

MP33.18 MP33.19 Evaluation of activity of Cs-137 seed Cs-137 Dosimetry Table for Asymmetric sources used in a remote after loading Source, R. Waggenert,J. Lange, J. Feld- system (Selectron LDR) meir, P. Eagan and S. Martin, The Univ. V.K. Sharma*, V. Poopathi, L. Aggarwal of Texas Hlth Sci. Cent., San Ant., USA MDOCT&RF, LUDHIANA ( INDIA ) A common Cs-137 brachytherapy source has In intracavitary treatments the radiation a 2 cm length with 1.38 cm activelength, dose delivered to tumour & its surrounding AL. The AL is not symmetric with respect normal tissues is proportional to the to the source center because of an eyelet activity and specific gamma ray constant in one end of the source. Current dose of the sources used. Hence it is tables assume symmetric source loading. important to ensure that there exists a A computer program was written to calcu- good uniformity in the activity of the late a asymmetric distribution using the sources when used in large number. The manufacturers source specifications. Cor- mean apparent activity of all the sources rections were made for attenuation and should be determined before actual use. obliquity through the source material, walls and ends. Table values have been A perspex calibration phantom was confirmed by ferrous sulfate measurements fabricated at our place to hold the using small volumes.Table values agree ionisation chamber and the plastic with published values at points greater catheters with sources and dummy pellets than a few cm from the source center. parallel to long axis of the chamber in a A knowledge of the asymmetric dose dis- cylinderical geometery at 120 °. The tribution should be useful for some ther- measurements were taken in water and the py situations. values were used to derive mean apparent activity and correction factors for attenuation in water and steel. This procedure was carried out as a part of the quality assurance programme in LDR Brachytherapy Dosimetery.

15 1 1988 World Congress on Medical Physics and Biomedical Engineering * San Antonio, Texas 129 Medical Physics Scientific Papers

MP33.20 MP?, t. Dosimetric Evaluation of Endobronchial Iridium-192 Implants Using Computerized Routine Calibration ofBrachytherapy Sources Tomography. S. Shahabi: Ph.D., M. P. with a Well-Type Dose Calibratorusing Stand- Mehta, M.D. , B. R. Thomadsen, M.S., T.R. ard Calibrated Sources. AzamNiroomand-Rad*, Mackie, Ph.D., University of Wisconsin Michael T. Gillin. Department ofRadiation Medical School, Madison, WI 53792 Oncology, Medical College ofWisconsin. 8700 W. Wisconsin Ave. Milwaukee,WI 53226, U.S.A. Conventional dosimetric evaluation of endobronchial implants has relied upon orthogonal radiographs. Brachytherapy, which refersto short- distanc treatment, requires an accurate knowledge ofthe dose dis- We present herein the application of computerized tribution in the implanted volume.A critical step tomography in ;.etermining the dose distribution in the determination of thedose distribution invol- around endobronchial implants and its advantages ves an accurate calibration ofbrachytherapy source over conventional dosimetry. The availability of strength. Following the guidelinesrecommanded by serial transverse slices enables possible three- AAPM Task Group No. 22(1),a routine calibration procedure for brachytherapy dimensional dosimetry. Better delineation of cri- sources Cs-137,Ir-192, and tical structures, especially the cord and esopha- 1-125 has been developed.Using a single potentio- gus allows better evaluation of the doses delivered meter setting, the calibration methodfor these sou- to these structures.Better definition of the pre- rces is described. Some of theproblems associated cise location of the endobronchial implant and, with the calibration ofsources are discussed. The therefore, a more accurate assessment of the vol- calibration of sourcesare performed with a well-type ume of lung and surrounding mediastinum irradiated dose calibrator usingstandard calibrated sources is also possible. The CT cuts also enable genera- obtained from The UnitedStates(NBS), British(NPL), and French(LMRI)national tion of dose-volume histograms and 3-dimensional laboratories. Selected cha- dose-volume complication probability curves. racteristics of dose calibratorare reported. By considering the NBS calibratedsource as a referance In conclusion, we feel that computerized tomogra- source, the calibration correctionfactors for Ir-192 phy may enable us to supersede conventional radio- and 1-125 are described.The result of intercompari- graphic techniques for dosimetric evaluation, and son of the calibrated sources bynational laborator- allow a more precise definition of doses delivered ies indicates some discrepenciesbetween the calibr- to a volume of interest as well as to critical ated sources. These discrepencieswhich are within structures. the stated uncertaintiesare sometimes larger than expected.(1)AAPM Report No. 13(1984).

MP33.22 MP33.23 Interstitial BrachytherapyPre-planning SET According To OF PROGRAMMES FOR RADIOTHERAPY Paterson-ParkerRules TREATMENT PLANNING. FacilitatedbyMicrocomputer, R.E. PART II - BRACHYTHERAPY TREATMENT. Drzymala,Washington University School M.C. Lizuaint C. Pino and C.Cinos. of lsoLadicine,St.Louis, MD, USA. Hospital de Bellvitge. Barcelona (Spain). A program, written in FORTRANfor use on a IBM compatible microcomputer, promptsthe user for the input parameters necessary to pre -plan Braquitherapy programmes calculate the interstitial implants according toPaterson- distribution of dose produced byimplants of Parker Rules. Allowingdifferential loading of linear radioactive sources of any form. The sources to achieve a homogeneous dose capsule filtration effectand the gamma ray distribution within thetreatment volume,the attenuation in tissue are takeninto account. code accommodates pointsources (seeds) as wall Two orthogonal radiographs are as linear sources. The logical flow of used to the localize the sources. In the program encourages theuser to think about and case of the curved sources, the two proyections of the enter the margin around thegross disease to source are fitted to cubic spline arrive at a tenor volume,to decide on the shape functions fromwhich, the spatial coordinate of of the volume to be implanted,to specify the a set of source points are obtained. In this way spacing between sourcesand toestimate the dimensions of the the curved source is consideredas a set of implant for appropriate small right sources. coverage. Violations of the Paterson-ParkerRules are checked in the flow of the The dose distribution is representedby means program, thereby of isodose curves on providing a learning tool any plane. for new residents or The computer junior physicists. is a PC (Vectra H.P.) with 640 The violation warningsare K-bytes RAM noted but can be overridden Memory, 80286 Intel to permit minor microprocessor deviations from the rules and 80287co-processor, 40 to accommodate unusual M-bytes harddisc circumstances. Additionally, driver. A digitizer as one can allow the input program to choose the spacing and loading device and one plotter andone printer of as ouput. sources for a quick treatmentpre -plan. r N a. -

130 1988 World Congress on Medical Physics andBiomedical EngineeringSan Antonio, Texas Medical Physics Scientific Papers

MP33.24 MP34.1 Detailed Calculation of Radiation Comparison of Variable and Fixed Modulation Energy Spectrum and Spatial Dose Proton Beam Dose Distributions in the Cranium, Distribution in Water for IslCs M. Urie*, Massachusetts General Hospital, and192Ir Point Sources, Y.Takaku*, Boston, MA 02155, USA. K.Takeuchi**, *Fukushima medical College, **Ship Research Institute, The dose to normal tissues in the entrance region of a proton Japan. beam can be reduced if the Bragg peak is spread out (modulated) to a different width at each point in the field to Radiation transport calculation is made precisely encompass a complex target shape, rather than being with a PALLAS discrete ordinates code, modulated to a uniform width determined by the greatest designated to solve a numerical solution breadth of the target Dose distributions calculated with fixed of the Boltzmann transport equation. The and with variable modulation proton beams were compared on calculation of primaryr ray transport is patients with chordomas in the base of skull. 3D treatment carried out for point sources ofl97Cs and planning allowed quantification of the dose differences I91Ir placed at the center of a 10cm rad throughout the head; dose-volume histograms of critical ius water sphere,resulting in:, ray angu normal tissues were calculated and compared. With target lar fluxes at each radial mesh at each doses of 66-72 Gy, variable modulation was found to reduce energy mesh.Thelfray angular fluxes gen the dose to some brain tissue by 15-20 Gy. Correlation of the erate secondary electrons arising from magnitude of the dose reductions with size, shape and location interactions of photoelectric effect, of the tumor will be presented. These data may aid in judging Compton scattering and pair production. whether or not the dose differences are of great enough signifi- Furthermore,these electrons generate cance clinically to justify the greater complexity of variably Bremsstrahlung radiations. All the proc modulated proton beams. esses are traced with the PALLAS codeto obtainirray energy spectra,dose rate distributions,secondary electron energy spectra,Xray energy spectra and Xray dose rate distributions within the water phantom.

MP34.2 HP34.3 Determination of differential w-values for Proton Radiation Therapy. Where will it proton, deuteron, 3-He and 60-Co gamma-ray he in the year 2000? J.M. Sisterson *, beams in several gases, T. Hiraoka*, K. K. N. Johnson, A. M. Koehler, Harvard Kawashima and K. Hoshino, 9-1, Anagawa- Cyclotron Laboratory (HCL), J.E. 4-chome, Chiba-shi, 260 Japan Munzenrider, Massachusetts General Hospital (MGH) and M.S.Z. Rabin, The differential values, w, the mean energy University of Massachusetts, Amherst. necessary to produce an ion pair have been determined for several gases in atmospheric pressure. The gases investigated were air, argon, Bythe endof 1987 morethan8000patients in nitrogen, carbon dioxide, methane, and methane theworld had been treatedwith protonbeams based tissue-equivalent. The tank gases were using facilities originally designed for nuclear used of pure grade. Six different ionization physics. The successesthat havebeen achieved chambers in six different sizes were used in will be illustrated by using the WARM this study. The measurements were carried out experience where 4139 patients had been in the irradiation of 70 MeV proton, 43 MeV treated The long term results fromHCL/MGH deuteron, 99 MeV 3-He ion and 60-Co gamma- .how that proton radiation therapy is indeed a ray beams. In each chamber, collected charge viable modality The superior dose distributions was measured for several applied voltages possiblewith proton therapy makethis the and averaged for both polarities. By dividing treatment ofchoicefor selected sites. the energy loss(stopping power) by the number By year 2000, there may be as many as 20 of ion pairs'produced(ionization measured), the centers in world, relative values of w were presented for the proton therapy the ranging gases and the radiation qualities under from large facilities routinely using protons to investigation. The use of these w-values for treat allsites,tosmallunits dedicated tothe the evaluation of absorbed dose with ionization treatment of a singlesite.Treatment philosophy chambers for proton beams gave good agreement as well as cost and real estate constraints with a TE calorimeter. Therefore, it is esti- dictate the final design of any facility,themix mated that these w-values are accurate to offacilitiesis determined bypatient populations. within about 2 %. Plans, cost estimates(rangingfrom $1-50 million) and possiblemarketingmix will be presented. 153 1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas 131 Medical Physics Scientific Papers

MP34.4 MP34.5 Fast Neutron Dosimetry Using KCCH- Cyclotron and Demonstration of Digital Autoradiography, J. Kleck*,J. Neutron Smathers, L. Myers and B. Stewart, Dosimetry, Hyun-Woo Chung, Kyoung-Hwan University of California, LosAngeles U.S.A. Chul-Koo Cho, and Seong-Yul Yoo,Korea Advanced Energy Res Institute, Seoul, Korea Many conventional techniques usingfilm have been developed for beam characterizationand The purpose of this presentation isto dosimetry in photon and electrontherapy. Film is of less value in eitherquantitative outline the measurement madeat Korea or qualitative dosimetry in fast neutron Cancer Center Hospital, KAERI, and t therapy beams. This is primarily due to the pre- relatively high film sensitivityto small sent the result obstained. amounts of photon and electroncontamination in the neutron beam. These measurements were. Designed to A technique has been developed at the UCLA demonstrate the compliance of the isocent- Neutron Therapy Facility wherebythin (3.18 mm) sheets ofacrylic plastic may be ric fast neutron facility. positioned in a water phantomand activated in the p(46)Be neutron beam. The resulting activation distributions inthe acrylic a) Neutron production and delivery characterize the high energy component of b) Pysical parameters of theneutron the neutron kerma. The principal reactions contributing to the activation are beam. C-12(n,2n)C-11 and 0-16(n,2n)0-15, both with c) Neutron Beam Calibration activation thresholds closeto 20 MeV. The including n activated acrylic plate is usedto expose a ratio and detector design Europium doped Barium Fluorohalideplate, which is subsequently digitizedbya d) Treatment planning Computed Radiography system usinga 15 mWatt HeNe laser. The spatial resolution of the e) Health physics considerations,etc. resulting image is 2.5 1p/mmwith 8 bits of contrast resolution. This researchis supported by NCI contract CM97315. Will be coverd the above topics.

MP34.6 MP34.7 Evaluation of the Gamma Componentin a Neutron Therapy Beam. Dosimetric intercomparisons of p(>60 MeV) /Be M. Yudelev and Neutron Therapy Beams. D.T.L. Jones*, D.T.L.Jones*, National Accelerator Centre, National Accelerator Centre, 7131 Faure, 7131 Faure, South Africa South Africa, S. Vynckier, Universit6 Catholique de Louvain, Belgium and S.W. Neutron beans are always accompanied by gamma Blake, Clatterbridge Hospital, U.K. rays and for therapeutic applications thegamma dose component has to be determinedbecause of In order to compare clinical and other its different biological effect. data from In this work a neutrontherapy centres itis tissue equivalent (TE) ionization chamber, essential for which dosimetry intercomparisonsto is equally sensitive to beundertaken. neutrons and gamma rays The centres involved in the and a present study were Ceiger-MUller (GM) counter, which is the National Accelerator Centre relatively insensitive to (NAC), South neutrons,have been Africa [p(66 MeV)/Be], The Universit6 used in a p(66 MeV)/Be neutron Catholique therapy beam to de Louvain (UCL), Belgium (p(65 MeV)/Be] determine the gamma and the dose. Using the"lead MRC Cyclotron Unit, attenuation Clatterbridge Hospital method" the neutron sensitivity (MRC), U.K. [p(62 MeV)/Be]. (ku) of the GM counter is In the first phase estimated to be 0.06 of the study measurementswere made with a NAC from preliminary data. Frommeasurements made in TE ionization chamber at UCL and MRC and a water phantom for various field sizeswith compared withmeasurementsmade usinglocal both the TE chamber and the GMcounter the gamma 60 equipment. The Co calibrations of the dose component can be evaluated.The gamma dose ionization chamber at UCL and MRC agreed was found to be 3.8% of the total dose within at dose 0.1% of that at NAC. In theUCL neutron beam the maximum for a 10 x 10 cm2 field. At a depth of response of the NACchamber differedby a 10 cm in water gammarays constitute 5.3% of the maximum of 0.2% for all conditions fromthe mean total dose for a 5 x 5 pn2 fieldincreasing to of the responses of all thechambersused. 10.1% for a 30 s 30cm field. For a 10 x 10 2 Measurements made with the NAC ionization cmfield at a depth of 10 cm inwater the gamma chamber in the MRC neutron beamwere between component at 10 cm from the centralaxis it is 0.6% and 1.0% higher than thosemade with the 30%. Measurements of the effects of different MRC chamber. The resultsare highly satisfactory beam filters will also be presented. as agreement of ±2% is more typicalfor such studies. Phase two measurements (at NAC)will rjalso be described. C)4

132 1988 World Congress on Medical Physics and BiomedicalEngineering San Antonio, Texas Medical Physics Scientific Papers

MP34.8 MP34.9 The Effects of Various Filters on a Fast Neutron Sources for Therapy-Current p(66 M6V)/Be Neutron Therapy Bean. Problems and Need for Basic Data, D.T.L. Jones* and M. YudeLev, National M.A.Chaudhri*, Austin Hospital -nd Accelerator Centre, 7131 Faure, South Africa University of Melbourne, Melbour.e, Australia. Several different types of beam filters may be required for the clinicaluseofa neutron In this paper a brief review of the source. These includeshapedmetalfilters historical development of the production whichflatten or rotate the beam profiles, of therapy neutron beams is presented, hydrogenous filters which harden the beam and with special reference to the author's reduce the OBE differential between surface and contribution, using the available nuclear depth and metal filters which absorb charged data. Different nuclear reactions and particles. The effects of such filters on dose target systems hw,c been critically rate and dose distribution have been examined regarding their suitability for investigated on the National Accelerator cyclotrons of different sizes. A few Centre's p(66 MeV)/Be (40 MeV) isocentric current problems in this field, neutron therapy beam. Iron filters flatten the especially relating to the non- beam to within ±3% at 10 cm depth in water for availability of appropriate nuclear data, ail field sizes while the maximum off-axis dose where the nuclear and radiationphysics enhancement at thesurfaceis +7%.Tungsten communities can greatly contribute, are wedge filters rotate the beam profile at 10 cm highlighted. Specific recommendations depth by 30', 45' and 60'. Fot a 10 x 10 cm2 are made as to what sort of nuclear data field the addition of 1,3 and 5 cm thick need to be acquired/compiled that would polyethylene filters reduces the dose rate to be most useful in the neutron therapy 93%, 80% and 70% and increases the 50% depth programme. dose frow 16.2 cm to 16.5,16.9 and 17.2 cm respectively. A few millimetres of leador steel has only a marginal effect on the build-up characteristics of an uncompensated beam. Adequate restoration of skin sparing is achieve.' if bolus material is lined with 2 mm of lead.

Mr 34.10 MP34.11 Calculated microdosimetry of primary and Calculating Radiation Dosc in HeterogeneousMedia With scatter neutron fluence components in a Different Atomic Numbers, T.R. Mackie,B. Thomadsen: S. muscle equivalent phantom for En<20MeV., Shahabi, B. Paliwal, Department of MedicalPhysics and R.E. Wallace* and P. Bloch, Univ of Penna, Human Oncology, University of Wisconsin, Madison, Wisconsin Dept of Radiation Oncology, Phila., Pa. 53106 The calculation of effective neutron dose requires knowledge of (1) the KERMA, The use of the convolution/superposition method for photon (2) the microdosimetry of charged parti- beam cavity calculations was explored. The dose per unit cles delivering the dose, and (3) the bio- fluencenear slaband three-dimensional phantoms, that logical response as a function of charged contair 1 heterogeneities with differences in atomic particle lineal energy (ie: LET). The number, were computed using convolution kernels generated microdosimetry of charged particles assoc- in a water phantom. The energy released at the interaction iated with primary and scatter neutrons sites was scaled with respect to the ratio of the mass were calculated as a function of depth in energy absorption coefficient of the medium to that of a muscle equivalent phantom. The micro- water. The dose deposited at the dose deposition sites was dosimetric spectrum for primary neutrons scaled tothe ratioof mass collision stopping powers (P) is spatially invariant.The micro- (Bragg-Gray) between the medium and water. The results were dosimetric spectrum for scattered neutrons compared to MAte Carlocalculations and measurements. (S) was found to be nearly independent of Very good agreement was obtained for air and aluminum slabs depth. The dose average lineal energy, located in the build-up region in a water phantom at photon d, for primary and scatter components energies of 1.25,10 and 20 MeV between the convolution/ are different, and for the total (P+S) superposition method and Monte Carlo benchmarks of Bielajew neutron dose, d varies significantly and Rogers (2). Reasonable agreement was obtained with the with depth. This behaviour will be measurementsof Ogunleye et al.(3) obtained inLiF TL discussed with regard to the variation oT dosimeters in aluminum, copper and lead phantoms irradiated d with neutron energy. A strategy will by Co-60 photons. be presented to compute a dose response factor using constant values for primary 1) Bielajew, A.F.and Rogers, O.N.O., Presented atthe and scattered neutron components. Workshop on the Convolution Method for Calculating Radiation Dose, Regina, Canada (1985). 2) Ogunleye, 0.T., Attix, F.H., and Paliwal, B.R.,Pte. Med. Biol. 25, 203-213. 1e 5 1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas 133 Medical Physics Scientific Papers

MP34.12 MP34.13 Separation of Primary and Scatter Doses in On Methods of Photon a Co-60 Photon Beam, P. Nizin*, K. Kase and Dose B. Bjarngard, Joint Center for Radiation Calculations, J.W. Wong* and J.A. Purdy, Washington University School Therapy, Harvard Medical School, Boston,MA 02115, USA, and University of Massachusetts of Medicine, St. Louis, MO, USA. Medical Center, Department of Radiation Oncology, Worcester, MA 10605, USA. We have examined eight approximate methods Results of theoretical considerations and for photon dose calculations: Ratio of TAR experimental measurements are presented, which (RTAR), modified Batho power law,and allow the separation of primary and scatter Differential SAR (DSAR) method, all devel- doses in Co-60 photon beams. If a small oped prior to CT; Equivalent TAR (ETAR) attenuator is placed in the beam, the dose from method, developed to make use of CT data primary radiation will be altered but the dose but needing increased computer power; FFT from scattered radiation will not change convolution, scatter ray-tracing Delta significantly. The ratio of the primary doses C Volume (DV), Dose Spread Array (DSA), and can be determined by relative attenuation Differential Pencil Beam (DPB)methods, measurements. It can be shown that the dose developed more recently, using theoreti- from primary radiation in the water is cally generated input data, but having P.(1-1/C)(D-DO, limited practical applications because of where D and D. are measured total doses without inadequate computer power. Horizontal slab and with the attenuator in the beam. phantom measurements were used to study Measurements were made with aluminum and photon transport corrections. RTARover- graphite attenuators. It was found that the estimated lung corrections by about 10%. results do not depend upon the material used. Not all scatter ray-trace approaches were The data for primary and scatter dosesare in better as DSAR was inferior to Batho. FFT good agreement with recently published values. A correction factor for possible perturbation of approximation omitted ray-tracing and led phantom-generated scatter is established, and to large artifacts at depth. Ray-tracing DV, DPB and DSA methods agreed with cobalt application of the method to megavoltagex-ray beams is discussed. 60 measurements mostly to within 2%. Non ray-tracingETARwas onlyslightly inferior for these geometries. (Supported by NCI CA-41574 and CM-47696).

MP34.14 MP34.15 Monte Carlo Studies of Small Inho- A 3D Implementation of the STAR mogeneity Dose Perturbations in Method Using FFT Convolution, C.X. Photon Beams, S.S. Hancock*, W.S. Yu*, P.Y. Ge, C.X. Yanand J.W. Wong, Yu and J.W. Wong, Washington University School of Washington University Schoolof Medicine, St. Louis, MO, USA. Medicine, St. Louis, MO, USA.

Photon dose is enhanced locally The equivalent tissue-air-ratio (ETAR) near bone- method, despite known weaknesses, provides like material. We have studiedthe rela- results that rival scatter ray-trace tive contribution of the variouscharged methods for photon dose calculations when particle transport components for6 MV and 18 MV x-rays. The EGS4 code electronic equilibrium exists. Although was used to fully 3D in formulation, an empirical CT calculate the angular distributionand energy spectra of the charged particles slice coalescing approach is used to reduce the calculations to a more manage- emerging from spherical inhomogeneitiesin able 2D form. As a result, the method is water. Compton and pair-productioninter- actions in the sphere and not suitable for irregular beam shaping. water are scored We have noted that the ETAR calculations separately. The results are then usedin a simpler CSDA model to delineate represent a density convolution which can the char- make use of the FFT technique. 3D ETAR FFT acteristic of the various doseperturba- tion components. There is calculations, more versatile than the 2D a net reduction version, have been implemented on our VAX of charged particle fluence emittedat small angles and a net increase 8600. The FFT approach compares favorably at larger with results using the earlier version and angles. Contributions due to thecharged particles originating from the sphere measurements. Calculations for 323 ve:ume and elements took less than 2 min., while the the water medium peaked atdifferent directions. The total charged particle direct convolution took about 1 hr. The approach offers a much improved option flux per unit area emerging fromthe over current 1D methods for practical 3D sphere is independent of its size.(Sup- ported by grant CA-41574). RTP until more advanced methods become practical.(Supported by NCI grant CA- 41574). 156 134 1988 World Congress on Medical Physics and Biomedical Engineering San Antonio, Texas Medical Physics Scientific Papers

MP34.16 MP34.17 Installation of the EGS4 Monte Carlo Optimization of CBEAM model parameters Code onaMtcrocomputer System, for linac photon beams, K. Ayyangar, S.Walker and D. Jette*, Institute I. raftari*, J.R. Palta, and of Appged Physiology and Medicine, N. 9mtharalingan, Department of 701 16 Avenue, Seattle, WA 98122. Radiation Therapy and Nuclear Medicine, Thomas Jefferson (him. Hospital, 1 We have installed the EGS4 Monte Carlo code on Fbiladelphia, PA 19107 the new Compaq 386/20 microcomputer system. This microcomputer, with its 32-bit wordlength, Measured bean profiles and central axis depth dose data 20 MHz clock speed, and Weitek mathematics co- for photonbeansfrom our linear accelerator are used processor, promises to rival a mainframe system to generate a dose matrix representing thefulltem. in its Monte Carlo run speed, and at very much Using the modified CRAM model, a corresponding dose less cost. The installation is by no means matrix is also calculated. The calculational model trivial, with an appropriate operating system requires a set of 3 parameters to define the be= and Fortran compiler required, and conversion of intensity at beam edges and 3 for the primary profile the EGS4 files from mainframe to microcomputer factor expressed as a ftmction of distance from the format is fairly involved.However, this work central axis.Anadditional parameter to account for should help other researchers to install EGS4 on the transmission of the bean through the collimator is their own microcomputer systems.We present also included.Anoptimization program has been adapted benchmark tests of running EGS4 on our system, to automatically adjust the parameters used in the vs. a VAX mainframe system. CEFAM model to minimize the chi -square between the measured and calculated data. The processof These investigations have been supported by a optimization is repeated for different field sizas. grant (07 RO1 CA41490) from the National Cancer Typically the average values of the parameters for Institute. small (6X6 cn2), medium (10X10 cm2) and large (20x20 cm2)field sizes are found to representthe bean 1. W.R. Nelson, H. Hirayama, and D.W.O. Rogers, adequately for all field sizes. The calculated and the "The EGS4 Code System", Stanford Linear measured doses at any point agree to within 2% for any Accelerator. Report 265 (Stanford University, field size in the range 4X4 an2 to 40X40 an2. The Stanford, CA, 1985). method of optimization and the agreement of the data with measurements will be discussed.

MP34.18 MP34. 19 Corrections to Scatter Dose in Applicationsof a Laplace Transform Heterogeneous Media for Photon Beams, PairModeltoDeconvolve theHigh R.K. Rice *, B. E. Bjarngard, and L. M. Energy Photon Spectra from Chin, Joint Center for Radiation Therapy, Transmissions Measurements. J.S. Department of Radiation Therapy, Harvard Ramaswamy*, W.F. Hanson, B.R. Medical School, Boston, MA 02115 Archer, and A.L. Boyer, Department of Radiation Physics, University of Texas To calculate photon dose distributions in System Cancer Center, Houston, TX. heterogeneous media accurately it is necessary to consider changes to the No direct method currently exists to measure the primary and scatter components of the dose energy spectra of megavoltage photon beams. Monte separately. Existing algorithms using this Carlo simulation and several indirect experimental procedure usually involve an integration methods have been suggested to quantify high energy over the irradiated volume. For a 3- photon spectra. We are exploring the Laplace dimensional integration the time required transform pair method as a practical method of is prohibitive for routine treatment determining high energy photon spectra. By modeling planning. We have developed an algorithm thetransmissioncurve andthespectrumwith to correct the scatter component of the analytical functions, one can use a Laplace transform dose in layered media by calculating the to deduce one from the other. Spectra reconstructed effective scatter-to-primary ratio (SPR) from transmission data measured using different along the central ray. The SPR is materials are intercompared to verify the uniqueness determined based on the radiologic depth of the Laplace transform pair model. Spectra on and and density-scaled field size at each off the central axis of 6 and 18 MV photon beam thus interface. By combining this effective SPR determined are compared with the spectra generated with a 2-dimensional scatter integration Carlo and primary dose c.,rrection accurate by Monte simulation. The selectionof results (<5% error in total dose) are attenuators, measurement of maximum energy of the achieved for a variety of geometries and photon spectra, and conversion of chamber response beam energies including a mediastinum to fluence are discussed. geometry for 4 and 15 MV photon beams. Measured results are compared with the new This work supported in part by P.H.S. Grants CA algorithm and Monte Carlo simulations. 43840 and CA 10953 awarded by the National Cancer InstiVerDHHS. J. (4.)ei

1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas 135 Medical PhysicsScientific Paper('

MP34.20 MP34.21 A Geometric Parameter for Represen- A High-Precision Method for Computing tation of Scatter in Photon Beams, the Missing-Tissue, N. Samsami#, F.R. Bagne and C.H. Obcemea4k , H. Rashid and B.E. S. Hoke, Medical College of Ohio, Toledo, Bjarngard, Joint Center for Radia- Ohio High tion Therapy, Boston,MA 02115 USA. Precisionradiation therapy requiresthat a uniform dose be deliveredacross the entire Before measured dose data for photon treatment volume. To accurately compensate beams are used for dose calculations for the effects of the missingtissue,it is based on scatter integration,primary imperative that all the necessary parameters and scatter dose components must be whichcan affect doseinthe treatment volume separated and the data edited. Wehave be determined. We havedeveloped a precise, examined how this may be aided by a yet practical technique for computing the simple geometric expression for the thickness of missing-tissue alongthe ray-lines. scatter-to-primary ratio:SPR=ALrd/(Azr+d). The technique takes This is a qualitative description into account the phantom of scatteringeffects, the first-order scatter as a function of missing-tissue scatter reduction and the scattering due to thecollimator depth d and average field radiusr. The jaws. To data used were tissue-phantom ratios assess the most suitable tissue-compensator material, a series (TPR) for 8 MV X-rays (d=2-30cm,square of measurements were done using polystyrene,lead, fields s=4-35 cm) and TAR datafor Co -60 Cerrobend and polyethylene-lead. All dose (d=0.5-30cm,s=4-75cm). The primary and measurements scatter components were separated by were made in tissue-equivalent phantom using 6MV (Clinac three different approaches, one basedon 6/100) and 10MV the formula above, with +1% agreement. (Clinac-18) photons. Our data indicate that The resulting TPR or TAR valueswere then tissue-equivalent materials providemore accuracy fitted to the expression witha least- than high-atomic number materials. Additionally, squares procedure. The accuracy was in using the described method, dose uniformity canbe general within ±2% for Co-60, somewhat achieved with less than a few percent less at d=30 cm. For 8 MV X-rays,only error over the entire clinically-used range occasional points at shallow depthsgave of depths, missing-tissue thicknessesand field discrepancies greater than 1%. sizes. This method is particularly suitable for generating cross-sectional layersof missing tissue using computer software.

MP34.22 MP34.23 An Algorithm for Design of Beam Compensators, SET OF PROGRAMMES FOR RADIOTHERAPY Renner, T.P. O'Connor, N.M. Bermudez, Dept. TREATMENT PLANNING. PART I Radiation Therapy, Community Rospital, 1503 - EXTERNAL HIGH ENERGY PHOTON Ritter Ave., Indianapolis, IN 1:211, USA. BEAMS. M.C. Lizuain*, C. Pino and C. Cinos. Hospital de A three dimensional computer propram Bellvitge. Barcelona was written (Spain). for the VAX computer that can produce isodose curves in any plane of the patient's body for virtually any beam setup that can be achievedon A set of Fortran-77 programmes the treatment machine, including any combination for Radiotherapytreatment planninghave of blocks, wedges, and beam compensators, provided been developed in a PC. The Cunningham and the patient can be scanned on a CT scanr-r in the Wilkinson method was used for calculating the treatment position. Computed doses are accurate to dose distribution in high energyradiation within at one standard deviation on comparison beams. The values of Co-60 SAR and TAR were with measured values in an anthropomorphicphantom. taken from the literature and RX-18 MV were Beam compensators are optimally designed to give a calculated from measures made with ionization uniform dose to any plane or midsurface that inter- chambers. sects a single beam, or to give a uniform dose to The programmes consist of three parts: the volume defined by the intersection of two or 1: Generation of tables of numbers giving the more teams. The primary and scatter components radiation distributions for beams of are taken into account separately, as well as the rectangular cross-section with patient's shape and internal heterogeneities. or without The wedge filters. design of the beam compensators is formulatedas a 21 Calculation of absorbed dose at points of linear programming problem and solved with a varia- one volume tion of the Simplex Method. irradiated with one or two Beam weirhtinp *'actors irregular sectior )-ams. are also obtained as part of the solution. 3: Calculation , dose distribution in a plane of a volume, irradiated by self-eat regular beams. (Six fixed or isocentric beams or two rotational beams).

136 158 1988 World Congress on Medical Physics and BiomedicalEngineeringSan Antonio, Texas Medical Physics Scientific Papers

MP34.24 MP34.25 Computing dose distribution for electron beams with The Effect of Oblique Incidence upon rotated collimator using Generalized Gaussian pencil Electron-Beam Dose Distributions, beam algorithm, S. Hy8dynmaa*, Technical Research D. Jette*, Institute of Applied Centre of Finland, Medical Engineering Laboratory, Physiology and Medicine, 701 16.t12. Box 316, SF-33101 Tampere, FINLAND Avenue, Seattle, WA 98122.

Generalized Gaussian pencil beam algorithm has been Surface curvature of the patient can significantly used for computing the absorbed dose for electron influence the dose distribution in electron-beam beams. In this work the algorithm was applied for teletherapy. In this study we systematically computing the dose distributions for fields with rotated examine the effect of oblique incidence upon collimator. The computation is based on calculating the electron-beam dose distributions, comparing the integration limits along the height of the field to the predictions of Fermi-Eyges multiple-scattering field edges at the position of each pencil beam. The theory and of second-order multiple-scattering computation was tested for a rotated 10x10 cm field for theoryl, with EGS4 Monte Carlo calculations. In = 11.4 MeV electrons (Microtron 14) in a water particular, we evaluate the finding of Ekstrand phantom. With collimator rotations up to 300 the and Dixon2, that the effect of oblique incidence agreement with measured values inside the field was is to increase the dose at small depths and to good. Outside thefield the computation overestimates decrease it at large depths. the dose (isodose curve shift < 5 mm). With collimator rotations higher than 300 the computation started to These investigations have been supported by a underestimate the dose inside the field (isodose curve grant (117 R01 CA41490) from the National Cancer shift < 5 mm up to 600 rotation). Outside the field the overestimation went up to 8 mm isodose curre shift for Institute. 600 rotation. The computation gives fairly good results inside the field for quite large collimator rotations. The 1. D. Jette, Med. Phys. 12, 178 (1985). approach can also be used for computing dose from 2. K.E. Ekstrand and R.I. Dixon, Med. Phys. 9, irregularly shaped fields, and computation in slices 276 (1982). outside the fields is possible.

MP34.26 MP34.27 Multiple-Scattering Angular Distributions A real time CT-linked 3-D treatment inElectronDose Calculation,D. Jette*, planning system, T.Nishidai*, Y.Nagata, Institute of Applied Physiology and Medicine, M.Takahashi and M.Abe, Kyoto Univ., 701 16th Avenue,Seattle, WA 98122. Kyoto 606, Japan

A problem with using Fermi-Eyges multiple- scattering theory in electron dose calculation is A real time CT-linked treatment planning that that theory predicts a Gaussian distribution system, called a CT simulator, has been in particle direction, at any particular depth in developed. The basic system consists of a CT a broad .:1,ectron beam. Such a Gaussian distri- scanner, a multi-image display component, a bution ignores the single-scattering tail of the treatment planning device with real time visual real angular distribution, leading to inaccurate optimization and a laser beam projecting dose calculation. On the other hand, Moliere component. All the components are combined on multiple-scattering theory includes single large- line and work simultaneously. With the angle scattering, but is valid only constant treatment planning device, visual optimization energy, and hence only near the surface of the and fast reconstruction of dose distributions, irradiated material. Our second-order multiple- oblique slices and simulation images are scattering theoryl does modify the Gaussian available. Accurate 3-D dose distribution is distribution so as to predict increased scattering calculated with heterogeneity correction by the at large angles, but like the Fermi-Eyges theory new equivalent TPR method on any interested it does not include the process of single large- slice. On two large multi-image monitors, the angle scattering. We compare the predictions of CT images and scannograms involvi.ig the target these multiple-scattering theories with EGS4 Monte outline, ROI, field s*.ze, block localization Carlo calculations, and discuss how to append a and dose distribution can be observed and an single-scattering theory to achieve accurate optimal planning is serched. The laser beam electron dose calculation. projector locates the planning result, e.g. beam center, field size. This new CT simulator These investigations have been supported by a can be used for 3-D planning and simulation for grant (117 RO1 CA41490) from the National Cancer radiotherapy within a reasonable short time. Institute. The research was done in collaboration with Shimazu Co. and NEC Co.. 1. D. Jette, Med. Phys. 12, 178 (1985). 159 137 1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas Medical Physics Scientific Papers

MP34.28 MP34.2.19 Implementation of a Three-Dimensional Planning System Using the GKS Graphics, Further Exploitation on Using Wedge Filters D. L. HcShan*, University of Michigan in Radiation Therapy Planning, Yi-minHu,* Medical Center, Ann Arbor, HI 48109. Ning Lin, Chun-li Zhang and Hong-zhiZhang Radiation Oncology Dept., Cancer Institute The GraphicsKernalSystem (GKS), an ISO/ANSI (Hospital), Chinese Academy of Medical graphics standard, has beenused to implement a Sciences, Beijing, China. full function three-dimensional radiationtherapy planning system. The standard provides a uniform On summarizing the experienc, of using interface between the treatment planningsoftware and a variety of graphical display devices. wedge filters iu radiation therapy plan- WhileGKS is presently only a 2-Dgraphics ning, authors derived theoreticallya form- standard, ourimplementation provides for device ulae which can in practice guide the plan- handler extension (via GKS escape functions) to ners who use Treatment Planning Systemor support overlay graphics, grayscale image help the radiotherapists who haven'tgot displays, and other image processing functions. Treatment Planning System at present to For 3-D graphics, a software layer written in choose independently the optimum wedge FORTRAN has beeninterfaced toGKS calls and provides the necessary tools to handle multi- angles and dose weightings of the beamsto format windows, 3-Dcoordinate transformations be used to achieve a very uniformdose (including perspective forbeam's eyeview distribution in target volume while sparing displays), clipping planes, axonometric displays effectively the normal tissues aroundand and surface rendering. An overviewof the the critical organs nearby. The dose dis- functionality of this implementation will be tributions deduced from the formulaeare presented along with a discussionof its performance in routine clinical operation. confirmed by RTPLAN and Film dosimetry in Anderson Phantom.

MP34.30 MP34.31 Comparison of Theraplan Electron Verification of External Beam Teletherapy Dose Calculations with Measurements Plans Produced by CT-Based PlanningCom- in an Anthropomorphic Phantom, puter, Lilian V, Rodriguez,* Jose R. J. Cygler* and J. Ross, Ontario Cancer Foundation, Reyes Memorial Medical Center,Department Ottawa Regional Cancer Centre, of Health, Manila 1003, and AgnettP. 190 Melrose Ave., Ottawa, Ontario K1Y 4K7 Peralta, Radiation Health Service, Depart- ment of Health, Manila 1003, Phillippines.

Electron dose distributions in an anthropomorphic The first CT-based treatment planningcom- (Rando)phantom were measured using TLD-100 rods. puter in the Phillippines has been installed CTscans of the head and chest wall sections of at the Radiotherapy Department of the the phantom were done using a GE 9800 scanner. Jose The measured doses for 6, 9, 13, 17, 20 MeV beams R. Reyes Memorial Medical Center. This were compared with isodose distributions development provides a new level ofaccur- calculated with Theraplan Version04 software acy and precision in the field of radiation (A.E.C.L.]. Thisprogram performs inhomogeneity therapy in the country. The present work corrections based on pixel by pixel data. describes the procedure performed to Discrepancies betweenmeastwedand calculated investigate the accuracy of treatmentplans doses fortheheadare of the order of 10-14%. made for external photon beam teletherapy. For the lungs Theraplan greatly overestimates the The study consists of two parts. dose. The differencesbetween and The first measured part is designed to detect any programming calculated data are larger for the larger error. electron energies (upto 40% for 20 MeV). The The second part involves verification reasons for such big discrepancies are discussed of dose calculations by actualmeasurement in terms oftheassumptions inherent in the in a phantom. algorithm employed by Theraplan.

138 160 1988 World Congress on Medical Physics and BiomedicalEngineering E3n Antonio, Texas Medical Physics Scientific Papers

MP34.32 MP34.33 QUANTITATIVE EVALUATION OF ALTERNATIVE Irregular Portal Planning in Radiation PLANS DURING COMPUTERISED RADIOTHERAPY Therapy - Use of CT Images for TREATMENT PLANNING PROCEDURE Target Volume Delineation, Preparation of Shields and Templates, Anil Kumar Sharma * S L Kapoor, TSG Integrations and R. Ravichandran, Kidwai Memorial Institute 202 Ashok Bhawan, 93 Nehru Place of Oncology,Bangalore- 560029, India. New Delhi The role of treatment planning in the Theaimof radiatic.i therapyplanningis radiotherapy procedure is to assist the to find an optimal way to deliver radical doctor in arriving at a prescription in radiation dose to the specified target order to give better quality of treatment volume while sparing the surrounding healthy to the patient. The prescription is a tissue as much as possible. A method is treatment scheme based on the clinical described to utilise Computed Tomography information about the patient, thera- data to provide accurateinformation for peutic objectives and the physical trec.t- delineating the target volume and to ensure ment plan. Alternative treatment plans reduction in the volume of healthy tissue are normally visualised by the doctor with in the radiation field. Tumour bearing the help of treatment planning systems to areas aremarkedontheindividualscans arrive at a plan which best meets the of the exposure series and projected onto overall objectives. This paper suggests thelateral and AP topogramsto getthe an easy-to-use scheme for quantitative irregular portals. The resultant irregular evaluation of treatment plans. portal is additionally superimposed with The scheme is based on comparing the the superior,middle and inferior cutlines desired dose to the various parts of the which help in positioning the patient and body (treatment objective) with the cal- decide the collimator, couch and gantry culated dose for a given treatment plan. angles.By employing suitable magnification A direct comparison of the objective with factor, individual shields are prepared the actual dose values to the points of from the topogram films carrying projected ROI's defined in the sequential tomograms. interest yields valuable quantitative These topograms are also used as templates data for evaluating the effectiveness of for adjusting the shielding blocks. alternate plans.

MP34.34 MP34.35 Objective Measures of Radiation Therapy Developmentofa dose calculation workstation ,J Treatment Plan Quality, Bruce A. Lulu, Radiation Viitanen*, J Heini1A, S Hyodynmaa and H Puurunen, Oncology, University of Arizona, Tucson, AZ Technical Research Centre of Finland, Medical 85724, USA. Engineering Laboratory, P.O. Box 316, SF-33101 Tampere, Finland and Dosetek Oy, P.O. Box 2, When comparing competing radiotherapy treatment 02131 Espoo, Finland plans, it is useful to abstract from the plans objective measures of quality (MOQ) more compact In connection with the Nordic Research Program than the usual multiplanar isodose CART (Computer Aided Radio Therapy) we developed distributions. Traditionally, concepts such as a new comprehensive and flexible concept for the dose minimum target dose, percent dose inhomogeneity calculation workstation for radiation therapy. The CT- throughout the target volume, and more recently based system, CADPLAN, has been implemented on a the dose-volume histogram, have been used. We modern workstation, Hewlett Packard 9000/350, present four measures )f plan guilty that can although the development work has been done on a be derived from a dose-volume histogram by VAX-11/750. inspection. These dimensionless measures contain two physician-adjustable parameters that have The menu and mouse driven system in,,,udes vast clear clinical meanings. The MOQ are applicable image processing facilities like beam's eye view to teletherapy plans and as well as presentation, multislice display, image reconstructions brachytherapy plans. The MOQ require the concept and 3D surface models. The photon dose calculation of target volume, which could in fact be an algorithms use pixel-by-pixel bases: generalized power isodose surface, as well as no other clinically law or equivalent TAR method for inhomogeneity significant structures other than "target" and corrections. For compensating filter calculations an "non-target". We show that all four MOQ are iterative method is used. The electron distributions necessary in the sense that a plan with three are calculated using the generalized Gaussian pencil good MOQ can be constructed that is clearly beam algorithm. Boluses can be inserted on the clinically unacceptable. Conversely, the four patient and included in the calculation and dose MOQ are sufficient to characterize all the histograms in the given ROI's can be generated. essential features of an ideal plan. Software for intracavitary calculations is availabz, in the system. The intracavitary distributions can be combined with external distributions and vdaapirpositioned on CT-slices

1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas 139 Medical Physics Scientific Papers

MP34.36 HP34. 37 5D Computerized Treatment Design Using Modern Some Experiences on Radiation Workstations,George W. Sherouse',Radiation TherapyPACS E.Takenaka,Y.Wada, Oncology, University of North Carolina, Chapel R.HoSaka and TFukui. Hill, NC 27514 National Defense Medical College In recent years the rapid evolution of high-performance workstations has given rise to new models in the computer Fifty national university hospitals have industry for the deployment of computing resources.For been year by year annoyed by piling up of departments where a great deal of computing power is Xray films due to the development of many imaging modalities. Study group of radio required, a network of single-user workstations with shared logy department of national university access to files and unique peripherals has become a cost- hospitals has been discussed PACS to be effective alternative to more conventional centralized com- made in their hospitals. As a member of puting. Moreover, even for sites where one or few worksta- group,we have checked and studied the func tions are needed, the ability to place very capable computers tionand performance of radiation therapy with high-resolution graphic displays on users' desktop PACS. Radiation therapy PACS treats all opens new possibilities for the design of highly interactive kinds of medical images in addition to software tools. A prototypical workstation network config- radiographs. This PACS are the most suit ured to support three-dimensional radiotherapy planning able to determine TNM and target volume,to will be presented. The richness of this approach will be detect early recurrence and metastasis and illustrated by description of some 3D radiotherapy software to teach radiation oncology to medical stu dents and post graduate students. A PACS tools, designed and built using the UNIXt operating system module with a laser scanning digitizer,two and the X Window System available on most modern works- high resolution CRTs,light discs etc was tations. set in radiotherapy section. Information contents of medical images,function and tUNDC is a trademark of Bell Laboratories performance of this PACS module and experi ences on module will be reported.

MP34.38 MP34.39 A PROJECT FOR AN EXPERT SYSTEM IN A PC BASED WORKSTATION FOR THE PLANNING OF RADIOTHERAPY STEREOTACTIC NEUROSURGERY L.Azario, A.Piermattei, G.Arcovito, Physics T M Peters; J A Clark, G B Pike, A Olivier. Inst.,Univ.Catt. S. C.,ROMA L.Andreucci*, McConnell Brain Imaging Unit, Montreal Neurological A.Schenone ,National Cancer Institue,GENOVA Institute, Montrcial, Quebec. C.Bacci Phys. Dep.,"La Sapienza" Univ.,ROMA

R. Bernabei, S.D'Angelo , M.P. De Pascale Physics Dep., "Tor Vergata" Univ.,ROMA Siereotactic surgery requires knowledge of cerebralstructures Data coming from CT, NMR and PET imagesare derived from more than one image source. We have developed well treated by means of optimized a software.A typical use of the computer for PC-AT based workstation which accepts patient images, made this kind of research is to store a with the stereotactic frame in place, from CT, MRI and DSA great deal of clinical data to optimize the modalities. Referenc^ marheis on the framemay be identified in tratement planning. The aim of our project the images to establish the coordinate geometry for each modal. isto make an expert system to contribute ity. to clinical treatments of different Target points may be identified on each image type and patologies, tra- expecialLy related to jectories of probe paths to these points defined. Targets iden- increasing quantity of the radiological tified on one set of images may be transferred automatically Informations. The realization will be to completed in this subsequent steps.In a n her images of the same patient, in order for example to guar. first time we will develope the software antee a vascular free path of approach to a target point deep for thetridimensionalreconstruction of within the brain. images comingfroma CT system. In the To date several handled patients have had stereotacticsur- second step, we will realize a software to gery performed on the basis of plans using this system. Pro- estimate the dose distribution an the cedures included biopsy and aspiration of lesions, implantation isoeffect to optimize the radioterapic treattment. The aim of this poster s to of electrodes for the recording of deep EEC signals, andradio show with some detail the purposes ofthis surgical techniques based on the use of a highenergy linear ac- project and the status of the our celerator. tridimensional imaging software. 162 140 1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas Medical Physics Scientific Papers

MP34.40 MP35.1 Maximum Utilization of Radiation Oncology Synthetic diamond crystals as 'state of the are dosimeters Computer System R.J. Reddy* and T.L. Nam Schonland Research Centre, University of the Witwatersrand Simon C.P. Lam: Steven H. Mahood, Larry L. PO Wits, Johannesburg, 2050 South Africa Windedahl, Michael L. Matheny, Morristown Memorial Hospital, Morristown, NJ 07960 Contemporary high speed, high capacity, multi- tasking systems can adequately perform routine clinical treatment planning, and other ancillary The main attraction of using diamond as an in -vivo or as an environmental functions. dosimeter Is that it is "tissue equivalent" to a high degree. It is however also non-toxic, robust and stable and can repeatedly be used. Only a rare We put together a local computer network, group of natural diamonds ad as efficient detectors of ionizing radiations centered around a parallel link of two PDP 11/73 however but diamonds with optimum quantities of selected impurities can computers. It is capable of simultaneously perfor- be synthesized to act as radiation detectors with performance responses ming active treatment planning in 2 stations. It that are superior to other tissue equivalent materials. The diamonds can be is capable of performing a multitude of tasks synthesized to act as therrnoluminescent dosimeters or as ionization cham- simultaneously in 10 stations. Its digital image bers. Therapy doses can easily be recorded and radiation levels approaching processing ability enables optimal combination of natural background values can be detected. Depending upon the applica images from different modalities for treatment Lion, the linearity of response extends over five orders of magnitude of dose within a useful range that extends to nine orders. The physical forms of the Video linkage between simulator and planning. synthetic material can be crystals, grits or powders. All types of ionizing physician's offices help streamline the patient radiations can be detected, particulate as well u electro-magnetic. throughput. Direct data link between treatment Because of the tissue equivalents, and stability of these specially synthesized units and main frames facilitate the accuracy of stones they could very appropriately be used as the detecting component treatment plan implementation. It also provides in instruments calibrated in ''absorbed" rather than "exposure"dose, ado recording and verification of actual treatments because of their small volumes ( mm3 ), they would lend themselves given. well to the miniaturization of such instruments. Ancillary tasks handled by this network include patient and staff scheduling, fiscal planning, project integration and analysis and remote communications to both the Hospital Information System and Radiology Information System. A direct link to the Nuclear Medicine MicroVAX computer is also provided.

MP35.2 MP35.3 A directcomparisonofwater Feasibilityof Non-Equilibrium calorimetry and Fricke dosimetry Dosimetry with aFricke Ferrous Sulfate System, L.D. Simpson, Washington University School of Ken It. Shortt', C.K.Ross, N.K.Klassen and Medicine, St. Louis, MO, USA. G.D.Smith, Division of Physics NItCC, Ottawa, Canada K1A OR6 Physicists' ability to measure dose under clinical, non-electronicequilibrium We have completed a direct comparison of water calorimetry and conditions is weak. Models to predict dose Frickc dosimetry for 20 MV x-rays. The calorimeter consisted of for external and internal sources thus a small sealed vessel containing 100 ml of stirred water saturated remain Incomplete and unverified. The most with a 50/50 mixture of 112 and 02 gases. Based on calculations sophisticated models affected by this and previous experiments involving different gases, we estimate weakness are the Monte Carlo codes which that the heat defect for this system is -2.1%. The calorimeter was have not been verified in some regions of non-electronic equilibrium. Limitations of filled with 100 ml of air-saturated Fricke solution and irradiated existir7 experimental techniques will be under identical conditions. Our Fricke system uses the Svensson reviewer': to include ionization chambers, and Brahme value for cG of 3.515 x 10-3 1 cm-1 J-1 and gives TLD, ani the Fricke ferrous sulfate dosi- agreement to within 0.2% with the dose to water for 60Co 7-rays meter. Alternative methods for overcoming obtained via, graphite calorimetry. After correcting the calori these limitations will bediscussed. metric measurements for the effects of the foreign mass (mostly Emphasis will be placed on my development glass), we find the dose to water determined by water calorime- of a new Fricke ferrous sulfate dosimetry try is 1.00610.004 times the dose determined by Fricke dosimetry. system.The water-equivalent detector We plan to make the same comparison for 60Co 7-rays and a range volume is <0.001 cc, configurable as a <4p of electron energies. Eventually, we hope to establish a calibra- thick disk or as a cylinder 1 mm diameter tion service at high energy based on this water calorimeter as a by 1 mm length. The detection limit is <0.1 Gv. Two important system components primary standard and Fricke dosimetry as a transfer standard. will be discussed: 1)a transition metal analyzer based on high performance ion chromotography and 2)the modified Fricke detectors. 183

1988 World Congress on Medical Physics and Biomedical EngineeringSat Antonio, Texas 141 Medical Physics Scientific Papers

MP35.4 MP35.5 The point of measurement for thick- Radiation dose distributionin walled ion chambers in 7-beams Agar gels imaged by MR, Alex F. Bielajew* & D.W.O.Rogers, Division of R.Renzi,Fisica Medica,Universita PhysicsNRCC, Ottawa, Canada K1AOR6 di Firenze,!taly. When ion chambersare used for determining exposure in point source 60Co fields, there is an uncertaintyon the point of mea- Radiation dosecan be measured through surement owing to the finite size of the chamberand the spatially the changes inducedinthe ferric ion varying photon fluence. In this paper we preseata theory for concentration of a Frickesolution. The the point of measurement building upon an "exact" theory of ion conversion of ferrousto ferric ions af- chamber response in uniform fieldsl. Thetheory is general enough fects the NMR relaxationtimes of pro- to allow for fields of arbitrary spatialconfiguration but for this tons in water, too. paper we consider point sources only (i.e. 1/r2dependence). We Some authors (Gore et al., demonstrate that for practical distances Phys.Med (SSD's >50cm or so), Biol.,1984,29,1189)havealready there is a negligibly small correction (<0.03%) round relative to the ge- the increase of longitudinaland trans- ometric center of the air cavity. This is versal demonstrated by Monte relaxation rates to be linear Carlo simulations for all standard icnchamber geometries, plane- respect to the cbsorbee dose. parallel, thimble-type, and spherical.This is at variance with the In this work, a solution c°Co exposure standards at the Pr of ferrous ions and BIPM2 where point-of- in Agar gel (Appleby measurement corrections of the order of 0.5% et al.,Med are used. Neither Phys.,1987,14,382) the NBS nor the NRC(Canada) apply such hasbeenirradiated a correction although with a X-ray beam (25 uncertainties of 0.1% and 0.2% ,respectively, MV) from a Linac. are associated with Ti and T2 values have this correction. We recommend that been evaluated by point-of-measurement cor- an Imaging MR machine. rections not be applied for thick-walledion chambers in 60Co Thisapproach is beams and that no error be associatedwith this correction. aimed to empl4y the spatial resolution Bielajew A.F., Phys. Med. Biol.30(1985)419-427. offered by theMRI technique in 2) Niatel M.-T. et al, Metrologia11(1975)17-23. the dosimetryofcomplex radiation treatments.

MP35.6 MP35.7 Stopping power ratios in electron beams. Dose Distributions of theFirst Clinical P Andreo*. A.Brahme and A.E.Nahura, Stereotactic RadiosurgicalUnit (Gamma Karolinska Institute. S-10401 Stockholm. Knife) in the UnitedStates, G. Lindner*, Sweden. and Royal Marsden Hospital, N. Smarra, University ofPittsburgh, Surrey Pittsburgh, PA SM2 5PT, England.

On August 14, 1987 the firststereotactic Most of the existing dosimetry protocolsagree in radiosurgical procedurewas nerformed in the United States using the recommended values for the water/airstopping the Leksell Gamma Knife at Presbyterian UniversityHospital power ratios in electron beams, i.e., the values from in Pittsburgh, Pennsylvania. Gamma Knife Bergerasgiveninthe AAPM protocol. radiosurgery refers tostereotactic locali- Experimental results have shown that zation and subsequent singletreatment, the use of closed skull irradiation s-ratios derived for mono-energeticand mono- of a brain lesion with 201 highly focusedbeams of cobalt 60 directional beams is adequate for radiation. beams with Treatment planning is doneon certain energy and angular spread,but not for a Digital MicroVAX II computerrunning soft- "clean" beams. ware developed at the Universityof Uppsala, Uppsala, Sweden specificallyfor the Gamma Calculations have been performed usingthe Monte Knife. The hemispherical Carlo method. Computed array of sources, swatimair for beams the large number of smalldiameter beams without energy and angular spreaddo not show and the steep dose gradientssurrounding significant differences from theBerger values, a target lesion make physicalcharacteri- especially for the electron beam zation of the radiationfield complex. In- energies more ph.intom film and LiFthermoluminescent commonly used in Radiotherapy (E0c20MeV). The dosimetry (TLD) measurementswere performed. Our studies inclusion of energy and angularspread in the have shown goodagree- calculations has shown a variation ment between measured andcomputer gene- of swater,air at rated field profiles and the depth of the maximum absorbed isodose distri- doge. butions for 4, 8, 14 and18 mm diameter fields. 164 142 1988 World Congress on Medical Physics andBiomedical EngineeringSan Antonio, Texas Medical Physics Scientilic Papers

MP35.8 W35.9 Measurement of Charged Particle Angular Emissionof Secondary Contamination ina Megavoltage Beam, Charged Particles from Small N.C. Ikoro*, E.C, Scarbrough, P.P. Cylindrical Volume for 6 MV and 18 Antich, U.T. Southwestern Medical MV X-Rays, S.S. Hancock*, C.X. Yu Center at Dallas, Dept. of Radiology, and J.W. Wong, Washington Unive.:- 5323 Harry HinesBlvd., Dallas,TX sitySchool otMedicine, St. 75235-9071 Louis, MO, USA.

The magnitude and penetration of electron contamination in 6 MV X-ray beam is investi- An understanding of charged particle gated using a specially built Charged particle productions and transport in finite size detector. The charge detector, with an volume element is important for accurate aluminium electrode,is used in the direct dose calculations. We have measured the measurement of contaminant electron fluence. angular emission of the charged particles The fluence is converted to dose by comparison from small acrylic, aluminum and copper withthefluence and dose measured from a cylinders in 6 MV and 18 MV x-ray beams to modified 6 MeV electron beam. Further analy- study the effects of size and composition sis of the electrons is carried out using the of irradiated volume. Electron contamina- detector in conjunction with a permanent tions, primaryandscattered photon magnet. contributions were minimized.Charged Present data suggest that contamination particle emission due to pair-productions electrons can be explicitly characterized; the increased with atomic number. Coulomb observed changes in the build-up region, with scattering shifted the emission to larger collimatoropening, are due to electrons; angles. The charged particle flux at each backscattered electrons from the medium are angle became proportional to surface arta responsible for as much as 35% of the observed when the cylinder diameter approached the surface dose at maximum collimator setting, at charged particle range. The measurements standard treatment distance. agreed with Monte Carlo simulations. The effects of multiple co;lomb scattering play a major role in the angular distribution of the charged particles. (Supported by NCI grant CA-41574).

MP36.1 MP36.2 The Light Propagation Physics of Breast Analysis of trace elements in seminal Ttansillumination Imaging, D.Wyman*, fluid by X-ray flourescence, R.CESARE0*, M.Patterson, V.Hansen, V.Peters, The Centro Ingegneria Biomedica - University Ontario Cancer Treatment and Research La Sapienza, A.MAGRINI & G. IZZO, Foundation, 711 Concession Street, Fisica Medica - University Tor Vergata, Hamilton, Ontario, Canada, 1,8V 1C3 G. D'OTTAVIO & A. LAGANA', Ospedale S. Carlo di Nancy - Rome - Italy,L.J. Breast transillumination imaging is a method of PEDRAZA, Centro de Estudios Nucleares - detectingabnormalities based on the differential Universidad Nacional Autonoma de Mexico, transmission through the breastof red and S.Q.B. MAHTABOALLY, Physics Department - infrared light. We quantitatively assess the Rajshahi University - Bangladesh propagationoflightthrough various tissues in anattempt to determine both the optimization and Energy dispersive X-ray flourescence induced by limitations of the method. monoenergetic photons was used in the analysis This assessment is basedondetermining the oftract elements in seminalfluid. Mono- fundamental optical properties of regular, energetic radiation of various energies was fibrous and malignant breast tissue byfirst obtained by using an X-ray tube with secondary measuring, fortissue samples, the reflected and external targets(Sr andZn).Fluorescence transmittedlightusing anintegrating sphere. radiationfrom eachsecondarytarget was These measurements are inverted to yield optical collimated and sequentially used to irradiate properties through iterative Monte Carlo fluid samples deposited onthin filters. A simulation. Phantom models have been constructed linear relationship between X-ray...ounts and for arange of optical properties. Both Monte content of each element was found. 44 samples of Carlosimulation and diffusion theory am used as seminal fluid from 22infertile males were mathematical models. analyzed. Potassium, calcium, iron, copper, and Presented results of imagingsimulations using z.J.nc were detected and quantitatively assessed, both models quantifythesensitivity of breast minimum detectable limits being 1 ppm (Fe, Cu transillumination imaging to geometry and naturr! and Zn) and 50 ppm (K and Ca). A close linear of theabnormality within the breast, wavelength correlation between Zn,K and Ca levels was of light, extent of scatter rejection and method found, regardless of the pathological condition. of detection.

.Ve)r 1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas 143 Medical Physics Scientific Papers

MP36.3 MP36.4 Light Scattering and Propagation in Tissue, H Key*, P C Jackson and Elastic Characteristics of the Self- P N T Wells, University of Bristol, Expanding Metallic Stents, Bristol BS1 6SY, UK B.G. Fallone, Diagnostic Physics University of Texas, M.D. Anderson The development of improved light transmission Hospital & Tumor Institute Houston, TX 77030 imaging requires more information on the scattering and propagation characteristics of tissues at visible and near infra-red wavelengths. We are especially keen to exploit the Elastic properties of the self- expanding full potential of transillumination imaging for metallic (Gianturco) stent were measured. Hooke' the detection and diagnosis of breast cancer. law was used to characterize the stent withrespect to its caliber, length and radius bya coefficient of stiffness. Volume scattering functions of normal and Although this coefficient was indep'ndent of the radius of the stent, pathological tissues have been measuredat we found that it was directly dependent several wavelength bandsthat correspond to on its caliber. high tissue transmittance. Experimental data For a stent of a particular caliber and have been incorporated into a Monte Carlo model length, the fractional change in radius determined the force exerted by the stent. for simulatingthepropagation of lightin For small three-dimensional tissue geometries. The model radial displacements of the stent, the forcere- is being used to predict transmitted intensity quired to compress it was directly proportional to the radial displacements; for large displace- distributionsfor different geometries and tissue albedos and to study the efficiency of ments, the force depended upon a power series of spatial and temporal image filtering the fractional radial displaL.ment. A hyperbolic techniques. function was empirically introducedto approximate this type of relationship between force and radial displacement.Tension and pressure exerted by the stent were calculated and the use of such results with normal vessels was suggested.

MP36.5 MP36.6

Influence of EEG by Static Magnetic Theprnpagation of light through Fields. L. von Klitzing*, G. Tessmann, biological tissue, J. Bews*, Med.Univers.Liibeck,D- 2400 Lubeck, S. Shalev, Manitoba Cancer Treatment and ResearchFoundation, Winnipeg, W. Germany Canada R3E OV9

Recently we reported about changes of somatosen- soric evoked potentials (SEP) by static magnetic A detailed study or light propagationthrough biological tiss.ie hasbeencarried out under fields used in clinical n.m.r.-equipments (von '20 min, 0.2 1?s- and Diffusion theories in regions distant "romthe source. The parameters required to generate the Ia), because the energy values are not decreasing theoretical distributionsare the meanfree path immediately. (MFP) and the scatter /attenuation coefficient (c) for the medium. We found that MFP for TEM varies from 0.206 to 0.495 mm over the spectralrange 550 to 900 nm, and that c is close to unity. Linear Transport theory was uses to demonstratehow light propagation through the medium dependson spectral variations in these parameters, and in particular the very critical dependence on the value of c. The results are interpreted in terms ofthe transport of light through tissue atdifferent wavelengths, and the effect of hemoglobin on the process. C 6 144 1988 World Congress on Medical Physicsw.fl"iomedicalEngineeringSan Antonio, Texas Medical Physics Scientific Papers

MP36.7 MP36.8 Microcompder Approach to Btomorphometry, THE GENERATION OF ULTRASONIC WAVES BY IONIZING F.Bersani, M.Nichelatti *, C.Pallotti, G.Pallotti, RADIATIONS, N.A. Baily*, University of California, P.Pettazzoni, San Diego, La Jolla, California 92093, USA Department of Physics, 40126 Bologna, Italy. Theories of radiation damage based on the assumption that high temperature thermal spikes were Stereo logy offers the microscopist a pratical methodol- produced in irradiated materials date back to the ogy by which parameters of structures can be inferred, work of Dessauer in 1923. Numerous investigations by calculations, from measurements performed on sec- followed, most of which showed indirect evidence tion. In the first part a system of data acquisition based for the presence of localized regions of high temperature on conventional etereological method (Point-Counting) Recent physical investigations involving the detec- is discussed. The morphometry program collects data tion and measurement of ultrasonic waves associ- from digitizer and calculates area, perimeter, volume ated with the irradiation of both liquids and and statistics. solids have shown conclusively that these result In the second part the image processing by a Real-Time flom high temperatures produced in small regions Video Digitiser Module for the IBM PC AT is discussed. (1 um3 or less). Calculations of temperatures and acoustic wave The digitised image is processed using a computer pro- pressures for various energy depositions in dif- gram that traces, by algorithms of various kind, the con- ferent size volumes will be presented. It will be tour of the object and extracts various features from the shown that utilization of the characteristics of image. The advantage of this device is that it measures this acoustic field (amplitude, spatial and time the micrograph features with greater precision that the distributions, frequency characteristics, etc.) will yield the ionization energy deposited, the manual point counting procedures and that one tracing direction of its development, track width (for operation yields several pieces of information. The discharged particles), and spatial distribution of advantage is that profile or features with density near to delta rays. the background may be missed, so that the information retrieved is not representative.

MP36.9 MP36.10 Clinical Measurements of Trabecular Bone Mineral Comparative study of the effects of differents Density in a Normal Population, and in a group insulines administrated by iontophoresis tec- of Osteoporotics, using the Ratio of coherent to nics.J.Bustelo: M.J.Medialdea and F.Vilaplana ComptonScattered Photons in the calcaneus; Avd. dela Marina,3-5.11007 CADIZ/I:SPADA. M.A. Greenfield, J.D. Craven'and S. Shukla, University of California, Los Angeles, The present paper treat of comparative study California, USA. in rabbits, of the effects of differents in- sulins administrated by iontophoresis tecnic Trabecular bone mineral density (TBMD) is s. the objetives of this investigation try - obtained in the calcaneus by measuring the ratio to evaluate wich type of insulin is the best of coherent to Campton scattered photons capable for this percutaneus application.Spe produced by acollimated beam of 60KeV photons cifically were developed the following expe- from an Am-241 source (44.4-GBq) with a riments: scattering angle of 71° and a highly collimated 1. Arch!ve that the insulin pass through the solid state Ge detector. Calibration standards skin "in Vivo" without it losing its biologi are comprised of varying amounts of ground bone cal activity, using an electrical field. mixed homogeneously ina volume of petrolatum. 2.To extend its activity. The method has an absolute accuracy of 5%, and 3.To ratiobalise the time of insulin aurninis has a reproducibility of 3.4%. To date tration. A study was carried out on a group of 125 - measurements have been made on 87 normal males mean TBMD = rabbits "in vivo".Were determinate the valu- (ages 22-77) with a 274 mg/cc es of glycemia after the insulin aplication (SD=46), and on 240 normal females (ages 18-85) TBMD= 250 mg/cc (0=40). In by means of iontophoresiz during 30 minutes. with a mean striking contrast the TBMD of 20 male and 9 female paraplegics are 162 mg/cc (0=56) and 138 mg/cc (SD=30) respectively. The TBMD of a group of 25 osteoporotic females is 189 mg/cc (0=33). Comparisons of these findings will be made with conventionalQuattitativ!, ComputedTomography (QCT) and Dual Ph-ton Absorptiometry (DPA).

1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas 145 Medical Physics Scientific Papers

MP36.11 MP36.12 Is Passive Microwave Radiometry (PMR) ComputedAticrotomography with Synchrotron a Medically Useful Diagnostic Tool? Radiation from the X-25C Beam Line at the A.B. Behrman*, and E.S. Sternick, National Synchrotron Light Source. Tufts-New England Medical Center, P. Spanne* and M.L. Rivers+,*Radiation Boston, HA 02111 Physics Dep.,Linkoping University,S-58185 Link6ping,Sweden,4Dep. Applied Science, Biological systemsemit thermally generated Brookhaven National Laboratory,Upton,N.Y. electromagneticenergy at all frequencies. In 11973, USA the microwave region, the intensity of this radiation is directly proportional to the absol- The very high X-ray photon fluence rates available ute temperatureof the emitting tissue. Since at synchrotron radiation sources now makes com- inflamed tissue, some malignant tumors, and puted microtomography with spatial resolution in organs with high metabolic rates have slightly the pm-range possible. Experiments at the X-26C higher temperatures than surrounding tissue, ex- microprobe beam line at the National Synchrotron ternal measurements of microwave emissions could Light source at Brookhaven National laboratory indicate the presence of underlying disease. have yielded images with spatial resolution down Whether this technique will prove useful, to about 20 pm and show the potential to improve depends upon a varietyof parameterswhich the resolution down to at least1 pm in imaging include the temperaturediscrimination and small dead objects. This creates a new microscopy spatial resolution of the microwave radiometer, method for imaging microstraetures in tissue tissue thicknesses, and the frequency of the samples. Since no cht.aical preparation is required microwaves being detected. Using two independent such a development will for the first time make tImo atical approacheswehavecarried out possible non-destructive studies of tissues. detailedcalculations of microwaveemissions Calculations show that CT imaging of small fromseveral types of layered biological tissue laboratory animals can be done without inducing interfaced with direct and remote sensing radio- radiation injury at 50 pm spatial resolution. By meters to investigate these factors. Our results use of a line-shaped radiation beam, a photodiode indicate that PMR is a promising diagnostic tool array detector, and CAMAC acquisition modules, an for a number of medical applications. image can be obtained in about a minute. Research supported by NTH USPHS Fellowship FO- TW03735, RR01838,DOE DE-ACO2-76CH00016, MFR 7557

MP36.13 MP36.14 A Methoa for RemovingProlongation of Fluid Therapy Consultation System(FLUIDEX) Mean Transit Time due to Response Time, M.Arakawa., K.Kambara, T. Segawa, Shoichiro Hareand Toshiyuki Furukawa H.Miyazaki,F.Ando ana S.Hirakawa, Gifu University School of Meaicine,40,Tsukasa- Institute of Medical Electronics, machi,Gifu 500,japan. Faculty of Medicine, University of Tokyo, 7-3-1 Hongo, Bunkyo-ku, Tokyo 113

Tne response time of tne indicator FLUIDEX is an expert system which can help in detecting system distorts dilution curve the diagnosis of fluid imbalance of thepatient and prolongsmeantransit time(MTT). In as wellas recommend the optimal prescription the thermal-dyetechnique for measuring for the infusion therapy. The characteristics of lungwater, correctingmethod forap- the system are: (1) a hierarchical inference parent prolongation of MTT associated engine which acts like a filter to selectthe with response time is unsettled. We most appropriate hypotheses from those initially obtained step function response curves of generated; (2) a knowledge base which can 5 differentthermodilution systems aria convert qualitative characteristics of the measured mean response i.imes(MRT). Using patientinto quantitative ones or compensate model witn flow loop system fillea with deficit data; and (3) a simulation model of the warmed 0.9% saline solution that Is electrolytes andwater metabolism. FLUIDEX's circulated at 6 different flow rates, we evaluation using actual patients has simultaneously obtained a pair of thermo- demonstrated its usefulness in the medical dilution curves from a pair of thermodi- environment. lutionsystems bothpositioned at tne FLUIDEX can not oaly propose thediagnosis same place. Multiple combinations of andprescription, butalsooffer simulation thermodilution systems and flow rates are facilities which are able to trace its made. Differences of MTTs(uMTT,y) yieldea diagnostic process using its input data and give a close approximatten of differences of computer -aided instruction of fluid therapy. MRTs( 4,MRT,x), y=1.17x0.04 (r=0.95, n=72). Therefore, we can remove apparent prolongation of MTT by subtracting MRT 1 r8 from MTT.

146 1988 World Congress on Medical Physics and Bimedical EngineeringSin Antonio, Texas Medical Physics Scientific Papers

MP36.15 MP36.16 Biological Soft X-rav Contact Microscopy: A New Triple-Pulsed X-ray Generator Having Variable Look at the Living Cell, G. D. Guttmann, Center Spectra, E. Sato*, H. Isobe and T. Yanagisawa, for X-ray Optics, Lawrence Berkeley Laboratory, Department of Physics, Iwate Medical University, Berkeley, CA 94720, U.S.A. 3-16-1 Honcho-dori, Morioka 020, Japan

Within the past decade, several technological Theconstructionand the fundamental studies for advances have contributed to the increased success the triple-pulsedx-ray generator having variable of imaging living cells by soft x-ray microscopy. spectra for biomedical radiography are described. The easiest approach to this technique is the This generator consisted of the followingcompo- contact method, which is analogous to x-ray nents: a high-voltagegenerating unit, a voltage lithography. The contrast mechanism is due to divider unit, three high-voltagepulsers, a absorption at the carbon and nitrogen K edges. triple-parallelimpulse switching system, a high- The region of interest, also known as thc. "water power gas diodehaving threeterminals, aturbo window", is 23.6 A

MP36.17 MP36.16 Reactions in Concepts of viscoelasticity A Study of Photo-Nuclear and thixotropy of blood. Tissue, P.D.Allen, M.A.Chaudhri*, Australian Radiation Laboratory and J.C. LELIEVReand C. LACOMBE. Austin Hospital, Melbourne, Australia. Unite de Biorheologie, CHU Pitie-Salpetriere, 91 Bd de l'Hopital, 75013 Paris, France. The total yield of photoneutrons from soft-tissue equivalent material, irradiated with bremsstrahlung of end At a normal hematocrit, blood is a point energy from 6 MeV to 28 MeV, has thixotropic gel. At rest, this gel exhibits been measured using the 35 MeV University viscoelastic and plastic behaviors. of Melbourne Betatron and a 477.Halpern- type neutron detector. The results are First, one indicates the principles which compared with calculations based on the underliethe generalfo"mulationof the cross-sections for neutron production of constitutive law of thixotidic fluids. constituent elements of tissue (H,C,N and 0). It is shown that the less common Then, one precises the experimental methods isotopes of the constituent elements (H- which are most suitable for characterizing 2, C-13, N-15, 0-17, 0-18) must be this law (hysteresis loopsmethod, step included to obtain agreement with the method).Different examples ofthixotropic measured yields. The dose equivalent behavior are presented (blood, fuel oil, Na- imparted to tissue by thephotoneutrons bentonit, etc). Theoretical methodsare is estimated for the energy range from 6 briefly discussed. Thesemethods lead one to MeV to 30 MeV and compared to other consider elastothixotropy as non-linear measurements/estimatiion (where viscoelasticity. available). Furthermore, the dose equivalent due to all photonuclear Finally, examples of experimental set-up are reactions is estimated for energies of 24 given and related difficulties are discussed. MeV and 30 MeV. Applications to bloodand to redcell suspensions are indicated and discussed in terms of RBC aggregation. 189

1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas 147 Medical Physics Scientific Papers

MP36.19 MP36.20 Fluoride Determination in Bones and Teeth of Experimental Animals, Radiographic Characteristics of a Single-Shot M.A.Chaudhri *, Flash X-ray Generator Having Three Types of Austin Hospital and University of Melbourne, Melbourne, Australia. Tubes, H. Isobe*, E. Sato and T. Yanagisawa, Deparment of Physics, Iwate Medical University, 3-16-1 Honcho-dori, Morioka 020, Japan In a programme to study the effect of water fluoridation on calcified tissues, the F concentrations of the bones and Theconstructionand the radiographic character- teeth of experimental animals, whichwere istics fora single-shot type of high intensity kept on fluoridated water were determined flash x-ray (FX) by prompt activation analysis. generator is described. This generator consisted of the following Powdered bone samples were pressed components: a into control board, two high-voltagepower supplies for the grooves of the aluminium target positiveandnegative voltages, a high-voltage holder, while whole teeth were used as coaxialoil condenser of 120kV-0.2pF, a gas gap such, for irradiation purposes. The switch witha highcurrentcapacityofabout samples and the standard were covered 100kA, a low impedancecoaxial transmissio line, with a thin aluminium foil and and three typesof FX tubes, irradiated, in turn, each of a different with 2 MeV protons type. The electric charge stored by the condenser at 20 nA intensity. The 6-7 MeV gammas was discharged to from the FX tube when the gap switch the F(p,00')0 reaction were counted was closed. For the Type A tube, a conical anode with a 5 x 6' inch NaI(T1) detector, andcathode electrode placed were employed in order to at 135 degrees to the beam obtain a smaller locus of less than 0.5mm direction. in diameter and to change the beam quality according to The bones of the animals who were given changesin the anode-cathode space. The Type B fluoridatedwater (250 ppm of F) for 15 tube utilizeda conical anode and 45 days, and a ring cathode were found to contain about for obtaining a higher intensity anda wider ener- 5 and 10 times more F than the bones of gy spectru: distribution. The Type C tube could control animals respectively. Similarly, produce slightlyhardFXbypermeating the teeth of the animal- on this fluoridated radiation through variouskinds of foil anodes. water (2 pp of F) contained twice as The cathodeelectrodewas changedto either much F as those of the control animals. a conical or a pipe shape according to the radiographic objectives.

MP36.21 MP36.22 A High-Speed Radiographic System Utilizing A High-Frequency Field-Emission X-ray Source A New-Low Impedance Flash X-ray Generator, for Soft Radiography, E. Sato*, S. Kawasaki, S. Kawasaki*, E. Sato, H. Isobe and H. Isobe, J. Takahashi, Y. Yasuda, H. Arima, T. Yanagisawa, Department of Physics, Iwate T. Akitsu, S. Oikawa and T. Yanagisawa, Medical University, 3-16-1 Honcho-dori, Department of Physics, Iwate Medical University, Morioka 020, Japan 3-16-1 Honcho-dori, Morioka 020, Japan

Ahigh-speed radiographicsystem utilizing a new Afield-emission x-ray source for soft biomedical type of single-shot flashx-ray generator with radiography is describe( . Thisx-ray source was exposure times ofabout 1ps isdescribed. This constructed in orderto obtain a higher dose rate system wasconstructed in order to analyze the using a cold cathode x-ray tube without a tungsten high-speed motion of biomedical objectsand filament. Thisx-raysource consistedof the consisted of the following components: two types following essential components: a DC-AC converter of delay switches, a delayed pulsegenerator, a with an output voltageof 1 to 10kV anda,rec- digital radiography (CR), anda flash x-ray tangular AC frequencyof 2-20kHz, a Cockcroft generator developedby the authors. The flash x- circuit, a fieldemission x-raytube with a rod ray generatorconsisted ofthe followingcompo- anode and a ri(ig cathode, and a vacuum pump. The nents: a high-voltage generator, ahigh-voltage tube voltage and the currentwere determined by pulser, a vacuum pump, and a combined remotex-ray the output voltageof the Cockcroftcircuit and tube withtwo sets of anode and cathode elec- the circuit resistance of the discharge line.The trodes. When the radiographic object trigge-id spectrumdistribution primarily varied according the delay switch, a shortelectric pulsewas to the tube voltageand the anode material. The produced and transmittedto the delayedpulse effective focalspot size wasdetermined bythe generator. An accuratedelayedflash x-ray was diameter of the anode rod and ranged from 0.5 to produced when the delayed eulse was transmittedto 3.0mm in diameter. The electric wave form of the the high-speedimpulse switching system inside of tube voltage, thetube current, and the x-ray the high-voltage pulser. Next, when the radiogra- output displayed a high frequencyripple corre- phicobject was exposedto the controlled x-ray sponding to the A-C frequency, but the deviation under the optimum radiographicconditions, the of te x-ray intensitywas a only few percent. permeating x-rays prodmed thedigital image. Various kinds of soft radiographs were obtained by controlling the radiographic conditions. 170 148 1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, 'texas Medical Physics Scientific Papers

W36.2:; MP36.24 Investigations on Developing Mouse Testis after Neurochemical Studies on Developing Mouse Brain an intrauterineLow-Level Tritiated Water (HTO) undercontinuous tritiated (111.0) exposure frcm Exposure. Sangeeta Pareek*and A.L. Bhatia,Depart- 17th day in utero. A.L. Bhatia* and Sangeeta ment of Zoology, University ofRajasthan, Jaipur - Pareek, Department of Zoology, University of 302 004, India Rajasthan, Jaipur - 302 004, India. Tritium, a radionuclide of hydrogen, may emerge as The effects of low doses of tritiated water on the fromthe a major radiopollutant of prime concern neurobiological system have not been sufficiently view point of health hazards. It is also used in explored, though it has been demonstrated that the determination of total body tracer study for after a protrated HTO-exposure, tritium (.511) is tritium in the gene- water. The incorporation of retained in larger amounts for longer duration in constituents (DNA) isof prime concern since tic mouse brain. F-esent study is an attempt to evalu- its turnover is expected only in the event ofcell ate the total cholesterol, glycogen and phospho- death. An attempt has been made in the present lipid of the brain during postnatal development to evaluate the changes in total choleste- study under a continuous tritiated exposure at the dose content in the postnatally rol, protein and DNA level 0.3duCi (11.1 kBq)/ml. of drinking water developing mouse (at 4 and 6 weeks) testis under from gestation day 17. Cholesterol level showed exposere at the dose level a continuous tritiated week which gradually in- kBq) /ml of a deficit by 20.92% at of 0.3juCi (11.1 kBq) and 3.01uCi (111 creased and reached at peak level by 3rdweek of drinking water from gestation day 18. Cholesterol age. Thereafter it decreased and achieved a steady content increased significantly at both the dose state after 4 weeks approximately 27.0% below the Protein concentration, how- er, decreased levels. control value. An elevated level of glycogen after irradiation. DNA contentincreased at 4 content was marked till 5th week pp p. which dec- did not show any significant change at weeks, but reased at 6th week of age. Phosphblipid level was 6 weeks after the exposure at both the doses. found decreased in all the studied age groups Besides this, the behaviour of spermatogenic cell which was pronounced at the early intervals (1- populations viz. Spermatogonia A,I,B; Spermato- 3 weeks). The results will be discussed in the cytes R,L,Z,P and Spermatids have al-o been eva- light of tritium incorporation and metabolism of Results will be discussed iii the light of luated. various neurochemical constituents. metabolism of tritium with special reference to testis.

HP37.1 MP37.2 HOLOSPECTRAL IMAGING IN NUCLEAR Chest-Imaging Using MEDICINE. D. Gagnon. MontrealHeart Self-Scanning Linear Diode Arrays. Institute, Montreal, QC, Canada. D Sashin*, BS Slasky. University of Pittsburgh, Pittsburgh, PA, 15261, USA Holospectral Imaging (HI) is a novel approach to digital imageprocessing in nuclear medicine to A high detail diode array digital chest unit has eliminate contamination resultingfrom Compton been developed which produces digital images ha,- scattering. Themethod requires that the energy ing a matrix of 2356 x 2000 pixels with a 14 x 16" information is made available for all photons, field of view. This technique achieves high de- irrespective to their primaryor scatter chara^- tail through the use of self- scannir, linear diide teristics. From the HIdata, a set of frames arrays (1024 /inch) that have a wide dynamic range representingimages a'different energies is and high spatial resolution, which separates this created which allows .ar amultidimensional method from other digital radiographic techniques. description of the imaging processes. In fact, by A very significant feature of our system is the multivariate analysis, the original data can be rejection of scattered radiation because of the transformed into a series ofeigenimages and fan geometry, resulting in increased contrast sen- eigenvalues that made passible the distinction of sitivity at a lower radiation dose (5 and 10 mR the specific contribution to theimage of both for PA and lateral views, respectively) than that unscattered and scattered photons, and noise. The obtained by area imaging techniques. The entire first eigenimage corresponding to the unscattered image of the chest is recorded in less than 2 sec- image, and the second the image of what cannot be onds with an exposure time of less than 3 ms. The explained bythefirst namely, the scattered present machine upgrades our previous model: a) image. The image of the Poisson noise, unstruc- the x-ray tube and detector assembly is moved; b) tured by definition, is represent by smaller erect imaging can be taken; c) the field of view eigenimage. A filtering process can be performed has been increased; d) the matrix size has been then to reduce the scattered contribution. greatly expanded; and e) a higher resolution rare Some holospectral images, fromcomputer simula- earth phosphor screen. After reconstruction, tion, phantom and from patient, will be presented the images are digitally processed to improve and interpreted with the multidimensional image quality. language. Preliminary results ofthe Compton removal capabilities will also be shown. 171 1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas 149 Medical Physics Scientific Papers

MP37%.3 MP37.4 Optical Image Processing with Liquid-CrystalDisplay On the automatic assessment ofImage Qualityin for 'ImageIntensifier/Television(II/TV)Systems: Digital Subtraction Angiogra5Hy DynamicRange Compression andSignal-to-Noise C.H.Slump, PhilipsandPicker MedicalSystems, Ratio Vmprovement. K. L. Lam*, H.-P. Chan,K. Doi, Eindhoven, the Netherlands. Kurt Rossmann Lab. for Radiol. Image Research, Dept. The important image quality parameters in DSAare: of Radiology, Univ. of Chicago, Chicago, IL60637 spatial, (low-) contrast and temporal resolution. Airthermorecontrast and spatial uniformity, We are developing an optical image processing contrast linearity, subtraction artefacts, (01P) device using a liquid-crystal display (LCD)to radiation exposure levels and noise. modify the transfer characteristic ofan II/TV imaging The image quality performance is usually evaluated system and to perform unsharp maskingon the output non-invasively by means of appropriate phantoms. image from the II.The LCD which is driven at TV The pertinent images are then visually inspected. frame rate is placed in the optical pathbetween the II Resolutiontest patternimages (line-pairs per and the TV camera, and is displaced from thefocal millimeter) and contrast-detail (C-d) images are plane of the optical system in order toblur the famous for their inter- andintra-observer displayed mask image.By controlling the transmit- fluctuations in measured values. tance of the LCD according to the outputimage from Theautomatic i.e. theobserver independent the II, we can compress the dynamicrange of the assessment ofthe image quality parameters from image detected by the TV camera. Ourexperimental thephantom imageshasgreat advantages for results demonstrated that compressing thedynamic quality assuranceprograms. The ease of such a range of low-quantum-noise image with the 01P device procedure for routine checks, and the improves the signal-to-noise ratio of thedetected possibilities for additional computer aided tests image.Furthermore, by an appropriate modification (expert system search trees) in case cf of the transfer characteristic, the detectivequantum malfunctioning may also have its impacton remote efficiency over the entire imagecan be maintained at serviceability. the level of the high-exposure area of the image. As example of the above approach thepresentation will focus on the automatic evaluation of the DSA Supported by USPHS Grant (NIRA) HL 36238,The system's low-contrast resolution. The procedure Whitaker Foundation, and USPHS Grant CA "4806. is basedonsignal-detection theory and image processing. A test-statistic is calculated which indicates the likelihoodof the presence of a certain contrastat specified false alarm and "miss" probabilities.

MP37.5 MP37.6 Large Field K-edge Subtraction System for FAST AUTOMATIC VESSEL TRACING AND STENOSIS Coronary Angiography using SynchrotronRadi- EVALUATION IN DIGITAL SUBTRACTION ANGIOGRAPHY ation III: High Speed Aquisition of Video USING SPECIAL PROCESSORS AND PARALLEL ALGORITHMS. Images for the Energies above andbelow K-edge K.Nishimuraf%K.Hyodo,R.Hosaka,M.Ando,F.Toyofuku, K. Barth; U. Bittner, B. Eicker, P. Marhoff, M.Akisada,S.Hasegawa,E.Takenaka,Saitama Medical SIEMENS AKTIENGESELLSCHAFT, Med. Engineering Group, School,Moroyamajruma,Saitama,Japan Henkestr. 127, D-8520 ERLANGEN, West Germany We are developing a image aquisition system for In the last years fully digitalangiographic K-edge subtraction coronary angiography using imaging systems at high resolution have beanintro- I.I. and video cameras. To obtain a K-edge sub- duced to clinical application. MaLrix sizes ofupto traction image of movingcoronary vessels as 1024 pixels squared and amplitude discrimination a quasi still image, the energy must be changed with 8 to 10 Bits are available asa standard now. rapidly in the order ofa few msec. Because This opens a new perspective for quantitative the two images appear in theoutput screen of assessment of fine image structures like stenotic I.I. subsequently within a few msec, the images lesions of the coronary arteries.However, the cannot be stored in different TV fieldsbut in algorithms known from previous systems with cine- a same TV field by a conventional method. film quantitation or with lower resolution have To solve this difficulties to we designed the system be basically revised and improved withrespect to consisting of a half mirror,a couple of video accu:acy and speed of computation. Research on the camera with optical shutters which were made best way for automatic tracing of vesselcenter- from PLZT. The optical output of I.I. was divided lines and subsequent border calculation hasbeen into two directions by the half mirror. Ihe carried out in this laboratory investigation. This shutters can be made open and close by applying has lead to refined kernel design for the tasksof a voltage to nem. With proper timingsignals, vessel detection and geometric or densitometric a video camera can see only the image abovethe measurement. Globally correct performance is K-edge and the other can see only ,he image below achieved with dynamic programming on the basis ofa the K-edge. Blanking techniquew..s uscd to avoid minimal cost algorithm. Filtering including median the adverse effect caused bysimultaneous writing filtering and kernel convolution has been imple- and reading of video signalon a video camera. mented wi special hardware, mostly with parallel The shutter were found suitable for our purpose algorithms, providing fast computationeven at the with respect to response time anddynamic range. langetroatrix size. J.

ISO 1988 World Congress on Medical Physics rnd BiomedicalEngineeringSan Antonio, Texas Medical Physics Scientific Papers

11P37.7 MP37.8 Topological Image Compression, H. Riml* Minimizing Artifacts In Eigenimage Filtering (1) McGill University, Montreal H3A 2B4, Canada J.P.Windham D.A. Reimann and D.O. Hearshen, Henry Ford Hospital, Detroit, Michigan 48202. Digital images typically contain one megabyte of data. The storage and transmission of such large Eigenimage filtering is a linearfilter which data sets strain the capacity of current tech- produces a composite image from a sequence of nology. Thus there is need to exploit structural images in which a desired feature is suppressed. properties of these data sets in any attempt to A potential source of error in the interpretation compress them. We present: of filtered images are projectionartifacts. 1) A FRAMEWORK which allows a classification of Prejaction artifactsare the result thatany traditional approaches to image compression, and pixelvectormay be represented as a linear arising from this framework combination of the eesired and interfering 2) A new technique called TOPOLOGICAL IMAGE COM- signature vectors, and a specific vector which is PRESSION. It uses homeomorphisms ("rubber-sheet- orthogonal to the subspace, R(d,u) containing the ing" transformations) on both geometrical and desired and interfering signature vectors. That brightness information of a "reference image" so component of thepixel vector parallel to the that it becomes identical to the "target image". desired signature vector will be projected into represents real 3) An ALGORITHM which implements topological the filtered image, and information or isan artifact.A method which image compression minimizes artifacts has been developed. This 4) : preliminary analysis of the EFFECTIVENESS method determines a vector which is orthogonal to of the algorithm for radiological images, which the subspace, R(d,u). This vector is used to form shows that for this very specific application a filtered image that contains only those pixels the algorithm outperforms all others by at least whose pixel vectors have a component orthogonal an order of magnitude tothe subspace, R(d,u). This image is used to 5) A demonstration that a failure of the algo- form a mask image. Any pixels contained in the rithm to work for "reasonable" distortion coef- mask image are not retained in the filtered image ficients is a strong indicator of potential medi- since they may represent an artifact. The cal lesions. In this "MASKLESS SUBTRACTION RADI- technique is demonstrated for simulated and OLOGY" the distorted reference image becomes the Magnetic Resonance Image sequences. mask image for the target image, and the difference between them the potential lesion. (1) J.P. Windham, et al, "Eigenimage Filtering in MR Imaging", JCAT 12 (1): 1-9, Jan./Feb., 1988

MP37.9 MP38.1 Digital Subtraction Angiography Using an ISODOSEAND ISOEFFECTDISTRIBUTION Efficient Method for Scatter Rejection, FOR BRACHYTHERAPEUTIC TREATMENT D. Plenkovich*, Department of Radiology, PLANNING OPTIMIZATION,A. M.Di Halle, Tulane Medial Center. New Orleans, LA L.Begnozzi, L.D'Angelo and M. Benas- si, Regina Elena National Cancer Scattered x-ray radiation and veiling glare Institute, Rome,Italy. reduce contrast and signal-to-noise ratio in digital subtraction angiography. The use To improve the therapeutic ratio and to of grids is trade-off between lesser amount reduce late effects in critical organs a of scatter and reduction in the percentage treatment planning optimization is o1dateCted primary radiation, thereby in- needed. Tumor control probability and creasing patient exposure. Grids, air gaps, complications depend on radiobiological and paired scanning slits are not effective andphysical factors. An attempt to against veiling glare because veiling glare represent the effect due to these fac- occurs afterradiation has passed through torshas been done; since simple dose any of this antiscatter devices. Highlyef- computations (Gy) are not suitable, ficient rejection of both x-ray scatter and particularly for combined modality veiling glare is achieved usingmultiple treatments, where intraca-vitary irradia- scanningfore slits and computer algorithm tion is used with external beams, an which keeps only portion of eachdigitized optimized treatment is possibleonly image in t1, projection of the slits. The calculating the "biological effect". resultant image is obtained adding a series A computerized program which worlts on a of slit images together. Digital subtrac- PDP11/34 and on an IBM compatible PC has tion angiography of a dogwas performed been created. It determines isodose and using 8:1 grid and combination of same grid isoeffect distributions for high dose and scanning-slit method. Visualization of rate brachytherapeutic sources. Treat- small blood vessels is better in the slit- ment parameters are of maximal importan- plus-grid image due to superiorrejection ce because of possible complications in of scatter and veiling glare. The digital critical organs observed with high dose scanning-slit technique increases the sig- rate irradiation technique. nal-to-noise ratio in subtracted image of a 20-cm-thick polystyrene phantom by 50%. 173 1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas 151 Medical Physics Scientific Papers

MP38.2 - MP38.3 A Critical Comparison of High and Low REPLACEMENTCORRECTION FACTOR IN BRACHYTHERAPY Dose Rate Remote Afterloading, Colin G. PHOTON DOSIMETRY BY IONIZATION CHAMBERS. Orton*, Gershenson Radiation Oncology G.Arcovito*, A.Piermattei,F.Andreasi-Bassi Center, Harper-Grace Hospitals/WA7ae Medical School, Catholic University, Rome, Italy State University, Detroit,MI 48201 Acorrectionfactor for absorbed dose measure- The recent surge in interest in high dose ments by means of ion chambers at distances-to- rate (HDR) remote afterloading and the availa- source of the same order of chambers' dimensions, bility of HDR equipment make it important to cri- as in phantom dosimetry of brachitherapy sources, tically compare HDR techniques with the low dose must bedetermined to account for effects rate (LDR) treatments they due are replacing. Even both to the replacement of phantom material by though the HDR modality has been used for many cavity air and to finite size of ion chambers. years to treat a variety of diseases, little We report measurements by four experience has been gained commercially with these techniques available ion chambers made in phantom, compared with conventional at LDR brachytherapy. distancesranging between 1cm and 10cm Hence there is a reluctance from a on the part of many 192Ir point source of 2mm in diameter and 1mm in radiotherapists to even consider adoption of this height. Absorbed dose values were obtained from new technology unless convincing evidence can be ionchamber readingsusing the Spencer-Attix given to show that it represents an improvement. formulation without introducing correction In this paper a critical comparison has been factors as mentioned above.Deviations made between these two from modalities of radio- expected dosevalues up to 20% were therapy, taking into found. We account both physical and present a formulation and a calculation model biological considerations of and, if available, a correctionfactor of chamber clinical results. readings which In order to quantitateany resulted dependant, at the shortest distances-to- potential benefits derived by conversion '1 the source, alsoon these distances, while at the HDR technique, a termtherapeutic advantage is proposed. farest ones it is dependent only on the chamber's Values of the therapeuticadvantage diameter and have been calculated photon energy. A good agreement using various bio-effect between our data dose models and these and results reported in the show that significant literature for longest distances -to-source advantages can be achieved was if the HDR treatments found. are sufficiently fractionated.

MP38.4 MP38.5 Continuous Vs. Discrete Sources in the Planning and Evaluation of Stereotaxic Brain- Paris System of Implants, J. M. Paul*, Tumor Implants of 1-125 Seeds, L.L. Anderson*, R. F. Koch, P. C. Philip and F. R. Khan. A.D. Osian and P.B. Sher, Memorial Sloan- NWCH, 800 W Central Road, Arlington Kettering Cancer Center, New York, NY 10021. Heights, Il 60005 An approach to planning stereotaxic 1-125 Since the Paris system of interstitial brachy- brachytherapy of brain tumors has been devel- therapy is based on iridium wires, the oped which involves least-squares optimization consequences resulting from the substitution of single-seed tube positions within the tar- of discrete point sources in afterloading get contour, followed by a combining of seeds ribbons have been analysed by the authors from neighboring tubes in order to reduce the for single ribbons, and single and biplanar total number of tubes. Seed orientation, which implants. The results of this analysis show Must be specified to include dose anisotropy, that the homogeneity parameters, suchas: is controlled by having all tubes emanate from total activity, treatment volume, total volume an extra-cranial virtual entry point. Typi- that receives over twice the stated dose(rate), cally, a standard deviation of 12-15% achieved average dose rate, etc., for seeds with among target-contour (CT enhancement margin) customary spacing of about 1 cm are comparable points for single-seed optimization is to those using wires. It is also found that increased to 15-20% in the final multi-seed 1 cm seed spacing often improves the quality tube configuration. Dose evaluation from of the implants, as measured byan objective post - implant CT images makes use of tube-air parameter called uniformity index, which takes as well as seed image data and may involve into account both the dose homogeneity and transformation of achieved seed locations total treatment volume within an implant. back to pre-implant scans on the basis of This paper is intended to dispel,any belief, specified anatomical points within the head. based on visual effect of isciose curves around discrete sources, that implants with line sources (wires) produce better dose homogeneity.

174 152 1988 World Congress on Medical Physics andBiomedical EngineeringSan Antonio, Texas Medical Physics Scientific Papers

MP38.6 MP38.7 Theoretical Evaluation of DoseDistributions in Interseed Effects on Dosefor 1-125 and Water about models 6711 and 6702 1251 Seeds. Ir-192 Implants, Ali Meigooni*, Jerome J.F. Williamson* and F.J. Quintero, Dept. of Rad. Meli and Ravinder Nath, Yale University Oncology, Univ. of ArizonaTucson, AZ 85724. School of Medicine,Dept. of Therapeutic Radiology, 333 Cedar St., New Haven, CT, We have used theM6nte Carlo method to calculate 06510, U.S.A. the two-dimensional dose distributions about the models 6711 and 6702seeds from first principles. In brachytherapy treatment planning,dose Our code models the detailed internal structure of at a pointis calculated asthe sum of the the sources as well as photonabsorption and dose contributed by each individual source multiple scattering inthemedium. At distances in a uniform medium. However, a typical implant of 2cm to 5 cm, our calculated data are in has many sources closeto each other,so that excellent agreement with the measurements of Lingl primary and scattered radiation from one source and Sche112. In contrast to their results, which reaching apointcould be affected by the were limited todistances greaterthan 5 mm, we presence ofthe other sources. Importance of report artifact-free dose estimates for distances this interseed effect was evaluated for two ranging from 0.1 cm to 7.5. A truncated Fourier dirferent multiseed configurationsin a solid series is used to describe the Monte Carlo data in water phantom: a)for 3 1-125 (Model 6702) terms of asmall number of measured coefficients seeds, aligned with their longitudinal axes allowing our data tobe accurately reproduced for coincident, it was found that the dose along the treatment planning purposes. Alimited set of axis ofthe seeds equals the sum of the doses experimental data was foundto be very useful in from theindividual seeds, However, for the 3 optimitizing and validatingthegeometric model seeds, aligned with their transverse axes assumed by ourcalculationsin the absence of a coincident, the dose along the transverse axes of sufficiently precisedescription of the seed the seeds is about 8% lower than the sum of the structure. These validated geometric models can doses from the indiNidual seeds; b) for a 2-plane be used with confidence to investigate more implznt, each plane containing a 3 x 3 array of dosimetry of eye complex phonomcna such as the sources in a 1 cm square grid, the results show plaques or of tissue heterogeneities. no interseed effect for Ir-192 sources and a 10% and 202 effect for 1-125 Models 6702 and 6711 1Med Phys 12: 652-655. 2Int J Radiat Oncol Biol sources, respectively. Phys 13: 795-799.

MP33.8 MP38.9 Evaluation of Planar Ir-192 Implants Dosimetric Parameters for 1-125 and Ir -192 thing the Three Volumetric Sources, K. Weaver*, C. Barnett, D. Huang, Irradiation Indices, C. B. Saw* and C. Ling, M. Schell, V. Smith, Radiation N. SunthuAdingan, Department of Oncology Department, University of Radiation 'Therapy and Nuclear California, San Francisco, 94143. Medicine, Moms Jefferson Uhiversity Hospital, Philadelphia, One task required by an NCI Brachytherapy PA 19107 Contract was to evaluate published dosimetricparameters for commonly used this The clinical approach for interstitial implants brachytherapy sources. To fulfill usually is an attempt to deliver a specified dose requirement we investigatedIr-192 seeds of 1-125 seeds. rate, typically 1000 coy /day, and also achieve an and two models acceptable dose Imogeneity in the treatment plane Measurementswerewith LiFpowder in located at the periphery of the target volume. In water-equivalent phantoms. Data were this study, planar Ir-192 implants configured w ,,.et corrected for background, sample mass, and volume. Results will be the desired clinical goals are qur.ntitatively finite detector evaluated in terms of the three volumetric, irradiation discussed in terms of a particular dose The planar implants are cc:,structed with parametrization, and willbe compared to indices.' as ribbons arranged in parallel, comprising of either MonteCarlo calculations aswell results agreed well single or double planes. Each ribbon is made up of published data. Our for relative uniform activity seeds spaced1.0 an apart. These with published data implants are used to treat 1.0, 2.0 and 2.5 an thick quanUties such as radial and angular dose The irradiation indices measure the dependence. Our measured value for the Ir- target volumes. 4.55 cGy cm2 effective coverage of the target volume, the relative 192 dose rate constantwas alsowas as expected. dose homogeneity within the target volume, andthe mCi-1 hr-1, which However themeasureddose rate constants volume of normal surrounding tissues being irradiated for 1-125 seedmodels 6702 and 6711 were above the prescribed dose. The optimalribbon derived from 1.16 and 1.05 cGy cm2 mCi-1 hr-1, values spacing for different size implants at well below those in general use. such an analysis. This worksupported byNCI Contract CM- 'C. B. Saw et al., Pied Thy la, 580, 1986. 57708.

175 153 1988 World Congress on Medical Physics and Biomedical Engineering San Antonio, Texas Section III

Biomedical Engineering Scientific Papers

176 Biomedical Engineering Scientific Papers

BE1-A.1 BE1-A.2 Channels as Enzymes. R.S. Eisenberg; Department STRUCTURE AND FUNCTION OF MEMBRANE CHANNELS. CHEM- of Physiology, Rush Medical College, 1750 W. ISTRY AT THE SINGLE MOLECULELEVEL. Olaf S. Harrison, Chicago IL 60612, USA. Andersen and Roger E. Koeppe II.Dept. Physiol. Biophys.. Cornell Univ. Med. Coll., New York. NY Ionicchannelsareproteins embeddedinbiological 10021. and Dept. Chem. Biochem.. Univ. Arkansas. membranes containing an aqueous tunnel that allows ions Fayetteville. AR 72701. to pierce the energy barrier of the low dielectric lipid of The function of transmembrane channels is the membrane. Ionic motion is driven by the gradient of determined by the amino acid sequence of the chan- electrochemicalpotentialacrossthe membrane and nel-forming proteins and by thethree-dimensional modulated by the potential of interaction between the ion folding of the sequence in the membrane. But it is and the channel protein. not obvious how functional properties of membrane It is useful to think of channels as catalysts for channelsshouldbepredicted from chemical and diffusion, as enzymes specialized for passive transport, as structural data. To obtaininformation on this molecules which pose many of zle same barriers to problem. we have used channelsformed by the understanding as enzymes.Topics in enzymology have linear gramicidins to examine how the functionof analogs n membrane biology, e.g. an enzyme's active site a channel is affected by alterations in the chan- 4-4 the channel's pore; enzyme isolation separation of nel-forming peptide sequences. The gramicidins are conductancesorchannels;enzymeclassification suitable for such studies because the structure of channel nomenclature; enzyme specificity 4-4 channel the transmembrane channel is known. This is impor- selectivity; enzyme activation 4 channel modulation. tant fot two reasons: one can establish functional The fundamental mechanism of channelsdiffusion criteria for structural invariance among channels; through a potential fieldis more amenable to physical and one can use constraints imposed byknowledge analysis than the fundamental mechanism of enzymes. A the channel structure to guide theinterpretation randomwalkinapotentialisdescribedbya f functional results. Channels formed by gramici- FokkerPlanck equation that can be solved to give the din analogues illustrate how single sidechain rate constant of traditional state diagrams of channels. substitutionscan result in ten-fold conductance The predictions of these diffusion models can then be changes without affecting the channelstructure. comparedwithexperimentalmeasurementsofopen The sidechains donot contact the permeating channel current under a variety If conditions. ions. emphasizing that there neednotbeclose proximity between theregion wherea chemical alteration is produced and theregionwherethe functional perturbation is induced.

BE1-A. BE1-A.4 Sodium - Calcium Exchange in Heart: Enzyme-mediated Ion Transport Regulation by tGa2471., W.Gil Wier: J.Mailen Kootsey; University of Maryland at Baltimore, Baltimore, MD, U.S.A. Duke University Medical Center, Durham, North Carolina 27710, USA Sodium-calcium (Na/Ca) exchange is aprocess believed to be important in control of contractility and in normal and abnormal Several ion transportmechanisms known to contribute to electrogenesis in the mammalian heart. Recently, ion regulation and electrophysiology in nerve and mus- changes in [Ca2 +]i have been recorded that can be cle cells involve the binding of ions to a large protein attributed to the operation of an electrogenic, voltage-dependent sodium-calcium exchanger in molecule, followed by a conformational change in the mammalian heart cells. Singlevoltage-clamped enzyme molecule. Examples of such transport mecha- guinea-pig ventricularmyocytes wereperfused nisms include the Na +, K+ ATPase, the Ca++ ATPase, internally with the fluorescent Ca2+indicator, Na+,Ca++ exchange, and Na+, H+ exchange. Both bio- fura-2, and changes in [Ca2+4 and membrane chemical and electrophysiological measurements have been current that resulted from Na/Ca exchange were isolatedthrough the use of various organic made on many enzyme transport systems. This paper re- channel-blockers and impermeant ions. views the theoretical methods used to describe ion trans- Information on subcellular localization of port by cnzy me molecules. State diagrams are used to de- [Ca2+) i was obtained by imaging the changes in scribe the transport process with rate coefficients derived furs -2 fluorescence with the use of afiber- from experimental data and thermodynamic constraints. optically coupled CCD camera and digital image processing. Depolarization of cells elicited Net rates are calculated for physiological processes and 'low, spatially uniform increasesin (Ca2+4, unidirectional flux rates are calculated for tracer experi- with the maximum increase depending on internal ments. Classical enzyme calculations were based on an sodium ion concentration,external calcium ion aly tic solutions of simplified reaction schemes, but nu concentration and membrane voltage. merit solutions by digital computer now make it possible Repolarization was associated with the extrusion to study the behavior of more complex and realistic re- of Ca2+ and an inward current. A mathematical model of Na/Ca exchange has been formulated that actions. can describe these results.

1988 World Congress on .iedical Physics and Biomedical EngineeringSan Antonio, Texas 157 177 Biomedical Engineering Scientific Papers

BEl-B.1 SIMULATION OF MOTOR UNIT ACTION POTENTIALS COMPARISON WITH CLINICAL RECORDINGS

Sanjeev D. Nandedkarl Ph.D. Duke University Medical Center, Durham, NC 27710

A motor unit action potential (MUAP) is the spatial and temporal summation of action potentials of all fibers belonging to oner for unit (MU). The MUAP waveform depends on the MU architecture, i.e. the number and size of muscle fibers in theMU, the distribution of muscle fibers in the MU territory, etc. The relationship bewteen the MU architecture and the MUAP waveform also depends on the typeof the recording electrode. We simulated normal MUs and thenaltered theirarchitecture to simulate various pathologic processes such as loss of muscle fibers, regeneration of fibers, reinnervation, increased variability of fiber diameter, etc, that are known to occur in nerve and muscle diseases. Single fiber EMG, concentric needle EMG and Macro EMGMUAPs were simulated fromsimulated normal and pathologic MUs. Although the numerical values of the features of the simulated MUAPs differred from the clinical recordings, simulations demonstrated many patterns of abnormalities seen in nerve and muscle diseases.

BE1-B.3 BE1-B.4 Phase Locked Loop Method to Process Surface Principal Component Analysis in Electromyo- EMG, M.Cesarelli; R.Castaldo,F.Clemente, graphy, N.B. Jones*, P.J. Lago and A. Parekh, M.Bracale, Cattedra di Elettronica Biomedica University of Leicester, LE! 70, U.K., Nucleo di Ricerca di Ingegneria Medica e del- University of Porto, 4000 Porto, Portugal. la Riabilitazione C.N.R., Faculty of Engi- Much of the progress indiagnosticmedicine de- neering, University of Naples, Via Claudio pends on the harnessing of computer power. It is 21, 80125 Napoli, Italy. now sensible to use extensive data reduction techniques for extracting the essential features A new method for evaluating EMG spectral parameters from images orsignals. Principalcomponent based on the analogue locked loop technique has analysis (PCA) is recognised as a potent method been proposed. of multivariate analysis whichdramatically reduces dimonnionality and is an efficient aid to As P.L.L. technique can be applied to follow the classification. The application of this technique instantaneous frequency of a narrow band signal, to electromyography is particularly attractive the EMG is SSB modulated to obtain instantaneous because signals arise in this area of clinical frequency waveform. The average of the instantane- practice which contain an ext. sive amount of ous frequency of the signal is an unbiased estima- information. Not only can PCA be applied directly to the classification of motor unit action poten- tor of mean frequency, i.e. muscle fatigue indica- tials but it can also be used for the classifica- tor, and the average of its absolute value is the tion of interference patterns by the prior use of variance of spectrum. a spectral analysis technique. Theprogress of neuropathic and myopathic disorders have been A prototype based on this method has been developci. followed in simulationusing these methods via This application offers the possibility of a real two and three dimensionaldisplays. Experience time measurement of suri,ce EMG spectral changes with a limited set of real data shows that guod during isometric and isokinetic suf mined tasks separation of cases may be achievable in clinical and is particularly important in applications for applications. studying the muscle fatigue.

158 1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas 178 Biomedical Engineering Scientific Papers

BE1-B.5 BE1-C.1. Magnetomyography: non invasive recordingof Leakage Conductances in Axons and intracellular currents in skeletal muscle, their Myelin Sheaths, W.L.C. Rutten *, and J.H.M. Put, University of A. R. Blight*, Purdue University, Twente, 7500 AE Enschede, The Netherlands. West Lafayette, IN 47907.

Magnetomyography is a non-invasive technique for Microelectrode recordings from the in vivo detection of very weak magnetic fields mammalian axons and theirmyelin of active skeletal muscle. In a configuration in sheathsindicate thatthere is a which the detection coil has a toroidal shape and descending order of myelinleakage in which the muscle is lead through the centre of conductance and a balanced ascending the toroid, magnetic responsescan be measured order of nodal leakage conductance in which correspond to the intracellular longitudinal theseries: centralspinal axons, currents in active muscle fibers.For thisthe peroneal nerve sensory fibers, peroneal muscle must be enclosed by the toroid as closely nerve motor fibers. From the aspect of as possible and both must be immersed in a well conduction velocity and energetics, the conducting fluid. The system has an average equi- relative distribution of conductances valent noise limit of 1 nA/jHz over a bandwidth of between node and,myelin appears to be 10 kHz. This value is sufficient to measure fields of little consequenceover a wide in the order of picoTeslas produced by single mo- range,as long as the combined input tor unitsin EDL and Tibial muscle of rat and impedance of thefiber isoptimal. mouse. Responseshave a biphasic shape which Relative variations in myelin sheath strongly indicates that the disturbing (triphasic) i nodal conductance appear to be contributions of fields dueto currents inthe --.gnificant for other aspects of axon volume conductor inside the muscle, but outside function, among which threshold the active unit, are very small. The method of- modulation and accomodation are the fered us the possibility to make fiber distribu- most obvious possibilities, though the tion profiles along the muscle, to determine con- broader functionalsignificance of duction velocities and to measure current-to-force differences in these properties is not relationships in a way much easier (and non-inva- yet well understood. sively) than by electrical methodS. Combined with EMG measurements and model simulations the method leads to parameter reduction in the models.

BE1-C.2 BE1.-C.3 ADatributed - Parameter Model of the Myelmated Magnetic Measuremmts of Action Currents in Bundled Nerve Fiber, J.A. Halter and J.W. Clark Jr., Depart- Nerves, J.P. Wikswo, Jr.* and R.S. Wiiesingbe, ment of Electrical and Computer Engineering, Rice Vanderbilt University, Nashville, TN 37235 USA University, Houston TX 77251-1892 We are using toroidal pickup coils and a low-noise current-to-voltage converter to record the magnetic Recent electrophysiological studies on both mammalian and field produced v the compound action current (CAC) amphibian myelinated nerve fiber have produced new informa- propagating along isolated nerve bundles from the tion about the distribution of ionic channels in the fiber, the earthworm and frog. By assuming the shape of the dynamics of the ionic channels as well as the electrical proper- transmembrane potential, we can use our volume con- ties of the myelin sheath and the underlying axolemmal mem- ductor model for an axon off-center in a sheathed, brane. These studies allow for the formulation of new more anisotropic bundle to compute the single-fiber action explicit models of the nerve fiber, particularily for the intemodal current (SFAC) and potential (SFAP). By coupling and paranodal regions where the passive electrical properties and this model with a non-negative least-squares algo- the function of ionic channels are under debate. For example, rithm, we can deconvolute both the CAC and the con- the mechanism for the depolarizing afterpotential has been pro- ventionally-recoraed compound action potential (CAP). posed to be storage of charge or, the intemodal axolemma requir- The resulting conduction velocity distributions 1% either a higher conductivity for the myelin sheath than tradi- (CVDs) are compared to the fiber diameter histogram tionally thought, or a significant conductive path through the (FDH) obtained histologically. In this inverse periaxonal space. 1' :w models can address these issues. In this calculation,the intracellular and extracellular study, a new distributed-parameter ir.:del is developed allowing conductivities, temperature, and the distance between for representation of both the amphibian and mammalian nerve stivtius and recording sites are the most important fiber... The model describes the myelin sheath separately from parameters; the extracellular conductivity and the the axolemma, includes a periaxonal conductance pathway, uses tnickness ano conductivity of the myelin sheath new ionic channel descriptions for the node and internode and affect the CVD from the CAN more than tie CVD from incorporates a more explicit description of the nodal and parano- the CAC. The slower conduction velocity classes dal .egion than in previous models. The relative influence of of both CVDs are affected by noise more than the these parameters on the behavior of the myelinated nerve fiber faster classes. The volume conductor model, the will be discussed. CVDs, and the FDH are used to reconstruct the CAC I his work cupponed in part by NSF Gram EC8405435. and CAP. Because the SFAC is biphasic while the SFAPis triphas'., the CAC has less cancellation than the CAP,suggesting that CAC deconvolution may be more accurate than CAP deconvolution. 179

1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas 159 Biomedical Engineering Scientific Papers

BE1-C.4 BEI-D.1 A Double-SFAP model for the Nerve Compound Effects Tissue Geometry on the Initiation of a Action Potential, G. Contento: M.R. Malisan Cardiac Action Potential, R.W. Joyner: and R. Budai, Depts of Medical Physics and Departments of Pediatrics and Physiology, Emory Neurophysiopathology, Ospedale S. Maria University, Atlanta,GA 30322 della Misericordia, Udine, Italy. We used rabbit papillary muscles (PM) and isolated ventricular muscle cells tocompare the The conduction properties of peripheralnerves can effects of Lidocaine on decreasing cardiac be characterized by the distribution of conduction excitability. For the PM we defined tissue velocities (DCV) of the constituent fibres. Methods excitability as the inverse of the current for the estimation of DCV rely _n a linear modelof required to initiate a propagated action potential frcm a local stimulus. the Compound Action Potential (CAP) whichcan be For the is,51ated cells we defined cellular excitability easily measured in vivo. A basic assumption of the as the inverse of the current required to model requires that a unique waveform of the Single initiate a membrane action potential. For PM, Fibre Action Potential (SFAP) is representativeof Lidocaine with elevated [1(1-)0 decreased V max' the whole class of fibres conducting ata given decreased conduction velocity, and strongly decreased tissue excitability. range of velocities. Failure of this assumptionmay For the isolated concern initial neuropathies where normal and pa- cells, Lidocaine with elevated [e], decreased Vmax but had little effect on cellular thological fibres co-exist in the samenerve and excitability. We interpret our results on the conduct at the same velocity but with different differences of effect on tissue excitability SFAF waveforms. The present model of the CAP allows versus cellular excitability as a consequence of a weakened assumption on fibre grouping and descri- the syncytial nature of the PM.The electrical bes the conduction of both kinds of fibres. Due to load of the surrounding cells makes the tissue excitability dependent upon the amount of inward the invariance of the model to a set of lineartran- current that the locally excited cells and the sformations, an external input other than the mea- surrounding cells can generate. We successfully sured CAPS is necessary in order to resoirethe DCV simulated these phenomena with numerical of both fibre groups and their relat.vecontfibution. solutions of action potential initiation inan isopotential cell as compared to a two dimensional disk of excitable tissue.

BE1-D.2 BEl-D.3 A Model of Reentry Based on Anisotropic EXCITABILITY, SAFETY FACTOR OF Discontinuous Propagation, M.S. Spach*, P.C. PROPAGATION AND PREFERENTIAL BLOCK Dolber, J.F. Heidlage, Duke University, Durham, NC, USA IN ANISOTROPIC TISSUE

Available models of cardiac reentry depend solely Bruce M. Steinhaus*, Ph.D., Carmen Delgado**, Ph.D. on the active membrane conductances to alter and Jose Jalife**, M.D., *Univ. of Utah, Salt Lake City, propagation. The purpose of this paper is to show UT, **SUNY/Upstate Medical Center, Syracuse, NY that different anisotropic passive properties produce different propagationresponses to the The success or failure of propagation in anisotropic my- same change in the sodium conductance. In uniform ocardium was studied using simulations of action potential anisotropic bundles, premature action potentials propagation and isolated tissue experiments. Longitudinal propagated as a smooth process until propagation ceased simultaneously in all directions. In (L), uniform transverse (UT) and nonuniform transverse nonuniform anisotropic bundles, however, early (NT) tissue Directions were modeled in a Beeler-Reuter premature action potentials produced either based one dimensional cable. Stimulation threshold was unidirectional longitudinal conduction blockora smallest in the NT direction.The safety factor during dissociated zigzag type of longitudinal conduction (a safer type of propagation similar propagation was determiucd as the total axial current dur- transverse propagation). Directional ing depolarization divided by the current just necessary for differences in velocity demonstrated that continued propagation and was largest in the L directi "n. anisotropic propagation was necessary forreentry to occur within an area less than 50 mm2. Sodium The strength-interval plot for the junction between L and channeA. blocking drop also produced dissociated NT direc,..ins identified a range of stimulus strengths which zigzag longitudinal conduction is nonuniform resulted in NT but not L propagation at the basic cycle anisotropic bundles, but not in uniform length while higher stimulus strengths resulted in L but not anisotropic bundles. These results show that obliteration of side-to-side electrical coupling NT propagation. Experimental strength interval plots from between fibers provides a primary mechanismfor 10 L shapped preparations of isolated sheep epicardial mus- reentry to occur within very small areas (1-2 mm) cles had similar characteristics. Results show directional due to propagation phenomena that donot occur in tissues with tight electrical coupling in all differences in excitability and safety factor of propagation in directions. anisotropic tissue. These properties can be a major factor in the genesis of reentrant arrhythmias. 180 160 1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas Biomedical Engineering Scientific Papers

BE 3.4 BEt-E.1. Contribution of active Na-K transport UPPER EXTREMITY ASSIST DEVICES, P.H. Peckham*, to membrane repolarization, F.A. Roberge* M.W. Keith, G.B. Thrope, and K. Stroh, and R. Lemieux, University of Montreal, Case Western Reserve University, and the Montreal, Canada. Veterans Administration Medical Center Cleveland, Ohio 44109

The small size of cardiac cells, the restricted Functional neuromuscularstimulation (FNS) myocardial interstitial space, and the relatively hasbeendeveloped to providecontrol ofthe important Na+ influx and K+ efflux create condi- paralyzed extremity of the cervical level spinal tions leading to substantial changes in [Na] i and injury patient. Subjects are implanted with [Kie during the action potential. An efficient either percutaneous electrodesor an imp'', nted Na-K pump, sensitive to both [Nali and [K]e, is stimulator to provide activation of the muscles. the. necessary to restore as much as possible of The electrodes are implanted in the major finger the original Na+ and K+ gradients at each cardiac and thumb flexor and extensor muscles, and cycle. These variations in [Nali and [Kie induce coordinated movement patterns are controlled by a important changes in the passive membrane currents portable external control module. Movement of the via the corresponding changes in equilibrium shoulder regulates hand opening and grasp in two potentials. These factors, together with the Na-K grasping configurations. pump current, determine the shape of the actin: Subjects with FNSsystemsdemonstrate the potential. A model of the cardiac cell membrane ability to p7-rform activities of daily living such incorporating active Na-K transport was used to as eating, writing,and grooming that they are study the ionic current changes occurring during unable to achieve without the assistive system. the action potential, following repetitive stimu- Such systems have been in outpatient use for over lation at rates between 0.2 and 3.3/s. The diasto- five years. Recent developments, including lic levels of INaji and [Kie increase at higher combining s4 lical reconstruction and the frequencies since the pump cannot maintain the implantab7a stimulator, have further improved concentration values absolutely constant. It is func.gon and reduced the user's care and observed that the contribution of active Na-K maintenance of the system. transport to membrane repolarization becomes increasingly important as the rate of activity Supported by the National Institute on Disability increases. and Rehab. Research and the Veterans Administra- tion Rehab. Research and Development Service.

BE1-E.2 BE1-E.3 FUNCTIONAL ELECTRICAL STIMULATION OF THE Cardiac Assist Device: Larry W. PARALYZED LARYNX, I. Sanders*, MD, Mount Stephenson, M.D.,* Dept. of Surgery, Sinai Medical Center, New York, NY 10029 University of Pennsylvania, Phila- delphia, PA 19104, U.S.A. In laryngeal paralysis the vocal cords come together in the midline of the larynx and cause a Skeletal muscle can respond adaptively to in- life threatening airway obstruction. To rehabili- creased functional demands. This adaptive training tate this condition function must be restored to capacity is commonly observed during the the muscle which controls vocal cord opening,the of athletes, but to a greater extent when skele- electrical posterior cricoarytenoid (PCA) muscle. The tal muscle is subjected to chronic Muscle conditioned in this way shows unique challenge is that the PCA muscle is in a sti.ilation. fatigue. During denervated condition and to date there has been a greatly increased resistance to transformed little research performed on the stimulation of t'a stimulation process, the muscle is from a muscle of mixed fiber type to ore of a denervated muscle. Initially 12 dogs were studied under varying uniform population of type 1 fatigue-resistant The changes are documented by immuno- conditions of electrode placement, electrode fibers. polarity, electrode orientation to muscle fibers, and histochemical criteria. Myosin isozymes and over a wide range of pulse width and pulse change. There is an enzymatic shift toward The bioenergetics of elec- amperages. From this original study an electrode aerobic metabolism. trically conditioned muscle are similar to those array was adopted for chrollift experimentation. Four dogs were alloyed to atrophy for six months of heart muscle. We have exploited this transformation phenomenon to develop skeletal muscle and at the end of that period two of the dogs ventricles capable of performing the continuous underwent three months of electrical stimulation work required to assist the failing heart. These by implanted Ilectrodes. ventricles generate greater stroke wJele than the Results indicated that a variety of stimulation parameters and electrode arrays are capable of right ventricle, and can pump blood in the circu- This research causing clinically useful vocal cord opening. In lation effectively for weeks. addition prolonged stimulation can restore func- indicates that a chrinic non-fatiguing biologic muscle tion to severely atrophied PCA muscle. Further oump motor const-uctel from skeletal This concept holds research is required into an electrode array appears to be acidevab.a. which localizes current to the PCA so that cur- promise for long-term augmentation or replacement rent does not spread to surrounding structures. of the failing caiiac ventricles. 181

1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas 161 Biomedical Engineering Scientific Papers

BEI-E.4 BE1-F.1 '.".ollision Block of Motor* Activity in Peripheral TheoreticalBasisforBrainMotor Control nerve. James D. Sweeney, J. Thomas Mortimer, Assessment, r;.V. Kondraske*, Human tonald R. Bodner and A. Stewart Ferguson, Performance Institute, University of Texas at AppliedNeuralControlLaboratory,Dept. Biomedical Engineering, and Division of Urology, Arlington, Arlington, TX, USA. (;17- 273 -2335) Dept.ofSurgery,CaseWesternReserve University, Cleveland, Ohio 44106. Sound conceptual bases which ind to testable hypotheses are essential for assessment of complex systems, such as in brain motor control. Generalizable Collision block of efferent motor activitymay offer a performance theory concepts are presented as a basis clinicallyviablealternativetoothertechniquesfor for defining basic elements of performance and for electrically blocking peripheral nerve transmission. A collisionblockwillresult whenartificiallyproduced structuring a quantitative approach to assessment in Aotidromic Unidirectional ly Propagating ActionPotentials this area. (AUPAPs) annihilate orthodromic action potentials.We are developing implantable "spiral' nerve cuffelectrodes that An elemental resource modeling approach is used to can reliably produce such AUPAPs. define structures functionally and dimensionalized These Asymmetric Two E: ,ctrode cuffs (ATECs)are performance measurands. Maximal stresses along specifically designed to produceacollisionblock of pudendal nerve mot:r activity.Pudendal nerve mediated isolated dimensions of performanceare used to spasticityoftheperierethralstriatedsphincterhas determine resources available.Task analysis is used hinderedattemptsbyotherstodevelopabladder to define a series of worst-cases unidimensional neuroprosthesis for spinal cord injured patients.In acute stresses on system resources.Resource economic animal experiments we have studied theeffectiveness of principlesare applied to a multi-dimensional and the ATEC design in producing trains ofpudendal nerve multi-unit system model to define simple criteria motor AUPAPs between I and 50 Hz. Both monophasicand balanced-charge biphasic quasitrapezoidal stimulihave been oriented rules for assessment. Combinatorial used. In chronic animalexperimentsthelong-term optimization to minimize stress across performance effectiveness and safety of pudendalnerve collision block resources is considered as a control strategy and is currently being studied.A finite-difference model of discussed in the context of assessment.Supportive the electric field generated by an ATECis also being developed as an aid for electrode design. experimental data are presented. Experimental approaches are discussed.

BE1-F. 2 BE1 -F. 3 ATask Analysis StrategyforModeling Assessmentof Healthy and Disordered and Assessment ofHuman Locomotion, Neurocontrol of Gait, M. Ma lez ic ,*M. James J. Carollo*, HumanPerformance Gregoric and Institute, J. Krajnik,J. Stefan University of Texas at Institute, and University Rehabilitation Arlington, Arlington, TX 76019 Institute, Jamova 39, 61001 Ljubljanr., Yugoslavia The problems associated with gait assessment are due in part to the absence of global models of human Central nervous sys;cm control of muscles locomotion, during incorporating both gait can be seriously altered in patients with biomechanical and europhysiological the upper motoneurone dysfunctions. information. Task In analysis has been incompletespinal cord lesions, the impaired employed to simplify model development, supraspinal control and results in diminished cotpled with resource based centrally programmed and suprasegmental reflex performance theory, provide a framework activity with increased segmental reflexes. for assessment. EMG recordingsofmuscle activityduringwalking Waning is described as a repetitive combined with measurements of forces and phases sequence of tasks where each task inthe of gait may so hap in understanding the sequence has specific objectives which pathophysiology of abnormal locomotor must be patterns, accomplished within the time as well as in evaluating the period allotted. restorative The CNS is modeledas procedures. Centrally programmed cont,:ol a controller is which translates these assessed in the tibialis anterior muscle, objectives into descending neural normally active when unloaded, while soleus, the commands Iresources), targeted for the antigravity muscle, is a monitor for functional units (muscles) associated suprasegmental and segmental stretch reflexes. with thecurrent task. A biomechanical Changes in the avaraged rectified and filtered model of the trunk and lower extremities EMG signals are studied before the footcontact, is used to determine theresultant joint within the latency ranges of segmental and angles and torques associatedwith the suprasegmental reflexes after it and befo: be neural input, with these values used to push-off. simulate afferent feedback. * Supported by Dallas Rehab Foundation

162 1988 World Congress on Medical Physics and Blm§dgalEngineeringSan Antonio, Texas Biomedical Engineering Scientific Papers

BE1-F.4 BE1-F.5 Polyelectromyography (PENG) as a Tool Comprehensive Assessment of Control of to Evalur.te the Efficacy of Therapeutic Motor Activity: Brain Notc- Control Manoeuvres in Motor Disorders, B. Assessment (BMCA), A.M. SherwoJd*, M.R. Cioni*, M. Meglio, Facolta' di Medicine Dimitrijevic, Division of Restorative e Chirurgia, Largo Agostino Gemelli, 8, Neurologyand Human Neurobiology, Baylor 00168 Roma. Italy College of Medicine, 7000 Fannin Suite 2140, Houston TX 77030 Surface PEMG was used to assess modifications of motor control following therapeutic manoeuvres in 40 patients. A percutanPnus test trial of spinal The objective of this projectis to develop a cord stimulation was performed in 33 of them. 13 clinical system for the automated assessment of with spastic hemiparesis due to ischemia of the the control c.. movement in humans when that middle cerebral artery, 13 with spastic parapare- control is impaired by severe trauma or disease. sisfrom a spinal lesion and 7 with multiple SurfaceERG signals offer available sclerosis.5 patients with spinal cord lesions method for understanding 1- ..,civity related receivedintrathecal injection of anti-spastic to movmement. Computerized methods for assess- drugs and 2 patients underwent surgical removal ment of neurocontrol of motor activity based on of a spinal compression. Surface EMG of the EMG recordingsfrom multiple muscle groups are upper and/or lower extremity muscles was recorded being 2eveloped. Along with 12-16 channels of (t.c. 0.03 msec, 15 mm/sec, 100 uv/cm) during surfaceEMG, the BMCA protocol utilizes event spontaneous relaxation, voluntaryandpassive markers and position transducers to record volun- movements, provoked clonus, phasic andtonic tary automatic and reflex maneuvers. The BMCA stretch reflexes, Babinski's response and gait. system is based on an Apollo 4000 workstation with The evaluation included neurological examination a SignifiCAT ADC board and sof-4are. Features of and the use of semi-quantitative scales to score neurocontrol in individual patients are documented motor activity. The protocol was done before and for studies of underlying motor control mechanisms after thetherapeutic manoeuvres.Theresults in humans. BMCA is also used in single or longi- were consistent with the clinical outcome in all tudinal studies of the properties of malfunction- patients. Analysers of the EMG recordings allowed ing nervoussystems. Specific goalsinclude. speculation on the mechanisms involved in the protocol assistance,comprehensive data logging, therapeutic procedu e and the neurophysiological quantitative results,and classification of the criteria for the selection of patients. type and severity of the functional disorder.

BE1-G.1 BE1-G.2 CaL-diacPotentialMappingintheUnited User-friendly body surface potential mapping in myocardial States. R. Lux', Cardiovascular Research and Training infarction. Fred Kornreich * and Mikhail Kavadias. Institute, University of Utah, Salt Lake City, Utah Free University Brussels (VUB), Belgium. Body surface potential maps (BSPM) provide more information Electrocardiographicpotential mapping inthe US. on the electrical activity of the heart then can be obtained from has expanded from the ,body surface mapping studies the standard 12-lead ECG or the VCG. Moreover, sequential of two decades ago to include cardiac surface and instant-by-instant maps allow direct association of surface transmural measurements of the present. Research is potentials with major wavefronts within the heart and have progressingalongtheoretical,experimentaland advanced the electrophysiologic bases of diagnostic ECG criteria. Despite these advantages, widespread use of maps as a lines. surface clinical Major emphasisof body diagnostic tool is still lacking; the non-traditional data display mapping isapplication to noninvasiva detection and and the insufficient accumulation of clinical-pathologic assessment ofcardiacdiseasesmissed by correlations are at present, the main factors hampering conventionalelectrocardiography,e.g.,coronary acceptability by the clinical community. Using 120-lead data diseasewithoutinfarction,cardionayopathies,and from 236 normal individuals and 258 patients with documented arrhythmia vulnerablestates.Availabilityof high first Q-wave myocardial infarction (MI), qualitative and quantitative interpretation criteria are determined and presented resolution torso and heart images and cardiac surface in a format that can be easily viewed and interpreted by the potential information has renewed interstin forward cardiologist. The method rests on the development of and inverse models for purposes of characterizing discriminant maps, obtained by subtracting at each instant normalaneabnormalcardiacelectrophysiology. group-mean normal maps from respectively anterior and inferior Intramural and cardiacsurfacepotential mapping MI group-mean maps and dividing the resulting values at each hasrapidlyevolvedforuseas atoolfor electrode site by the corresponding pooled standard deviation. to assessventricuiar Similarly, discriminant maps are produced for each single intraoperativemapping individual and indicate the presence, location and timing of arrhythmiasandtounderstand the 3D spread of significant abnormalities. By quantitatively extracting and excitation and recovery. The latterisof particular displaying abnormal features in individual patients and in interest for linprosring understanding of the effects of particular, significant losses of potentials in early and mid QRS, anisotropy on cardiac conduction. MIE could be identified and located at first glance by the cardiologist, however little familiar he may be with BSPMs. The 1:y/eloped criteria can easily be implemented in knowleAg6b d systems for computerized analysis of maps.

1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas 163 Biomedical Engineering Scientific Papers

BE178.3 BE1-G.4 Recent Japanese Contribution to Cardiac Substantiation of a Practical Lead Potential Mapping, K.Hartute, H.Stunakawa, System foi Electrocardiotopography, G.Nishiyama, S.Kanesaka, Y.Okamoto* and P. Kneppo and L.I. Titomir, Institute T.Musha *, Showa Univ. Mijigaoka Hosp. of Measurement and Measuring Technique Tokyo Inst. Technol *, Yokohama, Japan of the Slovak Academy of Sciences, 842 19 Bratislava, Czechoslovakia Since the body surface mapping systemswas introduced by Yamada to Japan, many research warks Three rows t.f eight electrodes symmetrica- were contributed to the field of the body surface lly distributed in horizontal planes and map(BSM) in our country, the detection of break- electrodes on the head and, lzf: legare through of right ventricular excitation, the lo- used. The instantaneous potential distribu- calization and size of myocardial infarction, the tion all aver the chest surfacecan be re- location of preexcitation, the site of conduction constructed by means of expansion of this block in RBBB, exercise BSM, and soon. Recently distribution in the series of spherical a standardization to enable the exchange of mutual functions up to the 3-rd order terms. The BSM data from different institutions was proposed expansion is calculated with respect to by Japanese group. As one of application of BSMto variables expressing the lengths of mutua- clinical use, we developed with Musha'sgroup the lly orthogonal arcs along the chest surface. method to evaluate the nondipolar contents(the Then the coefficients of theexpansion are residue) on the BSM. The following results have determined a:. linear functions of thepo- been obtained; two peaks of the residue during tentials measured by the unipolar leads of QRS and t /gle dipolarity during T in the normals, Nelson s system. Applying a similarexpan- the higher residue in the early part of QRS in sion also to the cartesian coordinates of anterior infarction and also in inferior infarc the chest surface, it is possible to obtain tion, which bins the higher residue in later part rather simple expression5 for low order of QRS, the inverse relation of the residueat multipole components of the caroioelectric middle of QRS and the size of anterior infarction, generator. Simultaneous use of Nelsons the higher residue in early QRS as the evidence of lead system and the techniques of spherical infarction with CLBBB and others. functions expansion significantly simplifies topographical measurement and multipole analysis of the cardioelectrir field.

13E1H.2 Simulation of Normaland Epileptiform Spontaneous Cellular Population Oscillations in Hippocampal Slices. EEGs with Chaotic Solutions of Static R.D. Traub, R. Miles and R.K.S. Wong, IBM Watson Research Ctr., Nonlinear Coupled Ordinary Differential Yorktown Heights, NY 10598 and Dept. of Neurology, Columbia Equations. Walter J. Freeman: University University, New York, NY 10032. of California, Berkeley California94720

Measurements of the fractal Oscillations in large neuronal populations (thousands of cells) occur in dimensions in vitro of normal and epileptiform EEGs recorded hippocampai slices, especially when synaptic inhibition is partially suporessed.Periods range from a few hundred ms to several from the olfactory bulb andprepyriform cortex of waking rats and rabbits indi- seconds. Treatments such as further blockade of inhibition that increase the amplitude of such oscillations also diminish the frequency. cate that the aperiodic and unpredictable These activities are detected by dendritic field potentials or by dual activity, previously thought to be intracellular recordings. "noisy", has low They also occur in a randomly connected dimensionalitycon- network model with a random distribution of spontaneous bursting sistent with deterministic chaos. Sets rates, provided that intrinsic cellular properties and synaptic actions of differential equations are construct- and connectivities are based on physiological data. In the model, the edin accordance with the known physio- period and amplitude are influenced by the time course of the long in- logical and anatomical properties of the trinsic afterhyperpolarization, the strength of recurrent excitatory bulband prepyriform and their neurons. synapses, and the time course of slow 1PSPs. While the population as Tire solutions contain domains of chaotic a whole oscillates (every few hundred ms, a cluster of about I% of the behaviorclosely resembling "spontane- cells fires a burst, Generating synaptic potentials in most of the rest of ous" EEG activities in a variety of nor- the cells), the eel:afar composition of the clusters has no apparent mal and pathological states, including periodicity. Gboa Ily coherent oscillatas also occur in a model with basal chaotic backgrounc activity, burstswith 9000 pyramidal cells wherein the synaptic connections are spatially inhalation, anA repetitive restricted (as siggested by recent data). spikes and waves that accompuy complex In such distributed models, thi. frequency )f thc oscillation does not depend on du total number partial seizures known also as petit mal of cells, but on,y on the local network properties. In vitio oscillations epilepsy. may provide insight into general rhythmic possibilities of isolated cortex. " 41

164 1988 World Congress on Medical Phy' cs and Biomedical EngineeringSan Antonio, Texas Biomedical Engineering Scientific Papers

BE1 -H.3 BE1-I.1 Dynamics of the EEGs as Signals of Neuronal Biphasic Waveforms for Defibrillation, J. Jones Populations (Alpha Rhythms): Linear and Non- and R. Jones, Georgetown University and V.A. linear Properties, F.H. Lopes da Silva*, Medical Center, Washington, D.C., U.S.A. Biological Center, Universiteit van Amsterdam, 1098 SMAmsterdam, The Netherlands Automaticinternal defibrillators have been developed for implantation in high-risk patients Models of neuronal populations can be useful to whose arrhythmias are refractory to drug treat- understand how EEG signals are generated and pro- ment. However,due tothe high defibrillation pagated in the brain. To simulate alpha activity threshold for the existing monophasic truncated as function of both space and time a model of a exponentialwaveform, these defibrillators are simplified cortical network was constructed. The heavy and bulky, and the high currents required time-dependent behavior of the neuronal membrane can induce ventricular dysfunction. Since defib- potentials was described by linear transfer func- rillation results from excitation of ventricular tions. Static nonlinearities relate membrane poten- cells, we studied relative excitation and dysfunc- tials to impulse densities. The spatial-dependent tionthresholds for biphasic waveforms in the behavior was described by exponentially decaying cultured cell model and possible mechanism-for functions of tne type (C(x) = Clexp (-clxl)) where threshold alteration with computer simulations of 1/C is the characteristic length. stimulated cariiac action potentials. The model generates alpha activity with character- Shaped biphasic waveforms produced a significantly istics in agreement with experimental data. From larger decrease in excitation threshold under " fibrillation conditions" (high Ko and short cycle the phase shift at the alpha frequency as function length) than under normal conditions in the cul- of space the characteristic lengths of the neuronal Computer simulations suggest network were computed. These were found to be of tured cell model. that the decreased threshold is due, in part, to the same order of magnitude as the basic cortical improved fast excitation channel activation with module estimated from histology. Linear and non- biphasic waveforms. Similar biphasic waveforms linear properties of this type of models including decreased defibrillation thresholdin the dog, the transition between linear and nonlinear dynamic calf, and perfused rabbit heart models. These behavior, as seen during epileptic seizures, are results suggest that biphasic waveforms decrease analysed. defibrillation thresholl by improving excitation of ventricular cells in the fibrillating heart.

8E1-4.2 BE1-I.3 Current Distributions in Bisyncytial Tissue, Simultaneous Potential Field and Cardiac J.P. Wikswo, Jr.*, B.J. Roth and N.G. Sepulveda, Recordings from the Heart During Ventricular Vanderbilt University, Nashville, TN 37235 USA Defibrillation. The electrical behavior of cardiac .issue has been studied using an idealized 2-dimensional, finite- F.X. Witkowski* and P.A. Penkoske, University element model that accounts for the bisyncytial of Alberta, Edmonton, Canada. nature of the tissue and the anisotropy of both the intracellular(IC) and extracellular (EC) spaces. Electrical defibrillation (DF) is the only Previously we examined current, voltage, and known generic form of therapy for ventricular magnetic field distributions from a circular defibrillation. We have developed a three - activation wavefront for various --isotropies. We dimensional electrophysiological mapping have now determined the IC, EC, and transmembrane system capable of simultaneously recoriing_ (TM) potentials resulting from injecting current both the actual high voltage field created (1) into the EC space, and (2) into both the IC and by a DF shock and the immediate post-shock EC spaces. In (1) theisopotential distribution activation. Using transmural directly-coupled (ID) is elliptical for both the EC and IC spaces. (DC) electrodes 120 left and right ventricular The TM ID is elliptical only for equal anisotropy sites were examined with a newly developed ratios (EAR), and complicated for other anisotro- DC mapping systen Each electrode was coupled pies. In some cases depolarized and hyperpolarized to 2 amplifiers (DC-500 Hz) one with a regions are adjacent, which may have implications dynamic range of ±130mV (electrogram) and for reentry, fibrillation and defibrillation. In the other ±500V (Di field). (2) theID for the IC and EC spaces is not elliptical; the TM ID has an unexpected pattern and Successfll DF required 1007. of the sites both depolarized and hyperpolarized regicns are examined to be depolarized, with the first present. These results are compared to n vivo beat post-DF originating in the area of measurements of virtual cathode sizes. We (...nclude minimum DF potential. Unsucc ssful defibril- that in 2-D anisotropic tissue. current injection lation demonstrated at least one site of results in spatially intricate potential and ineffective DF that correspond to the area current distributions that depend strongly upon of minimum DF potential. tissue anisotropy. Some of these effects were not previously recognized because other models used only equal anisotropy ratios. 185 1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas 165 Biomedical Engineering Scientific Papers

BE1 -I.4 3E1-J.1 3D FINITE ELEMENT ANALYSISOF INTERNAL EEG Evoked Response DEFIBRILLATION - N. Thakor, K. Kothiyal,B. Quantification and Detection Shenker, L. Fogelson,Biomedical Eng., Jorge I. Aunon* Johns Hopkins university, Baltimore,USA Colorado State University Ft. Collins, CO 80523 In internal defiarillation highenergy shock: are delivered by epicardialpatch electrodes or endocardial catheters. A Evoked brain potentials have been used for finite element method (FEM)is used to many years to evaluate the activity of numerically solve Laplace's equationfor selected sensory systems. Traditionally, the heart modelled as a volumeconductor. A averaging has been used as the tool of choice computer model is constructed from1473 in order extracttheaverage evoked brain elements and 2063 nodes. FEM solves potential from the on-going EEG. During the current density (CD) and energy (EN) dis- last 15 years a diversity of new techniques tribution in the heart for three electrode have been developed in order to quantify the configurations- patch-patch (PP),catheter- single evokedpotential. Some of these patch (CP) and catheter-catheter(CC) - and techniques will be described and compared. various geometries and placements. Ultimately, the questiml must be asked as to The simulations indicate that:1) PP and whether these techniques have really made an CP more effectively deliver ENto myo- impact on the knowledge of evoked potentials. cardium '82% aid 80%) as compared with CC ;68%). 2) Spec'-fic electrode placements do not significantly affect EN delivery to myocardium. 3) Electrode size significantly affects EN distribution, especially in the vicinity of electrodes. 4) Onlya small fraction of the myocardium receives high CD (50% at 0.1 A/cm2 and 4.3% at 0.5 A/cm2). 5) Very high CD (1.67 A/cm2 in CC and 1.14 A/cm2 in PP) are observed at the edgesof electrodes. These results suggest effective electrode designs and the roleof EN and CD in internal defibrillation.

BEI-J. 2 BE1-J.3 Quantitative Analysis of EEG Activity: Is SPATIAL DISPLAY AND ANALYSIS OF EEG/EP DATA. There Chaos In The Future? Ben H. Jansen; D. Lehmann *, Dept. of Neurology, University University of Houston, Dept. of Electrical Engi- Hospital, CH-8091 Zurich, Switzerland. neering, Houston, TX 770C4, USA.

Prerequisite for correct spatial display The quantitaive analysis of 'spontaneous' electri- ("mapping") of EEG/EP data is adequate spatial cal activity of the brain (EEG) is aimed at providing the sampling. Treatments (means, differences, electroencephalographer (EEGer) with 1) an objective tool correlations) over maps require locally unique to analyze EEGs, and/or 2) with a sensitive tool to extract values, obtained by spatial offset removal (ay. information from the EEG not obtainable otherwise. Early ref.) or high pass filtering (gradient, CSD). techniques were either heuristic or based on adaptations of Data reduction of map series should be space- methods originally designed for othe applications. More oriented. Spatial analysis aims at the recogni- recently, methods have been developed that take the spe- tion of the "electric landscapes" for which e.g., cific characteristics of the EEG into account. For example, the locations of the positive and negative syntactic methods and (adaptive) segmentation techniques extreme potential might be used. Adaptive were developed to deal with the nonstltionary nature of segmentation of EEG/EP map series into epochs of EEGs; the importance of contextual informationwas acknowledged through the use of expert systems to allow for stable map landscapes identifies brain functional reasoning involving complex spatial and temporal relation- nicrostates. Different microstates might manifest ships utilized in the visual EEG analysis. Evenmore recent different steps or modes of brain information aLe methods based on a model of the neural masses, involv- processing, e.g., different states at stimulus ing non-linear differential equations that exhibita 'chaotic' arrival are associated with different behavioral (as opposed to a random) behavior witha striking similar- outcome of processing (reaction time), and ity to the actually recorded EEG. This paper will present attending vs ignoring incoming information a critical review of these state-of-the-art techniques. produces landscape differences in a given segment of the event-related map series. J.S6

163 1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas Biomedical Engineering Scientific Papers

BEl-K.,^. The Design of Multiple Electrode Mapping Instrumentation.

F.X. Witkowski* and P.A. Penkoske University of Alberta, Edmonton, Canada.

Improved methods for assessing throe- dimen- sional distributions and rapid temporal changes of activation, repolarization, systolic and diastolic potentials in extracellular unipolar electrogram recordings would enhance studies of arrhythmogenesis. To assess all these parameters requires directly coupled (DC) mapping techniques.

Mapping systems to date have been limited by physical constraints of electrode design, number of channels simultaneously sampled, duration of continuous data fZorage and three dimensional representation of the data. Approaches to each of these problems will be discussed. A new system that has recently been developed will be presented capable of DC transmural . yocardialinterrogation from 384 channels simultaneously, each channel sampled at 2KHz, capable of continuous storage of 3,500 MBytes (60 minutes), with fully automated data analysis and three dimensional data presentation.

BE1-K.3 BE1-K.4 Mapping Electrical Impulse Propagation in the Heart Recording and Imaging Cardiac Electrical Activation for the Study of Cardiac Arrhythmia Mechanisms, Using Voltage Sensitive Dyes, Stephen Dillon*, Andrew L. Wit*, Columbia University, New York, Columbia University, New York, New York. New York. Myocardial infarction in humans results from Voltage sensitive dyes (VSDs) are a class of obstruction of a coronary artery that results in molecular probes which can linearly transduce var- death of cardiac muscle.Infarction is accompanied iations in membrane potential into changes in their by disturbances in the rhythm of the heart beat. absorption anrifeir fluorescence.VSDs bind to the Eventually ventricular fibrillation may occur that outer membrane leaflet of cells so that voltage is results in the absence of cardiac output and sudden detected as a component of the electrical field. death.We have been studying cardiac electro- With appropriate instrumentation it is possibw to physiological effects of coronary obstruction in the record "optical action potentials", i.e. the signal dog to determine how these rhythm disturbances developed by the variations in cellular voltage gen- occur.To do this we have designed a data acqui- erated during cardiac activation.This approach, sition system that enables us to record electrical therefore, allows one to study the det His of im- activity at 192 sites in the heart, digitize the sig- pulse propagation at a large number; sites, far nals and analyze activation times with a computer. more than is possible using microelectrode We have shown, using this system, that coronary techniques.While high density electrographic map- occlusion changes the conductile properties of car- ping can also provide such information, VSD diac tissi Ns which have a diminished blood supply. recording can provide data about the impulse itself This in turn results in a phenomenon known as re- (ex. upstroke velocity, refractory period). Further entrant excitation; the conducting impulse does not by using laser scanning or CCD cameras one can die out after activating the heart once as it does develop high resolution images of impulse propaga- normally, bct rather continues to propagate in a tion.This kind of data could be used to under- circular pattern throughout the ventricles activa- stand the mechanisms of arrhythmogenesis and ting them over and over again.This causes a their pharmacological treatment. rapid heart beat and eventually ventricular fibrillation.

187

1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas 167 Biomedictal Engineering Scientific Fapers

BE1-L.1 BEi-L.2 Long-term Monitoring in Epilepsy, Sleep Analysis as a Data Reduction Program J. Gotman*, Montreal Neurological lack R. Smith Institute, McGill University, Electrical Engineering Department Montreal, Quebec, Canada University of Florida Gainesville, FL 32611 The electroencephalographic (EEG) correlates ofep- ilepsy are sometimes rare and unpredictable,as are the seizures them'elves. It is therefore necessary Sleep analysiscan be interpretedas a data to perform long-term monitoring of the EEG and of reduction problem. Sleep staging, for example, the patient's behavior to obtain a full documenta- consists of reducing 30 seconds of data toone tion of tne disorder. Monitoring sessions may last byte and of reducing an eight hour record to less from 24 hours to several days. Behavior is moni- than 1000 bytes of information. Sleep staging tored by video. provides an incomplete and inadequate description Long-term EEG recordings can be performedon analog of a night's sleep. A subject reports that one tapes; video tapes are most commonly used at pre- night': sleep is much better thin the next, yet sent. Since it can be very tedious and expensive sleep staging does not discriminate between them. for somebody to review a 16-channel EEG recorded There are methods for extracting additional during several days, two strategies havepeen devel- information from the polysomnography data which oped to facilitate the process. Fast review systems more completely describe a nights sleep. The allow displaying the EEG on a monitor at 10 to 60 data reduction (sampling rates, number of times the recording speed; this can be avery effi- channels, bits/sample and feature extraction) cient process, particularly with 8 or fewer channels depend on the amount of information required by and when only major events are to be captured. Auto- the clinie,n, and the method plays an important matic detection methods, usually running on-line, patt in the system design and system cost. This allow processing 16 and sometimes more channels and paper provides system requirements for a variety detecting specific epileptiform activity suchas in- of sleep data reduction methods. terictal spikes, spike-and-wave bursts and seizure patterns; these methods are far from perfect but can be used effectively if biased towards minimizing false negatives. Possible avenues to improve the performance of existing methods will be discussed.

BEl-L.3 BEl-L.4 EEG Nbnitoring in Surgery Objective Assessment of Anticonvulsant Efficacy by R.A.F. Fronk*, A.J.R. Simons, E.H.J.F. Boezeman Analysis of Interictal EEG Activity, J.D.Frost, NicoletBic medicalInstruments, Madison, WI, USA Jr.*, R.A. Hrachovy, D.G. Glaze, and P. Kellaway, Continuous monitoring of thefunctioningof the Baylor College of Medicine, One Baylor Plaza, brain by means of therecordingand processing of Houston, Texas. the electroencm4Y.L:_pgrain (EEG) canbe clinically useful during surgical procedtres where the Although routine electroencephalography (EEG) pro- cerebral perfusion and metabolism are at risk, for vides definitive information essential to the example, during neurosurgery, carotidendarterec- diagnosis and classification of seizure disorders, bow or cardiac surgery. The main objective of it is not useful for assessing the effectiveness of EEG monitoring duringsurgeryis to prevent ir- drug treatment in individuals with partial (i.e., reversible brain damage resulting inpostoperative focal) seizures. We have demonstrated that seizure neurologiccomplications. For example, in carotid hazard can, however, be determined in a endarterectomy, there is a risk of cerebral quantitative manner through computer analysis of ischania when the carotid artery is clamped and interictal EEG spike waveform. Chiloren with dis- the collateral blood supply is inadequate. Then, crete spike foci and simple or complex partial sei- a decision to use atemporarybypass shunt is often zures have been studied before and during treatment based on significant ipsilateral slowing or volt- with sev--al anticonvulsant medications. Average age attenuation of the EEG data recorded during spike parameter, determinedduring 5- to test clamping of the carotid artery. Temporary 45-minute EEG samples obtained during NREM sleep. bypass shunt used or not, in both cases, it is Measurements included spike amplitude (A), normal- strongly recommended to continue the EEG monitor- ized sharpness (S), duration (D), anda mathemati- ing in order to detect aberrant EEG activity cal derivation, the coAposite pike parameter caused by cerebral ischemia. For the continuous (CSI), defined as CSP = A x D/S. In patients whose intraoperativeEEG monitoring, it is almost a seizures were controlled by the drug treatment but practicalrequirementto use a computerized system whose EEGs continued toshow interictal spikes, which provides the recording of the EEG of both there was a decrease in the average spike hemispheres, EEG data processing, feature extrac- amplitude, du-ation,and CSP, and an increase in tion and display, and the ..tutomatic geLeration of normalized sharpness, compared with the warnings at the occurrence of aberrant EEG activity. pretreatment baseline values. There findings indi- Experiences with such a cceTuter-assisted monitor- cate that computer-based EEG analysis o: interictal ing system will be discussed. activity may provide an objective means for 188monitoring anticonvulsant drug efficacy. 168 1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas Biomedical Engineering Scientific Papers

BE1-M. 1 BEi-M.2 Inverse Solutions in 2D Using Propagation Models Localization of pre-excitation and arrhy- Tadashi Mara and Roger C. Barr* thmogenic sites by a single moving dipole, Duke University, Durham, NC 27706 USA. P. Savard*, M. Shenasa, R. Nadeau, P. Page Hepital Sacre-Coeur, Montreal, Canada. If excitation proceeds away from a site of origin located somewhere within a two-dimensional sheet of excitable Epicardial mappingwas used to evaluate the tissue, can the site of origin be determined by examining the accuracy of a single moving dipole (SMD) for the waveshapes on a few extracellular electrodes? We exarniuoi localization of ventricular act' -tion in patients this question in a number of specific examples, with com- subjected to anti-arrhythmia surgery. Body surface puter simulations. The calculations assumed a stylized action potential maps (BSPM) and epicardial electrograms potential shape. Propagation velocity was assumed constant wererecordedwith 63 leads in 12 Wolff- but anisotropic (usually x :y was 3::1), although results for Parkinson-White (WPW) patients and in 8 patients withventricular tachycardia (VT) during 14 isotropic propagation were compared. Waveforms arising induced VT episodes. The SMDparameters were from the "true site of origin" were compared to the wave- computed from the BSPM using the same torso model forms arising at the same electrode sites when other "trial which included lungs. In WPW patients, the SMD sites of origin" were used. Most often the RMS error be- trajectoryduring the delta wave accurately tween trial and true waveforms, averaged over the electrode depicted the progression of pre-excitation away sites, was used to characterize the difference between sets of from the location of the accessory pathwayfound trial and true waveforms. For the examples we evaluated, the during surgery, and towards the rest of the heart. error surfaces giving the RMS error as a function of trial site In VT patients, during the first 40 ms of the QRS looked complex and were not simply extensions from known the SMD moved erratically whereas the SMD 1D cases. How many extracellular waveforms were required orientationwas grossly correlated with the the to find the origin, if everything else was known? Surprisingly, location of initial activation: upwards from left-wards from the right ventricle and with anisotropic propagation and low noise, the site of ori- apex, right-wards from the left ventricle. In conclu- gin could be fo'ind unambiguously with only two extracellu- sion, the SMD seemed to depict pre-excitation lar electrodes (waveforms), except for symmetric examples, better than VT, because: 1) ECG baseline is where three extracellular electrodes were required. Having difficult to identify in rapid VT, 2) cardiac more electrodes allowed recognition of the site despite higher activity is more complex during VTbecause of (simulated) noise levels. multiple wavefronts and myocardial infarction.

13E1-1.1. 3 BE1-M.4 RECOVERY OF EPICARDIAL POTENTIALS FROM Inverse ECG Model based 9nimum BODY SURFACE DATA IN A REALISTIC GEOMETRY Relative Entroav Rcconstruk TORSO, BJ Messinger-Rapport and Y Rudy, Case Western principle, C. L. Nikias, D. H. Brooks, Reserve University, Cleveland, Ohio, U.S.A. and J. H. Siegel, Northeastern University, Boston, MA and MIEMSS, University of Inverse electrocardiography has as its goal the reconstruction Maryland, Baltimore, MD. of cardiac electrical events from information obtained noninvasively at the torso surface. The epicardial potential Inresponsetotheshortcomingsoftime- domain distribution, as a solution to the inverse problem, mirrors least-squares(LS)bascdmethods, we introduceto the electrical events within the underlying myocardium in theinverseproblemofElectrocardiographyan more detail than is possible from the body surface potential established frequency domain method, the minimum map alone.Although in principle the epicardial potential relativeentropy(MRE)reconstructionprinciple, solution can be recovered from the complete set of body derivedfromaninformation-theoreticcriterion. surfacepotentials,theill-posednature of theinverse The MRE method reconstructs the power spectrum of problemmakesitnecessary toapply regularization eachepicardialsignal,startingfromtorso ECG techniques to stabilize the inversion. measurementsandincorporatingtheactualtorso gcomctryofasinglepatient. Althoughthis The inverse solution was tested on measured dataon a method does not have the advantage of being able realistic geometry, electrolytic tank with a beating dog's toestimatethetimedomainsignals,itcan heart placed in the anatomically appropriate position. 11.,-; estimate frequency domain parameters which have effect of error in estimating the placement of the heart beenshowninourstudiestobepotentially within the torso tank was examined,it was found that usefulinthedetectionandquantificationof Tikhonov regulartzLiion would permit a I cm errorin importantcardialdysfunctions. We compare estimation, but a 2 cm error could result in significant loss results obtained with the MRE and LS methods as a of resolution.Additional results are presented from a pilot function of both measurementnoi.,e and the size study of the effect of the number and position of leads on andshapeofsubarraysoftorsoscnsors. Our the inversion. Results of these studies suggest that a stylized results show thatthe MRE method appearsto be torso, based on a limited number of measurements,could morerobustandinadditionshowsimproved replace the set of measured electrode location.. An example performance even withasmall, compact subarray of astylizedtorsowithinverse cpicardialsolutionsis of sensors. presented. Work supported by Whitaker Foundation 8D 1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas 169 Biomedical Engineering Scientific Papers

BE1-0.2 l'issue Stimulation Using Magnetic Fields: Current An Introduction to Magnetic Stimulation of theBrain Technolow -Uses and Problems W.A. Tacker, Jr., and and Neuromuscular Tissue, A.T. Barker, LI. Freeston. J.D. Bourland Purdue University, West Lafayette, Indi- and R. Jalinous, Department of Medical Physics and ana, 47907 Clinical Engineering, Royal Hallamshire Hospital.Glos. Magnetic field stimula.1 is used experimentally in sop Road, Sheffield, UNITED KINGDOM. place of electrical field stimulation because it is less painful, and may be safer. Present design of themag- The technique of magnetic stimulation of neuromuscular netic coils produce large fields with diverse, andas yet tissue, developed at the University of Sheffield, England. poorly understood neurophysiological effects on tissue, is tecoming widely used for the clinicalassessment of especially in the CNS. Sites of stimulation, and path- both central and peripheralnerve pathways. This paper ways which mediate the responses have been proposed, will trace the historical development of the technique but to date are unproven. and describe the basic theory behind magnetic stimula- Stimulus duration, intensity and shape have not been tionIts advantages and disadvantages, compared to optimized. The stimulus intensity is proportionalto the con 'entional electrical stimulation, will be considered rate of change of the magnetic fi. :ld and stimulus dura- and the safety of magnetic stimulation of thebrain will tion is equal to the period during which the fieldis be discussed in terms of such parametersas induced elec- changing. The interaction of magnetic fields withtissue tric field, current density, charge andpower deposition. is weak, so very strong pulse,i magnetic fieldsare required. Consequently, due to coil heating, repetition rates must be low and averaging techniques to analyze the neural responses are impractical. Magnetic stimulators designed for high intensity, concentrated fields at repetition rates of several lizare needed to exploit this excitingnew technique.

BEI-0.3 BE1-0.4 Theory and Solutions for Magnetic Stimulation of Tis- Experience with Intraoperative Transcranial Magnetic sues, John Nyenhuis and J.D. Bourland, Purdue Motor Evoked Potential (MEP) Monitoring, 2.I3. Shields, University, School of Electrical Engineering, West ILL. Edmonds, M.P.J. Paloheinio, J.R. Johnson, Lafayette, IN 47907 and R.T. Holt, Division of Neurosurgery, University ofLouis- ville, Louisville, Kentucky. In magnetic stimulation of tissue, current pulsesare applied to a coil near excitable tissue. The currentin Intra nperative spinal cord monitoring withsomatosei, the coil generates a time- varying magnetic field which sory evoked potentials (SSEP's) dues nut monitor thitol induces a stimulating electric field in the tissue. pathways, is very sensitive to anesthetics andrequires Except for the simplest conditions, analytical solution time-consuming averaging techniques. With MB is intractable. Our approach is to calculateapproxi- approved informed consent, in 35 patients havingmap,' mate solutions by digital techniques, and to verify villal surgery, lower limb potentialswere obtained on a them by measurement of the voltage between two radwell MES 10 with transcranial motorcortex magnetic probe tips in a tank of saline solution placed above the stimulationENIG leg responses were obtained ultha coil. neurometrics analyzer (Cadwell). Respiratorygas concentration (02, 002, N2), blood pressure, heartiae, EEG Mudd results can be used to guide the design, ofmag- and upper facial muscle EM0'swere continuous13 111, 0111- netic stimulators. For example, our model slics that tnred, and 90% neuromuscular blockadeN as estinwted efficient energy coupling requires a coil radius ;,.eater with electrically evoked peripheralmotor responses. Pan-- than the coil-to-tissue separation. A second result less magnetic stimulation producedmotor 'espouses ui all from the model reveals a complic."on for theuoer of a awake patients preoperatively without spinalcord pathol- magnetic stimulator of conventional design. The peak ogyIntraoperative responses were obtainedin 71'1 ol intensity of the electric field is locatednear the radius cases MEP amplitude varied from 5-5.900 Mot. ati,1 of the coil, when the coil is near the target tissue. latency duration was 35±3ms. box dose volatile However, the peak intensity occurs further outside the anesthesics reduced MEP size. Wecan moult", it radius of the coil as the distance between the tissue sory (SSEP) and motor (MEP) functions of di. and coil increases. Use of multiple coilscan obviate cord in awake and anesthetized patients Notiniottdol this potentially confounding effect. effects of magnetic stimulation were noted. 9-studies are assessing effectson the brain

170 1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas Biomedical Engineering Scientific Papers

BE1-P.1 BE1-P.2 New methods to investigate the electrogenetic mechanisms Improved Diagnostic Performance of Computerized of atrial arrhythmias in the clinical setting. Multivariate ST/HR-analysis in Myocardial Ischernia, M.Allessie (*), R.Antolini R.Bonamini ( &), H. Sievinen", L. Karhumikit, I. Vuoril, J. Malmivuo, M.Disertori (°), W.Lammers (*), L.Mangiardi ( &), Biomedical Engineering Laboratory, Tampere Uni- F.Ravelli (§) versity of Technology, SF-33101, Tampere, Finland, (§) I.R.S.T., () Dept. Physics, Univ. Trento, (°) Dept. tKanta-H3me Central Hospital, H5.meenlinna, Finland, Cardiol., Trento, Italy , ( &) Ist. Cardiol. Univ. Torino, UKK Institute for Health Promotioi Research, Tam- Italy, (*) Dept. Physiol., Univ. Limburg, Maastricht, NL pere, Finland. In modem cardiological practice, many patients suffer from Regardless of its high sensitivity to detect ae myocar- disabling cardiac arrhythmias, such as atrial flutter or dial ischcmia the ST/HR-analysis is notapplicable as fibrillation. These patients are studied during catheterization a clinical diagnostic test due toits relatively low speci- of the heart and the therapheutical decisions are based on a imprcaed trial-and-error criteria due to the lack of information ficity. The clinical performance of the test is concerning the underlying electrogenetic mechanisms. We significantly by taking into account the propertiesof developed new methods of data analysis blued on the study the ST/HR-slope, the ST-segment valueand shape, of the atrial cycle length variations. We demonstrated that maximum hear'. rate and the ECG lead. By means during atrial flutter this variations :re due to mechano- of the specific multivariate approach the normal exer- electrical effects related to the influence of the ventricular from the contraction on the atrial walls. This physiological cise ECG response is accurately distinguished mechanism allows to find out the mechanism of the ischemic response. rrrhythmia(circus movement) aswellas some The computerized multivariate ST/HR-analysishas no eitctrophysiological parameters of clinical interest. During expectations about the patient material without any atria: fibrillation the analysis of the atrial cycle length loss in the diagnostic performance.Disregarding the fluttuactions allows the evaluation of the refractory period of manifestation of the underlying ischemic diseasethe the atrial tissue. The spatial dispersion of this parameter is sensitivity and the specificity of our method were100% an important factor in the induction of this arrhythmia. All results obtained with the clinical data are compared with and 87%, respectively.The corresponding clinical computer simulations. performance was 93.5%, which implies thesuperiority of the multivariate ST/HR-analysis in respect toother noninvasive ECG based diagnostic tests.

B El- P . 3 BE1-P.4 Computer-Assisteu Cardiac Electrophysiology Kno ledge Representation of Cardiac Electrophysiological During CardiothuracioSurgery,D. S. Buckles*, Phenomena in a Real-Tune Measurement Environment M. E.Harold andP. C. Gillette,Pediatric W.J.Irler (s), R.Antolini (ND, M.Kirchner 0), F.Ravelli (§) Cardiology, Medical University of South Carolina (§)I.R.S.T., 1-38050 Trento - Povo, Italy Conductionsystemaberrations,accessory 0) Department of Physics, University of Trento pathways,ectopicfociandotherelectrical disordersofthemyocardium are frequently In autumafic interpretation of on line measurements in clinical amenabletosurgicalintervention.Effective electrophysiological tests, the knowledge representation of surgeryrequiresspecific}more/ledge ofthe locationsandtypesofdisorders.This electric phenomena of the heart includes numerical and sym- informationisoftenobtainedduring electrophysiologystudiesinthecardiac bolical issues: sequences of time-intervals, phase-cycle catheterization laboratory, but mist be verified relations and stimulation response patterns flora different in the operating roan. Also, electrophysiology protocols. We integrated a knowledge-based test support mappingmustbeaccomplishedimmediately following the surgical procedure to ensure good system into a real time waveform recognition and measure results. A nucrocarputer-based system for mapping and graphic display of data has been developed ment device. It delivers descriptions of the beat-to-beat data for use during pediatric open heart surgery. plots and interpretation hints. The inferences about clear° Signals are acquired from reference and mapping leads with a modified PPG VR-16 recorder, and anc., genetic mechanisms of the arrhythmias and their pharmaco- then processed by an IEM PC/AT with a Data logical control arc obtained in real time by means of heuristics TranslationsDr-2801A A/Dboa? l.The capiter analyzes the relative timing of as many as eight that need only a couple of poi;:tr. rather than complete a reference and mapping signals, then displays the posteriori statistics. Clinical examples concentrating on the results in any of several color -coded cardiac atrial activity (atrial flutter and fibrillation) elucidate our surface maps. Aswellasbeingrapidand reliable thisnethodprividesaneasily "smart instrument" approach. canprehended display of the information required for effective surgical intervention. 191 1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas 171 Biomedical Engineering Scientific Papers

8E1-1%5 BE1 -P. 6 A portable system for digital ANALYSIS OF MULTI-LEAD EXERCISE ELECIRCC.ARDIO- recording of electrocardiogram, GRAFRY, USING SINGULAR VALUE DECOMPOSITION(SVD), G. Baudoin and M.Chaouche*ESIEE, U.Tachinardi* and S.S.FUruie, Heart Instqute of BP 99, Noisy le Grand(France) Sao Paulo, C.P. 8091, S.P.-BRAZIL. "Molter" systems are portable tape recorders which record 24 hours of electrocardiographic signals Although multi-lead ECGs (mIECG) proved to be use - (ECG).Reolacing them by digital recorders presents ful tool to enhance Exercise ECG (EE) accuracy, they several advantages: a) elimination of mechanical greatly enlarge the amount of data to be stored and problemes,b) possibility of simultaneous datapro- analysed, with a great level of redundancy. The oth cessing,c) a better quality of recorded signals. er problem in maECG is to choose the proper leads But the digital recording of 24 hours of ECG sig- to make measurements (like ST segment shift). We nals with a 200Hz sampling rate and 8bits by propose the use of SVD to analyze nIECG of EEs.1t4s sample requires a 14Mbytes memory area. method calculates the principal signals (PS) that The development and testing of several data com- can substitute the measured signals. eor this pur - pression methods applied to ecg have revealed pose we used an orthogonal array of EE signals (X, that the KarhunenLoeve expansion with a lear- Y,Z) of normal and ischemic patients, and calcu - ning phase is very efficient. This technique,used lated their PS in which several EE parameters were in conjunction with preprocessing and segmentat- extracted, i.g. the ST segment area measured in the ion of signals leads to a data rate reduction by first tw., PS suggests a good agreement with inva- a factor of about 10(the relative mean square er- sive angiograPhic findings even better than tradi - ror is less than 27. for a digital rate of200bs-1) tional EE analysis methods. The PS signals can be A new robust QRS detection algorithm has been analysed alone or plotting the loops of a Pair of developped which is particularly efficient oc them. As the first two PS represent more than 951, noisy signals ("sportif"). This algorithm is of the total information, the use of both together based on mathematical morphology. The on line is a good way to represent the electrocardiographic processing can be implemented on digital signal phenomena, allowing the display of the biggest part processors, and the new 1Mbits RAM makes it pos- of the energy in only one plane. An interesting fea sible to develop a digital bolter system. ture is to lower the noise level in the fiwrts PS. When applying SVD, the EE values are normalized, it means that the PS are less affected by anatomical discrepancies between the patients and by the dif - ferent technical protocols used.

BE1-(1.1 BE141.2 Underdamped Multiphasic Sinusoids in the Detrimental Sffects of an Epicardfal Patch on Defibrillation of 100 kg Calves, J. Schuder, T-anathoraclo Defibrillation, .J. P. Wixswo, Jr.°, W. McDaniel*, H. Stoecklo, University of N.G. Sepulveda and D.S. Echt, Vanderbilt Missouri, Columbia, Missouri, U.S.A. University, Nashville, TN, USA Implantable defibrillator systems use one or two Experimental studios indicate that asymmetrical bi- epiaardial patch electrodes insulated on or aide phasic rectangular waveforms are far superior to to minimize energy loss in the lungs and taoeix. critically (or near-critically) damped sinusoids and However, thereis a clinical on that. the to uniphasic trapezoidal waveforms. Nevertheless, presence ofinternalpatch electrode(3) may perceived problems in creating such biphasic shocks increase the energy required for transthoracic in clinical apparatus have resulted in the latter (external) defibrillation due to the patch two waveforms continuing to dominate the clinical insulation. ronically, this problem would be defibrillation field. In the present study, 120 greatest in patients whose nigh internal energy fibrillation-defib illation episodes in which trans- requirements mandate two large patchesfor thoracic defibrillation was attempted in six 92-109 successful internal defibrillation. We have kg calves with a very much underdamped sinusoid in therefore studied the effects of insulated internai which zel.a crossings were separated by 4.05 ms and patches on the external defibrillation threshold the successive peaks were +58, -16.5, +4.7...A were using a 2-dimensional mathematical model ofthe interlaced with 120 episodes involving a rectangu- canine thorax with the predictions that defibril- lar waveform having a leading half-cycle of +38 A, lation requiees 80 % of the heart to be rendered 4 ms and a lagging half-cycle of -18.5 A, 4 ms. The inexcitable by a current density d 35 mA/cm2. underdamped sinusoids, with resistance of 21.0'12.6 Without a patch electrode, 105 Joules were needed I"! and a mean delivered energy of 148 J were success- to defibrill,,e the model transthoracicully. The ful in 88 (73%) of the episodes. Values for the addition of a single 16 cm2 insulated patch to the rectangular waveforms were 21.1.12.5 151 J, and LV apex increased the required energy to 184 J. 91 (76%). Since the success rate for the two wave- A.arger patches and the use of two patches further forms are not significantly different, the results increased the transthoracic energy requirements. suggest that the effectiveness of the underdamped While our 2-D model can only approximate the multiphasic sinusoidal waveform is comparable with transthoracid current distributionsin a3-D the biphasic rectangular waveform and, hence, also thorax,we conclude that large epicardial patch appreciably superior to the r.hocks used clinically. electrk. s can significantly increase trans- 19h2acic dcfibrillation needs.

172 1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas Biomedical Engineering Scientific Papers

BF-1-9.3 8E1-8.4 FIBRILLATION THRESHOLD AiM VENTRICULAR APreliminary Study on the Index WEIGHT. E Ruiz, J Rodriguez, I Medina or the Excitability of Nyocar- and ME Valentinuzzi. Bioingenieria, dium,tin -yin Xu, Ye-cho Huang*, Universiaad Nacional de Tucumbn, cc 28 XianMedicalUniversity, Xian, 311C 2, 4000 Tucum6n, Argentina. China. The objective was to search for a Geddes reported that the membrane time cons- possible relationship between ventricular tant T of myocardial cell derived from the ex- we.lgth (VW) and fibrillation threshold perimental data ofpacing and defibrillation (FT). We used 14 small cats (average VW are reasonably similar, and T equalschronaxie 2.5 g, SD 0.4), triggering fibrillation c approximately. Huo Zong Lian advanced that with single 5 ms rectangular pulses the standard electrical quantity a he named was synchronized with the ECG and adequately an actual index for the excitability of tissue delayed. Bipolar hook electrodes were or cells. The experiments about excitabilityof placed on the basal region or the heart. myocardium (EM) have been doneon 5 patients FT was scaled to VW obtaining an overall and 7 dogs with the pacing system analyzer we averageof 7.9 mA/g (SD 3.7) for 52 developed. EM was changed by drugs and strength measu..ements. Average results from tnree -duration curves were measured by stimulation other groups of larger animals were: 5.9 threshold, varying rectangular currentpulse mA/g (SD 1.4) in 4 dog puppies of 5 g VW durations(0.01-2.5ms). Zinn the constantsband (SD 0.9),'.0 mA/g (SD 0.6) ir. another 4 a in WEISS f)rmula werecalculated, and c was puppies of 20 g VW (SD 3.8) and 0.24 mA/g derived from a/b. Results: (1) a decreases or (SD 0.08) in 11 adult dogs 64 g VW (SD increases as EM rises or reduces respectively, 14.8).We did not run any statistical butband c do not. (2) The average T of pa- comparison between groups, but these tients and dogs are 1.311+0.406(SD) and 1.069+ results show a tendency, i.e., as 0.322(SD) respectively. Considering strength- ventricular weigth increases, duration curve for energy and the function of fibril.,ation threshold decreases. Thus, energy compensation inpacemaker, thepulse hypertrophied hearts would fibrillate duration of pacemaker had betterbedesigned more easily than normal size ones. about T/2. CONICET and CIUNT grants.

BE1-a.5 A SEQUENTIAL ALGORITHM FOR DETECTION OF Tine Frequency Distribution Mappings of Event -Related VENTRICULAR FIBRILLATION - N. Thakor and Z. Potentials in Kornhuber Experiments. Yi-Sheng, Biomedical Engineering, Johns S.Salinat4+, A.Urbano Hopkins University, Baltimore, MD, USA. 1st.di Fisiologia timana-University di Roma "La Sapienza"-Piazz.le A.Moro 5-00100 ROMA Ventricular fibrillation (VF) is a life- +Dip. di Infcrmatica e Sistemistica-UniVersita di Roma "La threatening arrhythmia that must be imme- Sapienza -Via Eudossiana 18- 00184 ROMA diately and accurately detected.An automatic implantable defibrillator delivers a Spectral analysis is one of the most used tools in signal lifesaving high-energy shock to the heart processing, however the underlying hypothesis, for its use when VF is detected.An algorithm is is the steady state behaviour of data to be processed. In presented that reliably discriminates VF many cases, the biological signals are deeply non from other ventricular arrhythmias. stationary. this motaates the need for the data description VP signal is first compared with a in a time frequency joint domains. We attempt this non threshold (20% of peak) to obtain a binary staaonary spectral analysis to the signals recorded during a signal.The mean value of time interval Kornhuber experiment, we want to study changes in the between threshold crossing is smallerin amplitude of 'Fourier coeffiLents at different frequencies, tachycardia and VF than in aormal rhythm. dad tcmporal rc.lation5hips. At this purpose :ye The variance of these intervals is smaller recorded Kornhuber potentials in normal and parkinsonian in tachycardia than in VF.A test is subjects form 30 channels placed on the scalp.We constructed that discriminates between mean computed the time frequency distribution of the signals by and variance of threshold crossing different algorithms. Data windowing is performed by intervals of different arrhythmias. A Hamming or split-cosine functions. A PDP11/23 plus sequential algorithm, based on Wald's computer was used for data processing. hypothesis technique,is constructed. The 3D graphics shows the time frequency distribution for decision to detect VFis made for every signal recorded. Amplitude and power color successive, 1 second segments of sampled mappings were also performed, at frequency of interest, in data. The detection thresholds are order to follow the temporal evolution of Fourier established so as to obtain desired false components over the whole scalp. positive and negative rates. In a study of 179 episodes, VFdetection accuracy was 95% after 3 segments,97.6% after 5 segments and 100% after 7 segments.

1988 World Congress on Medical Physics andBgbn9talEngineering o San Antonio, Texas 173 Biomedical Engineering Scientific Papers

BE1-R.2 BE1-R.3 Estimation of ERP Parameters basedon the Time- ADAPTIVE FILTERING OF TRANSIEPT EVOKED sequenced Adaptive Filter. N.Takano* and K.Yana. POTENTIALS - N.Thakor*and C.Vaz, Hosei University, Koganei Cit.", Tokyo Japan. Biomedical Engineering, The Johns Hopkins University, Baltimore, MD 21205, USA. Accurate estimations of latency, amplitude and waveform of ERPs(Event Related Potentials) are Evoked potential (EP) monitoringmay be difficult becauseof the"- low signal(ERP)-toan important technique in monitoring the noise(back ground EEG) ratio. Averaging method aLd status of the brain in neurological surgery the adaptive correlation filter have been utilized and intensive care.This requires that for these estimations. As an improved method, we transient and time-varying events be would liketo propose the folloding procedure. captured.An adaptive filtering (AF) Steel: Average original EEGdata tomake a techniqueis presented that rapidly reference waveform For adaptive filtering. step2: improves signal-to-noise ratio (SNR). Compute time dependent filter coefficients on and EP signals are modelled bya Fourier around the peak point of the reference waveform. linear combiner (FLC). Fourier coefficients step3: Obtain the amount of time-shift for each for a truncated series are estimated. A original data which gives the best matchbetween least-mean-square (LMS) algorithm is the reference waveform and filteroutput. step4: employed to recursively minimize the mean- Average time-shifted original data according to squared error(MSE)between successive the amountgiven by step3 tomake an improved ensembles and the FLC model of EP.The LMS reference waveform for adaptive filtering. Repeat algorithm permits improvement in MSE,as steps2-4 until the convergence condition basedon well as adaptation to time-varying events. the amount of time-shift in the step3 will be met. Comparison with ensemble averaging reveals steps: Pe-form the ERP signal emphasis for each that AF improves SNR more rapidly and time-shifted data to obtain the estimate of single tracks transient changes in EP amplitudes. ERP waveformand it's amplitude, by the time- This research deals with two theoretical sequenced adaptive filtering technique. issues: 1) optimal selection of model The preliminary experiment showed the improvement order, and 2) selection of fastconvergence in the estimations of ERP latency and amplitude. rates while maintaining stability. This method may also give us the estimate of AF technique is currently being tested single ERP waveform in each original data. in two protocols: neurologic response to 1) induced hvnoxia in dogs: and 2)induced hypotension in human volunteers.

BE1-R.4 BE1-R.5 A Two Dimensional Approachon the Problem of Filter- Methods for single sweep processing of ing Evoked Potentials, 0. Svensson, Telecommunication evoked and event-related potentials Theory, University of Lund, Box 118, S-221 00 Lund, Sweden. D. Liberati, S. Cetutti, G. Pavesi CNR-CTS, Dipartimento di Elettronica. When it is of interest to follow changes ofan evoked Politecnico, Milano, Italy response it is nessesary to achieve an interpretablere- sponse with few trials. Filtering methods can be used The variability occurring in successive evoked (EP) both along the traces andcross the traces. The LMS or event-related (ERP) potentials obtained wizh the same protocol from thesamepatient adaptive filter algoritm is used for filtering alongthe becomes more and more evident with the increa- traces. The bias as well as the sensitivity to changes in sing of the complexity of the involved neurophy- the response for the LMS adaptive filterare analysed. siological or psychological task.Moreover, in Evoked responses often contain higher frequenciesat the monitoringduring neurosurgery, a variety of beginning of the response than towards the end.The sweep-by-sweep changes in the main parameters of response is therefore divided into different time slices the evoked potentials canprovide important feedback information about the correctnessof where different adaptive filters operate in thedifferent the surgery itself. In such slices. cases, a method developed fordetecting the sweep-by-sweep Averaging is the common method for filtering variability becomes an important tool. cross the The traces. This method can be viewed proposed techniques are based on theaddi- as a FIR-filter having tive superpositionof signal andnoise (bac- a uniform impulse response. By using an IIR-filter with kground EEG), and on a model of the EEG itself an exponential impulse response instead, higher weight as an AotoRegressive process. can be given to the latest trial and hence changes in the On the basis -f this model, a few parametersare response are easier to follow. The time constant of this sweep-by-sweep extracted which try todescribe cross trace filter can easily be tuned to obtain an optimal the dynamics of the neural response. Resultsobtained on both simulatedand real trade off between time constant and signalto noise ratio signals in the response. Results based confirm the capability of the model for on simulated data are singlesweep classification,besides showing presented where both white and EEG-likebackground improvement in the S/N ratio in respect to the noises are used. 194traditional filtering. 174 1988 World Congress on Medical Physics and BiomedicalEngineeringSan Antonio, Texas Biomedical Engineering Scientific Papers

BEi-R.6 BE2-A.1. A Linear System's Approach to the Rela- AUTOMATED DELIVERY OF I.V. ANESTHETIC tionship between Visual Evoked Potentials AGENTS: MODELING CONSIDERATIONS, James R. Jacobs* and Peter S. A. Glass, Department and the Prestimulus EEG. of Anesthesiology, Duke University Medical Center, Michael E. Brandt* and Ben H. Jansen, Univer- Durham, NC 27710 USA sity of Houston, Dept. c: Electrical Engineering, Houston, TX 77004, USA. Until the phannacodynamics of general anesthesia have been better quantified, reliable and versatileclosed-loop Various signal theoretic procedures (e.g., averaging, administration of intravenous anesthetic agents will not be adaptive filtering) can be uccd to improve the signal-to- possible. Alternatively, we feel that open-loop delivery of noise ratio of evoked potential (EP) measurements. It is these drugs is optimally realized with a pharmacokinetic conceivable that methods based on a model of EP gener- model-driven system, where the physician citrates the ation could lead to a dramatic increase in performance. theoretical drug plasma concentration to achieve the desired Specifically, knowledge of the relationship between the clinical effect.This approach requires techniques for EEG in'-lediatley preceeding the EP and the shape of the EP could be of major importance. This relationship is modeling both the pharmacokinetics of the drug(s) being largely unknown. Previous work in this area has generated infused and the plasma concentration-effect relationship. some support for the hypothesis that the EP is essentially With regard to the former, discrete and analytical approaches a modified version of the EEG present at the time of the to the real-time simulation of compartmental pharmacokinetic stimulation. models have been developed and are readily implemented in In the present study, healthy volunteers were sub- software; real-time sensors of actual drug plasma concen- jected to a long series of randomly presented, identical tration are not available and, in our opinion, are not flash stimuli over a one hour period. Clustering was done necessary. Although elegant concentration-effect models on the 1 sec prestimulus intervals and the poststimulus have been suggested for some dnizs, fcr i.v. anesthetic epochs co.responding to a given prestimulus cluste, were agents, the determination of Cp50's (analogous to M A.C. used to compute an average EP (AEP). The relationships for inhalazional anesthetic agents) by logistic regression between prestimulus EEG and VEP shape thus cli.ctnered modeling has proven to be the most useful in clinical practice. will be discussed.

BE2 -A. 2 BE2-A.3 tobdeling and Simulation of the Effects Hydrogels for Controlled Delivery of Drugs ofAntiarrhyttunic Agents, T. C. Jannett: T. J. Laughlin, Schering-Plough Corpora- L. C. Sheppard, and G. N. Kay, The tion, Memphis, TN, and R. L. Dunn*, Vipont University of Alabama at Birmingham, Laboratories, Inc., Ft. Collins, CO. University Station, Birmingham, AL 35294 Hydrogel alloys canbe prepared by blending select hydrophobic and hydrophilic polymers. The Pharmacologic control of a physiological varia- integrityof the alloy is determined by the ble requires knowledge of the dynamics ofthe noncovalent bonds (ionic, dipole-dipole, and response to drug infusion. Mathematical modeling hydrogen bonds) and hydrophobic interactions and simulation were employed in the development formed between polymers. of an automated system for lidocaine administra- We have been investigating the controlled-release tion. A model of the rate of premature ventricu- and swelling properties of blends of polystyrene lar contractions (PVC rate) and its response to maleic anhydride and polyvinyl acetate). The lidocaine infusion was developed in system iden- water sorption rate, water sorption capacity, and tification experiments. The model was used to thermoplasticity can be controlled by selecting design a closed -loop controlsystem forthe the proper composition and processing procedure. regulation of the PVC rate at a desired set We have observed sorption capacities for up to 30 point. Infusion rate constraints were developed timezthe w-;ght at Inv. We have algr) basedon the analysis of a phaxmacokinetic model observed an apparent zero-order release of to limit the predicted lidocaine serum concen- tetracycline and insulin from tablets made from tration to a safe level during operation of the these alloys. The release appeared to be closed -loop system. Operation of the system was controlled by an erosional mechanism. validated in computerskrailationsofthe PVC rate and the lidocaine serum concentration during closed-loop drug infusion.

195

1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas 175 Biomedical Engineering Scientific Papers

BE2-A.4 BE2-B.J. Automatic Computer Control of Prostacyclin - Releasing Sutures for Neuromuscular Blockade, T.J. Quill*, the Prevention of Venous Thrombosis P.S. Glass, J.R. Jacobs, and J.G. R. L. Dunn; C. A. Eddy, and L. E. Laufe Reyes, Duke University Medical Center, Ft. Collins, CO and San Antonio, TX Durham, N.C., 27710 The prostacycin analogue Iloprost from Schering AG, Berlin, West Germany was The design and development of an auto incorporated into polycaprolactone, a matic computer controlled infusion sys Iiiodegradable polymer. The blend of drug tem for the administration of either and polymer was extruded as a mono-klament vecuronium bromide or atracurium besy on a melt indexer and drawn to a size 9-0 late for human surgical relaxation is suture. The fiber containing 0.25 ug of described in detail. drug per centimeter of fiber was armed The system utilizes a PuritanBennett with a 100-um 3/S circle taper-point evoked EMG unit and an Intel 8088 micro needle. Control sutures were prepared processor based modified adaptive PID using polycaprolactone and no drug. controller driving an IVAC medical pre Bilateral transection and anastomosis of the common femoral vein was conducted in cision infusion pump, and successfully thirteen rats with the drug prcstacyclin- maintains the evoked EMG consistently within narrow tolerances. The controller releasing sutures on one vessel and the control sutures on the contralateral utilizes an internal pharmacokinetic vessel. An examination 24 hours post- model based upon measured drug charac operatively showed the prostacyclin- teristics in order to provide noise sutured anastomoses were patent with a detection and elimination. Results of uniform absence of thrombosis. The system perfDrmance upon 40 human sub control anastomoses gave five patent and jects are presented. The principles de eight occluded vessels with a dense, well- scribed could be easily utilized in the or-ganized fibrinous clot that filled the development of other automated drug entire lumen. These results demonstrate delivery systems based upon closed loop the effectiveness of the local delivery of pharmacodynamic feedback. prostacyclin from the suture in inhibiting thrombus formation.

BE2-B.2 BE2-B.3 Polymeric Implants for Delivery of Anti- Fuzzy Control of Arterial Pressure by Drug cancer Agents, J. D. Strobel*, T. J. Infusion Using a Fuzzy Expert System'FLOPS Laughlin, F. Osrroy+, M. B. Lilly+, B. H. Perkins, and R. L. Dunn, Southern H. Ying: L.C. Sheppard, and D.M. Tucker Research Institute and University of Alabama at Birmingham+, Birmingham, AL. 1) Department of Biomedical Engineering University of Alabama at Birmingham Birmingham, AL, 35294, USA. We havepreparedprototypeimplants forthe controlled release of three anticancer agents: 2) Kemp-Carraway Heart Institute Birmingham, AL, 35234, USA. misonidazole (MSND), adriamycin (ADM), and cytosine arabinoside. The studies were Fuzzy controllers are conceptually rule- undertaken to determine the feasibilityof delivering the drug directly to malignant based and have the ability to control proces- tissue to increase efficacy and reduce ses without explicit mathematical models. In this paperwe propose a new type of fuzzy toxicity. Itwas anticipated that the controllerwhich is not based on the existing drug-releasing implants. would beused in fuzzycontrol theory. Instead, this fuzzy combination with radiation or hyperthermia controller is based on general-purpose Fuzzy treatments. Therefore, for implants that LOgicProduction SystemFLOPS. Human mean contained MSND orADM, orthodontic wire wa., coated with the drug and a polymeric excipient. arterial pressure was regulated by this fuzzy Drug and polymer were matched to achieve the controller, in digital computer simulation, by controlling the infusion rate ofthe drug, targeted rate and durationof relt.se. A sodium nitroprusside. In the presence of non- variety of polymers were examined for stationary background arterial pressure varia- permeability characteristics, including a tion, mean arterial pressures ofdifferent polypropylene, a polyethylene, an ethylene patientswith different sensitivities to so- vinyl acetate copolymer, a polyether polyurethane; polycaprolactone, a hydroxyethyl dium nitroprusside were controlled satisfactorily. Furthermore, the simulated results using methacrylate/methacrylate copolymer, and a a first-order linear model demonstrated that DL-lactide/glycolidecopolymer. Fabrication processes included melt spinning and dip coat- the performance of the proposed fuzzy control- ing. ler was superior to that of a fine-tunedPro- 196 portional-Integral (PI) controller.

176 1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas Biomedical Engineering Scientific Papers

BE2-B.4 BE2-C.1 Spermicide-Releasing Vaginal Diaphragm. Automatic control of halothane anesthesia, F.J. Sasse *, G.F. Schils Benjamin S. Kelley*, Richard L. Dunn, Southern ResearchInsti- and V.C. Rideout, University of Amelia G.Potter, Wisconsin, Madison, WI 53792, USA. tute,2000 9th Avenue S., Birmingham, AL. a viable and The design and laboratory implementation of Vaginalbarrier diaphragms are a controller for halothane anesthesia in dogs is convenient means of reversible contraception, but acceptance is limited because of the messy described. Difficulties with the control of mean and maintenance. We have arterial pressure (MAP) during linear (PI) creamsor jellies control prompted the design of an ON/OFF studied the release kinetics of the spermicide frombiocompatiblethermo- controller. The ON/OFF controller provided fewer nonoxyno1-9 (NN9) This has included tradeoffs between stability and speed of plastic elastomers. that can be response. examining manufacturing processes Recognizing that the desired and undesired used to economically manufacture the elastomer- spermicide-lubricant/rate modifier blend. The effects of an anesthetic span several physiological systems, a method was devised to goal has been to develop a convenient barrier inexpensive, effective, coordinate the control ur two physiological contraceptive that is variables (MAP and an EEG measure) with the and requires a lower degree of user motivation. using either diaphragms or single anesthetic agent. Animal studies conducted in The design criteria of a controller for medicated rodshavealsobeen inhalation anesthetics are discussed in detail. baboons or rabbits to screen for efficacy and Clinical issues considered are the possible safety. We have found that effective doses of from several different clinical utility of an automatic controller, the NN9 can bedelivered a 24-hrperiod clinical timeframe, and the general biocompatible elastomers over the physical characteristics of the clinical control of the without greatly compromising properties of the elastomers. Our goals have depth of anesthesia. Related control issues been to release about 30 mg of NN9 from the considered are the choice of a control method, 1 hr and about 150 to 180 mg over the desirability and Fmktations of adaptation, diaphragm in We hope to soon begin Phase I the choice of output signals, and the user a 24-hr period. clinical trials. (Sponsoredby the interaction. human NationalInstitute of Child Health and Human Development.)

BE2-C.2 BE2-C.3 CACI II: PHARMACOICINETIC MODEL- Open and Closed7Loop Control of Premature DRIVEN DRUG DELIVERY, James R. Jacobs,* Ventriallar Contractions, T. C.Jannett: Peter S.A. Glass, Eric D. Hawkins, J.G. Reyes, G. N. Kay, and L. C. Sheppard, The Department of Anesthesiology, Duke University University of Alabama at Birmingham, Medical Center, Durham, NC 27710 USA University Station, Birmingham, AL 35294 CACI H (figure) is a pharmacokinetic model-drivendrug delivery system for intravenous anesthetic agents. The output An automated drug delivery systemwhich incorpo- (A) of a 3-compartment pharmacokinetic simulation:1 rates open -loop and closed-loop lidocaineinfu- compared with the setpoint plasma concentration (B).A sion for the suppression of premature ventricu- simple infusion algorithm acts on the error signal (C) to lar contractions (PVCs) was tested inclinical generate an optimal drug infusion rate (D) to[theoretically] trials. On the first day of the study, an open - maintain the setpoint plasma concentration during a specified loop fusion regimen based on a pharmacokinetic period of time. Drug is delivered to the patient (E) andthe model 9S used to quickly achieve andmaintain a model is updated (F). The physician may adjust the setpoint therapeutic lidoceine serum level. On the second as often as required to titrate a stableanesthetic. CACI H has day of the study, closed -loop control was used been used to deliver fentar.yl, sufentanil, alfentanil, propofol, to regulate the PVC rate at a setpoint level sufentanil/ midazolam, or fentanyl/midazolam as the primary after a therapeutic serum concentration was or adjuvant anesthetic to more than100 patients. The clinical achieved by open -loop infusion. The drug infu- utility of CACI H has been excellent, and we believe that sion rate was constrained during closed -loop devices of this nature will play a major role in the future of control to keep the predicted serum concentra- intravenous drug therapy in a variety of clinical settings. tion between upper and lower limits. Measured concentrations were cared to predicted values. During closed-loop druginfusion, satisfactory regulation of the PVC rate was achieved with predictable serum concentrations.

197 1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas 177 Biomedical Engineering Scientific Papers

BE2 -C .4 BE2-C.5 Insulin Administration with Implantable Automatic Control of Anesthetic Delivery and Pumps by Various Routes, W. Zingg* and Ventilation During Surgery, R. G. Ritchie*, E. M.V. Sefton, Research Institute, The A. Ernst, B. L. Pate, J. D. Pearson, L. C. Hospital for Sick Children, Institute Sheppard, University of Alabama at of Biomedical Engineering and Department Birmingham, Birmingham, AL 35294 of Chemical Engineering, University of Toronto. Use of a closed rebreathing circuit for anesthesia delivery offers several advantages: conservation of Parenterol administration of insulin is mandatory. anesthetic agent, natural heating and humidification of Traditionally it is given subcutaneously once a inspired gases, less pollution, and improved monitoring. day.To cope with individual metabolic require- However, the technique requires careful control of the ments, multiple injections and continuous infusion fresh gas and anesthetic delivery.An anesthesia by wearable or implantable pumps have been used; delivery system has been developed which the infusion can be subcutaneous, intravenous or automatically controls fresh gas delivery, anesthetic intraperitoneal.We have assessed these routes in delivery, and ventilation in order to regulate circuit a homogenous population of rats, and found a wide volume, oxygen concentration, end-tidal anesthetic variability. The route is not important because concentration and end-tidal PCO2 in a closed breathing of the presence of normal compensatory factors. circuit. The system makes available the advantages of In diabetics however, there is a deficit of these the closed-circuit technique without encumbering the factors and as a result the metabolic control is anesthesiologist with its more demanding control tasks. defective because of a lack of counterregulatory During animal testing, the system's response to action. Both constant iafusion rate and program- step changes in end-tidal PCO2 setpoint, end-tidal mable - telemetry commanded implantable pumps are anesthetic setpoint, and oxygen uptake were recorded. useful in selected patients, but without a All controllers were stable and exhibited clinically continuous glucose sensor to create a closed-loop acceptable response times.End-tidal anesthetic system the application is limited. Other major concentration could be quickly increased without problems are power consumption and the biological opening the breathing circuit.During circuit volume reaction to the presence of the implant. regulation, oxygen delivery tracked sudden changes in oxygen uptake, making the system a good monitor of changes in oxygen uptake.

Funding: Ohmeda Grant #RMR080283

BE3-A.1 BE3-A.2 The Distributed ' transmural Parameters of LL Integrity of myocardial fiber structure maintained Function, Metabolism and Perfusion, R. Beyar by fiber load induced local growth. T. Arts, P. and S. Sideman, Biomedical Engineering, Bovendeerd, F. Prinzen, R.S. Reneman, University Technion-IIT, Haifa, Israel. of Limburg, 6200 MDMAASTRICHT, The Netherlands.

A model approach to the inter-elationship between A change in hemodynamic load induces adaptations left ventricular (LV) mechanics, perfusion and ofthe heart within afew weeks. Despitethis ene getics in a symmetrical ellipsoidal shape is flexibility in local growth, integrity of global presented. The complex model is composed a of myocardial fiberstructureis maintained.In a hierarchical setof tightly linked models which computer simulation two hypothesis on the under- combine the geometry of the LV, its fibrous struc- lying growth control mechanism were investigated, ture, radial electrical activation, and properties both having in common that muscle fiber growth/ of the muscle fibers to yield the stress and pres- degradation isinduced by maximum musclefiber sure distribution within the LV wall, which in turn stress. When controlling end-diastolic sarcomere affect the perfusion and energetic models.The length (SL), on the long term the structure ap- energetic model links the mechanical parameters of 'arsto be unstable. However, when controlling an individual fiber contraction via its lengthand .ie number of sarcomeres along a myocardial fiber force to yield the local and global oxygen consump- by the amount of sarcomere shortening dining tion. The perfusion model utilizes the pressure systole (positive feedback), the transmural course distribution fror the mechanical model, combined of muscle fiber orientation is found to bevery with the oxygen demand as specified by the energe- stable selfcorrecting and close to anatomical tic model, and uses basic laws of microcirculation findings.Thus, the global,complicated cardiac dynamics, in order to simulate the distribution of structure can be maintained by the control signals myocardial perfusion across the LV wall. Combining fiber stress and sarcomere length, which are lo- these three models yields the temperature distribu- cally available at the cellular level. tionacross the LVwall. This model approach provides a tool to study the complex interrelationships between the different aspectsof cardiac contraction.

178 1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas Biomedical Engineering Scientific Papers

BE3-A.3 BE3-A.4 Accurate Prediction of Stroke Work from Sy:,tolic ASSESSMENT OF LOCAL MYOCARDIAL FUNCTION IN VIVO - and Diastolic Pressure-Volume Relations in the R.S. Reneman*, F.W. Prinzen, C. Augustiim and T. Isolated Canine Heart, W.L. Maughan ,D. Kass, K. Arts, Departments of Physiology and Biophysics, Sagawa,Johns Hopkins Medical Institutiors, University of Limburg, Maastricht, The Netherlands Baltimore, MD 21205. In vivo global left ventricular (LV) function can It would he valuable if cardiac performance be determined relatively easily, but reliable as- assessed under on set of conditions could be sesnment of local myocardial performance is still applied to predict performance under other problematic, mainly because the devices used I.:At:- conditions. If end-systolic pressure-volume fere with myocardial fiber function. Local fiber relationship is modelled by a linear function stress is generally estimated using mathematical (Pes-Ees(Ves-Vo) and end-diastolic pressure-volume r e lat opsh to by an exponential function models of LV wall mechanics. The values of intramyocardial pressure (IMP), a variable of importance (Ped-AeBOred-vo).), stroke work can be predicted in studying the relation between myocardial con- (SWp) under various filling pressures (Ped), traction and perfusion, are still under debate. afterloads (Ea) and contractile states (Ees,Vo). To test the model predictions, we compared SWp There is agreement about the positive gradient of However, it is still with measured stroke work (SWm) in 12 isolated IMP from endo to epicardium. canine hearts ejecting against a computer unknown whether IMP in the endocardial layers is about equal to or significantly higher than LV cav- controlled loading system. For Ped from -5 to 31 ity pressure, although recent micropressure meas- mmHg (mean 6.8+8), Ea from 2 to 32 mmHg /ml (mean urements and model predictions support the former 9.4 ±4.8), and Ees from 2.6 to 11 mmHg/m1,, (mean idea. In vivo epicardial fiber shorteninc can be 5.4 +1.3), SWm-1.09SWp + 20.7 mmHg-ml assessed relatively accurately from the deformation n-213, SEE-126). Multiple regression showed small parameters, as determined locally with x-ray or independent influences of Ea,Ped, and dog on the video techniques. With the latter, epicardial fi- residuals, probably related to slight changes in ber shortening can be assessed simultaneously in the shape of the pressure-volume loop with each 35 areas of the LV free wall, allowing investiga- variable. tion of, for example, the interaction between isch- Thus the systolic and diastolic pressure emic and non-ischemic areas. From the epicardial volume relations can be modelled and used to deformation parameters, endocardial fiber shorten- accurately predict cardiac performance over a wide range of different preloads, afterloads and ing can be estimated during normoxia and ischemia. contractilities for a given isolated heart.

BE3-B.1 BE3-B.2 ESOPHAGEAL CARDIAC PACING IN MYOCARDIAL ISCHEMIA: Fluid-Fiber-Collagen Description of the Myocardium EFFECTS ON LEFT VENTRICULAR AND ARTERIAL SYSTEM Applied to Different Left Ventricular Geometries, J. FUNCTION. J. K-J. Li: Dept. of Biomedical Engineer- Ohayon, M. Lewkowicz, and R.S. Chadwick*, ing, Rutgers University, Piscataway, N.J. 08855- Theoretical Biomechanics Group, Biomedical 0909 USA. Engineering and Instrumentation Branch, Division of Different modes of cardiac pacing and pacing Research Services, National Institutes of Health, sites have been shownto alter the performance Bethesda, MD 20892. of the left ventricle. This paper reports the effects of transesophageal cardiac pacing during myocardial ischemia. The microstructure of the collagen sheath or weave Experiments were performed on open chest anes- surrounding a myocyte and the collagen struts interconnecting thetized dogs. Bipolar pacing catheter was uti- neighboring myocytes is incorporated into a fluid - fiber - lized afterestablishing the strength-duration collagen continuum description of the myocardium. The curve for threshold pacing. Pacing was initiated sheaths contribute to anisotropic elasticity, while the struts after one hour occlusion of the leftanterior contribute to an isotropic component. Elastic moduli of the descending coronary artery. Left ventricular and composite myocyte-sheath complex and tile strut matrix are aortic pressures, aortic flow, segment length estimated from existing passive biaxial loading data from and the electrocardiogram were recorded. sheets of canine myocardium. The contribution of the sheath to Results show that a higher heart rate during the elasticity of the myocyte - sheath complex is critically pacing was accompanied by a slight decrease in dependent on the helical pitch angle. mean arterial pressure, flow, and segmentshorten - Calculations are presented for different geometrical configurations: a thick-walled finite cylinder; a thick -walled :Lng. Arterial pulsetransmission characterized by fawardandreflected travelling waves also sphere; and any thick-walled truncated body of revolution showed a small change. Thus, esophageal cardiac close to a cylinder, to show both the effects of collagen and assumed geometry on ventricular function. We found that the pacing alters regional myocardial and arterial system dynamics only to a limited extent. collagen strut matrix tends to : increase myocardial tissue pressure during systole, with endocardial tissue pressure exceeding left ventricular pressure; decrease tissue pressure during diastole, and thus facilitate myocardial blood flow, and; aid filling during ventricular relaxation by providing a suction effect relying on a release of stored elastic energy from the previous contraction. 199 1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas 179 Biomedical Engineering Scientific Papers

BE3-B.3 BE3-B.4 Analysis of the Left Ventricle Mechanics. Delayed end-ejection as a determinant of enhanced A. MOUTTAHAR, and C. ODDOU: relaxation rata during an increase in stroke volume

L.M.P, Universite Paris 12, Creteil, France. Masatsugu Mori , Masafumi Kitakaze, Yoshio Ishida, Hideo Kusuoka, Akira Kitabatake, Michitoshi Inoue+, Ananalytical model was developed considering the Takenobu Kamada, Hiroshi Abe++ left ventricular wall as an axisymetric and pres- The 1st Dept. of Med. and Dept. of Medical surized thick ellipsoid shell in mechanical equi- Information Science+, Osaka Univ. School of librium. Themyocardium is supposed to behave Medicine, Osaka National Hospital like a continuous medium essentially made up of contractile fibers embedded in a fluid matrix. Ventricular relaxation rate is enhanced when stroke Thepitch-angle of the fibers, viewed an geodesic volume is increased. To test the hypothesis that lines on the wall surfaces, is varying with the enhanced ventricular relaxation in this condition localisation within thewall. A linearrheo- is due to delayed end-ejection associated with an logical behavior for the fiber at their different increase in stroke volume, the left ventricular stages of excitation was assumed. volume and ejection of the perfused isolated canine Drawingfrom canine experimental data the geomet- hearts (cross - circulation) were controlled by a rical and rheological model input parameters, servo-pump, andventricularrelaxation ratewas this analysis helps determine the local stresses assessed by the time constant (T) of isovoiumic and strains within thewall. In particular, it left ventricular pressure decay in various was shown that the fiber length varies during the conditions. As stroke volume was increased, I was cardiaccycle in the range between -10% to 20%, decreased (relaxation was accerelated) only when relatively to the rest reference state. Moreover, end-sytolic volume was concomitantly decreased. an increase in the shell longitudinal curvature When ejection velocity was increased, the induces a uniform fiber straindistribution relaxation rate was increased in almost all within thewall. The systolic torsional motion experiments. However, when the timing of end- and structure eccentricitywere found to be ejection eas clamped, T was not significantly greater than thediastolic ones. In agreement altered even with a large change in ejection with ultrasound measurements, the model predicted velocity. These results indicate that neither the that the relative variations during the overall extent nor velocity of ejection affects the cycle of the internal minor diameter, internal ventricular relaxation rate but the delay of end- major axis and equatorial wall thickness are ejection enhances the relaxation rate when the respectively 337.,147. and 307.. stroke volume is increased.

BE3-C.1. BE3-C.2 Forces determining transmural distribution of Coronary Pressure-Inflow Relationships, myocardial blood flow, J.I.E. Hoffman*, A. Flynn, R. E. Mates: SUNY at Buffalo, G. Aldea, D. Coggins, W. Husseini, R. Austin, Buffalo, NY 14260, USA. University of California, San Francisco, California, U.S.A. The mechanisms responsible for observed coronary pressure-inflow behavior remain controversial, Left ventricular systolic intramyocardial tissue pressures particularly the observation that diastolic equal cavity pressure below the endocardial (deep) surface inflow ceases at pressures above right atrial and are near atmospheric below the epicardial surface. presstre. Because of blood vessel compliance, This implies that in the superficial left ventricular muscle, phasic inflow differs from flow in the microcir- there should be a large pressure gradient from extramural culatioo. Since the latter is difficult to mea- coronary arteries to tissue, and therefore that there should sure, models have been proposed to relate inflow be significant systolic perfusion of the superficial muscle to microcirculatory flow. We have utilized the from the extramural arteries. However, studies of the ca- flow response to coronary pressure perturbations pacitance of these extramural arteries indicate that little of to characterize pressure-inflow relationships. A the blood that enters the arteries in systole actually perfuses lumped model consisting of a viscoelastic com- the myocardium. The discrepancy can be explained by pliance, a resistor and a back pressure has been postulating that in systole, the increase in deep intramyo- found sufficient to describe sinusoidal and ramp cardial pressure raises intravascular pressures in the deep perturbations in prolonged diastoles. Rapid ramp microvessels above pressure in the extramural arteries, perturbations analyzed with a nonlinear optimi- whereas in the superficial muscle, intravascular microves- zationtechniquehavebeenused to evaluate sel pressures are below extramural arterial pressures. model parameters during normal diastoles, with These pressure differences cause blood to move retrograde similar results. As in the prolonged diastoles, from deep microvessels to superficial intramyocardial ar- parameters differed significantly from control teries, thereby competing with systolic inflow from the when the circulation was dilated with adenosine. extramural arteries. Experimental observations support Resistance was lower and capacitance higher in this hypothesis. the vasodilated bed. Back pressure was lowel: in thedilated bed,although stillsignificantly above right atrial pressure. We conclude that pressure-inflow relations do not vary significantly during diastole and that results obtained in prolonged diastoles can be anplied to dias- 2t toles in normal cycles.

180 1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas Biomedical Engineering Scientific Papers

BE3C.3 BE3-C.4 Quasi linear input behavior of non-linear model The Presence and Significance of Venous Level of intramyocardial perfusion,J.A.E. Spaan: Coronary Collaterals, J.M. Downeyeand M.V. Cohen, University of Amsterdam, Department of Medical University South Alabama, Mobile AL 36688 Physics, The Netherlands. To measure venous level collateral connections, the great cardiac vein (OCV) and the Compression ofintramyocardialvessels is anterior descending coronary artery (LAD) were apparant from the increased venous and decreased cannulated in an open chest dog. Dye studies arterialcoronary flows in systole. Moreover, indicated that under baseline conditions the GCV time averagedintramyocardial blood volume drained exclusively from the LAD (10/11 dogs) and dependson themagnitudeof left ventricular LAD and OCV flows were equal. Perturbations in cavity pressure. Hence, it is plausible that flow the LAD flow were not reflected in the OCV. how- patterns in all types of vessels but dependent on ever. LAD occlusion caused only a 56% decrease in the depth inthe myocardium are influenced by GCV flow, while peak GCV flow during reactive hyp- changes of volumeeither directly (capacitive eremia consistently underestimated peak LAD flow flow) orindirectly becauseofa changein by nearly 40%. Simultaneous LAD and OCV occlu- 'esistance. sions resulted in near equilibration of arterial Therelativecontribution of the different (13.6 + 1.3 mmHg) and venous (12.4 + 1.2mmHg) effects were assessed in a multi layer model of pressures.The data closely fit a model which the ventricular wall in which arterial, capillary assumed the presence of a low resistance venous and venous compartments are put in seriesfor anastcmosis. The resistance of the segment distal each layer. Based on a pressure volume relation to the venous anastcmosis was 10.0 + 1.5% of the foreachcomparaient reflecting stiffening of LAD resistance, while the resistance of the venous vessels athigher pressures and a relation collateral averaged 9.9 + 3.0% of the LAD resis- between resistanceand the inverseof volume tance. Following LAD occlusion, 133Xe washout squared the volume displacements and resistance decreased from 0.86 + 0.16 to 0.13 + 0.02 variation were predicted. ml/min/g. Retroperfusion of the OCV with arterial Theinput effects of the non-linear model blood during occlusion significantly enhanced the resemble the input of a constant RC network with washout by 85% to 0.23 0.03 ml/min/g (p<0.005)

small time constants ( < 0.2 s ) while the other showing that venous level collateral flow is system variablesvaryslowly (characteristic somewhat nutritional. (supported by NIH grant HL times > 1s). 20648)

BE 3D. 1 BE3-D.2 Myocardial Regional Flows and Metabolism: PET Studies, CARDIAC CONTRACTION AND CORONARY FLOW J. B. Bassingthwaighte: M. M. Graham, and R. B. King, Katsuhiko Tsujiokeand Fumihiko Kajiya University of Washington, Seattle, WA 98195. Department of Medical Engineering, Kawasaki Because first order compartmental models are applicable to Medical School, Kurashiki, JAPAN the parameterization of myocardial blood-tissue exchange processes from indicator dilution or PET or SPECT data We recently reported that the intramyocardial only under severely limiting circumstances, we are using capacitance vessels are composed of two functional usual capacitance and unstressed These account for axial gra- components: axially distributed models. volume. Thus, the phasic coronary flow may be dients within the capillary-tissue units and for the effects of caused by the mechanism that the blood entered axial diffusion. Het^rogeneity of regional flows is accounted into the intramyocardial vessels during diastole foron the basis of fractaldiAribution.The rates of would be squeezed out into the coronary vein by transmembrane transport (across 2 endothelial surfaces and cardiac contraction during systole. Tu clarify the myocyte surface) and intracellular reactions forming this squeezing out dynamics, we analyzed the metabolites can be estimated only from the kinetics since response of great cardiac vein flow(GFVF) after external detection does not distinguish either localization or sudden cessation of coronary inflow by laser molecular form of the radioactive tracer. Glucose models of Doppler velocimeter in open -chest anesthetized this form give quite different estimates of substrate metabol- dogs. GCVF decreased mono-exponentially after the ismthan do compartmental models.An argument for cessation of coronary inflow. Thus, the time preferring the estimates from distributed models is that they constant of GCVF decay was calculated and the blood within intramyocardial vessels during steady fit high resolution time-course data well, whereas compart- state(Vo) was obtained as time-integral of GCVF the mental models donot.Other arguments regarding after cessation of inflow. When GCVF during preference are inferred from the anatomy and from observa- steady state(Fo) was changed by altering coronary tions of spatial gradients in concentrations of other solutes. tonus, perfusion pressure, heart rate and Conflictsin estimates of metabolic rates from different contractility, the unique linear Fo-Vo relation experiments are probably resolvable by making use of the determined Fo, but there was no significant most realistic models. (Supported by NIH grants CA42045 relation between Fo and T. We conclude that the and HL38736.) coronary vein flow is caused by squeezing out of blood within intramyocardial vessels and is regulated mainly by Vo, not by T.

198k World Congress on Medical Physics and BiomedicalhglieerIngSan Antonio, Texas 181 Biomedical Engineering Scientific Papers

HE3-D.3 BE3-E.1 Assessment of Myocardial Perfusion and Magnetic Resonance Imaging Assessment of Myocardial Blood Volume Using Ultrafast-CT, Regional Ventricular Function, James T. J. A. Rumberger*, M. L. Marcus, Mayo Clinic, Willerson,* University of Texas Southwestern Rochester, Minnesota, and University of Iowa, Medical Center at Dallas, Texas Iowa City, Iowa. A quantitativenuclearmagneticresonanceimaging method has been developed for evaluating regional left ventricular function. Short-axis images of the heart were Ultrafast computed tomography (CT)uses a unique obtained atend-diastoleandat100 msec intervals electron beam technology to allow for thereafter. Regionaldiastolicleftventricularwall videodensitometric analysis of indicator clearance thickness and maximal percent systolic .vall thickening data from the great vessels and the myocardium. were measured at the level of the papillary muscles in Following intravascular injection of iodinated each of 6 segments.In 6 normal dogs, the mean end- contrast, 50 msec, high resolution images are diastolic wall thickness was 9+1.6 mm and the mean acquired at up to 2/cardiac cycle from multiple maximal percent thickening was 61+11%. In 8 dogs with 4 cardiac levels. Selection of regions over the left day old infarcts, maximal percent thickening was 5+8% ventricle, aorta, and myocardium allow (p<0.001) in the infarcted segments.In 10 normal human characterization of contrast clearance from each volunteers, the mean end diastolic wall thickness was vascular site using a gamma variatecurve. 10.1+1 mm and the mean maximal percent systolic wall Application of classic theory to these data have thickening was 60+18%. Using the mean value obtained in shown promise for intravenous injection techniques normal volunteers, reduced maximal percent systolic wall with regard to quantitation of regional flowper thickening was defined as a value 2 SD or more below the unit volume and the ratio of myocardial blood normal values. Seven patients with regional wall motion volume to central blood volume when comparedto abnormalities were independently assessed by nuclear simultaneous assessment using radiolabeled magnetic resonance imaging and biplane ventriculography. microspheres (r * .85, SEE = 50 cc/100 g/min, Compared to regions identified as akinetic or dyskinetin r ° .84, SEE = 5 percent, respectively). This by biplane ventriculography, nuclear magnetic resonance approach may have broad application to the imaging demonstrated reduced maximal percent systolic evaluation of patients with a variety of cardiac wall thickeningIIIthe same segments with a sensitivity of diseases. 94% and specificity of 80%.Abnormalities of regional systolic wall thickening are accurately identified using thisquantitativenuclear magnetic resonance imaging technique.

BE3-E.2 BE3-E.3 Cardiac Structure and Function Quantitated Measurementsof LeftVentricular Shape and

with Multislice Computed Tomography (CT), Function, S. Sideman , H. Azhari, U. Dinner N. Chung and E. L. Ritman*, Mayo Medical and R. Beyer, Biomedical Engineering, School, Rochester, MN. Technion-IIT, Haifa, Israel.

The fast, high repetition rate, volume A new method of detecting iegional left ventricular scanning computed tomography Dynamic (LV) dysfunction by 3-D reconstruction of the LV is Spatial Reconstructor was usedto quan- presented. In general, the 3-D shape of the heart titate cardiac structure and function in at end diastole and end systole is transformed to a intactexperimentalanimals. From the helical coordinate system.By that the 3-D infor- scans performed during a right atrial mation is transformed into a continuous one- injection of contrast agent, Iwasaki et al dimensional signal. Input datawere obtained by demonstrated 95% accuracy of estimates of various measurement techniques including echocar- myocardial muscle mass (Am J Physiol 246: diography, Cine CT and MRI. Three levels of analy- H138-42, 1984) and Hoffman et al obtained sis were performed: a) Shape analysis and a shape the same accuracy for volumes of all car- distortion analysis assisted by the Karhunen-Loeve diac chambers (Radiol 155:739-44,1985). transform, b) analysis of endocardial motion from Myocardial perfusion estimated from opaci- end diastole to end systole as a measure of region- fication of the myocardium during an aor- al dysfunction, and c) analysis of wall thickness tic. root injection of contrast agent was and thickening as a measure of LV regional dysfunc shown by Wang et al (Clinical Res 34:947A, Lion. The combined technique of assessing patholo- 1986) to match values estimated by micro- gical LV shape and fction represents a powerful spheres over a wide range of values. tool which, together with the fast developments of Chung et al (Physiol 30(4):54, 1987) a newgeneration of 3-Dimagingmethods, will demonstrated a good correlation between greatly enhance non invasive cardiac diagnostics. myocardial oxygen consumption and ventricular pressure volume loop area. We conclude that quantitation of multiple aspects of cardiac structure and function can beobtained simultaneously from a single volumetric CT scan during a C 4 contrast medium injection.

182 1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas Biomedical Engineering Scientific Papers

BE3-F.1 BE3 -F. 2 Assessment of Left Ventricular Systolic Dynamic Transverse Stiffness Can Estimate and Diastolic Mechanics Using Ultrafast-CT, Myocardial Wall Stress during Single J. A. Rumberger*, M. L. Marcus, Mayo Clinic, Contractions, H. Halperin*, J. Humphrey, P. Chew, Rochester, Minnesota, and University of K. Sagawa, J. Tsitlik,F. Yin, Johns Hopkins Iowa, Iowa City, Iowa. Hospital, Baltimore, MD 21205

We have shown that transverse stiffness (TS) Ultrafast computed tomography (CT) uses a unique (the ratio of indentation stress to strain during electron beam technology to allow for evaluation transverse indentations) is ?elated to in-plane of left ventricular (LV) function at multiple wall stress at the peak of contraction. Those levels. Cine framing rates of 17/second allow studies required 20-40 beats for one stop action images of the heart using intravenous determination of TS. We developed a new servo contrast injection .echniques. Quantitative system that can indent muscle cyclically at analysis of LV mass (SEE * 5 g) and LV stroke frequencies up to 100 Hz, while maintaining the volume (SEE 1.5 cc) are possible. contact stress constant throughout a contraction. Time-dependent ventricular volume information can In-plane wall stress was measured directly in be analyzed per tomographic level as well as for five isolated ventricular septa while TS was global analysis; curve fitting techniques allow determined during each contraction. For values cf for characterization of maximum emptying and in-plane stress from 0.05 to 2.0 g/mm2, TS varied filling rates. These CT techniques have broad from 2.1 to 25.7 g/mm2 and was linearly relotnd applications towards the noninvasive diagnosis of to the in-plane wall stresses in each septum (p < coronary artery disease and quantitative 0.001, ANOVA). The slopes of the individual investigations regarding LV response to surgical relationships between TS and wall stress from and pharmacologic therapy. systole were similar to those from diastole (mean t SEM) (9.1 t 1.6 vs 10 t 3.1, p - NS), but the intercepts with the TS axis were greater than those from diastole (11.8 -E 1.6 vs 4.2 t 0.6 g/mm2, p < .016). Thus, dynamically determined TS is linearly related to in-plane stresses and can be used for quantitative estimation of in-plane stresses throughout a contraction cycle.

TBE3=Fa- CET-A.2 The Use of Diastolic Sounds to Characterize Management and Maintenance of Health Care Cardiac Performance: Theoretical Studies Equipment - WHO's Global Effort. A. Issakov, World Health Organization, Geneva, Switzerland. W. Welkowitz,* J. Wang, B. Tie, S. Lewkowitz, J. Semnlow, Department of Biomedical Most developing countries are facing tremendous Engineering, Rutgers University. wastage of resources and severe limitations of all aspects of health care delivery due to the ineffective It has been proposed that diastolic sounds can management and maintenance of health care equipment. be used to suggest potential problems of cardiac The reason for this is the lack of awareness, policy performance. In particular, such sounds can commitment, effective infrastructure of health care indicate the presence of partial blockage in the technical service, manpower development programmes coronary tree and perhaps quantitate the blockage. and a comprehensive approach at a country level. In order to carry out theoretical studies per- tcoultu yr international action undertaken by taining to this problem it is necessary to first various organizations and agencies were far from produce an acceptable model of the coronary artery anticipated due to lack of coordination and tree which can be used to predict flow throughout duplication of efforts in some areas with others the cardiac cycle for any blockage location.Once being neglected. the flow through the blockage is calculated, then Responding to the growing concern over the existing it is necessary to construct a model of the sound situation WHO has initiated a global programme generation mechanism. Finally, the propagation of directed towards strengthening of national this sound to the outside of the body must be analyzed. capabilities for the development of comprehensive A model of the arterial tree which includes area systems for effective management of health care variations duringsystole(including occlusion) and equipment. peripheral regulation has been analyzed and The programme aims to meet its objecting by coronary flows in all parts of the system have building up common coordinated action, by been calculated for normal and constricted cases. facilitating collaboration between all agencies These agree well with measurements. Finally, concerned and by mobilizing resources for using these flows, sounds generated by turbulence comprehensive, integral and coherent action as part at these constrictions have also been calculated. of strategies developed with the aim of achieving the goal of Health for All. 203 19&8 World Congress on Medical Physics and Bioinedical EngineeringSan Antonio, Texas 183 Biomedical Engineering Scientific Papers

CET-C.1 CET-C.2 A PROFILE OF CERTIFIED CLINICAL Clinical Engineering in the U.S.: ENGINEERS. L. Fenniskoh, St. Luke's The Administrator's Viewpoint, M.J. Medical Center, Milwaukee. WI.USA. Shaffer; D.Sc. and M.D. Shaffer, M.S., George Washington University In an attempt to obtain a profile of Certified Medical Center, Washington, D.C. U.S.A. Clinical Engineers (CCE's) and their thoughts on the profession. a questionnaire was et It to 346 In a survey of 250 hospitals with 89 re- CCE's worldwide. A remarkable 72%retr.m spondents, it was found that administra- revealed. in part, the following: tors running clinical engineering depart- 57% of the CCE's are employed by hospitals. ments received in their college courses an 15% by manufacturers. 12Z by academia, and average of 0.6 hours of training in clini- 13Z are consultants in private practice; cal engineering matters, while they con- 50% ofCCE'shave been with their current sidered 4.1 hours are needed. They had an employer for 6-9 years; on-job experience of 6.5 years, while re- 48%of thehospital-based CCE's previously commending 4 years as an appropriate mir_- worked in hospitals.while36% of non mu.d. As to reporting hierarchy, 42% re- hospital-based CCE's left hospitals for commended both Clinical Engineering and their present jobs; Plant Operations should report to the same the CCE's average age is 44 years; administrator, 22% recommended they should 37Z of the CCE's are alsoregistered report to separate administrators, and 31% Professional Engineers; recommended that Clinical Engineering mediansalary range for hospital-based CCE's should report as a division of Plant Oper- is $40-45.000/yr. whilethat of their non ations. Concerning their attitudes, on a hospital-basedcounterparts is $S0- SS,000 /yr scale of 0-5 with 5 highest, the admini- (onaverage. hospital CCE's are also 5 years strators rated their clinical engineering younger than non hospital CCE's). responsibilities with little power (2.0) Although the general conclusionssuggest that and chance to make their name (2.3), with the certification process in particular and the moderate prestige (3.65), and with high profession in general is in distress. CCE's acceptance of their authority (4.01) and believe theprofession hesmade significant pride of achievement (4.4). contributions tohealthcare and continues to remain a viable cLreer choice.

CET-C.3 BE4-A.1 Preliminary Results of an International Clinical Engineering: A Focused Biomedical Survey cf Clinical Engineering Departments. Engineering and Career Program M. Frize, Regional Clinical Engineering Service, 330 Archiblad Street, Moncton Barry Feinterg, Ph.D., P.E., C.C.E (New Brunswick) E1C 2Z3, Canada. Center For Engineering Analysis Northbrook, Illinois 60062 USA In a survey of clinical engineering departments in several countries, with 80 respondants, 14% A clinical engineer is an engineer whe indicate that they do not have an engineer on applies the science of engineering to the their staff; 49% of departments report to admi- delivery of health care. He or she is concerned nistration; 30% to plant/maintenance; 6.5% to the with the application ofengineering theory and medical director and 11% to others. However, analysis to diagnostic and therapeutic systems as only 62% of departments reporting to plant find well as to patientcare andlife suppnrt in this satisfactory, whereas 100% of those report- general. ing to administration indicate satisfaction.The Clinical engineering has emerged as the only proportion ofdegartmentsperforming 90% of more areas of biomedical engineering that has in-house repairs were:68% for medical equipment; crystallized intoa unified body ofknowledge 19% for medical imaging; 132 for the laboratory; similar tothe othersiassical disciplines of 44% for anesthetic equipment; 6% for non-medical engineering. Clinical engineering will be devices. The proportion of departments being discussed from the view of career op,..'"xunities, "always" consulted prior to equipment acquisition both hospital and industrial; educational is higher where there are engineers on staff and preparation required of the clinical engineer- also higher in University 1.,ersus non-University courses;laboratories and internships, and hospitals. Several other variables will be certification in clinical engineering. discussed and variances between countries will be The emerging areas of medical radiation shown. These preliminary results indicate the engineering and clinical engineers as part of an presence of great disparities between clinical artificial heart transplant team and ether engineering departments in the same country and advanced areas of health care delivery as new from one country to another. paths in the career development of the clinical engineerwillbediscussed.

184 1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas Biomedical Engineering Scientific Papers

BE4-A.2 BE4-A.3 electronic State Orders to Clinical Engineering in the GDR Biomedical Engineering in an engineeringcurriculum:a ten yearsexpe- D.Kraft: Zentralklinik, DOR - 5303 Bad Berke, GDR rience in Spain,R.Pallas-Areny:Division The ministry of health in the GDR have devoted de Instrumentacion y Bioingenieria,ETS Ing attention for many years the spezialized field of Telecom, Box30002, 08080Barcelona, Spain. biomedical engineering in medical institutions. profes- 1. On April 1st 1981 the decree on the postgraduate Biomedical Engineeringis not recognized as a orschool study biomedical engineering for scientific and sion inSpain,nor are thereuniversity technological university graduates, which work in qualificationsin thissubject But the medical equip- need medical institute ms, came into force. This study ment industry,hospitals and enndistributors knowledge of medical instru- is an integral part in systematically organized engineers with a certain development training - analogously to the continu- mentation. Spain there are noelecti- ation training to become a medical specialist. In engineering curricula in must be introduced (Requirements for attendance, setup and course, ve courses and any change in them Takingad- study material, attendance of other medical insti- under the umbrella of an existing course. foralreadyexisting tutions as guest, final examination, professional vantageofourresponsibility andTransducers coursesand their bye-identification). Instrumentation practicals we changed our program 10 years 2. On January 24st 1982, the decree on the struc- laboratory and 50% of it now consists of Medical Instrumen- ture and organization of biomedical engineering by ago tation.In it,biomedical topicsare considered not as medical institutions came into force. of introducing, for The basic unit of clinical engineering has the an end in themselves but as a mean example,low levelsignalamplification, interferen- following duties: Responsibilities for guaranteeing flow measurements, electrical safe- readiness for use, efficiency and safety of medico- ces, pressure and energy interactionswithliving technical equipment, assignments in the care for ty, imaging,and compulsory laboratory coursecon- patients as direct partner and under the physi- beings.A parallel cians responsibility; assignments within the plan centrates on analog signal processing. component of science and technology at the institution, assignments in the field of educational development, assignments in the field of planing and economy.

BE4-A.4 BE4-A.5 EDUCATION AND TRAINING FOR CLINICAL BIOMEDICAL ENGINEERING EDUCATION IN THIRD WORLD ENGINEERING IN THE U.K. NATIONS - T. Krishna Murthy* and M. Dhananjaya, V C Roberts* Dept. of Biomedical Engineering, India Institute Department of Medical Engineering, of Technology, Bangalore, India. King's College School of Medicine, LONDON Compared to developed nations, the disease pattern No formal course of instrudtionexists and socioeconomic conditions among the majority of within the UK which leads towards an people warrents a modified approach regarding undergraduate or postgraduate degree (or planning and utilisation of trained manpower, certification) in clinical engineering. even use of automation and semiautomation though The presently preferred route is via one inevitable and desirable may have to be restricted of several MSc courses inbioengineering to certain centres of excellence having the (including medicalelectronics and necessary back-up facilities. There can be no are then biomechanics). These degrees doubt that biomedical engineers, clinical engi- followed by suitableperiods ofon-job training in the clinical environment. neers, rehabilitation engineers and medical equipment technologists are needed for the major hosp- Postgraduatetrainingprogrammes are itals, medical colleges and other specialised available for tht,se in health care who medical research centres. Courses have to be seek professionei registration as a planned depending on the utilisation of such Chartered Engin,.>:r. However, their trained manpower. Besides regular courses, short persuit is dependent onthe sympathy of and long term refresher courses, workshops and the head of aach department ofMedical symposia form a major part of Biomedical Engi- Engineering ..)r Medical Physics. A neering education and training. So, a cautious number of part timepost-qualification need based approach with necessary modification courses are either plannedor being run anu adaptation in consultation with the state and these will need to adeuess aspects medical and health care authnritia.s may lead to of businessmanagement as well as optimum utilisation of such manpower. scientific upda ".ing. There is evidence that the UK'sHospitalEngineering Service is increasing its involvement in clinical engineering, particularly in the area of equipment management. 205

185 1988World Congress on MedicalPhysicsand Biomedical EngineeringSan Antonio, Texas 41=1.4=MiNZIIKIK

Biomedical Engineering Scientific Papers

BE4-A.6 BE4-A.7 SUPPLEMENTAL TRAINING FOR CLINICAL ENGINEERS Estimating the Number of Clinical AND TECHNICIANS: INTERNSHIPS, G.I. Johnston' Engineers in the United States, and R.L. Morris, Oregon Health Sciences Univ. M. Baretich., University of Colorado Health Sciences Center, Denver, Colorado, USA.

Estimates of the number of clinical engineers Engineers and technicians graduate from four-year practicing in the United States range from a few and two-year programs with little or no practical hundred to several thousand. At least two meth- experience. This is understandable and notneces- odological factors contribute to this wide var- sarily undesirable. An electrical engineer or iance: (1) Different definitions of the term electronics technician has an enormous varietyof clinical engineer. (2) Different statistical career choices upon graduation, all of whichcan- estimation techniques. not and should not be addressed in the academic curricula. They should first of all be competent The definition of a clinical engiaeer should be in the fundamentals of their engineeringor tech- as broad as possible yet exclude personnel not nical discipline. It should then bc, 3nd histor- holding apprwiate professional credentials ically has been, the responsibility of the hiring (education, licensure, certification, experi- organization to provide training specific to the ence, etc.). Statistical estimation techniques company's product or service. For bachelors level should be carefully chosen and correctly ap- biomedical and clinical engineers and for many plied. This paper addresses these issues and biomedical equipment technicians this process has offers recommendations for future estimates. been subverted in favor of providing life science education at considerable cost to the engineering The author estimates that there are 500 to 1000 curricula. That subversion has been aided and practicing clinical engineers in the United abetted by hiring organizations failing nrovide States. However, a definitive survey should be apprentice or internship training programs ,s- undertaken to determine the number more precise- sary for the completion of the education process. ly. Without such basic information regarding Several solutions to this problem will be pre- the profession, preparation for the 21st century sented and discussed. will be difficult.

BE4-A.8 BE4-B.1. Expanding Horizons of the Clinical Standards: An Overview, A. Wald', Department Engineer:The Artificial Heart, of Anesthesiology, ColumbiaPresbyterian MedicalCenter, Michael S. Bernstein: M.S.E.,P.E.,C.C.E. New York, NY 10032 Methodist Hospital of Indiana, Indianapolis, IN 46202 In the most general terms, thereare four types of standards. The Clinical Engineer's Horizon of responsibility First, there are standards that relateto has been expanded to include the Total Artificial measurements. These standards permit the conductof scionce and technology and also foster tradeand commerce. Heart. The patient is totallydependent on A second type of standard would an more appropriately be electromechanical devicetherefore changing referred to as a uniformity document. clinical engineering'sresponsibility from in- Thus, there are direct stanaard screw threads, kitchen counterheights, and to direct patient care. Before, during, transistor characteristics. and after the implant surgery the clinical engineer has crucial responsibilities. The Clinical Then there is a more intangeabletype of standard. Engineeris responsibleforthe constant state A of readiness of standard that does not necessarily referto any physical the support equipment and the product, but rather to some type of nursingstaff. human behavior or Immediatelybeforeand during activity.This type of standard is often only written implant,the clinical engineer must set up the down in the broadest terms.Itis often based on some consensus equipment systems for their specific use. Because of what society as a whole expects, the artificial heart does not provide ECG signals, either for the ideal condition or for what can be expectedpractically. the operating room/intensive care monitoring Finally, we have what would best systems cannot be used in the conventional way. be called, product standards, that is standards that relatedirectly toa Support mustbe provided on a 24 hour aday particular product.Suchstandards can includecriteria basis during the implant period. such as materials, construction, operation, During performance, the explant/transplant, the clinical labeling, safety, and efficacy. engineer is responsible for the supervision ofthe artificial heart console and shutdown Standards must be addressed in theirown right, in terms of assisting thesurgeon in propertransfer from legal, economic, and technical considerations.Some aspects artificial heart to heart-lung bypassto trans- of standards on an international basis will plant. be presented. 206

186 1988 World Congress on Medical Physics and BiomedicalEngineeringSan Antonio, Texas Biomedical Engineering Scientific Papers

BE4-8.2 BE4-B.3 Multiple Roles of International Standardization The Impact of International Standard in Medical Engineering, S. E. Lindell ,University Development on Medical Devices of Lund S-2041 Mahnii, Sweden; and R. J. Mort Levin' Plaszczynski* SEE/CFTBM, F-75724 Paris, France Easy global communication is tending to The first author is chairman of Technical Commit- blur distinctions between U.S. and other tee TC62 "Electrical Equipment in Medical Practice" standards. What are the forces shaping of the International Electrotechnical Commission this phenomenon? How do these forces (IEC) and the second one has been involved for influence medical devices? the last 18 years in the creation of international standards within numerous IEC working groups. Marketing across international Their experience enables them to emphasize the boundaries requires that manufacturers role of standardization not only for the optimiza- design to standards acceptable to the tion of the characteristics and cost of equipment importing user. Economic forces but also for promoting a better mutual understan- encourage producers to participate in ding between the people concerned. Since 1906 the worldwide standard development in order IEC continues to develop an international collabo- to learn about events taking place and ration between manufactures and users of electri- to share in the decision process. cal equipment. Experts from 42 countries participate through their National Committees in the Safety of medical devices is a concern activities of about 200 Technical Committees& the of users, producers and regulators. Sub-Committees of the IEC. Its Central Office is Since users and regulators are also situated in Geneva, Switzerland. The TC62 was impacted by finances, economics and establised in 1966 and now comprises 4 Sub-Com- safety tend to bring all parties mittees with 30 working Groups. World-wide volun- together for the benefit of the patient. tary standardization facilitates a closer cooperation The impact of easy communication on between physicians and engineers involved in the international standard activity and the establishment of safety rules concerning patient, resulting influence on medical devices operator and surroundings and futhermore contri- will be discussed in greater detail. butes to the transfer of the know-how and technology from developed to developing countries.

BE4-B.4 BE4-B.5 Safety regulations for medical equipment in U.S. Electrical Safety StandarL,. Focus Germany on Leakage Current and OxygenEnrichment U. Boenick*, P.M.Schroeder, A.Miksicek A. LipschultePE CCE, Waterbury Hospital, Waterbury, Connecticut 06721 On the 1st of January 1986 the German "Medizingerateverordnung" (regulation upon There are two major U.S. Standards dealir. with safety of medical appliances) has obtained legal hospital electrical safety: AAMI Safe Current validity. This regulation devides technical Limits Standard; and NFPA 99, Health Care Facili- equipment used in the medical environment into ties. The chassis leakage current value in these four groups according to their function and standards of 100 microamperes is substantially purpose. lower than the IEC value of 500 microamperes. The To be allowed to sell equipment subjected to reason for this appears to be historical as well group 1 (life preserving device) or group 2 as a different assignment of risk values. Some (implantable energized parts) a type examination of the difference also is a result of the lower and an official approval by the Industrial U.S. line voltage. The rationale for the AAMI Inspection Board ist necessary. value is well documented. The values chosen by For devices conforming to group 3 and 4 (energy- the other standards is not well documented. driven and non-energized equipment) a voluntary One aspect of the NFPA 99 standard that I am very type examination can be conducted, permitting familiar with is the use of electricity in oxygen the manufacturer to hold the GS-label (GeprOfte enriched atmospheres. I will use this example to Sicherheit = proved safety). illustrate the process by which this standard was The responsibility of manufacturers, significantly modified to reflect the data I had distributors and customers, maintenance of the accumulated. The original philosophy in this mat- equipment and other details are regulated by ter was that all electrical equipment should be law. kept away from any portion of the oxygen delivery system. The new philosophy looks at the place where the oxygen enriched mixture is actually expelled to the atmosphere, and at the temperature aid /or energy. that the electrical device is capable of generating. 207

1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas 187 Biomedical Engineering Scientific Papers

BE4-B.6 BE4-C.1 PROPOSAL TO IMPROVE THE ELECTRICAL SAFETY STIMULATING EFFECT OF LASER AND NONCOHERENT RED SITUATION IN PUBLIC HOSPITALS (TUCUMAN) LIGHT ON REGENERATION OF RAT SCIATIC NERVES - Tan CJ Felice, CP Breitman, OJ Barrionuevo, Runchuit, Qiu Yiguang, Tan Yankong, and Ma Guo- OR Batallth, HD Gramajo, VI Rotgerand guang, Sun Yat-Sen University of Medical Sciences, MI Vercellone. Bioingenieria, Universidad Guangzhou, P.R.C. Nacional de Tucum6n, cc 28 suc 2, 4000 TucumAn, Argentina. ME Valentinuz zi* This is an important basal study of laser medicine on the biological function with laser. The target Without considering more or less mightcome from the energy, the wavelength, the incidental short courses, in Argentina coherency and the polarization. So far as present there are: A) Bioengineering (undergrad time is concerned, there were the differential level), six-year well-structured program, opinions, as some authorsconsidered the biologi- Universidad Nacional Entre Rios.First cal effect relate to coherency, wavelength, ener- graduates by1991.B) Residence in gy, etc. So wa took to coherency and wavelength. Hospital Engineering (grad level), at The following results: several hospitals, Buenos Aires City, three years.C)A few universities 1. The 633nm red light and laserof other wave- (Tucum6n, Buenos Aires, Technological ) lengthscould stimulate the regenerationof of fer courses at the undergrad level. nerves. However, although a good introduction, 2. No significant difference was found between the they are not enough for the full training effectsof coherent and noncoherentlasers on of Clinical Engineers. We propose:1) sciatic nerve regeneration. Talks about the subject. 2) Courses for 3.The differenceof wavelength had no influence hospital technicians.3) Courses for on the stimulating effectof laser upon sciatic hospitaldirectors,medicaland nerve regeneration. p:ramedica 1 personnel. 4) To prepare roster of professionals with a of training in CE to be presentedsome kind to hospital authorities for possible hiring. 5) Field visits and detailed analysis of the situation at all public hospitals to obtain a complete overview.

BE4-C.2 BE4-C .3 Electrophysiological Studies of Laser and Suture Biomedical Engineering Approach, Theory Anastomoses in Rat Sciatic Nerves, and Biomechanic Basis of AIDS Laser The- T.A. Benke*, J.W. Clark, and P.J. Wisoff, rapy and Curative Computed Tomography Department of Electrical & Computer Engineering, for Combined Diagnostic and Treatment. Rice University, Houston, TX 77251-1892; V. Vasilionkaitis: Lithuanian Cancer Re- S.A. Shehab, Texas Heart Institute, Houston, TX 77225; search Institute, Vilnius, Lithuania, USSR S. Schneider and 3. Laurent, Texas Children's Hospital, Houston, TX 77030 According to mechanism of human immunodefiency virus (HIV) persistence the system of anti-HIVla- ser traps together with electrochemical devices Conventional suture repair of peripheral nerves results in agains AIDS is proposed. Laser traps from I to 10 fibrotic reaction which is detrimental to nerve regeneration. classes depending selectivity installed with blood As an alternative procedure, a laser can be used to reanasto- or lymph separators (e.g.types of Aminco, IBM, Ha- mose severed peripheral nerves. This study presents electro- emonetic, Penwell, Ito - Flow - Through) together physical measurements of the right sciatic nerve of Wistar rats with original chamber with mobile walls in turn for which has been surgically cut and reanastomosed either by external exposure at curative screen ( contact means of sutures or a laser weld. The left sciatic nerve is left therapy, sterilization, photoradiation, immunos ti- untouched and serves as a control for that animal. Each nerve mutation, killing, elimination) and return of blood was placed in a lucite bath containing a linear array of elec- elements to natural flow. Also a curative compu- trodes and filled with Kreb's solution. The nerve was then ted (CCT) tomography (x-CT, SPECT/PET, ultra- stimulated with a pulse of .01 msec duration, and the response sound, MRI) was substantiated with collection of measured. Computer simulations of short lengths of nerve are CCT devices for immediately treatment after preci- compared with the measurements to account for the progressive diminishment in amplitude of the action potential with se diagnostic: mobile CT-microscalpel with cor- increasing distance from the stimulation site and to understand recting tracers, CT-selective drug deliverysys- the nature of the injury potential at the distal end of the nerve tem in scanning layers, CCT devices for remote fiber. ligation, clip or cutting, canals performing, angioplasty, dilatation, cytotherapy. Number of CCT programmes are unlimited at all organs or tissue for different CT or magnetic resonance imaging therapy, surgery or microsurgery. 208 188 1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas BiomedicalEngineering Scientific Papers

BE4-C.4 BE4-C.5 Laser Coronary Angioplasty Using CO Laser under Laser Induced Autofluorescence Analysis Visualization by Coronary Angioscope, of Arteriosclerotic Arteries, T. Arai, M. Nakagawa, M. Kikuchi ,K. Mizuno, G. Pettit*, S. Kubodera, F.K. Tittel, K. Arakawa, T. Shibuya, K. Satomura, M. Sartori, P. Henry, and R. Sauerbrey, A. Miyamoto, Y. Okamoto, K. Horiuchi, Department of Electrical and Computer Utsumil, K. Isojima, A. Kurita, H. Nakamura, A. Engineering, Rice University, Houston, TX and K. Takeuchi2, National Defense Medical College, Saitama, Japan. 1Mitsubishi Cable, 77251-1892 U.S.A. Indus. Ltd., Hyogo, Japan. 2Fukuda Denshi Co. Ltd., Tokyo, Japan. Laser angioplasty has the potential to be a treatment for A fundamental investigation of laser coronary coronary arterydisease.Currently, one of the main angioplasty using CO laser under visualization of problems with this approach is uncontrolled damage to the intra-lumina lesion by coronary angioscope is vessel wall, including perforation.Therefore, real time described in this paper. In vitro experiment of diagnostic techniques, such as tissue autofluorescence CO laser irradiation to human atheroma indicates analysis which could indicate when laser application strong ablation effect with tissue selectivity should be terminated, are highly desirable. without any thermal damage to vascular wall. CO laser light, 5um of its wavelength, can be deli- Argon ion laser radiation at 457.9 nm has been shown to vered by flexible infrared glass fibers. The elicit characteristic spectra which allow discrimination authors have been successfully reported flexible between normal arterial wall and arteriosclerotic regions. energy delivery of CO laser by As-S chalcogenide In addition,thisspectralanalysishas been used to glass fiber. According to a estimation using determine when a diseased region as been penetrated and experimental results of the energy delivery, 3.3W normal tissue reached by excimer laser ablation.Since CO laser power may be avairable in human left ideally a single laser system would be used for both coronrty artery.By every3.5sirradiation of 351 nm excimer laser radiation this power can vaporize 7mg atheroma. To analysis and ablation, targetting and positioning of the laser irra- delivered via optical fiber has been tested as a means of diation, we develope thin coronary angioscope eliciting autofluorescence signatures. catheter of which outer diameter less than 5F(1.66mm). A top small round balloon can occlude blood flow to reduce blood flushing fluid.

BE4-C.6 BE4-C.7 Laser Surgery and Medicine in Japan. ADVEN TITIAL TEMPERATURES AND EXTENT Atsumi: University of Tokyo, 7-3-1 Bongo, OF DAMAGE DURING IN VITRO APPLICATION Bunkyo-ku, Tokyo, 113, JAPAN OF A LASER THERMAL PROBE TO CANINE ARTERIES The research and development on laser surgery J.H. TORRES7A.J. WELCH and J. GIEDONI and medicine was started by usiag ruby laser to THE UNIVERSITY OF TEXAS AT AUSTIN study the irradiation effects on tissues of mice THE UNIVERSITY OF TEXAS HEALTH SCIENCE in the University of Tokyo in 1965. CENTER AT SAN ANTONIO The first clinical trial of ruby laser therapy A laser thermal probe with a 2mm dimeter was heated by an for a skin cancer patient was experienced in 1967. argon laser at powers of 1 to 4 watts and applied to femoral, In Japan, surgical laser units such as CO2, renal and carotid canine arteries in vitro in 5 to 10 second Nd-YAG, argon were developed and widely used in pulses. The probe was placed inside the arteries and its clinical cases, argon dye laser was used for PDT, temperature was measured by a thermocouple welded to it (the He-Ne and diode lasers are applied for low power probe). A 3-5 um thermal camera was used to measure the therapy and CO, Nd-YAG and excimer lasers have temperature rise at the adventitial side of the vessels. Three been started to use for angioplasty. cases were considered: 1)no liquid in the vessel, 2)the artery At the present state, 230 Nd-YAG, 170 CO2, and filled with saline, and 3)the artery filled with blood. The 30 argon lasers have been installed in the extent of damaged was determined by histoloy and correlated Japanese hospitals and medical institutes. with the temperatures achieved by the probe and the The Japanese Society of Laser Surgery and temperatures measured at the back surface of the vessel. The Medicine was organized in 1979 and the 1,012 temperatures obtained were also correlated with the applied members including the 800 medical doctors and the laser energy. 150 engineers and the others were registered at the date of March 1988. This work was supported by the Free Electron Laser According to the wide spread of the laser units Biomedical/Material Science Program : ONR contract number for medical use, the education of clinical N00014-86-K-0875. engineers for laser surgery and medicine and their qualification become the important problems. The safety guideline was also made by the safety committee of the Japanese Society of Laser Surgery and Medicine. 209 1988 World Congress on Medical Physics and Biomedical EngineeringSan Anionic), Texas 189 Biomedical Engineering Scientific Papers

BE4-C.8 EE4-D.1 Laser Speckle Displacement Cardiography, ADVANCES IN MEASUREMENT OF RESPIRATORY G. Rmnachandran*, Megha Singh,Biomedical GAS EXCHANGE IN CRITICALLY ILL PATIENTS, Engineering Division, Indian Institute Edward J. Truemper; Larry S. Jefferson, of Technology, Madras - 600 036, India. Yadin B. David, David L. Welding, Baylor College of Medicine, Houston, Texas

A new non-contact and non-invasive technique to The use of oxygen consumption, carbon monitor cardiac functional disorders has been dioxide production, and respiratory developed. It is based on laser speckle inter- quotient determinations in critically ill fercmetry in which speckles, produced by the patients are assuming increasing interference of laser light, scattered due to importance in guiding therapeutic the roughness of a surface are utilized for mea- interventions for cardiopulmonary, suring the displacement of the surface. An nutritional, and metabolic support. electronic shutter has been developed for cont- Despite the wide variety of commercial rolling and synchronizing the laser exposure and custom designed instruments with the electrocardiography(ecg) of the subject. available, serious limitations exist Recording is performed on a holographic plate which restrict their application and and the developed plate is analyzed by point- interpretation when used on certain wise filtering technique. The displacements are patients. Potential factors that must be interpolated statistically and are plotted to considered include: use of endotracheal chtain a perspective mapping of the activity as tubes, high fraction of inspired oxygen, seen on the chest wall. By this method, the certain mechanical ventilatory modes cardiac displacement activity during various (continuous flow, high peak flow rates) phases of the cardiac cycle has been obtained and water vapor. In critically ill and compared for healthy male subjects. Its children, the fast respiratory rates and application to the patients suffering fran cardiac small lung volumes often further limit functional abnormalities provides details of the the technical capabilities of most malfunctioning region of the heart and shows ex- instruments. An outline of these various cellent to good agreement with that of 2D - eCho- factors limiting instrument application cardiograpy. is presented. Methods for reducing equipment error and improving instrument performance will be discussed.

BE4-D.2 BE4-D.3 AutanationinthePediatric EXercise Application of the equationof motionto laboratory,FrankM. Galioto, mechanical ventilation.P. Richardson* and T. M.D., Children's Hospital National Hansen, 1 Baylor Plaza, Houston, TX. Medical Center, Washington, D.C. Passive exhalationlung mechanics (complianceC, resistanceRand time constant TC) are used to Increasingcapplexity in the pediatric exercise characterizethe pulmonary pathophysiologyof physiology laboratory has lent itself to infants with lung disease in newborn intensive automation. Measurements that are becoming care units. We use the equationof notion(P = increasinglyiriportanttofullyassessthe V/C +141 to describelungvolume (V) and gas exercise performance in children include: flow (v)during passiveexhalationwhenpressure electrocardiogram,blood pressure, respiratory P for the respiratory system equals 0.TC is gas exchangein the form of oxygen oonsunption, calculated fran the slope of flow-volume curve, C carbon dioxide productionandminut.eventilation isdetermined eitheras quasi-static C or dynamic and cardiac output typically in the form of C, then R is calculatedfLauTC/C. Two different acetylene re-breathing. These parameters cannot techniquesareusedindetermininglung mechanics be measured simultaneously withoutcarputer- and the results are usually different. For controlled data acquisition and onlineanalysis. mapping severity or stage of disease, we clamp The computer serves to interfaceinput devices the infant'sendotractealnear end-inhalation and such as therespiratorymassspectrometerand allow pressures in the lung to reach equilibrium pneumatachograph, acquires data and performs then release the clamp for passive exhalation to online carputationof the various parameters. the atmosphere;therefore,the ventilator is The computer prepares the final report in a momentarilyrenewedfrom the system for quasi- timely fashion so that data is available for static measures. For determiningthe effects of analysis by the physician at the completion of ventilator settings on dynamic lung mechanics, theexercisetest without delay. measurements aremade during mechanical ventilation. An IBM-XT ccaputer is rsed to This paper will emphasize the studies undertaken sample flow and airway pressure then calculate C, with discussionof the complexmethodology R and TC. Tidal volume, respiratory rate and required and the usefu lness of automation in the airway pressures are also calculated. Computer pediatric exercise physiology laboratory. math :modelingis applied to flow-volume curves that are not linear and a distribution of time constants is calculated. 2.10 190 1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas Biomedical Engineering Scientific Papers

BE4-D.4 BE4-D.5 The Use of Magnetic Resonance Spectroscopy in Portable device for continuous monitoring Hypoxia. G. Gutierrez*, Pulmonary Division, Univ of oxygen uptake. K.Sato*, T.Tamura, nd Texas Health Science Center, Houston, TX. T.Togawa, *Univ. of California,San Diego, Magneticresonancespectroscopy (MRS) is a new La Jolla, CA 92093, U.S.A., Tokyo Med. & technique with great potential in the non-invasive Dent. Univ., Tokyo, Japan diagnosis and treatment of tissue hypoxia. MRS works by directinghighfrequency radio pulses An apparatus for continuous monitoring of oxygen into tissuesamplesplaced within a strong and homogeneous magneticfield. Theradio energy uptake without discomfort was developed. The absorbedby certain nuclei in tissue produce a subject wore a hood through which air was drawn by spectrum where different peaks correspond to an exhaust blower. The oxygen concentration in the various important biologically active molecules. hood was maintained constant by a servo-controlled The area under the peaksprovidea relative so ratewas essentially measure of the tissueconcentration of these blower, thatthe flow molecules. Two important, naturally occurring, proportional to the oxygen uptake. The flow rate of isotopes are 1-H (proton) and 31-phosphorus. the main stream was measured by a time-of-flight Proton MRS is widely used to create tomographic flowmeter located in the exhaust hose. Flow rate images, but it also provides information on the and oxygen concentration signals were digitized and tissue concentration of numerous compounds, such device, as free fatty acids and lactate. 31-phorphorus stored on a memory chip. The entire MRS can be used to follow changes in the energy composed of these several parts, was assembled in a state of the cell as it measures ATP, small package (wt=3kg) which the sub'ect carried on phosphocreatine, inorganic phosphate and his back. Upon removal of the memory chip after intracellular pH. The possibleroleof this each experiment, data were transferredintoa techniqueincharacterizingcellulareventsin hypoxic tissues, as well as the relative microcomputer where trends of oxygen uptake were advantages and disadvantages of proton and calculated. The measurements made withthis 31-phosphorus MRS, will be discussed. instrument are comparable to those made with a Douglas bag collection and gas analysis.

BE4-D.6 BE4-E.1 Clinical Use of Continuous CardiacOutput Ethical Codes in Biomedical and Clinical Measurement System with the Modified Thermo- Engineering - M International Comparison, dilutionCatheter, T. Kano, S. Sekii*, Jan Persson*, Dept. of Biomed. Eng. and Center M. Katayama**, K. Miyasaka**. Dept. ofME- for Medical Technology Assessment, Linkdping Service, Mitsui Memorial Hosp, Terumo Corp*, University, S-581 83 Linkbping, Sweden. National Children's Med. Res. Center**, 1, Kanda Izumi-cho, Chiyoda-ku, Tokyo lUl, Japan

Recentlythermodilution (TD) cardiac output (CO) In 1987 the International Federation for Nbdical and Biologi- measurements are widelyused inmany clinical cal Engineering (Ing3E) started a project aiming at an interna- cases. However, it is impossible to get CO values tional comparison of the ethical standards under which blared- continuously in such a conventional TDmethod ical aid clinical engineers wort. An inquiry on these matters requiring frequent intermittent cold injection was sent to the affiliated organizations of the IFVBE. which may affect patients. Issues considered were existence and applicability of ethical We've developed Continuous Cardiac Output Measure- codes or guidelines for professional engineers, clinical and ment (CCOM) system to solve the above problems. biomedical engineers and physicians. Convents on the need for In this system, theModified TD catheter (KATS/ ethical guidelines were alsodenanded. catheter) is used. This catheter has two thermis- In February 1':4 one third of the miter countries of the tors; one is for the conventional TD method, the Federation have responded to the inquiry. F?.4 have a code at other is for detecting heat loss dependent upon present, but several countries consider devising a code or feel blood flow velocity (BFV). So, once CO is measured the matter of high importance. Pros and cons will be discussed. by the conventional TD method during the continuous measurement of BFV, the relation between CO and BFV is decided (CO/BFV=S corresponding to the sectional area of pulmonary artery). After this, CO can be measured continuously by measuring BFV. The correlation coefficients between CCOM values and electromagnetic flow meter values were 0.966 (1=18) in hemodynamic model, and 0.944 (n=29) in animal experiment. CCOM system is now clinically tested and good results are being obtained. 2 1 1

1988 World Congress on Medical Physics and Biomes cal EngineeringSan Antonio, Texas 191 Biomedical Ettgineetlng Scientific Papers

BE4-E.2 BE4-E.3 Biomedical Innovations:The Role Ethics in Clinical Medicine of Ethics, M. N. Maxey*, Biomedical C. William Hall, Southwest Research Institute Engineering, The University of Texas San Antonio, Texas, U.S.A. at Austin, Austin, Texas. Over the past two decades, life support systems have been developed which allow maintain- Professionals responsible for innovations ing life during a difficult period and give the in biomedical technology are increasingly patient an opportunity to return to a productive vulnerable to accusations by critics who life of respectable quality. However, these same claim that ethical considerations are life support systems bestow upon the physician rarely a factor in determining the merits and family a difficult decision. Once it is of a particular innovation. As a consequence obvious that the patient will be unable to return critics have captured the "moral high ground," to a quality of life that contains joy or produc- and the images of Faust and Frankenstein tend tivity, should the life supports continue to be to dominate public perceptions of technology- used or discontinued. Probably, the most famous out-of-contrql. clinical case which focused public attention to Biomedical technologists must take a more this problem was that of Karen Ann Quinlan. active role in restoring public confidence Because of its wide news media coverage, the on at least two fronts: (1) shaping the Quinlan case requires no discussion here. Because public debate by demonstrations of concern of the United States' cultural background, euthan- for ethical principles in anticipation of asia or "mercy killing" is not practiced in this social implications of innovations; (2) country. However, the rising costs of hospital- preparing students to deal more effectively ization atd medical bills are creating debates with emerging ethical issues. ou this issue. Currently, the U.S. government only allows the support systeLs to be disconnected but demands that the individual be kept comfortable. In many cases, the patient is allowed to make a decision about the use of life support systems by initiating the "living will". Whether the decision is made by the patient or family, most persons agree that an individual should be allowed the opportunity to die with dignity.

BE4-E.4 BE4-E.5 Should Bioethics Training Be A Part of Protecting the Rights of Human Research Subjects, Clinical Engineering Curriculum?, S. Saha*, H. wigodsky*, University of Texas Health Science P. Saha, Louisiana State University Medical Center, San Antonio, Texas78284-7750 Center, Department of Orthopaedic Surgery, Shreveport, Louisiana 71130. There are ethical and practical reasons for protecting the rights of human research The fact that we need to provide an education that can prepare doctors and other medical personnel to subjects. Appropriate laws have been enacted face ethical problems is becoming more recognized. and regulations have been published. A Yet, a similar need remains unrecognized for the regulatory organization has been established in government, research institutions, and elsewhere bioengineer. The evergrowing dependence of health care on technology is creating the need to instill to assure protection of the rights of human research subjects particularly in government an ethical consciousness and a strong understanding of personal responsibility in those Clinical Eng- sponso ed research including research with drugs ineers who provide that technology in the hosp- and devices. Research investigators are well ital. As members of the health care team, Clinical advised to familiarize themselves with these Engineers must be prepared to face ethical issues processes to make the proper responses. arising from defective of inadequate equipment, scarcity of resources, conflict of interest,con- fidentiality, clinical research, "truth-telling," and care of the terminally ill.Thus it is important that clinical engineers, during their undergraduate and graduate training, should be exposed to the problems of biomedical ethics and to the principles which should guide them to deal with these difficult issues. With the rapid development of medical technology, new ethical questions are bound to evolve and thus clinical and biomedical engineers should also take part in continuing education on such topics so that they will be well prepared to face these issues so that we can assist .f in the progress of modern medicine. Examples of fa some of these issues will be discussed.

192 1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas Biomedical Engineering Scientific Papers

BE4-E.6 BE4-F.1 Professional Society Impact on the Status of PATIENT-CARE TECHNOLOGY EVALUATION, Clinical Engineering as a Profession, Gerald Y. David, Biomedical Engineering, R. Goodman: University of Texas System Cancer Texas Children's Hospital, Houston, Center, Houston, Texas. Texas 77030 U.S.A.

In the United States at present, there are three The delivery of Health-Care Services are professional societies working to represent the becoming more technology intense than interests of clinical engineers, and an inter- ever before. In addition the rate of national certification commission for clinical the introduction of new technologies engineering. into the health care environment is A reasonable question to ask is if these socie- accelerating. Thus, hospitals are ties work to the best interest of clinical engin- facing pressures to allocate their eering, and are the best use of the financial limited resources to the acquisitions resources available to promote the clinical and operation of these newer engineering profession? technologies. An answer to these questions can be found Biomedical Technology assessment is through a review of the sociology literature in being debated at various decision making the area of role theory, where professional levels. Governmerts, Health Care societies are identified as only one element in insurers, and professional societies are a natural developmental history of successful participating in the search for the professions. optimal methodology for assessing Clinical engineering would be better served Biomedical technology. At the hospital through an examination of the professionalize- level, however, limited support is tion process, and the development of a single available to guide the selection. society that best meets the exclusive interests The fitting of new devices into the of clinical engineering. unique practice of an individual hospital seems to be a suboptimal process. Hospitals using decisions based on evaluations by an interdisciplinary team of experts can improve operations and thus patient-care.

BE4-F.2 BE4-F.3 A Public Database on Medical Devices Clinical EnginePring in Primary Care, Jan Persson*, Chr. van Nimwegen* and J. Boter, Lars Borgquist ,nd Claes Debourg, Dept. of Bicmed. Eng. Medical Technology Unit TNO, P.O.Box and Center for Medical Technology Assessment, Linkoping 188, 2300 AD Leiden, The Netherlands. University, S-581 R Linkoping, Sweden.

During the last decennia the growth in use of medical devices in health care has been tremendous. Industry offers a wide variety of Primary care is in many countries expanding with regard to makes and types to the consumer. Health care extent and content. Technologies originally intended for the institutions need assistAnce in purchasing hospital are transferred to primary health care centers, often suitable devices and solvth3 problems concer- without analyses of effects. In this study an inventory of ning reliable and safe functioning of devices medical devices, their use and need for maintenance and train- in critical areas. Therefore, there is a ing, was made ccnprising the health care centers in Sweden. growing need for objective information on Measures of standard level with regard to medical devices and commercially available medical devices. chemical laboratory tests were found to differ significantly We developed a database, which contains: between the counties. The standard neasures increase with size - an inventory of diagnostic and therapeutic of the care center as measured by nurber of district phys- equipment offered tot the Dutch market-place; icians, and to extent of specialities available. The standard - information on publications like evaluation level of medical devices also increases with distance to the reports, hazard reports, market surveys, hospital. Policy decisions within the county councils and the review articles etc.; physicians background in hospital care seen to be important to - important specifications of a number of the choice of technologies. types of medical devices; - information on national and international A nurber of devices are pointed out where there is a need for standards concerning medica. devices. increased efforts by clinical engineers concerning training Since the beginning of 1988 the database is for correct use and maintenance. directly accessible for health care A mcdel for the diffusion of technologies to primary care is institutions. The database presented has a proposed. relational file structure while the data are presented in a videotex environment. 213

1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas 193 manr .=4".1

Biomedica: Engineering Scientific Papers

BE4-F,4 BE4-F.5 Modeling the Impact of Technology A Programme for the Assessment of Health-Care -n Neon.tal Intensive Care, R.T. Technologies in Japan, Makoto Kikuchi; Dept. Almeida, M. de ^arvalho, of Med. Eng., National Defense Med. Col., 3-2, Lopes, and R.B. Panerai*, COPPE/ Namiki, Tokorozawa, Saitama, 359, Japan. FIOCRUZ, R.) de Janeiro, Brazil. Since 1975, high-level medical technologies have The relationship between multiple technologies begun to be widely used in hospitals and clinics routinely used in neonatal intensive care units in Japan. At the same time, national health-care (NICU) and outcome 1,driables such as length of expenditure has increased from 240 billion yen or stay and weight gain was assessed byconsideriri, 2.7 per cent of the gross national product (GNP) the utilization of 53 technologi_s in 82 pain 1955, to 18 trillion yen or 5.1 per cent of the tients with birthweight in the range 100-1999g. expected GNP in 1987. Since the 'nirden of these The technologies studied correspond to 22 diag- expenditures has increasingly shifted from nostic/monitoring technologies, 29methods o: individual patients to society, the Government has :reatment, and 2 techniques of psychological attempted to stop or at least slow down the rapid support. escalation of health-care costs. In this paper, The total number of technologies utilized per the author introduces the unique situation and case (TU) can be explained by the 5 min Apgar tendency of high-level health-care technologies in score and the index of severity of illness with Japan compared with western countries, and also a coeficient of multiple determination (R2) of discusses the role which medical technology has 0.86. The length of stay (LS) is explained by played in this transition of the health-care TU and birthweight (BW) w;hR2=0.70, and the system. He also draws attention to the urgent need percent weight gain by LS, BW, and gestational to establish the actual programme for assessment age with R2=0.49. The three models are signif- of health-care technologies appropriate for the icant at the p<0.001 level. These results can national health-care system,and then discusses be applied to technology assessment and the the new methods used in developing the evaluation of relative performance of different fundamental criteria and the effective programme NICU. for the assessment of health-care technologies, which has been under study for the past year in Japan.

BE4 -F. 6 BE4-F.7 Assessment of Medical Devices Heatth Equipment Harming and J.F. Dyro, State University Pueunement: Expenienee linom a of New York, Stony crook, Duetoping Count4y, B. Wang*, ASEQ/ _mew York 11794. Secretaria da Sadde, Sao Paulo SP, Brasil 05403. Through assessment of medical devices, biomedical engineersin health care facilities circumvent the familiar prob- Within the State of sao Paulo a new health lems of frequent failures, high cost of system is being formed by uniting federal, operation, incompatibility with other state, city, and philanthropic health services. devices and user misuse and misapplica- This new system is composed of approximately tion. Softgoods and disposable products 544 hospitals (with a .otal of about 84.500 are assessed by University llospital' 5 beds), 1000 clinics, and 10 research centers. Standards & !;valuation (S&'..1) Committee This system serves, together with private which includes members from most hospi- institutions, a population of about 30 million. tal departments. natient care equipment For this system, a special advisory office, acquisitions are reviewed b? biomedical directly subordinated to the Health Secretary, engineering (Pn). In-house assessments was created to establish and implement a policy supplement information available from on health equipment, including all stages of sources such as governmental agencies, its life cycle, i.e., planning, procurement, independent research laboratories, i-istallation, utilization, maintenance, shared service organizations,journals, refitting, and eventually, obsolescence. professional organizations and the old Particular emph-sis was given to the boys' network (ot' :).nisposable developmentof a methodology and supporting pressure transducers, physiologieal mon- databases for planning and procurement, because itors, defibrillators, infusion pumps, of the large amount of capital investment continuous noninvasive blood pressure needed for new facilities and for improving the monitors, data management syotems and existing ones. apnea monitors have been evaluated by biomedicalengineering. 214

194 1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas Biomedical Engineering Scientific Papers

BE4-F.8 BE4-F.9 International Medical Device Hazard Reporting Comparative Evaluation of Medical Network, Elliot B. Sloane., MSEE, ECRI, 5200 Equipment, B. van Eijnsbergen', Butler Pike, Plymouth Meeting, PA19422 USA TNO-Corporate Planning Department, P.O. Box 297, 2501 BD The Hague. ECRI, a nonprofit medical technology assessment agency since 1965, has been designated by the Within the European Community a project on World Health Organization (WHO) as the Collaborat- Comparative Evaluation of Medical Equipment ing Center .Ir Information Transfer on Medical is going on. The aim of the project is Devices. i fundamebtal task is extending and a) co-ordination on the activities in the maintaining the mulLlingual International Medical field of evaluation of medical equipment of Device romenclature (IMDN) system.Secondly, ECRI the Member Countries, b) to collect, store is collecting worldwide medical device problem re- and distribute publications on evaluations of ports and organizing them with the IMDN for rapid medical equipment and c) to draft protocols analysis. Finally, ECRI will disseminate reported for comparative evaluations of medical risks and hazards of commonly used medical devices equipment. to the WHO member nations. The global economy of the 1990's and beyond will cause significant In the known literature-data-bases reports changes in the distribution, use, and safety of of evaluations are mostly not present (grey medical technology. In over 20 years, ECRI's ex- literature), so that a special data-base perience is that virtually all medical products, no. to be set-up for this aim. In the whether diagnostic or therapeutic, have some level Netherlands such a data-base is already of risk. No country is able to test and certify functioning for some years. Since in all of the medical devices used within its different other countries similar data-bases borders. Also, it is unethical, uneconomic, and are in preparation, a network-system of the^^ unnecessary for each country to learn on its own, can provide (on line) information. building on its own tragic experiences. Growing international trade and interdependence makes the Protocols on evaluation of infusion-systems IMDN-based comprehensive ECRI/WHO International and ultra sound diagnostic equipment has been Medical Device Hazard Reporting Network essential drafted. The use of a same protocol in an for all countries and patients. evaluation has the advantage of possibility of comparison, of the results.

BE4-F.10 BE4-F.11 Technology Assessment: Small Volume A.H.C.T. - RFLEVANCE TO THIRD WORLD NATIONS Infusion Pumps, Barbara Ann Bukowskil T.G. Krishna Murthy*, Bangalore, India 10103 Prospect Hill, Houston, Texas The realistic assessment of Health Care Technology 77064 appropriate and relevant to these nations involves a scientific and planned need based approach. Ma- This research explored a methodology for writ, of people (about SO percent) in these nations technology assessment within the health which are highly populated are. So technology and industry. Phase I reports the development trained manpower utilisation implies setting up and testing of the Candidate Adoption-Use both centres of clinical excellence where the re- diffusion paradigm and the time series quired back-up facilities are ensured and also at analysis of external SVIP diffusion. Com- the majority of other centres where the essential pared to the classic diffusion model, CA-U diagnostic, therapeutic and btoanalytical instru- more accurately explained the empiric ments which are low to moderate cost and are simple diffusion trends. CA-U favors utilization and rugged. Clinical experience in the last decade as opposed to adoption behavior. indicates the application of integrated therapy SVIP chemotherapy is a technology applied, utilising both modern and locally prevalent systems monitored and maintained by nurses. Phase of medicine. Such a step ensures carrying medical II reports on the prospective survey of care to the doorsteps of remote rural areas. The 117 random patients who adopted and used GPAHCT of WHO located in Copenhagen in consultation SVIP. Their functional ability in self with clinical engineers in these nations can en- care, social role and interaction, lighten and educate the medical and health care employment and activities of daily living experts which can go a long way in attempting to was analyzed. There was sufficient achieve the really formidable challenging goal of evidence that the alternative inpatient HeALTH FOR ALL BY 2000 A.D. technology restricted activity whereas SVIP outpatient therapy permitted patients to function more independently and in a social lifestyle, thus adding quality to life. 215

1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas 195 Biomedical Engineering Scientific Papers

101.-S.1 BE4-G.2 CURRENT STATUS AND FUTURE CHALLENGES FOR PRESENT AND FUTURE OF CLINICAL ENGINEERING CLINICAL ENGINEERING IN THE U.K. DEPARTMENTS IN JAPANESE HOSPITALS, N. One, V C Roberts' Dept of Med. Eng. Service, Mitsui Memorial Department of Medical Engineering, Hospital, 1, Kanda-Izumi-cho, Chiyoda-ku, King's College School of Medicine, LONDON Tokyo, Japan 101 Clinical engineeringhas onlyrecently become "officially" recognised inthe The national certification law of clinical engi- United Kingdom following the neering (CE)technicians was recognized in the establishment, by the Biological Di ' in May '87. However, there are a few esta- Engineering Society, of a Scheme of blished in Certification. This applies to CE-departments Japanesehospitals. those Althoughthere areabout practising professionally as clinical 7000CE-technicians, engineerswithinthe UK'sNational most of themengagein hemodyalysistreat- Health Service, and fully accords with mut in small cli..ics. There are only about the IFMBE guidelines for registration. 20 all-round CE-departments inJapan. They engage in theengineeringservice similar to Many health care scientists workon those in U.S. and othercountries except for clinical engineering activities, but so repairs. At thesame timethey engagein the far only a small number have sought clinical service, which means operation of many certification, a situationwhich should kinds of medicalengineering equipment, such change as the certification scheme as extracoporeal machines, hemodyalysis machines, achieves credibility. The main respiratory assist devices and other life suppor- challenges for the future, dictated by a ting equipment. Intheverynear future, we diminishing national health care must establish the organization and re-education resource, will irmolvetheclinical system of CE-technicians, and also the relation- engineer taking a ill' role as a ships betweenin-houseCE-departmentsand the business manager andnot merely a manufactures. I would like to report on these scientist. The educational and future programs. training implications ofthis have yet to be properly addressed.

BE4-G.3 2E4 -G.4 Ceinicat Engineeting in &a it., B. CLINICAL ENGINEERING INITALY Wang*, Assoc. Bras. de Engenharia R. Herletti: Politer:deo CI Torino, Torino, ITALY e Manuteneao Hospitalar, c/oCEB/ D. Dravar, Rip. App. San., USL 1 triestina, Trieste, ITALY UNICAMP, Cx Postal 6040, Campinas SP, Brasil 13081. A five year research project in Clinical Engineering (CE) has been completed by the Italian National Research Council in 1988. Purpose of the project was to survey, coordinate andlink in a national For many years, medical equipment was purchased program theexisting local CE activities and to provide guidelines in Brazil withoutregard to itsmaintenance to the National Health Service (350642 beds) for estabilishingCE needs and costs. Preveutive maintenance was Departments and recognized CE positions at least in the major hospitals. almost inexistent and services were provided The results indicate that 16 clinical engineers and 70 BHET are actually operating in Italian solely by hospitals (most of them manufacturers' representatives at affiliated to the Trieste Local Health Unit), while350engineers high cost and with low efficacy. Often it was and 1750 DHETwould be needed in the country to meet the minimum requirement of 1 eng./(1000 hosp.beds) and 1 tech./(200 easier to buy a new equipment instead of hosp.beds). On thebasis of a survey on 12 hospitals the value of biomedical repairing the old one. Thus a critical equipment in Italian hospitalshasbeenestimatedas 18,000 situation was generated: about 20-40Z ofall S /hosp.bed while the annual maintenance cost is 1,120 $/hosp.bed existing equipment (i.e,, about US$ 2-3 billion (nationalaverage) with minimal values at 770Slhosp.bed in worth) is out of order. hospitals with CE activities. Among the research products ,described in over 2000 pages, In the last 5-10 years, a few hospitals are: a software package forequipment management (a possible national established their own clinical engineering standard), a set of manualsabout purchasecriteria ,operating teams to take care of technology planning, principles, standards and safety problems, maintenance requirements of OR ICU, radiological, clinical chemistry and selection, acquisition,maintenance and physical therapy equipment. Pilot evaluations of cost/effectivness of some operators' training. The good results obtained radiological and clinical chemistry services are also described. by them are spurring others to follow the oaue As experienced in other countries, the estabilishment of CE approach, and even thegovernment to modify Departmentsappeared to be the only way to control the inflation of it's health policy. modern Health technologywhile increasinghospital safety and effectivness. No training programs are offered by universities for Last year, a Brazilian Association of Hospital clinical engineers. Experimental tral ing of BHET hasstarted in 1.ngineering and Maintenance was creatnd to two cities with fundsprovided by the EECandby regional promote further advancement of this area. administrations. Hovever the lack of official positions andcareer prospects and the low hospital salaries make industry offers more attractive to thenewly graduates. Twonational programs are expected to be approved in the next few years for training clinical engineers and BHETs. This first coordinated national effort provides health authorities with very clear directions about the appropriateapproach to the issue of health care technology 216 management. 196 1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas Biomedical Engineering Scientific Papers

BE4-G.5 BE4-G.6 Biomedical Engineering Challenges for Southern C1inica1 Engineering in Hungary, M.Susanszky, National Institute Africa, MA Poluta*, GG Jaros, Department of Biomedical Engineering, Groote Schuur Hospital for Hospital and Medical Engineering and UCT Medical School, Cape Town, RSA. /ORKI/, P.O.B. 32, H-1525 Budapest 114 Hungary, N.Richter+, MEDICOR, P.O.B. 150, H-1389 Budapest 62, Hungary Southern Africa is a developing region facing both first and third world problems in health Being aware of the present situation of care. In the context of health caretechnology clinical engineering in Europe it seems a need-driven rather than atechnology-driven to be worth giving an account of the approach is desired. "High-tech" solutions are Hungarian achievements which could be a often prohibitively expensive from both capital model of an effective system in a small and maintenance cost considerations and are not country. In Hungary there are about always the best or most appropriate engineering 100.000 hospital beds and the health care solutions to meet given needs. It is also per- is entirely organised and financed by the tinent to ask whether a developing region can state. Within the health service there justify a local equipment industry. We propose are about 220 qualified engineers working that such an industry is both desirable and in jobs that are connected with clinical viable. In particular, biomedical engineers engineering. It is a local peculiarity together with clinical engineers, clinical that only a smaller number of them works technologists and other health care workers, for the maintenance, unlike in other have a key role in such an environment. Only European countries. At the same time they through their continued involvement will an have larger responsibility for preparing appropriate and cost-effective transfer of of decisions concerning the purchase of technology be accomplished. This presentation medical equipment and for maintenance will examine some identified needs as well as management. examples of solutions proposed and implemented by our department. The technical staff of the hospital, under leadership of clinical engineers, does only the less difficult maintenance work.

BE4-G.7 BE4-H.1 EVALUATION OF PROVINCIAL PUBLIC HOSPITALS The Relationship Between Clinical Engi- IN TUCUMAN FROM A CLINICALENGINEERING neering Activity and Financial Support - VIEWPOINT. CJ Felice, CP Breitman,OJ Report from an Investigation by Clinical Barrionuevo, OR Batallan, HD Gramajo,VI Engineers, Bengt Lindberg* and Bengt- Rotger and MI Vercellone. Bioingenieria, Olov Wieck, Clinical Eng. Dept., East Unive.rsidad Nacional de Tucumbn (UNT), cc Hospital, S-416 85 Gothenburg, Sweden. 28 sue 2, (4000) Tucuman, Argentina. Max E. Valentinuzzi* The increased use of medical technical equipment Objetive: To evaluate electrical in hospitals has been more rapid than the follow* safety in the five main public hospitals sap by administrative personnel involvedin the (about 1400 beds) of the Tucum&IProvince financial support to this activity. An investi- (1,000,000 people) of Argentina. Field gation was carried out to highlight the problem. visits were performed and a detailed The value of the medical technical equipment in cuestionnaire was given to one of the our hospitals was estimated and the costfor hospitals (321 beds). For the latter:(a) service and mainteronce to keep this eouipment No intramural maintenance is perfomedto work safely was calculated. We compared this medical equipments. (b) Electricalsafety cost with the grant that is spent on suchactivi- standards are not applied. (c) There is ties by the hospitals. no adequate consultingmechanisms for the A comparison was made between the cost and time acquisition of new equipments. (d) No to repair equipment from either an external com- training is offered to the personnel.For pany or by the biomedical engineeringdepartment all hospitals:(1) Even knowledge of of the hospital. electric safety standards isessentially The investigation showed a large imbalance bet- nonexistant. (2) There is no Clinical ween the use of biomedical equipment andthe Engineering Department. Maintenance financial support for its use. personnel takes over a few CE tasks. With the results we hope to increase the aware- Repairs are sent outside (3) Within ness of the hospital managementof the consequen- hospitals, the available maintenance ces there are in introducing newequipment into personnel does not have enough skill.At hospitals. a national level, thereis a hospital standard (IRAN 4220) which considersonly general aspects of medical equipments.

4,..74.1

1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas 197 Biomedical Engineering Scientific Papers

BE4-H.2 BE4-H.3 Profit from Organized Medical Engineering AProgram to Increase Equipment Center at Kitasato University Hospital Durability in Brazil, MagalltBes R.A*., Satoshi Watanabe,M.D.* Panerai R.B., Nobre F.F., Wiederhecker Medical Engineering Center & Departmentof N.G., Machado J.C., Vieira C.R.S., Anesthesia, Kitasato University Hospital, INAMPe and COPPE/UFRJ, Rio de Janeiro- Sagamihara, Kanagawa, Japan. BRAZIL.

Medical Engineering center at Kitasato The transfer of mediae' technology to developing University Hospital was established in countries has been characterized by the early 1981 to organize all medical equipments deterioration of costly equisment resuiring from excepting for those instruments which be- inadequate planning, installation, and mainte- long to clinical laboratory and X-ray de- nance. Attempting to change this pattern, and to partment. This center provides not only increase the medical benefits resulting from de- engineering service as maintenance andre- ployment of 48 modern imaging devices in several pair of equipments, but also extends to regions of Brazil, a special program has been promotion and education activity, and ex- set up to prevent the classical facto which mination and development of new devices. have been responsible for the reduced useful li- We have studied the amount of finan- fe of similar technologies. Before delivery, a cial resources which contributes to profit detailed contract was established with the man- of hospital in the past one year. Theac- ufacturer regarding shipment, insurance, guaran- cumulative cost of repairs was calculated tee, and maintenance Following the installation with labor, managing and time share. This of each device, a clinical engineering team in- amount was compared with those expenses spects several aspects of the physical which we ought to pay if they are repaired enviromental conditions, and operator by manufacturers. The total expensewas training. The protocol also includes basic elec- cut down to one eighth of the amount of trical measurements and tests of image quality. requisition by the manufacturer. All abnormalities detected were immediately reported for corr-crion. This program can be regarded as a conceptual model wh;ch can be useful for other countries facing the problems surrounding tethnologv transfer.

BE4-H.4 Clinical rnoineering Productivity BE4-H.5 Management: A Case Study, A KNOWLEDGE BASED MIXED ARCHITECTUREFOR A. Mahachek, The Johns Hopkins ELECTROMEDICAL EQUIPMENT MANAGEMENT Hospital, Baltimore, Maryland,21205 r.u. Branca, P. Cappa*, A. Mura,F. Vincili, Lab ArtLficial Intelligence,Rome, Italy Sim( 1913 the Chemical Engineering Serviceof the Children Described in chronological orderare Hospital "B:erbino Gesu'" of Bra, wtublished in1931, utilizes the steps taken to install a traditional data bye system the monagement of 2353 electro- productivity measurements and controls within the Cli-,:t1 rnedial equipmentr, (global value 21 MS). Fran theacoraired EngineeringDepartmentof a large opexience the necity of this traditional database teaching hospital. Motivationfor CerifilTred,and the growth of new problem thatcan not be solus this project involves both by team of these system appeared. preparation for going "for profit" In particular the inefldmss of Artificial Intelligence and theneedto Teal resolve an over- in the equipment management appeared. budget condition of labor cost. The In this paper a model project steps include current based on mixed architecture is prqxced. hospital management reports, This modeliscapable to identify climatically the siotifi=t database adequacy, supervisors' time parameters for life cycle strategy of each equipment. logs, computerized budget spread The architecture is based on: traditional ctnta base (inventaxy sheets, testing changes on a data, annuoasonce data, costs), algarithnic model(casts ond financial simulation model, developing labor task standards, and product ivitim prevision); are), finally, lerwledgeassociated departmental labor performance to an overt sysban. reporting. The project yields significant positive results for the department's performance versus budget and also leads t' growth in supervisory and management skills. 218

198 1988World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas Biomedical Engineering Scientific Papers

BE4-H.6 BE4-H.7 The equipment control system, MEDEQ, Funding strategies for clinical as a tool in the reinvestment process, engineering in the Maritimes, *M. Frize, L. Lofstedt1 C. Nerhed and 8. Eklund, A.L. Nickerson, Reg. Clin. Eng. Service, Medical eng. dpt, University Hospital, 330 Archibald St, Moncton NB/NSAHO, 5614 Uppsala, Sweden. Fenwick St, Halifax NS.

The University Hospital in Uppsala has medical Two methods of comparing workload and funding re- equipment for 80 million dollars value, quirements of clinical engineering services are calculated 1987. All equipment, and day to day considered. information about it is continousely registered A first method is the traditional managerial in our equipment control system MEDEQ. approach of assessing workloads, in total hours, and dividing by the number of hours representing a The hospital has need for 20 MUSD as capital full time equivalent (1100 hours).The total work- money 1988. But we have only 5.5 MUSD avail- load includes repair hours, scheduled periodic able, for both reinvestment and investment in maintenance and incoming inspections. The base of new technology. If all money is used for med- 1100 hours takes into consideration that 30% of the ical equipment it givs a reinvestment period total activities are not accountable and belong to of 15 years in our hospital. The lifetime of the "other "activities category. medical equipment normally ranges between 5 A second method is to assess the replacement value and 15 years. of the equipment supported. The budget is expected to be in the 5 to 7% range, including overhead. The reinvestment process must therefore be The budget value would be lower when appli_d to carefully done, and the need for adequate in- radiology, typically 3-5%. formation is obviouse. One important information The value of using both methods in parallel is to source is our equipment control system, MEDEQ, develop experience which could lead to more which we have been using for more than 10 years. effective resource management. One or both these We have found it to be a very useful tool in methods may enable a valid comparison of in-house, the reinvestment process. shared, and manufacturers services. Actual results and experience will be discussed.

BE4-I.2 Choosing The Optimum Network For Real-Time Information Distribution K. H. Okumura*, Hewlett-Packard Waltham, Massachusetts 02254

When choosing a network architecture and protocol, one must consider the nature of the information which is being transported by the network. Phys- iologic waveforms, parameters, and alarms require different considerations than journal articles and electronic mail. The equipment-hostile nature of the hospital environment relative to the office environment must be considered when designing devices for use in the hospital. And in the event of failure, it must be easy to replace instruments without disrupting function of the network.

The CareNet utilizes a highly reliable central controller with a 32-msec polling cycle in a star topology to provide real-time transmission of data fail-safe and reliable operation, environmental integrity, and flexible configuration. Devices may be interfaced to CareNet to provide gateways to other networks in the hospital.

219

1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas 199 Biomedical Engineering Scientific Papers

BE4 -1.3 BE4-I.4 Development of a totally integrated hospital Data management in the ICU by automatic means- information system: HUMANE SPEEDSCAN(c) . Michitoshi Inoue and Hiroshi Takeda* 11.Inglis*,A.Pannike,J.Ilueger, Unfallchirurg. Department of Medical Information Science, Klinik,J.W.Goethe Universitaet Osaka University Medical School Frankfurt(Germany/West) Fukushima-ku, Osaka 553, Japan A system for parameter-evaluation, -edition and -documentation with a scanner has been Owing to the chance of the hospital reconstruction developed as a new and error-free tool for data by 1992, the concept of the "intelligent hospital" processing in the Intensive Care Unit. is introduced to Osaka University Hospital. HUMANE Specifications: Using a scanner for data-entry (Human-oriented Universal Medical Assessment system there are no more keyboard entries to be made. by Network Environment) is named for the totally Data are processed after fixing the values on integrated hospital information system to be the data sheets that are configured like multiple core of the intelligent hospital plan. The infra- choice data sheets using pencil-marked items as structure of the HUMANE will be consisted of the regions of interest. There is no more time optic fiber loop LAN (100 and 400Mbps) for digital (and money )lost waiting for altered and data and digital image communication and star-type reprinted data sheets any more, the main factor LAN (10Mbps) for voice and analog image. In develop- in losing lots of time at the beginning of ing the system configuration, the micro-medium-main almost every study as the sheets are printed on frame link (MMML) is considered to handle the cent- the user's impact printer by himself. There ralization/decentralization problem. is no limitation applying the system to The order entry system will be the major part for logical, numerical and character-type data, the integration of the patient-oriented administra- therefore there also is no limitation combining tive and clinical data into a common file and to the data acquisited by SPEEDSCAN with those re- reduce the clerical errors and total work load. As sulting from other automatic tools as doctors and staffs must access the database directly laboratory computers and systems for patient- from the computer terminals, human-oriented man-ma- monitoring. chine interface should be installed. Picture Archi- ving and Communication System (PACS) will be one of goals of HUMANE. As many problems must be solved for the total image information system, mini PACS will be the choice at the initial stage.

BE4 -I.5 BE4-I.6 2n.-/*Mod of Ilealliny ltftkel Peery& trith laser Tahreology. Problems of V.edical Record Storage by a J.H.U. Brown. Ph.D. and Carlos Vallbona. M.D. Universitu of Laser Disk System, Kazunobu Yamauchi,* Houston and Baylor College of Medicine, Houston, Texas Takayuki Miura, Nagoya University Hospital, 55 Tsururnai-cho, Showa-ku, Nagoya, Japan. Working with Sumitomo /CSK of Japan we have been able to place 2 million bytes(700 pages)of information on a plastic credit Anoptical disc filing system is one of the new card by a laser imprinting method. The method is cheap(cards are methods to memorize the contents of medicalre- about $2) and rapid. The information can be read into the card at cords. This system can not only store largea- 64kb/sec and read out at the same speed with an error rate of mount of medical records, but also send informa- about I0- 11. Any piece of information can be located within one tion to several areas in the hospital. However, second. We have developed software to place a complete medical there are some problems of image quality and dura- record on the card formatted so that any portion(demographics, bility of the disc. The present study investigated procedures, treatment, drugs, etc) can be located and accessed. the problems with image quality as to the content of medical records. An Hitachi Optical Disc Filing The card can be read and updated at snit with a simple reader system was used, which is a stand-alone type with which attaches to a p.c. so no main frame computer is needed. The memories of 2.6 GB.Various contents of medical card can be copied or tape backup can be provieds Summaries of records such as discharge summary, present ill- all data can be easily produced on the computer andprinted out if ness, progress notes, ECGs and polaroid pictures .,,aired and x-rays and ecgs as well as identification photos can be were put through the irnage scanner,and were placed on the card. The system I: under test in the Harris County compared with output images. Image data such as Hospital District and at NASA and results will be presented. The discharge summary and present illness were clearly concept being tested is that the patient can carry his own medical reproduced by a resolution of 200 dots per inch record with him at all times and it can be accessed and updated at (dpi). However, polaroid pictures of UCGsor any contact with the health care system. Records are instantly sketches by doctors needed a resolution of 400 available in emergencies, in out patient clinics or the hospital. dpi with 32-64 gradation.ECGs needed a resolution of 400 dpi with 2 gradation. Needed administrative informationsuchasICI) codesand Records in more occupational codes are built into the program to be accessed than 2 colours of ink such as renograms, nurses' notes and doctors' sketches were reproduced in automatically. Billing can, be added as desired. monochrome images, so the colour signals could not be distinguished.In conclusion, the optical disc filing system is useful to store medical records, but it must be improved so as to distinguish 220 colour information. 200 1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio. Texas Biomedical Engineering Scientific Papers

BE4-I.7 Useof Microcomputers in Medical Informatics -Experience in Belgrade Health CareSystem, Z. Bo2oviol V. Brusid, D. Djakovid, V. Srdanovid, Z. StoJiljkovid and M. Trajkovid, Center for Multidisciplinary Studies, Belgrade University, Slobodana Penezida-Krcuna 35, 11000 Belgrade, Yugoslavia.

During the last three years the Group for Medical Informatics at the Center for Multidisciplinary Studies of the Belgrade University has been developing a system approachsuitableforthe needs of the Belgrade health care organizations. Using modern analyzing tools and new software concepts as well as having studied features of the local health care system, various applications have been developed. These applications include: - primary healti care information systems; - hospital information systems; - decision support systems. The paper also discusses some problems arisingfrom specific features of the regional health care system, development tendencies and the possibilities of implementing them in the Belgrade health region.

BE4-I.9 BE4-J.1 Kyoto University Hospital Information Holter Monitor in Coronary Angiography, Network System on Ethernet, G.R. Yu*, Zhen D.S., Y. Zhen and S.H. Bao. T. Takahashi, A. Hirakawa, Y. Okada*, Ren Ji Hosp., Shanghai Second Medical Department of Biomedical Informatics, University. Kyoto University Hospital, Shogoin, Szkyoku, Kyoto, JAPAN. 50 cases, 41 male and 9 female, aged 47- Network and database are the key points of 63, were monitered by Holter during coron- Hospital Information System in the nere future. ary angiography. 46 with EF over 0.6 and 9 CPU's and more than 300 terminals are 4 less than 0.4. Of 35 with fixed stenosis, connected on Ethernet in Kyoto University 32 had a stenosis over 50%; 7 had 1 VD, 21 Hospital. MUMPS manages the distributed had 2 VD and 4 3 VD. Of 15 without pace- database on this network system. maker, all had a HR over 60 during left In former system, there were 10 dedicated lines ventriculogram, 2 sinus bradycardia during to connect 8 CPU's and more than 180 terminals left coronary angiography and 7 below 50/ are connected individually to their CPU's. So min during right coronary angiography, it was difficult to expand the system without followed by 1 VF, 2 VT and 1 sinus arrest explosion of complexity. Ethernet and MUMPS lasting for 4.2 sec. No serious arrhythmia brought us the flexibility to connect new CPU's was observed in 35 cases with pacemaker. or new terminals anywhere on line. There was less change of QRS voltage, PACS and PBX networks will join this network to duration, ST-T and less ventricular pre- realize the distributed multimedial hospital mature beat in cases with than without information environment. pacemaker. Of 4 with EF less than 0.4, all had significant STT changes during angiography. Of 5 with marked sinus bradycardia, persistent VT and transient Q wave, 2 showed spasm and 3 2VD. It was concluded that pacemaker might prevent serious complication during coronary angiography. 221 1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas 201 Biomedical Engineering Scientific Papers

BE4-J.2 BE4-J.3 Source of False Deviation of S-T Segment in Ambulatory Electro- Microprocessor-Based System for Measuring cardiogram and its Elimination Lung Water, Biing-Nan Hung*, Yun-Bin Lin, by a Compensation Circuit Chih-Ming Shen+ and Kang Hsu44, Department K.Iwao*,A.Yokoi,S.Suzuki,T.Goto, of Biomedical Engineering, Chung Yuan Univ., K.Doniwa,K.Kamiya & M.Okajima Caungli, Taiwan (32023), Republic of China Fujita-Gakuen Health University, Toyoake,Nagoya Area,470-11,Japan + Biomdltcal Engineering Department In commonlyusedambulatoryelectro- Tri-Service General Hospital cardiogram through a DR magnetic tape Taipei, Taiwan, Republic of China recorder,false deviation of S-T segment is frequentlyobserved. Frequency re- ++ Chest Medicine Department sponses of the system were checked with Tri-service General Hospital commercial ambulatory ECG systems. The Taipei, Taiwan, Republic of China phase shiftwas assessed by Wagner's method. The characteristics of the systems were found much worse than that of ordinary ECG equipments both in gain and In this study, the IBM PC/XT was used as the phase responses. The decay of -3 dB in mainframe and was connected with two thermal gain was observed in 0.2 Hz and the lead of 45 deg in phase was observed in 0.6 dilution cardiac output computers and one dye Hz. In orderto investigate which of dilution cardiac output system. By applying these 2 responses play major role in the method of thermal dilution and dye dilu- generating false S-T deviation,a computer program was run with 2 frequency res- tion to heart-lung circulation system, three ponses models i.e.,i)real gain responses cardiac output curves were recorded. After obtained in the above test plus fictit- analyzing the dilution curves, the extravas- ious phase responses with no shift,ii) cular lung water (EVLW) were calculated accorc- fictitious gain responses with no cutoff plus the real phase responses. ing to the principle of thermal-dye and single The actual ECGs were fed into the math- thermal indicator method respectively. The matical filters. The false S-T deviation system has been tested in-vitro and in-vivo were found to be cased mostly by enor- to calculate the lung water. mous lead in phase. Intending to diminish The results were the lead in phase to t!.. point of ordi- as reliable as the gravimetric methods. nary ECG recorder,a compensation circuit was designed and inserted in the commercial ambulatory ECG system. In the trial by 20 ECGs,the compensated system showed no false deviation in S-T segment.

BE4-J.4 BE4-J.5 AWORK STATION TO STUDYAMBULATORY Frequency Characteristics of Volume BLOOD PRESSURE, F.del-Pozo7 F.Hal- Elastic Modulus in Finger Arteries berg(£), E. Gomez-Aguilera and J.L. H.Shimazu ,H.Ito, A.Kawarada(Dept Physiol, Zoreda, Dpt. Bioingenieria, E.T.S.I. Kyorin Univ, Mitaka-shi, Tokyo, Japan) & K. Telecomunicaci6n, Madrid, Spain; CD Yamakoshi(The Res Inst, Appl Electr, Sapporo Chronobiology Labs. Univ. of Minne- shi, Hokkaido, Japan) sota. USA Frequency characteristics of viscoelastic Based on the Blood Pressure (BP) ambula- properties expressed by an index, volume tory system we published recently, it has elastic modulus (Ev), were indirectly eval- been developed a BP Work Station, suppor- uated in human finger arteries. Ev was ted on a PC-compatible system, to identi- defined as AP/(AV/Va); where AP is pressure fy the new information content in the am- change and 1V /Va volume changeratio. A bulatory blood pressure recordings, in volume compensationsphygmomanometer was comparisonwith conventional casual BP used to determinebeat-to-beatarterial determinations; that could be pertinent prc.3ure in an index finger, and transmit- inprevention, screening, diagnosis, ted infrared photoelectricplethysmograph prognosis and treatment. to detect AV/Va in a ring finger, simulta- neov,sly. During the measurement of arterial TheBP profile data manipulation proce- volume change in the ring finger, transmu- dures is a menudriven collection of ral pressure (Pt) was varied with applied software packages oriented to: 1) the cuff pressure which was controlled by sinu- assessment of BP reference limits, homo- soidal variations (0 - 30 Hz) superimposed genizedac,ording to different factors; on the same pressure detected in the index 2) the statistical modelling andpara- finger so as to cancel the pulsatile volume metrization of the sistenatic temporal change in theartery. Thus, Pt in the ring variations, to account for a large part finger varied only by the sinusoidal pres- of the BP variability; 3) the estimation sure variation (AP). Therefore, frequency of BP excess indexes. characteristics of Ev were determined by this method among subjects with different Acknowledgments. CICIT. Grant No. 2911-83 ages and sexes during stress loadings. 222 202 1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas Biomedical Engineering Scientific Papers

BE4-J.6 BE4-J.7 Non-invasive Measurement of Instantaneous Detection of Fetal Atrial Activities Blood Pressure in Superficial Temporal and without Signal Averaging, M. Ishijima*, Radial Artery by the Volume-Compensation Y. Takeda, Y. Sakurai, Tokyo Women's Method, K. )Yamakoshi *, S. Tanaka, H. Ohnuma, Medical College, Tokyo, Japan A. Kamiya'', H. Shigazu, A. Kawarada, H. Ito4' & P. Rolfe'',1)Hokkaido Univ., Japan, 2)Kyorin Univ. School Med., Japan, Ensemble averaging is a common approach to 3)Keele Univ., U.K. investigate fine structures of fetal ecgs. The fundamental drawback of the ensemble averaging is We have previously devised an ingenious tech- the impossibility of single beat observation which nique called volume-compensation method, based may exhibit its maximum power under arrhythmias. on the vascular unloading principle, which al- We developed a new system to overcome the situation lows the beat-by-beat measurement of systolic and to observe single beat fetal ecgs including its and diastolic pressure and the recording of the pressure waveformin the human finger atrial activities without averaging. continuously and non-invasively. This employs a The method involved multiple body surface electrode transmittance-type photoelectric plethysmography and a brief estimation of electrical axis of the for the local detection of arterial volume fetal atria by a ultrasound image scanner. The changes in the finger. The present study con- electrodes were placed on the abdomen of expect- cerns the blood pressure measurement in the ant mother in such a manner that the electrical superficial temporal and radial artery based on axis was a center of the spread electrodes in this method. Accordingly, we have designed a The fetal ecg was calculated with the volume-servc system which comprises (i)locally circle. pressurized pneumatic cuff, (ii) reflectance- multiple signals by finite element method taking type photoelectric transducer fixed just under into account of distances among electrodes. the cuff,(iii) electro-pneumatic converter for The method was applied on the third trimester controlling the cuff pressure pneumatically and mothers in an electrically shielded room. The (iv) volume-servo control circuit. Using this calculated results showed distinctive P waves of system we have successfully measured the in- the fetal heart in consecutive beats. The wave- direct instantaneous blood pressure both in the forms were, however, contaminated by emg of the temporal and the radial artery, comparing with the finger blood pressure obtained simultaneous- fetus when it was awake, and diminished according ly by our previous instrument. Due to local to the rotation of the fetus. Otherwise, it was a arterial pressurization this system seems to possible method to observe single beat fetal ecgs allow more stable measurements of longer period from the body surface. of time.

BE4-.J.8 BE4-K.1 TBCHNIOUR EQ8 NONINVVUE DIAGNOSIS DE Future Technologies, Y. Sakurai, Tokyo Women's Medical College, 8-i Kawada-cho, laalUaLat N.B. Reddy , E.P. Canilang, M.B. Rane, and J. Casterline, Univ. of Shinjuku-ku, Tokyo, Japan Akron, Akron, OH 44325 and Edwin Shaw Hosp., Akron, OH 44312 Medical technologies are advancing so rapidly that medical system might change revolutionarily in the Dysphagia, a disorderof the near future. When new information media system swallowing mechanism, is a common problem will diffuse widely in the society, medical system encountered in the rehabilitation of will develop from hospital into "Hospitele", where stroke and head injured patients. We have medical facilities, social institutions and homes recently identified and developed techni- will be closely connected each other with ques to quantify several biomechanical information network on the basis of person to parameters that characterize oral muscula- person relationship. In this system, a mobile ture in dysphagic patients (Reddy etal medical facility is important to deliver a medical 1985). In the present investigation, we practice to the spot. have developed techniques, to noninva- The development of superconducting element might sively measure the paryngeal phase and give a strong impact to medical technologies. The coordination. Acceleration was measured operating theater is too big to accommodate only at the throat by placing two ultra one patient who should be surrounded with strictly miniature accelerometers on the skin over aseptic atmosphere. So the new concept of the throat, and swallow suction pressure "operating capsule" might be developed which will was simultaneously monitored. In normal be made of transparent wall and a patient will individuals swallowing gave rise to float in the air inside the capsule with a magnetic typal acceleration pattern andthere power coming from superconducting materials. was no time lag between the pressure and In this paper, several kinds of future dream in acceleration waves. In dysphagic patients, clinical engineering will be mentioned in SF way. the acceleration pattern was either absent or was delayed,and was significantly different. This technique can aid the clinician in diagnosis of dysphagia.

6'. f,

1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas 203 Biomedical Engineering Scientific Papers

BE4-K.2 BE4-K.3 In vivo Cerebral Metabolic Monitoring with Near Future Perspectives in Cardiac Assist: Infra-Red Spectroscopy (NIBS). Totally-Implantable Permanent Artificial Heart P. Rolfe, Maureen Thorniley, YABD Wickramasinghe. M.McGee *, R.Kung, S.Parnis, 0.H.Frazier University of Keele, Stoke-on-Trent ST4 7Q3, U.K. Texas Heart Institute, Houston, Texas and Abiomed, Inc., Boston, MA. The scarcity of donor hearts limit the benefits In vivonear infra-red spectroscopy (NIRS) has of heart transplantation to 2000-2500 of the emerged over recent years as a method with the estimated 17,000-35,000 patients needing cardiac potential for continuous monitoring of important replacement. Mechanical devices to support or biochemical phenomena. Characterisation of the replace temporarily the failing heart are absorptionspectra of all absorbing species is currently being used with encouraging results. essential if quantitative measurements are to be However, these systems require large power/control made. The relevant absorbers in the range 750 to consoles which limit patients' mobility and are 910 mm are oxy and deoxy haemoglobin, oxidised and therefore not suitable for long-term support. We reduced cytochrome oxidase, and a generalised are involved with the design, development and 'tissue' component. testing of a permanent, electromechanically- actuated total artificial heart, under a program Instrumentation for neonatal cerebialmonitoring supported by NHLBI. The design cr'teria for this comprises laserdiode sources at 4 or more system are: 5 years of tether-free operation; flow wavelengths, optical fibres to convey the rates of 8 L/min at < 120 B/min into a mean aortic radiation to and from the head, optical detectors, pressure of 110 mmHg and mean pu monary artery and processing facilities. Our studies have been pressure of 25 mmHg; autoregulated control carried out on purified enzyme, cell suspensions, responsive to varying circulatory demands; perfused organs, intact rat brain with both blood non-thrombogenic, non-hemolytic blood-contacting and fluorocarbon perfusion, human adult and materials which minimize or eliminate the need for neonatal limbs, and neonatal heads. The anticoagulation; anatomic and physiologic feasibilityof continuous in vivo monitoringhas compatibility; rechargeable external battery power been demonstrated. for >10 hours; and rechargeable implantable battery power for >45 minutes. These efforts will result in a mechanical system which will provide permanent circulatory support with the patient quality of life required.

BE4-K.4 BE4-K.5 Advances in instrumentation for Analysis of Sound at Urethral Flow, flow cytometric measurements, H. Terio, P., Ask, Departrnent of Biomedical J. Eichmeier and V. Kachel, Tech- nical University Munich, D-8000 Engineering, University of Linkoping, 581 85 Munich, FRG Linkoping, Sweden A new flow cytometer has been developed which combines electrical cell volume The idea behind this project is to try to find a measurements with simultaneous multipara- noninvasive method to detect obstruction of meter fluorescence detection. The flow urinary flow by sound measurements at the chamber is suitable particularly for fast body surface. In the vascular system it is sample exchange and kinetic measurements. known, that turbulent flow through obstruc- A mercury high pressure lamp with broad tions produces sound, the spectrum of which light spectrum allows the excitation of can be used to assess the obstruction. Low a large number of fluorescent dyes. The data are evaluated by an IBM-AT-Personal frequency peaks in the sound spectrum are Computer, stored in list mode and pre- related to the elasticity of the artery whereas sented on a color display in the form of the changes at the higher frequencies are re- one-, two- or three-paramter histograms. lated to the degree of the obstruction and the Areas of application of such devices are flow velocity. tumor diagnosis, immunology and basic me- We have measured generated sound at various dical and biological resaarch. degree of obstruction in an urethral flow model. In these experiments we have used latex and silicon tubes with different wall thickness and diameter. The obstructions have been produced by squeezing the tube with a rod connected to a micrometer screw and the sound was recorded with a condenser microphone. The sound spectrum shows clear intensity increase with increasing obstruction. Some changes in the frequency spectrum c:laracte- risticz could also be observed. 224 204 1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas Biomedical Engineering Scientific Papers

BE5-A.1 BE5-A.2 Mechanical vs. Metabolic Factors in Effect of Spaceflight on 114 Rat the Development of Osteoporosis, Skeleton, David J. Simons , University of Texas Medical Branch, Manmohan Singh*, Michael Reese Medical Center, Chicago, IL 60616 Galveston, TX 77550 In the postmenopausal female patient In young growing rats, spaceflights of 7-18.5days retard the rate of bone formation without osteoporosis results from two basic eti- ological factors: (i) the bone may resorb affecting bone resorption. Total bone mass is reduced, particularly in the trabecular regionsof in response to a metabolic need for cal- The vertebrae are weaker than normal cium, or (ii) the bone may adjust to theskeleton. lowered mechanical stresses as activity when tested incompression,but the torsional strength of the long bones are not compromized. le"el and muscle mass decline with age. The ratio of mineral to matrix is notabnormal It is possible to separate two types of (calvaria, ribs, jaw, femur, incisordentin). bone loss by measuring the bone density of radius at two sites in the forearm. Gradient density analysesdemonstratethat the global decrease in bone formation in microgravity Distal bone density (DBD) is measured at is associated with qualitative changesin bone the "5 mm site" which contains about 50% compodition, and that these disturb the maturation trabecular bone. Middle bone density (slowly) forming matrix and mineral (MBD) is measured at the "2/3 site" which of the contains 98% cortical bone. The two moieties. Spaceflight compromizes the formation of collagen intramolecular links and the growthof readings are expressed as percentages of apatite crystals. Fourier analysis of dentin adult mean values and compared. Two Type A: if electron microprobe scans indicate thatthese types of bone loss are seen: maturational deficits do not reflect an abnormal MBD is greater than DBD by 10% or more. phasing of the circadian and ultradian Ca-P It indicates an accelerated loss of biorhythms. Rather, they point to the influence of trabecular bone probably due to bone Type B: persistently high matrix sulfur concentrations resorption for metabolic needs. (proteoglycans) on bone biomineralization.We if MBD equal DBD to within i 9%. It indicates a balanced loss of trabecular conclude that microgravity affects the maturity of the growing skeleton and, thereby,its and cortical bone probably due to adjustment of bone to lowered mechanical mechanical strength. stresses.

BE5-A.3 BE5-A.4 A Cellular Based Computational Skeletal Biomechanics and Aging, Subrata Model for the Depostion and Saha: Dept. of Orthopaedic Surgery, Lou- Resorption of Bone Tissue, S. C. isiana State University Medical Center, Cowin: Tulane University, New P.O. Box 33932, Shreveport, Louisiana Orleans, LA 70118. USA. 71130. Although many investigators have tested the load- Tie transduction of mechanical information carryingcapacities of whole bones and the mech- to chemit...1 information, the mechanism underlying anical properties of bone tissue, only few have the functional adaptation of bone tissue (Wolff's focussed their attention on how these properties Law) has not been established.Furthermore, the change with age. Similarly, only a few authors means by which cell populations are subsequently have attempted to quantify the relationship be- controlled has not been identified. In this tween the changes in the microstructure and chem- contribution the evidence concerning the biological ical composition of bone tissue and its mechanical strain transduction mechanism is briefly reviewed. properties. Also, little is known about how ch- A model for the surface deposition and resorption anges in the shape of a whole bone (either in dia- of bone tissue based on the cellular level concept meter or cortical thickness) affect its load- of a cell activity function is revi wed.The camIng capacity. A complete picture of the sim- applications of this model to the predection of ultaneous changes in the biomechanical, structur- bone modeling and remodeling are briefly al, and microstructureal variables is almost never summarized. available because authors who test whole bones often do not examine the microstructural or chemical changes in the bone tissue. Similarly, investigators who test small, machined bone specimens often do not consider the overIll size or shape of the whole bone. Available data on the changes in the mechanical properties of whole bones and bone tissue will be presented. How the load-carrying capacity of whole bones changes with decreased cortical thickness (with aging) will also be discussed. k4-6 r.

1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas 205 Biomedical Engineering Scientific Papers

BE5-B.1 BE5-B.2 In-Vivo Mechanical Response of the Human Growth Micromechanics Modeling of Compact Bone Plate Close to skeletal Maturity, J Kenwright, A J Tissue, H. A. Hogan , Mech. Engr. Dept., Spriggins and J L Cunningham*, Oxford Orthopaedic Texas A&M University, College Station, Engineering Centre and Nuffield Orthopaedic Texas 77843-3123. Centre, Oxford 0X3 7LD, U.K.

Recent studies on leg lengthening using epiphyseal Compact bone tissue can be treated as a fiber distraction have suggested that lengthening of the reinforced composite material with at least growth plate without fracture is possible in three phases. Osteons form the main fiber immature subjects using low rates of distraction. component and are surrounded by a concentric If this process were possible in patients close to layer of cement line material. These structures skeletal maturity it could be advantageous since are assumed to be arranged in a close packed most lengthenings are performed at this stage when hexagonal array embedded in a matrix of inter- an accurate assessment of leg length inequality is stitial lamellar bone. A micromechanics possible. In this study the distraction forces analysis of this system has been studied using a have been measured and are related to in-vivo finite element model of a representative volume biological events. element of material. A two-dimensional mesh Axial force during distraction was monitored daily with linear elastic constituent material using a purpose built loadcell attached to the properties has been considered initially. The external fixator. Slow distraction rates of 0.5mm/ micromechanics approach allows study of the day were applied as were constant distraction relationship between constituent properties and loads well below those found in-vitro, from an overall macroscopic material rzsponse. Also, amputation specimen, as leading to fracture. "micro-level" stress distributions are examined. For all the distraction regimes applied it has not For a given macroscopic, or average, stress or been possible to lengthen the limb without there strain level, the finite element. model gives the being evidence of fracture, as assessed both detailed stress and strain distributions amongst radiographically and by a sudden decrease in the the separate model constituents. Refinements to peak distraction force. With both constant include more realistic constitutive relations distraction rates and loads growth plate failure for the discrete components are also considered. occured between 600-800N, considerably below that seen in-vitro of 1700N. It would appear therefore, that fracture must always occur in this age group.

BE5-B.3 BE5-B.4 REGIONAL BONE DENSITY IN THE PROXIMAL FEMUR Stress Relaxation Behavior of AND ITS CORRELATION TO BIOMECHANICAL PROPERTIES Three Human Bones, A.M. Ucisie MD Smith, DL Dickie*, A Cooperman, SA Goldstein, and C. Bindal, Bogazici University MJ Flynn, Universityof Michigan, Ann Arbor, and and Istanbul Technical University, Henry Ford Hospital, Detroit, MI. Istanbul, Turkey The cost of hip fractures is enormous, in both human suffering and financial respects. In order Three human bones, tibia, femur, fibula, to assess the significanceof regional bone which were amputated from same person, mineral den'sity in the proximal femur, twenty-two were isolated from soft tissues and kept cadaver specimens were scanned using a density in deep freeze until experiments run. specific computed tomography protocol. They were Samples taken were close to the knee then embedded into steel collars and loaded joint. Loads were applied with constant anatomically until fracture occurred. Subsequent crosshead speed on the several different hip screw pull outtests were performed, which parts of single crossection of each bone measures the ability of the bone to be successfully repaired using this fixation device. by a diamond indentor specially attached Five regions in the femoral head and neck area to the Instron Universal Testing Machine. were measured using interactively positioned After reaching to the previously decided regions of interest on a coronal plane (parallel value of load, by keeping crosshead level to the neck and shaft surfaces). Traditional constant, stress relaxation started. Singh values from standard radiographs were also Comparisons in terms of stress relaxation determined for each femur by an orthopaedic as well as elastic properties were made resident, an experienced orthopaedic surgeon, and for each single bone depending on test a radiologist. test location and among three bones. Due The correlation of resident's Singh index to to strong anisotropy, microstructural fracture load was r..593, the orthopaedic differences, relaxation time and the a- surgeon's and radiologist's were r =.463 and mount of load relaxed were not same ev- r=.333, respectively. The regional densities erywhere on a single crossection of each nearthesuperior and medialsurfaces of the bone. Differences among tibia, fibula femoralhead correlated to fracture loadwith and femur were stronger. Thus, mathemat- r=.798 and r..719, respectively. The screw pull ical modelling of bone with available out resistance and lateral surface regional data is almost impossible. density were correlated with r =.837. 226

206 1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas Biomedical Engineering Scientific Papers

BE5-B.5 BE5-B.6 Anisotropic Viscoelastic Constitutive A Simple Constitutive Equation for a Equation of Trabecular Bone of Human Cortical Human Bone,Jia-Ju Sun* and Spine, Cheng Peilai*, Dept. of Mechanical Jie Geng, NW. Inst. Archit. Engng., Engn, Tianjin Institute of Technology, Xian, Shaanxi, P.R.China. and Chen Hongxing, Tianjin University, Tianjin, P.R.C. The anisotropic viscoelastic constitutive This short study presents a simple viscoelastic equation has been established by uniaxial constitutive equation for the cortical human bone constant loading creep experiments on 37 lcm3 with haversian systems. The equation is obtained. cubic specimen cut from human spine. The The mechanical model is considered in accordance trabecular bone has been assumed as a with the testings. transversely isotropic body. Two kinds of lcm3 cubic specimen were cutfrom L1 -L5:one alongthevertical and treansverse direction andanother one along 45° and -45°direction to vertical axis. Altogether 37 cubes were cut. The load was applied on vertical, transverse and 45° directions of trabecular respectively. The displacements along the loading and lateral directions were masured simultaneously. The experimental results showed that trabecular bone behaves as anisotropic viscoelastic material Andthecreep ratevaries with its direction. The two-, three-, and four-parameter Kelvin models were used to model the experimental results and the three-parameterKelvin model was showntobe excelent with average errors of 1.26% .The time-dependent anisotropic viscoelastic constitutive equation of trabecular bone of human spine was established andall parameters in constitutive equation were obtained.

BE5-D.1 BE5-C.2 Quantitative Evaluation of Bone Mineral Redistribution of Bone Tissue in the Humeral Cross- Density. C.C. Gluer*, J.E. Block, Section of Growing Rats Following Swimming Training, P. Steiger, H.K. Genant, University of A. Siskin, I. Leichtera'and A. Swissa. The Jerusalem California, San Francisco, CA 94143. osteoporosis Center, Dept of Orthopaedics, Hadassah University Hospital, Racah Institute of Physics and A variety of techniques for non-invasive Cusell Center. Hebrew University, Jerusalem. Israel. quantitative evaluation of bone mineral density are available for assessing metabolic bone Theeffect of non-habitual physical activity on diseases. Quantitative Computed Tomography (QCT) hone remodelling at two sites along the rat humeral has the capability for precise three -Idimensional shaft was examined in this study. Two groups of anatomic localization, providing a selective rats were trained to swim for one hour a day. for measurement of highly responsive trabecular ..0 weeks at, tw6 training levels. The control group bone. Mainly applied to the assessment of the .onsistedof sedentary rats. Parameters of cross spine, techniques for hip mineral quantification ,iectionai bone morphology were used to evaluate the are also being developed. Results obtained on resuonse ofbone architecture, to mechanical. 538 healthy women illustrate the use of QCT I.,ading. The strewth of bone wasasssessedby spinal measurements in the delineation of normal hy,tatirihe theultimate compressive force and age-related bone loss. For peri-menopausal women -,tress. The cortical crosssection areaand accelerated bone loss of -4.4 r4/m1/year orincit.al Art-.."1 nr.ments of inertia were found tobe (p<0.0001) was found as compared to a non- sianifioantly hiaher. byupto13% and h7%. significant decrease of -.45 mg/ml/year for pre- rest.,ectivelv. in the swimmina groups than inthe menopausal women and an average rate of change ontiN Examination ofthe ratiobetween the of-1.9mg/ml/year (p<0.0001) for post- ma3or andminor area momenth of inertiarevealed menopausal women. Other studies comprise the that tio, bone became more ourvcmemic. The ultimate evaluation of estrogen effects on bone, the ,omnrsive fon-e was ,-aanificantly hiaher. by on assessment of fracture threshold and risk, and in the swimmino rats. while the changes in the study of the association of various exercise altimate .:tees;; were not sionificant. Our results regimens to bone mineral density. QCT will be indicate a aain of hone strenath due t, increa:-.ed put in perspective to other quantitative aprosition andmodified bone tissue techniques such as Dual Photon Absorptiometry Ji-trihution. The marked change:. in bone morphology and Quantitative Digital Radiography. are attributedtothe differed. nature ofthe forcesand moments exerted on thehumenis during ,g4imming. orwared to loccxnotion. 22 1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas 207 Biomedical Engineering Scientific Papers

BE5-C.3 BE5-C.4 CARBON ARTIFICIAL BONE REPLACEMENT FOR BONE TUMOUR MECHANICAL PROPE3TIES OF THE TRABE- TREATMENT (7 CASES REPORT; - Chen Lantian and Jian CULAR BONE CF THE VERTEBRAL BODY, Kunyuan, Central Hospital of Jilin City, Jilin Pro- ZHU DCIGMING*,CUI JIAN,SUN CHANGZHU, vince, P.R.C. XU YANGHE,SU YONOJUN,BIAA,PRC. With the development of bd-medical engineering, An experiment on the mechanical properties of people paid more attentic. to Artificial bone, the vertebral cancellous bone has been perfor- through the further study to carbon material, we med. About 314 cancellous specimens from the started making Artificial bone with Iow Temperature humap vertebrae and 139 specimens selected Isotropic carbon in 1979, we had done the treatment frompig's vertebrae were compressed to break of bone tumour with replacement of Artificial bone along the vertical axis,the transverse axis for seven cases until 1984, two cases of 2-3 degree and the sagittal axis respectively. 8 intact giant cell tumour of lower end Radius, three cases vertebrae were compressed,too. The strength, of 2-3 degree giant cell tumour of upper Humerus the elastic modulus,the Poison's ratio and (one case relapse), these five cases were serious- the apparent density ofthe cancellous bone ly eroded and damaged by swelling giant cell tumr of human vertebrae and the pig's vertebrae our, they could not be treated with excision and were determined. The pattern and the feature bone graft or bone cementing because the soft tis- of the trabecular bone were observed and sue was eroded by tumour.One case of Fibrohaeman- the mechanical properties of the cancellous gioma of upper right Ulna with pathological fract- bone affected by its trabecular structure ure. One case of fibrous dysplasia of right upper were discussed. The results showtthere is no Ulna with pathological fracture. senificant difference in the mechanical pro- Follow-up - the longest for these cases was eight perties between the parameters in the direction perpendicular to the vertical axis. However, years, the Shortest was two years and two months, the average was five years and ten months (three the streneth and the elastic modulus in the axial 4irection are significantly erenter than cases follow-up time more than six years). In all cases, no loosing, no relapsing of the tumour, and those in their perpendicular directions. Both no metastasis would occur. the arparent density and the trabecular structure are two important factors affecting the mechanical properties of the cancellous bone. Follow-4p examinations revealed that the results were satisfactory.

BE5-D.1 BE5-D.2 A Review of Total Hip Prostheses, J.A. Albright*, laterial and Engineering Technology in Orthopedic LSUMC, P.O. Box 33932, Shreveport, LA 71130 Implant; Yesterday, Today and Tomorrow. Dane A. Miller: Ph.D. biomet. Inc.. P.O. box 587, Warsaw, Indiana, 46580 USA. The useof implantsin the human hip joint has increased dramatically overthe past 75years. Archaeological evidence from prehistoric times The recordincludes: experiencewithnumerous suggests that fur many centuries, attempts have materials at the turn of the century; the been made to treat orthopedic injuries and Smith-Peterson cup arthroplasty; the Austin diseases with implanted foreign materials. With Moore centuries of historic experience, it has been only prosthesis; the Judet prosthesis; the in the past two decades that the successful total CharnleyTHP, together withinnumerablesecond joint and trauma products of today have tvolved. generation modifications by others. Cemented Of special note are the total hip and total knee hip prostheses were universally accepted in prosthesis which have revolutionized the treatment the United States withouta serious evaluation and prognosis of todays arthritis patient. Until of efficacy. Now there is a growing disaffection the advent of contemporary total joint replacement, the arthritic patient was subjected withcement and a current shift toward porous to a series of pharmacological and surgical surfaces, but limited information on results. treatments of limited success which ultimately resulted in the disability of the patient. Upon review,several points emerge:1) The body Today's success rates, often reported above 95% isolates most foreign materials anddoes not good or excellent, are a result of a combination of new materials, new design information and act as if they were an integral part of a improved surgical techniques facilitated by functional unit. Those materials which have improved instrumentation f- the surgeon. The been utilized to date appear inadequate. 2) arthritis patient of today .an, almost without Technical design, material properties, and reservation, look toward a return to full activity surgical skill determine short-term success. following total joint replacement. As we look ahead to the next decade, similar "improvements" 3) Physiologicalresponse determines long-term in bothevolutionary and revolutionary technology success. 4) The response of bone cannot be can be anticipated. Several examples of this wave determined with certainty for an extended r.riod of future technology will be presented. of time.The number of years needed is dependent on the magnitude of the response. 5) The acceptance of new devices and procedure by orthopaedic surgeons is very responsive to marketing strategies. 228 208 1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas Biomedical Engineering Scientific Papers

BES-D.3 BE5-D.4 "The Orthopaedic Implant Manufacturer: Challen- COMPOSITE MATERIALS FOR TOTAL JO1RP REFIACEMEM ges for the Year 2000," Richard W. Treharne: Michael T. Manley*, Osteonics, Allendale, Richards Medical Company, Memphis, Tennessee Composite bicmaterials are the next major step in total joint replacenr_nt technology. At present, All US-based orthopaedic implant manufacturers orthopaedic coupcsites fall into the categories of will have many challenges to face by the year coated metals (particularly ceramic coated titani- 2000. The first of these challenges will be an um), polymers reinforced with chopped fibers, de- increase in aiiiind for products. The aging of sign specific carposites and laminated ccaposite the "baby boomers," the gradual increase in life structures. Clinical experience has been gained expectancy, and the advances being made in all from the first two materials while the latter two other non-orthopaedic areasof medicine means have been used only in animal studies. that the year 2000 will have an increased usage of orthopaedic products. A second challenge The major advantages for composites in orthopaedic will be regulatory. The powerWEithority by applications is the new possibilities available to the FDA over the orthopaedic implant manufac- implant designers. Material anisotropy and the turer continues to increase and will pose a se- tailoring of material properties are now a realis- vere challenge by the year 2000. A third chal- tic possibility. Biologically active ccaponents lenge will be financial. Third-party reimburse- can be deliberately incorpo-ated into the material ment will always be a critical issue and even design. Ion release to tissue present in all metal more challenging in the future. A fourth chal- systems can be obviated and the radiolucent char- lenge will be legal. Product liabriTTFissues acter of composites will allow more careful moni- will be a challenge for all companies in the toring of bony interfaces. Composite materials future. A fifth challenge is technology. What have the potential for long term joint replacement will be the new technical horizons in the fu- rather than limited term joint substitutes. This ture? Where willthe advances in ligaments, presentation dill highlight the use of composites composites, externalfixators,and other bio- In orthopaedic applications with illustrative clin- medical devices come from?These challenges and ical and animal studies. questions of the future will be addressed.

BES-D.S BE5-D.6 A Biodegradable TM Nail with The Design and Testing of External FixatorBone Interlockinr. Crosspins: Screws. J L Cunningham'', M Evans and J D Harris, Computational and Experimental exults, Oxford Orthopaedic Engineering Centre, Nuffield * H. A. Hoganand D. A. Hulse, Texas ACM Orthopaedic Centre, Oxford 0X3 7LD, U.K. University, College Station, Texas still 77843-3123. In external fixation bone screw loosening is a major problem. In this studythe design factors affecting the mechanical strength at the bone experimental A biodegradable It! nail with interlocking screw interface were assessed, and intended to crosspins is under consideration. Because the screws designed and tested which are stiffness of such materials is much lower than maximise the holding power of the screw and that of steel, a series of interlocking cross- minimise screw loosening. pins are needed for sufficient stability. Push-in, pull-out and bending tests on various commercially available 6mm screws and the Understanding the mechanics of the deformation experimental screws were carried out both on a and load transmission behavior of the device is synthetic material and in bone. Fatigue tests were an essential element in its ultimate develop- also carried out on all of the screws todetermine cent. A simplified finite element model of the susceptability to loosening. Photo-elastic tests system has been developed and studied. Initial also trends in the variation of overall implant on different screw threadforms were stiffness with various system parameters have performed. The results of the push-in and pull-out tests been established. A wooden model of the system indicate that the cutting head and threadform of has been tested experimentally and reveals of the screw have a marked effect on theholding highly nonlinear behavior due to local material strength. The bending stiffness of a screw can be failure near component interfaces. The current increased by 33% by a small (10t) increase in core model, therefore, only models the initial diameter. The number of fatigue cycles to loosen a response of the system at very low loads. screw can be influenced by both thethreadform and Modification of the model to include these by an increased core diameter. Thephotoelastic effects has been initiated and is continuing. tests show that most of the applied load is Developing an adequate, yet computationally distributed over the first few threads closest to cost-effective, model is the overriding goal of the applied load, and that the area betweenthe these efforts. thread crests is subjected to high shear stresses. 229 1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas 209 Biomedical Engineering Scientific Papers

BE5-D.7 BE5-E.1 Carbon-Titanium-Combine Femoral Head Prosthesis A Biomechanical Analysis of the SteffeePlated (A Clinical Report 100 Cases), ChenLantian*, Jian Thoracolumbar Spine, Stephen P. Courtney*,John Kunyuan and Wang Shiqing, Central Hospital ofJil- M. Lipka, Subrata Saha, andJames A. Albright ian City, Jilian Province, China. Orthopaedic Surgery, LSU MedicalCenter P.O. Box 33932, Shreveport, LA 71130-3932 Since January 1980 through November 1986,replacement arthroplasty with Carbon - Titanium-Combine Patients with unstable thoracolumbar spine femoral head prosthesis was done in 100 cases; fracturesbenefit from surgicalreductionand there were 64 fresh (subhead fc-Jral neck frac- rigid internal fixation. The Steffee spinal ture), 20 old, 3 pathologic (Aneurysmal bone plates represent one type of transpedicular cyst) femoral neck fracture, 13 Avasailar necro- screw fixation device currently utilized for sis of femoral head. stabilization.

The longest follow-up was 7 years and 8 months The torsional, flexural, and lateral bending and the shortest was 1 year and 11 months.The stiffness of 2 embalmed and 2 freshthoracolumbar average was 3 years and 10 months. spines were measured before and aftera simulated fracture of the 12th vertebra fixed withSteffee Follow -up examinations revealed that the resulLs spinal plates. The Steffee platedthoracolumbar were satisfactory (only 2 cases occur sinking of spines did not differ significantly inbehavior the prosthesis, 1 case periarticular Calcification, from the normal thoracolumbar spines intorsion. 2 cases suffered traumatic femoral shaft fracture). In flexion the plated thoracolumbarspines were significantly less stiff (p<0.05) than thenormal thoracolumbar spines. In lateral bending, the Steffee plated spines wereless stiff than the normal thoracolumbarspines, but thedeviation fromnormal was not significant. Removal of the anterior plated nuts had no significant effect in thebiomechanical performance of the Steffee plated spines.In summary,the Steffee plated thoracolumbar spines behavedvery similar to normal in torsion.

BE5-E.2 BE.5 -E.3 The Biomechanics of L-Rod Interlock to The Biomechanics of Neck Trauma-Injury Prevent Sidesway of theScoliotic Spine, Envelopes and G. McNeice*, and B. Allen Jr., Divisionof Applications to Gymnastics, Orthopaedics, University of Texas Medical Branch, G. McNeice*, T. Oxland and P. Scott, Civil Galveston, Texas, 77550. Engineering, University of Waterloo, Waterloo, Ontario, N2L 3G1, Canada.

This study was prompted by clinical evidence of post-surgery progressive sidesway of some Catastrophic cervical spine injuryamounts to pediatric spines during the fusion period. A approximately 10,000 new cases each year in the U.S.A. spine model was constructed to provide simulation Estimates indicate each of these cost society about $400,000 and of the ligamentous spine.A test table was that about one-third is due to falls. designed to allow incremental apical rotation with The present paper summarizes the lateral load application. Standard 3/16" diameter injury tolerance data and proposes injury envelopes which can be used to predict 4.njury L-Rods were wired over the thoracolumbarspine and thenloaded laterally to demonstrate the severity following accidental humanimpact. kinematicsof wire-rod interplay and subsequent Application is made to the study of safety matting slippage. An interlock device was designed and ns used in the sport of gymnastics. Only vertical component tested to provide inter-rod locking. head-first impacts were simulated experimental:. The device was constructed using MP35N metalfor The test f(,:ility comprised a drop impact mechanism wit use of either 316L Stainless Steel or MP35N spinal variable mass and impact velocity. rods. Additional instrumentation including A combination of impact force,moments, n acceleration and torque device and operating room calibration was displacement weremeasured designed. The interlock device was designed to simultaneously to allow full assessment of resist a 2000 N shear slip force in order to impulse, energies and deformation.Additional resist a lateral load of 500 N placedat the top tests were conducted using a Hybrid III of the spine. This was estimated to provide a anthropomorphic test dummy for impactson mats safety factor range of 2 to 4 times the expected currently on the market. lateral loading. Impulse values remained insensitive to impact severity, suggesting that it is not a good parameter for injury assessment. Protective matting has alimiting percentage of energy absorbency and appears to remain constant forany 230 one mat construction. 210 1988World Congress on Medical Physics and BiomedicalEngineeringSan Antonio, Texas Biomedical Engineering Scientific Papers

BE5-E.4 BE5-E.5 Space Truss - A Systems Approach RESEARCH ON MECHANICS OF THE SPINE, Zhu Dongming*, To Cervical Spine Mechanics, Sun Changzhus Xu Yanghes Zhang Hua Sheng Jingfangs S.M. Levin, Potomac Back Center, BeiJing Languages Institutes BeLlings P.R.C. 5023 Seminary Road, Alexandria, VA 22311 The study on the mechanical Properties of each section of the vertebral column is the foundation of A space truss system is proposed as a biomechanics of the spine. Research has been con- structural model for the ce.vical ducted on the mechanical properties of the inter- spine to replace the present post and vertebral disc and the cancellous bone of vertebrae. lintel or lever model. Lever models The anisotropy of the vertebrae and the intervert- are reductionistic and present apparent ebral disc was calculated, analyzed and given ex- solutions to local problems but fail to planations. The method of finite element and its give a comprehensive understanding of application in biomechanics was discussed.Three biomechanical design. Systems theory models of the motion segment were introduced: 5?11- modeling is used to develop a rational ker's axially symmetric models Shiraziadi's model basis for the use of a space truss with an of the 2-3 vertebrae of lumbar sPine and its inter- internally vectored icosahedron as vertebral disc from a 29-year old female body and the finite element of a heirarchical Simon's model of Porous materials. truss system. Some advantages of a truss system are omni-directional stability and Also, the biomechanical model of the sPine has been within the same instant, flexibility, discussed. local load distribution and low energy requirements. Physical properties of biological material and icosahedral truss systems coincide and allow for unique biologic characteristics such as nonlinear stress-strain curves and Poisson's ratio approaching unity. One basic model is applicable to all cervical spines of any species under any condition of loading.

BE5-E.6 BE5-E.7 The Biophysics of the Spine in the Etiolation Biomechanic Research of Treating in the Thoracic of Man on the Earth,H.Po.anc!',' "Functional Pe and Lumbar Vertebrae Compressed Fracture education and Rehabilitation",La :Sapienza Gu Yun-Wu, Wang Xiao-tone, Han Hui University, 00185, Rome, Italy Tianjin Hospital Orthopaedic Institute Tianjin, P.R.C. The A. shows a "human model" selected from sone case studies which consider subjects An a state of AnsuffAcAent reireene- Functional exercises with a pillow under the part ment In the sagittal plane 4hypers.l.photAs with invol%ement of lumbar compressed fracture, mean while a spinal of the non-rigid cervico-thoracic spine+ which reset rack is putted on the part of thoracic and lumbar favorabily to a vertical co"pression applied by the opera- fracture, are a good method of treating in thoracic tor's hand on the neck/trapezius angle.Peactionmeans and lumbrate compressed fracture. "functional verticality" which according to the biomechans- According to the animal experiments, fresh spinal cal low invariability may also have evoked a dynamic erect specimen experiments and clinical test,itis evi- posture as it should ha" been experienced by the evolutiona dent that the more serious compressed fracture is, ry families between gene.a Hominoid and Homo.The above func- the smaller reduction force is needed. tional adjustments should have been compulsorilyassumed Although using the method of function exercises while the evolutionary families carried huge quantities nf heavy materials towards the open plains to defend themselves with a pillow, the function of the pinal would still against aggressors throught the construction of rudimentary recover slowly due to staying on bed SG !Ong time. but effective walls.The A. studying the "human model" defi- For solving this problem the spinal rack was used nes the weight (about 12,5 Kg) and location (neck/trapezius to keep a fixed force so that the patients could do angle) of these materials which a transitional being between functional exercises on the ground. genera Hominoid and Homo should have favorabily carried.In this paper the A. analyzes the compressive/tensile forces which from the necessity to mantain the erect posture may have guided towards the structural changes observable in fossil remains.

231

1988World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas 211 Biomedical Engineering Scientific Papers

BE5-E.8 BE5-F.1 Experimental Study on Effect of A Theoretical Analysis for the Spine TransmittingVibration, Characteristics of Articular Joints, Liu Zhen-tian,and LiuRun-nan, M. Zaki ,Lept. of Industrial Prod., Nanjing Institute of Technology, Mar ara University, Mansoura, Egypt. P. R. China

A simplemodel of two porousparallel plates In this paper, the caused vibration in with reference to the human articular joints the roads ofdifferent flatness were was suggestedandstudied in the presence of chosento be as the exciting source of synovial fluid. Variation of pressure, load capacit), vibration situation. The Vibration time of approach, concentration of lubricant, information of the vechical and haman minimum film thickness and the effect of cartilage buttock and human head wascollected. porosity upon lubrication, have been investigated. Using signal digitalprocess technique The mathematical analysis wasbasedon the to analyse the power srectrum, in the continuity of pressure and velocities at the inter- viewpoint of energy, we discussed the face of thesynovial fluid film and thethree transmittingeffect of the vibrated cartilage layers possessing different porosities. human body,and studied the transmitting The study brings out many valuable and important situation apong the cars chassis, the results which are in good agreement with earlier chair and the head. Vibration strength experimental observations. transmitting through the spine,frequenee beyond 6fiz is absorbed by the spine and other physiological parts of the body. Absorbed energy protecl,s the head, but this part of energy is dissipated in some physiological parts (like heart, lung, stomach,liver and four limbspct.), excessive energy will injure these parts.

BES-F.2 BE5-F.3 Measuring System for Contact Pressure The Biomechanical Analysis of Distribution in Artificial Knee Joint, H.EUrosawa*, K.Yunakoshi,T.Yamanoi and Removal of the Meniscus, Wang A.Kamiya,Pesearch Institute of Applied Yi-jing; Shanghai University of Electricity, Hokkaido University, Sapporo, Japan. Science and Technology, China.

Although a number of biomethanical studies of the human knee joint have been carried out, the kine- Neniscectom7 is one of the most frequent- matics of the knee joint is not fully understood. ly performed orthopaedic operations. In The artificial knee joint replacement surgery is usually indicated for the severely damaged knee. this thesis, removal of the meniscus is TO datemany kinds of artificial knee joints have analysed biomechanically with Finite Ele- been designed.In some cases reoperation is needed to treat complications such as loosening of pros- ment Nothod (FEN) . A two dimentional thesis. We have developed a microcomputed-based plane stresses FEN has been developed by system which allows the measurement of contact pressure distribution inartificial knee joint author, using the method of combining FEN during operation in order to prevent complication with strain gauge measurement with the due to malalignment of the joint. The system is.L.-posed of (i)measuring apparatus model, FEN analyses are carried out under where fvelve press' re sensors are buried in the several pathological condition. The re- trial of tibial component and (ii)graphic display tc visualize the changes in pressure distribution. sults are compared with that in normal In vitro experiments Shcwsthat only a limited knee. The results are compared with that area of the joint.gas subject to the loading in the malaligned knoe. When the quadriceps force in pathological knee. The results of com- was applied, tha pressure force of tilt. femoro- puting show that it is influential on tibial joint varied according to the magnitude of the for as well as the flexion angle. stress field in the knee after meniscectopy.

232 212 1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas Biomedical Engineering Scientific Papers

BE5-F.4 BE5-F.5 The Biomechanics of the Intracranial Transient Pressure Changes in the Spinal Pressure Response to Closed Dynamic Canal under Rapid Headmovements. M Svensson*, Impact Injury to the Spinal Cord, B Aldman, P L6vsund, T Seeman, T Ortengren. Maurice S. Albin*, Leon Bunegin, Chalmers Univ of Techn. S-412 96 G6teborg University of Texas Heal.h Science Sweden Center, San Antonio, Texas 78284 After whiplash injuries in rear end car collisions at low impact velocities, car occupants often cord and In severe trauma, combined spinal suffer from pain in the neck and pricking fingers. injury is notunusual. craniocerebral They also often lose consciousness immediately studiesonevaluationofforce Previous after the impact. Mostly neither injuries to the fromsite of spine/cord injury propagation tissues nor signs of impact to the head can be fluid pressure to brainvia/cerebrospinal diagnosed. The injury mechanism has not yet been waves has been intrinsically limited because given a satisfactory explanation. directly on impact injury was accomplished The spinal canal changes its length when the head exposed spinal cord. Thus, such energy the is pivoted in the sagittal plane. Assuming the attenuating structures such as skin, muscle, cross-sectional area of the canal constant, an and ligaments were removed due to the bone, exchange of fluid between the spinal canal and the This problem was laminectomy itself. ambient tissues must take place by an alteration obviated using a dynamic, closed, impact in- of either volume of cerebro spinal fluid in the instrumentdelivering 1.0 x 107 jury meningeal tubes or blood volume in the veinplexa dynes/kg force to the Tio vertebral of the epidural space or both. Measurements of interspace in the cat. A fast rise time fluid pressure in the spinal canal in both anes- semiconductor transducer was connected to an thetizied pigs and in a mechanical model show intracranial subarachnoid bolt andimpact marked differential pressure gradients in the force calculated using an optical-electronic canal under simulated whiplash. This is prolably detector. CSFwaves weredigitizedand a result of the expedited fluid flow. averaged after subtraction of impact We are now studying the detailed effects of these artifact. After impact, the CSF pressure pressure-flow phenomena to be able to understand wave in the cranium reached an amplitude of 30 mmHg over a period of 8.0 msecwith the the injury mechanisms. pressure peaking 3.8 msec after impact.

BE6-A.1 BE5-F.6 ventilatory fluctuations and dynamics Biomechanical Research of Function Shelves in Treat- of the respiratory chemical control ment of Articular Fracture Wan Xiao-tone, Gu system, A. M. ElHefnawy: Ph.D, Yun-wu, Han Hui, Tianjin Hospital Orthopaedic Inst. Biomedical Engineering Dept, Cairo Articolar fracture is common in patient with frac- University, Giza, Egypt. ture and difficult in treating. These external fixations have been applied to clinic and biomechanic Ventilatory fluctuationsarestudied by characteristics have been concluded. The external incorporating noiseinto amathematical fixations were made of small bulk, less weight and model of the human respiratory chemical small structure. After reduction, the fixations control system. Uncorrelated Noiseis were fixed on part of fractures and acted on a re- imposed on various system parameters duction force. The strain forces for reduction, including metabolic rates, receptors, and which were used to treat humeral transcondy and Ben- blood flows. The resulting fluctuations nett's fracture, were about 4-10kg. The compressed show the differential sensitivity of forces, which used to treat olecranal, patellar and ventilation to the studied parameters. calcaneal fractures, were about 4-27kg.The static In addition, the central component of moment was about 9.5kg-cm for treating olecranal ventilation is negatively correlated with fracture, about 3kg-cm for treating Bennett's frac- the peripheral component, resulting in a ture, about 18kg-cm on calcaneal fracture, about total ventilation which is less variable 7.4kg-cm on humeral transcondy fracture, about 13.1 than the algebraic sum of its components. kg-cm on patellar fracture when the patient stand- Spectral analysis of the fluctuating ven- ing. These external fixations were made of the good tilation shows a peak around 0.025 Hz, elastic materials, so the elastic recovering forces mainly due to the peripheral component of were very well. During patient's exercise, the ventilation. This is the same frequency strain and compressed forces float up and down about encountered by increasing controller gain 1-3kg. the moments float between 21-20kg-cm. The to destabilize ventilation. The spectral early function exercises were good for heal of frac- peak is diminished in both cases by in- ture and could prevent stiff joint. During exercis- creasing the level of inspired CO2. The ing, the static strain force float between +/-0.43kg, _results support -that respiratory -oscilla- -and-stati.c-compressed-force-between-T/-T.3kg. After tions might be due the system dynamics. 3-4 days, the static strain and compressed forces went down to 50-70% but the float forces went up one to three times as much as float forces of beginning. 233 1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas 213 Biomedical Engineering Scientific Papers

BE6-A.2 BE6-A.3 DEPENDENCE OF UPPER AIRWAY CALIBER ON Sleep-Induced Effects on Chemoreceptor INSPIRATORY MUSCLE ACTIVITY DURING Responsiveness and Its Implications, SLEEP, D.W. Hudgel*, Case Western J.A. Dempsey*, University of Wisconsin Medical Reserve University, Cleveland School, Madison, 53705. Metropolitan General Hosoital, Cleveland, Ohio, U.S.A. During NREM sleep the ventilatory response to It is hypothesized that chest wall muscle hypercapnia or hypoxia remains substantial, while contraction during inspiration, by producing an small reductions in arterial P strongly intraluminal p22 subatmospheric pressure, has a J:hibit breathing, so that at = 40 mmHg co2 collapsing effect on the compliant segments of expiratory time increases 2-3 X and at Pa the upper airway in man during 02 sleep. This 35 mmHg Tincreases 10-15 X. This combination collapsing effect E is counterbalanced by a of sleep effects on chemo responsiveness has the dilating force generatedby contraction of following effects: a) causes periodic breathing muscles r---7;unding the upper airway. To in hypoxia; b) causes augmentation of diaphragm examine t- s hypothesis, we recorded upper and accessory muscles of respiration--- including airway and chest wall muscle inspiratory tonic abdominal expiratory muscles--when airway and phasic electrical activity and upper airway resistance is high during eupneic breathing; c) resistance during periodic breathing in healthy permits complete "resting" of inspiratory and subjects and obstructive sleep apnea patients in expiratory respiratory muscle activity when Pac02 NREM sleep. A disproportionate decrease in is reduced 2-3 mmHg, (via nasal mechanical upperairway relative to chest wallmuscle ventilation); d) increases upper airway inspiratory activity resulted in an increase in resistance--in "susceptible" subjects--when Pa upper airway resistance or occlusion. In is low (such as during unstable hyperpnea) and contrast, an increase in upper airway muscle decreases upper airway resistance when Pco, is inspiratory activity out of proportion to chest high and P02 low (such as at end-apnea during wall muscle activity caused a decrease inupper periodic breathing). The site of transient airway resistance, or resulted in airway opening hypocapnic inhibition and hypercapnic excitation from an apnea. These findings demonstrate the resides primarily at the carotid chemoreceptors. importance of the relative inspiratory (Supported by NHLBI). activities of upper airway and chest wall muscle contraction during inspiration in settingupper airway caliber during sleep.

BE6-A.4 BE6-A.5 Periodic Breathing: A Model with Experimental Spectral and Coherence Analysis Applied to Sleep Observations, D.C. Shannon, D.W. Carley and Apnea, D.A. Silage* and A.I. Pack, Department of Ch. Maayan, Massachusetts General Hospital, Boston Electrical Engineering, Temple University and the Massachusetts 02114 Department of Medicine, University of Pennsyl- vania, Philadelphia, PA, USA. A minimal model of periodic breathing was developed. It defines the conditions for PB in terms Spectral analysis by digital comb filtering of of: 1) COsensitivity, 2) cardiac output, ventilation and the coherence between ventilation 2 3) mixed venous CO 4) circulation time, and 5) and the frequency content of the electroencepha- 2' mean lung volume for CO2. In each of 38 healthy logram (EEG) are used to describe the hypothesis subjects, awake (15), asleep (23), 8 with familial that there are temporal relationships between the dysautonomia and 2 lacking CO2 sensitivity, we periodicities in ventilation and oscillations in studied the relationship between relative stabil- state affecting the "wakefulness: input to the ity (R) and loop gain (LG) of the CO2 control ventilatory control systems. The analysis system. LG was estimated using measured values of indicates that thdre is no appreciable lag the 5 model parameters.We tested the hypothesis between the oscillations and a significant (p.05) that LG deLermined R.We found a positive correl- coherence (0.8) at a frequency of approximately ation betweenRand LG (r = 0.74, p, .01, N = 85) 0.012 Hz or a period of 83 seconds. This result and between measured R and predicted R; (r = 0.09, is simila- to the observed cycle time of venti- p = >..1 serial correlation of residuals); the latory periodicities during Stage 1-2 sleep in mean residual was not different from zero (p our elderly subjects. The long cycle time .1). Sustained PB occurred at LG >1 and never at suggests that they are not produced by the carotid LG., 0.7. The average duration of a PB cycle was body and in our studies the oscillations in venti- 29.0 sec; predicted = 36.2s (r = 0.57, p > 0.5 lation persisted during oxygen breathing, although comparison of slope with unity). Results were apnea was abolished. similar awake and asleep. Given the above 5 parameters LG was estimated and this model predicted relative stability and, cycle duration of PBin_healthy subjects awake and asleep, in -familial dysautonomia and in congenital central %),poventilatioa. 34

214 1988 World Congress on Medical Physics and BiomedicalEngineeringSan Antonio, Texas Biomedical Engineering Scientific Papers

BE6 -B.1 BE6-B. 2 ANALYSIS OF THE RESPONSES TO ABRUPT Quantitative Analysis of Upper Airway and SPONTANEOUS VENTILATORY DISTURBANCES, Respiratory Muscle Electromyograms and M.C.K. Khoo*, Univ. of Southern Electroencephalogram during Sleep, E. Cnal*, California, Los Angeles, CA 90089. D. Carley, R.M. Aronson and M. Lopata, University of Illinois at Chicago, USA Deep breaths that abruptly raise ventilation (hyperventilatory sighs) or spontaneous pauses In studying the effects of sleep on respiration and the sleep- of a few seconds duration represent approximate induced disorders of respiration, quantification of upper impulse disturbances to the respiratory control airway and respiratory muscle activity provides important system. By analyzing the ensuing changes in information for the interpretation of sleep-induced changes ventilation and alveolar PCO2, the dynamic in minute ventilation and respiratory system mechanics. properties of the peripheral chemoreflex loop We record genioglossus muscle activity by a bipolar needle may be inferred. Denoting the changes in ventila- electrode inserted perorally and diaphragmatic activity by a tion or alveolar PCO2 as x(t), a large class of gastroesophageal catheter, while the sleep stages are responses may be characterized as x(t) = A exp monitored by electroencephalogram (EEG). (-bt) cos(wt) or x(t) = A exp(-bt) (1 +bt). In The electrical activity of both muscles (EMG) is amplified, this study, we propose the simplest physiologi- band-passed between 30 and 600 Hz, rectified and averaged cally plausible model that can yield the above over consecutive 200 ms intervals, providing a "moving types of responses. We derive analytical expres- time average" measurement of activity. Similarly, to derive sions that relate the best-fit coefficients (A, a continuous "state index" (SI), a 5 s time window is moved b and w) to the parameters of the closed-loop through each EEG waveform in 1.25 s increments and the control system, which include loop gain and the power spectrum is computed for each increment. The total time delay inherent in the loop. This computa- power between .5 and 10 cis is derived from each spectrum tional procedure represents a simple method for to provide the individual points of the SI function. Studies in assessing the degree of stability of the respi- normal subjects with sleep-induced periodic breathing and ratory control system without the need for patients with occlusive sleep apnea reveal a close external interventions, such as inspired gas association between SI and EMG. However, the assumption mixtures or forced ventilatory maneuvers. that state changes are the sole determinant of periodicity, apnea onset and resolution cannot be confirmed. Supported in part by the Whitaker Foundation and NIH Grant RR-01861.

BE6-B.3 BE6-B.4 VARIATIONS IN THE SIGNAL STRENGTH AND UPPER AIRWAY PRESSURE-FLOW RELATIONSHIPS DURING CONSISTENCY OF RESPIRATORY NEURONAL SLEEP: EFFECTS OF CHEMICAL AND MECHANICAL STIMULA- ACTIVITY AND THEIR PHYSIOLOGICAL TION. R.A. Parisi*, N. Krawciw, D.A. Hutt, N.H. SIGNIFICANCE, John Oreml Department of Rdelman and T.V. Santiago. Department of Medicine, Physiology, Texas Tech University Health UMDNJ-Robert Wood Johnson Medical School, New Sciences Center, Lubbock, TX 79430. Brunswick, N.J. Upper airway (ua) narrowing results in activa- Respiratory neurons are, by definition, tion of pharyngeal and laryngeal dilator muscles active during a particular phase of the via mechano- and chemoreflexes. However, the ex- respiratory cycle. However, the tent to which these reflexes may modify airway relationship between the activity of a resistance to overcome obstruction, especially neuron and some phase of the respiratory during sleep, is unclear. We measured inspiratory cycle is not always obvious, and some airflow (V) and pressure differences (mask to hypo- quantitative techlique is necessary to pharynx to trachea) to assess supraglottic (sg), detect the relationship and to define the laryngeal (lx) and total ua resistance (R) in 4 strength and consistency of it. This goats during wakefulness (W), slow wave (SWS) and presentation will show 1) that the analysis REM sleep. Changes were measured during single- of variance is a powerful tool for detecting breath flow-resistive loading (FRL; 18 cm H20/1/s) respiratory activity; 2) that an effect size and during asphyxic chemical stimulation (CS) pro- statistic can be used to quantify the duced by rebreathing 10% 02 in N2. Rsg and Rlx respiratory signal strength and consistency both increased from W to SWS to REM (p<0.05). of that activity; and 3) that classification During CS in each state, Rsg and Rlx decreased at cf neuronal activity according to its signal comparable V. FRL resulted in a large increase in strength and consistency is important to an Rsg over the range of V achieved, while Rlx did understanding of the effects of sleep on the not change. For all states at any level of CS, respiratory system and of the mechanisms for loaded Rsg was a function of unloaded Rsg (r=0.86). behavioral control of that system. We conclude that partial collapse of the pharynx occurs during FRL. Thus, defense of the upper airway by dilator muscles is not sufficient to overcome pharyngeal narrowing. This is more critical during sleep due to the greater baseline Rsg. (Support: HL16022, HL07467, HL01549, UMDNJ #20-87) 235 1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas 215 Biomedical Engineering Scientific Papers

BE6-B.5 BE6-B. 6 Acoustic and Magnetic Resonance Imaging of Mathematical Analysis of Periodic Breathing Extrathoracic Airways, Kingman P. Strohl*, in NREM Sleep, N.S. Cherniack* and G.S. Peter L. Hoekje, E. Mark Haacke, and Lee J. Longobardo, Case Western Reserve University, Brooks, Case Western Reserve University, Cleveland, Ohio Cleveland, OH 44106 We used a mathematical model of the respiratorycontrol We compared two noninvasive methods for measure- system to systematically investigate the effect of severity of ment of extrathoracic airway size and response to disturbances to breathing and systemparameters on periodic positive transmural pressure. By acoustic imaging breathing during sleep. The model consisted ofmulti- (AI) with airborne sound introduced through the compartment representation of 02 and CO2 stores, a peri- mouth, a one - dimensional measure of airway cross- pheral controller sensitive to 02 and CO2, anda central section over the travel path of the sound is controller sensitive to CO2. The effects of hypoxiaand obtained 5 times per second, tracking dynamic hypercapnia on the upper airway muscleswere not changes in area. Magnetic resonance imaging (MRI) considered. An episode of hyperventilationor asphyxia was provides information in two and three dimensions used to disturb the control system. Simulations withthe about the composition and configuration of the model produced the following findings: the number ofapneas tissues surrounding the airway, but at the pre- associated with periodic breathingwere greater as circu- sent time requires imaging times longer by 2 to 4 lation time increased; controller gain increases alsomade orders of magnitude than does AI. Images were the number of apneas greater, but at each level ofcircu- obtained by both methods in six supine normal lation time there was a range of controller gainchanges subjects breathing through a mouthpiece, with and which caused little change in apnea number;there were without administration of continuous positive more apneas with hypoxia; the number of apneas increased airway pressure (5-10 cmH2O). Both methods with sleep associated reductions in metabolicrate and the provided similar values for airway size along the more rapidly resting pCO2 rose at sleep onset; and themore extrathoracic airway, and identified similar intense the disturbance, the moreapneas there were. regional variations in static airway size in It is known that hypoxia and hypercapnia havemultiplicative response to an increase in airway pressure. With effects on controlling gains. The precise form ofthese AI, changes in airway size were observed to cor- multiplicative interactions had important effectson the relate with inspiration and expiration, while MRI occurrence of periodic breathing. revealed specific anatomic de-tails about the The results suggest that the occurrence of periodic yieldl:ig and unfolding of the airway wall in breathing depends on a number of different factorsincluding response to static pressure changes. system parameters and disturbance intensity.

BE6-C.J. BE6-C.2 BLUED BALANCEAND TRAMEOFTINIIBE Transcapillary Transport Pathways LUNG. Robert J. Roselli: Departmentof in Skeletal Muscle, P.D. Watson* Biomedical Engineering, Vanderbilt University. and H.B. Wolf, Dept. of Physiology, Nashville, ZN 37235. 0 University of South Carolina, A multiple pore model of steady-state fluid Columbia, SC 29208, USA. and macrcmplecular transport has been applied to measurenents of vascular pressures, lymph flow, lymph and plasma concentrations of several In the conventional view of transcapillary proteins, and the permeability surfacearea for movement, water and solute cross by the urea in sheep lungs.Two pore sizes (approximate- extracellular route. Both pore and fiber ly 5nmand 26nm, in proportion 10,000:1)are matrix models of pathways assume this. Using sufficient to describe lymph data collected during osmotic pressure of small molecules (urea, baseline conditions and following microvascular NaC1, sucrose, and raffinose) to drive water pressure elevations, saline infusion, and plasma across the mierovessels of the isolated cat infusion. Atwocanpartrnentplasma - interstitial hindlimb perfused at high flows (>100 ml.min- nbdel which uses the above pore structure has also 1.100g muscle-1), water flowrates (>0.005 been developed to describe unsteady-state fluid ml.min-1.mmHg-1.100g muscle-1)were obtained and maw:molecular exchange in the lung. Add- which were comparable to flows driven by itional model parameters, such as interstitial hydrostaticpressure orlarge molecule canpliance and filtration coefficient of the osmotic pressure (0.01). Using these data to lymphatic barrier, are estimated using lymph data calculate reflection coefficients for these collected after a step increase in vascular molecules gave values >0.5, close to that for pressure. Model predictions indicate that onlya serum abumin, 0.8. These data cannot be small fraction of the interstitium is accessible easily fitted by models where water and to exchanging protein - a phenomenon verifiedby solute use thesame path,and are best experiment. Thismodelis currently being used to explained if most of the dater flow passed interpret vascular permeability fromtracer through the endothelial cells rather than experiments where radioactive macromoleculesare around them. Supported by USPHS Grant Hi. injected intravascularly and transvascular 24314 and the South Carolina Heart Assoc. exchange measured with external detectors. (Supported by NHLBI 19153, NIB BRSG S07-RR07201) 236

216 1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas Biomedical Engineering Scientific Papers

BE6-C.3 BE6-C.4 AND pH PHYSIOLOGICAL REGULATION OF CAPILLARY KINETICS OF CAPILLARY CO2 EXCHANGE PERMEABILITY, V.H. Huxley*, Dept. Physiology, EQUILIBRATION, Akhil Bidani: Department of Univ. Missouri Medical School, Columbia, MO, Medicine, UniversityofTexas Medical Branch, USA Galveston, TX. 77550, USA

The study of transcapillary water and solute In vivo kinetics of capillary CO2exchange and has concentrated on factors influencing intravascular pH equilibration are determined by hydrostatic and osmotic forces at the the complex interaction ofmultiple factors. capillary wall. For the most partthe The most important of these are (a) the rate of exchange vessel wallis viewed as a passive CO2 hydrat:,n-dehydration reactions within barrier which may be disrupted under plasma and in red cells, and (b) the kinetics of pathological states. Recent studies of water red cell anion exchange. The role and influence and solute flux in individually perfused of thesefactors has been examinedusing a microvessels suggest a role for active detailed mathematical model of the reaction and modulation of the capillary barrier. transport processes that occur in blood during Specific examples, from this laboratory, and after capillary gas exchange. Furthermore, include conditions of hypoxia and anoxia or this model has been used to analyze and explain infusion of atrial peptide or sodium experimental data obtainedby us in intact nitroprusside. In these cases weobserve animals andin isolated perfused animal lungs. changes in hydraulic conductance and/or Thesestudies indicatethat bloodmight not permeability of the capillary wall to reach electrochemical equilibrium during solute. Theresponses are specific to the capillary transit. Additionally, several treatment, dose and duration. They can be commonly used drugs adversely affect the rate of modified or inhibited by a variety of capillaryCO2 exchangeand intravascular pH agents. The mechanisms whereby the exchange equilibration, by modifying the kinetics of barrier propertiesare altered appear to be specific reactions and/or transport processes regulated by the endothelial cells in consort involved in gas exchange. Such detrimental with at least one other cell type (pericyte, effects on gas exchange and acid-base nerve or mast cell?) under normal homeostasismight be particularly important in conditions. patients with limited cardiopulmonary reserve. VHH is an AHA Established Investigator

BE6-D.1. BE6-13.2 INCLUSION OF AIROY WALLS IN LUMPED ELEMENT MODELS Modeling the Low-FrequencyMechanical Behaviour of OF THE RESPIRATORY SYSTEM. A. C. Jackson: Dept. the Respiratory System, J.H.T. Bates*, McGill Biomed. Eng., Boston Univ.. Boston, MA. University, Montreal, Quebec, Canada. A 6-element model consisting of a central branch (with a resistance and inertance) separated The mechanical properties of the respiratory from a peripheral branch (with a resistance, system are conveniently assessed in terms of the inertance, and compliance) by a shunt compliance relationships between three easily measured (C ), thought to represent alveolar gas quantities- pressure and flow atthe airway compression, has been used for respiratory system opening, and pressure in the esophagus. However, impedances (Zrs). This model was fit to Zrs in therelationshipsbetween these quantitiesare nine dogs (2-96 Hz) but Co estimated using the dependenton the frequency of ventilation. It is model was not correlated Uith C. computed from clearlyimportant to understand the mechanical independent measurements of thofecic gas volume. behaviour of the respiratory system at naturally One reason for this may be because of the occurring frequencies, which are virtually always distorting effects of airway wall shunt lessthan two Hertz.We consider data from the impedance, Zwsli To test this,two methods of following three types of experiments on compensating Zrs for the effects of Zw of the anesthetized paralyzed dogs, allyieldinglow- upper airways were investigated. Both methods frequency (0-2 Hertz) information: a) passive rely on a measure of impedance with the airway expiration, b) the cyclical ventilation of the occluded at some point. This provides a separate lungs, and c) the sudden interruption of flow at measure of ;gall which is assumed to be a shunt theairway opening during passive expiration. It impedance in parallel with the impedance is apparent that both the lungs and the chest wall peripheral to the closure. Zser.In each of the ofa normal animal are well modeled as systems dogs. Zrs was measured (2-56 Hz). the tracheas with two mechanical degrees offreedom. Further were occluded jest above the carina and Zwall was experiments in whichalveolar pressures are computed by both techniques flow then measured. Zper measured directly in open-chested dogs during were physicallyunrealistic in that the real part interruption demonstrate that one of these degrees became negative above 56 Hz. It is concluded that offreedom in the lungs is the inflation of the neither of these techniques can be used to account lungs as a unit, while the other is accounted for for the shunt impedance of the tracheal wall. by theviscoelasticpropertiesof the lung Supported by NIB grant HL-31248. tissues. A similar model structure is invoked for the tissues of the chest wall. 237 1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas 217 Biomedical Engineering Scientific Papers

BE6-D.3 BE6-D.4 WHAT MEANINGCAN BE GIVEN TO RESPIRATORY MECHANICS PARAMETERS WHEN ESTIMATED THROUGH ISSUES WHEN FITTING LUMPED MODELS TO RESPIRATORY CURRENTLY USED MODELS? IMPEDANCE DATA. K.R. LutchenBoston University, Boston, MA 02215, P. Barbini*, G. Cevenini, University of Siena,Italy G. Avanzolini ,A.Cappello,University of Bologna,Italy Many studies have tried to estimate respiratory system propertiesby fitting lumped models to impedance data. A nine-element RLC model of the respiratory system. Herewereview she consequences and issues which takes into account both low and high with this approach. The frequency behavior, is used as a quantitative overall theme is the trade-off between model tool for investigating themeaning of parameters identifiability and beingable to maintain a identified in artificial ventilation and forced modelstructure withphysiologicaly pressure. oscillations. This model, characterized relevant by a sews airway and two parallel tissue compart- parameters. We examine how, for a given clinical ments. snables us to generate realistic transpulmo- condition, data from a specific frequency range narypressures by using mouth flow signals typical limits the complexity of proposed models. These of these two respiratory conditions and parameter models, in turn, inherently values representative of dogs. limit the ability to Two and four lumped parameter models are used for resolve specific system properties or features. estimation from low frequency signals, while a six- To demonstrate this weuse "forward-inverse" element model is applied in high frequency condi- modeling. Several kinds ofdetailed forward tions. Parameter estimation procedure is performed in the time-domain and the Powell algorithm is used mcdels (FMs)` are used to generate data from 0-4 to minimize the root-mean-square error between Hz, 4-128 Hz, 0-128 Hz, and 0-200 Hz. The FMscan simulated and model-predicted pressure. have a structure to purposely distinguish At very low frequencies (less than 1.5 Hz) the anatomic regions (eg., a common airway leading two-elementmodel is preferred and its compliance to reproduces the total one, while the resistance as- parallel pulmonary pathways then a common pleural sumes an intermediate value betweei, the two tissue node and then parallel chest wall pathways). For resistances. When the input signal spectrum reaches other FMs, theelements can be systematically 3 Hz, the four-element model is preferred. All parameters are satisfactorily estimated even if the altered to indicate specific types and amounts of resistance in series with the lower lung compliance disease (eg. inhomogenous airway disease only). is overestimated. At high frequencies (4-200Hz) the Then, the limitations of simpler inverse models six-element model allows a good estimation ofair- required for uniquely way parameters. The tissue resistance and inertance fitting these data in the are the parallel of the respective elements in the presence ofmeasurement noiseare explored. We two tissue compartments,while the tissue compliance also examine constrainingsome parameters based assumes an intermediate value.In conclusion,differ on independentmeasurements. Supported ent signals require different models which by NIH provide Grant 31248 and The Whitaker Foundation. complementary information on respiratory mechanics.

BE6-E.1 BE6-E.2 OneandTwoDimensionalWave Anesthesia Ventilator Mechanics Models for Oscillating T. Cadena and F. Prieto* Flows in Respiration, M.A.F. Institute Nacional de la Nutricion Epstein*, F. Moslehi, M.A. Pisani and J.R. Ligas, Univer- Universidad Autonoma Metropolitana sity of Connecticut Medical Mexico School, Farmington, Connecticut, U.S.A. A practical, low-cost anesthesia ventilator is one of the notable medical equip- Recent interest in the physical phenomena ment needs in Mexico and probably in other underlying inhomogeneities in gas distribution within the lung has identified developing countries. the need to predict pressure and velocity We propose a pneumatic-power, electronic distributions throughout the airways. Wave control, constant flow ventilator, built mechanicsprovides a potentanalytical with low cost, locally available industrial tool for characterizing such oscillatory pneumatic, measuring and electronic parts. flows. The equations of motion for a gas undergoing small amplitude sinuousoidal Flow rate, respiratory rate, relative ins- oscillations with negligible viscous pira_ory and pause periods and end expira- effects were expressed as a one dimensio- tory pressure are set by the user. The nal waveequation. This model is capable pneumatic circuit provides the pressure to of describing the velocity andpressure drive an anesthesia bellows. fields (transient and Gteady state) in a number ofgeometries including slightly The ventilator has been put to test during compliant ducts and horns. A two dimensio- animal experimental surgery and has proved nal model which includes viscous safe and effective in a variety of surgical dissipation has been solved analytically procedures. to predict the time dependent radial and axial velocity and pressure distributions. This approach has great promise for understanding airway flow fields. 238

218 1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas Biomedical Engineering Scientific Papers

BE6-E.3 BE6-E.4 Alicrocantroller far a Valuzetric Respiratcr Respiratory Drive Monitor A. Martinez M.*. C. I..bpez A., J.Ji lignIeZA. Integrated in a Ventilator, thiversiden Autberma Metropolitana - Izt. L. G. Hellstrtim*, B. Hallen, Department° de Ingeni er I a Elbctr ca and D. Linnarsson, Av. Purisima V Michoacan. tib3rico, 09340 DF Karolinska Institute, Box 60400, 104 01 Stockholm, We havedesigned a microcontroller-based Sweden system for the display aria cxrrtrolof all the functions involvedin a volumetric ventilator designed by Instituto Nacionalde la Nutricibn A criticalproblem inintensive care incollaboration with IJAM-Iztapalapa. It was and anesthesia is to decide when the designed around.the8031 microcxntroller,and patient can take over his own breathing auxiliary chips such as the 8254program:able from aventilator. For clinical appli- timer and the AD00809 8charnel.8 bit A/D cation, we have integrated a system for carves-ter. non-invasive determinationof respiratory driveinto a Siemens-Elema Servo The systemis divided.into tlree parts: Ventilator 900C. The system determines display, monitoring anal control, and alarms. The the respiratory drive by evaluating the mcnitoring and display subsystems are based pressure drop during the first 0.1 sec openthecoupling of floe andpressure of an occluded inspiration P0.1). This transducersto the A/D converter,ern the can bemonitored, togetherwith other measuring of02 concentration. Different relevant respiratoryparameters, on a parameters such as respiratory frequency. tidal breath bybreath basis. No additional volume and inspiration/expiration ratio can be valves or sensors are required. A small displayedacidcontrolled.Several alarms are personal computerconnected to a Servo int:Iv:led, such as apnea, airway pressure, minute Computer Modulecontrols the measuring wirme. absence. of electrical energy anfl absence process, presents the result and stores of oases;default volutes for operation are it on file. Respiratory occlusion pres- inclule:l to enable the system to assure positive sures agreewell with simultaneous va- safety tarter all normal and -failure conditions. lues obtained with a standard method.

BE6-F.1 BE6-F.2 Propagation and Damping of Flow Disturbances in the Axial FlowDevelopment during,High Frequency Bronchial Tree in High-Frequency Ventilation. P.Niederer: Sinusoidal Oscillations in Airways. D. Isabey, E.Bush, Inst. Jf Biomedical Eng. and Med.Informatics, University finite de Physiologic Respiratoire IHSERM 0 296, of Zurich and Swiss Federal Inst. of Technology, Switzerland; Hopi tal Henri liondor, Creteil, France. D.Spahn, E.Schmid, Inst. of Anesthesiology, University Hospital, Zurich, Switzerland; P.P. Lunkenheimer, Dept.of Exp.Surgery, From detailed measurementsofsinusoidalflow University of Munster, FRG. resistance in a cast of central human airways,He partitionedfrequency( f) tidalvolume ( VT) Unstable flow processes such as turbulence production or eddy- diagrams for peak flowandfloesreversalinto shedding caused by turbulent flow conditions, bifurcations, abrupt separate regions in which the level s of diameter changes, and entrance effects in general have been unsteadiness are predictible ( JAP 61( 1):113-126, suggested to be mechanisms for gas exchange enhancement in 1986). The dependence of resistance on f and VT in High-Frequency Ventilation. The question arises whether such each region Has character zed by non-dimensional flow uisturbances actually reach the periphery of the lung in spite parameters describing the relative effect of the of the damping provided by the viscosity of the air and the forces governing the floes motion in each.One of viscoelasticity of the lung tissue. them is an al ternativeStrouhal number Propagation and damping of flow disturbances were modelled ( 4L/d)( cc2/Re). coswt,varying throughout the floes mathematically onthebasisofaone-dimensionalflow cycle which seems to be the key to understanding approximation taking into account the above mentioned damping floes resistance in arange offandVT commonly mechanism. The results were compared to measurements made on used during high frequency ventilation.He propose tubes of various lengths, diameter and quality, as well as on dried extending the Blasi us solution for Flom pig lungs. The dependence of the damping is demonstrated as development along a flatplatetotheboundary function of frequency and tube diameter, and the ratio of viscous layer equation Kith a non-steady term.This showed to viscoelastic damping isdiscussed. It is found that high- that 6 always remains proportional to the frequency flow disturbances are strongly damped out in the small instantaneous and local ratio of the unsteady tetm terminal airways, in spite of the short propagation distances over the convective tcrm confirmingthe perti nence associated with the peripheral lung structures. of this parameter for oscillatory flow.when this parameter is appliedtoaxialvelocity profiles previously measured during sinusoidalflowin a 3: 1 model ofthe human centralairway:;, Lila velocityprofile dependence- on thediffer nt 239 level s of unsteadiness is confirmed.

1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas 219 Biomedical Engineering Scientific Papers

INE6-Fa BE6-F.4 A REVIEW OF GAB TRANSPORT PROCESSES IN HIGH Computer Simulation of CO2 Gas Exchange FREQUENCY VENTILATION R.D. Eamml Department During Mechapical Ventilation of Neohanioal Engineering,. Hassaohusetts P.W. Scherer, G.N. Neufeld, Institute ofTeohnology, Cambridge, Massa- J.D. Schwardi, S.J. Aukburg ohusetts 02139, USA. Department of Bioengineering and Department of Anesthesia, University of Pennsylvania An overview of thedifferent physioal Philadelphia, Pennsylvania 19104 meohanismsresponsible for gas transport USA during highfrequenoy ventilation will be presented and oast ina unified and sim- Conventional mechanical ventilation differs from plified form that more olearly identifies spontaneous breathing mainly in the flow waveform produced by those faotors that limit the rate of gas tha ventilator and the manner in which the lung is expanded by transport. One simplifioation introduoed positive pressure. We have developed a computer model of simulatesthe enhanoement ofmoleoular steady state CO2 gas exchange based on the solution of the one diffusion due to themovement of gas baok- dimensional convective diffusion equation In the Weibel and-forth through regions of varying oross- symmetric trumpet bell model of the human bronchial tree. The rsotional area. Prediotions are made of computer model allows us to determine the position of the tidal the looal resistanoe to gas transport asso- volume (VT)-FRC Interface as it moves during Inspiration and oiatedwithdifferentpulmonary airways. expiration and to determine the amount of CO2 which has been Two resistive zonesare identified; one in transferred into the VT from the FRC by molecular diffusion. By the vioinity of the traoheaand another using the computer model, we can investigate the effect of many more peripherally inairways having diame- system variables including: acinar airway structure, anatomic ters in the range of 1.0 to 2.6 mm. This dead space, breath flow waveform, breathing frequency, tidal suggests that new efforts should be devoted volume, breath hold duration, and acinar blood flow distribution to betterunderstanding thenature of on the amount of CO2 expired per breath and the form of the CO2 transport in these zones and to the devel- washout curve seen at the mouth on expiration. The model opment of variations inthe method of ven- predicts that In the absence of breath holding, frequency and tidal tilation that have the greatest influence volume are much more important than breath waveform shape In there. determining the efficiency of CO2 gas exchange during mechanical ventilation. Supported by the NHLBI, grant HL 33009.

BE6-5.1 BE6-G.2 Connection between Microscopic Structure Finite Element Model of Alveoli andDucts and Macroscopic Mechanical Properties of R. Lowel, D. Hitchings2, F.L. Matthowsl, the Lung, Y.C. Fungl University of Cali- and R.C. Schroters. Physiological Flow fornia, San Diego,La Jolla,CA 92093, Studies Units and Aerona,tics Dept= U.S.A. Imperial College London SW7 2AZ

Knowledge of the mechanical behaviour of To derive the constitutive equations and the lung lE important for understanding material constants of the lung tissue from the its function and monitoring abnormalities. microstructure,the widths and curvatures of Recent work by 'owe et al(J. Biomechanics the collagen and elastin fibers in the pulmon- 19,541-549,1986) on effects of the ary alveolar wallsand alveolar mouths are constituent tissue and surface lining of measured. Together with rheological data of the alveolar structures has shown that a these materials, the morphometric data are used mathematical model of a single alveolus of tocompute the stressstrain relations of the p.escribed geometry and assumed material alveoli. The geometry of the alveolar ducts is and surface tension properties can be used analyzed and compared with known data. The to yenerate pressure-volumecycles for validated duct model isusedto compute the whole lungs In vItro. Finite element macroscopic stressstrain relationship of the methods were used to determine the lung tissue. In the calculation,the micro- distribution of stress for a system of strain of the duct is allowed to be different forces due to the applied pressure on the from the macroscopic strain. Among other structure. In an e;,tension of this model things, itis shown that a usual hypothesis we aim to reproduce the network of alveoli about soft tissue mechanics: that the elastin surrounding the duct. The chosen shape of fibers are straight andtake up load at low the alveolus lends itself ideally for this stress level, whereas collagen fibers are purpose, enabling s:mplification of the curved initially and takes on an increasing model due to symmetry. By producing an share of load only when an increasing portion accu-ate functional representation of the of them are straightened, is wrong,at least duct ye+ hope to examine the role of for the lung. (Research supported by NIP. NHLBI interdependence of alveoli and theeffect HL 26647.) of non-homogenous alveolar units. 240

220 1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas Biomedical Engineering Scientific Papers

BE6-G.3 BE6-G.4 Mechanical Properties of the Lung Parenchyma Deter- Surface Tension Driven Flows in Thin Liquid mined from 3-D Reconstructions.R.R. Merceeand Linings D.P. GaverIIII and J.B. Grotberg, Evanston, IL. J.D. Crapo, Duke University, Durham, N.C. Northwestern Univ. The surface tension on the thin walled alveolar We study the events following the impaction septa and the collagen-elastin connective tissue of an aerosol droplet containing an immiscible network within the septa are major factors in pa- surfactant on the lung's liquid lining. Surface renchymal micro-mechanics. 3-D reconstructions tension gradients drive a "Marangoni" flow from serial sections have been used to study the within the film, spreading the aerosol, and mechanics of alveoli and alveolar ducts in lung pa- providing a larger area for the transport of renchyma.The results indicate that connective substances carried by the aerosol. We used tissue elements act to enhance lung stability at lubrication theory to derive non-linear partial volumes greater than end-expiratory volume by dis- differential equations descri..4n; the evolution tributing stress between adjacent regions. Recon- of the surfactant concentration and the film structions demonstrate fine collagen fibers (5-30 thickness, and found their solution by numerical fibrils/fiber) interconnecting the collagen rich integration. The driving mechanisms in this alveolar entrance rings (80-300 fibrils/fiber) of problem are surface tension gradients, surface adjacent alveolar ducts. Elastin fibers are pre- diffusivity of the surfr.ctant, and gravity. dominately located on the edge of alveolar entrance Several dimensionless parameters, which mainly rings and interconnect adjacent rings. This indi- depend upon the thickness and viscosity of the cates that stress distribution in the lungs is due film and the composition of the aerosol, in- to the elastin interconnected system of alveolar fluence the behavior of the solutions. entrance rings and the fine collagen fibers inter- Our results show the possibility of a wide connecting the alveolar entrance rings of adjacent range of behaviors depending upon the magnitudes of the parameters. However, for all cases in alveolar ducts. The elastin interconnected al- whicn the gravitational parameter was non-zero, veolar entrance rings are under tension at all we observed an initial spreading of the aerosol levels of lung inflation while the fine collagen accompanied by a localized thinning of the film. fibers in the alveolar septa are under tension and A recirculation region in the form of a ring distributing stress when the alveoli through which vortex develops and the film finally relaxes to they pass are near full inflation. a flat interface.

BE6-H.1 BE6-H.2 Conceptual Problems in Rib Cage Kinematics. Intrathoracic Pressure Dynamics with Forcing R.C. Saumarez, Dept. Cardiological Sciences, at the Airway Opening and Chest Wall St George's Hospital Medical School, London S.A. Ben-Haim* and G.M. Saidel SW17 ORE, United Kingdom. Technion, Haifa, Israel and Case Western Reserve University, Cleveland, Ohio 44106 The kinematics of the rib cage have proved hard to analyse. These difficulties stem fr'a: We have developed a mechanistic model of ventila- a) Unrecognised functional differences beLueen tory dynamics which accounts for movements of the the human and quadruped rib cages and the func- lung, rib cage, abdominal wall, and diaphragm. - tional and morphological differences between The applications of the model include the analysis the human upper and lower ribs. of positive-pressure ventilation (PPV) and b) The transformation of a rib between twc cardiopulmonary resuscitation (CPR). To determine I.Dints may be expressed as a rotation and a tran- optimal strategies for these therapies, we have slation. The trajectory of the rib cannot be simulated th.: dynamics of the ventilatory system deduced from this transformation yet ribs are with various forcing functions. With respect to often assumed to rotate about a single axis. PPV, we addressed the question "For a given tidal c) The rib cage probably moves with its compon- volume and breathing frequency, is there a pres- -ents fitting together. With monoaxial rotation sure or flow input waveform that produces minimal there are small misfits at the four joint sur- airway and pleural pressures?" The intent is to faces on the rib. While these are unegli6ible" deliver sufficient ventilation, but with minimal compared to the dimensions of the rib, very large pulmonary and cardiovascular side effects. With forces would have to be created to overcome these a variety of waveforms we found that none could misfits. Also, with monoaxial movements, varia- minimize airway and pleural pressures at the same tions in rib cage motion cannot be accounted for time. With respect to CPR, we addressed the except by rib deformation which would require question "Can occlusion of the airway opening very large forces. These problems are resolved during chest-wall compression provide a positive if the rib cage is treated as a multi-element effect on cardiac output by increasing intra- mechanism which can move within the constraints thoracic pressure?" Both model simulations and imposed by the geometry of its joint surfaces. animal experiments showed that airway occlusion during CPR could provide significant improvements.

2421.

1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas 221 Biomedical Engineering Scientific Papers

BE6-H.3 BE6 -H.4 STRUCTURAL MODEL OF THORAX ANDABDOMEN FOR RESPI- RATORY MECHANICS, S.H. Loring, Mechanics of Compartmental Models of the Department of Chest Wall, Theodore A. Wilson, Department Environmental Science and Physiology,Harvard of Aerospace Engineering and Mechanics, School of Public Health, Boston,MA 02115. Respiratory mechanics includes bothpulmonary University of Minnesota, Minneapolis, mechanics and mechanics of the"chest wall", by Minnesota 55455 U.S.A. which we mean the mechanics ofthe thorax, the abdomen and the diaphragm. The geometrical com- The methods of mechanics require thata plexity of the chest wall and its associated linear elastic system with two degrees of muscles of respiration makes rigorous analysis of freedom be described by three compliances. the mechanics of the respiratorysystem difficult. In most of the literature, the compliance We have used a structuralmodel of the that describes the coupling between rib human rib cage and abdomen thatincorporates many geometrical and functional featuresimportant to cage and abdomen is implicitly assumed to the understanding of respiratorymechanics. This be zerc. Deschamps et al. find that in structural model of the respiratorysystem with seated subjects, the coupling compliance two degrees of freedom of motionincorporates is small, but consistently positive, realistic geometries and mechanicalproperties of contrary to current models. intercostal and abdominal musculatureand can predict the actions of respiratory The methods olso provide definitiors of muscles. The the effective forces exerted by rib cage, diaphragm, abdominalwall and lungs are respiratory muscles. characterized by nonlinearstress-strain charac- These forces are proportional to teristics that dependon muscle activity. The the rate of change of muscle length with positions, volumes, forces andpressures for each compartmental volume. A simple model of set of initial conditions are found by determin- the geometry of the diaphragm yieldspre- ing mechanical equilibrium byan energy equation. dictions that are consistent with Hubmayr's Many such determinations allow data on the changes in pleural pressure predict3ons for large displacements. The model predicts passive respiratory and chest wall shape caused by diaphragm system behavior and twitch. the action of individual respiratorymuscles that are qualitatively consistent withphysiological Supported by National Heart, Lung, and data. (Supported by HL33009and HL19170). Blood Institute Grant HL-21584.

BE6-H.5 BE6-H.6 Influence of Body Position on Ribcae Rib Cage Aft !cling from X-Ray Measurements, and Diaphragm Shapes. S. Margulies , C.M.Kenyon', T.Pedley, and T.Higgenbottam, Dept. of J. Rodarte, P. Bergsland, E. Hoffman. Applied Mathematics and Theoretical Physics, Cambridge Mayo Foundation, Rochester, MN, 55903. University, Cambridge CB3 9EW, England. vie evaluated the effect of lung (L) voiume (V) and A continuing problem in respiratory physiology is in relatingmea- dorsoventral pleural pressure gradieni. (Ppl/dx) surements of the chest wall to the underlying anatomical structures. on the shape of the ribcage (R) and diaphragm (D). We present here an adaptive kinematic rib cage model which takes Six anesthetized does were scanned in a 3-D Cf as input the positions of a small number 10) of points on the imaging system at TLC and FRC while prone and thorax measured on lateral and posterior-anterior x-rays, and then supine. After pmdumethorax (PH) was introduced, 5 adapts mathematical representations of the bones and joints to fit. were studied supine and I was also studied prone. In this way the positions and shapes of the manubrium, sternum, Cross-sectional areas of L, mediastinum (M), and vertebrae C7L1, and the po":n of the sikth ribon each side are subdiaphragmatic contents (SD) were detIrmined determined from measuremeu from 1.6 mm transverse slices. .d mathematical models. The shapes of D This technique relies on two hypotheses. and R (R=L+SD+M) were evaluated from the cephalo- 1) that the constraints of a functional rib cage are such that the variation in rib caudal distribution of cross-sectional area cage shape between individuals can be described by a small number (A(z)). In uoth positions, A(z) of R at FRC could of variables. 2) that the bones of the rib cage fit together such that be modelled as they can more with only a small number of degrees of freedom. AFRC(z) = MATLC(z)) The model has been validated using x-ray data from 12pa- tients befor, and after median sternotomy required for where K = (VFRC /VTLC coronary ofR). PN reduced dPpl/dx to artery bypass surgery. Errors were found to be ±5rIm and ±3° zero and thus increased the transdiaphragmatic for dimensions and orientations respectively.Full e..e computer pressure (Pdi) gradient. At FRC PN had no effect plots of the rib cage were compared with actual x-rays tomeasure on D shape while at TLC, the dependent portion of the error, and also where there were discrepancies the shapes the D moved cephalad and the independent region were altered interactively to examine the effecton the numerical predic- moved caudal. These results are consistent with tions. In addition the technique showed itself to be able the presence of a dPpl/dx in the intact dog which to detect the small change in rib cage configuration whichoccurs after the influences the shape of the diaphragm at TLC when operation in conjunction with a decrease in the total lung capacity the average Pdi is small, and has little offect at and residual volume (11 and 0.51 respectively). FRC when average Pdi is comparably large. 242

222 1988 World Congress on Medical Physics and BiomedicalEngineeringSan Antonio, Texas Biomedical Engineering Scientific Papers

BE6-I.2 Energy Dissipation in an Airway Bifurcation Mechanical Effects Induced by Jets in Airways. D. A. Tsuda, B. Savilonis, R. Kamm, and J. Fredberg, Isabey, Unite de Physiologic! Respiratoirc INSERM The Biomech. Inst., Worc. Polytech. Inst., Univ. (U 296), U6pital Henri Hondor, Cretoil, France. Mass. Med. Sch., Mass. Inst. Tech., Harvard Univ., Massachusetts, U.S.A. Tracheal jets have recently been shown to We measured energy dissipation associated with contribute to airway gas transport and are large amplitude periodic flow through airway bifur- frequently used in different types of mechanical cation models. Each model consisted of a single pulmonary ventilation s.,h as High Frequenci Jet asymmetrical bifurcation with each branch terminat- Ventilation or Constant FlowVentilation (CFV). ed into identical elastic loads. Sinusoidal oscil- The behavior of such sets in the airwain resemble% lations were applied at the parent duct such that that of confined sets in a tube.The latter have the upstream Reynolds number (Re) varied from 30 strong mechanicaleffects which have never been to 5000 and the Womersley parameter (a) from 2 deeply investigated nor quantified for the to 30. Pressures were measured continuously at the respiratory system.Ho thus studied the mechanical parent duct and at both terminals, and instanta- effectsof using a newendotracheal tube that neous branch flow rates were calculated. allows gas injection in the lower trachea.Fresh Time-averaged energy dissipation, Ev, in the bifur- gas, which flows through several th capillaries cation was computed from an energy budget over a molded intothe tubewall, is injectedathigh control volume integrated over a cycle. We found velocity near the distal end of the tube, that when flow amplitude was small (tidal volume producing local air entrainment. He applied the to dead space ratio, Vt/Vd<>1), the influence of entrainment mechanism createdeithcr a positive free stream convection was significant and Ev pressure gradient (as occurs during CFV) and/or became quasi-steady as a increased. The behav- an overall entrained flow through the core tube, ior of Ev suggested an analogy to steady curved with apredictible quadratic relationship between tube flow. This study demonstrates that only two airway pressure and injected Clow rate.Lastly we nondimensional parameters, a and Vt/Vd, are present the mechanical and physiological results necessary and sufficient to describe Ev in a given obtained with this system in patients ventilated geometry during sinusoidal flow. Supported by with CFV or assisted by pressure support. HL33009 and HL34616.

3E6-1.3 BE6-I .4 The Expiratory Pressure-Flow Effect of Osculatory Freauencv and Insniratorv:Expiratory Ratio Character of the Lung on Tracheal and Alveolar Pressures in a Rabbit Model, D. Winter,* J.M.Collins, R.D.Kamm and A H.Shapiro SwRI, San Antonio, Texas, D. Gerstmann, A.F. Taylor, K Nose, Massaohusetts Institute of Teohnology J.M. Fouke, R.A. deLemos, SR:JR, San Antonio, Texas. Cambridge, Massaohusetts During high frequency oscillatory ventilation (HFOV) In an Experiments were oonduoted to study the intubated, open-chest rabbit model, we measured pressures In pressure-flow relationshipduring steady the trachea and alveoli using a stainless steel tracheal monitoring expirationin amulti-generation human tube and a lowmass pleural capsule. Ventilator pressures were lung model (J.Hammersleyand D.Olson). A also measured proximal to the 2.5 mm ET tube.Oscillatory subtraotion teohniquewas developed to pressure amplitude (Pamp) and mean airway pressure (Paw) allow measurement of thepressure drop were recorded for the 3 airway locations in each animal for each aoross asingle generationwithin the of 6 randomly chosen Inspiratory:expiratory !:E) ratio (1:2, 1:1) network. Experiments, first oonduoted in and frequency (10, 15, 20 Hz) combinations. Nine animals were a planaridealized airway, oovered the studied using a right thoracotomy to place upper lobe capsules Reynoldsnumbers from 50 to 8000. on 4 and middle lobe capsules on 5 animals. Analysis of variance Sensitivity studieswere then performed tests showed no change In tracheal or alveolar On both lobes) to determine the influenoe of variations Pamp that was related to changes In ventilator frequency or I:E in geometry and flow distribution. ratio. There was, however, a marked Increase In tracheal Paw as The results fora single bifuroation I:E ratio was lengthened from 1:2 to 1:1. Tracheal to ventilator were then used to synthesize the pressure PawratiosIncreasedfrom0.795+0.132to1.163+0.133 -flowoharaoter of aoomplete airway (rnean+SD) (p=0.0001, West) In the middle lobe capsule group network. The numerioal prediotions are in when I:E was Increased.Similarly, In the upper lobe group, olose agreement withpublished data, and tracheal to ventilator Paw ratios Increased from 0.731+0.111 to demonstrate that branohingasymmetries 1.158+0.265 (p =0.0004) as I:E increased. We speculate that enhanoe the oontribution of the lower higher tracheal Paw atI:E1:1could result from unequal airways to the total flow resistanoe, proximal /distal airway Inspiratory-expiratory Impedances, partioularly at lowflowrates and with resulting In volume redistribution to the large always. Casts of more visoous gases. the large airways showed that the upper lobe airway branch had a much smaller radius of curvature than the middle lobe branch Supported by HL-33009 (approximately 3 mm vs 15 mm) and a much larger turning angle (approximately 150° vs 500). The Increased pressure drop In the up2r4rss the middle lobe may be a result of this geometry.

1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas 223 Biompdical Engineering Scientific Papers

BE6-I.5 Br6-T.6 Theoretical Considerations of Nasal AIRFLOW DYNAMICS IN THE CONDUCTING AIRWAYS OF THE Airflow Mechanics, G.O'Neill and N.S.Tolley, LUNG: S.K. KESAVAN, AND N.P. REDDY* BIOMED. University Hospital of Vales, Cardiff, ENGR. DEPT., UNIV. OF AKRON AKRON, OHIO 44325 U.K.

Three-dimensional air velocity patterns were A mathematical model relating nasal airflow measured in a 2:1 plexiglas model of the conduct- and pressure, using Bernoulli's equation ing airways of the human lung using a hot-film applied to flow through the nasal valve anemometer.A plyisobutylene solution in region, has been shown to give values of decalin was used as a viscoelastic sumulant of nasal resistance which correlate well with tracheal mucus. Both steady and oscillatory flows those measured experimentally. It is were studied. Air flow rates were varied to concluded that the nasal valve is the primary simulate conditions varying from normal to heavy location for nasal resistance and that the breathing. The experimental results show the nasal vault contributes a negligible amount skewed and complex nature of the velocity profiles to the total trans-nasal pressure drop. as air cascades through the branches. Model Flow limitation is predicted in the model by geometry was found to exert a greater influence incorporation of a term describing nasal than flow Reynolds number on the velocity profiles. valve stiffness. The interrelationship of Results of flow visualization experiments conduct- the alar wing and nasal valve are demonstrated ed to study the interaction of a turbulent and the model is discussed in terms of aiding air flow with a lining of polymer mucus analog the understanding of normal airflow physiology illustrate the formation of interfacial waves, and what implications this may have to nasal and the alteration of the turbulent boundary lay- surgery. er. (Supported by Univ. of Akron Faculty Research Grant).

BE6 -J.1 BE6-J.2 Resmodynawic ModellingofRapid Vascular Occlusions On theUse of ArterialDensity in theLung, JULLinehan*, T.A. Bronikowski., and Fluctuation to Assess the Variation in C.A. Dawson; Marquette University, Medical College the Blood Volumeof thePulmonary of Wisconsin, Zablacki VA Medical Center; Capillaries, J.S.Lee'andL. P. Lee, Milwaukee,Wisconsin University of Virginia, Hemodynamic analyses of the information contained Charlottesville, Virginia, U.S.A. in the transient arterial and venous pressure curves following vascular occlusion hasproveduseful in When the rabbit or dog is ventilated by u attempting to understand the relationship between the respirator, we find that the gravimetric density structure and function of the pulmonary vascular bed. of arterial blood has a component which In an experimental setting wherein the lung flow is fluctuates at the same frequency as the constant, rapid occlusion of the pulmonary vein for a ventilation. We havehypothesized thatthis short time or simultaneous occlusion of both the artery fluctuation is due to the presence of the and vein provides quantifiable pressure data which can Fahraeus effect in the capillary blood flow and be interpretated using models which consider the dis- the ventilatory induced deformation of the tribution of vascular resistance versus vascularcom- pulmonary capillaries. pliance. In particular, the locus of vasomotion with The density fluctuation due to the ventilation respect to the site of major vascular compliance, the and the density dilution due to the injection of microvasculature, is detectable.We have further shown a low density saline bolus into the vena cave art that the pressure of the microvasculature can be readi- measured by a Density Meter. Regardingthe ly determined from the simultaneous occlusion dilution curve as the transfer function of experiment. Recently we have explored the impact of indicator dilution in the central circulation, we fluid inertance on the interpretati,:n of the arterial use the Fourier transform of this transfer pressuredecay curveinilowin; arterial occlusion. function to assess the attenuation and time delay Since the arterial occlusion maneuver can be done inin the transport of sinusoidal concentration vivo by inflation of the balloon of s Swan-Ganz cath- through the central circulation. With this eter, we have extended our models to interpret the assessment and the Fahraeus effect, we determine information content of the restating arterial pressure the density fluctuation in blood leaving the decay curve to provide an uprer bound of the micro - pulmonary capillaries to quantify the volumetric vascular pressure. change of the capillaries induced by the Supported by Grant HL -19298 and the Veterans ventilation. (Supported in part by HL 36285) Administration. 244 224 1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas Biomedical Engineering Scientific Papers

BE6-J.3 BE6-J.4 ModelingthePulmonary Input Impedance Waling the Structure and Blood Circula- Spectrum, C. Lucas*, B. Ha, G. W. Henry, tion in the Lung, M. R. T. Yen and Y. C. J. Ferrier° and B. Wilcox, The UNC School of Fung, Memphis State University, and Univer- Medicine, Chapel Hill, North Carolina, USA. sity of California,San Diego, La Jolla, California, 92093. U.S.A. The simple shape of the pulmonaryinput impedance spectrum has ledtothe development of models of varying complexity which mimic itssalientfeatures, To modelblood flow one hasto modelthe including the frequencyandamplitude of thefirst structure first. Thus one has to validate both modulusminimum,characteristic impedance, and the geometric and theological models of arteri- fluctuationsindicative of a bedwithreflections. al and venous trees, the microvasculature, and Models are used 1)to produce realistic afterloads for the alveoli, alveolar ducts, and airways.New therightventricleand2)tointerpretvascular data have been obtained on the collagen and responses to system perturbations or alterations,e.g., elastin bundles in these structures; and used druginfusions,structuralabnormalities, or surgical to derive theoretical values of the distensi- interventions. Interpretationisoftendifficult; bility of capillary sheets. The theoretical success depends not only on the validity of the model, values are compared with experimental values. but also on 1) the quality of underlyingpressureand Further, new data on the coupling between the flow signals--which aredifficult toobtainin vivo in arterioles and venules with the alveoli and thecardiorespiratorycycle-sensitivepulmonaryar- ductsare obtained.These improvementsare tery, 2) the technique used to calculatethespectrum incorporated into a theory of pulmonary circu- of a system which may be quasi-linear atbest,and lation. The validated theory is used to probe 3) the technique usedtoestimate model parameters the effect of a variety of parameters on given the calculated spectrum. The pressing need for pulmonary pathophysiology. (Supported by NTH improved methodsfordiagnosing pulmonary vascu- NUM grants BL-26647, HL-34440, and HL-07089.) lar impairment, particularlyininfants and children, emphasizestheimportance of continueddevelop- ment of spat Amodels whichareageappropriate and which weparameterssensitivetophysio- logicalchanoesassociatedwithpulmonaryvascular disease.(Supported by HL35389.)

BE6-K.1 BE6-K.2 Clinical and experimental measurements of the VENTILATION STUDIES WITH RADIO-AEROSOLS pulmonary clearance of inhaled Tc-99m-DTPA, G. Coates, M.D., McMaster University P Wollmer*, E Evander, B Schmekel and B Jonson, Hamilton, Ontario, Canada, UN 3Z5 University of Lund and AB Draco, Lund, Sweden. The perfusiajung scan performed with a gamma camera and -macro aggregated albumin (MAA) The transfer of water soluble substances across the produces an image of the regional distribution alveolocapillary barrier can be studied by means of of lung blood flow. The scan is sensitive to a radiolabeiled tracer substance in the form of an changes in regional blood flow but is not specific aerosol and measurement of the pulmonary clearance since both primary defects in perfusion such as of the tracer by external detection. Measurement of pulmonary embolism and defects secondary to a the pulmonary clearance of inhaled Tc-99m-DTPA i3 a regional decrease in ventilation such as asthma simple and non-invasive procedure, which can be or chronic obstructive lung disease all can performed with standard nuclear medicine equpiment. produce similar scan pattern. The ventilation In animal experiments, we have shown that the rate ign produced with inhaled radio-gas such as of pulmonary clearance is related to the functional Xe was introduced to increase the specificity integrity of the surfactant system. There are pro- of the perfusion lung scan for pulmonary embolism. bably also other rate limiting factors, e.g. the There are inherent difficulties in the routine use permeability of the alveolar epithelium. of radio-gas, many of which are overcome by using In normal man, the pulmonary clearance of inhaled radio-aerosols. The ones most commonly oed are Tc-99m-DTPA follows a monoexponential course with a submicronic aerosols of technetium -99m ( mTu) half-time of approximately 70 minutes. In smokers, labelled DMA or sulfide colloid. This paper the clearance curve follows a biexponential course will discuss the use of submicronic aerosols in and the abnormality is related to the smoking patients with suspected pulmonary embolism and habits. In patients with interstitial lung disease, in children with a variety of conditions including the pulmonary clearance of inhaled Tc-99m-DTPA also asthma, inhaled foreign body and congenital lung follows a biexponential course and is related to defects. other indicators of disease activity. This simple technique promises new insight into the pathogenesis of lung diseases and may be useful for diagnosing and monitoring lung diseases. 245 1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas 225 Biomedical Engineering Scientific Papers

BE6-K.3 BE6 -K.4 Aerosol Dispersion Used as a Test of Imaging of Radioaerosols to Measure Pulmonary Function, Mucus Flow within Lung Airways. M. McCawley, NIOSH, Morgantown, WV and H. Abronst West Va. Univ., Morgantown, WV, USA W.Michael Foster, VAMC-Northport & SUM-Stony Brook, NY USA 11794. The dispersion of an aerosol bolus was found to Lung airways of humans have be an effective means of differentiating groups mucus lining that Is continuously transported to the larynx from of smoking (average exposure of 20 pack-years) lung periphery by ciliary activity of epithelial cells and non-smoking subjects (p<0.0001). Their mean of the mucosal surfaces. Mucus i3 differences for standard respiratory function replenished primarily by secretory products of cells indices obtained from spirometry (FVC, FEV,, and interspersed among mucosal cells or located within submucosal FEF25_75%) were smaller and non-significant. glands. Mucus prevents desication and buffers These tests were also done before and after underlying tissue from inhaled irritants; bronchodilation on subjects with a history of convergence of mucus to the larynx not only maintains reactive airways. The changes detected were luminal airway patency for convective consistent with caliber changes in the small flow of gases but provides deliverence from the airways. A gamma distribution model, simulating continued pres- a single mixing chamber (with incomplete mixing) ence of inhalational hazards which have deposited onto airway surfaces. in parallel with a variable by-pass fraction, was used to account for the differences seen in the Noninvasive methodology is utilizedto study mucus various groups that were tested. transport in man. Radioaerosol (Fe203 spheres la- This model beled with 99mtechnetium) allows a parameter to be derived for both series are inhaled into the lung and subsequent clearance of insoluble spheres and parallel ventilation. Measurement of the is Peclet number for the alveoli obtained in this measured by Anger camera as index ofmucus flow. In addition to recording the collective fashion agrees with the value previously function of all airways, this technic provides estimated by other researchers. a measure of the velocity of mucus flow within centralairways. Kin- etics of mucus clearancecan be assessed under control conditions and compared to experimentalper- turbations in healthy and diseasedairways. Effects on mucus flow are described fortherapeuti agents which improve conductance of lunggases, and for noxious gas, ozone,present in pollutant smog.

BE7-A- BE7-A.2 Qualitative Reasoning With Incomplete Knowl- Simulation Methodology in edge in Medicine, B. J. Kuipers-, University of the DEWS-Scheme Environment Texas at Austin, Austin, Texas 78712 USA.

* Incomple:4.1knowledge is a fact of life in medicine the de- Bernard P. Zeiglerand Tag Con Kim* tailed structure of physiological mechanisms involved ina patient's disorder may be only partially understood, and many Dept. of Electrical and parameters of the patient state are unknown or even unmea- Computer Engineering surable in a clinical setting. Model-based reasoning methods such as numerical simulation require specific numerical values University of Arizona for each model variable, and specific curves to be assumedfor Tucson, AZ., 85721 each functional relation, even where this level of precisionis not justified by available knowledge. Qualitative simulation allows qualitative descriptions of We review the concepts of knowledgerepre- the possible behaviors of a mechanism to be derived from in- sentation and modelling and simulation methodology which facilitate complete, qualitative descriptions of its structure and initial computer exploration of alternative artificialworlds, state. Where the information provided is insufficient tospec- such as self-sufficient bio-regenera,-ve ify a unique behavior, the prediction isa branching tree human habitats. possible behaviors. We have recently developed severalnew An object-oriented computerenvironment methods to improve our QSIM algorithm: (1) application of which supports such simulationstudies is higher-order derivatives; (2) use of a global phase-spaceper- briefy described. A simplifiedexample is spective; (3) abstraction by time-scale to decompose complex given to demonstrate thepower of the approach. mechanisms into a hierarchy of simple mechanisms; and (4) techniques for incrementally incorporating quantitativecon- straints into the qualitative model. The goal of this work is to permit model-based diagnostic and therapy-selectionrea- soning in the face of incomplete knowledge.

t 4-4 k la0 226 1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas Biomedical Engineering Scientific Papers

BE7-A.3 BE7-A.4 An AI Approach to Modeling in Biology Qualitative Nets as Components of Brain Models, A. de CallataS7 IBM A.K. Watson Alain Pave.- Laboratoire de Biometrie IEC, La Hulpe, B-1310 Belgium Universite Cl. Bernard, Lyon 1, 69622 Villeurbanne Cedex Francois Rechenmann Laboratoire ARTEMIS/INRIA The studied qualitative nets have one-way connections Universite J. Fourier (Grenoble) - 38402 St Martin d'Heres with all-or-none irreversible synapses (latches).Each To construct or to choose Ind finally use mathematical group of cells works as a recording content addressable models in Biology requires many different skills which memory (RCAM) storing simple production rules. make this approach complex. The french Group EDORA, RCAMs are nodes of a well-formed directed circular concerned by modeling problems in biology, works in the network system (DNPS) solving problems as a logic conception and the development of new softwares to aid program interpreter.In the simulated robots, the fea- such modeling approaches. An AI solution was chosen, ture extractors are sampled to produce the symbols based on an Object Centered representation of knowledge. processed in DNPS. The output symbols are the set- In this framework declarative and procedural knowledge points of servo-mechanisms.Timing is provided by can be represented and handled by an inference engine internal clocks. The invariant repetitive rules learned called SHIRKA. The representation, the actual structure by experience are:Action C changes the state from A of the knowledge base and the basic inference processes to B. The processing phases are:1) the intention B is will be presented and discussed. EDORA approach illus- chosen, e.g. by a perceptron; 2) the DNPS finds which trates the interest of AI in the software design, for the C can change A in B; 3) the set C of muscular com- utilization by non specialists of complex tools at different mands is executed; 4) the evoked salient events enable levels, and also as formalization tool of knowledge. This a well-connected lattice of RCAMs; 5) the rules de- last remark, i.e. AI as modeling framework itself, suggests scribingthe actual episode are addedintothese that it will be complementary or even alternative to mathe- RCAMs. Simulations have shown that learning is fast. matical modeling when itpresents drastic limitations The general system permits chunking, subgoaling, rea- which is often the case in biological sciences. soning by analogy, adaptation, habituation, and pars- ing. DNPSs are mapped to mammalian brains, predicting how biological circuits may work, e.g. latch closure is mapped to dendritic spine resorption.

BE7-B. 1 BE7-13. 2 Fast Computation of Convection-Diffusion-Permeation- Edora Methodology for Biological Systems Modeling Consumption Algorithms for Blood-Tissue Exchanges, J.B. Alain Pave.- Biometrics Lab. UCBL Bassingthwaighte: J. I. S. Chan, and C. Y. Wang, Univ. Wash- F - 69622 Villeurbanne Cedex ington, Seattle, and Michigan State Univ., East Lansing The main goal of the french EDORA project is the The complexity of solving partialdifferential equations for design of a software to aid mathematical modeling in capillary-tissue exchanges in multiceilular organ models is ordi- Biology, based on AI concepts and on high level narily sufficiently handicapping to obviate the use of such graphic capabilities. It led to the development of AI models for standard approaches to analysis. Yet the demands tools,particularlyan inferenceengine SHIRKA, for accurate parameterization are such that the simplest ade- working on anObject-Centeredrei resentation of quate models are of this form. The use of a Lagmngian sliding knowledge. A first prototype of the software enables fluid element approach allows the coupling of diffusion and to choose a model in a model base following different criteria, to estimate parameters from experimental data convection with permeation and transformation, giving compu- and to recognize the mathematical structure of a tational efficiency without sacrifice of accuracy. Linear models model (i.e. is it known in the base ?) ... A short des- gain further efficiency by computing eigenvalues during the ini- cnption the software architecture and of the structu:P tial condition segment of the program. Diffusion is handled by of the knowledge base give ideas for the future compu- computingaxialdiffusiontermsaftersolvingforradial ter aides modeling systems. We note that the necessity exchanges, in effect using local expressions rather than global of the formalization of knowledge has a positive matrix inversions. The results for a capillary-endothelial cell feed-back effect on the domain of expertise itself, e.g. interstitium-parenchymal cell model is an approximate speed up mathematical modeling m biology, which is at least so of about 107 compared to the analytic solution, an expression interesting than the software itself. Finally, we also requiring solving roots, calculating Bessel functions and duuble note that formalisms and adapted reasoning processes and triple convolution integrations.Speed is essential to the developed on AI concepts are themselves good tools optimizing of the fits of such models to data for outflow dilu- for modeling. As mathematical models, particularly in tion or for residue function data acquire.' by external isotope biological sciences, present important limitations, AI formalizations could be a new approach. The extension detection, as with positron tomography or emission computed of modelling is one of the new center of interest of the tomography or reconstructive tomography in general. (Sup- EDORA group. ported by NIH grant HL38736.) 247

1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas 227 Biomedical Engineering Scientific Papers

BE7-B.3 BE7-B.4 Complex Models Development Simulation oj Catheter-Manometer Dynamic Response Methodology in Biomedicine, Dietmar P. F. Moller: OrOgerwerk AG, R.P. van Wijk van Brievingh+ and F. Pasveer, Medical Technology, 0-2400 Lubeck, W.-Germany Delft University of Technology, NL -2600 GA Delft The Netherlands. Modeling, simulation and parameter estimation are of growing importance for the determination of non- Blood-pressure signals from Catheter-Manometer measurable parameters and state variables of dyna- systems are distorted by the system response. mic biomedical systems. The common problems arising Simulation as a transmission-line in the time do- from modeling and simulation in engineering and in main includes the non-characteristic termination biomedical science, especially the possibility of by the transducer and by air-bubbles as causes of solving them by applying the same methods to a wide wave reflection. System parameters are: catheter range of systems research, have improved the coope- compliance, fluid column inertance and flow resis- ration between these different areas. It is of im- tance per unit of length, determining the charac- portance to note that in biomedicine in general teristic impedance Z and the transmission factor y modelbuilding is an essential approach to estimate which are complex functions of frequency. The dis- clinically important but not directly measurable continuities are accounted for by compliances, system parameters and internal system states where determining complex reflection coefficients. This even an essential excitation of the system, as re- model is mathematically described by two hyperbo- quired for efficient parameter estimation, is nor- lic partial differential equations with time and mally not allowed for medical reason. position as independent variables, and appropri- Model development is generally based on the ate initial and boundary conditions. - a-priori knowledge of the system being modeled The equations are solved by the method of charac- (system must be observable) teristics, leading to two ordinary differential - experimental data consisting of measurements equations describing the forward and backward available from patients or from literature. travelling waves. Excitation with signals obtain- From the methodological point of view there are ed from catheter-tip measurements shows good three types of concepts which depend on the agreement between model and reality. The simula- a-priori knowledge based on the knowledge of the tion may also be used for signal distortion com- inputs or the knowledge of the outputs or the know- pensation purposes. ledge of the system state for the decision of the unknowns, which will be discussed in detail.

BE7-B.5 BE7-B.6 Qualitative Simulation of Biomedical Bond Graph based Physiological Models. J. LefEvrem, D. Roucou, B.Debande, Processes - An Aid in Decision G. Ellinghaus and S. Huberty, Making, F.E.Cellier: University of Louvain University, Brussels, Belgium. Arizona, Tucson, AZ 85721, U.S.A.

Physiological models use various formal- Biomedical and clinical processes are frequently characterized isms : energetical ( electric analogs, by relatively large numbers of different variables interacting mechanical and hydraulical networks, with each other in a poorly understood fashion.E.g. in open thermal and chemical systems) others are been surgery, the surgeon measures a number of tody more simple : compartments, biochemical functions of his patient that he believes to be related to the maps , block diagrams . We present a current status of the patient.He would like to simulate the unified Bond Graphs approach to design interactions between then variables on line to be able to separate and/or coupled mc..1els using any of these formalisms . forecast potential problems early on in the game. However, the Physical parts are thermodynamically basedand phenomeno- equations governing these interactions are largely unknown, logical partsuse a kinetic approach. and there can be noticed a tremendous variability in the Biological structures are clearly dis- behavioral pattern exhibited by different patients. playedand submodel compatibility is This parer describes a methodology that lends itself to the evaluated by algorithmic causality chec- modeling and simulation of such types of processes One of the king. A package will be presented which major characteristics of this methodology is the fact that the implements modular modeling andautoma- model building and model validation activities are not separated tic coding of model specifications into from each other, and that the obtained models are optimized for routines suitable for simulation . maximum forecasting power. Coupling of Bondgraphs and AI tools for model analysis ( causal qualitative analysis ) will be discussedusing muscle models as examples.

61; 4.4 6

228 1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas Biomedical Engineering Scientific Papers

BET -C.1 BE7-C.2 ERACTALS IN PHYSICAL SCIENCES, Fractal Kinetics of Ion Channels S. H. Liu*, Oak Ridge National L. S. Liebovitch: Columbia Univ. NY, 10032 Laboratory, P.O. Box X, Oak Ridge, TN, 37831-6032 Currents through individual cell membrane ion channels can be resolved by the patch clamp Most natural objects do not conform to Euclidean technique. The observed sequence of open and geometry, which is the backbone of modern science. closed states provides information on the kinetics of As a result the study of irregular objects has the changes in conformation occurring in a single eluded physical scientists until recently. The turning point occurred a decade ago when molecule. We found that the statistics of this Mandelbrot published an impressive collection of open.--closed switching is self-similar when viewed evidence that nature often prefers fractal geo- at different time scales. That is, the effective kinetic metry instead. As new examples of fractals are rate constantkeffmeasured at time scaletellhas the being found in almost every branch of science, we are confronted with two tasks: (1) to understlnd form keff = Ateffl-D where A is the kinetic setpoint and the natural processes by which fractals are form- D the fractal dimension. Such kinetics are predicted ed, and (2) to uncover the new physical properties by a model where there are many conformational resulting from fractal geometry. This lecture states of the channel with approximately the same consists of an introduction to the basic concepts of fractal geometry, a brief discussion of modern energy minimum and where the activation energy theories of fractal growth, and examples of how barriers between these substates are characteristic well-known physical processes are modified, often of a slightly non-linear oscillator. Processes at beyond recognition, by the new geometry. different time scales are linked because the same Research was sponsored by the Division of underlying physical mechanism produces both the Material Sciences, U.S. Department of Energy small energy barriers (of the fast processes) and the under contract DE-AC05-840R21400 with Martin large energy barriers (of the slow processes). Marietta Energy Systems, Inc. Supported by NIH grantEY6234and the Whitaker Foundation.

BET -C.3 BE7-D.1 BIFURCATIONSIN A 113DEL OFTHE PROPAGATING Nonlinear Dynamics and Percolation Processes in Cardiac CARDIAC ACTION POTENTIAL. M.R. Guevara: Arrhythmias, BoE.H. Saxberg, DanielT. Kaplan, Joseph M. Department of Physiology, McGill University, Smith, and Richard J. Cohen*, Harvar.r.tviiT Division of Montreal, Quebec, Canada H3G 1Y6 Health Sciences and Technology, Cambridge, MA.

We have shown that finite element models of cardiac An ionicmodelofastrandof cardiac conduction processes can account for a wide range of muscle is investigated.The strand is assumed disturbances of cardiac conduction processes including toobeytheone-dimensionalcableequation, reentrant tachycardia and fibrillation. which is numerically integrated. The proximal We have found that spatial inhomogeneity in local end of the cable is stimulated with a periodic electrical properties of the myocardium as well as the nature of train of current pulses.For sufficiently high the dynamical behavior of the local elements can play ratesofstimulation,blockofpropagation important roles in determining the electrical stability of the occurs;i .e.noteverystimulus produces an system. We have applied concepts from percolation theory actionpotentialthattraversestheentire and nonlinear dynamics to characterize the behavior of this length of the cable.Various rhythms of block system and to develop new techniques of identifying latent suchasWenchebach,reverseWenchebach, electrical instability of the heart. Some of these techniques alternating Wenchebach, Mobitz II, and al terans have been successfully applied experimentally and clinically. are seen as the stimulation frequency or some For example, we have identified subtle electrical alternans, an other parameter is changed.We associate the ABABtype pattern of variation in QRS and T wave appearance of a new rhythm as a parameter is morphology of the surfaceECG, tobe a quantitative indicator changed witha bifurcation. Each rhythm in of enhanced susceptibility to fibrillation in simulations, animal time is associated with a specific structure in studies and human studies. Electrical alternans represents a space,leadingtotheconceptof a period doubling behavior in this system. We are addressing spatiotemporal pattern. the question of whether higher order periodicities regularly occur and whether fibrillation can be characterized as a 'chaotic Supportedbygrantsfromthe MRC and behavior' in the framework of the theory of nonlinear NSERC. dynamics.

249

1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas 229 Biomedical Engineering Scientific Papers

BET-D.2FRACTAL BE7-D.3 NETWORKS AS THE BASIS FOR Fractal models of vascular beds with HETEROGENEITY OF MYOCARDIAL, BLOOD FLOW and without anastomoses. J. Lefavrem J.B. Bassingthwaighte:J.H.G.M. van Beek, and R.B.King, Louvain University, Brussels, University of Washington, Seattle, WA, 98195, Belgium. and Vrije Universiteit, The Netherlands, Local myocardial blood flows range over six-fold in the normally functioning heart in anesthetized sheep and rabbits and in Fractal models may be used to study awake baboons. Flow distributions have relative dispersions blood flow in vascular beds with a func- (RD = SD/mean) that closely follow a fractal relationship with tionally required uniform perfusion . respect to voxel size, the mass, m, of the pieces into which the Three steps are necessary :to find an heart is divided: approximate topology ,to model spatial RD(11)=RDOTO(mtIN)" constraintsanddimensions of vessels, to choose flowequations and implement where the fractal dimension D is about 1.14, and mt, isa computations . We introducehere topo- chosen reference voxel mass. The regression coefficients were logical approximations of large -beds 0.99 in sheep and baboons, and 0.98 in rabbits. The values of based on minsquare error between RD(rno) varied considerably from animal to animal; in the Horsfieldcountings in the model and in awake baboons the range was 0.15 to 0.30. The shapes of the real samples . Themethod, based on an distributions were relatively stable over time. A useful expla- angiogenetic simple model , reproduces nation is provided by a fractal recursive branching of the structural aspects( proximal topology, arterial tree. Purely deterministic dichotomous branching with number of terminals , distribution of specific asymmetry at the branch points can fit the data So average lengths ) andresults in very doessymmetricalbranchingwithGaussianvariationin simplecomputation for bloodflows . daughter radii.Neither recursion reduces tothe equation Anastomoses maybe introduced and their above. Two dimensional fractals appear the most suitable. effect evaluatedonvery large scale models. The quality of the approximation Supported by NIH grants HL19139 and HL38736. is assessedbyHonteCarlo simulations andmay be increasedbyconstructing higher order approximating beds . Examples discussed are peal and pulmonary beds.

BE7-E.1 BE7-E.2 Parameter Estimation in a Model of the Electrical CHAOTIC BEHAVIOR IN TOAD VENTRICLE. GV Activity of a Single A-V Nodal Pacemaker Cell, Savino, DL Gonzalez and L Romanelli, C.R. Murphey*, J.W. Clark Jr., W.R. Giles* and Bioingenieria, Universidad Nacional de G.V.Naccarelli*,RiceUniv.,Houston TX Tucuman and Centro Arg Estud Radiocom 77251-1892 *Univ. of Calgary Medical School, Compatib Electromagnttica (CAERCE), cc 28 Calgary, Alberta, Canada T2N 4N1, *Univ. of suc 2,4000 Tucumtn, Argentina. Texas Medical Center, Houston, TX 77225 Max E.Valentinuzzi* In the past decade suction-pipette whole cell voltage clamp We recorded monophasic action potentials with small suction electrodes techniques have been used quite successfully to obtain from toad ventricles (Bufo arenarum) electrophysiological data from single, isolated cardiac cells which were paced by pulses of slowly of practically any type in the mammalian heart. In this increasing frequency. Resulting time study we successfully apply a nonlinear least-squares series, attractor reconstruction and parameter estimation technique to a model of the A-V spectral analysis gave a transition to nodal pacemaker cell and action potential data from single, disordered behavior which is compatible isolated A-V nodal myocytes. The model is baseon vol- with Feigenbaum scenario (Physics Today, tage clamp data reported from a variety of sourcts in the Dec 83,p 46). Subharmonic components literature.Itcharacterizes (f/2, f/4) was systematically detected in thetransmembraneionic the power spectra. The disordered currents, the membrane Na+IK+ pump, the Na+Ca2+ solutions yielded a continuous spectrum exchanger, changes in compartmental concentrations of which allowed their characterization. It Na+, K+, and Ca2+, intracellular buffering of Ca2+, and has been conjectured that Feigenbaum's uptake and release of intracellular Ca2+ by thesarco- route can be a precursor of fibrillation plasmic reticulum. The estimation technique has been used in the mammalian heart.In such case, to determine a number of key physiologic parameters asso- generalized ischemia due to fall in blood ciated with this mathematical model, and is capable ofpro- pressureappears as an additional viding a close model-generated fit to action potential data. complication. Rufo arenarum does not h.ve In addition, parameter sensitivity analysis is used a coronary system; thus, this to deter- complication is minimized. Conclusior,: mine the influence of changes in important modelparame- Cardiac syncitium presents a chaotic ters on the action potential waveform. behavior when driven at increasing rates. CONICET and CIUNT grants. 61, ft, t.)V

230 1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas Biomedical Engineering Scientific Papers

BE7-E.3 BE7-E.4 3-D Simulation Model for theElectrical The Simulation of the Cardiac Activity Characteristics in the Ischemic Excitation Propagation for Heart, E. Barta and S. Sideman*, the Analysis of Arrhythmias Technion- Israel Institute of Technology, K.MitsuiI N.Shibata, Y.Fukui, Haifa, 32000, Israel. System Engineering, Tokyo Denki University, A macro-global level approach is utilized to Saitama, 350-03 Japan simulate the electrical activation process in the ischemic left ventricle (LV). The propagation A new simulation model of the ventricular cita- pathway is chosen so as to minimize the time of tion propagation was established, which enables activation, subject to the followingsimplifying the analysis of the mechanism of the cardiac constraints: 1) Only 3 orthogonal directions of arrhythmias. The ventricular model consists of propagation are permitted, 2) the geometry is an 361 cell units, formed as two dimensional 19 x ellipsoidal one, and 3) the activation potential 19 cells. Each cell unit has three phases of the signal is a rectangle function with 2 possibleon activation cycles represented by depolarized, and off values.The proposed model results in refractory, and non-refractory periods. The simula- instantaneous activationmaps throughout the tion analysis is for the case that the strong myocardium in sequential heart beat cycles. The electrical energy is introduced for the induction relatively short computing time, typical to this of arrhythmias. model, enables to investigate the formation and As a simulation results, re-entrant pattern of development of electro-pathological phenomena, the the cardiac conduction propagation was observed, dependence of the activation characteristics on when refractoriness was prolonged in a specific the geometry of the ischemic region, the effect area of the model, and the specific timing of characteristic signal propagationspeeds, the the coupling interval of the shock was used. refractory periods, the rate of "shooting" of the This results was compatible with the experimental pacingcells and the rate of "shooting" and the results previously. location of t::e ectopic points. Thus, once the properphysiological conditions are defined, this model may eventually serve to evaluate the effects of variousdrugs on specific parameters involved in cardiac performance.

BE7-E.5 BE7-L6 Forwardproblem sinulation of un the Resolution Power of the Inverse Electrocardiographicbody surface ECGSolution Method, T. Furukawa*, T. potential nap using individualized Aoki and S. Hara, Institute of Medical torso-heartmodels, K.Kamiya*, Electronics, Faculty of Medicine, K.Sakakura, A.Iwata, N.Suzumura, University ofTokyo, 7-3-1 Hongo, K.Ohta, T.Goto, K.Douniwa, K.Iwao Bunkyo-ku, Tokyo 113, Japan. andM.Okajima, Nagoya Inst. of Tech., Nagoya,466,Japan As an easily understandable example of the applicationofinverse solution in clinical In order to interpret the variation diagnosis, a method to estimate the epicardial of body surface potential mapdepending potential distributionthrough an inverse upon configuration ofthe torsoand solution process using the source data obtained heart, a forward calculation system by frombodysurface potential measurements is the finite elementmethod using the presented. The inverse solution is based on the torso-heart model so built as to fit to boundary element method. each individual was developed. The resolutionpower of this methodwas Electromotive force moments of the carefully investigated with animal experiments ventricleswere expressed bymultiple in which two foci of myocardial injury were made lipoles. The torso was divided into on the ventricular anteriorwall using a about5,000 tetrahedron elements with specially designedfreezing device, and the intentionto give smaller element size results of the inverse solution were compared to to the regions withhigher lurrent those actually measured. The resolution power density. Such heart - tcrso models were was estimated to be 3.5 in the animal built for five individuals with experiments; however, it washigher in the differentbodyfigures inheight and computer simulations (2 cm). weight. Bodysurfacepotentialswere In another experiment, the freezing injury calculatedusing theseriodels,bringing was made on the endocardial myocardium. In this aboutbody surface maps which were experiment, the resolution power was much lower different fromeach other butmatched due to the masking effect ofthe myocardium. fairlywell to the actual mapobtained However, in all cases,the centers of the foci with respective individual. were exactly estimated. 251 1988 World Congress on Medical Physics and Biomedical Engineering o San Antonio, Texas 231 Biomedical Engineering Scientific Papers

BE7-F.1. BET-F.2 Measurement and Modeling of Virtual Cathode Effects Simulation of the effects of myocardial in Cardiac Muscle, J.P. Wikswo, Jr.*, J.P. Barach, anisotropy on electrocardiographic potentials, W.A. Altemeier, and D.M. Roden, Vanderbilt R.M. Gulrajani* and M. Lorange, Institute of University, Nashville, TN 37235 USA Biomedical Engineering, University of When cardiac tissueis stimulated with a small Montreal, Montreal, Canada. cathodal electrode, an activation wavefront propagates away from the stimulus site.At threshold, A heartmodel thatincorporates myocardial the size of the region initially depolarized by the anisotropy with a reduced conductivity in the stimulus is a minimum and is known as the liminal endocardium-to-epicardium direction is length.As the stimulus strength is increased, the described. The effects of this anisotropy on initially-depolarized region enlarges. Since the resultant dipoles of the model, and on the propagation begins at the edge of this region, subsequently-computed vectoccardiogram, rather than at the electrode, this region is termed electrocardiogram and body surface potential the virtual cathode (VC).Measurements of the mapare considered. The heart dipoles are arrival time of the wavefront at multiple elec- oriented oblique to the outwardly-propagating trodes can be used both to determine the conduction excitation fronts, with a consequent narrowing velocity and to estimate the size of the VC.A 10 of thesimulated vectorcardiographic loops. mA, 0.5 ms stimulus pulse applied to a 250pifi During late depolarization, when the epicardial electrode will produce a VC ranging from excitation fronts are partially tangential to 2 to 4mm, depending upon fiber direction and the intraventricular blood masses, the oblique tissue excitability. We have used a two-dimen- dipoles are even more so. Introduction of sional electrode array on the in vivo canine heart these high-conductivity blood masses results to study thevirtual cathode effect and to in Brody-effect decreases in the second peak determine how activation propagates away from the of the scalar vectorcardiogram, and in virtual cathode.Certain features of the data can diminished S waves in the electrocardiogram, be explained using a one-dimensional, non-linear, as has been experimentally observed. Finally, core conductor model for cardiac propagation, while the simulated body surface potential maps other features appear to require a two- or three- support, on a qualitative basis, the dimensional anisotropic bidomain model. The hypothesis of a partial cancellation of existence of the virtual cathode effect may have myocardial anisotropy and intraventricular significance in cardiac pacing, cardiac defibrilla- blood mass effects on surface potentials. tion, and the stability of reentrant arrhythmias.

BE7- F.3 BE7-F.4 STATISTICAL DERIVATION OF AIMED ELECTRO- Reliable Detection of Myocardial Ischemia by Modi- CARDIOGRAPHIC LEADS, Gy. Kozmann: R.L. Lux and fied ST/HR-analysis, H. Sievinen*, L. Karhumalit, I. L.S.Green,CardiovascularResearchandTraining VuoriI, J. Malmivuo, Biomedical Engineering Labo- Institute, University of Utah, Salt Lake City, UT 84112 ratory, Tampere University of Technology, SF-33101, A statistical methodissuggestedtoderive"aimed Tampere, Finland, tKanta-115..me Central Hospital, leads"(AL)-linearcombinationsof unipolarleads- Hameenlinna, Finland, IUKK Institute for Health Pro- which respond to cardiac sourcesinlocalized regions motion Research, Tampere, Finland. of theheart,whilesuppressingthecontributions of otherregions.Bodysurfacepotentialdistributions (BSPD) of a The maximum ST/HR-slope analysis has been shown representativegroup of normal (NOR) to improve the reliability of the exercise ECG in de- subjectsandthatofthepatientswithlocalized myocardialinfarction(MI)wereused.TheAL tection of the coronary artery disease (CAD). However derivation is based zn the determination of thatlinear the statistically significant slope is not always calcula- combinationoftheBSPDs whichmaximizethe ble, which limits the clinical applicability of the test. separationbetweenthe NOR andtheMI group A supplementary slope based on the two last measured considered according toa mean squarecriterion. By ST-segment levels is comparable to the statistically sig- systematicallycomparingthe NOR groupwith MI groups of different location, ALs can be derived for nificant slope as an accurate indicator of CAD. anatomicalregionsofthemyocardium. The same As for the ECG leads required for optimal ST/HR- learningsetsallowtheestimationofBayes-type analysis chest leads V4 and V5 are the most sensitive decision schemestogiveregionalinformationabout to detect the ischersue response regardless of the site thestrengthofelectricsources.Theexamples of the perfusion defect. presented were computed by the use of 40 each NOR Furthermore, to provide and anterior and inferiorMI patients. The an all-inclusive detection of CAD leads II and aVL computations were carried out inthe Karhunen-Loeve are incorporated to the analysis.This lead system domain representation of BSPDs and theresults were guarantees the appearance of the indicative slope transformed back intothe measurement domain. (>1.3pV/bpm) practically in all ischemic cases. The performance of our method in detecting the underlying ischemia was 97.5% within our patient material (40 patients with various MIs). 252 232 1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas aMilammr.

Biomedical Engineering Scientific Papers

BE7-F.5 BET-F.6 Simulation of electrocardiogram (EKG) Enhanced Methods for ECG Signal Averaging Applied by a finite state automat. F. Prieto.. to Detection of Ventricular Late Potentials. R. Jane*, P. Uni versi dad Autonoma Metr opo I i tana Carninal, J.M. Dominguez". Institut de Cibernitica, Diagonal Iztapalapa. 09340, Mexico, D.F. 647, 08028 Barcelona, Spain. "Hospital Sta. Creu i S. Pau.

A computer program is used to simulate Thepurpose of this study is to obtain the ventricular ie genesis of the EKG by means of an late activity from the surface ECG. Signal averaging of array of cells each of which can be in the High Amplification ECG is the technique used to one o three possible electrical states, recover these micropotentials. In this work, the signal is i.e. resting, active or refractory, and measured by means of a two-channel electrocardiographic change state accordingto certain rules system:a low-gain amplifier for the reference ECG (order of change, conduction of action acquisition and an isolated low-noise high-gain amplifier potential) and parameters (duration of for the late potentials obtention.The selected leads each state). The shape of the array and are a standard bipolar limb, lead for referenceand a the number of cells in it can be special bipolar precordial lead for a high amp:ification. designed by theuser, arrays may be The detection of the smallhigh-frequencypotentials stored in or read from disk. Most after the QRS complex is limited by inaccuracy in the parameters are assigned according to beat synchronization. Recently, some papers have been cell locationbut may be changed if the published about the incidence of the alignment errors in user so wishes. Three frontal plane EKG signal averaging. Probably, the low prognostic accuracy leads may be chosen for display. A reported is due to these technical problems.In order dipole is formed wherever a cell in the to obtain useful results several synchronism methods 'active" state is an immediate neighbour have been used in the present work, including double of a cell in some other state. Each level estimation and matched filter estimation.Results individual dipole contributes to the obtained by signal averaging of the ECG from patients potential at each 'electrode' according withdistinctcardiac disease (myocardialinfarction, to the 'square of the distance" rule and ventricular tachycardia) are presented. A comparative simple addition of effects. Interesting analysis of the different temporal alignment methods has experiments and demonstrations are made been performed. The incidence of the bipolar precordial po5ible by run time parameter changes. lead position and the selected bandwidth has also been evaluated.

BE7- G . 1 BE7-G.2 Identification of Multiparametric Signals A ParametricApproach for theAnalysisof Membrane Current Fluctuations, K. Yana*, in Time Domain, C. Druzgalslci: Hosei University, Koganei City,Tokyo 184 Japan, California State University,Long Beach, H. Mino, Juntendo University,Hongo, Tokyo 113 Ca..90840, USA Japan.

Characteristic patterns ofrespiratory airflow and correspond- Membrane current fluctuations have been analyzed ing respiratory sounds with their peculiar spectral composi- to reveal the kinetic properties ofsingle ion tion reflect conditions of pulmonary ventilation. These channels at biological membranes. Thepower combined signals are used to develop partly linear regression spectrum pattern is associatedwith kinetic models. Specifically, synthesizedmodels, ACF and PACF properties of the single ion channel. Therefore, has patterns and their relations with frequency analysis as well as the accurate estimation of the power spectrum crucial task in theanalysis. Discrete dynamic interrelationships between time series provide im- been a Fourier Transform (DFT) method has been proved analysis of acoustical signals and their correlation with conventionally used for the estimation. However, underlying patterns of rf.spii story airflow. In particular, the the parametric method would be more efficient when implementationof these cotcepts of computerized rather parametric auto regressive (AR) or auto regressive than visual model determination on a PCbased system enhan- movingavarage (ARMA) models are to be held for ces on-line analysis of these signals. the phenomena. We would like to note thatthe membrane current fluctuations could be modeled as In order to accomplish this time domain characterization of (n-1,n-2)ARMA model, where 'n' is the number of signals, an analysis based upon the classification of the kinetic states of the single ionchannel. Since ARIMA (p,d,q) models which specifyautoregressive and ARMA parameters are explicitly associated with moving average factors with a corresponding degree of kinetic properties, we could estimate them by the stationarity is conducted. Obtained patterns of autocorrela- ARMA parameters. Computer simulation revealed the tion functions (ACF), and partial autocorrelation functions parametricmethod requires lees than 25%record (PACF) are automatically compared with the theoretical pat- length to achieve the sameestimation accuracy terns of these functions. Spectral signatures and correspond- with the conventional DFTmethod. The proposed ing ARIMA models provide supplementary characterization method would be an alternative method for the of various conditions of ventilation. analysis of membrane current fluctuations.

2 5' 3 1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas 233 Biomedical Engineering Scientific Papers

BE7-G.3 BE7-G.4 An Adaptive Mechanism for the Regulation of Blood Fast Algorithm of Prony's Method and Glucose, E.A. Woodruff* andR.B. Northrop, University of Connecticut, Storrs, Connecticut. Application to lhysiological Analysis, L. Wang, N. Xu, F.S. Yang', The purpose of this research is the closed-loop regulation of the blood glucose concentration Tsinrhua University, Beijing, I.RG. (BGC) in a Type-1 diabetic when sparse measurements of the BGC are available. A model reference technique is implemented to provide an From physiological point of view, it is estimate of the BGC. Whenever a BGC measurement is not available the BGC estimate is used in the often an adequate assumption thatmany bio- controller realization. After developing the model referenceoff-line, medical signals are coposed of complexex- a computer simulation is realized for the closed- ponentials with various time delay.Prony's loop regulation of the BGC in a Type-1 diabetic. An on-line parameter estimation scheme isimple- method can be used for the decomposition. mented to update the parameters of the model ref- erenceduring the closed-loop simulation. These The problem is that the conventional algo- parameter estimates can be used to adjust some of rithms are noise sensitive and timeconsum- the controller parameters to enhance the robustness of the controller. ing. A method based on an extensionof DFT Since a linear model reference was evolved off- is proposed. It operates much faster line from a complex. nonlinear glucose-insulin than plant, a Kalman filter is utilized as the on-line the conventional methods and performs fair- parameter estimation scheme to update the parameters of the model reference. The Kalman filter ly well with respect to noise. Applications is used, since thismethod isexcellent in dealing with the transient properties of the to compartment analysis and to evokedpo- parameters when there is a short data record. In tential and Doppler blood flow signalana- addition, UD factorization algorithms are imple- mentedto ensurethat the covariancematrix lysis will be presented. remains positive semi-definite during closed-loop operation.

BE7-G.5 PE7-G.6 COMPARTMENT MODEL ANALYSIS OF CREATINE KINETICS The Analysis Of Multiexporential Curves iN MAN. By Pade Approximants And Size H. Thierens(1)*, J. Delanghe(2), J. Seuntjens(1), Identiflability Of Compartmental Models M. De Buyzere(2) and 0. Segaert(1), Phys. Lab. (1) Z. Bajzer; A. Myers and F.G. Prendergast, Clin. Chem. Lab. (2), Faculty o1 Medecine, Dept. of Biochemistry and Molecular Rijksuniversiteit Gent, Ghent, Belgium. Biology, Mayo Foundation, Rochester, Minnesota 55905 Important transient rises in serum and urine creatine concentrations are observed after myocardial A common problem in mathematical modeling of infarction and damage to skeletal muscle. As crea- dynamicalbiological systems isthe analysis tine is released by the damaged cells within a of curves described by multiexponentialfunc- few hours following injury, enhanced levels in tions. The method based on Laplace transform serum and urine are indications for an early aiag- and Pade approximants [Nature, 326(1987)169] nosis of a myocardial infarction in patients. offersthe possibility to detectthe number A compartment model is developed to describe the of componentsin multiexponentialcurves, as creatine concentration in serum and urine as a well astheirparameters. We present two function of time following an infarction or muscle generalizations of this method: 1) when the injury. In the model the damaged tissue, the curve isthe convolutica of multiexponential plasma pool, the skeletical muscle tissue and the function and experimentally knownUstrumen- liver are considered as different compartments. tal impulse response function;ii)when the The values of the parameters describing the kine- signalis not time dependent but is givenas tics between these compartments are derived from Fourier transform in the frequency domain. basic data on the creatine metabolism. The crea- Examples are provided by the analysis of tine concentrations in serum and urine are obtai- tryptophan fluorescence in proteins. ned by numerical integration of the differential ThePade-Laplace method is discussedas a equations resulting from the model using the tool for determining h.: number of compart- Runge-Kutta approximation method. ments in compartmenta models. In addition Effects of variation in total amount liberated by to formal developments, we perform simulatior the injury, sex, age (renal clearance size) and studies, especially with respect to the effi- body weight are investigated. The results of the ciency of the method when the noise levelin simulation are compared with clinical data. output functions is varied. 254

234 1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas Biomedical Engineering Scientific Papers

BE7-H.1 BE7-H.2 Vowel Recognition using an Auditory Model, Beat-by-Beat Analysisof Cardiovascular T.S.Yoon/,' J.H.Lee, S.H.Park, Yonsei Univ., Control withMultivariate Autoregressive S.H.Beack, Myongji Univ., Seoul, Korea. Modeling, S.Kalli, V.Turjanmaa. S.Majahalme, M.Sydanmaa. V.-M. Haggman. The characteristics of human peripheral auditory Technical ResearchCentre of Finland and system are applied to Korean vowel recognition. Tampere University Central Hospital. The model of auditory systemconsists ofthree Tampere. Finland parts ; the band pass filter bank withdelay tabs, which represents the resonance of basilar .3pontaneous oscillations and fluctuations membrane, the interaction part between filter occur in blood pressure and heart rate bank and the cochlear fluid, andmechanical- are related to the cardiovascular neural transduction of haircell part. The output control. This paper proposes to study of this model takes the shape of auditory nerve them with multivariateauoregressive firing probabilityas a function cf time. The modeling technique, which describes each firing pattern of each nerve is synchronized to signal as a linear combination of the the input signal and only the frequency compo- past values of its own and the other nents corresponding to CF are intensified. Con- signals. Themodel is used to determine sequently, the amplitudes of formant frequencies the interaction between the variables. are manifested in running spectra, which arrange The method was applied to the analysis of nerve firing pattern as their CF's. In addition, blood pressure and heart rate inter- pitch of the vowel can be estimated from the ma- actions in normotensive and borderline xima of low CF's firing patterns. Therefore, the hypertensive subjects in supine and output of the model represents the major feature standing positions. The interaction of the vowel. Speaker-independent recognition increased significantly in standing test using these features shows the accuracy of position in both subject groups, but no 96 % for closed, and 96.47 % for open test. significant difference could be fauna between the groups. The method proved to be a feasible way in studying complicated interactions between heart rate and blood pressure.

BE7-H.3 BE7 -H.4 Analysis of Feedback Systems in the Body Simulation using SPICF2 of an Enzyme with Use of Autoregressive Modelling. System which captures Energy from an Takao Wada;' Kidney Center, School of Oscillating Electric Field. D, C. Medicine, Keio University, Tokyo, Japan Mikulecky, Department of Physiology and Biomedical Engineering Program, Medical Autoregressive modelling has been applied for College of Virginia Common wealth feedback regulation of various mechanical systems. University, Richmond, VA 23298-0551. Recently, we demonstrated the usefulness of Akaike's approach of autoregressive modelling foe Austumian et al. (P,N,A,S. 84:434-438, 1987) have analyzing immunologic systems in the body. In the introduced a model for an enzyme system which present study, I tried to confirm its usefulness captures energy from an oscillating electric for analyzing other systems, such as endocrino- field. The energy from the field drives the logic and metabolic systems. reaction in the direction of ATP synthesis. Time series studies were done on 6 albino rabbits Using the cirsuit simulation program SPICE2 as a with or without administration of angiotensin general-purpose simulator, the system is easily converting enzyme inhibitors. Although the iaw simulated. Furthermore, a D.C. transfer curve blood chemistry data showed seemingly bizarre can be generated by SPICE2 which clearly fluctuations, the analysis with Akaike's method demonstrates that all the steady states available enabled to describe the impulse response curve to the system result in ATP breakdown. ATP proper to each system under study. Also, Akaike's synthesis occurs only when the system is forced relative power contribution, which was calculated to oscillate. SPICE2 is also used to generate from the matrix of autoregressive coefficients and the frequency dependence of certain aspects of the covariance matrix of the white noise term, was the system. The use of SPICE2 as a general- found helpful to disclose the cause-and-effect purpose simulator will be discussed. Also, the relationship between the variables composing the modeling of chemical reactions in membranes which system. Thus the analytical results clearly depend on electric-fields across the membranes demonstrated that the well known discrepancy will be explained. (Invited presentation). between the in vivo effects and in vitro effects of angiotensin converting enzyme inhibitors is due to the existence of multiple feedback pathways in the body. 255

1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas 235 Bic-medical Engineering Scientific Papers

BE7 -H.5 BE8-A.1 Characterization of Nonlinear Biological DEVELOPMENTS IN BLOOD OXYGENATION Systems by Wiener and Volterra Kernels, Peter D.Richardson' Naohiro Ishii*, Katsumi Yoshine, Haiwen Chen, Division of Engineering and Biology and Medicine Nagoya Institute of Technology, Showa-ku, Brown University, Providence, RI 02912 Nagoya 466, Japan.

Many nonlinear characteristics are found in the Blood oxygenation science, technology and application have passed through several stages. The science has focussed biological systems. More often, nonlinear biolog- on understanding the principles of transport in the convective ical systems are avairable only as "black boxes". flow setting of many different geometrks, and on the Then we can only obtain information about systems avoidance of excessive protein deposition and thrombus inner structure by the relation between systems formation on surfaces. Flow in curved channels has attracted input and output. By using the input and output, much attention because of increases in transport that are we characterize the nonlinear structure of the possible. The effects of blood pulsation superimposed on a biological systems by applying the Wiener's G- steady mean flow over membranes have been investigated in functional expansion and Volterra series. In bio- many flow passage shapes. Passages which are not flat or logical systems, the linear and nonlinear subsys- straight, and which have ridges or curvature instead, give tems are interconnected in cascade or with feed- improvedtransportrates.Theidealsituationwhere back. Then, conditions of the nonlinear biologi- anticoagulation begins at entry into an extracorporeal circuit cal systems are derived, particularly the linear and ends when blood emerges from it has not yet been and nonlinear subsystems in cascade and with feed- attained. Technologyhasconcentratedonmaterials, back loops, by using the Wiener's and Volterra's especially membranes, and construction methods, and much kernels words. These conditions are constructed attention has been given to detail in circuits and monitoring as algorithms of characterization of black boxes. capabilities. The failure to demonstrate significant clinical Then we applied these algorithms to the computer benefits in the extracorporeal membrane oxygenation trials simulated black boxes and retinal neurons in the (ECMO) has been followed by questioning of the berh.:it of visual system, by using input and output measure- vigorous respiratory therapy being applied simultaneously. ments. Application developments have been centered especially on seeking improved concurrent treatment modalities for the natural lungs when support during acute respiratory distress (ARDS) is provided. There is encouraging progress.

BE8-A.2 BE8-A.3 Blood Purification for Chronic Renal Augmentation of Detoxification by Failure, Prakash R. Keshaviahl Ph.D., Hemoperfusion, B. Brian*, W. Dorson, Regional Kidney Disease Program, V. Pizziconi, Arizona State Minneapolis, Minnesota University, Tempe, AZ 85287. Significant developments have taken place in Hemoperfusion is an effective therapy blood purification techniques for end state for the removal of soluble toxins up to renal failure over the last three decades. the column breakthrough limit. Many of The population on chronic dialysis has grown the most dangeroussubstances are enormously with over 90,000 patients currently protein bound and the removal using on dialysis. In addition to hemodialysis and adsorption media, such as coated peritoneal dialysis, therapies such as hemo- activated carbon (CAC), is limited by filtration, hemodiafiltration and high flux the inability of the bound toxin to dialysis have been developed. These therapies diffuse and adsorb within the adsorbent have become much more efficient and treatment matrix. However, the removal can be times have decreased from 36 hrs/wk to between significantly augmented by the addition 6 and 9 hrs/wk. Equipment used for blood purifi- of a biocompatible unbinding agent cation in renal failure has become much more (solutizer) proximalto thecolumn sophisticated with incorporation of microprocessor entrance. The action of the solutizer technology and other innovations. Quantitative is to increase the unbound fraction of modeling approaches such as urea kinetics, heparin toxinavailable for intraparticie kinetics and kinetics of vascular refilling are diffusion and subsequent adsorption. being used more extensively. In spite of the This concept has beenproven for population on dialysis being older and having numerous drugs of abuse, cholesterol, more co-morbid conditions, significant improve- and bilirubin. A unique combination of ments in morbidity and mortality have occurred. a pore diffusion adsorption model and a The correction of the anemia of renal failure by simplified protein binding model is used human erythropoietin along with reductions in to predict the solutizer and toxin treatment time and symptoms during dialysis are dynamics within a CAC column, a priori, expected to contribute to the improved well-being with parameter estimation from and rehabilitation of the patient with chronic simplified experiments. renal failure. 26 236 1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas F

Biomedical Engineering Scientific Papers

BE8-A.4 BEti-B.1. Solute Transport in a Hybrid Rio- Artificial Pancreas, Renal epithelial cells grown on permeable membranes as M.R. Pillarella *, A.L. Zydney, Department of Chemical a potential ultrafiltration processor, P. Aebischer7T.K. Engineering, University of Delaware, Newark, Delaware Ip, Brown University, Providence, RI 02912, USA

Many of the long-term complications associated with Dialysis has dramatically changed the prognosis of renal failure, yet it remains burdened by considerable morbidity the current treatment of diabetes by insulin injections because of the non-physiologic manner in which it performs have been linked to the lack of physiologic glucose and some of the most critical renal functions. The combination of insulin control. Recent studies have demonstrated the po- an ultrafiltration device with exchangers that actively tential of the hybrid bio-artificial pancreas in controlling transport solutes may lead to a device that could purify blood in a more physiologic manner. We are presently studying the blood glucose concentration, but the insulin response to a transport capacity of the exchangers. Our approach consists glucose stimulation remains unacceptably slow. of growing functional kidney epithelial cells in a tight We have developed a detailed theoretical model for monolayer on permeable membranes. The cell lines solute (glucose and insulin) transport in a hollow fiber LLC-PK1, a proximal tubule-like cell and MDCK, a distal membrane bio-artificial pancreas. The model explicitly tubule-like cell, can be grown to confluence on nitrocellulose, cellulose, polyacrylonitrile and polysulfone accounts for convective transport arising from Starling either in hollow fibers or on flat sheet permeable membranes. flow in the device, including the effects of protein and Attachment factors must be used in order to obtain this red cell polarization on the flow patterns. Also included coverage. Dual label tracer studies reveal that inulin and is the reduction in solute iiffusivity in the porous mem- glucose move across all unseeded membranes with similar rates in eithcr direction. LLC-PK1 cell seeded membranes are brane as well as the enhancement of solute transport in impermeable to inulin but demonstrate a net apical to the lumen arising from shear-induced red cell migrations. basolateral glucose flux. Depending on the membrane used Model predictions agree well with in vitro data taken this glucose flux varies significantly. The effects of in Amicon hollow fiber cartridges (30,000 MW cutoff). attachment factors, substrate and cell types on the mass Results indicate that increasing convective transport sig- transport capacity of the seeded membrane are under study. We conclude that renal epithelial cells can control the nificantly improves the insulin response. The model pro- permeability of a membrane to various solutes, thus opening vides a fundamental understanding of the important phe- the way to the development of a biologically controlled nomena governing solute transport and can thus be used ultrafiltration processor. in the design of improved clinical devices.

BE8-B.2 BE8-B.3 Membranes for plasma Fractionation, EXTRACORPOREAL IMMUNOADSORPTION P.S. Malchesky ,M. Zborowski, and Y. Nos6, George Buffaloes Ph.D, COBE Laboratories, Inc. Department of Artificial Organs, Cleveland 1201 Oak St. Clinic Foundation, Cleveland, Ohio Lakewood, Colorado 80215-4407

Extracorporeal immunoadsorption (El) is defined Advances in the extracorporeal on-line treatment as the use of immoblized immunologic reagents of blood for therapeutics have made it an effec- for the treatment of blood outside of the body, Further tive technique in clinical applications. for therapeutic purposes. After many years of development to increase its cost effectiveness and laboratory and clinical studies, El is emerging is required. Membrane simplicityof operation as a therapeutic modality that today can be appl- separation technology is ideally suited to meet ied to a number of disorders. Membrane fractionation of plasma has these needs. Recent advances in fabrication engineering and been studied for the selective removal of macro- biotechnology are being applied to create new Soluteselectivity and molecules fromplasma. systems. There are certain core criteria that removal amounts have been shown tobe strongly must be met in the development of El systems related to the choice of the membrane and tlAe which include materials biocompatability, flow/ Operation in the cold temperature of operation. pressure/volume parameters, and circuit design. (cryofiltration) is particularly effective in the Systems can be designed to treat whole blood, solutes formingmicroaggregates at removal of plasma, or cellular fractions. below physiologic temperature such as cryoaggre- Because of the highly selective nature of El, a and immune getableproteins (as cryoglobulins major impediment to the application of this tech- Opera- complexes) and very large macromolecules. ology to treat current medical problems, is lack tion at above physiologic temperature (thermofil- of definition of the etiology of a significant tration)is more effective in lipoprotein frac- number of disorders to which El might be applied. tionation of the low and very low density lipopro- Thus, in the near term, El will likely play a Clini- teins from the high density lipoproteins. role in the elucidation of disease mechanisms, cal studies have shown that these techniques are and in the future, El has the potential of be- In addition, easy to perform and safe to use. coming an effective form of therapy. expensive plasma replacement productsare not required making membrane plasma fractionatLa a preferred alternative to plasma exchange.

tz)

1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas 237 Biomedical Engineering Scientific Papers

BE8-B.4 BE8-C.1 Critical Look at Controlled Studies in CCNIPLEMENP ACTIVATION BY BICMTERIALS , Plasma Exchange, U. R. Shettigar; Richard Johnson and Kenneth Burhcp. Department of Suregery, Building 535, University of Utah, Salt Lake City, 84112, Since the pioneering work of Craddock et al (New USA. Eng. J. Med. 22.6, 769 (1977) implicated complement activation as the probable cause of granulocytopenia during hemodialysis, there has This study makes an attempt to explain poor been a growing awareness of the role of results of controlled trials in plasma exchange complement in the analysis of device (PE) and better results of uncontrolled trials performance.This presentation will review reported in the literature.This is done by studies of complement activation by various looking at different treatment variables, namely: materials employed in clinical settings, in frequency and length of treatment, volume and animal model studies and inini vitro type of plasma exchanged,early or late experiments. Based on these studies sever:Al approaches have been identified to fabricate institution of PE and suppression of protein materials that display a diminished capacity to rebound phenomena. Controlled trials of PE activate ccmplement. Factors believed to be have been reported in Chronic Progressive important in this regard include surface Multiple Sclerosis (MS), Rheumatoid Arthritis nuclecphilicity and the presence of negatively (RA) and Acute Vascular Rejection (AVR). charged groups that influence the binding of C3a and C5a anaphylatoxins. In addition, surface chemistries that influence the involvement of Outcome of PE appears to be influenced by the control proteins (Factors H and I) have been above treatment variables. Regarding MS and found to affect the complement activating RA, the controlled studies,except one, potential of a bicmaterial. Analysis of performed only a few PE treatments as materials currently employed in hemodialysis compared to the longer and more frequent illustrates the importance of Factor H in uncontrolled studies. Such short studies may limiting the extent of activation and suggests give erroneous conclusions. that materials which facilitate involvement of this protein will display improved biocctpatibility.

BE8-C.2 BE8-C.3 Effect° a;' Flow on White Blood ,7aZ Adhe- CONTINUOUS MONITORING OF PLATELET ACTIVATION sion, Aggregation and Arachidonic Acid REACTIONS IN A CONTROLLED SHEAR FIELD. T. D. AttabuZiem, L. V. McIntire*, M. B. Giorgio and J. D. Hellums1 Biomedical Engineering Lawrence, Biomedical Engineering Labora- Laboratory, Rice University, Houston, TX 77251, tory, Rice University, Houston, Texas, USA. 77251, USA. A rotational viscometer has been fitted to permit monitoring of: (1) light trai mission as an The effects of flow on the adhesion of leuko- indicator ofplatelet aggregation, (2) light cytes to endothelial monolayers and biomaterial emission as an indicator of release, and (3) surfaces was investigated using a parallel plate fluorescence to indicate intracellular calcium ion flow chamber. In the wall shear stress range concentration. seen in venules, the dependence of adiv.sion on Results show thatthe shear stressintensity flow is extremely strong, leading to little ad- inflences both the extent of platelet aggregation hesion at wall stresses above 4 dynes/cm2, even andthe size distribution oftheformed aggre- on IL-1 activated endothelium. gates. In addition, using concentric cylinder viscome- Intracellular calcium ion concentration ter, the effect of stress on leukocyte arachi- increases only at moderate and high shear donic acid metabolism and aggregation was deter- stresses. Platelet aggregation which occurs in mined. There is an important role of platelets response to low dosages of added ADP and/or low in determining the metabolites formed by intensity shear stressis reversible and isnot stressed leukocytes and the aggregation kinet- accompanied by a rise in intracellular calcium ion ics. PMN leukocytes produce mainly 5-HETE when concentration. scraned in purified suspensions, butproduce Platelet release increases with shear stress. leukotriene By when stressed in the presence of Addition of ADP accelerates the release in a platelets. concentration-dependent way which is also poten- tiated by increased shear stress intensity. Continuous monitoring of release makes it possible to estimate platelet surfaceconcentrations of releasedchemicals. The platelet surface concentrations correlate well with expected rates and 258 extents of aggregation.

238 1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas Biomedical Engineering Scientific Papers

BE8-C.4 BE8-D.1 Influence of Surface Chemistry on Biological DYNAMIC VISUALIZATION OF VELOCITY PROFI- Interactions at Polymeric Interfaces, LES IN THE ASCENDING AORTA IN HUMANS. A.S. Hoffman; Center for Bioengineer ing, FL-20, P.K.?aulsen, J.M.Hasenkam, O.Albrechtsen, University of Washington, Seattle, WA 98195 Department T, Arhus University Hospital, 8200 ArhusN,Denmark.

When polymeric biomaterials are exposed to a biologic environ- As late valve related complications such ment, it is generally agreed that events begin with adsorption of a layer of proteins. The proteins in this layer can be displaced as thromboembolism and hemolysis are thought to be caused by hemodynamic dis- and/or biologically, chemically or physio-chemically modified during the proceeding seconds to minutes. Thus, the adsorbed turbances, the present paper focused on proteins can change in relative abundance, organization and the study of these downstream of normal antigenicity. Enzymatic reactions can also occur at the interface. aortic valves (n=8, prior to coronary by- It is this protein layer which mediates subsequent cellular interac- pass surgery) and St. Jude Medical (n=4) tions, such as the inflammatory response or thrombosis, as well and Starr-Edwards Silastic ball valves as embolization. The surface composition of the polymer sub- (n=3) in humans. strate has a significant influence on this sequence of events. A hot-film anemometer system with a 1 mm There have been many different approaches taken to modify the conical L-shaped needle probe was used surface composition to make it more or less responsive to the to register blood velocities at 41 almost biologic environment. For example, surfaces have been modified evenly distributed measuring points in by radiation grafting or plasma gas discharge treatments to depo- the cross sectional area of the ascending sit new surface compositions, which are sometimes later modified aorta. Later data analysis using a computerized drawing technique enabled us to byatmclunent of biologicallyactiveagents such asanti- thrombogenic agents.Surfaces have also been modified by make three-dimensional plots of the velo- adding surface active agents to the polymer, which accumulate at city profile at 100 time intervals during the surface during the formation process.Surfaces have also one mean heart cycle. By successive pho- been modified by direct chemical modification of the original tography of each plot animated films were made. In patients with normal aortic val.,. material. In this paper, I will briefly review the influence of such treat- ves flat but skew profiles were seen with ments on biologic interactions, and then I will present our own a counterclockwise rotation of the skew- recent results on a rather surprising and exciting effect of surface ness during systole. After insertion of treatment on such interactions. the artificial heart valves, their design was reflected in the velocity profiles.

BE8-D.2 BES-D.3 Color Doppler Studies of Normal Mitral Valves In Vivo Measurement of Turbulence M. Jones*, E.E. Eidbo and R.E. Clark, National downstream of Aortic Valves. Heart, Lung and Blood Institute, Bethesda, MD J.M. Hasenkam*, H. Nygaard, J.H. Ostergaard, Transepicardial color-encoded Doppler (Aloka SSD- E.M. Pedersen, P.K. Paulsen and B.A. Schurizek. 880 system) studies were performed upon the native Dept. Thorac. Cardiovasc. Surg. Skejby Sygehus. mitre' valves (MV) of 20 sheep.Heart rates were Aarhus Univ. hosp. and Instit. Experim. Clin. varied between 60-120 and cardiac outputs between Res. 60-120 ml/min/kg. Studies were performed in real Aarhus University. 8000 Aarhus, DENMARK. time and with electrocardiographic gating at 10 ms intervals. During early diastole, aortic closing Turbulent shear stresses are considered to cause he- volumes and forward displacement fronts of blood at molysis, thrombo-embolism and endothelial damage 20-40cm/s, lasted 10-40 ms prior to MV opening. At downstream of artificial valves. Quantitative turbulen- the onset of passive left ventricular (LV) filling, ce studies in vivo have been made only scarcely. the MV began to open, blood flow accelerated, be- Therefore, the numerous in vitro fluid dynamic studies com$ab mildly turbulent with regions of vortex are difficult to compare with the clinical situation. shedding behind the MV leaflets. By mid diastole, We made 33 aortic valve replacements in pigs and u- the leaflets were widely open and broad, laminar sed a1 mm hot-film anemometer probe, inserted 1 profiles (velocities < 60 cm/s) extended to the LV diameter downstream of the valve ring in the ascend- apex; flow occupied 100% of the MV orifice. There- ing aorta, for blood velocity measurements in the en- after, blood displaced from the LV apex appeared as tire cross sectional area. Twenty heart cycles were regions of flow reversal along the LV free wall and recorded succeedingly in each of 41 measuring points. septum producing partial closure of the MV. After The ensemble averaged velocity energies were sub-. atrial systole, LV inflow again accelerated (often tracted from the instantaneous velocity energies in > 60 cm/s) and the MV again opened. In late dias- frequency domain for extraction of mean systolic tole, flow reversals produced by displacement turbulence energies, and 3-dimensional visualization caused near closure of the MV prior to the onset of of turbulence profiles were made. Maximum Reynolds LV systole; thus the MV had a minimum closing Normal Stress (RNS) was approximately 500 Nm 2 volume, lasting only 10-20 ms. During isovolumic downstream of artificial porcine valves. Mean spati- systole, the MV leaflets shuddered with turbulent al mean systolic RNS ranged from 20 to 70 Nm 2. velocities akiacent to them.These observations Difficulties in comparing experimental data from may be compared to flow disturbances caused by pros- different centers and the need for international thetic cardiac valves and do have implications re- standardization will be discussed. garding future valve designs. n .4") 1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas 239 Biomedical Engineering Scientific Papers

BE8-D.4 BE8-D.5 Flow Mapping of Biopircetimelo and Tilting Disc Heart EFFECT OF PROSTHETIC HEART VALVE Valves in the Mitral Positiat In Vivo and In Vitro ORIENTATION;IN VITRO, STUDIES A.P. Yoganathari*, H. Jones, H-W. Sung, E.E. J.D.Walker, W.M. Phillips., W.G. Lederman Eidbo, S.T. McMillan and R.E. Clark, Georgia BIstitute of Technology, Atlanta, GA 30332-0100. Purdue University, W. Lafayette, IN 47907

2-D imaging and calor-encoded Doppler studies were Influence of tilting disc heart valve orientation on performed on 9 types of clinical and preclinical flow through a curved-tube aortic model was studied. bioprcsthetic valves and 5 types of dicing disc (TD) The primary criterion for evaluating performance was mechanical valves in the mitral position in sheep. The turbulentshearstress levels. Simultaneous, 2- TD valves were oriented with the major orifice "zo the component laser velocimeter measurements were made left ventricular (LV) free wall (FW) or to the LV septum (5).Porcine valves had high velocity (150-200 cm/s), using a mock circulation loop to simulate physiological turbulent, non-central, eccentric jets during diastole. pulsatile flow. A 22 mm diameter glass tube with a 41 Areas of flow separation and vortex formation were mm radius of curvature modeled the average dimensions observed beneath the leaflets.Pericardial valves had of the adult aorta.Results at mid-arch showed low broad central, but often asym metric, velocity jets with turbulence levels in early systole and no influence of in-orificeturbulencewhich decayed about 2 cm valve orientation.Fluid from the ventricle reached downstream. The observations in vivo were comparable mid-archat mid-systole shovingstrong valve to those obtained in vitro by laser Lopplem TD valves orientation influence and increased turbulence levels. had two high veloaty jets (max velocities 1-2 cm/s in With the major orifice of the valve adjacent to the both orifices). In FW & S orientatio-31vortex shedding and entrainment of adjacent blood produced eddies of inner curved wall, the peak turbulent shear stress was velocity reversal measuring 20-30 cm/s just distal to the 307 dynes/cm2.Rotating the valve 180 degrees, the disc. Greater intraventricular turbulence occurred with peak turbulent shear stress was a reduced value of 91 S than with FW orientation.In S orientation, large dynes /true. At the exit of the curved section, the flow areas of reverse velocities (25-50 cm/s) directed against .vas independent of valve orientation aad the turbulent the minor orifice inflow occurred along the LV FW in mid and late diastole. In contrast, valves in FW shear stress levels were an order of magnitude lower. orientation had velocity profiles which became laminar These studies show that the major orifice of tilting disc and directed to the LV apex 1-2 cm distal to the disc, valves should be adjacent to the outer curved wall to measuring 50-90 cm /s in mid and late ilastcle. minimize turbulent shear stress leve;..

BEO-E.1. BE8-E.2 Comparative Study of St. Jude and Edwards- FLUIDDYNAMICS AND FUNCTIONAL PERFORMANCE OF Duromedics Bileaflet Valves Using Laser Anemometry. BIOPROSTHESES H. Reul*, E. Knott, M. Giersiepen, R. Fatemi* and K.B. Chandran,University of Iowa, U. Steinseifer, Helmholtz-Institute for Biomedical Iowa City, IA, 52240. Engineering, Aachen, West Germany An in vitrocomparative study of St. Jude (SJ) and Since the number of different types of Bioprostheses Edwards Duromedlcs(DM) Bileaflet valvesof size is continuously increasing it becomesmore and more 27mm was performedunder steady(18Ipm: difficult for the implanting surgeon to differen- approximate peak systolic flow rate) and physiological tiate the advantages and disadvantages of specific pulsatile flow (time averaged flow rate of 6 !pm) valve types. Therefore, it may be of great help for conditions in an axl-symmetric chamber using Laser the surgeon to have comparative independent testre- Doppler Anemometry (LDA). LDA measurements were sults which are obtained under carefully controlled obtainedIntwodifferentorientations;o n e conditions and are relevant to the clinical situa- perpendicular to the leaflettilt axis (01) and one tion. parallel to the tilt axis (02). The axial and transverse This paper covers laboratory testing methodssuch as velocities as well as turbulent shear stress were pressure drops in steady and pulsatile flow, closure measured in both orientations at two different locations and lukage volumes, energy losses and velocity distal to the valves. The velocity profiles at peak systole field evaluation by means of Laser-Doppler-Anemometny. show presence of stronger vortex in 01 orientation In Pressure drop measuremerhs under various steady and the sinus region for flow past SJ valve compared to the pulsatile flow conditions serve asa reference for DM valve. Velocity profile distal to the di valve in similar clinical situations.Energy losses and reorientation OP was relatively flat where as for the DM gurgitant volumes also have an immediate clinical valve,a jet-like flow was present. This may be due to impact since they are a measure for the additional the curved geometry of the leaflets of the DM valve. We work load of the heart.The detailed analysis of measured higher magnitudes of maximum turbulent the down.r.eam velocity fields allowsa localization shear stress for theDM valve (765 dynes/cm2) of high shear-stress regions, wakes, flow separation compared to the SJ valve (675 dynes/cm2). The higher and wash-out of the valves in context with blood turbulent shear stress of the DM valve can be attributed demi 2e, platelet activation and potential thrombus to Its thicker leaflet, smaller angle of leaflets opening deposition. Comparative test results for about 15 and the presence of the leaflet curvature. different valve makeswill be presented and relative valve performance will be discussed. 260 240 1988 World Congress on Medical Physics and Biomedical EngineeringSanAntonio, Texas Biomedical Engineering Scientific Papers

BE8-E.3 BE8-E.4 NOVEL FLOW VISUALIZATION TO DETECT SITES OF Effects of End Diastolic Ventricular Volume on the THROtIBUS GENERATION AT ARTIFICIAL HEART VALVES Dynamic Performance of Prosthetic Heart Valves, K. Affeld*, P. Walker, K. Schichl, Universitatsklin - M.C.S. Shu and N.H.C. Hwang? Cardio'iascular Flow ikim Rudolf Virchow, Berlin, F.R.G. Dynamics Lab, University of Houston, Houston, DC The flow through artificial heartvalves has been 77004 studied at great length using various methods. LOA has especially rendered man:, results. Unfortunately, we still do net have a detailed picture of how the The dynamic performances of eight clinical prosthetic overall instantaneous flow looks.Although the flow heart valves (PHV) in both mitral and aortic positions are has been visualized and photographed, the pictures studied by using three fixed left ventricular end-diastolic show little details, mainly because of the small volume models (FVM), and a supplemental varied left size of valve and the high velocity of the flow. ventricular end-diastolic volume model (VVM). The To overcome these obstacles a mock circulation has experiment is carried out to observe a relationship been developed, which generates a flow for a 10 to 1 between the PHV performance and the three fixed left enlarged model of a heart valve (Bjork Shiley). This ventricular end-diastolic volumes, under otherwise large size and the use of water reduces the velocity identical conditions. The results are compared to those to 1/35 of a real one. Together with the enlarged obtained from the VVM. Systematic analysis of the in scale this improves the lighting condition by the vitro data indicates that changing ventricular end-diastolic factor of 350. volumes does exhibit significant increase of pressure The flow is kept similar by ensuring drops across the valves,alterationof regurgitation that the Reynolds-number, the Strouhal-number and a valve area" third describing the buoyancy of the occluder are patterns and decrease of "the effective (EVA). the same. The flow is made visible with suspended particles, injected dye, hydrogen bubbles and particles generated directly in the boundary layer.The interest is concentrated on flow separation - "stag- nant flow" - known to be the preferred sites for thrombus generation.These sites are found in the wake of the "ing and behind the struts. In the real valve they are of minute size - about 1x 2 mm - and hardly detectable. In the enlarged model they are clearly visible. Their location coincides with the sites of observed thrombi reported by others.

BE8-E.5 BE8-E.6 The Interaction of Wall Shear Stress and Prosthetic Studies of the Flow Field of Bi Valvular Cusp, S. Einav, , D. Stolero, J.M. tivider, leaflet Mechanical Valves with D. Elad, and L. Talbot, Tel Aviv University, Cambered Profile, Zhenhuang Kang, Tel Aviv 69978, ISRAEL, and University of California, Bioengineering Institute, Berkeley, CA 94720, U.S.A. Chengdu University of Science & Technology, Chengdu, Sichuan, China High levels of wall shear stress on the surface of valvular cusps can cause mechanical damages to the blood cells and Bileaflet mechanical valve with cambered the cusps' surfaces. The shear stresses are also responsible for profile was designed as a new valve type mechanical failure of prosthetic heart valves.Quantitative based on thin airfoil theory. it is the measurements of wall shear stress in the vicinity of the leaf- purpose of this design to improve the lets is thus esential for diagnosis of suspected complications closing behavior of mechanical cardiac and an important information for design and fabrication of prostheses so that the impact and regur bioprosthetic heart valves.For this purpose we measured the gitation ran be reduced. In addition to velocity distribution along the inside wall of the cusps of a the design of abovementioned valve, in tri-leaflet heart valve with a two color laser Doppler ane- order to get some features of the flow mometer system. The wall shear stresses on the cusp surface field and situations of thrombus forma were computed and found to range from 80 to 120 N/m2 tion, we have conducted experimental during the ejection phase.Wall shear stresses of up to 180 studies about the wall shear stress and N/m2 were measured in loci of cusp flexure and the boun- the flow field of the valve by using dary layer accelerated.The results of this study show a electrochemical technique and hydrogen correlation between the high shear stress loci and the clini- bubble technique respectively. Moreover, cally (animal) observed regions of cusp clacification. a numerical study by means of FEM was also done. Finally, a quantitative con cept about the flow field was obtained, and some useful data for the valve design were provided through comparativestudies for different leaflet cambers and opening angles. 2 1

1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas 241 Biomedical Engineering Scientific Papers

BE8-F.1 BE8-F.2 HEART VALVE ENGINEERING: STEPS OF QUALITY ASSUR- Simple Device For The In Vitro Measurement of Regureation Of Prosthetic ANCE - J. Kohler, Aerodynamisches Institut, Aachen, Heart Valve In Mitral Position. Long-Sheng Yu! %Worn J. Kottf. Division of West Germany. Artificial Organs, Durrke Building. Univ. of Utah, Salt Lake City, Utah 84112

Measurement of regurgitation of prosthetic heart valve is kni)nnarrtin Steps of the development of test standards forar- evaluating the performances of artificial valves. The electromagnetic(E-M) tificial heart valves (AHV) withina project spon- !toy/meters are most commonly use, but with certain disadvantages. Our recently developed device, calledel-Inflow sored by the German Government are described.Pro- Valve Regurgitation Tester(BIVRT) is simple, cheap and easy to operate with low error in baddlow measurement as vided are acceptance tests for new types of valves compared to that of E-M flowmeter and without the disadvantages of ;rase -fine on the one hand, and random tests for already ap- shift, and axial asymmetry, disturbance of flowpath, and the !feces* o' proved valve types on the other. computer data processing. The variation in measured results by the BIVRT is The following test steps are under work: Develop- about 2.5%, which Is more accurate than that by the E-M flowmeter. By the use ment and definition of the medical and technical of BIVRT, comparison of baddbws of disc valve, ball caged valve and VI-leaflet valve were found to be that the latter is about 15 of that of the former two. basis for the quality assurance of AHV; development Further experiments showed that the baddlow measurements of a disc valve and definition of test devices and test procedures; by the BIVRT and the E-M flowmeter versus filing volume under the same comparitive assessment of already marketed valve condition were tom !to be that the readings by the former method is 20 % types by present day and new test methods. lower than those by EM ibwrneter. For the comparison of *the one-inflow valve mode testing- with a single disc valve and -the valve mode testing" A list of test parameters and test steps has been with one disc valve and one tricusp valve, the E-M flowrneter was used for defined in the project with regard to test devices gc.Jd correlation. The results of these two modes were found to be quite close urgently needed (geometrical parameters, flow para- to each other. Rtiasorts for the lower bac:blow measurements in the BIVRT meters, deformation- and force-parameters). A num- method as compared with those by the E-M flow meter are due to : 1) the ber of corresponding test devices is under work or resistance of the vent tube and its one-way valve; 2) the leakage of the additional competent valve; 3) the reverse suction of the vent tube during available as prototypes. diastole for the BIM set-up: and 4) measurement error by the E-M flowmeter. It can be summarized that quality assurance of

prosthetic heart valves - new and current types- RELATIVE REGURGITATION has not yet been sufficiently developed. Not only the qualities to be measured but also the procedures and the test devices have to be defined for rational valve testing. The volume loss, the deformability of the valve rim and the safelty of the valve structure lave to be taken into account more deeply in quality assurance of AHV.

BE8-F.3 BE8-F.4 Quantificaticn of Mitral Regurgitation from Turbulent Flutter in Bileaflet Cardiac Jet Theory Valve Prostheses. T. H. Reif* and E.G. Cape*, A.P. Yoganathan, E.G. Skourds, and RA. Ned H. C. Hwang. Republic Medical Levine, Georgia Institute of Technology, Atlanta, GA Prodnct. Inc.. Dayton. OH= and 30332-0100 Ca.diovascular Fluid Dynamics Laboratory. Houstin TX Due to well - recognized disadvantages of invasive diagnostic techniques, there is great need for a method to assess mitral valve incompetence noninvasively. The rotational form ofNewton's second Using currently available technology such as ultrasound law of motion is solved based on the Doppler velocity measurement and Caor Doppler flow dimen-.ios oc a sire2, mm Carbomedics mapping,attemptshavebeenmade tocorrelate Prosthetic Heart Valve ttm). The lift regurgitant volume with color jet area, etc.We have force and pitching moment areestimated taken a more fundamental approach and derived a from potential flow theory, while the theoreticalmodelfor theflowratethrougha drag force is estimated from the lift regurgitant lesion, based on turbulent jet theory. The force and a blunt body empirlCISM. equation expresses the flow rate through the lesion as a liluoaoc ond gravitational effects are function of ultrasound Doppler measurable parameters; based on the assumptionof homogeneou., In order to test the validity of the equation, w leaflets. Other assumot ons include constructed acylindrical flow chamber with small ,uniform flow. negligiblefriction at circular orifices of varying sizes.A 2 MHz transducer the pivot amts, and a symmetry (continuous and pulsed wave) interfaced to a Vingmed condition. Oscillations are observed in SD-100 ultrasound Doppler system was placed distal to the opening dynamics of theorder of and parallel to the regurgitant jet and the appropriate A-1 Hz. The frequency of the observed velocities measured at varying flow rates. oscillations ts dependent upon the Thirty-six combinations of volumetric flow rate and orientation of the gravitational field orifice size were tested and the calculated flow rate and the flow rate. These results suggest compared to the actual. An average deviation of only that the natural frequency of the 6.0% was found.The model shows high potential as a rotational second order system may noninvasive method and has the additional advantage of significantly affect the flutter of being theoretically grounded. bileaflet cardiac valve prostheses. 262 242 1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas Biomedical Engineering Scientific Papers

BE8-F. 5 BE8-G.1 Effect of Flow Rate on the Prediction of MASS TRANSPORT IN PERITONEAL DIALYSIS Effective Valve Orifice Area from Hemo- J. M. Won% M. J. Lysaght, and J. W. Garber, dynamic Measurements, FL Scnoephoerster* Baxter Healthcare Corporation, Round Lake, IL and K.B. Chandran, University of Iowa, Iowa City, Iowa. Peritoneal dialysis is a blood - cleansing process widely applied for the treatment of patients with end stage renal A study was performed to test the prediction disease. Hypertonic diatysate is periodically infused into capabilities of theclinically used Gorlin formula as well as the Aaslid and Gabbay the peritoneal cavity by gravity through an indwelling formulas for the effective orifice area(EOA) of catheter. Plasma water and solutes v. '''ch would prosthetic heart valves.Pressure gradient, normally be cleared by the kidney are transported from flow, and valve opening area measurements blood to dialysis solution through a rather complex were performed onfour 27mm valve series of resistances including capillary endothelium, prostheses(Edwards-Duromedics, St. Jude, interstitial tissue, peritoneal membrane and fluid Edwards-pericardial,and atrileaflet boundary layers. Fluids traverse this " membrane" in polyurethane valve) each mounted in the aortic response to osmotic pressure differences; sob..tes are position of an in vitropulse duplirw..,..With rem,,ved by diffusion; both may be returned to the the known valve orifice area. a discharge bloodstream via the lymphatic circulation. coefficient(Cd) was computed for each of the four valves and three EOA formu as.All Cd's were observed to increase with increasing Despite its complexity, the process may be adequately systolicflowrate(SFR). An empirical described by the analogy to a simple planar membrane relationship of Cd as a linear function of CFR separating two well mixed pools. This allows the was determined through a regression analysis calculation of equivalent transport parameters which for each of the valvegeometrieswhich characterize the system including: hydraulic improved the accuracy of the prediction of the permeability, sieving coefficienand mass transfer EOA.Further work is needed to test for the coefficients. These transport parameters may be readily effects of valve size on the Cd, as well as extrapolated and interpolated so that the behavior of a determining Cd's forprosthetic heart valves of wide variety of solutes may be predicted from a other geometries. relatively small number of experimental measurements.

BE8 -G. 2 IsE8-G.3 Fluid Mechanical and Transport Factors in Slow THE EFFECT OF MASS TRANSFER LIMITATIONS ON Continuous Renal Replacement Therapies. Thomas INSULIN SECRETION IN HYBRID ARTIFICIAL L. Pallone*, Division Nephrology, Univ. PANCREAS Rochester, Roch., NY Clark K. Colton, Keith E. Dionne, and Martin L. Yarmush, Department of Chemical Continuous arteriovenous (AV) hemofiltration Engineering, Massachusetts Institute of (CAVH) and hemodialysis (CAVHD) are important Technolgy, Cambridge, MA 02139. modalities foi treatment of hemodynamically un- Blood flow stable patients with renal failure. One approach to diabetes therapy is im- rate in these circuits is governed by the AV plantation of xenogeneic islets immunoso- pressure gradient, total circuit resistance and lated by interposition of a semipermeable Distribution of hydraulic blood viscosity. membrane between the islets and the bloodpressure is governed by the relative location stream or tissue. Internal vasculariza- Filtration of circuit component resistances. tion normally present is lost after iso- rate by the hemofilter/dialyzer is governed by lation, forcing the islet to depend in- transmembrane hydraulic and plasma oncotic stead on external and internal diffusion pressures and membrane hydraulic permeability. to provide an environment wherein the bet, The many variables involved make circuit opti- cells survive and secrete insulin. Two mization difficult. problems are described. (1) The effect of To aid proper circuit construction, mathemati- glucose and insulin diffusional imitations cal models have been developed which, given on observed insulin secretory dynamics. patient and circuit characteristics as inputs, Results from mathematical models define predict system performance. Assumptions include design conditions required to preserve in- the modelling of components as Poiseuille re- trinsic insulin secretion kinetics. (2) sistors and neglect of concentration polariza- The effect of p02 on intrinsic insulin se- tion of plasma proteins. S,mtial variation of cretion. Measurements in perifusion and blood viscosity is considered. Predictions of static culture demonstrate markedly re- performance agree with results of in vitro per- duced second phase secretion under reduced fusion experiments. p02 conditions. These results correlate with a lower islet core p02, as predicted by a mathematical model, for glucose-stimulated as compared to basal conditions. :283 1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas 243 Biomedical Engineering Scientific Papers

BE8-G.4 BE8-H.1 GLUCOSE AND INSULIN KINETICS IN A BIOARTIFICIAL Current Issues of medical Device Policies PANCREAS in Japan, M.Saito* and M.Kikuchi, Inst. M.Y. JAFFF N: UTC, UA CNRS 858 Medical Electronics, Fac. Med., Univ. Tokyo, Dpt de Genie Biologique, BP 649, Tokyo 113, Japan 60206 COMPIEGNE CEDEX, France A bioartificial pancreas is an implantable device The advent of complex systems, costly devices and connected to the diabetic host as an arteriovenous technologies requiring extremely high reliability, shunt. It contains insulin secreting cells is not making the role of policy-making agencies (Langerhans islets), separated from the difficult and complex. circulating blood by a semi-permeable membrane to The task should not simply be considered as a regulating mechanism. avoid immuno rejection. We present here a model Pre-ass- which describes the operation of such a device essment of coming technology, guidance of industry from a fluid mechanical point of view and evaluate and medical professionals, education and standardization are some of the important items to be added the respectivecontributions of diffusion and to regulatory functions. ultrafiltration to the glucose and insulin mass The reliability of technology should be secured from a large-scale stand- transfer. It is shown that the pressure drop along ardized production as well as from flexible small- the blood channel produces across the first half scale production. In international cooperation, of the channel an ultrafiltration flux towards the the diveristy of requirements from different count- islet compartment followed in the second half by ries should be considered. an equal flux in reverse direction from islets to A simple-minded uniform guide or regulatory i 2chanism will not work. blood. The mass transfer analysis is carried out A quality assurance program may work for a certain for an optimal geometry in which a U-shaped blood level of technology but may be hazardous to others. channel rounds closely a very thin islet compartment formed by a folded flat membrane. A Sometimes, technology is too advanced and standard quality assurnace does not pay for the danger. complete model of insulin release by this device Some of tne topics that require urgent study are: is developed and is compared whith in vitro data obtained with rats islets. internationally translatable but flexible nomenclature, pre-assessment of technology, man-machine Satisfactory kinetics is achieved with a polyacrylonitrile membrane used in hemodialysis. safety standard, testing method and devices, etc. But the model shows that the membrane hydraulic Some of the recent effort in Japan ,.11 be described. permeability should be increased by a factor of 10 to significantly improve the performance.

BE8-H.2 The Medical Device Regulation in the Federal Repub- lic of Germany, H. Hutten, Institute of Physiology, Johannes Gutenberg-University, D-6500 Mainz, RFG

The Medizingerateverordnung (MedGV) has been enacted on Jan. 1st, 1986. Its objection is the protection of users, patients and third persons against any avoidable risks that might be related with medical devices (m.ds). In separate paragraphs, the MedGV specifies the regulations both for market admission and employment of m.ds. The m.ds are classified in one of four classes. the 1.class consists of a list with 25 high-risk devices. The 2.class comprises all energy-powered implants. The remaining energy-powered m.ds belnng to the 3.class and all other m.ds to the 4.class. The most relevant regulations of the MedGV are: (1) m.ds of class 1and 2 need a type-related market admission )n which fur-

ther particulars are specified. (2) m.ds of class 1 and 3 have to be listed in an inventory. (3) Only specifically trained users are allowed to employ m. ds of class 1and 3. (4) m.ds of class 1 are only free for the first operation after a complete functional check and operating instruction have been made by the producer or supplier. (5) A detailed book must be kept for each class 1 m.d. (6) Acci- dents with personal injuries must be reported to the authorities. (7) m.ds supplied before Jan. 1st, 1986 had to undergo a safety check by authorized experts within two years if proper maiotenAnce in the past could not be proved by the owner.

e 1 244 1988 World Congress on Medical Physics amicBfeini&lical EngineeringSan Antonio, Texas Biomedical Engineering Scientific Papers

BE8-I.1 BE8-I.2 Current Status of Biomedical Engineering in the Segmented Polyurethanes as Basic Material Research and Development on Artificial Heart. for Cardiovascular Artificial Organs K. ATSUMI, University of Tokyo, 7-3-1 Hongo, T. Matsuda, Res. Inst., Nat'l Cardiovasc. Bunkyo-ku, Tokyo, 113, JAPAN Ctr., Fujishiro -dal, Suita, Osaka, Japan One of the crucial tasks for guarantee Artificial heart is aiming to construct the of the reliability of blood pump is to simulation model which substitutes cardiac ensure its system's antithrombogenicity. functions of biological heart. It is divided into This paper assesses to define determinant two modes - ventricular assist device(VAD) and factors and mechanistic aspects of total artificial heart (TAH) - by replacing the short-term and long-termantithrombo- cardiac functions partially or totally. genicity of segmented polyurethanes used At the present state, 300 clinical cases of VAD for blood pumps.The short-term anti- and 7 cases of TAH were experienced in the world thrombogenicity was found to be markedly and artificial heart is highly evaluated as the influenced by species of adsorbed plasma essential therapeutic method to cure the proteins. The adsprption of cell-adhesion deteriorated patients with severe heart allure. promoting proteins such as fibrinogen and In the symposium, the R & D on TAH is mainly fibronectingreatly enhanced platelet discussed. adhesion and aggregation. Both coagulation TAH is constituted from three parts - blood and complementsystems were activated pump, driving unit and energy source - . Until directly on polymer surfaces. Segmented now, air driven type of TAH were studiedin the polyurethanes were found to be relatively animal experiments and clinical applications, inert toward these biological response3. however, in future, implantable TAH will be On the other hand, longer - implantation implanted in human of which project has recently resulted in the formation of proteinaceous been started in the several eminent facilities. multilayer, which proceeds as a time-lapse The problems and solutions of biomedical process. The stability of multilayering engineering on TAH will be discussed in the was found to be the determinant for long- symposium which is formed from the two sessions, term antithrombogenicity. Besides these, one is material, blood pump, monitoringand blood flow rate and implantation period control and another is energy conversion and were alsomajordeterminants. Thus, energy source. The topics of the discussions will time-variant multi-factorial interactions be concentrated to the R & D on implantable TAH. determine system's antithrombogenicity.

BE8-I.3 BE8-I.4 EFFEaS OF HEAT DISSIPATION FROM MECHANICAL DRIVEN Development of A Multi-Function Control System ARTIFICIAL HEARTS ON BLOOD AND TISSUE COMPONENTS. for Artificial Hearts A SYSTEMATIC IN VITRO STUDY, H. Harasaki*,and T.Yamaguchi*, T.Shimooka, Y.Mitamura Nose, Cleveland Clinic Fdn., Cleveland OH44106 Section of Me&cal Electronics, Res. Inst. of Appl. Elect., Hokkaido Univ. Sapporo 060 Japan Management of waste heart imposes a new problem inherent to currently developed implantable energy We have developed a portable control system converters for artificial hearts. Effects of for pneumatic artificial hearts. This control predicted surface temperatures in coagulation, system enables automatic blood pressure control (APTT, PTT, TT) platelet % max., aggregation(AGR) without invasive blood pressure measurements as retension (RET), and DTG and PF4 release and well as conventional pumping methods. viability of various tissue cells were studiedin Blood pressures are measured indirectly in vitro. The results are summarized in the Table the control system. When driving pressure is below. changed gradually, the pressure when a diaghragm Epos. in the pump begins to move equals to arterial or 47°C Time 37 °C 41 atrial pressure. Waveform of drive pressure is 69.5 + 12.8* sec APTTIS in 41.0 t 7.7 49.8 t 2.2* regulated by electronic air pressure regulator and 27.712.1 sec PT IS min 19.4 t 0.6 20.8/ 1.8 30.6 15.2 *sec diaghragm movement is detected by optical sensor. 15 min 16.11 1.9 19.81 0.9* TT Based on the indirectly measured values, * 1 hr 77.818.6 37.216.6* 0 blrod pressures are controlled. The microcomputer MM 32.4 4 12.0 Z 1 hr 83.7 t 0.9 80.1 t 6.7 based control system maintains the blood pressures RET 195 t 52* neml BTG 15 min + AIR272 t 31 294 I 36 constant regardless of changes in peripheral ET-4 15 min + AIR 113 t 22 120 t 19 73 t n* Mara resistance or circulatory blood volume by changing *stetisticAlly sisnifivmt as com±ared to37°C drive rate,%systole, positive pressure and This data indicates that blood becomes less negative pressure. thrombogenic on the heated surface, andexplains In vitro experiments disclosed accuratebeat- out previous in vivo observationthat neointima by-beat estimation of arterial and atrial pressure formed on the heated surfaces is thinner thanthat and precise regulation of blood pressures.The on a un-heated surfaceof the blood pump. control system is useful for driving artificial Fibroblasts were more heat resistant than hearts optimally without the troubles of infection endothelial and smooth muscle cells and 91% and thrombus formation. viable even after 2 hr. exposure to 47°C. 2 C 5 245 1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas 1319medical Engineering Scientific Papers

BE8-I.5 BES-J.1 New Control Method of Artificial Heart DEVELOPMENT OF AN ELECTRIC MOTOR K. Imachi, I. Fujimasa, and K. Atsumi DRIVEN TOTAL ARTIFICIAL HEART, Inst. of Med. Electr.,Faculty of Med., W. Weiss, M.S.*, G. Rosenberg, Univ. of Tokyo, Hongo, Ph.D., Bunkyo-ku, A. Snyder, Ph.D., W.S. Pierce,M.D. Tokyo, Japan Pennsylvania State University, Hershey Medical Center, Hershey, Pennsylvania Total artificial heart(TAH) control methodis too primitive stage for the patientto enjoy a high An electric motor-driven total quality of life like a healthyperson. One of the artificial heart utilizinga roller difficulty to establish an ideal TAHcontrol screw drive is being developed. A method, is that there isno adequate detecting 3-phase brushless DC motor produces method of the cardiac output(CO) demandof the bidirectional rotary motion which living body. is As a step to the ideal TAH control translated to the rectilinear motionof method, a new control method of TAH duringthe the roller screw shaft. Pusher exercise was developed in plates at our laboratory. either end of the shaftalternately CO change curves for variuos grades of treadmill compress left and right bloodsacs exercise(TE) are obtained from controlgoat as a contained in pump cases locatedat time function(objective function curve:OFC).The either end of the motor housing. correlation between tradmill speed and physical Bjork-Shiley convexo concave tilting activity rate(PAR) which is definedas the time disc valves are used. A compliance average of upperward acceleration of the body,is chamber permits left-right stroke obtained as a linear function. volume These functions differences, which areutilized by the are memorized in a microcomputer. The TAH control motor controller to balance left-right procedure is as follows. 1)When the TEstart, its pump outputs. speed is detected as PAR by an acceleration sensor, Electrical power is deliveredto the 2)The microcomputer choose an OFC according to the implanted motor controller viaa pair of PAR. 3)TAH driving parameters of bothblood pumps coils, one subcutaneous andone are controlled predictively for the pump outputto -xternal. An external battery packwill fit with the OFC under a control algorithm operate up to 10 hours, whilean installed into the microcomputer.:ne system implanted backup battery willoperate up worked well in the an nalexperiments, and its to 45 minutes. effects on metabolism was alsoproven.

BE8-J.2 BE8-J.3 Solenoid Energy Converter of theNovacor Energy Source and EnergyTransmis- Clinical Left Ventricular AssistSystem sion for Artificial Hearts, V. P. Miller*, J. Jassawalla, H. Chen, Poirier*, C. Sherman, D. LeForge, P. Portner. K. Dasse, Novacor Medical Thermedics Inc., Woburn MA01801 Corporation, Oakland, California, USA Artificial hearts and heart The Novacor Left Ventricular Assist assist System (LVAS) systems require significantpower uses a balanced pulsed solenoid electromechanical levels which cannot be met energy converter for blood pump actuation. with im- Under plantable power sources. The two development since 1970, the current version of this most widely used energy sourcesare simple lightweightdriver hasessentially air and electricity. only two moving parts The air sys- -- the identical armature/ tems, generally spring assemblies used for short -- coupling directly to the term, require an external air blood pump pusherplates. When energized, com- the pressor while the electricsystems two solenoid armatures close rapidly, efficiently require chemical storage batteries. convertingelectrical input power into mechanical energy stored Transmission of energy from theex- in the "decoupling" springs. The ternal sources can be accomplished springs then deliver their storedenergy to the by two means. pump at a rate governed by the physiologic load. For electrical energy transfer, transcutaneousenergy The balanced solenoid closure resultsin no reaction transfer systems which utilizetwo torques, gyroscopic effects, or momentum coils of wire to inductively transfer to the body. Because the entire LVAS couple the energy into the bodyare used. is symmetrical, there areno unbalanced loads on In this form, energy is the energy converter frame or bloodpump housing. transferred without penetration of theintegu- The system has been extensively tested invitro ment. For transfer of airor elec- andin vivo,in multiyear endurancetestsand trical energy, percutaneousenergy animalexperiments as longas nine months. A transfer systems are utilized, multicenter clinical trial of the NovacorLVAS has which can becomean integral part been ongoing since 1984. System efficacy has been of the body. demonstrated in successful bridge-to-transplant applications of up to 90 days. r rs 4 k_so

246 1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas Biomedical Engineering Scientific Papers

BE8-K.1 BE8-K.2 Fluid Mechanical Analysis of Blood Flows Optimum Design of a Compact Centrifugal with an Intra-Aortic Balloon Pumping Blood Pump: Casing Configuration and Pump T.K. Hung and H.S. Borovetz, University Performance, S. Tanaka*, K. Ito, K. Yama- of Pittsburgh, T.E. Natan, Jr., Carnegie- koshi, A. Kamiya, K. HoriuchiA & M. Kusa- Mellon University, Pittsburgh, PA 15261 kabei Res. Inst. Appl. Electr., Hokkaido Univ., Sapporo, & ATerumo Inc., Tokyo, For three-dimensional blood flow in the descending Japan. aorta with an intra-aortic balloon counter pulsation, the radial axis, r, for the Navier-Stokes For the use of .33ctracorporeal circulation, we equations in the cylindrical coordinates, (r,e,z), have previously developed a compact centrifugal is transformed to a dimensionless axis: blood pump driven directly by a miniature DC servomotor. Through in vitro tests, it was found r - b(0,z,t) R + 1 that the configuration of the impeller having a(0,z,t) - b(0,z,t) the blade angles at the inlet of 20 deg and at where the radii of the balloon (b) and aorta (a) the outlet of 50 deg was the most appropriate and their motions are the prescribed boundary for this pump with respect to the pumWIng capac- conditions for the computational flow simulation. ity and the decrease in hemolysis."In this Local acceleration components were employed in the study, the effect of the volute casing on the determination of the transient pressure field from pump performances was evaluated to determine an the divergence of the Navier-Stokes equations.The optimum configuration of the pump casing. It was pulsating spiral flow processes were analyzed by demonstrated that the pumping capacity and the correlating the longitudinal velocity pr ';les, degree of hemolysis produced by the pump with the secondary flows, and the pressure ana shear volute casing were significantly improved at fields with the deflation and inflation of the high flow rate region compared with those by the balloon during a cardiac cycle. pump without volute casing. It was suggested The study was supported by an NIH Grant HL33934 through this study that the optimum design of from the National Heart,Lung,and Blood Institute. the casing configuration should be made accord- The Cray X-MP supercomputer at the Pittsburgh ing to the range of flow rate required. Supercomputing Center was used for this in.esti- gation. Reference: 1) S. Tanaka et al, Trans. ASME, J. Biomech. Eng., 109(3), 272 (1987)

BE8-K.3 BE8-K.4 Development of Nutating Centrifugal Blood Optimization of Pulsatile PumpFlow Pump, T.Akamatsu, Kyoto University, Kyoto Parameters, R.Fumero, F.M.Montevecchi, 606,Japan S.Scuri°, Politecnico di Milano, Italy.

Conventional centrifugal blood pumps have weaknesses The differentiation between effective and of infection through shaft seal for direct motor ineffective forms of pulsatile flow during drive and blood degeneration in journal bearir,-; for cardiopulmonarybypass has beenoften stressed. closed magnetic coupling drive. A centrifugal pump Factors such as ejection phase duration, peak flow, with a nutating(not rotating) impeller, so called etc, must be chosen in order to achieve the energy Tea Spoon Pump, can overcome these weaknesses. This content of flowwave whichmaintains capillary pump is allowed to have only one impeller for its patency and, on the other hand, to avoid -ixcessive mechanism. Therefore its pump efficiency was not mechanical trauma on alood cells. The influence of so high. But by our followir 4nventions it is much arterial line and aortic cannula, whichmodify improved; (I) annular pump rt.._ surrounded with considerably the pulse, must be considered too. inner guide wall and outer casing wall,(2) spiral A pneumatically driven and microprocessor inlet passage, (3) well-contoured outlet passage controlled pulsatile pump for pediatric and diffuser. Rectified flow is obtained as if a extracorporeal circulation has been developed. In piston (impeller) pushes blood in an endless cylin- order to optimize flow parameters, both energy loss der (annular pump room),and the pressure difference in the aortic cannula, which is a reliable index of between outlet and inlet is sustained by centrifugal blood damage, and output energy past the cannula force of rotating blood flow. mustbe evaluated. For this purpose, a finite At present, the pump efficiency is raised up to 45% differenceprogram has beenused to calculate compared with the previous 32%. Undesirable hemu- transient flowand local pressure drops in the lysis in the previous pump was larger than that in arterial line. The program has been also validated the roller pump. But in the present improved pump, by comparing numerical results to sperimenlal we can expect the lower hemolysis. Pulsatile flow measurements assessed in e simple mock circulation. can be obtained by the on-off run of the motor and Numerical simulations have been performed under the the addition of a check valve at the outlet of the sameconditions of mean flow rate for different pump. pulse rate (60,70,80b/min) andejection phase duration (40,50,60% of pulse period). The results haveprovided useful directions for optimum choice of systolic to diastolic time ratio. 267 1988 World Congress on Medical Physics and Biomedical Engineering 0 San Antonio, Texas 247 Biomedical Engineering Scientific Papers

BE8 -K.5 BE8 -K.6 STUDIES ON ELECTRONIC CONTROL PNEUMATIC DRIVING Mechanical Studies on the Left Ventricular Recovery SYSTEM FOR ARTIFICIAL HEART Process from Induced Heart Failure by Cardiac Yang Zi-lain, Zhu An -ping, Ln song -fang, Assist. T.Nakamura*, K.Hayashi, J.Seki°, H.Noda°, Han yen, Huang Wen-mei, Wang :Ciao-Nig T.Nakatani°, H.Takano° and T.Akutsu°.Hokkaido Shan Hui-min University, Sapporo, and °National Cardiovascular Dam. of Meat-JAI Ecaccestet renter, Osaka, JAPAN. lestitete et Bette Mtd.cat S4w1.4% Chime AoiLecT Of Medical Sc leaks Pneumatically driven, diaphragm-type leftven- Beijing China tricular assist devices (LVADs)were chronically Highly reliaile electronic control and pn .instic driiing system for arti- implanted between the left atria and the descending aortas of goats. ficial heart has heen ue'ieloped, It was colsisted of tao subsystems which Myocardial infarction was in- are electronic eortrol subsystem and pnearnatie power supply. The compres- duced in the anterior myocardial area by the multi- sed air adjusted from 0--.100 mmHg was supplied by compressor or by com- ple ligations of coronary arteries. Left ventric- pressed nir station. The ,aeuum pressure ad,usted from 0--00 mmHg was ular diameter, segment length and wall thickness supplied by a %aeutm pump. were measured with sonomicrometers implanted in the In the electronic co:troller. the dro tig stval isrecei,ed from either ventricle. Bulk and regional mechanical work of ECG for s)cchronous drive with the heart beats or signal generator for asy- the left ventricle were calculated fromthese nchronous. The coltroller moittors ECG 313431 le.ei avi also the dro so; dimensions and the left ventricularpressure. pressure to check the (wactloa of pressure re;alatore satuutt. regulator and LVAD pumps were removed when the cardiac function magnetic akes. seemed to be recovered. Cardiac and hemodynamic After a series in iitro tests, the AH elm,ing system has been used in data were obtained until around 1 month afterthe animal experiments. Darius the period of animal caytertmeat. the AH dri%- pump removal. ing system has ,corked all together for mor than 500 :tours and continuous- The LVAD reduced the bulk and regional mechani- ly moiled for 227 hours, The functions of the AH dri,ing system was sho-- cal work. In cases that the hearts were recovered weS to be normal add stable, so it is suggested to be used in AH eXperime2 tal study successfully, the difference between thebulk work done during pumping-on and -off decreased gradually, concomitantly with recovery, and theMO pumps could be weaned off within 1/2 month. In some of these cases, the ischemic regionswere healed by the LVAD pumping and became to be able to dome- chanical work.

BE9 -A. i. BE9-A.2 PLASMA MACROMOLECULES IN HUMAN AORTIC INTIMA F tacellularPathwaysforTransportof Low Density Eispeth B. Smith; Dept. of Chemical Pathology, Lipoprotein,J.Rutledge, F. Curry,P. Davis and J. Foresterhill, ABERDEEN, AB9 2ZDScotland. Lenz, University of California, Davis, CA

Endothelium transports macromolecules into the Low density lipoprotein (LDL) istransported across endothelium intima. In normal adult intima the concentration by receptor-mediated and vesicular pathways. We have evaluated paracellular pathways of of low density lipoprotein (LDL-Mr 2x10 ) is twice transport the plasma concentration and for most macromole- ofLDL in highpermeabilitystates in isolated -cules relative concentration is directly related perfusedfrogmesentericmicrovessels. LDL wasfluorescentlylabeled to Mr. High concentrations are not necessarily withDiI and20,000 MW dextran,amolecule pathogenic; pathogenicity probably results from whichlacksreceptor-mediated processing within the intima, and this may be transport, was labeled with FITC. By alternating sets oflightfilters, we influenced by haemodynamic factors. LDL may be were able to image these internalized by macrophages, subjected to extra- fluorescent molecules almost simultaneously. cellular proleolysis, or tightly bound. Fibrinogen Fluorescence intensity of the microvessel and is accompanied by degradation products (FDP) that surrounding tissue was measured by photometer. The appear to be derived from fibrin. This implies endothelium was perturbed with the calcium ionophore, A23I87, which continous conversion of fibrinogen to fibrin, and effects a transient increase in permeability. :ontinuous fibrinolysis. Release of tightly bound LDL by incubation with plasmin is correlated with Our studies have shown that permeability of LDL the amount of cross-linked fibrin degraded and dextranhavesimilar magnitude and time course aftertheinfusion (r=0.688, p<0.01). These complex interactions must of A23187 which suggestssimilar pathwaysforthesemolecules. be dependent on factors such as residence ti and Inaddition,I.DL transporthasbeen mixing which will, in turn, be influenced by showntoincreasewithincreased haemodynamic forces. capillarypressure. Becausewater flowincreasesas Some proliferative lesions capillary accumulate a core of extracellular LDL-derived pressureincreases,thiscouplingof LDL transport to lipid, whereas others in the same aitery do not.' water flow suggests a paracellular pathway for LDL transport. Could haemodynamic factors influera oxygen and lactic acid concentrations deep in the lesion, and hence the release of 1)somal proteases? 268

248 1988 World Congresi;on Medical Physics anc' Biomedical EngineeringSan Antonio, Texas Biomedical Engineering Scientific Papers

BE9-A.3 BE9-A.4 Optimal Design of Experiments to Quantify Effect of Hemodynamics on the Rate of LDL Transport and Degradation in the Arterial Walt, Transport in Lesioned and Nonlesioned Aorta E.D. Morris, G.M. Saidel, and G.M. Chisolm*, S.A. Berceli,* S.J. Durham, D.P. Showalter, Cleveland Clinic Foundation and Case Western V.S. Warty, S.K. Tanksale, H.S. Borovetz Reserve University, Cleveland, Ohio 44106, USA. LDL transport and catabolism were studied in A dynamic, one-dimensional model is used to depict lesioned and uninvolved abdominal aortae of 10 the transport and metabolism of low density lipoprotein WHHL rabbits and in abdominal aortae of 10 NZW (LDL) in arterial tissue. The model includes decreas- rabbits perfused in vitro for 20 hours under ing concentration of tracer in plasma, permeabilities at "normotensive" (P 70/40 wm Hg, Q - 50 ml/min) both luminal and media-adventitial boundaries; or "hypertensive" conditions (P - 100/70 mm Hg, diffusion and convection thr nigh interstitial space and Q - 100 ml/min). Freshly ey.cised arteries were degradation by smooth muscle cells of the media. perfused with normothermic i:abbit serum to which Model parcmeters representing these processes can be was added LDL doubly labelled with 1131 andI125- estimated from transmural concentration profiles of tyramine cellobiose (I125 -IC). Post-perfusion labeled LDL generated in vivo across the arterial wall. tissue biopsies were analyzed for intact 1131-LDL The sensitivity of the model output to changes in and accumulated I125-TC-labelled degradation parameter values was examined. An optimal experi- products. ment design was defined to be one that maximized Under normotensive conditions, the plasma LDL precision of the estimates. Three designs were consid- flux in WHHL lesioned aorta was 4.28 nl/mg/hr, ered: 1, a single bolus injection of *I-LDL, 2) sequen- 1.5 times larger than in uninvolved WHHL aorta tial injections of *I-LDL and **I-LDL (*I and **I are (p < 0.001) and 2.9 times above that in NZW aorta distinguishable isotopes of iodine), and 3) a simul- (p < 0.001). Under hypertensive conditions, the taneous injection of *I-LDL and **I-TC-LDL; TC is an plasma flux in WHHL lesioned aorta was 7.54 iodinated tyramine cellobiose adduct that accumulates nl/mg/hr, 1.7 times larger that. in univolved WHHL in cells after LDL degradation and is not removed. The aorta (p < 0.001) and 2.5 times above that in NZW optimization procedure compared the three designs for aorta (p < 0.001). The study suggests that a wide range of circulation times.Our results hypertensive hemodynamic conditions exacerbates prescribe a simultaneous injection of *I-LDL and **I- the already elevated flux of LDL into TC-LDL and a circulation time of 18 hours. Funded by atherosclerotic tissue. NIH (HL29582)

BE9-A.5 BE9-B.1 Chronic hypertension decreases the in vivo albumin Cultured Endothelial Monolayers: A Model for uptake by the rat aorta, A. Tedgui*, J. Belmin, J.B. Michel and Transport Studies, R. Bizios*, R.S. Cargill, L. Juan, INSERM Unite 141, Hopital Lariboisiere.75010 Paris, R.D. Iveson, and L.A. Holleran, Department France. of Biomedical Engineering, Rensselaer Changes of protein transpt.:1 across the arterial wall may facilitate Polytechnic Institute, Troy, New York, atherogenesis. We have studied the effects of chronic 12180-3590, U.S.A. hypertension and its therapy by a converting enzyme inhibitor (CEI), on the in vivo albumin uptake by the aortic media of hypertensive rats (2 K-1 C) (HR), 8 weeks after application of the clip. Blood Cultured endothelialmonolayersprovidea model pressure of some HR was lowered to a normal level by injection of iur investigating the transport properties of nicardipine prior to the albumin uptake measurement, in order to endothelium under experimental simulation of separate the effects due to hyperpressure from those due to various conditions (such as hyperoxia and anoxia) structural changes. which induce and/or contributeto cell injury. The transmural distribution of the relative albumin concentration Measurements of mass flux across the monolayers In across the wall was obtained using a frozen serial sectioning vitro provide aquantitative index ofthe time technique (10 gm), 90 mn after injection of 1251albumin. course of cell injury. Hypertension decreased the mean medial relative concentration Bovine pulmonary artery endothelial monolayers (MRC) :1.53 ± .38 % (n=8) in HR and 2.02 ± .65 % (n=10) in control were grown to confluence on polycarbonate (p < 0.01). This decrease was even larger in HR with acutely lowered (0.8 pm pore diameter)membranes. Transport of blood pressure : 1.16 ± .32 % ,n=7 (p < 0.01),In the CEI treated studied a 4 gm percent animals, MRC (1.56 ± .41 %, n=10) remained lower than in control albumin was under (p<0.01). The medial thickness was increased in HR (120 ± 14.1 concentration gradient at 37"C and over a 2 hour gm, n=8) as compared trtontrol (92 ± 6.5 gm, n=10, p < 0.01), and period. Abluminal albuminconcentrations were was normalized by CEI (95 ± 8.9 gm, n=10). The collagen density, determined by the Lowry assay. measured by histomorphometry, was increased in HR .11.5 ± .8 %, Permeability coefficients of the endothelium were n=15 vs 8.2 ± .7 %,n=10 in control (p < 0.01), but remained calculated fripm the mass flux data and were in the increased after CEI (11.9 ± .9 %, n=10). order of 10-J cm/ sec. The results indicatetthat Structural changes of the aortic wall induced by chronic inadditionto endothelium other subendothelial hypertension seem to be responsible for the important decrease of tissues and/or cells contribute significantly to the medial permeability to albumin in HR. The increase in the the resistance to transport across the blood collagen density of the media observed in HR was not reversed by a vessel wall. (Supported by the Whitaker 4 weeks CEI treatment. This accounts for the maintained decrease Foundation). of the albumin uptake in CEI treated animals with a normal pressure and thickness. 269 1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas 249 Biomedical Engineering Scientific Papers

BE9-B.2 BE9-B.3 Mass transport properties of different Time dependent filtration in a blood vessels. M. J. Lever* and porous elastic cylinder-Applica- M-T. Jay. Physiological Flow Studies tion to Water transport in the Unit, Imperial College, London,5W7 2AZ artery wall, Girija Jayaraman, Centre for Atmospheric Sciences, We have re-examined the nature of the I.I.T., Hauz Khas,New Delhi-16. variation of the transport of materials between the blood and vessel walls at The time dependent filtration of water different sitesin the circulation. This through the wall of a soft porous elastic suuject has beenexamined in many previous cylinder is studied theoretically. The ar- studies and differences found in the rates and tery wall is modelled as a mixture of an extent of traceruptake atdifferent sites incompressible liquid and porous elastic have been attributed to many factors, solid whose permeability is a function of including differences in the shape or turnover the applied pressure. The luminal pres- of endotheial cells,in the properties of the sure is considered to be harmonic in time glycocalyx, in the number of plasmalemmal and the matrix is assumed to be elastical- vesicles and in local blood pressureand ly and hydraulically isotropic. Analytic flow. solutions are presented for some simple Our studies have included measurements of both relations between pressure and permeabili- theshort-termandsteady-state uptake of ty but numerical results are found for tracers by different vessels in vivo and also more general relations. Special features of the equilibrium uptake of tracers by those of the pressure distribution across the vessels and their hydraulic conductivities in wall are discussed for the important par- vitro. Steady-state tracer levels in ticular case when the characteristiccon- different vessels in vivo appear to be related solidation time is large compared to the in part to thedistribution volumes of the characteristic load time. tracers in the wall tissue and in part to the permeabilities of the different layers of the vessel wall.

BE9-B.4. BE9-B.5 Microarchitectural Changes of Human Aortic The study on the role of hemodynamic forces in Bifurcations with Reference to Atherogenesis, the development of atherosclerotic lesions, Y, Yoshida*, T. Oyama, T. Yamane, M. Mitsumata, N. &Licata; K. Kawamura and S. Takebayashi, T. Yamaguchi', and G. Ooneda 2, Yamanashi Medi- Fukuoka University, Fukuoka, Japan cal College, 409-38, Yamanashi, 1 National Cardicascular Center Research Institute, 565 Osaka, 2 Geriatrics Research Institute and At the present time, there are conflicting hypo- Hospital, Maebashi 371, Japan theses concerning the hemodynamic mechanisms of the atherogenesis. One isFry's "shear stress hypothesis"; the other Caro's "boundary layer Apical and outer wall intimas at branchings of in- diffusion hypothesis ". In order to clarify the ferior mesenteric arteries from aortas of human role of hemodynamic forces in the development of autopsy cases were investigated. The outer wall human atherosclerotic lesions, we have studied the intima, a low shear stress region, was thicker than localization of atherosclerotic lesions and mor- the apical intima, a high shear stress region,:xce- phological changes of endothelial cells over the pt in cases younger than one month. Deposit of 13- atheromatous plaques. Early atherosclerotic le- lipoproteins in the outer wall intima increased sions occurred predominantly at the outer walls of with age, but not that in the apical intima.Elect- bifurcation8 and distal to the center of the inner ron microscopic studies revealed th.t collagen fib- curvature at curves, where wall shear stress is ers increased markedly in the apical intima in the considered to be low. In contrast, the flow divi- third decade. Among the dense collagen fibers, int- der of bifurcations and the outer curvature of the imal smooth muscle cells (SMC) exhibited contracti- bends, presumably high shear regions, were free of le phenotype. The outer wall possessed mucinous the lesions. Endothelial cells over the athero- intima accompanied with synthetic SMC. Atheroscle- sclerotic lesions were significantly less elon- rotic rabbit intimal SMC cultivated on type I col- gated and had more prominent microvillous projec- lagen gel increased cAMP production to change pheno- tions as compared with those in lesion-free areas. type from synthetic to contractile and to suppress These results suggest that focal hemodynamic for- DNA synthesis. As a result, blood flow dynamicsmay ces, presumably low shear stress, are implicated induce cellular and microarchitectural changes in in the development of the atherosclerotic lesions vessel walls which are favorable or resistantto as related to morphological changes of endothelia. atherogenesis.

250 2 7 0 1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas Biomedical Engineering Scientific Papers

BE9-C.1 BE9-C.2 LOCALIZATION AND SIZE OF EARLY SUDANOPHILIC Flow Patterns and Distribution of Atherosclerotic LESIONS IN CHOLESTEROL-FED RABBITS Lesions in Human Coronary and Cerebral Arteries, Margot R. Roach*, Christine Zeindler, and Ralph Kratky T. Karino*, S. Mabuchi and T. Asakura, Montreal The University of Western Ontario General Hospital, Montreal, Quebec, Canada. London, Ontario, Canada. N6A 5C1

To investigate the potential role of fluid mechan- We have developed a method to cast aortas of rabbits with ical factors in the localized genesis and develop- Batson's methyl methacrylate.Casts made at 100 mm Hg ment of atherosclerotic lesions in man, the exact shrink to a size equivalent to that at 78 mm Hg (diastolic anatomical locations and the flow patterns at such pressure). The ti4bue is digested off with 0.1 N KOH, and sites in the human coronary and cerebral arteries then analyzed with the scanning electron microscope. were studied in detail by means of flow visual- Segments of cast are mounted on a spit so all sides can be ization and high speed cinemicrographic techniques viewed.Composite photographs are analyzed to show the using isolated transparent coronary and cerebral location and size of the plaque which is identified by: i) arteries prepared from humans postmortem. It was small fragments of material, probably elastin, ii) cells, and found that in both coronary and cerebral arteries, iii) a rough surface. The length of the plaques is measured atherosclerotic lesions were localized exclusively at 10° intervals on a polar grid. at the outer walls (hips)of major bifurcations The earliest le..ions, after one week on a 2% cholesterol and T-junctions, and at the inner wall of curved diet, were lateral. By two weeks they were spreading more segments where flow was either slow or disturbed distally.The area of the lesion was related to the area of with the formation of slow recirculation and sec- the branch in any one rabbit.Since the size of the branch ondary flows and where wall shear stress was low. is roughly related to the flow, this suggests that branch In no instance, were atherosclerotic lesions found flow determines the size of the lesions. at and around the flow divider where fluid velocity and wall shear stress were high and where the Supported by the Heart and Stroke Foundation of Ontario. formation of early atherosclerotic lesions have been observed in experimental animalsfed high- cholesterol diets. The results indicate that there is a strong correlation between the sites of low flow velocity (low wallshearstress)and the preferred sites for the genesis and development of atherosclerosis in man.

BE9-C.3 BE9-. Localization of Human Atherosclerotic Lesions Ath4erogenesis as an adaptive process Lope J. Fredrick Cornhill*, Edward E. Herderick wrong? The roles of hemodynamics and arterial The Ohio State University, Columbus, Ohio geometry, M. H. Friedman*, The Johns Hopkins University Applied Physics Laboratory, The topographic distribution of atherosclerotic lesions has Laurel, Maryland 20707 USA been a prime factor in suggesting a role for hemodynamic forces in atherogenesis. Probability-of-occurrence maps of the human Hemodynamic measurements made in our laboratory aorta for sudanophilia and raised lesions and average thickness and others' show that the fluid dynamic shear maps were created using image processing (n=237; B & W; 15-34 stress at the vascular wall is remarkably simi- years). In the thoracic aorta, high probabilities of lar in different segments of the human arterial sudanophilia were associated with the dorsal surface (from 30% tree. The possibility that this similarity to > 50%) while the ventral surface was virtually spared (< could reflect adaptive processes in the wall is 10%). The regions of the highest probability of sudanophilia consistent with a model of intimal thickening (> 50%) were located midway between the origins of successful recently developed to explain the complex pairs ofintercostals. The intercostal flow ," were dependence of intimal thickness on mural shear spared. In the abdominal aorta, the dorsal distribution of seen in a collection of human arterial bifurca- sudanophilic lesions observed in the thoracic aorta continued tions. According to this model, the intima with regions of high probability being located midway between eventually becomes thickest at sites exposed to the lumbar ostia (> 3011. Regions of high probability were relatively low or oscillatory shear stresses. associated with the inflow tracts of' the celiac, superior and If the thickening process is mediated by sensed inferior mesenteric (IM), and right and left renal ostia (> shear and the regional vascular geometry is 50%). Regions distal to all ostia and the apex of the aorto- such that the shear is locally reduced, then iliac bifurcation were spared (< 10%). A region ofhigh inappropriate focal thickening could result. probability was also observed along the ventral aortic surface Thickening would be further enhanced if, in extending from the level of the renal ostia to the origin of contrast to the situation in straight segments, the IM. The association of the disease with specific geometric intrusion into the lumen results in a further entities suggests that hemodynamic forces probably play a role reduction of wall shear. This situation would in the localization of some of these lesions; however,this be exacerbated in vascular segments whose role is probably played in concert with many other important specific geometric parameters are such that localizimn biochemical, cellular, mass transport, and wall shear is inordinately reduced at susceptible foci. structural factors. 2, 7 1

1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas 251 Biomedical Engineering Scientific Papers

BE9 -C.f BE9-D.1 WALL STRESS AS A MAJOR CONTRIBUTOR TO Platelet Adhesion to Endothe141 ATHEROSCLEROSIS. M. Thubrikar*, R. Eppink, Molnolayer in vitro, T. Takano , S. Nolan, University of Virginia, Cnar- C. Mineo anU-IC-ffighida, lottesville, Va. Department of Microbiology and Molecular Pathology, Faculty of Atherosclerotic lesions generally occur at Pharmaceutical Sciences, Teikyo arterial branch sites, which are also the University, Sagamiko, Kanagawa sites of stress concentration produced by 199-01, Japan. intra-arterial pressure. Wall stresses were determined at the bovine circumflex Platelet adherence and subsequent aggre- coronary arterial branch using finite ele- gate formation on damaged vessel walls may ment analysis. The arterial branch was be fundamental events in the initiation of modeled as intersecting cylindrical shells an atherosclerosis. incorporating variations in the branch To study the mechanisms of platelet adhe- geometry, thickness, and material proper- sion to endothelial injury, we developed a ties. In rabbits fed a high cholesterol novel model in tissue culture system. The diet, we explored whether atherosclerotic porcine aortic endothelial cells cultured lesions could be inhibited by reducing wall on gelated type I collagen supported by a stress. To reduce the wall stress, peri- dacron sheet were injured according to a arterial casts were placed at lowered removal of tefi9n rings. Rabbit platelets arterial pressure at the left renal arteri- labelled with '1Cr were counted to measure al branch and at the aortic bifurcation. the amount of platelet adhered to injured The results indicate that in the bovine part of the cell monolayer quantitatively. coronary artery the wall stress was 3-4 In this system it was found that both times greater at the branch than in fibronectin and fibrinogen were essential straight segments. In rabbits, athero- to platelet adhesion in addition to colla- sclerotic lesions were inhibited at selec- gen. This in vitro model should be useful ted sites only when the wall stress was to elucidate the mechanisms of platelet reduced by cast placement. Hence, wall adhesion, which may be connected with stress appears to be a major contributor atherogenesis induced by blood flow re- to atherosclerosis. lated stresses.

BE9-D.2 BE9-D.3 Enhanced LowDensity Lipoprotein (LDL) Receptor EFFECTS OF SHEAR STRESS ON ENDOTHELIAL CELL Expression in Cultured Bovine Arterial Eidothelial PROTEIN SYNTHESIS, S. G. Eskin*, K. J. Cells (BAEC) Exposed to Elevated Fluid-Imposed Angelides, B. Wible, L. T. Navarro, L. V. Laminar Wall Shear Stress. EA Sprague., SA Logan, McIntire; Depts. of Surgeri and Physiology EH Edwards, CJ Schwartz and RM Nerem. University and Molecular Biophysics, Baylor College of of Texas Health Science Center, San Antonio, Texas Medicine, Dept. of Chemical Engineering, and Georgia Institute of Technology, Atlanta, Rice University; Houston, Texas, U.S.A. Georgia Our previous studies have demonstrated that cultured Cultured bovine aortic endothelial cells were BAEC exposedto a high laminar wallshear stress subjected to steady shear stress (25 dynes/cm2) (HS, 30 dyne/cm2) exhibit an enhanced in an in vitro flow system, to investigate bio- receptor-mediated endocytosis of LDL relative chemical changes resulting from hemodynamic forces. to either low (LS,<1 dyne/cm2) or no shear (NS) After shear stress periods from 6 to 24 hrs., the stress. The objective of this study was to cells were labeled with 0.25 mCi /ml 35S- methiorine determine whether this enhanced LDL endocytosis for 30 min. followed by a cold chase with complete was due to a change in LDL receptor affinity and/or medium for 30 min. Radiolabeled synthetic proteins number. Using new largecapacity parallel plate were separated by 2 dimensional gel electrophore- viscometersconnected in series, confluent BAEC sis. Isoelectric focusing gels were run in the on polyester sheetsare exposed to HS, LS, or first dimension under nonequilibrium conditions. NS in the presence of circulating lipoprotein SDS polyacrylamide slab gels with a 4-15% linear deficient medium to maximize LDL receptor gradient were run in the second dimension. Autora- expression. After 24 hr, the sheets (98 cm-) diograms of gels from shear stressed cells show are removed and subdivided for 4°C binding studies differences in location and ihcensity of proteins using increasing concentrations of 125I-LDL. BAEC synthesized. Differences to be analyzed are pro- exposed to HS exhibited an increase in LDL receptor teins from shear stressed cells which are not pre- number without a changein affinityrelative to sent in controls, indicating the synthesis of new cells exposed to LS or NS. These data were proteins, and clear changes in spot intensity confirmed in similar binding studies using a between experimental and control cells which we radiolabeled monoclonal antibody (IgG-C7) specific have observed. for the LDL receptor. Our results indicate that shear stress signals are recognized and transduced to influence important cellular functions including LT, receptor expression. 2 7 2 252 1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas Biomedical Engineering Scientific Papers

BE9 -D.4 BE9-D.5 Effect of Hydrodynamic Stress on Intracell- MECHANISM OF HYODYNAMIC SIEAR Siff_SS STIMILATION OF ular Metabolism of Human VMbilical Vein EtCOTIELIAL CELLS Endothelial CeZZs, L. V. Mclntir t, M. U. Nollert, S. Diamond and S. G. Eskintp, Bio- J.A. Frangos'and A. Biagyalakshmi medical Engineering Laboratory, Rice Uni- The Pennsylvania State University, Department of Chemical versity, Houston, Texas, USA, 77251 and Engineering, University Park, PA Department of Surgery, Baylor College of MedicinelP, Houston, Texas, USA. Blood rheology is a majcr modulator and effecter of vascular cell response.One of the aspects of blood rheology that appears to be important is fluid shear stress on the endothelium. The role of steady and pulsatile shear stress on Changes in cell morphology, cytoskeletai, pimcytosis and rnetabolisn are well dommented. Using human umbilical vein endothelial cells (HUVEC) intracellular metabolism was investigated using prostacyclin production as a sensitive marker of the a parallel plate flow chamber. HUVECs were cul- endothelial cell response, we investigated the nechanisn of tured as monolayers which formed one side of the shear stress stimulation of endothelial cells.The onset of shear stress leads to a transient burst in frostacyclin chamber. Cells were incubated with 14C-labeled arachidonic acid prior to flow exposure. HPLC production which plateaus to an elevated steady state level. coupled with scintillation counting was employed Increasing levels of shear lead to increased long -term production rates. to quantitate the kinetics of metabolite forma- Chelation of extracellular calciun by BM EGTA inhibited the steady state phase, while chelatiai of tion. In some cases, specific metabolites were confirmed using radioimmunoassay. Stresses in intracellular calciun by 30 }II quin -2 inhibited both initial the physiological range (0-30 dynes/cm2) have and steady state phases.Inhibition of intracellular dramatic effects on HUVEC arachidonic acid me- calciun mobilization by 1143-8 inhibited only the initial phase.The calciun channel blacker nifedipine had no effect. tabolism. Both the magnitude and rate of change These results suggest that the onset of shear leads to of stress appear to be important. Possible biochemical mechanisms of stress signal trans- intracellular nx while extracellular duction and implications for protein synthesis calciun influx occurs in the long -term, but not via voltage sensitive channels. regulation are presented. This work was supported in part by NSF-EET 8708692 and PHS/Bianedical Research Support Grant.

BE9-E.1 BE9-E.2 Cytoplasmic Calcium Response to Fluid Shear RESPONSES TO SHEAR STRESS OF A CULTURED VASCULAR Stress in Vascular Endothelial Cells. JAndo' ENDOTHELIUM and A Kamiya. Research Instituteof Applied L.Tamisier, A. Caprani, P. Pdronneau*. Electricity. Hokkaido University. Sapporo 060 INSERMU.256, HOpital Broussais, Paris, France. Japan Endothelialcells from human umbilicalvein, cultured under hydrodynamic stress are endothelialcellsmodulatetheirstructureand Vascular investigated in ordertospecify the role played functions in response to changes in hemod'namic forces by the flow.The stress conditions are applied such as fluid shear stress. We have studied how endothelial as soon as the beginning of the culture in order cells perceive the shearing force generated by blood flow to beinsteady state whenthe celllayeris and the substances(s) that may mediate such a response confluent. Electrical impedance measurements We identify cytoplasmic free calcium ion (ca++), a major of the layer are performedtocharacterize the component of an internal signalling system, as a mediator of different structures of the layer and to detect the cellular response to fluid shear stress. theirmodifications. Thus we obtain a value Culturedmonolayersofbovineaortic endothelialcells of the junction resistance and of the membrane capacitance,andan electricalcharacterization loaded with the highly fluorescent Ca++-sensitive dye Fura 2 of the extracellular matrix and of the were exposed to different levels of fluid shear stress in a macromolecular layer adsorbed on thecells. specially designed flow chamber. and simultaneous We follow the time variation of these parameters changes in fluorescence intensity. reflecting the intracellular during a shear stressstep and after comeback freecalciumconcentration((Ca++1) were monitored by totheinitialvalue.We show thatallthese photometric fluorescence microscopy.Application of shear parametersaremodified bytheshearstress stress to cells by fluid perfusion led to an immediate several- variation with high time constantvalues(several fold increase in (Ca++) within 1 minute, followed by a rapid hours).In particular, we show at the junction decline and finally a plateau somewhat higher than control level an increase of the permeability that leads levels during the entire period of thestressapplication. to an obstruction by the macromolecules of the medium. Repeated application of the stress induced similar peak and plateaulevelsof [Ca++) but at reduced magnitudesof response. 2 7 3

1988 World Congress on ,iedical Physics and Biomedical EngineeringSan Antonio, Texas 253 Biomedical Engineering Scientific Papers

BE9-E.3 BE9-E.4 Relation between Wall Shear Stress and Micro- Viscoelastic Property of Cultured filament Bundles in Endothelial Cells in Canine Porcine Aortic Endothelial Cells Coronary Artery in vivo, A.Kitabatake*, measured by the Nicropipette J.Tanouchi, M.Uematsu, K.Ishihara, K.Fujii, Technique, M. Sato*, N. Ohshtmaand Y.Yoshida, H.Ito, M.Hori, M.Inoue, T.Kamada, R. M. Nerem, Univ. of Tsulv.lba, Ibaraki The First Dept. of Medicine, Osaka University 305, Japan and Georgia Inst. 'jf Tech., School of Medicine, Osaka 553, Japan. Georgia 30332, U.S.A.

To elucidate the mechanism of atherogenesis respect to local flow dynamics in the coronary Viscoelastic properties of cultured porcine artery, we investigated the relation between aortic endothelial cells were measured using the shear stress (SS) in vivo and the amount of micropipettetechnique for normal cells microfilament bundles (MF) in the endothelial (mechanically or trypsin detachedfrom the cells (EC) it the left anterior descending substrate)andthe cellsof which cytoskeletal coronary artery (LAD) of 5 mongrel dogs. We elements were disr,Tted by cytochalasin 13or obtained phasic flow velocity profiles using a 20 colchicine. In the micropipette technique, a MHz multigate pulsed Doppler flowmeter. SS was part of the cell was aspirated into the tip of determined as the velocity gradient at each wall the micropipette under a microscope by applying multiplied by the blood viscosity. The velocity stepwise an aspiration pressure, AP. The time profile at its peak velocity in LAD skewed toward course of the aspirated length, L, intothe OP outer side. Thus the peak SS at the outer pipette was measured from a TV monitor screen by side was greater than that at the cardiac side replay ig the recorded video tape. The data was (44.8+5.5 vs 21.2+4.3 dyne/cm', p<0.01). The anelyi J in the form of the time course Ghanges amount of MF in EC was evaluated by transmission in the stiffness parameter of the cell, 10=Rx electron microscopy as the ratio of the area AP/(L/R), where R is the pipett radius. It was filled with MF to the total area of EC (F/C). found that the creep deformation process of the F/C at the outer side was greater than that at cells was composed of two dift,swit phases. The the cardiac side (18.9+9.8 Z vs 14.4+7.1Z, stiffness parameter was expressed as an p<0.05). These results suggest t'iat inthe exponental function of time in each phase. canine coronary artery, MF in EC increases under the condition of high shear stress in vivo as an adaptation to its mechanical stress environment.

BE9-E.5 BE9-F.1 Pulsatile Shear Stress Effects on Flow Induced Proliferation of Endothelial Cells Cultured Vascular Endothelial Cells. H. Masudg, K. Tohda, K. Kawamura, T. Shozawa, M.J. Levesque and R.M. Nerem, Georgia & A. Kamiya, 2nd Department of Pathology, Akita Institute of Technology, Atlanta, University School of Medicine, Akita, Japan, and Georgia 30332, U.S.A. Research Institute of Applied Electricity, Hokk- aido University, Sapporo, Japan Theinfluenceof pulsatile shear stress on vascular endothelial monolayers has been It is known that vascular diamter is dependent on studied using cultured bovine aortic blood flow, although its mechanism has not been endothelial cells ona Thermanox plastic cover fully understood. We observed the morphological slip positioned in a parallel plate, laminar change of the endothelial cells which are loaded flow chamber. Both steady and pulsatile shear by high blood flow. Arterio-venous anastomosis stresses were employed, w!thpeak values between the common carotid artery and ipsilateral ranging as highas approximately 100 dynes/cm2 external jugular vein was constructed in dogs and and with exposure times up to 24 hours. The rats. Endothelial cells in the shunted common pulsatile flow environmentconsisted of a mean carotid artery proximal to anastomosis where blood shear stress upon which a 1 Hz, sinusoidal, flow increased were observed histol J.cally and non-reversing component was superimposed. ultrastructurally. After 4 weeks, they increased Measurements included bothcell shape and cell from about 3,000/mm2 to 6,000/mm2 either in dogs mechanical propertiesfor confluent monolayers or rats. A tracer study using Bromodeoxyuriaine and cell proliferation for subconfluent (BrdU) indicated that they proliferated with cell monolayers. The results indicate that division. Relative artery stenosis, about 2/3 of pulsatile shearstress produces significantly internal lumen, was constructed to the conuon enhanced responses compared to that due to a carotid artery of dogs. After 4 weeks, in the steady shear stress. The results also suggest stenotic portion where blood flowed in high velo- that vascular endothelia' cells cannot only city, endothelial cells were also much increased. recognize a fluid mechanic signal but can We think elevated blood flow induces endothelial differentiate between steady and Alsatile cell proliferation, which should be important to flow. This ability to recognize differencesin understand the role of blood flow in atherogene- hemodynamic environment maybe an important sis. factor 'in the determination of the focal pattern of atherosclerosis. 274

254 1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas Biomedical Engineering Scientific Papers

BE9-F.2 BE9-F.3 The Evidence of a New Method for Screening of the Multi-channel Doppler ultrasound studies Atherosclerosis Based upon the Pulse Wave Velocity of arterial blood flow in human subjects: effects of interventions, including vaso-

Tan 'ging , Datong Jiang *, Institute of Biomedical active agents. C. G. Caro*, N. Watkins, Engineering,Xi'an Jiaotong University, Xi'an, China K.N. Parker, M. J. Lever, C. J. H. Jones, Physiological Flow Studies Unit, Imperial According to Moensiorteweg equation, pulse wave velocity ( PWV ) College, London, SW7 2AZ will rise with the increasing of Young's modulus and thickness of blood vessel wall and decreasing of the vessel radius. It suggests that PNV will rise in patients with atherosclerosis and can be The occurrence of atheroma in arteries appears used to diagnose such disease .It is perferable that the vessel to ,a correlated with the blood flow pattern. between aorta valve and carotid artery is chosen for measurement Inadult humansubjects the diseaseoccurs for many reasons . We define a new parameter as pulse wave preferentially at sites where the time-average transmission time ( PITT ) in this vessel segment and it can be wall shear is low. It has been found that measured on phonocardiogram ( PCG ) and CPT simply . Because the these are also sites where the shear undergoes true length of the vessel is very hard to measure noninvasively small pulsatile excursions and large variations but it is closely related with subject's height, we prefer PNTT in direction during the cardiac cycle. It is as an indicator rather than PNV where height is considered. Also possible to alter the arterial bloodflow several other factors are thought to have influence to PUTT in pattern by means of pharmacological agents and normal subject, such as blood pressure ,body weight,sex,age, and cigarette smoking. We are now extending our mass index which should be considered.By a large amount of mea- studies on blood flow in the superficial surements on healthy volunteers(more than 2001,5 nonlinear multi femc- artery of healthysubjects using a variate stati. tic models were given and one of them is thought to mul, _nannel Doppler ultrasound device. The be practi-il in diagnosis where PNTT is non-linearly related with device provides velocity profiles during the height and age only.The difference between PUTT measured and the cardiac cycleas wellas spatially averaged expect healthy PUTT given by our model is considered to be the values. We are attempting to examine the diagnosis criterion.It is proved that can distinguish the man effects of various interventions. healthy or ill by many measurements with patients who were din- nosisd with atherosclerosis.This result shows the new noninvasive diagnosis method for atherosclerosis disease is feasible,especially for mass screening.

BE9-F.4 BE10 -A.1 Atheroscerlosis Research Using MRI with Three-dimensional Vessel Heart Rate Variability inAdults Reconstruction, J. Hale, J. Rapp, and Infants - An Overview R I Kitney, X-M. Pan and L. Kaufman, UCSF-RIL, Department of Electrical Engineering, 400 Grandview Drive, Imperial College, London SW7 2BT, UK. South San Francisco, CA 94080 USA The field of heart rate variability MRI is known to be effective in showing (HRV) analysis in adults and infants vascular pathology. However, it has not has expanded rapidly over the last been used extensively in clinical situa- few years. The paperwill review tions. One reason is the difficulty for thesedevelopments both in terms of those unfamiliar with tomographic images signalprocessing and modelling and to understand them as they are most willprovide an introduction to the commonly displayed. HRV workshop. We have developed softwa.e that converts reyareas to be covered are normal cross-Jectional images into the angio- and abnormal HRV patterns in adults, graphic formats surgeons are more and the use of HRV analysis in the accliltomed to analyzing. The software studyof autonomic neuropathy. HRV also reveals information not shown in the development in babies will also be standard magnitude displays and films. discussed and the differing patterns Using three-dimensional reconstruction we of activity during the first year of have demonstrated a method of characteriz- life described. The final section of ing lesions that is potentially more the paper will consider the role of accurate than angiography, at least for HRV analysis in the study of SIDS. large vassels. Our most recent work has been to explore the ability of MRI to detect the presence of thrombus. Early results are encouraging, and we are incorporating this new data into the three-dimensional display software. 275

1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas 255 Biomedical Engineering Scientific Papers

BE10-A.2 BE10-A.3 Spectral analysis of variability in neuro-- 24-HOUR ANALYSIS OF HEART RATE AND ARTERIAL BLOOD vegetative control systems-shy and when? PRESSURE VARIABILITY SIGNALS John M.Karemaker," Dept. of Physiology, AMC S.Cenitti:G.Baselli,M.Bertocchi,S.Civardi,A.Guzzet- Meibergdreef 15, 1105 AZ Amsterdam, NL ti,L.Furlan,A.Pagani,M.Pagani Department of ElectronicEngineering,Politecnico Recent literature has shown a revival of interest and Medical Pathology, University of Milano, Italy in the analysis of biological oscillators. Especially thequantificationof variability in an easily measurable parameter such as heart rate Aprocedure for the analysis of 24-'/ourECGand has found many supporters. Parameters extracted arterial bloodpressure variability signals is from such spectra have been used to quantify presented. Thesesignals areprocessedthrough alertness, psychological stress or cardiac spectral analysis carried out via parametric health.However, the widespread application of techniques. Pcwer density spectra are obtainedon this technique seers more likely linked to the 512consecutive cardiac cycles for the whole24- distribution of computing power than to our hour peiic.d and a few parameters are extracted increased knowledge of the in underlying physiology. each spectrum(normalized power in low Basic issues still frequency are:is an observed and high frequency bands and their ratio periodicity due to together with the classical mean and variance values). a. a central nervous system instability in the The plotting of such parameters provides 'circuitry'that controls the observed quantitative indeces for the evaluation ofthe phenomenon (temperature, respiration, blood sympatho-vagal balance in the control of heart rate pressure,heart rate), or and arterial blood pressure. b.the fortuitous outcome of a set of coupled In particular, during the nocturnal hours there is oscillators, both in the CNS and periphery,or a marked increase of high frequency components and c. the 'tuned behavior of a filter' controlloop, a correspondent decrease of low frequency set in motion by noise-like perturbations ? components. This factmay be explainedbythe As long as these questions are not settled we predominantrole elicited by the parasympathetic should refrain from too much interpretation of system during the night. too scanty data on too complex a system as is an Thisprocedureof signal processing is actually iunsedated subject, sitting quietly in a chair'. routinely applied in a medical pathology department of a major universityhospital in Milanofor clinical purposes.

BE10-A.4 BE10-A.5 A model of short-term blood pressure and heart ASSESSMENT OF AUTONOMICNERVOUSRESPONSES DURING rate variability. AUTONOMIC STRESS TESTS BY POWER SPECTRALANALYSIS J.J. Settels: K.H. Wesselirg, BMI-TNO, Academic OF HEART RATE FLUCTUATIONS IN PATIENTS WITH Medical Center, Amsterdam, the Nehterlands. IMPAIRED LEFT VENTRICULARFUNCTION A.Tani*, Y.Ishida, Y.Yoshida, H.Kusuoka, M.Hori, A continuous blood pressure recording showssurpri- A.Kitabatake, M.Inoue, T.Kamada, K.Kimura, sing variability. Short term variability includes T.Kozuka, The 1st Dept. of Med., Dept. of components with periods of a few seconds to several Radiol., Osaka Univ. Osaka 553 Japan hours. The most powerful short term bloodpressure stabilizing mechanism is the baroreflex, and the To noninvasixaly assess the autonomicnervous question is raised whether bloodpressure variabi- activities, we developed a computer assisted lity exists in spite of the baroreflex or is media- method determining the autoregressive power ted by the baroreflex. Bloodpressure and heart spectral density (PSD) of ECG R-R interval rate variability were studied using noninvasive fluctuations. Using this method, we studied the and invasive registrations on patients and healthy autonomic responses to orthostatic stress (OS) and subjects in conjunction with a model of a barore- cold pressor stress (CP) in 6 normals (NL) and 5 flex and cardiopulmonary reflex controlledcircula- patients with prior iwocardial infarction (MI). tion. In the time domain, the significant level Mean LV ejection fraction (EF)was 62-15 % shifts in mean pressure sometimes observed inthe (mean±SE) in NL and 48±8 Z in MI. 24 hour day are difficult to reconcile PSD contained with a pro- two major components in power at low (0.1Hz :LF) perly functioning baroreflex, bLt can be reproduced and high (0.25Hz :HF) frequency which have been by baromodulation i.e. modulation of baroreflex assumed to ,flect sympathetic and parasympathetic loop gainnly, in terms of inhibition or facilita- activities, respectively. With OS, LF;HF ratio tion of nervous activity. Since the baroreflex as is a marker of sympatho-vagal balance increased from fast acting changes in blood pressure can be fast 1.7±0.6 to 9.6±4.8 in NL but remained almost also. In the frequency domain, the proper random baro- same in MI. With CP, the ratio increased in both modulation together with statistically independent groups, 1.7±0.6 to 4.5±2.5 in NL and 1.1±0.5 to random modulation of the vagal reflex gain on heart 6.5 ±2.0 in MI. Thus in MI patients with decreased rate explains the pressure and rate power density LVEF, the autonomic response to OS is spectral components around O. blunted Hz and below 0.05 indicatingdecreased baroreflex Hz taken in resting subjects. Periodic responsiveness, baromodula- although the response to CP is preserved. tion explains spectral phenomena ofrespiration. ET r ; t;

256 1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas Biomedical Engineering Scientific Papers

BE10-B.! BE10-B.2 RESPIRATORY SINUS ARRHYTHMIA: A PROBE On respiratoryvariations in bloodpressure OF AUTONOMIC CONTROL OF THE HEART non-related to sinus arrhythmia,P. Rill*, J. J. Phi!ip Saul, Ronald D. Berger, Richard J. Cohen Colominas,J.Rosell,and R. Pallas-Areny, Harvard Medical School and MIT, Boston, MA Div.Instrumentacion y Bioingenieria,ETSI Modulation of both er-'lac sympathetic and parasym- Telecom, Box 30002, 08080 Barcelona, Spain. patheticefferentactivityoccurswithrespiration. We developed and tested an efficient new technique to noninvasive- .,cdpressure is subject to thesimultaneousinflu- ly investigate autonomic cardiac control by using respiration as ence ofseveralactions,including respiration.This a frequency probe of the autonomic nervous response. The produces a sinus arrhythmia (RSA) which isthe main heart rate response to respiratory activity was determined in 14 factorconsideredwhen analyzingrespiratoryvaria- normal adults during 6 minute periods in which the respiratory tions in blood pressure.There is a frequency modula- rate was voluntarily controlled in a predetermined but erratic tion of the arterial pulsewave which is apparent also fashion.f3elective autonomic blockade, with either atropine in the ECG signal. (0.03 mg/kg) or propranolol (0.2 mg/kg), and changes in pos- Ourresults innormal subjectsindicate that there is ture were used to dissect the sympathetic and parasympathetic aconsiderabledifference in thecardiacfrequency contributions. The gain and phase of the frequency response measuredbeattobeat depending onwhether itis functions describe how the autonomic nervous system responds calculatedfrom the ECG or the external pulsewave. to a stimulus, in this case respiration, as a function of frequen- Thisdifferencepersistsindependently of thealgo- cy.We found thatthesympathetic (standing + atropine) rithmsapplied for themeasurement.By subtracting response is characterized by markedly reduced gain above 0.1 bothfrequencies,andinterpretingthe resultas a Hz and a phase delay, while pure vagal (supine + propranolol) timesignalsampleu atnon-.egular intervalscorre- modulation of heart rate is spondingtothe instantaneouscardiacperiod,this characterized by higher gain showsa goodcorrelation with the respiratorysignal (nasal thermistor).Thisindicatesthat,asidefrom atallfrequencies and no phase delay(figure). The theF.SA,there are llerrespiratory influenceson data provide the basis for blood pressure,perhap.,through the variation of the noninvasively quantifying pulse wave velocity as pointed out by several authors. the sympathetic and vagal 180r compon'ents of cardiac control in an individual, and the technique should be useful in the diagnosis and management of autonomic -180 Frequency (hertz) dysfunction. BE10-B.3 BE10-B.4 A study of the human Baroreflex System by Heart Rate SOME RECENT RESULTS IN THE ANALY. TT RATE Variability spectral analysis during anesthaesia VARIABILITY IN DIABETICNEt Rompelman, B.J.ten Voorde*, G.J.Scheffer, Th.Faes, O.Rompelman Th.J.C.Faes*, P.Lanting, B.ten Voorc J.J.Heimans,Medical Physics Group, Free University. Medical Physics Group, Vrije Universiteit Amsterdam POB 7161,1007 MC Amsterdam, The Netherlands POB 7161, 1007 MC Amsterdam, The Netherlands Diabetes MelliLus may affect the autonomic nervous Short term regulation of arterial blood pressure is system. An observable effect is the change in heart largely a function Jf the autonomic system through rate variability (HRV), caused by the reduced in- baroreceptor reflexes. Two effect°_ mechanisms are nervation of the SA-node. Different techniques of heart ra:.e, mediated para- and sympathetically and HRV anal! is are used to diagrJse diabetic neur o- vasomotor control, mediated only sympathetically. pathy. Spectral analyses of HRV offers an usefull Heart Rate Variability (HRV) spectra reflect these tool for analysing both sympathetic and parasym- autonomic functions in low frequency (LF) power, pathetic influences on HRV. related to blood pressure variations and high fre- Instrumentation has been developed that allow for quency (HF) activity, respiratory sinus arhythmia. on-line bedside analysis of HRV oiffarent statistical HRV-parameters as well as spectra are We developed a bedside HRV data-acquisition/proces- estimated; the results are available immediately sing system to perform clinical HRV studies. We after finishing the measurement. studied HRV spectra during selective blockades of So far we measured HRV in 74 healthy adults and in autonomic efferent pathways: - Blockade of vagal 44 adults suffering from diabetic neuropathy. baroreflex control of heart rate caused a dramatic decrease of both LF and HF HRV power; - Blockade of We found e.g. that in adults with diabetic neuro- the sympathetic baroreflex control loop of vasomotor pathy the energies in frepency bands round the activity through eidural analgesia did not alter HF respiratory frequency an the Baroreflex Oscilla- nor LF power in HRV significantly for the group of tion Frequency (BOF) are reduced. The obse.vation that the BOF is shifted ;c1 lower frequencies and patients studied. that the BOF and the IE -'alue are correlated The strong vagal influence on heart rate control is (R.0.8, p 0.02) suggests ,.hat the BOF is also a well known phenomenon. The latter result though under parasympathetic infi,cxnce. raises new questions around the origin of LF HRV, Spec,a1 analysis of HAT has proven to be possible which is believed to be inherent to the vasomotor in clinical practice aLd provides a method to baroreflex loop. Possible explanations will be dis- analyse the sympathetic and uarasympathetic cussed. influence on HRV separateiy. 2 7 7 1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas 257 Biomedical Engineering Scientific Papers

8E10-8.5 BE10-C.1 Neonatal Vagal Tone: "HRV in Healthy, Fullterm Neonates", Diagnostic and Prognostic Implications K. Baldzer, F. Dykes, S. Jones,D. Davet and Stephen W. Porges: Ph.D. D. Giddens: Emory UniversitySchool of Medicine Numin Development and Psychology Georgia Institute of Technology University of Maryland Atlanta, Georgia 30332. We havestudiedHeartRate 'ariability Heart rate patterns havelong been used in (HRV) in 20 normal fullterm infants during neonatal assessment. Underlying this research quiet sleep, between 30-60 hrs. postnatally. is the assumption that heart rate variability Using spectral analysis, weexamined the fol- is a manifestation of central lowing: mechanisms low frequency power(.04-.20Hz),res- mediated primarily via the vagus nerve. Our piratory sinus arrhythmia (RSA) power (i.e., laboratory has conducted a number of studies power at peak respiratory frequency +/-.2Hz), demonstrating that accurate quantification of total power, and respiratory rate. For data respiratory sinus arrhythmia provides a obtained from stable 100 second sensitive measure of cardiac records our vagal tone. This infants fell into twogroups. Group A infants presentation will focus on onr application of exhibited a greaterRSA powerand lower res- the vagal tone measure with normal and high piratory frequencies (.5-.7Hz),whereas Group risk neonates. Studieswill be presentA B infants had very littleRSA powerand describing: therelationship between neonatal frequencies of .7Hz orabove. Slopes of low vagal tone and clinical outcome; the frequency (LF) and RSA powees vs. totalpower developmentof vagal tone in the were pre-term significantly different forthe infant; the influenceof two sleep state on vagal groups: GrA: LF vs. total .32 (r=.89); RSA tone; the relationship vs. betweenvagal tone and total .4-17-Tr=.,93). GrB: duration of LF vs. total .73 hospitalization; the relationship (r=.99); RSA vs. total .04 (r=.72). between vagal tone and autonomic reactivity to Our experience with HRV in healthy adults medical procedures suchas circumcision, lumbar and in preterm infantssuggests that the Group Punctures, and gavage feeding. A fullterm neonates have spectral characteristics similar to adults whiletheGroup B neonates tend toward preterm behavior. The quantitative indices derived from HRV spectral analysis in this investigationmay thus be of value in the study of maturationof CR control.

BE10-C.2 BE10-C.3 A STUDY OF CARDIOVASCULAR REGULATION USING HARMONIC DETERMINATION OF SLEEP STATE USINS ANALYSIS David Gordon*, Div Cardiology, Children's RESPIRATORY VARIABILITY Hospital of Philadelphia, Philadelphia, PA Gabriel G. Haddad*, and Tze L. Lai, Dept Pediatrics, Columbia Univ,Coil of Physicians A retrospective clinical study of 44 infants and and Surgeons, New York, NY children that had serious heart disease conducted It is known that cardiorespiratory June 1983 to July 1985 at The Children's Hospital variables in rapid eve movement (REM) sleeP in Boston, and The Children's Hospital of Philadel- are different from those in quiet sleep (ES). phia using a bedside monitor that performed heart We thus asked whether such variables can be used forthe determination of rate and respiratory activity spectral analysis re- sleeP state. We studied nine normal fullterm in- vealed that specific spectral natterns of 18 pa- fants at 1 and 4 months of life. tients wno sustained a cardiac EEG, ECG, and post- -rest differed sig- ural muscle EMB were recorded. nificantly from others in the study group (p<0.0001). Means of RR interval, respiratory cycle time (Ttot) We report here th results of a double blinded clin- and tidal volume, and coefficients of variation (CV) of ical study from December 1986 until January 1988 on the same variables, were obtained for 30-s intervals throughout 404 patients which constituted approximately 90% of each sleeP study. We developed a statistical decision the postoperative cardiac surgical patients. The rule using CV of Ttot for the patients were monitored for approximately 24 conse- classification of REM and RS in blocks of 5-min Periods. cutive hours after surgery. Of 05 epochs The fundamental premise staged as QC by neuroPhYsiologic of these investigations is rooted in the conjecture and behavioral criteria, 79 epochs (or 93%)were classified correctly that quantifying fluctuations in heart rate and res- as as using our decision rule. piratory activity are sufficient to provide sensitive Of 85 epochs staged as REM sleeP, 84 were classified and specific indicators of impending and in process as REM sleeP and only 1 misc.assified as S. cardiorespiratory life threatening events. We conclude that in Implicit ming infants 1) the variability of in this investigation is the impact that evolving cardiorespiratory parameters :an separate quiet from patterns of surgical and intensive care management nonquiet sleep better thai the mean values and 2)sleeP staging will have on the indicators discovered in the pre- using a decision rule based vious study. This conjecture will be tested, and on CV of respiratory cycle time can be Performed withhigh degree of the evolutionary changes in patient management will accuracy. be correlated with the dynamical process studied using methods which have their origins in the harmonic analysis of multiple time series.

258 1988 World Congress on Medical Physics and BiomedicalEngineeringSan Antonio, Texas Biomedical Engineering Scientific Papers

BEIO-C.4 BE10-C.5 The Transmission of Circulatory Disturbances to RESPIRATION ANDR-R INTERVAL TIME SERIES IN SLEEP the Neonatal Brain. APNEA - Toshi-Hiro Nishimura*, Dept.of Elec. Eng., Oita Univ., Oita, JAPAN S Bignall,* R I Kitney and R P A Rivers This study describesa method of evaluation Dept of Paediatrics, St. Mary's Hospital Medical about the information ofrespiration and heart School, * Department of Electrical Engineering, beat obtained from the newborn. It is possible to Imperial College. London, UK. estimate sudden infant aeathsyndrome in advance Abnormal patterns of cerebral blood flow by investigating cross correlation between respi- velocity (CBFV) have been shown to be associated ration and R-R interval in stationary state and with brain damage in the preterm infant. We sleep apnea. have previously demonstrated that sign The alternation of interval on wave form is cal- processing techniques applied to ble essure culatedbythe information of ECG which is detected from the newborn. The envelope on respiration and CBFV waveforms can reveal disc . of the waveform is calculate(' by the informationof res-_ regulation of brain blood flow innfants piration which is detected from the foregoing new- undergoing -7.-Gensive care. The CBFV disturbances born. This envelope is sampled every R-R interval arise from modulation phenomena produced by on the information of ECG. Cross correlation be- respiration and spontaneous motor activity. tween the alternation of interval on waveform from Hitherto these events have been considered the foregoing information of ECG and the envelope to be stationary phenomena. However, more on respiration which is sampled are calculated. detailed analysis has revealed complex inter- This method is constructed so that the relation actions between blood pressure and CBFV. between advance and delayof respiration and R-R interval inwhich thevalue of cross correlation The elucidation of these interactions requires is large is evaluated. the use of signal processing techniques which Byforegoing results of cross correlation, it is are capable of tracking non-stationary events shown thatcross correlation between respiration on a short time basis. The methods which will and R-R interval is changed fromstationary state be described are also capable of investigating within a few minutes before nerve center type the degree of chaos in the system under study. sleep apnea. By calculaUng automatically cross correlation between respiration and R-R interval, for example, it is possible to find just a few minutes before sleep apnea state of newborn and to inform the state.

BE11-A.1 BE11-A.2 Control-Space Diagrams ,,or Plow Through Effect of Entrance and Outlet Rigid Collapsed Tubes, C.D. Bertram; Pipes on Flow in Collapsible Tubes, University of New South Wales, Australia 2033 Y. Matsuzaki* and T. Matsumoto, Nagoya Universit, Dept. Aeronaut. Engng., The operating points for a Starling resistor can Ciikusa, Nagoya, Japan 464 be described on axes of pressure drop along the tube, flow-rate, and transmural pressure at the downstream end (sPe2), as a curved surface which is In order to study self-excited oscillations of a independent of external conditions, i.e. para- finite-length collapsible tube, a simple two- meters up- and downstream of the tube. The dimensional flow channel is used together with stability of an operating point depends on both Cancelli and Pedley's one-dimensional flow internal and external conditions, and so at each theory which may account for the pressure loss parameter combination the surface is partitioned due to flow separation. A flexible part of the differently into regions of distinct behaviour. flow channel consists of two elastic membranes, No special advantage is then possessed by this and entrance and outlet rigid parts are made of presentation. Position on a given surface is pairs of parallel rigid walls.The analysis determined by two control variables, here upstream includes the effect of longitudinal membrane pressure pu and external pressure pe. The surface tension caused by an initial strain and membrane can be mapped onto a plane with these axes. How- deflection. For the entrance and outlet rigid ever, for a given pu, the interesting behaviour channels, the inertia of fluid mass and viscous regions are traversed in a relatively small pe resistance due to the walls are also taken into range about a pe increasing linearly with pu. The account. Numerical results show that the resulting control-spacL diagrams contain zones entrance and outlet rigid channels have a great which cannot be attained experimentally by setting influence on the flow characteristics. For pu then altering pe, because f2 changes discontin- initially unstrained membranes, the amplitudes uously as the tube collapses or changes oscillat- of membrane and flow oscillations are small. On ion mode. Other regions of behaviour show three the other hand, when the membranes are less distinct frequency ranges plus, at unexpectedly flexible because of the initial strain applied, moderate pu and Pet, yet-higher-frequency oscillat- the oscillations become large and chaotic. ions which may be modulated by the low-frequency modes. 279 1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas 259 Biomedical Engineering Scientific Papers

BE11-A.3 Bill-A.4 Characteristics of Shock-Wave propagating Flow Limitation and Flutter Phenomena in throt.gh a partially collapsed Rubber Tube, Collapsible Tubes J.B. Grotterg: N. Gavriely, T.R. Shee and D.W. Cugell, Northwestern M. Snimizu*,Tbkyo Institute of Technology. University, Evanston, Illinois 2-12-1 Ookayama Meguroku Tokyo, Japan. We studied flutter in collapsible tubes as As observed in Korotokoff sound generation, a steep a possible mechanism for the generation of asending limb (shock-wave)is formed when a respiratory wheezes. The pressure-flow relation- compression wave is travering down a partially ships and the wall oscillations of thick-walled, self-supporting latex and Silastic tubes mounted collapsed blood vessel or rubber tube. Behavior of between rigid T.ines were measured. 1.1a fou,Id loud the tube in the shock -wave regicn has not been known honking sounds ant- tube wall oscillations that well and many hypotheses are proposed. To make clear occurred only wbsn the tubas w.re buckled and the relation between the vessel cross-sectional area flow-limiting, i.e. when the flow became con- and the transmural pressre (tube law) in the shock - stant and independent of downstream driving pressure. The overall range of oscillatory wave regeon, an :%perimental study using a thin frequencies was 260 to 750 Hz for air flow, walled silicone rubber tube has been conducted, in presenting as sharp peaks of power on the fre- which rubber tube diameter and internal pressure are quency spectrum. The oscillatory frequencies measured at the same location and simultaneously. were higher at higher fluid velocities and with From the experiment, it has been known that, when the na,..:wer distance between opposing flattened degree of collapsion is low, the tube law obtained walls. The experimental data were compared with predictions of our fluid dynamic flutter theory with the assume don of no energy loss at the shock and the vortex-induced wall vibrations region agrees well with the real one. mechanism. The results of the tube experiments and the theory predictions are consistent with measurements of flow limitation and whee-ing breath sounds made by our group on excised dog lungs and in normal human subjects.

BEll-A.5 BE11713.1 Stability of the flow through a Flow Induced Instability of Collapsible Tubes collapsible tube mounted in a rigid P.M.Webster; and C.Walsh hydraulic circuit, J.W. Reyn, Delft Univ. of Toronto Inst. for AErospace Studies, University of Technology, Department and Faculty of Medicine, Toronto, Canada. of Mathematics, P.o.Box 356, 2600 AJ Delft, The Netherlands. eroblems in physiologic fluid flow have stimulated the development of continuum mechanics There is investigated how wave propagation in the from its infancy.Euler described the gover- flow through a nearly cylindrical collapsible tube ning equations for one dimensional flow in filled with an incompressible, inviscid fluid in- collapsible tubes. Their solution models flow teracts with the flow in the upstream and down- limitation in the caval and cerebral veins, the stream rigid segments of the hydraulic circuit in membraneous urethra, the airways and elsewhere. which the collapsible tube is mounted. This leads The dynamic instabilities accompanying flow to the study of eigenmodes of the classical wave limitation, most obvious as the wheeze of equation satisfying somewhat unusual boundary con- forced expiration, are not modeled by these ditions dictatsd by the matching of rigid and equations. Two dimensional linear models of flexible parts of the circuit. The characteric uniform channels do yield dynamic instabilities equation, determining the frequencies of the ei- but not flow limitation. Other approaches genmoC.s, is of a transcendental quasi-polynominal include relaxation oscillations, vortex shedding, type ano the location of its roots is analysed. It and one dimensional models with longitudinal appears that for subcritical flow ks<11 all eigen- tension. Experimnts with a latex membrane modes are damped, whereas for supercritical flow stretched over a rigid channel Oth a local (S>1) also growing eigenmodes exist. Here S is the minimum depth snow a pressure pxyale similar to index being the ratio V/C, where V is the fluid open channel flw limitation and aerodynamic speed in the collapsible tube and C the speed of chokes. Dramatx, growing, waves move down- wave propagation. stream from the constriction suggesting aern- elastic flutter in the recjicn of supercriticai flow. We will present hip speed photography (5000 frames/s) and preliminary numerical solutions in the context of a general review of the medical implications of such phenomena.

() i'N A.. 0V 260 1988 World Congress on Medical Physics and Biomedical Engineering San Antonio, Texas NIZINIMMESEENSS61.=.12MIZIEr

Biomedical Engineering Scientific Papers

BE11-B.3 Simulation of Flow Pattern During Forced Expiration, Urine flow through the ureter, D.Griffiths, D. Elads, R.D. Kamm, and A.H. Shapiro, Tel Aviv Erasmus University Rotterdam, Netherlands University, Tel Aviv 69978, ISRAEL, and Massachusetts Institute of Technology, Cambridge, MA 02139, U.S.A. The ureter is a collapsible, muscular tube that conveys urine from the kidney to the Flow limitation during forced expiration is simulated by a bladder. Peristaltic contractions move mathematical model which is based on methods of analysis successively along it towards the bladder. for one-dimensional flow in collapsible tubes.Airway com- A fundamental property of a collapsible pliance is modeled using a semi-empirical, lung-volume dep- tube is the "tube pressure" at which it endent tube law with the tethering effects of the parenchyma collapses. At a "compressive zone" (a lumped into the wall properties of the airway. This mimicks region of elevated tube pressure) the velo- the parenchymal-bronchial interdependence for collapsed air- city of a flowing fluid may reach a criti- ways, and provides a mvie realistic representation of airway cal value and in this ease the compressive collapse. The maximal flow-volume curves generated by the zone controls the flow pattern. A peri- model compare well with physiological data.For the first staltic contraction represents a moving time, the pattern of steady flow downstream of the flow lim- compressive zone. Lubrication-theory iting site (choke point) is computed, including regions of sup- analysis shows that it is potentially able ercritical flow and elastic jumps.The model examines the to pump urine tL_Jugh the ureter at flow effects of effort it: excess of that required to limit flc, In rates up to a maximum "carrying capacity". addition, the effects of airway closure and fluid compressibil- Th reality is more complicated: (a) at ity are incorporated into the model and their contribution to low peristaltic frequencies sometimes no the flow-volume curve is analyzed. contraction is present in the ureter and uninterrupted discharge from kidney to bladder may occur; (b) urine is discharged into the bladder against an outlet resist- (SupportedbyHL-33009 and Bat-ShevadeRothschild ance, including the bladder pressure itself Foundation) Many different flow patterns are possible, in most of which peristalsis acts as an inefficient pump, allowing bladder pressure changes to be transmitted to the kidney, or even hindering the urine flow. Experiments in pigs illustrate these conclusions.

BEll-B.5 BE11-8.4 A Non-Homogeneous Model for Human Transitions Between States of Subcritical Flow in Expiratory Flow, R.K. Lambert; Massey Tubes That Do Not Close at Negative Transmura] Pressure, Theodore A. Wilson; Departmeni, xr Aero- University, New Zealand. space Engineering and Mechanics, University of Minnesota, Minneapolis, Minnesota 55455 U.S.A. The merging of two dissimilar flowsina tubejunctionisunderspecifiedbythe Certain features of maximum expiratory flow curves of some canine lungs suggest that 3-11-ating flow conservation laws of mass and energy. One through the airways is not unique. mcre relationshipis needed. The empirical There is a maximum flow that can pass 6:.-ough a equation of Wilson etal (JAP 59:1924.- collapsible tube if the tube is sufficiently 1928,1985) has beenusedinconjunction compliant. A lower flow is allowed at either of two values of lateral pressure and tube area. At with the computational model of Lambert et the higher pressure, flow is subcritical; andat al (JAP 52: 44-56,1982) to develop two-lobe the lower pressure, flow is supercritical. modelsforexpiratoryflowfromnon - Transitions from supercritical to subcritical homogeneous human lungs. Cases flows are allowed; reverse transitions arenot. investigatedincludebothmaximaland If the tube does not close at large negative transmural pressures, a given flow can pass partial forced expirations as well as steady through the tuve at three values of lateral sub-maximal flows. The investigations were pressure. It appears that transitions between carried out with asymmetric peripheral and the subcritical flow states, the states with middleairwaysconstrictionswithboth the lowest and highest values of lateral pressure may be allowed. If -nose transitions occur, the central and more peripheral mixing of the flo, through a collapsible tube with axially flows and with and without.tntral flow varying properties is not unique and flow-limit- limitation.Theeffectsofvariationsin ation-at-wave-speed could occur for different regional lung compliance were also studied. flows. Supported by National Heart, Lung, and Blood Supported in part by grant HL 21584. Institute Grant HL-2158h.

2I 1988 World Congress on Medical Ph7sics and Biomedical Engineering o San Antonio, Texas 261 Biomedical Engineering Scientific Papers

BE12-A.1 BE12.A.2 Dynamic Measurement of the Geometry of Fluid The extitence of steady flow in a collapsed tube Filled Collapsible Tubes, D. Elad*, M. Sahar, O.E.Jensen* and T.J.Pedley, Department of Applied S. Einav, J.M. Avidor, R. Zeltser, and N. Rosenberg, Mathematics and Theoretical Physics, Silver Sreet, Tel Aviv University, Tel Aviv 69978, ISRAEL. Camhridge CB3 SP,W, England. Self-excited os, illations arise during flow through a pres- Physiological flows such as blood flow in veins or air flow in surised segmett of collapsible tube, for a range of val- the lungs can be simulated by flow throus.'i collapsible tubes. ues of the time-inZvendent controlling pressures (up- When the external pressure exceeds the internal pressure, the tube is partially collapsed and attains a multitude of complex stream, downstream and external). They come about cross sections. The contour of the tube is determined by the either because there is a steady flow corresponding to local dynamic equilibrium between the moving fluid and the these pressures, but it is unstable, or because no steady properties of the tube wall.Since any physical contact with flow exists. We investigate the existence of steady flew etube alters the dynamic equilibrium, existing techniques in a one-dimensional collapsible-tube model, which takes are capable of measuring cross-sectiwal areas only. The account of both longitudinal tensioiTand jet energy cross sectional contours have not yet been experimentally det- loss Edownstream of the narrowest point. For a given ermined.Experimental data of the tube geometry is particu- tube, the governing paran.eters are /', flow-rate Q, and larly important for better understanding of large amplitude transmural pressureP at thedownstream end of the col- wave propagation and the overau fluid dynamic behavior lapsible segment. For any fixedPandTthere are two across an elastic jump. This work presents experimental res- critical flow rates Qt and Q2. For Q < Qi, two col- ults of fluid flow through a collapsible tube (latex tube), in- lapsed solutions exist whenE 0but only one when cluding the measurement of the tube geometry by a novel E0; for Qi < Q < Q2 there are no solutions when non-contacttechnique,based ontheRasterstereography E = 0,but one solution (with the upstream part of the method. The three-dimensional geometry of the tube is reconstructed from the images of the tube by advanced tech- tube distended and the downstream part collapsed) when niques of image processing. Here we report on representative E 0;forQ> Q2 there exists only a solution with the results of steady flows at different combinations o; flow hates tube everywhere distended andE = 0.The results are and chamber pressure.However, the technique can success- compared with those of Reyn (theoretical) and Bertram Lily measure dynamic changes of the geometry, of 4 fluid- (experimental). filled elastic tube and provide new experimental data which is useful for analytical predictions.

BE12-A. 3 BE12-A.4 Expiratory Flow in a Collapsible Bifurcation, usciautins in a Collapsed -Tube Analog of the D. Katz , D. Elad, and S. Einav, Tel Aviv University bruohial iftery under a Sphygmomlnometer Cuff', Tel Aviv 69978, ISRAEL. C.D. Bertram* and C.J. Raymond, University of New South Wales, Australia 2033.

The bronchial tree is a complex system of br --!hing compli- becaust the dynamics of a Starling resistor (SR) ant tubes.During expiration, air is flowi., from the peri- are strongly influenced by the diverging flow at rheral daughter tubes into the parent tubes, until it converges the downstream end, where the most Lollapsed cross- to a single stream in the trachea and expelled from the section rapidly changes to the (usually) circular mouth. The airways are flexible tubes which undergo major attachment, it has been suggested that SR oscill- geometrical changes, especially during forced expiration.In ations have little application to physiological this study we present a one-dimensional simulation of air vessels. To demonstrate that the oscillations do flow through a simplified model of a single Y-bifurcation. not depend on the ptesence of an unphysiologically The bifurcation is assumed elastic and its mechanical proper- abrupt change from compliant to rigid tubing, a t:es are defined by a modified tube law.The model is silicone rubber tube of length 30.7 inside diameters numerically analyzed by the two steps MacCormack method and wall thickness-to-radius ratio 0.3 was set up which was found suitable for such type of fluid-structure passing through a larger-diameter compliant tube problems.Numerical calculations were performed for two stretched between the rigid pipes of a short SR different streams that enter the bifurcation with either sub- pressure chamber, so that under external, pressure critical or supercritical flows.Preliminary results show that the outer tube pressed on the tube and the two the interaction between the daughter streams, while merging collapsed together. The outer tube stretch ensured in the parent tube, final flow pattern.It is also that the compressive force applied to the inner observed that the angle between the daughter tubes has a sig- tube over its central 10.3 diameters was reduced to nificant effect on the degree of merging of the flows. The ze.:o smoothly at the ends of the central section. results indicate that flow limitation (choked flow) as well as The inside diameter of the SR chamber pipes was elastic jumps may occur in the parent tube due to the initial sufficient that the Jima tube passing through did conditions at the entrance. not contact opposite sides even if collapsed. Operating points have been mapped as for a SR. Oscillations at frequencies from 3.2 to 320 Hz have been observed.

2 8 2

188 World Congress on Medical Physics and Biomedical EngineeringSan Antowo, Texas Biomedical Engineering Scientific Papers

BE12-A.5 BE12-B.1 Numerical Simulation of Flow in the Trachea, The Effect of Respiratory Static Volumes Incorporating a Membrane Wall Model, C. Walsh*, on Heart Valve Sounds, C.S. Lessard, P. A. Sullivan and J. S. Hansen, UTIAS, 4925 J.J. Im", and J. Lay, Texas A&M University, Dufferin St., Downsview, Ont., Canada, M3H 5T6 College Station, TX 77843 Flow limitation and wheezing occur during forced The effect of respiratory static volumes on heartNil%e expiration and the quantitative observatio:s of sounds as obtained from a tracheal microphone may pro- flow rate, pressure and frequency have been matched vide an insight into the relationship between respiratory by previously published :xperimental results from a mechanical model of the human trachea. This pres- static volumes and heart sounds. Spectral analyses of entation deals with a nonlinear, one-dimensional, the first heart valve sounds(HVS1), the second heart numerical .imulation of the -lechanical model. In valve sounds(HVS2), and the combination of the first contrast to past analyses using tube law plus and second heart valve sounds(HVS3)\Jereobtained at constant axial tension models, the present analysis the four different respiratory st, 1c volume conditions, leads to a natural dependence of axial tension on transverse displacement. Further, the axial i.e.. breath held following normal tidal volume inspi- tethering constraint implicit in the tube law ration, normal tidal volume expiration functional in- approach can be dropped, allowing membrane inertia spiratory capacity, functional expiratory capacity. The to enter through the wail axial equation of motion. Cooley -Tukey FFT algorithm was applied to the HVS We have examined axisymmetric flow through an data to obtain an estimate of HVS spectra.Descrip- annular duct with an axially tethered membranous inner wall and the dependence of axial tension on tive statistics, i.e., peak frequency(FPK), mean power radial displacement causes only minor modifications frequency(MPF), bandwidth(BW) were calculated from to the predicted behaviour. It is anticipated the frequency sp .tra of HVS at the four conditions. Fi- that dropping the tethering constraint will have nally, the respiratory effect on heart valve sounds was more dramatic consequences. obtained by the analysis of variance procedure for a two- way classification resulting from a factorial experimental design. The results indicate that the HVS2 spectra are affected by the respiratory static volume conditions. The HVS1 spectra and the HVS3 spectra are unaffected by the respiratory static volume conditions.

BE12-B.2 BE12-B.3 Characterizationof Hemorheology Non-invasive study ofhemodynamic in Total Artificial Heart (TAH) effects of respiration on barcreflexes, Implantation, T. Hung; D. Butter, M. GUERRISI* & A. MAGRINI,Fisica C. Yie, R. Kormos, H. Borovetz, Medica, University Tor Vergata- B. Griffith and R. Hardesty, Rome-Italy, J.J. SETTELS & K.H. University of Pittsburgh WESSELING, BMI-TNO Academic Medical Pittsburgh, Pa., USA Centre - Amsterdam - The Netherlands

Beat-to-beat vailiona inheartrate and Blood rheology was examined in 14 patients arterialblcodpressureduringcontrolled who underwent from 2 to 35 day- of Jarvik- respiiatory manouvers performed by 13 heal,.;:y 7 TAHsupport as a bridge to cardiac subjects at rest were analyzed using a model transplantation. Rheological abnormali- the short-term controlled cardiovascular system, ties observed included increased RBC including nervous reflexes from both high and rigidity, increasedblooa viscosity and low pressure receptors. Arterial blood pr--ssure elevated fibrinogen concentration. Using was measured non-invasively in a finge by a the non-Newtonian Casson fluid model, calibrated servo-pletysmo-manometer (FIhiteRES). blood viscosity was expressed in terms of Mechanical effects of the respiratory manouvers ultimate Newtonian viscosity (UNV), were simulated ona computer by varying the yield shearstress (YSS) and shear rate. value of intrathoracicpressure in the Experimenta.1. resultsshowed that UNV for circulatory model. our TAH patients waswithin normal range, Model analysis of experimental results shows however, YSS was from 13 to 38% above that most of the complex hemodynamic and heart normal. Theseresults indicate that rate effects due to inspiration can be explained increased RBC rigidity and fibrinogen con- by making a single hypothesis of temporary centration are majorcausesof increased baroinhibition. Effects of expiration on heart blood viscosity, specifically high YSS, in rateappearto be weakerand mediated by these patients. btroreflexes viaa different mechanism whinh remains to be further investigated.

283

1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas 263 Biomedical Engineering Scientific Papers

BE12-B.4 BE12-B.5 Theoretical approach to the analysis of Artificial Heart Geometry Determined by passive mechanical interactions between NMR Imaging for a Patient, C.Uyama, heart chambers, M.GUERRISI & A. S.Takatani, Y.Taenaka, H.Takano and MAGRINI*, Fisica Medica- Universita Tor T.Akutsu, rational Cardiovascular Center Vergata - Rome - Italy Research Institute, Fujishirodai 5, Suita, Osaka, 565 Japan The concept of compliance of an "open elastic wall" ("wall compliance") is introduced inorder When transplanting a total artificial heart to to establish a theoretical basisfor the a subject, it should be connected without pulling analysis of global mechanical propertiesof or pushing the tissue of the remaining heart and adjacent elastic chambers sharing acommon wall. thorax wall. In order to do so, a heart geometry, Itis shown that the proposed definition of especially, geometry of the tricuspid and mitral "wall compliance" includes, as a limitcase, the valves, aorta and pulmonary artery should be classical definition of compliance of a "closed estimated to obtain a good connection of the elastic cavity".Theory was applied to the artificial heart. We took 12 transveroe thoracic analysis of passive ventricular interaction on NMR images triggered by ECG. Each NMR image was the basis of a model formulated in terms of taken 1cm apart, perpendicular to the longitudinal "wall compliances". Inordertotestthe axis. Contours of pericardial heart, ventricles, applicability of the method proposed, reported atria and inner wall of thorax were entered by data on excised and in situ canine hearts a digitizer into a computer. We calculated, by without pericardium were fitted with the model interpolation method, contour of each part of assuming non-linear wall compliances. Thus, the the heart on a nlane of every 2mm apart mechanical properties of the ventricular free longitudinally, three- dimensional heart and thorax walls and septum were separately assessed. Itis model were made by accumulating 2mm thick plastic con.-luded that the reported need to define a plate whose shapes were determined by the contours. single "effective external pressure" forone of Furtheremore, a remaining heart model which was the ventricles, besides not being theoretically made in the same way gives us sizes of the justified inthe beating heart,would be tricuspid and mitral valves anddirections of overcome with the proposed approach. these two valves and so on. These geometry thus obtained will be useful to produce the total heart suitable to a patient.

BE12-B. 6 BE12-B. 7 A CONTRIBUTION IN ULTRASOUND PULSED DOPPLER On the Transport of Sinusoidal Dens ity HUMAN ARTERIAL BLOOD FLOW MEASUREMENTS DURING Fluctuationin Catheter Flows, Q.Ye, VASCULAR SURGERIES. J.S. Lee*, L. P. Lee, and M. Hunter, V. PUGLIESI, P. GARCIA, N. BOCCI, R. RIEU, R. PELISSIER UniversityofVirginia, Institut de Mdcanique dean lides - UM 34 du CNRS Charlottesville, Virginia, 22908 U.S.A. 1, rue Honnorat 13003 Marseille, France P. BERGERON, M. REGGI, R. COURBIER ;Then a sinusoidal concentration is the input to 116pital St Joseph. Service de chirurgie vasculam. Marseille the dispersive flow in a catheter, we show mathematically that the output is the input To evaluate some criterion of quality and optimum statein attenuated by A and shifted by td.This function arterial bypass grafting, despite routine arteriography, we have Acxp( -iwtd) is theFouriertransform of the developped a perioperative methodology to measure spatio temporal transfer function characterizing the dispersion blood flow velocity profile, and thus calculating the volume flowrate, of indicator through the catheter flow. assuming that the vessel is circular and velocity profile is quasi To verify this analysis, we imposed a unit step axisymmetric with respect to the measuring plane. This technic.; input to the catheter and measured the density involves an easy to use elastic probe support setup close to the artery in a very good way positioning, and a low cost velocimeter which change to establish the transfer function of the processes the crystal signals. catheter. Using the same catheter, we sampled Emission frequency is 8 Mhz or 15 Mhz, pulse repeuuve the rabbit's 'Aortic blood and monitored its frequency is selected from 50 khz to 70 khz, doppler shift frequencyto density fluctuation. I:. this experiment,the velocity signal conversion is obtained bya zero crossing method sampling flow rate and the ventilation frequency which is adequate since the sample volume is small, maximum were varied to change the product wtm of which w measured velocity is about ± 1.2 m/s. A data acquisition and is the frequency and tm themean transit time of processing system computes in 11 to 15 channels all the numeric and the catheter. The attenuation and time delay of graphical results in order to view them in quasi real time. The system the dnnsity fluctuation was determined in take into account a cardiac pulses synchronization and breathing effects reference to the ventilation pressure. during measurements; so we had to integrate the velocity profiles for We found thatthe dependence of the attenuation and the 30 s when processing is in progress to achievea minimal dispersion value(<5%). time delay on wtm agrees with the predictionfrom In vivo measurements have been madeon 30 patients (>100 data th" Fourier transform. This agreement supports flow measurements) before and after grafting principallyat the level of the nse ofthe transform analysts to determine carotid and fe.moral arteries. the transportof density fluctuation occurring in the vascular system of the lung.(Supportedin part by HL 36285). .2 8 4

264 1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas Biomedical Engineering Scientific Papers

BE12-B.8 BE12-B.9 TheProperties of Frequency Selection of Investigation of the Controller Models of the Sane Arterial Beds, S.T. Young* and W.K. Human Cardiovascular System Wang, Yang-Ming Medical College, Taipei,Taiwan, It I Kitney7M A Chabbani, Biomedical Systems Group, Dept of Electrical Engineering, Imperial The deformation of the pressure wave asthey College, London SW7 2BT, UK travel through the circulation provides a way of measuringthe characteristics o. arterial Keywords: Modelling, Simulation, Cardiovascular System, tree. T tube model still is a generally accep- Control of Central Nervous System. ted explantion of the deformation of thepul- stile pressure waves, and it can easily explain Becauseofitscomplexity.describingthehuman a lot of phenomena of circulation system. Howe- cardiovascular system in quantitative terms is particularly ver, the property of the frequency selection of difficult . As a result, there have been a number of attempts arteries suggestedby thispresentsupplies to produce computer models of various degree of complexity more reasonableexplanation, especiallythe over the last twenty years. These models have varied from impedance patterns of arteries and the change simple teaching simulation with only few segments to very ratios ofeach harmonic ofwaves whenthey complexmodelswhichnotonlymodelthesystcrnic traveling through the aorta. This study would circulation in terms of large number of elements, but include like to emphasized a observedphenomena that cardiovascular dynamics and physiological control. It is in eachartery have different prefered frequency the area of control where significant disagreement still exists. band, andwork as filters of electricalcir- What is required is a model whit h is capable of simula, ing cuit. Highpass in supermesenteric artery, ban- physiological responses in both normal subjects and patients opass in renal artery, and complex filterin suffering from cardiovascular diseases, which includes an gastric artery were observed in this study. we accurate model of the control function . expected there would be more filtering types identified. The experimental data supported our This paper investigates some of the key controller models suggestion, and to offer a acceptable explana- available and their integration into human cardiovascular tion for some debted issues of the circulation models of varying complexity. The controller models which system. willbediscUssedarecapableofproducingaccurate simulationofcardiovascularcontrolforarangeof physiological situations including various standared validity tests.

BE12-B.10 BE12-B.11 Inner Wall Extraction of a Blood Vessel on a Computer Simulation of the Hu. Cardiovascular Coronary Cineangiogram for a Vessel Diame:.er System under Lower Body Negative -,essure

Estimation,Y.Yanagihra',T.Sugahara,N.Sugimoto, V. K. Sud , All-India Institute of Medical Sci- C.Uyama, National Cardiovascular Center -es, New Delhi, India, Research Institute, Suita, Osaka, 565 Japan. Bungo, J. B. Charles, NASAJohnson Space Center, Houston, Texas, and Measurement of a coronary vessel diameter is R. Srinivasan, KRUG International, Houston, Texas. utilized to estimation of artery stenosis. Cineangiography is the best way to show us Tin present work simulates the distribution coronary arteries in a living body. As X-ray of blood pressures and flow velocities in the scattering makes cinepictures noisy, an accurate entire cardiovascular system ofa normalhuman. measuring method is required. The cardiovascular system is modeled as an At the first step, in order to determine which arrangement of interconnected elastic vessels. The filter is the most accurate to measure blood pumping action of the heart is represented by vessel diameter on the coronary cineangiogram, time-varying elastance ,liaracteristics of the several sizes of vessel phantoms on tine film ventricles. The finite t-,ent modelis direct taken under different energy and contrast and versatile in its applicability. substance densities were calculated by using Solution of the model equations is obtained several edge detection filters. The entropy and from a knowledge of the dimensions of the vessels, the Gaussian filters were found to be twice as peripheral resistances of the arteriole-capillary- accurate as the derivative filters. venule junctions, and the rheological properties At the next step, in order to extract blood of blood. The results include changes in steady vessel area on a coronary cineangiogram, and pulsatile flows due to transition from normal locations of vessels on the periphery of image is to stressed conditions. Of particular interest is first extracted. Then, a vessel area over the the percentage redistribution of pressures and image is selected. The entropy filter is flowvelocities in the various regions of the applied to this area. The detected vessel edges cardiovascular system due to lower body negative were as accurate as that determined by pressure stress with both normal and reduced a radiologist. This method can be applied to amounts of circulating blood volume,the latter coronary cineangiograms, including stenosed sites. representing the effects of weightlessness encountered during space flight. Wort: supported by National Research Council, USA. 285 1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas 265 Biomedical Engineering Scientific Papers

BE12-B.12 Ba2-B.13 Measurements of SVC and IVC Vascular Diastolic compliance oftheleft atrium Resistance andComplianceDuring and the left ventricle in man, T. Nagano'; Cardiopulmonary Bypass in Open Heart M. Arakawa, T. Tanaka, M. Yamaguchi, Surgery T.Tsuji*, K.Suma and T.Togawa, T. Takaya, K. Kagawa, T. Noda, H. Miwa and Inst.for Med. & Den. 'ngng., Tokyo Med.& S. Hirakawa, Second Department of Internal Den. Univ., 2-3-1., KandaSurugadai Meaicine, Gifu University School of Medi- Chiyodaku Tokyo 101, Japan cine, 40, Tsukasa-machi, Gifu 500, Japan Vascular resistance and compliance during cardiopulmonary bypass(CPB) was studiedin During slow filling period, 1) theleft 21 cases of open heart surgery. atrium (LA) and the left ventricle(LV) The venous return (Fsvc, Fivc) fromthe become a common chamber, and the internal superior and inferior caval veins (SVC, pressure equals, and 2) blood flowing from IVC) were measured by an electromagnetic the pulmonary veinwillbe distributed flowmeter. The aortic, SVC and IVC according to the magnitude of each chamber pressures (AP, Psvc, Pivc) were measured compliance. We studied normal relationship by pressure transducees. of chamber compliance between the LA and The vascular resistance in the SVCarea LV in 8 patients withcoronaryartery (Rsvc) was obtained from (AP-Psvc)/Fsvc, disease.We performed biplane cineatrio- and that of the IVC area (Rivc) from (AP- graphy coup.ed with direct recording of LA Pivc)/Fivc. The systemic resistance(Rsys) pressure, and single plane cineventriculo- was calculated from 1/Rsys = 1/Rsvc+ graphy. Calculating volumes of the LA and 1/Rivc. The SVC and the IVC drainagetubes LV with area-length method, and construct- were clamped alternatively forabout 6 ing the pressure-volume relations cf the seconds. The compliance of SVC area (Cavc) LA and LV, we fitted data to the expormn- tial was obtained from Vsvc/dPsvc (Vsvc : function and determined diastolic compliance accumulated volume in SVC, dPsvc : of the LA and LV at the increased SVC pressure). The Civc was ena of diastasis (CIA°, CLva) The CLAd obtained in the same way as the Csvc. The ana CLva were 1.64 ±0.41 (mean+SD) systemic compliance (Csys) was calculated ml/mmHg/m2and 4.21 ±1.13, respectively. We

from Csys = Csvc+Civc . conclude that compliance of the LV is two Resistance and compliance changed to three time larger tnan that of tne LA. depending upon the time course of the CPB.

BE12-B.14 BE12-B.15 Increased rignt atrial contraction in patients with myocardial infarction, Simulation Studies on the Myocardial Oxigen Supply- T. Nodal M. Arakawa, T. Takaya, T. Nagano, Demand Relation During the Valsalva Maneuver. P. K. Kagawa, S. Kojima, H. Miwa, and Caminal°,M.Vallverdti°,R.Jane°, J.M.Aug", C.Creixelts ° °. S. Hirakawl, Second Department of Internal ° Institutde Cibemetica, Diagonal 647, 08028 Barcelona, Spain °° Medicine, Gifu University Scnool of Medi- Hospital Sta. Creu i S. Pau. cine, 40, Tsukasa-machi, Gifu bob, Japan Intrathoracicandabdominalpressures areseverely increased as a result of the performance of straining activities. 7uch activities may predispose to potentially In the infarcted heart, right ventr1Lular lethalarrythmiasinpatientswithcoronaryartery function is varyingly impaired, but rignt disease.In order to analyse the effects of intrathoracic atrialfunction has never been studied. and abdominal pressure patterns on coronary flow,a First, we made 6 castsof thecadaver cardiovascular system model has been developed. The right atrium(RA) ana measured their vol- model includes CNS controls of heart rate, peripheral umes with Simpson's rule, filming them by resistance, coronary resistance, myocardial contractility biplane cineatriography.The true volumes and venous tone. A simulation program has beer (x) agreed with measured volumesky), written and the model has been validated by means (y=0.94x+2.1 (r=0.99)). Second, in 9 of invivo dataunder two experimental conditions: paticnLs with normal heart(C-group) ana 9 Equilibriumconditions(Pre-Valsalvamaneuver) and patients with myocardial inf,.rction Valsalva maneuver (VM) phases(ItoIV). Intin (MI-group), we constructed pressure-.olume parameter estimation stage an optimization module has loopofthe RA.The volume expe'ied by been developed. Itisacollection of subprograms atrial contraction (ejection volume (f the tailored for use with the digital simulation language.The RA, RAEV, y) correlated isitively with effects of higher intrathoracic pressures, longer duration of the VM and two consecutive maneuvers have been volume at thebeginning v.:: atrial contraction(preload of the RA, RAVd, x) investigated using the modal.Results show that the (y=0.42x+2.8(r=U.54)).TheRAEVforthe evolutions of myocardial oxygen supply-demand relation comparable levels of RAVd was significant- during different VM patterns are not the same for all ly higher in MI-group than in C-group. We patients. The estimated parameters of the neural control conclude that the right atrial contraction are different in each patient, with a high variability. This is augmented homeometrically and hetero- is the reason why a VM may be convenient for some metrically in the infarcted heart. coronary artery disease patients but very dangerous in other patients with the same pathology. 6.; (;) CS6 266 1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas Biomedical Engineering Scientific Papers

BE12-B.16 BE12-B. 17 The Effect of Intra-aortic Balloon Freauencv DonlakAnelvsis of the Radial Alm/ Pulse;Aae Pumping on the Myocardial Ischemia in Related Changes, D. Winter, B. Mather*, Southwest Research Institute, San Antonia, Texas, M. O'Rourke, R. Kelly, St. Vincent's Goats, Z.N.Xue*, H.O.Chen, L.Li and L.R.Zhang, West China University of Hospital, Sydney, Australia. Medical Science, Chengdu,Sichuan,China We have monitored tho radial artery pulse pressure waveform Intra-aortic balloon pumping(IABP) can improve the noninvasIvely using a hign fidelity pressure transducer applied to the skin directly over the radial artery. The tranciucer has a small balance of oxygen demand and supply for injured h- pressure sensitive area (.5 mm`) with a frequency response from eart. After ligation of a coronary branch,the ser- 0 to 2 KHz. Applanation of the artery allows the direct rr lording of ious problen is genesis of cardiac arrhythmia, As In a series of anirr , tests, the arrhythmia occured,lidocaine vas used. The total the arterial pressure waveform. significant dose of lidocaine given was used as estimation for waveforms recordednoninvasIvely showed no arrhythmic incidence. The incidence of arrhythmia difference In amplitude or ph...30 when compared to direct was influenced by dosage of amobarbital. IABP vas intraarterlal recordings. started 0.5hrs after coronaryligation and maintai- Radial artery pulse waveforms were recorded nonInvasively In 1005 subjects agsr.i. 2 to 91.Pulse waveforms for each subje^t ned for 3hrs. Under similar dosage rap,,,e of this The anesthetics,the incidence of arrhythmia during 24 were ensemble averaged and grouped by age decade. waveforms in each decade were truncated to the shortest heart hrs after coronary ligation was much lover in IABP cycle In the decade 'end an average waveform for each decade group than in the untreated group. Average necrot- Normalized power spectral densities were ic weights of myocardium.herodynamic parameters a- was computed. estimated using fast Fourier transform techniques. Power nd ECG chances sere no significant difference bet- spectral densities showed significant differences between the first ween these tt.'o groups. These results suggest the and second decade, the fourth and fifth decade and the fifth and IAB). could reduce the incidence of ischemic arrhy- sixth decade.The amplitude of the fundamental harmonic thmic;. between each of the agogroups Increased and the energy ;it the higher harmonics decreased with age. In the first decade, energy in the third harmonic was significantly higher than In the second. By the second decade, the harmonic amplitudes decreased monotonically. The establis;iment of normal aging patterns allows the identification of abnormal pulses which may be related to cardiac or vascular diseases.

BE12-B.18 BE12-B.19 Three dimensional reconsbruction of the human carotid Volume change of the coronary circulation on a heart rate step, J. Dankelman*, I. Vergroesen* hi furcation,Naoki Suzann'`,RouOrMoui2,Souichirou Mochio2, and J.A.E. Spam**. Yukibide Isogai2,Mikic llou4,Sachio Miyamolo'Sinpei Tadas, *Department of Mechanical Engineering, Delft Fu"io Tsulimote.Sinpei Tadas. Tetsuo Okamura', ME Reserach University of Technology, Delft. Med.' Dept. of Radiology' Jikei Unit' Dept. of 3rd.Internal Department of Medical Physics, Academic Medical 105 Minato-ku Univ. School of Med., Nihon Kogaku K.K.4 Centre, University of Amsterdam, the Netherlands. Tokyo,Japan Intramyocardial blood volume depends on the A noninvasive measurement technique Producing three compressionby heartcontraction. The volume dimensional( 3D ) images of tht carotid artery bifurcation change of the coronary circulation on a heart has developed.Series of transverse images of the carotid ratestep has beeninvestigatedin seven arteries with 2D echography are degitalized and translated anesthetized goats. Heartratestepof30 150 to values in 3D coordinates with distance data among each beats/min were performedbetween 60 and beats/minunder constantpressureandunder images. constant flowperfusionconditions. Arterial Processed data are able to display on CRT as a wireframe (CAF) and venous flow (CVF) were measured. The model or a surface model. By this method we constructed 3D intramyocardial volume change was determined from images of the town carotid artery (CCA),internal caro- the digitized flow signals by (CAF - a *CVF)dt, tid artery (ICA),external carotid artery (ECA) and skin with g. the ratio between arterial and venous flow surface. After some clinical neasuremeat,we confirmed in the steady state. Blood volume decreased with following properties of this technique. Rotation and move- an increasing heart rate. ment technique were useful to recognize the shape of Volume changewith constant pressure bifurcation and the relative location with skin surface perfusionconditions was0.31 + 0.03 m1/100g and helpful for the vascular operation. Disection techini- (mean+ S.E.) and with constant. "low perfusion que was suited for the observation, of the inner shape of was 0.13 + 0.02 m1/100g. simulations, in which the venous vessels and the strucure of Plaques in it. Model resistance is dependent on transmural pressure, We also add the information about the distribution of could explain the difference in volume change blood velocit: a on 3D image obtained by a pulse Doppler between constant pressureandconstant flow tech-nique. These images help us to recognize the relation perfusion. with the morphological irregurality and the rheological changes in the vessel. 287 1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas 267 Biomedical Engineering Scientific Papers

BE:12-D.20 HE12-B.21 Optimal Design of Vascular Grafts DeL6.itomettie Method bon Dekivation 1,6 A. I. Rachev; Institute of Mechanics Letit VentkieutaA Ejection Fueti.on Otom and Biomechanics, 1113 Sofia, Bulgaria. Uneangiogkaphie Fitm6. M. Niczan*, D. Sapoznikov, M. Mossery, Y. Mahler and M. There exists an increasing interest in Gotsman, Jerusalem College of Technology, fabricating synthetic grafts withcom- Jerusalem 91160, Israel pliance similar to that of the naturalar- teries. The purpose of this paper is to Left ventricular (LV) volume and ejection fraction propose a method for optimal, from mecha- (EF) are routinely derived from cineangiographic nical view point, design of grafts made of films obtained during catheterization. By inject- synthetic material whose elastic proper- ing contrast medium LV volumes can be calculated ties can be controlled. IL is supposed from LV silhouettes, using various assumptions that the initial dimensions and the mecha- regarding LV geometry. However, for patients nical behaviour of the artery to which the having abnormal ventricular function, these ass- graft is anastomosed are given. The inter- umptions are not always justified. nal diameter and the wall thickness of the In cinehngiographic films, the grey level of each graft as well as the values of thecon- pixel in the LV region is proportional to the stants in the constitutive equations for depth of the LV cavity at the corresponding pint. its material are to be determined bymeans EF can be derived from the values of the integral of multiobjective optimization. Twopar- grey level all over the LV region, obtained at tial criteria of optimization are defined- two stw.cessive diastoles, assuming that: 1. The the distance between ,he functionsrepre- contrast medium is distributed homogeneously all senting the internal diameters of thear- over the ventricle. 2. The amount of contrast tery and the graft over the physiological medium at the second diastole is the sameas that pressure range and the distance between at the previous systole. the characteristic impedances. A global EF was calculated by the two methods: densitometry criterion for selection of the preferable and the area-length method. Good agreement was solution is proposed. Examples of grafts obtained, indicating that the assumptions which the made of composite material are consedered. two methods are based upon are generally justified.

BE12-B.22 BE12-C.1 RECENT PROGRESS IN CORONARY FLOW MONITOR- One-Compartmental Model and Evaluation ING. G. Tomonaga, K. Boni, Y. Ogasawara*, of A Hybrid Artificial Kidney(HD /HP K. Tsujioka*, F. Kajiya* System) Tenri Hospital, , Y.L. Chang Tenri, Nara and C.J. Lee, Dept. of *Dept. of Medical Engineer- Chemical Engineering, National Tsing ing and Systems Hua Cardiology, Kawasaki University, Hsincnu, Taiwan, R.O.C. Medical School,Kurashiki, Okayama Japan A mass transfer model with recycic(1 flow is proposed for evaluating the Great advance in medical engineering has made performance characteristics of a system composed ofa it easy to get real-time data about cardiac hemodialyzer (HD), and a hemoperfusion patients. But measurement of coronary blood cartridge (HP)solutions for the recycled flow has been very difficult in clinical fields, systems of two different sequences ofar- although it is essential to monitor the condition rangement, i.e, HD-HP and HP-HD areobtained of the patientwit:coronary arterydisease. and discussed in term of their performance In this paper, several recent progresses of of clinical monitoring ofcoronary blood flow will Theoretical computations reveal that the be discussed, especially the development in our concentration oscillation phenomenonmay be country. Wo willtalk about the results of our observed with recycled flow system under recent research: 1) 20 MHz 80 channel dual mode certain conditions. This oscillatory be- (zero -cross and Fourier transform) ultrasound havior is related to the competitivead- Doppler method andits application to evaluation sorption/desorption characteristics ina of the coronary artery flow velocity during fixed bed with recycled flow buthas noth- cardiac surgery and 2) laser Doppler catheter ing to do with the dialysis operation. and its clinical applicability. In addition, The results of mathematical analysiscon- we will touch upon somo results of other Japanese clude that during early stage ofoperation, research groups, e.g., the transesophagoal Doppler the HD-HP arrangement has higher efficiency method to quantitate coronary blood flowand for favorable adsoring solutes, andthat the combined two dimensional and Dopple. ,he HD-HD has higher efficiency for unfa- echocardiography to detect coronary arterial vorable adsorbing solutes. flow non-invasively. 8

268 1988 World Congress on Medical Physics and Biomedical Engineering San Antonio, Texas Biomedical Engineering Scientific Papers

BE12 -C.2 BE:12-C.3 ANALYSIS BY DIGITAL SIMULATION OF THE ANTIDIURETIC HORMONE (ADH) A Model for Renal Control of Long Term IN,BURN PATIENTS Blood Pressure; Richard B. Davis*, Paul Laura Roa Gomez-Cia T.,Cantero A. A. Taiganides*, Herman R. Weed E.T.S.I.I., 41012-Sevilla, Spain. A model has been introducedwhich simulatestheresponseofthe renal Burn injury provoke a protein and liquid system to changes in saltintake. It isbasedonthe renal function curve shifts from plasma to interstitial compart which describes the relationship ment, so plasma volume decreases. ADH in- between salt intake and blood pressure. crementz water reabsortion at renal collec ItIs proposed that the renal system ting ducts and diuresis decreases. is the sole control ofbloodpressure ADN control mechanism has been incorporated inthelongterm. Themodelalso to a mathematial model for digital simula- offers an explanation for salt tion of burn injury. ADH secretion as a sensitiveand insensitive hypertension function of nervous stimulus, extracellu- by analyzing two types ofpathologies: lar osmolality and circulatory mean pressu reduced renalfunctionandreduced re has been quantified. renalpressure. The normal model The simulation results have been tested limitsthebloodpressuretoa 40% against experimental and clinical data, increase despite a 2400%increasein showing a reliable model behavior. salt intake.Inducing a 70% failure of Diuresis analysis has importance in burn thekidneys,themodelbecomes salt patients by its facility as clinical moni- sensitive, allowing the bloodpressure toring parameter of patient response to tovary 90% forthesame change in reanimation treatment. intake.With a 50 mmHg dropinrenal pressure, simulating stenosis,the model increases the blood pressurebut maintains its salt insensitivity, limiting the blood pressureto a 25% change due to salt loading.

BE12-C.4 BE12-C.5 HAEMODYNAMIC MODELLING OF Intracellular pH in Regulation of Gap Junctional THE RENAL CIRCULATION ConductanceinFrog Proximal RenalTubules. IN THE MANAGEMENT OF TRANSPLANTATION W. RehwaidA. C. Hallbrucker.and F. Lang C R Deane*, D Taube, D Goss, V C Roberts Dpts. of Medical Physics & Physiology, University Department of Medical Engineering, of Innsbruck,A-6020 Innsbruck, Austria King's College School of Medicine, LONDON To test for the influence of intracellular pH (pHi) on Experience with themeasurementof gap junctional conductance in proximal renal tubu- vascular hydraulic impedancein the les, three conventional microelectrodes were impaled reconstruction of theperipheral into different cells of the same tubule and potential vasculature, and the use of these difference across the peritubular cell membrane (PD) measures inconjunction with lumped parametermodels of the arterial continuously recorded. Current was injected into one circulation, has now been applied to the cell and PD deflections determined at the other two. problem of monitoring renal The results allowed calculation of the lumped resis- transplantation. tance of luminal and peritubular cell membrane (Ro = 53 ±7kQcm) as well as the resistance of the cellular Measurements have beenmade, using an core (Re=164±14MO/cm). pHi was recorded with pH Acuson Colour Duplexsystem, of blood sensitive microelectrodes. When extracellular pH is flow velocity inthe renal artery, reduced from 7.8 to 7.4 by increasing CO2 from 1to immediately following transplantation, 2.5%. pHi decreases from 7.74 ±0.04 to 7.45±0.04, PD and every two daystherefter for a decreases from -62±2 to -44±3 my, Raiincreases by maximum period of 10 days. In cases 31±6% and Rc by 31±11%. Inhibition of Na/K-ATPase of rejection and nephotoxicity, these by 0.1 mmob 1ouabain reduces PD,alkalinizes the flows havebeen correlatedwith the cell, increasesRCM but does not significantly increase results of biopsy. This data has been Rc. 1 miro1/1 cyanide depolarizes thecell and in- used to establish atentative model of creases toth, Re,and Rc. Potassium free extracellu- the renal circulation which we believe lar perfusates hyperpolarize the cells and increase would be of value in predicting impeding both, Riean Rc. 100 umoUl mepacrine depolarize and rejection without the need for more alkalinize the cell but increase both Me and Rc. In invasive measures. conclusion,Rcis sensitiveto alterationsof OIL However, Rc and pHi can be dissociated pointing to additional determinants of gap junctional resistance. 899 1988 World Congress on Med,cal Physics and Biomedical EngineeringSan Antonio, Texas 269 Biomedical Engineering Scientific Papers

BE12-C.6 BE12-C.7 INLET OXYGEN-SUPPLY HEMODIALYSIS(RO CASES-TIME REPORT) THE EXTRACORPOREAL CIRCULATION AND P. WANG DA -YONfl et at ULTRAFILTRATION FIRST HOSPITAI OF HUBEI MEDICAI OF Asrair FLUID - INTRODUCE 1 NEW TECIINNUE OF TH1 COLLEGE, WUHAN. HUBEI. CHINA. TREATMENT FOR OBSTINATE ASC!TIC Some having chronic renal failure patients have P. 4AVG DA YONG:Pt at been proceeded in 80 cases-time by FIRST RnSPITAL OF HUBEI %DUAL treatment of he- COLlEGF. WUHAN. HUBEI. CHINA modinlvsis with 140, in eialsate. The hydro-dioxide ranlippl) oxygen to the blood directly through (ABSTRACT) the According to the priruipte of ultrafiltration. dialyse membrane.it called inlet oxygen-supply heron authors designed a new technique of asciticextra- dialysis. The biologica1 reaction isa catalysis owing to the hydra - dioxide corporeal circulation and ultrafiltration. enteme in the blood to The medical engineering principle of the work. extra- the hydro-dioxide decomposed intowater, oxy- corporeal ultrafiltration of ascitic isan extra- gen and thermal energy quickly 211,0i 0, 1-474 corporeal material tranfer. The ascitic, through the Cal) The most of these oxygen integrateshemoglobin fiber line of dialyzer and (leaningpartly. As the conditinn of dialysis is stable state. to form oxygenated-hemoglobin,that improves PaO, the cleaning quantity of ascitic is directly proportional and San, concentration and provide thetissues to to the use,the few of other dissolve: in the pressure of ultrafiltration (IMF) and the time of plasma and ultrafiltration. pases the pulmonary circulation through the blood- The average cleaning rate of urea(mlimin)= oxygen defence regurgitating diffusible to breathout. According to the expirement. wheninlet oxygen - Cb supply hemedialvsis, the safe amount ofi1,0, using (annotation): The urea concentration of the is VH,D, (0.1% ?Wv0.98). This numerical ascitic in front of begining ultrafiltration(mg%). valume (5-10% of all blood rapacite) is in the scope C.: The urea concentration of the ascitic after the of the body's capacious vessel thatcan be regulate, so itis proper. end of ultrafiltration (mg').(1W: The average discharge of the ascitic at the course of the The inlet oxygen-supply hemodialysismakes PaO, ultrafiltration. and SaO, remarkable rising,so the complicate rate of the dysequelibrium syndrom inthe dialysis is Not only this method can relive the suffering low. in addition,this therapy can provides an effec- by eliminated the ascitic largely andpromptly, tive way that it surmount thelung to supply the patients had not the evil of aggravating failure oxygen directly to the blood, it can be improvethe owing to the protein discharge, but also thedoctor had not the worry that have to frequent hypoxemia promptly, to retax the cardiafailure and ultrafiltra- pulmonary edema, and to remove the poverty-stricken tion, because the output quantity of the ultrafiltration is more than the increasable quantity of symptom of the respiration, so it has thefine effect the of the cure for the uremia that worthto promote. pausal period of the ultrafiltration.

BEUI-D.1 BE12-D.2 Downsizing and Clinical Engineerinf!, Strategies in the United States ADMINISTRATIVE COMMITMENTS FOR CLINICAL ENGINEERING DEVELOPMENT, Michael J. Shaffer: D.Sc., M.J. Shaffer, D.Sc.* and M.D. Shaffer, The George Washington University, Washington, M.S., The George Washington University D.C., U.S.A. Medical Center, Washington, D.C. High technology and nanaged U.S. hospitals are facing cost-cutting health care in the United States is challenging pressures due to new reimbursement prac- clinical engineers to expand and developnew oppor- tices, increased reliance on out-patient tunities. clinics and pre-paid systems, and As a prerequisite, clinicalen- increas- gineers need determine the ed competition from other hospitals. Oc- extent of hospital administration commitmentto their sup- cupancy data shows in-patient care ison port of: the decline. Hospitals are thus looking a. personal computer consultation, to reduce costs and develop new lines of program- business. ming and maintenance servicesat the clinical department level, b. maintenance integration,with automatic In these changing conditions clinicalen- funding of expensive spare gineering strategies include parts kits and adjustment of maintenanceoperating a) maintaining the integrity of theser- buuget with the purchase of vice through professional associations capital equipment, with service manualneeds in- and standards, performancemeasurements, cluded in purchase orders, and comparison of departmental values and with service contracts regularlyreviewed, with those of other providers, and c. in-service technical training b) contributing to new business prioritiz- opportuni- ed and scheduled bymanagement, ties such as shared services, technolo- d. staffing adjustments for gy assessment, microcomputer applica- specialized tions, and training. technology assignments, and e. unrelated business incomeincluding tel- ecommunica,-ions voice/data,switching and cabling issues, sharedservice maintenance, etcetera. 9 (1 0

270 1988 World Congress on Medical Physics and Biomedical Engineering San Antonio, Texas Biomedical Engineering Scientific Papers

BE12-D.3 BE12-D.4 TECHNIQUES FOR ACHIEVING CREDIBILITY WITH A Financial Simulation Model for HOSPITAL ADMINISTRATORS, G.I.Johnston Clinical Engineering, A. Mahachekl and R.L.Morris; Oregon Health Sciences Univ. The Johns Hopkins Hospital, Baltimore, Maryland 21205

Hospital administrators have difficulty assessing the impact of the enormous potential of a disci- Budgetary performance is becoming as pline, clinical engineering, for which thereare important as technical performance few management models.Furthermore those which for many clinica: engineering have bean developed are written in "engineer- departments. Building afinancial ingese" and reported in the clinical engineering model using a personal computer literature rather than the hospital administra- spreadsheet software package is an tion literature. They have even more difficulty ideal way to prepare a budget, track allocating adequate resources which would enhance actual performance, and test policy the value of the department to their organiza- or personnel changes. The clinical tion. Any department not census driven is auto- engineering manager gains new matically suspect! Since it is unlikely adminis- insight and understanding of the trators willhave, or take the time to,learn department's operations while more about this strange new department,it be- building and manipulating the hooves tt: clinical engineer to seek methods for model. The existence of a financial achieving the support andrespect of hospital simulation model instills confidence administrators. Clinical engineers can(must) and projects competence when work both inside and outside their hospitals to communicating with hospital senior improve their credibility and status with hospi- administration. This customized tal administrators. They must begin publishing financial model is in actual budget their management articles and models in journals use (for 3 years) at a major which hospitaladministrators read. And they teaching hospital. Also described must developinternal reporting andbudgeting are the simulated changes in policy systems which hospital administrators can relate and/or personnelthat resulted in to. Some suggestions will be presented and dis- actual management changes in the cussed. clinical engineering department.

BE12-D.5 BE12-D.6 Development of "wide use" PlID(Personal Health ASSESSMENT OF BIOMEDICAL TECHNOLOGY Data) management system. 11.1shikawa*, II.Takeda, IN A UNIVERSITY HOSPITAL IN NICARAGUA, Robert P. Patterson*, UMHC Box 297, H.Inadd, 1.1noue.dnd S.Sakai, Kagawa Ha. School. Univ. of Minn. Hospital, Minneapolis, MN 1750 1Kit:, liki Cho. Kagaha. 761 07. Japan 55455, USA Tn Nov. of 1987, a medical team consist- WP are developing a 11111(per:wna1 heal th data) ing of physicians, nurses, a respiratory therapist, a biomedical engineer, and a management system hhivh hill monitor the indii.idual's rehabilitatior engiheer visited the health throughout his life. University Hc.:pital and medical school in Leon, Nicaragua to assess needs and The basic system configuration is organized around tht, provide assistance. The hospital had 360 beds and the medical school graduat- personal computer with d new audio visual telephone. ed 250 students annually. Some of the IC memory card system link up. The health information obvious deficiencies were no working is qualified into three levels, which are "personal ventilators, no working ECG or blood pressure monitors, and no blood gas ID informations", "basic physical informations" (i.e. measurement capability. The repair blood pressure, pulse rate, body temperature etc.), technicians faced a lack of parts and Spanish manuals for U.S. made equipment, and "technical informations" (i.e. medication, lack of a medical understanding of the surgical specific medical histories etc.). The IC function and use of the equipment, and high ambient temperatures. Three dif- memory card will be used as the major information ferent brands of U.S. made ECG monitors transport media. The development of a PhD system is failed apparently due to the high tem- perat res in the non-air conditioned not only concerned the net, health information hospital. Two of the monitors were restored to use by installing fans. management system, but also contributes the formation In one of the top hospitals in Nicaragua, the of the regional synthetic medical service system. level of technology available was low.

2 9 1988 World Congress on Medical Physics and Biomedical Engineering San t otonio, Texas 271 Biomedical Engineering Scientific Papers

8E1271).7 BE12-D.8 HEALTH CARE TECHNOLOGY - RELEVANCE TO THIRD WORLD NATIONS - T. G. Krishna Murthy* and D. V. S. Raju, MAINTENANCE OF BIOMEDICAI EQUIPMENT- T.O. Krishna Murthy Bangalore, INDIA. and N. Krishna Swamy, Bangalore,INDIA The impact of advancements of technology Maintenance and utilization ofequipment used in Particularly in the last decade has been medical institutions and hospitals phenomenal. Instruments utilising mini and have been a micro computers have led to optimum major problem in all third worldnations. Various utilisation of the very limited resources factors have led to the under utilizationand non- utilization of equipment. available for health care delivering The major factor has programmes in the highly populated and been the total lack of essential back-upfacilities poverty ridden third world nations. in the form of trained manpowerand availability of Planning and Co-ordinating of such spares, majority of equipment are importedand due Programmes have to be based mainly on the to a very rapid rate of obsolescencein developed disease pattern prevalent and actual needs. nations, the suppliers in majority ofinstances regret to provide the badly needed So a need based approach for theappropri- spares. Health ate utilisation of technology is inevitable. care is the responsibility of the states andthere The %HO group on GPAHCT located in appears to be total ignorance about the need of Copenhagen has initiated steps in this trained manpower to ensure optimumutilization of regard and are building up their database equipment. However, in the last decade,a number at the university of Linchoping. The of service and maintenance centershave been set target of HEALTH FOR ALL BY 2000 A.D. is up. In India, plans are under way toset up the a really formidable challenge. Besides medical instruments maintenance corporationson a modern medicine, local systems of medicine national level to apply necessarycorrectives. whose clinical therapeutic potential has to beevaluated in a planned and scientific manner along with YOGA therapy can be gainfully utilised to provide medicare at very moderate ccsts to the millions even in the remote rural areas in the third world nations.

BE12-13.9 BE12-D.10 Audiovisual Material aimed at the Prevention of+ Protocol for Maintenance and Repair to Clinical Accidents, R.P. van Wijk van Brievingh Suction Apparatus potentially exposed and F. Koornneef, Delft University of Technology, to High Risk Infections. D. Naughton, NL-2600 GA Delft, the Netherlands. B. Murphy, L. English, J. Mahady and J F Malone; F.D.V.H., St. James's Hosp.& Accidents in hospitals have to be analyzed in a Trinity College, Dublin 8, Ireland. system-analysis approach. The subsystem "Human- ware" is the most difficult to describe. Human Suction apparatus is commonly repaired or subject error, however, is governed by internal proces- to schedule maintenance in hospital Clinical ses which can be mastered by awareness of criti- Engineering Units. It is possible for body fluids cal situations ana by organizational measures. to be drawn directly into the machine, including Before this is possible, a positive attitude to- the pump, thereby contaminating it. Under normal wards the discussion of accidents and incidents circumstances the associated maintenance problem has to be breeded. The use of audiovisual aids is unpleasant and time consuming, as the unit is of value here, as by the repeated discussion must be dismantled, cleaned and reassembled. of situations shown, objectivation as well as However it is no longer merly unpleasant, it reference to own experiences is induced. After may be potentially lethal, given the associated an introduction of a general system model with increased risk of exposure to a Category 111 indication of the entities and relations, Virus the such as Hepatitis B or HI". To cope with the background of the "attention problem" is iluci- above we have established a protocol to protect dated with reference to every-day hospital situ- Biomedical Engineering Staff who must perform ations. For example, in the operation theatre repairs/preventive maintenance on suction units. the team has to rely on their own resources as The protocol is both Preventive and Procedural well as on what the "outside world" has prepared. in that it (a) aims to prevent faults/behaviour Awareness of the limited attention capacity has that will result in the ingress of fluids) and to determine task-division and protocol. (b) sets out a formal procedure to be followed Attitude changes have to be supported by quality- during transport and repair of the unit that assurance measures by the management. Part of address the infection risks if the unit (i) need these will be registration and proper discussion not be opened, and (ii) if it must be opened. of accidents and incidents in a "guilt-free" situation. 2f32 272 1988 World Congress on Medical Physicsand Biomedical EngineeringSan Antonio, Texas Biomedical Engineering Scientific Papers

BE12-D.11 BE12-D.12 MEDEQ - A Medical Equipment Performance Indicators for Equipment Information System, C. Nerhed*, L. Servicing Teams. D Whelpton*, King's Lofstedt and B. Eklund, Dept of Medical Engineering, University Miil Hospital, Sutton-in-Ashfield, Hospital, S-751 85 Uppsala, Sweden. Nottinghamshire, UK. MEDEQ is an information system for medical equipment and it has been used at the Univer- All departments of the British National sity Hospital in Uppsala for almost ten years. More than ten thousand equipments are stored Health Service are required to show in the database and it is used by over 30 efficiency. Accordingly, maintenance persons on a daily basis. MEDEQ is also used by six other ht.spitals in Sweden. MEDEQ has departments are increasingly required to been developed in a VAX/VMS environment using measure performance. It is difficult to the relational database MINER. A fourth generation programming tool has been used to make valid cost comparisons. In any implement the system. Great emphasis has been case, it may be that financial cost is made to make the system versatile and easy to use. MEDEQ consists of functions for inventory, only one of several parameters to be maintenance and cost analysis. The data stored are used for a variety of purposes e.g. inven- considered. Whilst in-house servicing tory, investment plans, and to find weaknesses must be competitive, other criteria, and security risks in the equipment. such as rapid response, availability of staff education and advice on when and how to replacement equipment are also important. Experience gained in operating an equipment management system will be discussed.

BE12D.13 BE12 -D.14 EHC Heasurement and Evaluation in The NorwegianInspectoratefor Biomedical HE Equipment, Z.Ohta: NEC San-ei Equipment and the Natismal System for Accident Instruments,ltd., Tokyo, Japan. Reportabd Investigation, Johansen A.*, F. Nilsson, E. Fonstelien, STEM, 1320, Stabekk. As a result of recent increase in the use of HI equipment. the consideration of EHC (Electromagnetic Compatibility)in Norway (4.o mill inhabitants - approx 100 dif- hospitals has become very important for the safety of pati- ferent hospitals of a high medical and tech- ients. ENC is classified roughly into LHI (Llectro-magnetic nical standard) has had compulsory governmen- Interference)and Immunity. In Japan.we organized a committee tal inspection and control of electrical medi- for study of EHC in HE equipment underthe Technological cal equipment in health institutions since 1980. The inspection has been carried out by Committee of Hedical Electronics (Chairman:z.ohta). EIAJ( 8-12 highly skilled engineers on the basis of Electronics Industry association in Japan),with the assis- a comb stionof a total and a random selec- tance of the Hinistry cf International Trade and Industry. tion o' atient rooms and/or equipment cathe- EHI was measured in 9 kinds of equipment in accordance with gories. CISPR/B(S)42. Some equipment exceeded the limit for the Since1980 there alsohas been an enforced immediately report, radiated field strength. Immun.ty was measured in 5 hinds obligation to day and night, any incidents or near-accidents where of equipment in accordance with some standards published in electrical medical equipment may be involved. Japan and abroad,because no international standard is avai- The Inspectorate investigates accidents in the lable for Immunity now. The ability of rejection differed actual hospital if necessary. from equipment to equipment, showing the property of each Asteady decrease (around 40% to 10% of the engineering technique. equipment installed) of the faults and the defects found and a parallell increase in the To satisfy the standards,it is necessary to improve the number of accidents and near-accidents repor- environment such as the shield of equipment especially for ten is elucidated and theoretical and practica the therapeutic equipment, as well as the technical design explanations are proposed. of each equipment. The administrationof the inspection and the reporting system will be thoroughly exp)ained.

6:: Pi :3 1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas 273 Biomedical Engineering Scientific Papers

BE12-E.1 BE12-E.2 A Computerized Strength EvaluationSystem A Computerized ERG Monitor for Children using the Cybex-IIIsokinetic Shiro Usui,Ei -i chi ro Nagasaka* Dynamometer. C.A. Rodriguez*, D.S. Glass+, Manabu Sakakibara and Yozo Miyake+ *Texas AO University,47exas Scottish Rite Hospi- Toyohashi Univ. of Tech., Toyohashi440 tal. and 'Nagoya Uni., Nagoya 464 Japan

A PC-based control system for theCybex-II Iso- kinetic Dynamometer is being developedto assist A computerizedportableERGmonitor was in the evaluation of strength andendurance of developed for a clinicaluse with the aid of knee extensor and flexor muscles. The system modern digital signal processingtechnology. will be used to observe thecourse of the disease It isknown thattheERGis a sunumative and the effects of surgery, drug therapyand potential composed of a-,b-wave and oscillatory physics/ therapy in children with neurologic and potential, which are sensitivetovarious typesof orthopaedic disabilities. Several computer- eye disease and thus the amplitude,latency and controlled isokinetic evaluationsystems have inter-peak interval of eachcomponent provide a been developed in the past, but primarily for good cue for a diagnosis. However,each component adults in sports medicine and rehabilitation. is being mixed together both in time and The computer system controls the speedof the frequency domain and thus itis impossible to exercise, and collects torque and position data. isolate the components by theconventional methods The data is plotted in real time and such as an adjusting time biomechan- constantof amplifieror ical parameters of patient performance (e.g., a so-called manual method. torque-to-body weight ratio, work, timeto fat- Here wedevelopeda totally igue, maximum torque, position computerized at maximum torque) portable ERG monitoringsystem which is able to are determined. isolate oscillatory potentialsand detect each parameters automatically with utilizingdigital filtering, techniques suchas moving average and digitaldifferentiation.Thiscomputerized electroretinometer was tested in daily clinic and found enough beneficialto ophthalmologists with easy use and high precisions.

BE12-E.3 BE12-E.4 A COMPUTER AIDED PROCEDURE FOR THE Measurement MANAGEMENT OF DIABETIC of spontaneous PATIENTS, anaesthetic parameters by a E.G6mez-Aguilera* and F. del -Pozo, Dpt de Bioingenieria, E.T.S.I.Tele- microcomputer system, M.A.Tooley, S.A.O'Connort and D.J.Wilkinson, comunicacion,Madrid, Spain. St. Bartholomew's Hospital, London, England. A new procedure, runningon z PC-compatible, has been designed to helpthe diabe- tological training of patients and phy- A microcomputersystem for the measurement of sicians; and facilitate the precise, and error free, therapy selection procedures. parameters of ventilation and bloodpressure in the spontaneously breathing To allow aparsimonious approach to anaesthetised the patient is described. The system task of defining the optimumtreatment continuously for each patient. calculates minute volume, respiratoryfrequency and mean end-tidal carbon dioxide This therapyselection tools concentration andthe averaged over a patient follow-up graphic-analytical minute. Also four aids, cardiovascular parameters aremeasuredwhich are supported on a minimum metabo- are lic model for the estimation updatedevery60 seconds. The inputs are of blood obtmaned from glucose, that works on the a wet-wedge spirometer, a information capnographand a provided on diet, insulin sequence, and blood pressure monitor. The methods of computer calculation are described exercise regime . Thus, this procedure, in detail. The computer results are compared during trainingsessions or, in real with those obtained by manual follow-up situation, can be used for the calculation and areall in agreement to within 3%. The detection and identification of thecau- system ses described has the capacity to record and (including behavioural sequences)of display any metabolic accident or the a wide variety of further measurements evaluation made during anaesthesia. This of therapies; which will finally provide systemprovides invaluable information for to the patient a progressiveknowledge of research and her/his own teaching but may also be useful during routine illness, anda treatment anaesthesia. individualizedandadaptable to daily circumstances. Acknowledgments.-CICIT. Grant No.2911-83.

274 1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas Biomedical Engineering Scientific Papers

BE12-E.5 BE12-E.6 Intelligent Computer Control of Blood Measurement and Analysis:Mechanics Pressure on Induced Hypertensive Chemotherapy for the Cancer Treatment of Body Sway, Din Jianxin, Wen Xiao- Y.Fukui*, M.Suzuki yang* Huazhong University of Science *Tokyo Denki Unuv., Saitama, Japan Tohoku Univ., Sendai, Japan and Technology, Wuhan, P.R.China

Chemotherapy with Angiotensin II (AII) induced Reported a new mothed to research balance hypertension has been expected as a powerful cancer treatment in recent Japan. The idea of mechanics of human.A moritoring system of the therapeutic strategy is based on the experimen- body sway was developed.Body sliay inform- tal findings of the micro-circulation:"AII infusion causes the vaso-constrictive effects on the normal ation was obtained from the crystal rate blood vessel, but not the tumor. As a result, gyroscope and the data were processed by for AII induced hypertession(150 mmHg mean arterial pressure), most of the blood flow to the tissues TP-801 single board computer. The charac- pass through the tumor, which enables the effective tors of body sway in normal state and delivery of the anti-caLcer drug to the tumor ataxic patients were given. selectively." The current problem of the clinical hypertensive A series of new results on body ,alance chemotherapy is how to perform the safe and smooth mechanics based on PDS ( power density elevation as sell as the steady regulation of the blood pressure by the proper AII infusion. spectrum technique to analysis Romberg To overcome this, intelligent closed-loop control posture( orthostatic reflex ) were found. system of hypertension has been developed. Intelligent means in a sense that the computer behaves like the e.:pert physician's treatment way, where optimal control and sensitivity adaptation algorithms are used to infuse the AII automatically. Satisfactory blood pressure control were realized on the animal (100 rats) experiments, which may be applicable to the clinical cases.

BEJ2-E.7 BE12-E.8 A Systems Approach to Microsurgery The Vision-Aided speech Training System Based Steve Charles, M.D., 6401 Poplar on Avenue, #190, Memphis, TN 38119 the Microcomputer for the Deaf. J.Yao. Z.H.Gao and R.C.Chen, Shanghai Jiao Tong University. Ophthalmic microsurgery has progressed Shanghai. P.R.China. from hand-positioned and hand-actuated to hand-positioned and remotely-actuated ocular instruments. The next revolution The speech training sysfem can be usedto help the people will produce remotely-positioned and re- who can not speak because of congenital loss of hearing motely-actuated instrumentation, i.e., to train pronunciation themselves. microrobotics, for dexterity augmentation. The speech training system A parallel revolution in the downsizing proposed in this paper is a of tools is occurring, resulting in the new type of visible speech training device.whichis based need for MEMS technolcgy. Ophthalmology on the technique of isolated word speechrecognition. The is a machine-dependent specialty embrac- system ran compare the pronunciation produced bythe deaf with the ing both the patient-machine interface standard pronunciationto decide if the input and the surgeon-machine interface. A pronunciation is ccrrect and how its qualityis. And then multidisciplinary approach to systems in- the system can convert these informationinto the visible tegration has been developed for the signals.to instruct the deaf to do training. ophthalmic microsurgery operating room The mayor difference between this visible speecn training system and the office, utilizing technology and the other now available ones is that the system we transferred from military applications. developed These include expert systems, speech does not use the features extractedfrom the recognition and synthesis, programmable speech waveform, but directly usesthe results of speech pump and valve arrays, and virtual con- recognition as the visible signals. Thus. this system is more convenient trol and display. The computational to use and provides an environment of aspects of the operating room system de- self-tutor for the deaf wno have some knowledge in pronunciation. veloped, referred to the Ocular Con- The simulated systemof this kind of speech nection Machine, consist of machine man- training device has beenaccomplished on the IBM-PC/XT agement functions and information man- microcomputer. agement functions.

e) r- A./

1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas 275 Biomedical Engineering Scientific Papers

liE12-E.9 BE12-E.10 Thl Engineering on Computer Diagnosis and IDENTIFICATION OF THERMAL MODEL FUR HIMII Adjustment of color Perception TISSUE Abnormality DURING ULTRASONIC HEATING & ITSMICPCCOMPUTER-CON- Chen Xiaoguang , Wei Jizhou, Liu hongjian, WALLED MEASUREMENT DEVICE MIN-HE CVENG* AND HONG- and Jiang Chunxiao, Bethune MedicalUni- ZHANG WANG, Precision Interment Dept, ShanghaiJiao versity, Xin Min Avenue 6,Changchun, Jilin Tong Univ, Shanghai, P.R. China P.R. CHINA Hyperthermia by ultrasonic, asa method for cancer treatment, has been applied to clinic trial for Based on a vast amount ofexperiments, the many years. Its effect has been confirmed.But a non- topological model for human colorperception invasive method of monitoring temperature is developed. distribu- It is discovered that its tion during heating still remains unsolvedso far. process is surprisingly similar to thatof Present method of temperature estimation dependson the, signal processing in colorTV systems. skin temperature and doctor's experience. In this hence it is possible of theadjustment of paper, a simple model structure of temperature dis- its perversion. We have developed micro- tribution in human tissue, which is basedon one- computerized device which output datasuch dimension bio-heat equation, is developed. The model parameters for real tissue as the classisification of the abnormality are estimated from patient treatment through the and its level, and the optimum use of a system iden- adjustment tification technique. The temperature distribution curve. By Auto CAD operation system we de- during hyperthermia can be predicted bythis model, veloped a software including 35sepectrum and non - invasive temperature measurement,can be re- carts for perversion checking. Thecharts alized. To verify the feasibility of this method, can be used by will and displayedin high the simulation temperature field is madeby computer. resolution. In the process of man-machine The result is more satisfactory andindicates that the simple control model structure is dialogue, the oplimum effect achieved,then adequate the device will display the through identification of phantom modelparameter. diagnosis and An identification-prediction measurement the adjustment curve in print form. device is Finally, made by microcomputer TP-801 andbetter results are and appropriate adjustment glasses willbe obtained by experiment heating fresh excisedpig made for each client, tissue. It's possible to apply this method for clinic non - invasive temperaturemeasurement during hypothermia. But there are still a lot of work to do.

BE12-E.11 BE12-E.12 3D/Multiplanar CT-Guided Stereotaxis: An Adaptation System- Henry Ford Hospital -MI. Computerized rhinomanometry: equipment Zamorano Dujovny M, Flynn M, Block R. and models

3D/Multiplanar data obtained from CT interfaced Fi lip Versnick (*), Marc Nyssen, Peter Clement with stereotaxis is becomingan ii,dispensable tool in the diagnosis and treatment ofbrain lesions. ENT dept. Academic Hospital Jette and Size, volume, shape, axis of lesions,definition Medir,a1 Informatics Dept. Vrije UniveisiteitBrussel of coordinates, angles, andareas can be obtained B-1090 Jette, Belgium and corrected with the aid of multiplanar or 3D Rhinomanol,etry is a technique to obtainan objective and quan- images. A major goal is the transposition of these titative value of the functioning and/or theobstruction of the imaging generated data into thestereotactic space. nose. An adaptation of the CT compatible frame"Komai" is During the experiments, the transnasalpressure difference is presented which allows transfer of30 /Multiplanar CT generated data with high grade measured simultaneously with the correspondingtransnasal accuracy (+-1m4 airflow, for each side of the nose. An adapter to the scanningroom table was designed allowing isocentric relationships to thegantry and We designed a system by which the analoguepressure and parallel to the scanning plane. The designwasmedi corresponding flow signals, generated by therhinomanometer of fiberglass avoiding CT artifact andallowing a are sampled at a definite frequency and digitized bya real-time complete range of scanning so that with inverted computer system, connected to the local UNIX workstation. positioning of the frame it is possible to aim tar- The workstation handles the mathematicalprocessing but also gets near the skull base or posterior fossa. An provides a lot of graphic possibilities,to visualize the results. angular localization system attached tothe Xscale In th's paper, we present the methodology allows translation of plotted trajectoriesof re- that allows investiga- formatted or scoutviews into the Y-Z plane.A tion of rhinomanometric characteristics indepth, such as fre- computer program allows considerableaccuracy in quency content, determination of th?, optimal sample frequency. the translation of CT data into stereotactic space The long term goal is the development andtesting of mathemati- with special reference to preselectedplotted cal models that describe transnasal pressure/flowcurves and trajectories. These modifications and adaptations parameter analysis, to understand and determine the enable performance of diagnostic and characteris- treatment of tics of transnasal breathing in view of future brain lesions. clinical applications. The first results are presented. 286 276 1988 World Congress on Medical Physics andBiomedical EngineeringSan Antonio, Texas Biomedical Engineering Scientific Papers

BE12-E.13 BEI2-E.14 Mult/cal Sitznal Analysis Actznizition A DISTRIBUTED MONITORING SYSTEM FOR S.-stem V. elina B.A. J. Aqiiroz L. ARRHYTHMIA AND ISCHEMIA MONITORING Tlniversidad Autonoma Metropolitana-17.t. G. Passariello*, F. MorkJ. L.Coatrieux(1) and Av. Artsima Y Michoacan M&Alco 09340 D.F. 1).Almeida(2), UniversidadSimonBolivar, MEXICO Grupo de Bioingenieria y Biofisica Aplicade. Apartado 89000, Caracas, Venezuela 1080 We have de.signel a system fir theacquisition andanalysisof biomedicalsignals, based on anTTII-P(1 (XT orAT) compatible cortronter.Tt Recent development and demonstration of efficacy of consists of a16 channel. 12 bit A/D converter antiarrhythmic drugs and other cardiac agent, have can& designed ail built atam laboratory. aria suggested the need for more quantitative, individualized a digital signalprocesing package which diagnosis end treatment, increasing the complexity of contains such fiwictions as signalacquisition. instrumentation systems.Current advances in f lter ing, fast.Fowl ertransform. microinformatics. hardware and software, digital signal spectral density. anl cornrolution. togetherwith processing, pattern recognition and artifitial intelligence, sem al moles of displaying the signals. have resulted inthe possibility of developing more sophisticatedcardiacmonitoringsystems, without '711P menndriven. modular design of thesystem necessarely involving a large centralized computer by permits the easy addition ofother signal ant distribution of the intelligence from the bedside unit to the analysisroittines,aswell as the easy central station A general philosophy of a distributed transference of sane ftrictiorks to othersoftware intelligent system for a Coronary Care Unit (CCU) is packages.The software was written in pascal. describedThisincludes:ECG analysisfor arrhythmia with several routines in BOBB assemblylanguage. detection and ST segment quantification, event counting, ti plotting, and patient supervisionduring drug Applicationsof this system now incluie the therapy The tasks mentioned are integrated withinan spectraldensity analysis of electr nexicephilo- expert systemAlai the development of an intelligent graphicsignals for train mapping ani the bedside unit for two beds bas: -n a personal computeras a acquisitionanifilteringoftranstoracic preliminary step toward the complete system is presented. impedancesignalsforthe determination of cardiac output. (1) Thuversite de Rennes I, Groupe de sigitaux et Imagesen II6dicine, Rennes -Francis. (2) HospitalUniversitario deCaracas.LTtudaddeCuidados Corotierios, Caracas -Vettecuele.

BE12-E.15 BE12-F. 1 PC-based equipment for the assessment of Estimation of turbulent shearstressesin puls- strength during various hand grips. atile flow downstream of aortic valves. 0. Lindell , T. Backlund, S. Karlsson, T. Stiderberg, University Hospital, S-901 85 H. Nygaard*, M. Giersiepen, J.M. Hasenkam, UmeS, Sweden. D. Westphal, P.K. Paulsen, H. Reul. Engineer- ing College, Arhus, Denmark. Dept. of Thorac. & Cardiovasc. Surg., Skejby Sygehus, Aarhus Endurance and motor skills of the hand grip are University Hospital, Denmark. Helmholtz-Insti- important parameters in the pre and post operative tute for investigation of patients with injured hands. We Biomed. Eng., Aachen, FRG. have constructed an equipment based on a PC to Measuring turbulent shear stresses is of major which we have interfa ed specially designed strain importance in artificial heart valve evaluation. gauge sensors for hand grip, pinch grip, twist Bi- and unidirectional fluid velocity measurements grip, wrench grip and push. enable calculation of Reynold's Shear Stress During the Motor skill test, the patient is in- (r = -pLiTin. and Normal Stress (a= pur7). Tis structed to adjust the grip force, displayed as important due to the relation to hemolysis and the hight of a moving bar on the PC screen, to thrombus formation, but a is the only obtainable keep the force within 45% to 55% of maximal parameter in vivo. Therefore, determination ofa power. Failures during 30 sek of test are regi- correlation factor between r anda is pertinent. strated. In the Endurance test, the patient is in- In a pulsatile flow model, Laser Doppler (LDA) structed to grip with maximal power and keep this and Hot-Film (HFA) anemometry were used for si- grip as long time as possible. Maximal force, multaneous bi- and unidirectional fluid velocity fatigue, area and raise (force/sek) are calculated measurements downstream of a Metronic Hall and a and displayed together with the force curve on the Hancock Porcine aortic valve. Velocitieswere re- computer screen. All results are documented on a gistered in two flow field locations and at four plotter. cardiac outputs. The velocity signalswere sub- The preliminary results shows that this method jected to analog signal processing prior to digital givs high reproducibility and is a first step toturbulence analysis, as a basis for calculation of wards a standardized method for evaluation of Tand a. grip strength in injured hands. A correlation factor of OA& witha correlation coefficient of 0.98 was found. ImplyingTx: 0.48p.7 Estimation of turbulent Shear Stresses in theregion of aortic valves, based on the axial velocity component alone, seems possible.

1988 World Congress on Medical Physics andlian7dical EngineeringSan Antonio, Texas 277 Biomedical Engineering ScientificPapers

BE12-F.2 HE12-F.3 Some Mechanical Methods for Design Three-dimensional flow mapping down and Improvement of CardiacValve Prostheses,Diao Yingmin:Qian Zhong- stream mechanical mitral and aortic valveprosthesesusingultrasonic fan,Xia Zhigao, and Shi Defang, velocimetry, F. Cassot; D. Morvan, Tongji University, Shanghai, China. G. Tonietto, R. Pelissier,C.N.R.S.-L.M.A and I.M.F.M., Marseille, France. This paper proposed some mechanicalcri- teria for design and improvementof cardiac valve prostheses; reasonable Athoroughhemodynamiccharacterizationof stress artificial heart valves involves a detailed study of distribution both in the stent andon the the leaflets,because the stress concentration tridimensional and unsteadypostvalvular may cause deformation and rupture of the velocity field. Cardiovascular simulatorsnow exist stent, and it can also make the leaflets which enable to duplicate in vitro the physiolo- calcified and torn. gical conditions oftransvalvular flow (LV and For this reasonwe applied appropriate aorticshapes, elasticity,pressure and flow waveforms). Instantaneous velocity mechanical methods to a unileafletvalve maps were set as an example.The photo-elasticity stress behind disk and ball valves in mitral (anatomically analysis is used in the stent to find the shaped LV) and aortic position, from ultrasonic location of stress concentration. Byim- Doppler measurements. Among the considerable proving the design the maximum stressde- amount of information gathered, the more striking features concern, as : creased to1/2-143 of the originaldesign. yet the LV vortex flow For the leaflets, in view of thedefinite pattern formation (related to the rolling up of the relationship between the stress distribu- shear layers at the value level) in mitral position; tion and design parameters(D,H,a,b,c),cal- damping of the post-valvular disturbances (aortic wall culating the stress distribution byfinite elasticityeffect), skewnessofvelocity element analysis, we determined that the profiles in the near wake of disk and ball valves, object of optimization should be min(max inaorticposition;reverseflowandvortex 01(H,a,b,c))(D was given). Soagroup of structures much more important fordisk valves optimized parameters was obtained.These than for the ball ones are strongly dependentupon methods can accelerate the ripe ofa new disk orientation. These data may alsoserve as a type of valves. background knowledgeintheinterpretation of blood flow velocity observations (comparisons with color Doppler will be presented).

BE12-F.4 BE12-F.5 High Strain RateViscoelastic Behavior of Bovine Pericardial Xenograft Materials:Effcr:of ScanningElectronMicroscopic (SEM)Eacaminations of a AlternativeCrosslinkingMethods, CeramicProstheticHeart Valve for Use as a Sean A. Haberer*, J. Michael Lee, and Cardiac Prosthesis Christopher A. Pereira, Centre for Biomaterials, Y.Mitamura*, K.Otaki,K.Hosooka, T. Yuta,and University of Toronto, Toronto CANADA T.Mikami,HokkaidoUniv., Sapporo 060, Japan Clinical failure of bioprosthetic heartvalves Fractures and thrombus generation has been linked to fatigue of leaflet materials of in regions of high stress. prosthetic valves have led toa development of a While the high more durable and thrombo-resistant heart valve strain rate mechanical properties of the leaflet comprising a single crystal alumina material are required for stress analysisand disk and valve design,these have not been measured. titaniumnitride valve ring. Blood compatibility Bovinepericardialxenograftsamples were was examined by SEM examinations of thevalves prepared by: (1) chemical treatment in implanted in sheep for 35(J/1), 26(1/2),20(1/3), glutaraldehyde (GA) orcarbodiimide (CD); and 23(N4) days. The ceramic valves were (2) physical treatment: freeze-dryingor incorporated into apneumaticassistpump.The heat/vacuum drying; or (3)a combination of the pump wasparacorporeallyimplantedbetween the above. The resulting materials were tested using left atrium and the descending aortain asheep. a computer-interfaced tensile testing apparatus, There was little fibrin and few piggy-backed onto an Instron servo-hydraulic platelet colonies on the alumina disks (01, 2, 3, 4), testing machine. Measurements included cyclic except for a small amount of loading, stress deposition of relaxation, and forced fibrin and vibration. Loads of 20,40, plateletson the outflow alumina disk and 80 g were (111). applied under loading frequencies of 0.01, 0.1, Therewas no gross platelet deposition on the TiN rings, 1, and 10 Hz. The stress-strain curves forfresh exceptfor some deposition of fibrin and platelets on limited regions pericardium (10% strain at 80 g)and GA or CD (111, 2, 3, crosslinked materials (20% strain at 80 g)were 4) both nearly independent of strain rate; however, The adhered platelet numberwas few and the in either case, the hysteresis loop at 10Hz was shape change was mild on both thesingle crystal markedly enlarged. Stress relaxation depended alumina disk and the TIN valvecage. The ceramic strongly on loading rate - reaching 40% stress valve is a promising heart valve forcardiac loss at 100 s (10 Hz). Previous low strain rate prostheses. studies have underestimated the viscoelastic response of these materials. 298 278 1988 World Congress on Medical Physics andBiomedical EngineeringSan Antonio, Texas Biomedical Engineering Scientific Papers

BE12-F. 6 BE12-F.7 The Role of Chordal Geometry in the Development of HYDRODYNAMIC PERFORMANCE OF BEIJING BOVINE PERICAR- Systolic Anterior Mittel. Moticc DIAL VALVE (BN2) AND STEADY FLOW TESTER USED IN E.G. Cape*, E. Gieseldng, D. Simons, A. Jimoh, A.E. CLINICAL Liu Yinglong*, Zhu Xiaodong and Xi Baoshu, Weyman, R.A. Levin'., and A.P. Yonanathan, Georgia Cardiovascular Inst and Fuwai Hosp, Chinese Academy Institute of Tecimolegy, Atlanta, GA 30332-0100 of Medical Science, Beijing, China In patients with hypertrophic cardiomyopathy, the Beijing bovine pericardial valve (BN2) was compared mitral valve moves anteriorly, assuming a unique shape in vitro with CE, HK, LK, BS s 60, HC and GZ. The withmitral- septal contact centrally and preserved open pressure drop of BN2 is lower than that of por- orifice area laterally. The purpose of this in vitro study cine valve, the AP is lowest among the valves tested was to test the hypotheef that anterior motion of a and the EOA of BN2 23, 2.7 CM2 is the largest at 12 membrane in a flow field can be generated by altering L/M in steady flow, also at 2-6 L/M in pulsatile the distribution or effectiveness of chordal tension flow. The CD of BN2 is about 70-85% and PI 50-65%. tethering the membrane.Accordingly, a horizontal The backflow are 7.7% in LK, 5.6% in BN2, 4.4% in GZ, leaflet mounted in a flow chamber was attached by 3.3% in CE, 2.9% in HC. After replacement, the mean cords at its distal end to a series of upstream screws. mitral EOA of 14 pts increased from 0.97 to 3.45 2 Chordal tension was varied by turning the screws or CM , MvR from 139 to 243 ML/S. CO doubled while shifting them anteriorly within the model. Leaflet mean LAP decreased from 21.1 to 14 mmHg and AP cross motion was recorded photographically and mitral valve from 14.7 to 3.5 mmHg. All of the CO, echocardicgraphically. CI, SI, PA, PCWP, SR, PR, PAR, LVSWI and rvswi re- Anterior motion with the same unusual configuration k..overed to normal levels 91 hrs after operation. seen in patients was reproduced by decreasing central The steady flow tester used is small, nonpoisonous, chordal restraint while tension on the leaflet edges was easy cleaning and sterilis.tion. Vanes turning is maintained.Directing chordal tension anteriorly caused driven by electrical machine pump with which the greater degrees of anterior motion at earlier stages in flow rate can be over 30 1. /H. Our experiment also the release of chordal restraint; increased flow rate had showed a good correlation between ROA and EOA in 134 a similar but less marked effect. These studies suggest samples and fairly good correlation of EOA between that primary geometry alterations in the papillary in vitro and in vivo. muscle- mitral valve apparatus are capable of playing an importantroleindeterminingthepresenceand geometry of systolic anterior mitralleaflet motion.

BE12-F.8 BE12-F.9 Amorphous Coating Materials for Heart Valves, Pulsatile Turbulent Formation in the Human M. Schaldach, Zentralinstitut A. Bolz* and Aortic Arch in Responseto Valve Orientation fur Biomedizinische Technik, Universitat of ProstheticHeart Valves - In-vitro Model Erlangen, D-8520 Erlangen, FR of Germany Assessment -, H.Minamitani*, E.Okada, T.Niyazawa andA.Shinoda, Hiyoshi3-14-1, Yokohama, Japan Pyrolytic carbon coating represents today's state of the art in the manufacture of cardiovascular Pulsatile turbulent formation in the human aortic implants such as artificial heart valves. How- arch was investigated in response to valve orienta- ever, comparative studies with other materials tion of prosthetic heart valves by using laser have demonstrated the possibility of improving Doppler visualization technique and pulsatile mock haemocompatibility. Based on an electro,:hemical system. The used system mimics humanaorta including model considering the activation of fibrinogen as three major branches and left heart. Turbulent inte- an redox process in the solid-blood interface, nsity, turbulent enrgy density, per spectra of new materials such as semiconducting Ta doped the turbulent flow components as well as shear rutile ceramics have been experimentally evaluat- stress were measured for three different valves. edAn electron concentration of approximately Turbulent formation was greatly dependent on the 1020 with a band gap of3.2eV have met an- vessel geometry and the valve orientation, advanced tithrombogenic requirements. Technological prob- as far as top of the aortic arch and affected the lems related to the ceramic structure resulted in flow in the major branches such as brachiocephalic, the evaluation of coating processes such as reac- and common carotid arteries. Total hydrodynamic tive sputtering and chemical vapor deposition assessment regarding the turbulence and systolic (CV')) of TaN, SiN and SiC. These materials are energy dissipation suggested that major orifice of deposited as amorphous doped semiconducting lay- Bjork- Shiley disc valve should be oriented to the ers using processes proven in the manufacture of aortic outer curvature, and that commissure of SIM amorphous photovoltaic cells. Preliminary results valve setting in parallel to the curvature gave the have demonstrated the feasibility of producing a best performance. As for Bioprosthetic valve, plac- variety of electronic structures to meet blood ing one of the three leaflets on the inner curvature compatibility requirements. The technology of site was found advisable. These orientations plasma activated chemical vapor deposition offers provided less obstructive effect on the outflow advantageswhich matchelectrochemical, solid tract, smoother flow formation, and less turbulent state and tribological requirements in the de- energy density in the aortic arch. sign of artificial heart valves. 2 9

1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas 279 Biomedical Engineering Scientific Papers

BE12-F.10 BE12-F.11 A NEW MOCK CIRCULATION FOR ARTIFICIAL HEART VALVES MECHANICAL BEHAVIOUR OF THE OMNICARBON HEART K. Affeld*, K. Schichl, A. Mohnhaupt, A. Spiegelberg VALVE PROSTHESIS Josef Kohler; Peter Guntermann, Paul Universitatsklinikum Rudolf Virchow, Berlin, F.R.G. Holtmann, Aerody amisches Institut, Aachen,West In a mock circulation designed for the testir ,of Germany artificial heart valves one has to imitate tne flow The study is focussed on the development ofstandar- through the valve and the aorta. The first task car, dized test methods for artificialheart valves. be achieved without great difficulty, while these- Three series of seven original packed CC-valvesof cond is quite complicated. The generation of the size 19 up to 31 could be tested(21 valves). Mech- pressure pulse in the real aorta is very complex. anical and fluidmechanical testparameters: Geometry, As a result mock circulations vary from labto lab deformability, pressure- and volume-loss. and consequently the test results. In the new mock Reeults/Geometrical parameters: The smaller valves circulation the artificial aorta is ..voided altogeth- show a bigger textile ring diameter and thebigger er. Itis instead simulated by the movement ofa ones a .,aller one than indicated by the manufactur- piston, which combines two elements in one- the er. The differences are mostly smallerthan 1 mm. heart wall and the aorta. During systole, the piston The valves of size 27 and 29 show thesame disc dia- moves according to a physiological flow curve and meter. Deformability and disc gap: The deformation forces the fluid through the valve. Then it stops results due to a radial force actingon the valve and begins to retract. When the valve is closed, a rim show a good compromise for the widthof the differential pressure across the valve is generated, bearing gap with relatively smallvariations. The which causes the piston to yo to a different control sensitivity of the valve against these forces(stick- mode -it now is retracted with a certain elasticity ing) is a lot smaller than that of theSt. Jude following the diastolic pressure curve. The imita- valve. Pressure losses and effective valve area: tion of the aortic elasticity is thus transferred The off. valve area A, is approximatelyindependent from hardware to software which means that the model of the volume flow. The relatiee ostium crossection aorta now is well defined. The movement of the pis- Ae/AT (AT c ostium crossection)amounts 35 to 44%. ton also serves as a very precise flowmeter. It increases slightly with increasing valvesize. In the new mock circulation valves can be tested with Volume loss: The total volume lossper beat VR 0 VC more precision and greater speed. Unlike the tradi- + VL (closing volume + leakage volume) increases tional mock circulations it is a system open at one with valve size. It has - as average for all pump- end so a valve can be changed with a twist of the ing conditions and sizes - with 3.4 cm3 about that hand. It therefore may be useful for the production value, which was obtained for othertechnical control of valves. valves, except bileaflet valves.

BE12-F.12 BE12-G.1 A mathematical model for the study of THE PLATELET THROMBUS FRMATION IN A MOVING SEREAM flow through stenosed valve FOR EVALUATING HEMOCOMPATIBILITY OF BIOMATERIALS Chang-song Gene, Wen-Yu Kong, Tianjin University Xi Tingfei*, Tian Wenhua, Wang Chunren, Lei Xuehui NICPBP, Temple of Heaven, Beijing, China Stenosed valve or artificial valve produces very different flow behaviour from normal in We study the process of thrombus in normal fresh both the valve and the ascending area. Some rabbit platelet rich plasma (PRP) on revolving of characteristic features are jet flow, polymeric loop and discover that this process is stagnation zone, large pressure drop across divided in five steps, first particles, snow storm, the valve. Thesecharacters havemuch to do aggregation, plug, thrombus. The time formed snow with clinical aetected problems, suchas thrombus storm is called Characteristic Thrombus Formation formation, endothelial damage, tissue over-growth. Time (CTFT) and is a parameter of evaluating hemo- Stenosis(natural valve with stenosed disease compatibility of polymers. We study 8 different or bioprosthetic) and clg(mechanical prosthetic) time of PVC exposured PRP (0, 10, 20, 30, 40, 60, is the main cause of producing these flow 80, 12n min) in order to ensure that testing are charateristic features. For this reason, we actually measuring activation due to the test ma- usedefiniteelement numerical method to study terials, results of CTFT (sec) are 0', 1441+109); the flow through a stenosed valve anda mechanical 10', 194( +50); 20', 204(462); 30', 192(4.22); 40', prosthetic. Unsteadyflow modelexperiments are 97(+22); 60', 79(+30); 80', 129( +58); 120', 109 conducted for three types valve(one bioprosthetic (+47). The best activable contact time is 60 min and two mechniacal prosthetic)to measurepressure The CTFT cf five polymers are measured with fresh drop under varing flow. rabbit PRP at 17 rpm for 60 mit.. Results of CTFT Theresultcan help usto understand why (me, s.d., n -5) are PV0,79(+30); Silicone (Dow and where problems such asthrombus information Corning co. USA), 205(431); Biomer, 212(442); Pell- occur. So we can improve the valve design. This ethane 2363- 80 A, 231(+31); Cardiot.hane, 277(+83). study is also lay a foundation for computer Therefore, t%e blood materials interactions and aided valve design. mechanism of thrombus formation will investigated with the method developed in this paper.

280 1988 World Congress on Medical Physics and BiomedicalEngineeringSan Antonio, Texas Biomedical Engineering Scientific Papers

BE12-G.2 BE12-G.3 AN INSTRUMENT TO ASSESS VASCULAR 1HEORY OF PULSE PHOTOMETRY OF HEMOGLOBIN CONDUCTANCE IN VIVO. C.P. Burns, AND DYE IN BLOOD S.A. O'Connor; S. Frampton and L.M. Wood, Smith Kline & French T. AOYAGI: M. FUSE, M. KANEMOTO, Y. SUGIYAMA, Research Ltd, The Frythe, Welwyn K. MIYASAKA*, M. SAITO** Hertfordshire, AL6 9AR, U.K. NIHON KOHDEN CORP, *NATIONAL CHILDREN'S HOSPITAL, ** TOKYO UNIVERSITY; TOKYO, JAPAN. Many disease states of the cardiovascular system can be attributed to the inability of blood to per- The pulse oximeter measures the photo- fuse regions of the circulation, e.g. the heart, plethysmogram in two optical wavelengths and exercising skeletal muscle or mesentry. By simul- calculates 35(ratio of the change rates of taneous measurement of blood flow to the region, absorbances) to find the arterial oxygen satu- and arterial blood pressure, the effectiveness of a ration. This principle can be applied to drug to increase the vascular conductance of the measure the concentrations of other dyes in region can be measured. plasma. We have designed, developed and tested, in vivo in As there are red blood cells in blood, the a variety of animals, an instrument which continu- reflection on the surface of the cells and the ously and accurately divides flow and pressure sig- flattening effect caused by limited distribu- nals to give a signal proportional to conductance. tion of hemoglobin and dye makes the relation Pulsatile input signals of flow, from an electro- between i6 and absorption coefficients or dye magnetic blood flow meter, and pressure, from a densities different from that of usual solu- blood pressure transducer, are smoothed by a low tions. pass filter prior to passing into a variable gain We found theoretically and experimentally stage which allows optimal signal inputs to be set that the3glof blood is approximately equal to for the divider stage. The divider used is an iq of solution even in blood with dye in plasma, Analog Devices AD535 which accepts signals inputs when optical scattering is high enough and in the range of 0.2-10V. The instrument can accom- absorption by the cell is low enough. This modate flow and pressure signals from a variety of will give the base for non-invasive photo- species. The instrument is cheap, reliable and metry of blood contents for many purposes. easy to use. Vascular resistance may also be simply obtained.

BE12-G.4 BE12-G.5 On-line Measurement of Red Cell Velocity Power SpectralAnalysisof in Microvessels With an Auto-tracking Vasoactivity in Cutaneous Window on Television, Wang ZhongRui*, Microcirculation Dept. of Scientific Instrumentation, following Zhejiang University,Hangzhou,China. Thermal injury, R. da Costa*, S.J. Aggarwal, and K.R. In order to measure the red cell velocity(Vrbc) Diller, University of Texas, Austin ,Texas 78712 . in microvessels of vital subjects, an auto- tracking window on TV screen has been designed to track the crooked microvesseB wobbling The responseof microvasculature to caused by heart beating and respiration. The thermal trauma has been studied in dorsal Vrbc is measured on-line by sampling optical skin-flap chambers of hamsters. density line by line from a microvessel image Following recovery and normalization of in the window and calculating AMDF(Average tissue post surgery, fluctuations in the Magnitude Difference Function) of two optical diameter of selezted arteriole-venule density waveforms sampled from successive pairs ranging in size from 10 gm to 50gm frames in a microcomputer. Pulsed Vrbc in and at different Lrznching order sites mensentery capillaries of frogs have been were measured several tines prior to burn recorded. The range of Vrbc and calculating and at the same sites post mild contact rate available in our system are +4.5mm/s and burn. The experimental vasomotion data 7-9 times/s respectively. A formula is derived was characterized in termsof mean, to modify the calculated velocity of vertical standard deviation and bytheProny movement on TV screen. method of spectral power analysis. The Al (37-50gm), A2(30-50gm) and A3 (25- 41µm) arterioles dilated an average of 58%, 20% and13% respectively. The fundamental frequency components of most vessels were shifted toward lower values, with some of the higher frequency components disappearing. Supported by a grant from the Whitaker Foundation.

4:k

1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas 281 Biomedical Engineering Scientific Papers

BE12-G.6 BE12-G.7 Measurement of Platelet and LeukocyteDistribution Measurement of Non-symmetric at Branches, J. Perkki6*, L.J. Wurzinger Velocity and Profiles in Microscopic BloodVessels: Formation H. Schinid-SchOnbein, Dept. of MedicalPhysics, of Compaction Stasis, W. University of Helsinki, SF -00170 Mel, J. Perkkib" and Helsinki,Finland II. Schmid-S.:hiinbein, Abt.Physiologie der RWTH Aachen, D-5100 Aachen,FRC As blood flows through inicrovessels,erythrocytes and plasma are partly separated from eachother. This phase Blood samples with strongly separation gives rise to non-uniform aggregating erythrocytes ha- distribution of cells ve a reduced suspension stability andto sediment unin blood vessels. We have studiedthe concentration of der the influence of gravity.We report about the phe- platelet and leukocytes in blood flowthrough T-junctions nomenon of intravascular erythrocyte cast in polyester blocks. The main aggregate sedimen- branch diameter was tation of human blood samplesduring maintained flow 200 or 100 pm and the side branchdiameter 100 or 80 pm, perfusing micro-channelscast in polyester blocks. Blood respectively. Human blood sampleswith different 1.-vels of flow was examined froma lateral position recording the fibrinogen were used. Human fibrinogen (Sigma, U.S.A.) flow on video tape. Erythrocyteswere suspended in iso- was used for preparation of fibrinogen solutions.Either tonic buffered saline containingbovine fibrinogen (7 g/1), fibrinogen solutionor isotonic buffered saline were added albumin (1 g/l) and heparin(2000 1E/1). The velocity to platelet rich plasma. Concentratederythrocytes were profiles were measured byusing a dual slit system adapted mixed with these solutions to yield four different samples: to the video record. Non-symmetricvelocity profiles and hematocrit 0.35 or 0.55 and highor low fibrinogen con- pronounced phase separation centration. between erythrocytes and perfusion media were observedat perfusion speeds of 0.5 The results showed that the concentrationof platelets in minis, the place of maximumvelocity moving progres- the blood from the side branchwas about 1.2 - 1.5 times sively out of the center towardsthe top of the channel. higher than in the reservoir. Theconcentration of leuko- Along the length of the channelthe speed of the lower cytes in the blood from the side branch was about 0.9 erythrocyte layers progressivelyslows down and thenon- 1.4 times the value of the bloodfrom the reservoir. The symmetry grows. The resultssuggest that the enhanced ratios were found to dependon the feed hematocrit and erythrocyte aggregatabilitymay affect the stasis forma- fibrinogen concentration. tion via gravitationally inducedphase separation leading to non-symmetric velocity profiles.

D. 2-G.8 BE12-G.9 Fundamental Light-ScatteringCharacteristics of Red Blood Cells, A. P. Shepherd* andJ. M. Photon Diffusion Model of the OpticalAbsorbance of Whole Blood, J. M. Steinke* and A. Steinke, Department of Physiology,University of P. Shepherd, Department of Physiology, Texas Health Science Center, San Antonio,Texas, University of 78284, U.S.A. Texas Health Science Center, San Antonio,Texas, 78284, U.S.A. Two important optical properties ofred blood cells (RBCs), Photon diffusion theory has had limited their microscopic scattering cross-sections success in modelling (as) and the mean the optical transmittance of whole blood.Therefore, we have cosine of their scattering angles (g),contribute to the optical developed a new photon diffusion behavior of whole blood. Therefore, model of the optical the adequacy of Mie absorbance of blood. The model, whichwas derived from theory in predicting values ofas and p was tested by the diffusion equation underappropriate boundary experiment. In addition, the effect of redblood cell size on conditions, has benefited fromexperiments to test its as and p. was investigated in two fundamental assumptions. Predictions ofthe model have ways: 1) by studying been compared extensively with transmittance eryth -^ytes from the dog, data from goat and human, three species whole blood, and the model is consistentwith experimental known to have different RBC sizes and2) by allowing the findings. Furthermore, when all RBCb 'root each species parameters associated with to shrink or swell osmotically. a given optical geometry are known, the modelneeds no Values of as obtained by measuringthe collimated variational parameters to predict the absolutetransmittance of whole blood. transmittance of dilute RBC suspensionsilluminated with a However, even if the exact value of the He-Ne laser agreed with thosepredicted by Mie theory. incident light intensity is unknown (whichis the case in many situations), only a single additive Moreover, measured as valueswere directly proportional to constant is required to scale experiment to theory. Finally, the modelis shown to be RBC volume. By contrast, values ofp from Mie theory useful for simulating scattering effects and were consistently greater than those obtained experimentally for delineating the relative contributions of the diffusetransmittance and the by making angular scatteringmeasurements in a goniometer. collimated transmittance to the total opticaldensity of whole Thus, Mie theory appears to yieldadequate values for the blood.Applications of the model include RBC's microscopic scattering cross-section,but by treating oximetry and the RBC as a sphere with an equal volume, measurements of the arteriovenous oxygen differencein Mie theory fails whole, undiluted blood. (Supported byHL 36080) to take the RBC's anisotropy into account andthus yields spuriously high values for (Supported by HL 36080)

282 1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas Biomedical Engineering Scientific Papers

BE12-6.10 BE12-G. 11 Erythrocyte Deformability, Total White Blood Cell Court, EFFECTS OF HEMATOCRIT ON RHEO - OPTICAL Platelet Count, and Transcutaneous Tissue Oxygenation as ERYTHROCYTE AGGREGATION KINETICS, Predictors of Survival in a Canine Septic Shock Model Abel Gaspar-Rosas* and George B. Thurston, Bio-Medical TW Chun2,t, MA Smitht.U, TL Whitsettt, EA O'Reart, LB Engineering Program, Department of Mechanical Engineering, Hinshaw°. tDept.ofChemicalEngineering,Univ.of The University of Texas, Austin, TX 78712 U. S. A. Oklahoma, Norman, OK, °Depts. of Physiology and Biophy- sics, *Dept. of Cardiovascular Diseases, Univ. of Oklahoma Accurate characterization of the aggregation properties of Health Sciences Center, P.O. Box 26901, Oklahoma City, OK erythrocytes is of great importance in the clinical evaluation of 73190, USA. blood. While erythrocyte sedimentation rate is strongly dependent on these properties, more precise quantitative The effect of shock developing subsequent to gram-negative measurement is provided by analysis of the rheo-optical septicemia on the milieu of hematologic and rheologic factors transient following cessation of a flow induced state of cell which govern microvascular tissue perfusion may provide util- disaggregation. Conventional clinical rheo-optical methods ity in predicting survival in a canine Fhock model. Hemorheo- give a single characteristic aggregation time, while more precise logic parameters measured at 6 hours after the infusion of live analyisis shows the need for two or more aggregation times to Escherichia coli which exhibited predictive value of survivabil- fit the exponential optical transients. ity of greater than 24 hours included a 1) less decreased defor- The aggregation kinetics are strongly dependent on the mability of canine erythrocytes, 2) decreased total peripheral hematocrit and this dependence must be accounted for before white blot,d cell count, 3) decreased peripheral platelet count, clinical observations can be standardized. The effect of and 4) increased transcutaneous tissue oxygenation. variations in the hematocrit on aggregation times and weighting Differences in mean systemic arterial pressure, heart rate, core factors is presented for both reflected and transmitted light, temperature,respirations,hematocrit,serumglucose, pH, using a physiological range of hematrocrit values of .30 to .60. arterial oxygenation, erythrocyte sedimentation rate, fibrinogen The kinetics of erythrocyte aggregation in a thin layer of concentration,and plasmaviscosity were notdiscernible blood (approx. 0.50 mm) during the first 40 seconds of the between both survivors of greater than 24 hours and non- transient is analyzed using two aggregation times. The survivors.Pretreatment of dogswithpentoxifylline did transients of both the light transmitted through the layer and improve erythrocyte deformability but did not improve survival reflected from the surface of the layer are monitored at an LD1c3 challenge.Greater erythrocyte defonvibility, leu- immediately following cessation of an oscillatory flow that kopenia, and thrombocytopenia are associated with greater tran- produces the initial high degree of disaggregation of scutantous tissue oxygenation and improved survival likely erythrocytes. As hematocrit is increased the first (shorter) indicative of a less altered state of systemic microvasculature aggregation time decreases more rapidly than the second perfusion. (longer) time, with the second time approaching a constant value as the hematocrit exceeds 0.40.

BE12-6.12 BE12-O.13 The Determination of Hematocrit(Hct) and ofTotal TUMOR CELL DAMAGE BY SHEAR STRESS Proteir. Concentration (TPC) in Flowing Blood from the Wang Junjian*, Liu Yongge, He Xi, Bioeng. Dept., Measurement of the Acoustic Wave Velocity NJ, Huazhong Univ of Science 6 Technology, Wuhan, Hubei, P.R. China D. Schneditz, M. Moserand T. Kenner, Department of Physiology, University of Graz, A-8010 Graz, Austria Experiments testing the ability of tumor cells to withstand shear stresses in Couette rheometer are The dependence of v on density p and on adiabatic carried out.Above a threshold stress level, EAT In blood, (Ehrlich asci tes tumor) cells are extensively dam- compressibility xad is given by v = pxad 0.5 aged due to shear stresses.We find that tumor both p and )(ad mainly depend on TPC. v is determined cells are killed 30% to 60% of the population at by the measurement of the acoustic wave travel time shear stresses of 25 dynes/cm for one to two hours, (the resolution isbetter than 50 ppm). xad, which is which shows fluid shear stress in the physiological calculated from p and v values, is determined at differ- range can be the major source of damage to tumor cells in the circulation.At stress below 10 dynes/ ent TPC, at different Hct and at different temperatures. cm, there is little direct damage because of shear For porcine blood the coefficient of correlation is > 0.99 stress.Based upon the assumption that all viable the range and the relation between xad' V'TPC (in cells have the same specific death rate in shear from 90 to 340 g/1) and Hct (in vol%) at 20.0°C is: field,it is easy to induce the function between viability of tumor cells and shearing time.Viabil- ity is decreased exponentially with the duration of xad= 4'586'10-1°- 3,226.10 -13 TPC C Pa 1] shearing time, which would be described as V=eKT v = 1483,07 + 0,273 TPC + 0,000346TPC2 (m/s] (V stands for Viability,t means time and k is char- xad=4'218'10-10-6,994.10-13 Hct EPa acteristic of shear stresses and properties of tumor v = 1517,31 + 0,8996 Hct (m/s] cells).Also, the threshold stress can be determined by extrapolation.All these derivations are consistent with the results of our experiments. The relation between v, Hct and TPC may therefore be used to continuously monitor blood cell concentrations in circulatory systems. 3 o 3

1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas 283 Biomedical Engineering Scientific Papers

BE12-G.14 BE12-G.15 Effects of Hematocrit on Visco-Elasti- Pulsating flowsofinelastic fluids city of Human Blood, H.Q.Chen, G.H.Zhong, in slowly varyingporous tubes L.Li, Z.N.Xue, West China University of - Modelisation of the blood rdcrocirculation Medical Sciences,Chengdu, Sichuan.China - G. Zeggwagh', D. Pellet,Groupe de Rheologie I.M.F T Avenue du Professeur Camille Human blood exhibits .he visco-elastic behavior Soule, 31400 Toulouse at France low shear rate. The viscoelasticity of humanblood can measured with sinusoidal oscillatory shear fl- The main purpose of thispaper is to present a ow. Using the Low Shear-30 Sinus Rheometer,we stu- theoretical and experimental studyconcerning the died the effects of hematocrit(HCT)on apparent v- pulsating flow of inelastic iscosity of blood at lov(0.512sec-1),moderate(5.96 fluids through porous conical ducts. The modelcorresponds to blood flow sec-1) and high(51.2sec-I) shearrate, and on vis- through porous conical vessels. coelasticity of blood. The results demonstrated t- The theoretical approach isbased hat with increasing of HCT,the apparent on animplicit viscositi- difference method, withCrank-Nichdson scheme. The es ar three different shear rates increased expon- experimental method basedon laser-doppler enema entially. The fitted curve between HCT and the - ap- metry is applied to pseudo -fdasticsolutions. parent viscosity at low shear rate was steeper th- Same as many authors,we consider fluid flow in an the others. We found that viscous component(') tubes, representing blood and elastic flow as fluids of one ho- conponentOnalso increased exponent- mogeneous phase in porous tubes.We assume ially when HCT was increasing and HCT affected that the fluid is described by eithernewtonian or power law qi" values more obviously than the arvalues. This model. study developed the empirical equationsto express The influence of geometrical,hydrodynamical and the relationship between HCT and gi',9r and discus- structural parameters issystematically examined sed the physical meaning of the elasticcomponent. and related to velocity profiles, We suggested that both II' and flow rate filtra- the aggregationlaid- tion and hydrostaticpressure. ex(AI) of RBC are objectivequantitative indices of aggregation of RBC,but theory'value is more accurate than the other one.

BE12-G.16 BE12-G.17 Study of red cell deformabilitywith electrical conductivity method, Viscoelasticities of NormalHuman Blood in Rotational Rheometer, Wen Zongyao*, Wang Hongru, F.Liao*,W.Li,Z. Chen Sheng Ding and X.Zhou, China Academy Zhao Yuheng and Shen Qiang, of Tradi- Beijing tional Chinese Medicine, Medical University, Beijing Beijing100700, China. China

A new type of instrument hasbeen designed by the authors. Though the principleof this instrument To imitate blood flow in is based on electrical arteries,viscoelastici- conductivity method, the ties of normal human bloodwere investigated by instrument differs from traditional resistive an oscillation at a frequency of1.3 Hz and amp- pulse spectroscope which can also measure red litude of 40 superimposedto a constant shear cell deformability by measuring its electrical rate of 201 s-1 in HAAKERotovisco RV100/CV100. impedance change. The blood sample is put into For comparison, viscoelasticitiesof the same the gap betwen a stationary inner cylinder anda samples were determined bypure oscillation as rotating concentric outer cylinder, therefore the mentioned above. The resultsshowed that the red cells are subjected todifferent shear stress. complex module G*,storage module The change of electrical G',loss module conductivity which G" and complex viscosity at at rotation were all results from red cell deformation can be measured significantly higher than thatat pure oscilla- through two poles on inner cylinder, namely, red tion (P < 0.001).The otherparameters at rota- cell deformabilitycan be measured. tion,loss factor tg 6 andrelaxation time X,were This instrument may measure static and dynamic significantly smaller than thoseat pure oscilla- results for red cell deformation with advantages tion (P < 0.001).The analysisshowed that the of satisfactory reproducibility, time-saving average elastic increment was 35.1compared with performance and economic in cost. All results can the viscous increment of 23.5when flow condition be given by a microcomputer. shifted from pure oscillationto rotation plus A series of red cell suspensionwere tested, The oscillation. obtained results were analysedand demonstrated that the instrumentcan indeed measure red cell deformability, Furthermore,a formula of exponential relation between electricalconductivity and red cell deformability has beenfound.

'.1

284 1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas Biomedical Engineering Sdentific Papers

BE12-G.3.8 8E1278.19 Permeability of Cultured Endothelial CeZZ Angular Distributionsof Forward Mono layere: Mediating Role of cAMP, S. A. and Backward Light Scattering from Casnocha*, L.V.McIntire, andS. G. Flowing Blood,Munehiro Niwa*, Eskin ,Biomedical Engineering Laboratory, Hiroshi Kanai,Jikei Univ.Sch.of Med. Rice University, Houston, Texas, 77251 8-3-1 Kokuryo,Chofu,Tokyo,Japan. and Department of Surgery, Baylor College of Medicine, Houston, Texas 77030, USA. The scattering and absorption characteristics of bloodis very important for the Endothelial cells (EC), which line blood vessels applicationof light to medicaland bio- in a monolayer, function as a permeability bar- logical engineering. rier between blood and the vessel wall. Monolay- Light irradiate the blood is chiefly ers ofhuman umbilical vein EC cultured on a scattered by the red blood cells. When the permeable polycarbonate substrate were used to blood flow, the intensity of scattered light study the effect of intracellular cAMP on per- from the blood changes from that of blood meability. Agonistsstudied wereiloprost, at rest. It is related to scattering angle, thrombin and bradykinin. Iloprost is a stable shear rate of the flow and other blood analog of prostacyclin, whichstimulates EC parameters. In the rednearinfrared region, intracellular adenyl cyclase activity by binding the intensity of backwardscattering in- to anexternalreceptor. Iloprost decreased creases greatly with the increase of shear permeability to 24% ofbaseline. A similar rate, and it becomes about double the decrease in permeability was observed in mono- intensity for blood at rest when the shear layers treatedwith dibutyryl cAMP and IBMX rate is 50 1/sec. Then the angular distri- (which acttogetherto increase intracellular butioncurve changes fromcircular to cAMP), suggesting that iloprost's effect on asymmetrical flattened ellipsoid.The forward permeability is mediateu bycAMP. Thrombin scattering intensity decreaseswith the treatment of monolayers resulted in an 11-fold increase of shear rate,and it also becomes increase in permeability, which was nearly double when the shear rate is 65 1/sec.Then completely inhibited by first treating the cells the angular distribution curvedoes not with dibutyryl cAMP and IBMX. Bradykinin had no change very much. significant effect on EC monolayer permeability.

BE12-11.1 Ambiguity of Determining Elasticity of "13E12-H.2 Blood Vessel According to Transit Time Poststenotic Shear Layer Instabilities, Arterial of Pulse Wave ,Ye GuoFan*, Qin Jie and Wall Vibraisons and Coupled Nonlinear Oscillators Chai Ming, Dept. Electronic Engineering, Fudan University, Shanghai, P.R. China. J Treiber, R I Kitney, Biomedical Systems Group, The means of estimating the elasticity of Dept of Flectrical Engineering, Imperial College, London SW7 2BT, UK a blood vessel segment by measuring the transit time (TT) of the pulse wave is widely adopted in diagnoses of arterio- Detailed analysis of sn vivo flow velocity measurements sclerosis. Based on Moens-Korteweg's downstream of an inducedstenosishasidentifiedthe formula, the faster the pulse wave speed existence of oscillatory flow disturbances, thought to arise C is, the larger the Young's modulus E is. from vortex shedding and arterial wall vibrations. However, the formula is derived from the As the evolution of shear layers is highly susceptible to nonviscous case different from the practical low-amplitude disturbances,wallvibrations cangreatly viscous situation, so the phase speed mea- affect the vortex shedding process. sured Cf is dependent on the Womersley num- Inthisstudythesensitivityof poststenoticflow Forced flow ber . For a definite harmonic the Cf/C- instabilitiesis studiedexperimentally. curves keep closely constant when 2. At measurements with an LDV downstream of modelled the early-schlerosis with no significant stenoses show that complete flow synchronisation may be stenosis and for a vessel like the femoral achieved at very low forcing levels in steady flow, whereas pulsatile flow is considerably leso susceptible. This explains artery, 3 is true, so M-K formula is the separability of the two oscillatory flow components applicable. When stenosis degree of vessel radius S begins to observed in vivo. 40%with E rising, 2 Finally, a mathematical model is described comprising a becomes true, which implies that Cf may be chain of mutually coupled nonlinear oscillators, that explains no difference from the normal. In the case flow synchronisation and the differences in susceptibility for of severe case with S 70% ( 1), Cf can be steady and pulsatile flow.Interestingly, this model also slower than the normal instead. Our exper- describes vortex pairing and exhibits a form of turbulence iment in vivo on a dog's commom iliac artery based on the concept of diffusion-induced chaos. with artificial sclerosis and stenosis verified the above - men -ioned law and proved the ambiguity of TT method. ( Supported by the National Science Fund of PRC)

t )) 1988 World Congress on Medical Physics and Dioniedical EngineeringSan Antonio, Texas 285 Biomedical Engineering Scientific Papers

BE12-H.3 BE12-H.4 Two- and three- dimensionalvisualization of The mathematical distributed parameter velocity profiles measured invivo in the model for arterial system with wall aortic trifyrcation in pigs. viscosity and tethering, H. Hirayama EM Pedersen3,JH Ostergaard& VE Knudsen and H. Yasuda, Hokkaido University, Dept. Thorac. and Cardiov.Surg., Skejby Univ. N 15, W 7, Sapporo, Japan Hosp. and Instit. of Experim. Clin.Research, Arhus University, 8000 ArhusC, Denmark To analyze the selective independent effects of Atherosclerosis is often locatedat arterial bran- the components of the arterial system including ches, probably due to localhemodynamic factors. the surrounding interstitial tissue, we have con- We have studied the velocitydistribution in the structed a mathematical distributed parameter aortic trifurcation in six pigs(90 kg). Point ve- model for a segmental artery such as the femoral locities were succec.sively measured in the distal artery or coronary artery. By inputting the bio- aorta and in the aortic trifurcationin 21 measu- logical data into this model, the flow wave trans- ring points distributed across the vessel area u- mission was reconstructed. The computed values of sing a 1 mm hot-film anemometer probe. Using ECG- the several kinds of the impedances were almost triggered ensemble averaging and computerized plot- ::oinsided with the experimental data. The fluid ting, 2- and 3-dimensional visualization of lami- dynamical viscous resistance, inertial resistance nar velocity profiles were made. and the longitudinal impedance decreased with the The velocity profiles proximally in the iliac arte- radius of the artery. The transverse impedance ry were asymmtric with very low velocities near increased with the elastic modulus of the arteri- the lateral wall includingsigns of retrograde flow al wall E. The characteristic impedance decreased during diastole. Velocityprofiles distally in the with the radius but its frequency dependance was iliac artery were less asymmetric. small. The complex impedance (local impedance) Small variations ( 10%) in heart frequencyduring increased with the E and the distance from the the measurement series producedsignifikant arte- entry zone of the arterial tube x. The reflection facts in the profiles. To avoid this in the future coefficient was also changed with the E andx. it seems pertinent eitherto: 1) keep a constant However the effects of the wall viscosity and the heart rate using cardiac pacing2) to measure ve- tethering of the surrounding tissue was not so locities in different pointssimultaneously. marked comparing with the effects of the radius and wall elasticity.

BEIZ-H.5 BE12-M.6 The pressure-volume diagram of human Numerical Simulation of Coronary Circu- aortas is unaffected by sclerosis. lation During Coronary SinusOcclusion. G. Langewouters*, K. Wesseling and W. W. Schreiner *, F. Neumann, W. Mohl, GoedharC, BMI-TNO, Academic Medical E.Wolner. 2nd Surgical Dept. University Center, 1105 AZ Amsterdam, of Vienna, A-1090 Wien, Austria. The Netherlands. In order to study the effect of coronaryvenous Segments of 45 human thoracic aortas, aged 30-88 interventions on coronary circulation, a model yrs at autopsy, were perfused with 37 C Tyrode ofcoupled differential equations is setup and at in-situ length. Two degrees of sclerosis were solved numerically. Aortic pressure and myo- classified, according tothe WHO definition: 26 cardial contraction, obtained by Fourier synthe- aortas were normal or only intima affected (ath- sis, enter as driving forces. Coronary arteries, erosclerotic), 19 aortas were also media affected capillariesand coronary veins are given re- (arteriosclerotic). sistances and compliances dependent on flow and Pressure-volume (p-V) relationswere computed volume, respectively. Integration yields pres- from measured diameter changes due to20 mmHg sures, flows and volumes within the coronary pressure steps between 20 and 180 mmHg. A circulation as functions of time. three-parameter mathematical model forthe p-V Simulations are carried out for normal perfusion relationship explained over 99% of the variance conditions andintermittent coronary sinus in volume with pressure for each aorta. occlusion. The venous compliance is modelled as Interindividual differences in the model parame- a function of increasing coronaryvenous volume ters could for the greater part(>60%) be ex- during coronary sinus occlusionso as to repro- plained by the factor age alone. Arteriosclerotic duce experimental data obtained from dogsand aortas had a slightly larger volume than atheros- humans. Resultsare obtained forchanges in clerotic aortas, over the entire ,,ressure range .phase ofcoronary flow and the amount ofre- studied. However, at physiological pressures, no distribution of blood as aconsequence of coro- differences betweenthe compliances (dV/dp)of nary sinus occlusion. thetwo groups of aortas could be demonstrated (analysis of covariance, p>0.05).

286 1988 World Congress on Medical Physics and BiomedicalEngineeringSan Antonio, Texas Biomedical Engineering Scientific Papers

BE12-H.7 BE12-H.8 Pressure Variation on the Walls of a Model Viscoelastic Properties of Arteries in Human Arterial Branch During Pulsatile Flow, by Forearms in Vivo, P. Gizdulich, Fisica Medica J. M. Khodadadi, Auburn Univ., Auburn, AL Universita 1-50134 Firenze 36849-5341 and N. S. Vlachos, Univ. of The viscoelastic pattern of the arterial pressure- Illinois, Urbana, IL 61801 volume diagrams was recorded.in a forearm segment of 8 healthy subjects. Arterial pressure was noninvasively measured by using a special waveform Numerical predictions of pulsatile laminar flow Blood pressure was in a plane 90-degree bifurcation have been pressure recording system. recorded simultaneously in the medium f4nger of the carried out. The cases of Reynolds numbers of left and right hand. Arterial volume was measured 38 and 102 with the corresponding Stokes numbers by using impedance plethysmography. Transmural of 4.1 and 4.9 were investigated. The results pressure and arterial blood volume in a segment of show that one separation zone is established on the forearm were controlled by using the unloading the bottom wall of the main duct and another on technique. Shifted loops in repeated subsequent the upstream wall of the branch. The numerical cuff inflation-deflation phases were obtained. In results were also compared to the LDV velocity measurements of the pulsatile flow field, the past, the importance to select very slow cuff deflation rate has been discussed in order to showing excellent agreement in regions outside analyze the whole forearm segment. In that protocol the separated zones a venous increased stasis was induced in one arm. As an immediate consequence, pressure waveform The pressure field was also obtained during the Our aim here has been course of the present study. The detailed distorsion was documented. to evaluate the influence of the increased venous variations of the pressure along the walls of pressure on local arterial blood pressure. A blood the arterial branch model were determined. It pressure variation between the two hands, related is clearly evident that the presence of adverse pressure gradients in the vicinity of the two to cuff pressure trend, is occasionally observed. recirculation zones lead to local flow No trend of the differences of the mean pressure separation. in the 8 subject group is :ocumented. Assumptions on linearity of the hand's impedance arediscussed.

BE12-H.9 BE12-H.10 Numerical Analysis of Fluid-Particle Dynamics An Experimental and Theoretical Simulation in Artery Bifurcations with Stenosis am Study of Blood Flow in Arteries, Y. Haggag*, Surgical Reconstruction, C. Kleinstre..er, 4. King Saud University, Biomedical Technology NazemP"and J. P. Archie, Jr., NC State Department, P.O.Box 10219 Riyadh 11433, University, Raleigh, NC, USA Kingdam of Saudi Arabia.

One of several hypotheses for the progression of The ;flow of blood through human arteries is atherosclerosis involve hemodynamics and mass tran- simulated by the flow of an analog solution sfer where fatty deposits represent tissue and wall circulating through an elastic rubber tube, under shear stress responses. It is of interest to find steady state flow conditions. The pressure drop/ out ifatherosclerotic lesions occur at arterial flow rate relationship is presented from theore- sites because of locally low or high wall shear tical point of view and canpared with the experi- stress, laminar/turbulent or recirculating flow or mental findings Obtained from a closed flow because of a combiaation of several momentum and circuit. TWo theoretical approach are introduced mass transfer processes. a) tube wall material obeys Hooke's law, b) tube As a first step in the analysis of the localiza- wall material do not obey Hooke's law. tion, developmentand physical effects of athero- TWo specimen of elastic rubber tube, smooth and sclerosis, we solved the Navier-Stokes equations soft, were used to conduct experiments. The first for steady laminar/turbulent flow in a two- specimen resembles the thoracic aorta or the dimensional axisymmetricallybranching channel abdominal aorta dimensions. The second specimen representingan aortic bifurcation. A parametric simulates the femoral artery dimensions. The sensitivity analysis concentrated on the stenosis experimental and theoretical results of the severity (ie, 0%, 35% and 65% lumen reduction), the pressure drop/flow rate relationship are in good bifurcation angle (ie, 7.5° and 22.5°), the agreement, particularly when the tube wall Reynolds flambee (ie, 250 and 500), and the widened material is considered not to obey Hook's law. In lumen geometry due to surgical reconstruction. In conclusion, the elastic rubber soft tube may be adding deposits in local areas of low and high wall used to simulate the blood flow in arteries, shear stress, pseudo-transient,localized develop- especially for the relatively large diameter and ments of stenosis were achieved in the Short length vessels such as the thoracic aorta. parent/daughter tube near the bifurcation. Interestingly enough, these sites coincide with actual fatty plaque locations found in humans. 3 0 7

1988 World Congress on Medical Physics and Biomedical Engineering o San Antonio, Texas 287 Biomedical Engineering Scientific Papers

BE12-H.11 Wall Shear Stress Calculations in BE12-H.12 Constricted Ducts NUMERICAL SIMULATION OF NONNEWTONIAN 3.A. Ekaterinaris, D.P. Giddens, A.K.MazHer" Ga Tech, Mechanical Engineering, BLOODFLOW IN VESSELS Atlanta GA 30332 r. STEFFAN; G. BACHLER, R. PUCHER AVL LIST GmbH., Theoretical and experimental studieson hemody- KleiststraBe 48 namics consider blood as an incompressibleNewtonean - 8020 Graz fluid.In the present work the applicability of a low For the Mach number compressible approach isinvestigated in numerical simulation of blood order to circumvent the difficulties flow in vessels the considerationof Non- associated with the newtonion numerical solution of the elliptic incompressibleNavier Flow-Effects influences the Stokes equations which describe blood flow. flow field. We used the method of finite differences to solve the equations of The full compressible flow equationswere employed, continuityand momentum for three dimen- and the solution was obtained forlow Mach numbers so sions. As we were looking at vesselswith that the incompressible flow limitcould be satisfacto- "largediameter" the mayor nonnewtonion rily approximated. Solutionswere obtained for steady influence is comming from the Rouleaux- flows in axisymmetric constricted du.; modeling condi- effect, which can be described mainly in tions relevant to atherosclerosis studies.The numerical a connection of viscosity and shearrate. solution was performed witha finite difference implicit For this link we used the equation of scheme using approximate factorizationand Alternate Cross. Using an interactive solution-al- Direction Implicit tecniques. The wall shearstress of gorithm we solved these equations for the low Mach number calculationswas in good agree- some typical stenotic narrowings. Acom- ment with both, numerical solutions and theexperimen- parison of newtonian andnonnewtonian tal LDV measurements for incompressible flow fieldshows considerable differen- flows. The ces. present approach is therefore useful for theinvestiga- tion of near wall phenomena in certainsituations which are'important in artherosclerosis studies,and it can be extended to treat unsteady and three dimensionalflows.

BE12-H.13 BE12-H.14 Ambulatory Measurement of Beat-to-beat AN EXPERIMENTAL INVESTIGATION OFPHYSIOLOGICAL TYPE Indirect Arterial Pressure in Human Fingers FLOWS IN LEFT VENTRICULAR AND AORTIC A.Nawarada*, H.Shimazu, H.Ito, H.Kobayashi, ARCH MODELS J.Masuda & K.Yamakoshi, Dept Physiol, R. RIEU*, F. CASSOT**, D. MORVAN*, R.PELISSIER* Kyorin Univ, Mitaka-shi, Tokyo; K.Yamakoshi (*) I.M.F., 1, rue Honnorat, 13003Marseille,France is in Res Inst, Appl Electr, Sapporo-shi, (**) L.M.A., 31, Chemin Joseph Aiguier, Hokkaido, Japan 13009 Marseille, France We had designed volume compensation tech- Using a cardiovascular simulatorallowing nique to indirectly determine beat-to-beat the production of physiological flowconditions, arterialpressure and its Waveforms in ultrasonic velocity measurements human fingers. were made in va- It comprises a compression rious elastic singularity models ofthe cardiovascu- fluid-cuff and its pressure controller and lar system such as aortic arch a or ventricular cavi- photo-plethysmograph to detectarterial ties. Two velocity componeitswere recorded in a volume change. Applied cuff pressure onto sample volume of about 1 mm ; then the correspon- the finger can be controlled bya servosys- ding velocity vector was obtained tem at each instant connected to the cuff around the basal of the simulated cardiac cycle.Using this no clas- phalanxof a finger so as tokeepthe sical method, it was possible to obtain arterial new impor- volume constant at the reference tant original results 3n the main featuresof the level which to flow. correspond the arterial In the aortic arch, skewing effectsof the volume in an unloaded state. In this study, velocity profile toward the innerwall of the bend we have developed an ambulatory system to and rotational creation near this wallillustrate determinearterial pressure using this unsteadiness, entrance conditionsand cur'ature ef- technique. Pneumatic pressure control sys- fects. Flow characteristics in theanatomically sha- tem is adopted in it: The frequencyres- ped ventricle are associated with ponse an important is flat up to 60 Hz. The weightof intraventricular vortex. The analysisof the instan- the cuff and actuator is 180 g, and the taneous velocity field allows, duringthe cardiac total weight including a recorder andpower cycle, the follow up of thisstructure whose inten- supply 2 Kg. Using this system, arterial sity and position are highly depending pressure of the mitral in ambulatory condition weresuc- and aortic valve specifications andin particular cessfully recorded during various exercises of vorticity creation on the mitral and stress tests. valve leaflets. 3 .,tr' 8

288 1988 World Congress on Medical Physics andBiomedical EngineeringSan Antonio, Texas Biomedical Engineering Scientific Papers

BE12-H.15 BE12-H.16 Mass Transfer to and from the Vessel Wall distal CONSTANT SHEER RATE IN HUMAN CEREBRAL to a Sudden Tubular Expansion, X. Deng*, and T. ARTERIES, A BASIS FOR A MDDEL OF THE Karino, Montreal General Hospital, Montreal, CEREBRAL CIRCULATION, R.K.Pucher+, F. Quebec, Canada. Anderhuber, St.Schuy, L.M. Auer University Graz, A-8036 Graz, Austria

To elucidate the possible role of fluid mechanical Exact knowledge of the structure of human factors inthe localization of atherosclerotic cerebral arteries is necessary for the lesions in man, the effects of disturbed flows or evaluation of a mathematical model of this mass transfer to and from the vessel wall distal complex vascular bed. 725 bifurcations in to a sudden tubular expansion (where an annular arterial casts of human cerebral arteries ring vortex was formed), were studied both theor- were investigated. Results show that the etically using computational methods and experi- 3rdpower of the diameters of the parent mentallyby measuring thedissolution rate of vessels equals the sum of the 3rd powers donut-shaped benzoic acid pellets which formed the of the diameters of the daughter-vessels. wall of the tubular expansion. It was found both Sheer rate is therefore only changed mini- theoretically and experimentally that mass trans- mally. Based on constant sheer rate in fer from the vessel wall to the flowing fluid is cerebral arteries, a model of the human much higher within the vortex near the reattach- cerebral circulation was developed. The ment point than anywhere else. Theoretical anal- model includes some 30 million arteries, yses on mass transfer from the blood to a perme- arterioles and capillaries. Non-Newtonian able vessel wall showed thatfor macromolecules behaviour of blood viscosity is taken into such as lipoproteins, having small diffusin co- account. Cerebral autoregulation is simu- efficients,a species concentration polarization lated by means of diameter variation. occurs at thesurface ofthe permeable vessel After numerical solution of the flow wall. The accumulation of species was predicted equations, pressure gradient, flow veloci- to be tke lowest at the separation point and the ty, Reynolds numbers and volume flow can highest at the reattachment point where the mole- be seen in all vessels. Theoretical limits cules are carried towards the vessel wall by the of cerebral autoregulatiot are calculated radially directed curved streamlines of the vortex and compared with measured data. existing in the vicinity of the reattachment (stagnation) point.

BE12-H.17 BE12-H.18 ANALYSIS OF HUMAN CEREBRAL ARTERIAL CASTS Pulsatile Flow of Blood Through an RK Pucher, F Anderhuber+, St Schuy, A-teriosclerostic Vessel, T.Usha LM Auer, University Graz, A-8036 Graz, and R.Ponnalagar Samy; Centre for Austria Mathematical Sciences, Vazhuthacad, Trivandrum- 695014,India:*Depart- Compared to other vascular beds the ment of Civil Engineering, Chuo cerebral circulation shows differences in University, Kasuga 1-13-27,Tokyo. the geometrical structure. Large cerebral arteries themselves and also their bifur- The effects of rheological behaviour of blood and cations are exclusively situated on the pulsatility on flow through stenosed arteries surface of the brain. Therefore, these have been investigated. The theological behaviour arteries are rather long to be able to of blood is described as non-Newtonian fluids reach all points of the cortex. obeying Casson models. The perturbation technique 27 arterial casts of human cerebral has been adopted to investigate the flow. The arteries, containing 1307 single arterial most notable result of pulsatility is the phase branches, in the diameter range from 3mm lag between the pressure gradient and flow rate, to 250um were investigated. The relation which is further influenced by the non-Newtonian of arterial diameter to arterial length nature of blood and the height of the arterio- was found to be similar in the vascular sclerostic plaque. It is found for the first time beds of the anterior (ACA), middle (MCA) that the phase lag between the pressure gradient and posterior (PCA) cerebral artery. The and flow rate, flow resistance, radius of plug data is used to develop a mathematical core region and velocity changes along the axial model of the cerebral circulation. The (flow) direction in the stenosed region. It is of relation of mean intravasal pressure to importance to know the correlation between the arterial diameter was calculated for all change in the phase lag and the growth or altera- three vascular beds. It is shown that the tion of arteriosclerostic plaque on the arterial pressure head loss in the large arteries wall but it is not still clear what the role this of he MCA territory is higher than in the phase lag can play in the growth of stenosis or territories of the ACA and PCA. This the development of cardiovascular diseases. It is result shows that additional pressure therefore suggested that a series of experiments drops (vasospasm, stenosis) are more is to be conducted to identify the role of the likely to cause hypoperfusion in the MCA. phase lag concerning cardiovascular diseases. 3 0 9

1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas 289 Biomedical Engineering Scientific Papers

BE12-H.19 BE12-H.20 An Experimental. Study on the flow characteristics Evaluation of velocity waveform inintramyocardial in a Se serai..:1 Bypass Grafts; S.Kanazawa., artery and vein by laser Doppler method T.Fujiwara, H.Inaua, S,Matsuoka, H.Yoshida, K. Mite, Y.Ogasawara, O.Hiramatsu, Y. Wada, M. Goto, T.Katsumura, Dept. ei Thoracic and Cardiovascular S. Tadaoka, K. Tsujioka, F. Kajiya, Surgery, Kawasaki Medical School, Kurashiki, Japan Kawasaki Medical School, Kurashiki, 701-01, Japan. ?he sequential bypass Jurgery is performed To evaluate theintramyocardial blood flow pattern, ` requently for the coronary revascularizationto we employedthelaserDoppler velocimeterwithan multiplecoronary vessel diseases.Toinvestigate optical fiber (0.062mm). The optical fiberwas inserted hemodynamics in a sequential bypass graft, we intotheintramyocardialartreyandveinoffive provided a sequential bypass graft-n c.1. 1 between anesthetized dogs, and the blood flow velocitynear the the left and rightexternal iliac artecict using rounding fibertip was measured. The blood° flow autologous external carotid vein in tecogs. We pattern of the deep septalartery (10-20mm depth) was measured the blood flow rate and velocitypro- exclusivelydiastolic, accompanied by a sharpreverse file across the graft using electromps.tic flow- flowduringisovolumiccontraction. Following meter and 20MHz 80 channel pulsed Doppler velocim- administration of nitroglycerin,itincreasedtheboth eter, respectively. The flow rate into the side- forward and reverse flows, and the second peakof the side anastomosis is always higher than that into reverse flow appeared during mid systole. Theblood the end-side anastomosis. However, enough blood flow pattern of theintramyocardialvein(depth:1.0- flow was supplied through each anastomosis, indi- 3.0mm) which emptiesintoanteriorinterventricular cating the usefulness of the sequential bypass veinincreasedrapidlyduringthe isovolumic graft for the blood supply to ..schemic heart dis- contraction phase. Itdecreased with the beginning ease with multiple vessel diseases. Regarding the of diastole and a transientreverse flow was observed velocity profile, directional skewing of theve- frequently during early (or mid) diastole. locity profile was recognized in the region just Administration ofnitroglycerinresulted inthe proximal to the side-side anastomosis. The revel ,e increase inthesystolicflow which showed adouble flow near the wall opposite to side-side anastomo- peaked pattern. Italso increasedthereverseflow. sissuggests the existence of flow separation Thesystolicforwardflowintheveinshoweda and recirculation zone in the region. Thus,we reciprocal relation with that of thereverse flow in the considered that the flow condition of the side- artery. It was indicated thatthe phasic blood flow side anastomosis is better than that of end-side patternsintheintramyocardialarteryandveinare anastomosis from the view point of fluid dynamics. more reflective to the effect of myocardialcontraction and relaxation on theintramyocardial vascular beds.

BEIZ-H12.1 BE12-H.22 Hemodynamic disturbances at arterial stenoses invivo BUCKLING OF A THIN CYLINDRIC4L SHELL J.H. Ostergaard*, E. Morre Pedersen, V.E. Knudsen, SURROUNDED BY AN INFINITE MEDIUM: J.M. Hasenkam, H. Nygaard & P.K. Paulsen. APPLICATION TO CORONARY MICROVESSELS Dept. Thorac. Cardiovasc. Surg., Skejby Univ. Hosp. & G. SAMAKE, D. GEIGER; C. ODDOU Instit. Experim. Clin. Res., Aarhus University. Denmark Laboratoire de Mecanique Physique, Universite Paris XII - 94010 Creel!, FRANCE. As the progression of arterial stenoses is considered caused by induced local hemodynamic disturbances,we A theoretical model was developed in order to assess the changes have attempted to characterize the velocity field in the lumen area of coronary microvessels during a cardiac downstream mild arterial stenoses. cycle. A segment of small blood vessel is assumed to be Fifty and 85% area-reducing stenoses wereconstruc- embedded in an isotropic and infinite medium which is subjected ted successively by external banding of the abdominal to a time and space varying tissue pressure, as calculated from a aorta of 6 pigs. Using hot-film anemometry and analog left ventricle mechanical model. It was emphasized that, close to filtering technique, laminar and turbulent blood velo- the endocardium where the tissue pressure is high and for a city parameters were measured and computed for 21 sufficiently thin wall vessel, a buckling phenomenon can occur in points distributed in the vessel cross-sectionalarea the occlusion mechanism. at 3 axial positions. Such an instability process was analyzed under the hypothesis of While the peak systolic velocity was highest 1 diameter a linear, isotropic, and elastic behaviour of the medium assumed downstream 85% stenoses, the absolute and relative to be strained in the plane perpendicular to the vessel axis. For a turbulence intensities were most prominent 2 diameters given pressure, our theoretical analysis gives a formula for the downstream. Velocity profiles revealed a typical post- critical pressure at which an elastic instability appears, depending stenotic jet, whereas just downstream the stenoses on the theological parameters of the medium. the relative turbulence intensity demonstratedan "M "- The results were validat d by comparison with those of shaped radial distribution. Furthermore, the analyses experiments where a PVC tube was either inserted or tethered indicated that the turbulence became more low-fre- within a large piece of latex foam, itself submitted to a controlled quent further downstream the stenoses. pressure. This study shows, that the peak systolic velocity, the absolute and relative turbulence intensity and the frequency content of turbulence in the post-stenotic velocity field strongly depends of degree of stenosis and radial and axial position of measurement. 0

290 1988 World Congress on Medical Physics and BiomedicalEngineeringSan Antonio, Texas Biomedical Engineering Scientific Papers

BE12-H.23 BE12-I.1 A NUMERICAL APPROACH TO PROPAGATIONS OF NONLINEAR Ultrasono-Tomography Using FM Chirp PULSE WAVES IN ARTERIES S.G. Wu*, Beijing Poly- Pulse Compression Technique, technic Univ, Beijing, P.R.China, G.C. Lee, and T.Moriya*,S.Kiryu,H.Matsukawa,T.Fuse, X.S. Ma*, State University of New York, Buffalo, NY Y.Tanahashi+, Tokyo Metropol. Univ., 1-1,Fukazawa 2-Chome, Setagaya-ku, The present paper mainly deals with methods of num- Tokyo,158,Japan ,+Tohoku Univ. erical solutions for the problems of propagations of nonlinear arterial pulse waves. The artery is as- To improve resolution of ultrasono- sumed to be a thin-walled vessel with the small tomography and to remove bioeffects tissue thickness, and its wall is locally orthognal, due to the peak intensity of the anisotropic, elastic and incompressible. In addi- ultrasonic pulse, we devised a scanning tion, the blood is modeled to be an incompressible system incorporating FM chirp pulse Newtonian fluid with the axisymmetrical flow. In compression technique. this numerical experimentation, the problem of pro- The B-mode images for tissue mimicking pagations of nonlinear pulse waves in the aorta of fantoms and abdominal B-mode images ar a dog is studied, and the calculation is being per- obtained using the system and compared formed in IBM-4381 computer. The complete numerical with those obtained using the results are obtained, they involve solutions of pro- conventional pulse method. pagations of the following nonlinear pulse waves: We have found that (1) the images pressure, velocities and flowrate of the blood, as obtained by this method are similiar to well displacements, velocities and stresses of the those obtained by the conventional vessel wall. These results obtained are very useful pulse method,when the center frequency of to explain the mechanism of propagations of pulse the chirp signal from the region of waves in arteries. interest is the same with the :enter frequency of the pulse signal from the same region,(2) without changing the transducer, we obtain the B-mode images with higher lateral resolution only by shifting the center frequency of the chirp signal to the higher frequency.

BE12-I.2 BE12 -I.3 The Acoustic Radiometer Measured Experiments for a new quantitativere- Ultrasonic Therapeutic Instrument flection imaging, Jian-Yu Lu and Yu Wei: Jin ShuweZhejiang Un',versity Dept. of Biomedical Engineering, Nanjing Hangzhou ,China Institute of Technology, Nanjing, China

A new quantitative reflection imaging method which According to acoustic radiated force theory reconstructs the sound speed distributions of an Shotton's method,a new type of acoustic biological soft-tissues using the outline informa- radiometer with float tetnered chians for tion and the phases of rf echo signals on these measuring output mean sound power from the outlines provided by commercial B-scanner has been ultrasonic therapeutic instrumentations and developed by the authors recently. Thispaper others ultrasonic equipments is developeu. reports the The basic priciple,construction experiment studies for image recon- design, structions using the method introduced above. sesitivity cali,,rated and measured method The rf echo signals are acquired from the are described and remakable advantage commer- of cial Japanese B-scanner SSD-256 using the the radiometer is introduced, wave- too.. form-storage oscilloscope HP-1980B with a sampling The reflected target with air backing in rate as high as 30 MHz. An interface between the diameter 6.1cm has the edflected coeffici- B-scanner and the oscilloscope was specially de- ent 99.9%.The vernier gauge and the optical signed to eliminate a random waveform shock caused readout system used for measuring displace- by big tine-delay of the waveform from trigger. ments by acousitic radiated force. The signals sampled are transferred to IBM-PC Nine errors of the radiometer is analysed. computer using The measured repeatability IEEE -488 parallel inter$ace for error forcon- image reconstructions. Several agar phantoms were fidence level 99.7% is better than 2.3%. prepared and were put into an echoless tanker for The sensitivity constant in 0.5-10Watts experiments. Compared measured range is 10.90mm/W. to the images displayed on the screen of the B-scanner, the images reconstructed are more helpful in the understanding of the structures of the test objects.

311

1988 World Congress on Medical Physics and Biomedical Engineering o San Antonio, Texas 291 Biomedical Engineering Scientific Papers

BEM-J.1 Early and late microvolt potentials BE12-J.2 Measurement of ventricular activity /iV/ in patients withventricular tachycardia/fibrillation. F.Walczak: in multi-directional high resolution R.Kepski, W.Rutyno, W.Poplawska, e.c.g. R.KapskitF.Walczak,M.Hoffman National Institute of Cardiology M.Hoffman. National Institute of 04-628 Warsaw, Poland. Cardiology 04-628 Warsaw, Poland Thanks to the spatial electrode In patients after myocardialinfarction location and ventricular tachycardia/fibrillation developed in our laboratory,multi-directional 'Ay noise /0.5pVp-p/recording /VT/VF/ existence oflate potentials /LP/ with may be often found. Ventricular uV could analog averaging was appliedto the pV be also observed insome pts, within /as signal detection. In thisway the ventricular activity detection due "notches"/ and even before /as"early po- to 4 diffe- tentials" - EP/ of the QRS rent measurement directions wts achieved. complex. Advantage of this method:1.Improvement In this study relationshipbetween degree in and quality of myocardialdamage and occu- small signals detection /likelate poten- rence of pV, in the low noiseecg recor- tials, QRS-notches, pre-QRSwaves etc.!, ding was evaluated. In 14pts with primary 2.Estimation of the onset andtermination VT/VF, LP occured in 71%, EP in 1 pt, the points of the QRS "viewed" fromdifferent total ventricular activitytime /TVAT/ was directions, 3.Indication of vectorial x121 ms and theejection fraction /EF/ orientation of ;IV signalson the basis of was R=49%. In 62 pts with secondary VT/VF, their magnitudes in subsequenttraces, LP had 89%, EP- 32%, TVAT 5Ln134 ms .and 4.Possibility of a single-beatrecording EF R=44%. The longest TVAT x=147 ms and or registration with limited number of the lowest EF- 33% had pts Ic.th secondary averaaing cycles. Clinical experience: VT/VF.aod with EP. method applied as a routinetest for pts Conclusion: Appearence of early with different stages ofthe coronary potentials heart disease, dilated and within pts having latepotentials indica- hypertropic cardiomyopathy,arrhythmooenic right tes higher degree of myocardialinjury. ventricular displasia, mitralprolapse and with ventricular arrhythmias ofunknown origin.

BE12-J.3 BE12-J.4 THE COMPUTER SIMULATION OF THE PASSIVE PROPERTIESIN A Method For Nei:invasive Deter- THE CULTURED MYOCARDIAL CELLS 'Tining The Cardiac Electrophy- Qin Chengde*, Hu Yang, Wang Junjian,Dept Bioeng, siolocical Parameters Huazhong Univ of Science & Technology, andLi Fang Zuxiang: Fudan University Li Xiaoguang, Dept of Medical Elec Eng, Xi'anMed- Jiang Wenpino, Suzhou No.1Hosp. ical College, Wuhan, Hubei, P.R. China CHINA

Using glass-electrodes (tip <0.5 um)intracellular recording technique, we measured the parametersof Measuring of cardiac the passive properties in cultured Myocardial cells electro-physiological parameters is a primal thing for study from neonatal rats as follows: membrane capacitance: arrhythmia and is an important Cm, 1.6 uF/cm2, membrane resistance: An 8.9 Kilcm2, evidence to rake a exact diagnosis.and decisionhow to membrane time constant: Tm, 14.2 mS. use anti - arrhythmia drugs. The RM and Tm of the cultured myocardial cellswere Usually.2-4 Catheters were insertedtrans- decreased by adding sodium fluoride (NaF, 5ppm)to vencusry into heart for stimularing.Now the perfusing eagle's fluid. we instead it of using a single quadripolar The effects of fluoride on the passive propertiesin :acrer electrode transesopngeally put a the cultured myocardial cells were simulatedby means oair cf electrodes neat the of Apple-II microcomputer system, according ventrin,so as to BR to stintaate heart effectively to ier-,t (Beeler and Reuter) model. BASIC language, Euler's the electrical acting of heart. method and a self-adaptive procedure were used. mne instrument for this purpose was Simulation results suggest that the effects of de- fluor- s-gneg which is a multifunction ide on Tm and Rm of myocardial cells program- were realized -Able stimulator. Itcan deliver four kind through increasing the potassiam conductance (GK)of of premature stimuli. Si, S2. S3, S4. heart cell membrane. Each one may be programmed indepedently and synchronizedwith P or R wave picked up trot esophagus by the same electrode. The stimulator is also ableto determine reentry tachycardia both by automatic and manipulation. More than 50,000 cases have been examined 3 1 2 in this way in the past 3years in China.

292 1988 World Congress on Medical Physics andBiomedical EngineeringSan Antonio, Texas Biomedical Engineering Scientific Papers

BE12-J.5 BE12-J.6 Studies on High Frequency Electrocar- The direction and distribution diogram in the Early Diagnosis of Coro- of His Bundle potential on the nary Heart Disease, Zhu Wei -ping; body surface, J. Jin*, Xian Zhang Yun- feng, Zhang Chun-hua, Medical University, P.R.China; Tang Lin, Wang Shu-chun, and Gao De-en, G.Schoffa, University Karlsruhe Shandong Teachers' University, Jinan, China D-7500 Karlsruhe,F.R.G.

We conducted the comparative studies between The purpose of this work is to search the opti- applied high Frequency ECG and selective coro- mal lead positions and reasonable space-average nary arteriography to the clinical diagnosis of methods for non-invasive beat to beat His Bun- CHD in three groups of 43 patients with angina dle potential recording(HBPR). Anew IBM-PC XT pectories, myocardial infarction and cardiomyo- (or AT)-controlled multichannel ECG system with pathies. Our data showed that high frequency very high resolution has been developed for this notches could separate patients in CHD group research. The measurement was carried out by 10 from patient° in cardiomyopathies group and that health men in normal environment. The direction significant differences were observed and the distribution of His Bundle potential on between groups (p<0.05). By counting notches in the body surface were investigated with PQ-Seg- from whicl% the 6 leads, it would be expected that patients ment map and PQ,-segment vectorgrtA, the reasonable -oho had more than 6 notches would be associated the suitable lead positions and with 50Z or greater obstruction of single or mul- space-averaging method for nor. - invasive beat to tiple-vessel, demonstrated on coronary cineangio- beat HBPR could be concluded. These new leading methods were then examined by the beat to beat gram. BAsed on the creterion, the correspondence in normal environment by 10 people. rate between applied HFECG and coronary arterio- measurement The result of this work is a of graphy to teh diagnosis in CHD was 74%. recommendation the reasonable leading system for beat to beat Using the high frequency technique, we have stu- HBPR, died: 82 new subjects in Ult middle aged and the aged, results showed that a total of more than 3 high frequency notches in the 6 leads suggested a high probability of CHD.

B BE12-J.8 Effect of blood - flow - induced electromotive force on A Low-Cost Real-Time System for Detection of heart in a hxdr. talpet c field, 0. It* and K. :lakino, Late Ventricular Potentials, MA Poluta*, IWP Kyorin University s:hool of Health 5cicn oes, Mayasiuta Obel, Johannesburg Hospital and University of 476, liachio3i, lekyo 192. t'e Witwatersrand, Johannesburg, South Africa

There are many conflicting papers about the magneticfield Non-invasive detection of late ventricular poten- effect. In this study to discussed one of these problens. 6hal tials (LVPs) is being used as a screening test for aconductivefluidmoves acrossaintpeticlield, cardiac patients at risk to recurrent sustained electranotive force (if) Is induced perpendicular both to Ur ventricular tachycardia, a life-threatening arrhy- roving directionofdr_ :coductive fluid and thenup,netio thmia. An LVP measurement system comprising the field.A study was nnde to investigate theeffectofthe following modules has been used on over 200 pat- electrical potential generated by blood now in the heart in a ients : high magnetic field on the sinus node which provides a rhythm - A high-frequency ECG module with a gain of up of the heart. to 500,000 and a bandwidth of 30-2000 Hz. (The Inthe first place an experiment was nrde onrabbits.11.o upper frequency limit is effectively reduced electrodes were placed for the EU or heart rate monitoring. A by trigger time-jitter) ringlet energized by an electrical current generated a nngnetic - A conventional ECG monitor module for process- fieldintensity of4(XXJ gauss at the centerbetweenthe ing of the reference ECG signal poles.he investipted the clines in heart rate in response - A trigger module for generation of the fidu- to a magneticfield atvarious angles(11)hetKenthe cial point, with adjustable refractory period lorsofrontal direction of the chest andthe north-to-south for rejection of extrasystoles direction of the field.After a rabbit has placed in the high - A Gould 1425 digital storage oscilloscope with magneticfield,we determinedheart rate undervarious waveform processor option for display, averag- conditions;them W = 0 degree,heart rate was fewerthan ing and manipulation of the reference and high control;then14 z 180 degrees,heart rate was mare than frequency ECG waveforms. control;utiai W was 93 and 270 degrees, heart rate vas almost A 200-millisecond time window is used to display of the sale as control.Fran these results it uns found that the QRS-complex and ST-segment. The test procedure therecipryall potentiaLs of these blood-flow-induced EV includes 4 bipolar lead configurations and typi- acted on the sinus node to elicit an increase or a decrease in cally lasts less than 15 minutes. All patient heart rate.The sane results was yielded in the experiment on Llnnections are isolated. nnn. 313 1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas 293 Biomedical Engineering Scientific Papers

BE12-J.9 BE12-J.1.0 Recognition and Termination ofParoxysmal Surface-activated LTI Carbon Tachycardia. Jiang Ding" and Zu-Xiangrang Endocardial Pacing Electrode, Fudan University, Shanghai. P.R.China Ye-cho Huang*, Jie Jin, Wei- xing Sun, Chang-zong Cui,Wen Fu, Xian Medical University, Paroxysmal tachycardia (PT) isan abnorimal rhythm which Xian, China cat, be treated by electric therapy. After analyzing the mechanism of PT.we set up a new threshold for automatic A new endocardial pacing electrode recognition.That is: material- TC:i/Z(cardiac period)+constant. surface activated lower temperature Practical application shows that pyrolytic this threshold is much isotropic (LTIA) carbon has been simpler, more practicable and studied. Its effective than those cur- electrical properties were measuredin vitro, rently used inHollers or monitors. It can distinguish compared with Pt-Ir, Elgiloy,Ag-AgCl. It pos- with high accuracy between PTand sinus tachycardia. sesses very good electrical rroperties andbio- Premature Extra-Stimulation (PES) is a safer method to logical compatibility. The terminate PT. experiments of LT -A In order to quickly find the"Terminating carbon endGrardial pacing electrode Window (TW)".we develop were car- a new searching algnrithm fumed ried out in two groups- LTIA and Elgiloy. Two "Half Division". it can adaptively modify the searching electrodes were implanted in each dog'sright direction by detecting the differentresponse to stimu- ventriule and the differlknce between lation in different sides of electri- the Mend hat reduce the cal parameters of these paired intervals of successive stimuli.Thus electrodes were it is probably far analysed. Statistical analysis of more efficient than the traditional stimulation Linear Scanning algorithm threshold, attenuation rate of QRSwave, endo- where the searching directionand stimuli in- cardial ECG and cardiac tervals are both fixed. Suppose function after elec- H and N are the equiva- trode implantation showed thatLTIA carbon lent stimuli number of HalfDivision and Linear :warm- electrode had a better sensing ing algorithm. theoretical character and cntrulation shows that they lower chronic stimulation threshold. A reduc- have the following relation: H:1.06 N. tion in locr.1 tissue reaction hasbeen verified by, in-vivo tests. This new developed pacing electrodes has been used in clinicsince 1985.

BE12-J.11 BE12-J.12 Performance of a sintered titanium electrode, A Monitor to detect His- and Early Atrial R. Thull* and M. SchalOach,Zentralinstitut Poten- flit- Biomedizinische Technik, Universit4t tials from the Surface ofa Langendorff Perfused Erlangen, 0-8520 Erlangen, FR of Germany Heart, E.H.sfer ',G.Stark,H.A.Trit than, Due to electrochemical properties, thestimula- Karl-Franzens-Universitat Graz, A8010 Graz, tion electrode material does not influencenormal Austria. thresholds as long as pulse duration isshort compared with the time constants ofany electrochemical processes, and the electrodesurface Detection of His-Bundle- and early atrialpotentials re- layer is either electron conductingcomparable to quires intracardiac measurements.W.I. developeda the bulk, or the dielectric is suitable thin and method to measure continously thosesignals by a single of large area. bipolar epicardiac lead in Langendorff Contrary to the threshold performance of the perfused hearts. ma- Simultaneousintracardiac terial, its structure has a significantinfluence (HBE)andepicardiac on the detection capability of small intracardiac (SECG) electrogramswere recorded and compared in signals. In continuation of earlierdevelopment the time and frequency domain. Thepower spectra were studies, an electrode was made of sintered evaluatedtodesignamonitor,consistingof a TiAl5Fe2.5. In vitro test results show a very low preamplifier and a mainamplifier withappropriate filters cutoff frequency and a shortrecovery time after and offset adjustments (overall gain 10', pulse delivery. Characterization shows cutoff frequen- both a cies of 30,80 and 900,3000 Hz). The small deflection from the equilibriumstate of trigger input for the the surface structure and ion distribution,which oscilloscope is switched to the R-waveof the SECG or to are prerequisites for low energy losses inthe the stimulus signal. With this monitor,the regular SECG electrode tissue interface. Optimalsensing and and the windowed His-or early atrial potentials can be stimulation properties have been confimedby a recorded in a beat to beat multicentric clinical study. manner on t' - screen of a simple digital oscilloscopeto evaluate impulse generation and conduction in the intact heart.

314 294 1988 Sorld Congress on Medical Physicsand Biomedical EngineeringSan Antonio, Texas Biomedical Engineering Scientific Papers

BE12-K.1 BE12-K.2 Cardiac output measurement: Accuracy and precision ALGORITHM FOR CALCULATION OF INFARCTION of thermal dilution and ultrasound techniques SIZE AFTER THROMBOLYTIC THERAPY, measured against a calibrated standard and J. Delanghe, L. Cluyse, H. Thierens*, guidelines for clinical use. G Hutchinson'and L I. De Scheerder, M. De Buyzere and Straton, ECRI, Plymouth Meeting, PA. R. Wieme, Dept. Clin. Chemistry, Medical Physics, State University Gent, The thermal dilution and ultrasound techniques Belgium. for measuring cardiac output have comparable accuracy and precision but the specific application Recent advances in myocardial infarction therapy may dictate the use of one or the other because of resulted in an early reperfusion of the infarcted the characteristics and limitations of each. Under zone. As a consequence release of enzyme occurs our controlled conditions, we found both techniques earlier than in untreated infarctions resulting in had errors of 5% or less and had precision of about different release patterns. As released enzymes 2.5% or less; the thermal dilution was technique show differences in catalytic proporties upon slightly superior. We tested over a range of 1to aging, classical algorithms for size estimation 10 L/m and heart rates of 40 to 180 bpm. For this cannot beused as they overestimate infarction study, we developed a pulsatile flow, left heart size (up to 30 %). Taking into account changes in and circulatory system simulator on which we can activation energy of the catalyzed reaction, serum Several flow studies verified calibrate the flow. activity of cardiac enzymes (CK, CK-MB) were cor- that the flow profile within the simulated aorta rected for loss of specific activity in function Calibration was accomplished by was rjalistic. of time. A computer algorithm was designed in having the fluid collect in a resevoir and which time needed to reach peak activity was used measuring the time to collect a known volume. as a corrective parameter. Elimination constant In practical usage, errors arising from user of disappearance of CK from serum was calculated technique are more controllable with thermal from each patient's individual data. Distribution dilution. In choosing to use one method over volumes of enzymes were corrected for age. Using another, it is necessary to consider whether these corrections, clinically useful and reliable absolute accuracy or simple trending is required, estimations of infarction size are possible ,_ad if continuous measurement is advantageous, whether efficiency of reperfusion can be calculated. simultaneous wedge pressure measurements will be needed, if infection risk must be minimized and the skill of the personnel available.

BE12-K.3 BE12-K.4 Cardiac Output Assessmentby theSingle-Breath Estimation of Mixing Volumes from Bolus Method: An Evaluation Indicator Dilution Curves, E.Fonstelien*, R. Srinivasae, KRUG International, Houston, Texas. STEM, 1320 Stabekk, Norway.

It can be demonstrated that abolus indi- The single-breath method of determining cator response after n mixingcompartments cardiac output is a simple noninvasive procedure may be expressed as requiring meal irements of only 02 and CO2 concentrations at the mouth during an expiration. c(t)=A(r(n))-1(kt)n-lexp(-kt) 1) However, the method is highly sensitive to measure where k is the rate constant flow/volume, -ment errors. Also, any lung tissue absorption of r(n) is the Gamma function of n, and t is CO2 that may be present can introduce significant time. A is a constant governed by the in- errors in one estimated cardiac output, if not jectate volume, contra:A, and injection taken into account in the calculations. Recent time. Vascular flow is calculated accor- analysis based on computer simulation has demon- ding to strated the feasibility of obtaining acceptable F=cV./7c(t)dt=A -1 kcV. (c,V. are injec- estimates of cardiac output by the single-breath o 10 o1 o 1 technique in the presence of both normal measure- tate properties). 1) may be transformed: ment errors and lung tissue exchange of CO2. This work extends the above analysis to ln(c(t))=1n(A/r(n))+(n-1)lnk+(n-1)1nt-kt evaluate the effects of the underlying assumptions and with lnt and t as independent variab- of the single-breath method on the estimation of les in a bivariate regression model, it is cardiac output. The assumptions include linear CO2 possible to calculate k, n, and A. The dissociation, constant 02 consumption, and homo- total mixing volume V, e. the vascular geneity of the lung. The likely errors from these space between the inj6ctate inlet and the assumptions were assessed using a gas exchange concentration reading device can be found model with a multi-compartment representation of as V=nF/k. A series of thermal dilution the lung. The results indicate ways to minimize measurementson humans at right heart cat- the errors through appropriate selection of data heterization to find V, and n is reported. segments from the expiratory interval. The results are believ4d to carry diagnostic information. Normally the right heart Work supported under NASA contract NAS9- 17721). cavities are V=210m1 and n=2,5. 315 1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas 295 Biomedical Engineering ScientificPapers

BE12-K.5 BEI2 -K.8 Personal Computer Based Cardiac Output Impedance Cardiographer. A PORTABLE SYSTEM FOR BEDSIDE COMPUT- ERIZED ANALYSIS OF CARDIAC CYCLE IN S.Charleston*,T.Aljama,M.Cadena. CRITICAL Universidad Autonoma Metropolitana PATIENTS, S.G.Lage, M.A.Gu- -Iztapalapa,E.E Deparment,09340 tierrez, J.L.Santello and S.S.Furuie, Heart Institute-Faculty Medicine-Sao Mexico,D.F.,Mexico. Paulo University, C.P. 11.450, Sao Paulo - BRAZIL. In this work we present design criteriato obtain a cost-effective single card,bus "pc" compatible, The paper describes the applicationof a portable system developed in a microcomputer impedance cardiographer for measurement of cardiac IBM-PC compati- output in patients at rept and in exercise. ble to perform a complete analysis ofcardiac cycle This cardiographer in addition to detect thoracic events. The simultaneous and continuouscarputerized impedance changes less than 0.1 ohm,using 350 evaluation of the ECG and the firstderivative cur- microampers at 50 kilohertz,is able to correct ve of thoracic electrical bioimpedance (dz/dt) al- artifacts induced by respiratory and electrode lows determination of: pre-ejectionperiod, ventri- movements. cular ejection time, isovolumetricrelaxion time The system is designed to take advantage of the and ventricular filling period incritical heart failure during rhythmic and digital signal processing support found ina per- arrhythmic conditions. sonal computer,in order to perform filtering and The software was implemented inPascal and Assembly languages and consists of the following averaging algorithms correlated with ECG and phono modules: signals. configuration, acquisition and analysis.The first modules permits to set up The performance of this system has been testedon a sample rate and the number of the signals. The acquisition twenty subjects during upright cycle exercise at routines per form the process involved in 0,50,100,150 and 200 watts,using a CO2 rebreathing signal conversion, dis play and storage during 30 seconds method and the Kubicek model to compute the per minute for a car- total 15 minutes of monitoring. diac output.Evaluation of the first results shows These parameters can be changed according to the good agreement between both methods. different clinical demands. The analysis modulecorrelates the signals using cross-markers tomeasure the time of systolic and diastolic events. In theCritical Care Unit the system permitted a bedside quantificationof normal and abnormal rhythms.

BE12-K.7 BEI2-K .8 Automated Evoked Potentials Analysis 1 for Cardiac Surgery Monitoring, Portable Arrhythmia Monitor- Gutierrez*, S.S.Fuxuie, U.Tachinardi ing System, Fan Xiaodong",Pang and C.P.Melo, Heart Institute of Sao Tianci and Huang Yecho, Xian MedicalUniversity, Xian,P.R. Paulo UniversityMedicalSchool, C.P. China 11.450, Sao Paulo - BRAZIL.

A long-term arrhythmia monitoringsystem for The response of the Central NervousSystem (CNS) to ambulatory patients, composed ofa microcom- external stimulii, Evoked Potentials (EP),is an im puter and a tape recorder, portant parameter for neurological diagnostic hasbeen develop- aid. ed. Derivation-Sum method is used The analysis of some characteristics of to recogni- the EP al- ze QRS complex and measure its width. Arrhy- lows the clinicians to evaluated possibleCNS dam- thmia is discriminated accordingto heart rate ages during cardiac surgery. The calculation ofthe and QRS width. CMOS NSC800 mean EP (MEP) using averaging techniques, microcomputer fami- enhances ly (including CPU and RAM-I/O-Timer), the low level of the S/N ratio of EP. The other EP moni- CMOS chips (including 6Kmemory? ADC, LCD) and toring system available, areausually dedicated micro-keyboard are employed to reducethe po- equipments composed of one stimulatorand averaging wer consumption, weight and size of module, that do not permit software changes. the system so that it is suitable for battery-powered It is introduced in this papera microcomputer- and portable. Micro-tape recorderreplaces fa- based system for EP analysis in cardiacsurgery mon shionable telephone to establishrelation be- itoring. The software was developed inPASCAL and tween monitor and central station.Therefore, ASSEMBLY Languages and consist of the following md replay system is simple andlow costs. The ules: Acquisition and Display, Disk-Storage and evaluation is performed accordingto recorded Analysis. The former uses the stimulator synchro - ECG including various arrhythmia,noise and nism pulse as a trigger to the EP signals and thus 256 PVCs. Detection rate ofR wave is 96.89%. MEP calculation, the MEPsarea then stored for for- Negative false detection of PVCis 22, and po- ther analysis. The analysis includeresponse-latercy sitive false is 54. and amplitude estimation as a functionof the sur - gical (from pre-to-post-operatory) procedures. Although the aim of this system isto attend (zodiac surgery acts, it can also be used in other experimental and clinical studies.

296 1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas Biomedical Engineering Scientific Papers

BE12-K.9 BE12-K.1.0 Battery-type Sequential Pulse Defibrillator by using Continuous Blood Mass Densitometry A Single Capacitor Bank, H. Makino, Y. Saitoh, in awake humans, M. Moser *, D. Schneditz, Y. Mitamura, and T. Mikami, Niigata Univ., Niigata and F. Pittner and T. Kenner, Physiolog. Inst., Hokkaido Univ., Sapporo, Japan A-8010 Graz, Austria

A battery powered sequential pulse defibrillator was A method is presented for the continuous arterial designed and tested for the development of implantable blood density measurement in awake subjects. The sequential pulse defibrillators by modifying the output resonance frequency of a tiny U-shaped glass tube circuit. Two GTO thyristors were used as the switching (ID 0.8 mm, sample volume 0.03 ml) perfused with device in the output circuit, and charged energy in a blood is measured continuously. From the oscilla- capacitor bank was divided into the first and second pulses sequentially. tion frequency the mass density of the blood (pbi) An electrode catheter which has can be calculated very precisely two glassy carbon electrode elements was introduced (10 -2 g/1). We applied this techniquein through the right jugular vein. The distal electrode(+) healthyvolunteersto observe arterial was located in the rightventricular apex. and the blood density variationsdue to breathing and proximal electrode(-) waslocated in the superior vena salineinjections.NormalPIAis cava. A plate electrodewas used as a subcutaneous around 1049 g/1 and ppia., around 1019 g/Iat electrode(-) under the left chest. 37"C. Inspiration is followed by an increase ofpbi, In animal experiments, two mongrel dogs (10.4 & 11kg) exspiration leads to a decrease of pbi . Depending were anesthetized and mechanically ventilated. The on breathing volume and frequency, blood density thresholdof simultaneous-pulse(5ms) and sequential- oscillations up to 1.5 g/I were observed. Injections pulse(5-1-5ms) defibrillationwere 400V(1.44J) and 260V of isotonic saline solution produce dilutioncurves (0.69J) respectively. In oursystem, the defibrillation suitable for calculation of blood volume and blood energy was 50% less thansimultaneouz defibrillation. flow. Theoretical considerations to understand the Furthermore, the electrodesystems do not require observed results are presented. thoracic implantation surgery.

BE12-K.11 Br12-K.12 A SAMPLED BOLTER ECG TAPERECORDINGSYSTEM FOR Determtnition re Cardiac artput2nrPick's MONITORING ARRYTHMIA, Li Ming Liang*, Tian Anmin, IrrlireetNether?. rls.i.-747 a Hangar Micro- Ministry of Electronic Industry, Shannxi, Baoji, art:touterSystem Yafiez S.0.*, MedinaB.V.. China; Zhang Yiming, Physiology Teaching & Research c.'arrasco S.S.':JAM-Tztapalapa Av.Ptrlsima V Office of Fourth Military Medical Inst, Shannxi, Michoacth. Miaxico. 09340, D.F. China This paper describes a Holter ECG tape recording One of the clinicalindexes most frequently system which can record the amplified cardiogram used to evaluatetheperformanceof the signal obtained from the second electrode on the cardiovascular system is the cardiac output. there surface of the body through tne (a) low noise am- invasive techniques are usually employed for its plifier and by -he frequency modulation and photo- (let erminat mr.However.non-invasive methods are electricity coupling methods. By adopting the am- preferredifthe determinationis to be Male plifier with a high-resistance and high-common-su- drrinq exercise. ppression data as well as the photoelectricity In this work a system far the autcmatic coupling technique, the electric network interfer- determinationore thecardiacoutputdurinq ence and other environmental interference have been exercise.used on 0)2 retreathinq method.is eliminated, thus making the recorded cardiogram sig- presorted. This system consistsof a modular nal fidelity and reliable. The adopted Holter tech- Mc-rot-molter.based on a Z-80 microprocesscr. nique has also eliminated the error caused by un- with acquisition.processing ard liquidcrystal stable tape speed and the flutter of the recorder. display capabilities.connectedto a -Past002 With many advantages such as its multifunction, low analyzer arri toa flow transducer. The consumption, small size and low price, the system microcanputer automatically performs fcnrr stages is really an applicable medium for widely expanding required to obtain the parameters needed for the evaluation of the cardiacoutput..These steps the Holter dynamic electrocardiogram. incltde the acquisition and presentation of the followinq variables:expired minute volume. CO2 ehliminstion,end tidalmiceand PCO2 equilibrium in relr eathinq maneuver. Ba.sed on these parameters,theresultingcardiacoutrartis comportedand presentedontheliquidcrystal 7 display.

1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas 297 Biomedical Engineering Scientific Papers

BE1'2-K.13 BE12-K.14 Development of Portable 12-lead ECG Simulator TimingModule for a COPAL UA-751 using 3-Wave-Data Synthesis. B.P. Recorder, D.J.Stewart, M.A.Chaudhri*, H. Makino. Y. Saitoh,I.Ishii, andJ. Yamato, Austin Hospital, Melbourne, Australia Niigata Univ. Niigata 950-21, Japan The Copal UA751 is a low cost, digital blood pressure recorder. For the development of a 12-lead ECG simulator,a new The user initiates a recording by pressing the method for restoring waveforms inless data memory was start button. The unit inflates the cuff proposed. From the geometrical relationship of electrode and measures systolic, diastolic and arrangement for the detection of ECG, we selected three heart rate via the oscillometric method. basic waveforms (II,V1,V3) which can be used to estimate The readings are displayed on an LCD display other waveforms by linear combinations. The basic and the user then has the option of printing these results waveforms were decided from the results of calculating on a small thermal printer, or storing up to 14 the method of least squares, and the calculated weight recordings in memory and printing them coefficients were used for restoration of other waveforms. out at a later date. In the system, the basic waveform data and the This is OK for "one off' recordings, but in many coefficients were stored in a ROM. A 8-bit microcomputer, diagnostic or research applications (such as evaluation of and two D/A converters (quad type) were used the to effects of drugs on apatient's blood synthesize the eight channel waveforms. Four other pressure) periodic recordingsare needed. waveforms (lll,aVR,aVL,aVF) were combined with thetwo Tra5itionally, this has been therealm of waveforms(I,ll) by using an analogue circuit. Original ECG machines such as the Johnson & Johnson Dinamap, waveforms were obtained from the ECG worth in the vicinity of tape recording A$5,000. To achieve this or from the ECG chart using an optical scanner. The we have designed and constructed a low cost (ca.A$50) and sampling frequency was 2.5kHz to input pacemaker spike. simple 'Timing Module", with which Finally, the data memory became 7Kbytes, including three automatic periodic measurements can be basic waveforms and coefficients, andit was 62% less madeusing the Copal recorder. The design and functioning of thismodule are than memories from the original eight-wave data recording. described.

BE12-K.1.5 Microcomputer Assisted BE/2-K. 16 Cardiokymography, D.A.Guan*, Signal Processing Techniques and L.Huang, J.W.Huo, and J.I.Yan, Detecting-Recognizing Algorithms China Medical Universit:o, of the characteristic points for Shenyang, P.R.China. ECG, PCG and CPT, Zhou Guang-Huf Yu Jia-Xian, Chengdu Inst of Radio Cardiokymography (CKG, is a contactiess Engineering, Chengdu, PR China. method of recording chest wallmovement related to displacements ofmyocardium. To overcome its technical One of the significantmeans for exami- difficulties ning the left ventricle functions is in the diagnosis of coronaryartery di- to sease, we have developed a suitable CKG measure the heart Systolic Time Inter- transducer and a microcomputer-based vals (STI) of ECG, PCG and CPT. The signal digital processing technique multiple CKG system. The CKGtransducer of the- which weighs 2 o: is supportedon the se three physiological signals and the algorithms of detecting and chest by a two-layer belt withoutlimit- recognizing ing the chest wallmovement. The CKG their characteristic pointsare discuss- signal is demodulated througha phase- ed. A twice-filtering algorithm for ECG, locked loop. Differentiated CKG a reckoning algorithm with adaptive dual signals -thresholds for QRS wave and from two sites on the chest,together a sample-by with a ECG signal anda respiration sig- -sample base line correction algorithm nals, are processed in realtime by the with piecewise linearization forCPT we- system. To eliminate breath interference re proposed to eliminate the main interferences. two digital processing methodsare adopted: multipoint averaging This attractive engineering method requ- and adap- ires much less computation work using tive filtering. Recordings duringtread- a mill exercise show both ofthe methods microcomputer system then the ordinary. are efficient and that it ispreferable to use the differentiated CKG signalin the clinical diagnosis.

318 298 1988 World Congress on Medical Physics and BiomedicalEngineeringSan Antonio, Texas Biomedical Engineering Scientific Papers

BE.12-K.17 BM-LA Non-invasive Cardiac Output Estimation Some Analyses on the Mechanical by using Inert g9s Inhalation Behaviour of Elastic Splint for Yuichi'. Upashi Koji. Pelvis Reduction (ESPR), Gao Ruiting*, Uchiyama Akihiko. Suma Kozo. Xian Rekai, shen Jifen, Dept. of +Sch.of Sci.&Eng. VASEDA Univ. ++Tokyo Mecahnics, Tianjin University, Shang Vomen's Medical Col.. Tokyo. JAPAN Tianyu, and Zhao Wenkuen, Tianjin orthopaedic Institute Our method is to measure cardiac output Presented in this paper are some (CO) in non-invasive way. To use this biomechanical analyses and measurements method. we can easily obtain information of the mechanical behaviour of (ESPR) which on pump function of the heart. In our has been used with satisfactory results method. the inert gas of only 5% N20 is by the orthopaedic Dept. of Tianjin Hospital used as a tracer. If the concentration of in curing fracture ofpelvis. The tensile No0 in the air inhaled by the subject force exerted by the elastic band of ESPR changes stepwise. the ratio of the con- was measured with a specially designed centration of the N0 in inspiration and 2 transducer. In order to understand the in expiration is formed by the sum of two mechanical behaviour oftheelastic band, exponential curves because of the system tension, Creepand relaxationtestswere function of cardiovascular system. CO can conducted; relevant curves with the be estimated from the amplitudes and time relationship of constants of these curves. tension VS. elongation were obtained and the decrease of reducing Our system is designed to change the gas force with -oncentration stepwise between 0% and 5%. time caused by the stress relaxation under constant ld can calculate CO in a few minutes. All elongation was dertermined. or these processes are done automatically. Static imitations weremade on 10 normal adults, clinical We measured the CO of 15 dogs to compare monitoring onmalepatient with pelvis with our method and electromagnetic blood fracture was carried out flow meter. There was coincidence between for 30 hours. dynamic these two methods (r=0.8). We can conclude measurements were made on a normal and a invalid respectively when they that our method is effective as a non- walked invasive CO measurement. and did exercises; and finally, the reducing forces provided by the splint were analysed.

BE12-L.2 BE12-L.3 MechanicalMeasurements of the stability Fracture Toughness of Human Enamel of Reduction Fixator for Treating and its Measurements, Y.Zhang*,M.He Tibio-Fibular Fracture, Gao Ruiting*, and C Kan. Institute of Mechanics Dept. of Mechanics, Tianjin University, Chinese Academy of Sciences Beijing Meng He, and Jin Yang. Institute of orthopaedics and Traumatology, China Academy of Traditional Chinese Medicine The indentation technique was employed to After fixation with the reduction and fixation de:ermine the fractu_-e roughness Kc of human device, the fractured tibia and fibulacan enamel Four teeth were selected for testing: form a new mechanical system with the device. a maxillary incisor, a mandibular incisor, a Thissystem should havea certain ability of maxillary canine, and a mandibular molar. in resisting external distubance. From the order :o verify the results, the calculation standpoint of clinical practice, there are of Kc was performed with two different formulas, four types of minifestation of the forces in one of which a parameter was modified so that subjected to the wounded leg, namely, the :ne formula used for ceramics would be more antero-posterior bending, lateral bending, reasonable for enamel. The results obtained from axial compression and torsion. In order to this study indicate that there exists a slight understand the stablity of the system, the difference in Kc between four different teeth, effect of the reduction and fixation device and for natural tooth the Kc of emamel is about for fixing the fractured tibia and fibula was 1.0 MN/M-1.2 MN/M. An error analysis of this simulated with two specimens of cadaver bones method was given in :his paper. and tested mechanically in this experiment. By means of thetransducersmounted on the two supporting rods and under the lateral pads, the forces provided by the rods of the reduction device were measured when the obilique or transverse fracture was compressed axially. Meanwhile, the force variationsof the pads induced byapplying bending or torsionwere also measured. The stiffness of thedevice is comparedwith samemultipin-type fixation devices used abroad. 319

1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio. Texas 299 Biomedical Engineering Scientific Papers

BE12-L.4 BE12-1..5 Three-dimensional Analysis of A Study Stress Distribution in Periodonal onFinite Element Stress Analysis ofEndodonti- Ligament of Humaii Tooth by Finite cally Element Method, Y.Zhang*, M.He Restored Tooth, W.H.Changl and S.Vnot;, Institute ofWchanics, Y.S.Ueng, and Chinese Academy of Sciences J.S.S.Wu, Dept. of Biomedical Engineering, Chung Yuan Chris- tian University, Chung Li, In order to determine thephysiological Taiwan, 32023, R.O.C. stress values in the periodonal ligament We use finite element (PDL) effectively,a new calculate model method to analysis and method were established. the stress magnitude and With this when distribution model. the tooth and the several types of loading are ap- alveolar bone plied were considered to be rigid, andthe stress on endodontically restoredtooth. distribution in PDL were analysed A two-dimensionalplane strainfinite by three- element model is constructed dimensional finite element method.In the and applied to evaluate the calculation, the basic solutionsof stress stress magnitude of distribution in PDL were obtained differentlength of post anddifferent at first. material Based on them the stress of crown for a givenload. In distribution in order to get a resonable PDL under arbitrary loadcan be obtained accuracy, 190 readily with the principle of plane strain quadrangleelements and 216 superimposition. nodes Three typical load cases are used. We get the following were selected in this results: study. And the results ofstress distribution (1)The dentine ismore highly under these loads were discussed. stressed in restorationwith both longer posts and shorter posts. (2) The crown is more highly stressed in with restoration resin crown than porcelain crown. (3) The dentine is notinfluencedin restoration with either resincrown or porcelain crown. (4) Both dentineand crown are more highlystressed in restoration when a horizontalloading is applied at the top oftooth.

BE12-L.6 A new device for the in vivo monitoring BE12-1-7 of Three Dimensional Finite Element fracture healing by vibrationanalysis. Stress Analysis of the SAD Implant G.Nikiforidis*, A. Bezerianos, Skoudis, Coated with Bioactive Ceramics. B Rang Shaoan*, Chen Anyu, and Huang Proimos Medical Physics Dept. Suhua, College of Stomatology, Rest University of Patras, 26001,Patra, Greece China University of MedicalSciences Chengdu, Sichuan, China Uncertaintyinthediagnosisofthehealing progressof fracturedbonesrendersthe The design optimization development of a quantitative methodfor such of dental implant diagnosis useful.In recent years the authors is one of the criticalelements for suc- conductedanalytical and experimental studies cess. The objective of this study was to regarding the bending vibration examine the stress distributions around offractured the dental implants with or bones in vitro. On the basis of theaforementioned without the investigation, a new device intramobile shock absorbing device(SAD). was constructed for The characteristics of this newimplant monitoring the fracture healingin vivo. system are:(1) the metal In the present worka schematic description of core coated with thedevice bioceramics( Hydroxyapatiteor Bioactive anditspossibilitiesaregiven. Glass Ceramics );(2) Specifically:1) The functional a shock absorbing criteria, the device within the implant;and(3)two-pha- components of the apparatus and the se dental implant.The model experimental conditions of its optimal was constru- response cted based on our previousstudies.Verti- are analyzed. Special emphasisis given to the cal, horizontal, and inclined influenceofthechanging loads were inthesupporting applied on the occlusalsurfaces of the conditions of the bone underexamination. 2) The repeatability of the measurements implants,respectively. The results indi- is tested and cated that the implant withthe SAD had a comparative evaluation between the a better stress distribution around spectra of the intact right and vibrational the left tibiae of the interface than that of the implantwith- same subject was performed.3)Finallythe out the SAD. The authors differencies in the vibrational suggeste that spectrum of an more implants with the design of the SAD intact left to that ofa healing right tibiae, of the will be used in clinic in the same subject are discussed. future. 320 300 1988 World Congress on Medical Physicsand Biomedical EngineeringSan Antonio, Texas Biomedical Engineering Scientific Papers

BE12-L.8 BE12-L.9 Comparison of X-Ray and Isotope-Based Dual- Site-Specific Assessment of Bone Mineral Energy Bone Densitometers. C.C. GlUer*, P. Density at the Proximal Femur by Quantitative Steiger,R. Selvidge,H.K. Genant, University Camputed Tomography. C.C. Gluer*,P. Steiger, of California, San Francisco, CA 94143. S. Steiger,J.E. Block,H.K. Genant. University of California, San Francisco, CA 94143. Traditional Dual Photon Absorptiometry (DPA) bone densitometers use an isotope radiation Quantitative computed tomography (QCT) provides sources whereas the newly introduced for selective measurement of trabecular and Quantitative Digital Radiography (QDR) device cortical bone, and various shapes, sizes and uses an incorporated X-ray generator. We contours of regions of interest (ROI) can be compared these technologies via phantom and chosen with flex1bility. Drawing upon these patient studies performed on a Nor land 2600 DPA features to assess bone mineral density (BHD) at scanner and a Hologic QDR-1000 QDR device. the proximal femur, we explored an approach that Reproducibility errors obtained on an uses an ;.P. Cr scout view to determine the region anthropomorphic spine phantom over a period of 2 for axial scanning. Contiguous 3mm slices are months were 1.9% (DPA) and 0.44% (QDR) at skin obtained encampassing the femoral head, neck and dose levels of 1.5 mrem (DPA) and 2.6 mrem trochanters. Because of the inhomogeneous (QDR) . For QDR. the error rose by about 50% when distribution of BMD at this site, a technique increasing the lucite thickness of the phantom that allows for reproducible placement of the friar 7 to 8.5 inch. For DPA this was avoided by ROI is mandatory. Using contour tracking the adjustable scan speed option. The fat algorithms to assess trabecular BMD of an sensitivity was identical for DPA and QDR (1.2% anthropomorphic hip torso phantom yielded a reduction of bone mineral density per 10% by reproducibility of 1.3% for a volume of 1.25 ml volume fat content). Cross-calibration results of trabecular tissue at Ward's triangle (dose of the first 16 patients yielded a coefficient about 300 mrem). Corresponding in-vivo results of correlation of r=0.98 (p<0.0001), a standard were 1.9-2.5% for volumes of about 10m1 (at the error of the estimate of SEE=0.043 g/cm2 neck) and 0.23-0.7% for volumes of 40-45m1 (CV=4.9%) and QDR results lower than DPA by an (whole proximal femur). Techniques for comparing average of 0.019 ± 0.045 g/cm2. The improved corresponding ROIs of serial measurements will resolution of QDR facilitated the identification be presented. of bone edges of osteoporotic patients.

BE12-L.10 BE12-L.11 THE SPINOSCOPE: TECHNOLOGY FOR OBJECTIVE Anisotropic Viscoelastic finite EVALUATION OF SPINAL CAPACITY, S. Gracovetsky*, Element Analysis of Vertebral Body, Cheng S. Levy, V. Zeman, R. Ben Said, I. C. Pitchen, J. Mlle, Peilai*, Dept. of Mechanical Engn. M. Kary, J. Lewis, Concordia University and Diagnospine Tianjin Institute of Technology, and Research Inc., Montreal, Canada. Chen Hongxing, Tianjin University, Tianjin, China An axisymmetric finite element model is A new device,the Spinoscope (Spinex Medical empolyed forthe study of thecreep behavor Technologies, :nc. Montreal) has been developed for the of an vertebral body under axial compressive objective evaluation of spinal capacity. The Spinoscope loading.The trabecular bon.; of vertebral body collects and processes two basic types of data as the patient is considered as a transversely isotropic performs simple activities, such as lateral bending, flexion, viscoelastic material and cortical bone and extension, lifting of weights, etc. First, spinal motion and endplate as a isotropic viscoelastic material. geometry are monitored using an advanced infrared A three parameter Kelvin solid model is empolyed three-dimensional motion analysis system. This system is in the finite element model. The material capable of tracking the position of 24 LED skin-surface parameters of the anisotropic viscoelastic markers with an accuracy of 1 part in 500, at a data constitutive relationfortrabecularbonewas acquisition rate of 180Hz. This large dynamic range allows obtained from experimental data. TLe material the effects of the time-dependent, viscoelastic response of the parameters of cortical bone is from Yang. soft tissues of the spine to be observed. Second, and The stress distribution and the deformation performed concurrently with the above, the activity of eight of vertebral body under axial compressive loading muscle groups is monitored bilaterally via skin-surface wereobtained. The resultsindicate thatthe electromyography. Subjective information (in the form of the magnitutde of stress in trabecular bone is patient's responses to a pain questionnaire) also collected, smaller than that of thecortical bone and and processed using artificial intelligence techniques. By this thecortical bone carriesa major percentage combination of subjective and objective information, the of the comparessive loading.Thy results also Spinoscope can greatly assist the physician in the diagnosis showthat the stress in vertebral body will and follow-up evaluation of spinal injuries. Sensitivity is redistribute withtheincreaseof time. The great enough to allow patient progress to be monitored on a vertical deformation have a significant changes weekly basis. during the creep time while the radial deformation have a little. 3 2

1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas 301 Biomedical Engineering ScientificPapers

BE12-L.12 BE12-L.13 Joint Stiffness Following Physiotherapy, J. Bromley, The influence of lowmolecular weight A. Unsworth* and I. Haslock, University components on the of Durham, DH1 3LE, England. viscoelasticity of synovial fluids. G. Ribitsch,V. Ribitsch and D. We have Schneditz-t Dept. of Physical designed and built a microcomputer Chemistry Univ. Graz- Austria controlled arthrograph tomeasure stiffness in the mp joint. The joint isoscillatedabout its Healthy human synovial fluids equilibrium position SF exhibit and the arthrographmeasures high viscosity and elasticity the due to a elastic resistance and theenergydissipated component, of high during flexion/extension. molecular weight This is used to assess hyaluronic acid (HA). In the effectiveness of the case of treatment. rheumatic diseasesboth Stiffness was assessed in viscosity and rheumatoid patients, to elasticity decrease see remarkable. Mwand the effect of physiotherapy. Four modesof the molecular weight physiotherapy distribution (MWD) were investigated over 6weeks of of HA isolated fromthose SF does not treatment. These were wax baths, pulsed ultra show significant alterations sound, compared to exercise and a combinationof wax together HA from healthy SF. with ultra sound. Short term changes in stiffness Investigations of model substances(Guar, were investigated (after one singleapplication of HAfrom umbilical the mode of physiotherapy), cord) indicated that as well as long term the viscoelastic propertiesof solutions effects (repeated applicationsover 6 weeks). Wax plus ultra sound produced are changed remarkable if polymerswith a significant short term molecular weightsmaller then 105 are reductions in stiffness (torquep < 0.001, energy p < 0.05). present. Nevertheless Mw and theMUD were However over the 6 week period only slightly oftreatment, shifted toward lower no significant change wasseen in values. The same findings thestiffness were observed before and after the course. In with human SF containing otherwords small amounts of each application ofwaxplusultra low sound molecular weight HA. 5,. of low reduced stiffness but thisincreasedagain molecularHA by the following application. added to an healthy SF Pulsed ultra sound decreases the viscous alone was similar but wax component about 30% alone increased the and the elastic component stiffness about 60%. The in the short and long term. Exercise presence of low molecular components produced no change at all. may inhibit the formationof entanglements and weaken the network.

BE12-L.14 BE12-M.1 Reference Material for CT- Measurementson Bones RESPIRATORY EFFECf ON THE BLOOD VOLUME OF PULMONARY CAPILLARIES, M.Hunter*, Faust, U., Institut fUr BiomedizinischeTechnik Jen-shih Lee, Timothy Fallon, Liwii-pin der Universitat Stuttgart,Seidenstr. 36, D-7000 Lee, UniVersity of Virginia, Stuttgart 1. Charlottesville, Virginia 22908

For the quantification of bone minerala nonto- We measured thedensity variations of aortic xic, solid water-equivalent reference system blood from the rabbits ventilated bya positive with long term stability has beendeveloped. end inspiratory pressure of 6 mmHgor a negative The reference systems consists of polyethylene, box pressure ofthe same magnitude. For both magnesium oxide and calciumcarbonate. The pro- ventilation methods,the fluctuationin aortic duced material has a water-eqivalent x-ray atte- blood density decreased from .5 g/1 at a nuation at voltages of 96 and 125 kV andcan be ventilation frequency of 20 cyclesper minute to used for automatic recalibration of CT-machines. .2 g/1 at a frequency of 35 cyclesper minute. To determine mineral content of bone, a referen- Calculated from the densityvariations, the ce material, which is equivalent to bones, was fluctuations in bloodvolume ofthe pulmonary produced bymixing thewater reference with capillaries within one cycleas induced by an apatite. The linearity between CT-numbersand intermittent positive pressure ventilationwere apatite concentration was nearlyideal. found to be similar to theones induced by an The mineral content of cattle bones (femur) has intermittent negative pressure ventilation. The been determined with quantitative CT and the corresponding percent changein capillary blood water- and bone- references. To study the error volume decreasedfrom 6.5% at 20 cycles per of varying fat/water ratio inmarrow, they nave minute to 4% at a ventilationfrequency of 35 beenexaminedin the native state and after cyclesper minute. Since thetranspulmonary substituting the marrow by water and paraffine. pressure and the transmural pressureacross the The mineral content first has beencalculated pulmonary capillaries are thepressures deforming fromthe CT-numbers of bone and the reference-mate- capillaries, our result suggests thatthe rial without a correction for thiserror. changes in these two pressures are similar for Finally the mineral content was determined with the two ventilation methods. (Supportedin part single- anddual- energy-methods witha correc- by HL 36285). tion for this error. 322 302 1988 World Congress on Medical Physicsand Biomedical EngineeringSan Antonio, Texas Biomedical Engineering Scientific Papers

BE12-M.2 BE12-M.3 Axial and Secondary Plow Velocity Patterns in a Human ANALYSIS OF THE CHANCE OF RIGHT VENTRICULAR Pulmonary Artery Model EJECTION FLOW PATTERN IN PULMONARY HYPERTENSION Hsing-Wen Sung* and Ajit P. Yoganathan, Georgia BY LUMPED PARAMETERMODEL institute of Technology, Atlanta, GA 30332-0100 Y.Yoshida*, A.Kitabatake, J.Tanouchi, K.Ishihara, K.Fujii, M.Uematsu, H.Kusuoka, M.Hori, M.Inoue, The purpose of this study was to characterize the T.Kamada, M.Fukushimat, H.Abet pulsatile flow velocity patterns in an in vitro model of a The 1st Dept. of Med. Osaka Univ. School of Med., "normal" human adult pulmonary artery. A two- Osaka National HospitalT, Osaka, Japan dimensionallaser Doppler anemometer system was used to map the axial and secondary velocity fields in the To elucidate the mechanism of the change of main (M Pt.), left (LPA) and right (RPA) branches of the rightventricular ejection flow (RVEF) pattern pulmonary artery model in both the bifurcation plane and shortening of the time to peak flow (AcT) of and the perpendicular plane. RVEF in chronic pulmonary hypertension (CPH), we In the MPA, a broad central flow field was observed. analysed RVEF pattern by a lumped parameter cir- In the LPA, the axial velocity profiles in the bifurcation cuit model. The model was composed of four ele- plane were slightly skewed toward the inner wall. In the ments, characteristic impedance (R1), inertance of perpendicular plane, the axial velocities were relatively blood mass (L), pulmonary artery compliance (C) evenly distributed.In the RPA, the axial velocities in and pulmonary peripheral artery resistance (R2). the bifurcation plane were very unevenly distributed. These parameters were determined from cardiac Higher velocities were observed near the inner wall, catheterization and Doppler data. RVEF pattern while lower velocities were observed near the outer wall (dome-like,triangular and double peaked pattern) and in the center.In the perpendicular plane, very was well simulated by the lumped parameter model marked M-shaped velocity profiles were observed due to as the response toright ventricular pressure secondary flows. A pair of helical vortices were input. Step response of this circuit model re- observed in both the LPA and RPA in which the fluid vealed oscillation pattern. Eigenfrequency moved outwardly along the side walls and then circled 4] was increased and AcT [-i(n/2)' AT] was shor- back inwardly towards the center of the vessel. A tened with the decrease in C in the patients with simplified one-dimensional axial flow model is used to CPH. These findings indicate that the decrease in interpret the observed axial flow distributions.Also, a pulmonary artery compliance brings about the in- three-dimensional flow model is used to explain how the crease in w, resulting in the shortened AcT and secondary flows occur and how they affect the flow double peaked pattern in RVEF in CPH. velocity patterns in the two daughter branches.

BE12-M.4 13E12 -M.5 DETECTION OF LARYNGEAL PATHOLOGIES Asthmatic Response to Sulfur Dioxide THROUGH CEPSTRAL/SPECTRAL ANALYSIS A.D.Wolin*, K.R.Lenner, 1.R.Silbert, E.R.McFadden,Jr., and J.M.Fouke Can C. Akkoc*(1), Robert L. Norton(1), Case Western Reserve University, Cleveland, Ohio and William G. Lavelle(2) Cold-air hyperventilation challenges cause a decrease in FEV1 in (1) Worcester Polytechnic Institute asthmatics and direct measures of intrathoracic airway temperatures (2) University of Massachusetts medical School indicate that the airways cool .ppreciably during these challenges. Worcester, Massachusetts, USA Similar challenges with warm-air do not decrease FEY', and the airways do not cool as much. Cromolyn sodium, an anti-asthma drug, also limits Voice samples from normal volunteersand bronchoconstriction during a cold air challenge and causes the airway to patients with laryngeal pathologieswere stay warmer. These studies demonstrate that a change in FEV1 can be analyzed using modernsignalanalysis correlated to a change in temperature in the airways.Hyperventilation techniques. Contributions from the and air temperature have also been explored as parameters which might larynx(signalsource)and the vocal aggravate the effects of SO2 on FEV1 in asthmatics. tract (resonator) were deconvolved using We challenged asthmatics with cold air and 0.4 ppm SO2 in cold air long/shortlifteringinthecepstral and measured RV'.Each mixture was pulled through a heat exchanger, domain to detect various pathologiesof where it was cooled to -10 °C. The subjects expired into a reservoir thelarynx.Long liftered autospectra balloon from which air was being drawn at a constant rate and were providedinformationontheacoustic coached to breathe at a rate sufficient to keep the balloon filled.In characteristicsofthesoundsignal this manner, YE could be controlled throughout each trial.Subjects generatedatthllarynx,whileshort were asked to breathe on this apparatus for four minutes and FEV1 was liftered spectra displayed theformants measured five minutes later. This protocol was repeated at increasing in the vocal levels of YE until a 20% decrease in FEV1 was produced. tract.Analysis ofthe Our results demonstrate that in both challenges FEV1 dropped with cepsrum in combination with long/short increasing levels of YE, but that the decrement was significantly less liftered autospectra provided signifi- when the challenge was performed with SO2. We hypothesize, therefore, cant clues to distinguish betweenthe that the challenge induced intra -airway temperature decrease was less two populations as well as among pathol- with S02.This could occur through either of two mechanism.Since the ogies.Laryngealdisordersdisplayed conversion of sulfur dioxide to sulfuric acid is an exothermic reaction, themselves in the long lifteredauto- a warming of the airways could occur.Alternatively, the generation of spectra, while the formants for patholo- sulfuric acid could cause an increase in pH.Such changes in pH are gical cases were not significantlydif- known to be associated with increases in blood flow, which could also ferent from normals. increase the temperature of the airways. 323

1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas 303 Biomedical Engineering Scientific Papers

BE12-M.6 I3E12-M.7 Magnetopneumographic Evaluation of Effects of Silica on Behavior of Magnetic Particles Applications of wave Spectrum in Rabbit Lungs, Y. Aizawa*, T. Takata, Computer Analysis and Laryn- K. Hashimoto, M. Tominaga, M. Kotani+, gostroboscope observing in K. Chiyotani++, Kitasato University, studying Laryngeal Function. Shao Qingyu; Zhong Ziliang, Kitasato, Sagamihara, Kanage , +Tokyo Denki Pan ian, Huang Xiaoyan, qi University, ++Rosai Hospital for Silicosis, Yan Japan Laryngologic diseases affect laryngeal function therefore vibration Silica is known to give a damage of to pulmonary vocal cords and soundwaves are chang- tissue and give rise to silicosis. This study was ed. Acoustic waves and spectraof 31 carried out to evaluate the effects of silica on normal subjects and 31 patients theclearance of iron particles from the were rabbit plotted by a IRM m....crocomputersystem lungs and on the relaxation. with 12 bit A/D converter, Male analysed rabbits were intratracheally instilled with and comparedwith results of obser- 20 mg of magnetite ( Fe304 )in 1 ml of sterile ving with DDJ-2 laryngostroboscope saline with 30 or 300 mg of silica ( Si02 )in 2 developed by ourselves. t'eatures ml of sterile saline. An anesthetized rabbit of 7 was kinds ofdisease:, (inciJding cancer) magnetizedwith0.5mT of magnetic field for 15 of vocal cords were studied. Further seconds and theremanent magneticfield was problems andmethods were put for- measured with a fluxgate magnetometer. ward. Intratracheal instillations of both high andlow To ensure higher precision ofanaly- doses of silica caused delayed clearance of iron sis ofspectra the FFF with Hanning particles from the rabbit lungs after 7 days of window technique was used. To get sui- exposure. The initial relaxation of rabbits table signals every examinee voice exposed to high dose of silica was disturbed for his hithest "i:" then "a:". Thesam- the first one week after instillation. pling frequency is 13,000HZ. After data processing the number ofreson- ant peaks, noise level (in db) and other parameters can be gotten.

BE12-M.8 BE12-M.9 Lung blood flow and lung gas volume determina- Mathematical modeling of measurement tion during open-circuit breathing, H. Larsson* errors in gas analyzers during exer- and D. Linnarsson, Dept of Medical Engineering, cise, V. H. Carbaial *, E. G. Urbina- and Dept of Baromedicine, Karolinska Institute, Medal and S. Carrasco-Sosa, Departa- S-10401 Stockholm, Sweden mento de Ingenieria Electrica, UAM- Iztapalapa, Apartado Postal 55-534, The pulmonary blood flow and the functionalresidu- 09340 Mexico D. F. . al capacity of the lungs can be determinedconcomi- tantly by the rebreathing of a gas mixturecontain- ing one blood soluble and one blood insolubleindi- In studying cardiorespiratory physiology and evalu cator gas. We have revised this classical methodso ating patients, it has been found very helpful to that tracer amounts of two gasesare dispensed into analize breath-by-breath patterns the inspired gas stream during open circuit breath- of ventilation and ventilatory gas exchange responses to exercise ing. The tracer gases are dispensed ata rate pro- and other physiological stresses. Most of the avail portional to the flow, thereby providinga constant able techniques thatare using gas analyzers to inspired percentage, equal for the two gases. The evaluate gas exchange measurements have underesti- gasuptakes of the two tracers are measured breath mated the transport delay and the dynamic response, by breathby means of a flowmeter-gasanalyzer which in turn introduce a significant departure assembly. Wehave used SF6 as insoluble gas and from the real gas behavior for further analysis. Freon 22 as soluble gas. Both gases can be analyzed We have developeda mathematical model for compen- in tracer concentrations by a simple IR technique. sating the on-line calculation of oxygen consump- Thedifference in uptake between the two tracers tion [S02] and carbon dioxide production [vC0] for during wash-in is proportional to pulmonaryblood 2 both the transport day and dynamic flow. The lung gas volume at the end of an expira- responseof the respective gas analyzers used during exercise. tion (FRC) is determined from the ratio betweenthe The implementation of the model was carried out in amount of SF6 washed out from the lung and the ini- a computer PC compatible system attached tial SF6 concentration before wash-out. to our cardiopulmonary laboratory. We concludethat the Our preliminary results show thatboth pulmonary model compensation should be performed for theac- bloodflow and FRC can be determined during rest and exercise during open-circuit breathing. curate measurement of breath-by-breath gas exchange in order to avoid losing significant information.

304 1988 World Congress on Medical Physicsand Biomedical EngineeringSan Antonio, Texas Biomedical Engineering Scientific Papers

BE12-M.10 BE12-M.11 Volume-Pressure "Curve" of Human Human Pulmonary Venous Return Pulmonary "Venous" System Curve -Effect of Nitroglycerin- Response to Nitroglycerin and K.Gotoh*, Y.Yagi, H.Takatsu, Mental Arithmetic Stress Test- T.Tsukamoto, Y.Terashima, K.Naga- ing K.Gotoh, Y.Ohsumi, Y.Yagi, shima, S.Hirakawa, Gifu Univer- H,Takatsu, T.Tsukamoto, Y.Tera- sity, Tsukasa-machi 40, Gifu, shima, K.Nagashima and Japan S.Hirakawa*, Gifu University, 40 Tsukasa-machi, Gifu, Japan. Cardiac output of left ventricle means an equilibrium point determined by cardiac output curve of left ventricle and Pulmonaryblood volume was estimated pulmonary venous return curve(PVRC).We with RN-angiocardiography, and mean calculated the compliance of human pul- pulmonary artery wedge pressure(PAW) was monary "venous" system(Cpv) and that of measured by Swan-Ganz catheter. When human pulmonary arterial system(Cpa)by ordinate represented volume(V) and our method using RN-angiocardiography and abscissa pressure(P), a short segment of Swan-Ganz catheter. Using these para- volume-pressure(V-P) curve of human meters, mean pulmonary filling pressure pulmonary "venous" system could be drawn (Pmp) was calculated from the equation: by performing passive leg elevation(LE). Pmp= (EVa+EVv)/(Cpa+Cpv) where EVa=Extra The following resultswere obtained: volume of pulmonary arterial system, (1)Nitroglycerin caused the V-P curve of EVv=Extravolume of pulmonary "venous" human pulmonary "venous" system to snift system, in 29 patients with various toward the left and become steeper, cardiac diseases. Next, drawing a part of suggesting pulmonary venodilation(n=50). PVRC from Pmp on abscissa and the (2)Mental arithmeticstress testig(MA) equilibrium point, we examined the caused the V-P curve of human pulmonary response of Pmp and PVRCto sublingual "venous" system to snift toward the administration of nitroglycerin. The right and become near horizontal, following results were obtained: (1) Pmp suggesting pulmonary venoconstriction was significantly decreased with NTG (n=23). (11.7±6.2-8.5±5.5mmHg) (2) PVRC was shiftedhorizontally to the left with NTG.

BE12-M.12 13E12-N.1 Reconstruction of the Alveolar Ridges Numerical mass transfer model in a bioartificial pancreas, with Bioactive Glass Ceramics. D. MORVAN and M. Y. JAFFRIN *. UTC, UA CNRS 858, Wang Shaoan*, Yu Zineng,and Chen Anyu Dep. Genie Biologique, B.P 649, 60206 Compiegne FRANCE College of Stomatology, West China University of Medical Sciences,Cheng- du, Sichuan, China Implantation of microencapsuleted islets of Langerhans in The purposes of this research were to in- peritonealcavityhas beenproposed a3a bioartifimal vestigate the biomedical fundamentals and endocrine pancreas. This numerical study investigates the the clinical results of the alveolar rid- kinetic of glucose and insulin mess transfer between a ge reconstructionswith particulate Bio- microencapsulated islet and an arteriole. active Glass Ceramics. Four beagle dogs were used in animal studies. Three months The mathematical problem Is the evaluation of the glucose and insulinconcentration fieldsinside and after extractions of the mandibular pos- outside the terior teeth of the animals,the materials microcapsule during the time following an augmentation of were placed into the formed subperiosteal the glycem la in the arteriole. tunnels in the ridges. Theanimals were Therefore these concentrations must verify a boundary value sacrificed at different time intervals. problem composed by a set of transient diffusion equations The specimens were examined by light mi- coupled with boundary conditions .The insulin secretion in the croscope,SEM, and EDAX, respectively. The islet is represented as the liberation of insulin granules histological studies showed that the par- stored in two compartments.The rate of Insulin production is ticulate materials were connected with controlled by the local value of glycemia in the islet. The the tissues firmly.EDAX results indicated numericalresolutionisperformedusingtheboudary that the compositions both calcium and integrate equation (BIE) method. phosphate of the new bone at the interfa- The results show that the insulin secretion is amplified when ces between the materials and the alveolar bone were equivalent to those of the the diameter capsule decreases.In all cases the time lag original bone.The clinical results demon- between the rises of glucose and insulin levels in the blood strated that the bioactive glass ceramics does not exceed 10 minutes which is enough to normalize the material is good for reconstruction of the glycemia in diabetics. artrophic alveolar ridges. 325 1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas 305 Biomedical Engineering Scientific Papers

BE12-N.2 BE12 -N.3 Technical Determination of Permeation Oxygen Transfer of Liquid-Liquid Oxygenator Resistance in Blood Treatment by Membrane H. Suzuki*, T. Tsuji, K. Tanishita, T. Tamura Distillation, T. Koyano*, M. Tamura and K. Sakai, Waseda University: Ohkubo and T. Togawa, Fac. of Sci. & Tech., KEIO Univ. Shinjuku-ku Tokyo 160 JAPAN. Kanagawa, JAPAN, Tokyo Med. & Dent. Univ., Tokyo, JAPAN TWO fluids of different temperaturesare caused to flow on either side of a microporoushydrophobic membrane in membrane distillation, andwater vapor The use of hemofiltration moduleas a liquid- is transported from thewarm to the cool side of liquid oxygenator applied for the longterm the membrane. In this work, effectsof polarization perfusion is promising because of its excellent layers formed on either side of themembrane on and reliable blood compatible fzztures. In the separation of solution components frombovine blood presentstudytheoxygen transfer were clarified. performance of hemofiltrationmodule was Distilled water, bovine plasma andbovine blood evaluated by in vitrowater and blood test in were pAmped into a module (surface area: 0.3m2) order to reveal the feasibility of liquid-liquid composed of polypropyrene hollowfibers. Water oxygenator. The hemofiltration vapor permeability was measured at membrane warm tested washollowfiber of microporous co- temperatures at inlet ranging from 307to 316K, polymer ofethylenevinyl cool temperatures at inlet from 288to 293K and alcohol fl as rates from 1.67 EVAL, KURARAY Co. Ltd.). Eoth to 3.33x10-6m3s-1. studiesshowedthat Values for permeation resistance of themembrane the oxygentransfer were 1.6x1010, 1.3x1011 and 2.1x1011m2sPam-3 for occuredsimply inproportiontothe distilled water, bovine plasma andbovine blood, concentration gradient of oxygenacross the respectively. Higher valueswere obtained for membraneand oxygen uptake by thebulk bovine plasma and bovine bloodbecause of a flow acrossthe membrane wasnegligibly lowering of vapor pressure and adhesionof serum small. This signfied that the oxygen transfer proteins on the membrane. Valuesfor permeation performancein hemofiltration module could resistance of warm-side polarizationlayer were be evaluated similarly with that of gas-liquid 1.4x1010, 4.0x1011, 8.9x1011m2sPam-3for distilled membrane oxygenator. The oxygen transfer water, bovine plasma and bovine blood ata flow rate of the hemofiltration module turned out rate of5x10-6mis-1,respectively,and increased to be quite small (<15 ml/m2min) due to the with the viscosity of test solution. extremely large membrane resistance.

BE12-N.4 BE12-N.5 STUDY OF THE INFLUENCE OF THE SHEAR STRESS ON THE ADSORPTION KINETICS OF ALBUMIN ON A PLATINUM Hemoperfusion Treatment in Pigs Experimen- ROTATING DISK ELECTRODE. tally Intoxicated by Paraquat, T.S. Yang P. Bernabeu, A. Caprani, P. Pdronneau* (Dept. Genetics and Physiology, Pig Rese; arch Instituter, Y.L. Chang and C.J.Lee INSERM U. 256, Hbpital Broussais, Paris, France. Dept. of Chemical Engineering, National Tsing Hua University, Hsinchu, Taiwan, In order to better understand the interactions R.O.C. of proteins withartificialmaterialsin shear Crossbred gilts (Landracex Yorkshire), stress conditions, the adsorption of albumin weighing around 70 kg were experimentally ona platinum rotating disk electrode has been poisoned with the herbicide paraquat.When studied. Using electrical impedance measurements, animals were intoxicated by intramuscular it has been possible to record the kinetics injection of the herbicide at levelsof 74 of the adsorptionthroughthe changes of the mg/kg, 17 mg/kg and 6 mg/kg, theparaquat electrochemical double layer capacitance, within content in the serum, in all threecases, the range0-1700 s-1. This kinetics couldbe rose sharply, reaching a peak around 1.5 h modeled in two consecutive reactions taking after injection and then declining. place at the interface,this providing 5 para- Hemoperfusion was conducted for 2 hon 6 meters : twotimeconstants, Tiand T2,and gilts beginning 2 h after gasticloading of three capacitances, Cd, Cdi, Cd2, of the surface 70 mg/kg paraqua . The mean turnover rate naked, -overed with the reversibly adsorbed of circulating pi. aquat was 28.1%. Hemo- pro'.e i and covered with the irreversibly perfusion treatment for 2 h couldremove adsorbedproteins, respectively. The influence about 5% of the toxin loaded intothe sto- oftheshear rate on each of these parameters mach. Prolonged treatment for 6 h didnot hasthenbeen studied. It hasrevealedthat, enhance the turnover rate of circulating although the shear rate variation hasno signifi- paraquat. Fuller's Earth and Na2SO4 were cant effect on the three capacitances,it does given to another 2 orally poisonedgilts 30 influence thetwo timeconstants, evenafter min prior to 6 hr hemoperfusion andthey correction for themass transfer contribution. survived. Pathological examination revealed Thus, Ti monotonouslydecreases as the shear capillary congestion in lung, heart,liver rateincreases. Concerning T2, thevariations and infectional mucosa. versus the shear rate are more complex and are Use of hemoperfusion in treating paraquat calling for additional experiments. poisoning might offer protectionto lungs from9 excessive accumulation of the toxin.

306 1988 World Congress on Medical Physicsand Biomedical EngineeringSan Antonio, Texas Biomedical Engineering Scientific Papers

BE12-N.6 BE12-N.7 Effect of control parameters on exchanged Applications of a Physicochemical Model of the substances during hrodialysis - A simulation Fixed Caa rier. Hugo A. Massaldi:Instituto de analysis, 0. Thewsand H.Hut ter ,Johannes Investigaciones Cardiologicas, Universidad de Gutenberg Universita Mainz, D-6500 Mainz, West - Buenos Aires, Argentina. Germany. Different variants of the mobile and fixed carrier hypo- A mathematical model is presented that describe: thesis are traditionally used in the biological litera- the exchange kinetics of urea,creatinir.t., ture to explain the facilitated transport of substances vitamin 812, sodium, potassium, acetate, oxygen, across cell membranes. Of these, the fixed carrier is carbon dioxide, bicarbonate, hydrogen ions and getting generalized acceptance as a more feasible trans- water between the dialysate and different body port device. A recently developed physicochemical model ccmpartments during hemcdialysis. The model has (J. theor. Biol. 106537; 110:35,19841 of mediated been employed for simulating the effect of transport provides a contemporary view of the fixed- control parameters (a) transmembrane pressure, carrier mechanism. The model, which postulates migration (b) bloodf low through the dialyzer, and (c) the of polar solutes through interstitial channels of intrin- dialysate sodium concentration. sic proteins, is shown to lead to the classical equation Results: (a) The transmembrane pressure controls of Wilbrandt and Rosenberg for the transport of hexoses the amount of total body water as well as the as a particular case. Derivation of other expressions, watts distribution between plasma and such as that for saturable transport with interactive interstitial space, furthermore the sodium sites for ion exchange, is also possible through this kinetics, howcver not the potassium kinetics. model. (b) Blood flow through the dialyzer influences A distinct feature of the present approach is that it the acid-base-state,but not the water provides the spatial character and displacement pro- distribution.(c) The dialysate sodium perties of the species, which are not present in the concentration effects mainly the water purely enzymatic treatment of the problem. Thus, the distribution between extra- and intracellular conventional parameters TmAx and KT for hexose trans- space, the potassium kinetics and the acid-base port are defined in terms of a multicomponent permeabil- state. The oxygen pressure is independent of all ity coefficient, the effective cross area of protein three control parameters, whereas oxygen channels and the sorption constants of the solute onto saturation of arterial bloodis controlled by the protein. By correctly accounting for the initial all threeparameters. conditions, the model is capable of describing the active transport of ions as well.

BE12-N.8 BE12-0.1 Rheotogical hope:lie% of Gallbladder Bile Detection o Surat Cancet by Meameang A Challenging Problem in Pathophysialogy R.F.Yang:X.Y.Weia.P.Wu Amery:. Stood Ftow. M. Nitzanl Y. Mahler Institute of Bioengineering and Z. Gimmon,Jerusalem College of Technology, Chongqing Bniversity,P.R.China Jerusalem 91160, Israel

The theological properties of gallbladder bile are itosely related to the biliary secretion,flow,and The higher metabolism of a malignant tumor demands, etiology of gallstone and other diseases in the generally, higher blood supply, which in the par - biliary system.The relationship between the rheo ticular case of breast cancer, can often be de- lagical properties and the chemiral components of tected by skin temperature measurements, using gallbladder bile is very important also. however, thermography. However, the small pathological those subjects have not been investigeted so far. In present paper,behavraur of uatIbtadder bite temperature differences are not clearly distinct is discussed. Various pathaingicel specimens of from normal temperature fluctuations resulting gallbladder bile,included IL acr.te,chronic the from various physical and physiological reasons. lecystitis and chatelithiasis,are collected and The non-invasive transient thermal clearance measured by using Low Shear 30 rheometer under the shear rale range of U.01607 118.2s-I. The method was used for the detection of breast can- experimental results have showed that their rheo cer. It differs from thermography since it mei- logical behaviour is like nun newtanian fluids sures directly skin blood flow instead of mea- when shear rate is less then ISs ',and like new suring its temperature, As breast carcinoma de- Ionian fluids when shear rate is more then IS 20s i. Their viscosities are remarkably different among velops in the milk ducts, which are directed todifferent pathological specimens. wards the nipple, the blood flow measurements were B' measuring the vi.caelasticity of gallbladder performed in the areola. bilt,Its storage modulus simply increases, its The measurements have shown clear correlation to modulus first increases and then decreases, with the angle frequency w increasing. Besides, between breast cancer and increase in a:eolar the positive correlation exists between the apps blood flow, especially in the left breast. The rent viscosity at same shear rate in high shear areolar blood flow in n cancerous breast was high- rate range and the contents of cholesterol C, and er than that in the contralateral breast, as well mucin C2.Regression equations are respectively q 4.35910.0268C, regression coefficient rl 0.9304, as higher than the average blood flow in normal andq-2.5048+0.0078C regression coefficient breasts. (1.9304. According ta those results,we can quanti tatively investigate pathaphysiology in the bi tiary system. 327

1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas 307 Biomedical Engineering Scientific Papers

DE120.2 BE12-0.3 Permeability Measurements of Mouse Embryo Temperture prediction during Membranes at Temperatures Above andBelow hypertherm_ zteezing, Douglas B. Walcerz*, Shanti , Min-he Cheng & J. Hong-Zhang Wang*, Precision Aggarwal, and Kenneth R. Diller,University of Texas at Austin, Austin, Instrument Dept. Shanghai Jiao Texas, 78752. Tong Univ. Shanghai P.R.China The ability to preserve mammalian cells by freezing Hyperthermia by ultrasonic, depends greatly on knowledge of the flows ofwater as a method for cancer treatment, hasbeen applied to and cryoprotectant into andout of the cells. For clinic trial Tor this reason, measuring the many years. But a non- permeability of cell invasive method of monitoring membranes, especially attemperatures below temperature distribution during heatingstill remains freezing, has become an importantmatter of unsolved so far. investigation. A cryostage has been designed to In this paper, a simplemodel structure measure the permeability of cells towater and of temperature distributionin human cryoprotectant at temperatures both aboveand tissue, which is based below freezing. In addition, the stage allows on one-dimension an bio-heat equation, isdeveloped. The entire cryopreservation procedureto be performed model parameters for realtissue are esti- while continuously observingthe cells under the mated from patient treatmentthrough the microscope. use of a system identification The permeability of technique. mouse embryo cell membranes to The temperature distribution water and glycerol has been measuredat during hyperthermia can be predictedby this model, temperatures from -30C to 35C. The survival of and non-invasive temperaturemeasurement cryopreserved mouse embryos at freezingrates from can be realized. .1C/min to 5C/min has a'sobeen measured. Graphs showing membrane permeability Identification of phantom modelparameter vs. temperature, indicates that the simple embryo survival vs. freezingrate, and embryo control model structure is adequate. Anidentification volume vs. time duringcryopreservation will be presented. measurement divice is made bymicro-com- The design of the cryostage willalso be presented. puter TP-801 and better resultsare obtained by experiment heatingfresh excised pig tissue.

BE12-0.4 BE12-0.5 HEATING TUMORSWITH EXTERNAL MICROWAVES: THE ANALYSIS OF RISING TEMPERATURE 1D CURVE MEASURED SIMULATION, Thomas R. Canfield, BY A DEEP BODY THERMOMETER, Ph.D.*, F.L. Lan* and J.W. Loyola University of Chicago, Huo, China Medical University,Shenyang, P.R.C. Stritch School of Medicine,Maywood, IL. The Deep Body Thermometer (DBT)based on zero A simple finite elementmodel using the heat-flow principle was firstdeveloped by R.H. Fox, 1970. transient one-dimensionalbio-heat trans- Now, the DBT can measure thetempera- fer equation was developedto investigate ture1 cm deep under body skinwith an error of 0.1-0.2 c. heating tumors withan external waveguide Nevertheless, a much longer periodis needed for a complete measuring applicators. It has been used to model (usually 20 minutes or more), and until heating of the canine larynx withexter- now, all the data treas- nal waveguide applicators. In these cal- ured were abandoned exceptthe steady value. culations the applicator was modelled by Actually, the shape of risingtemperature curve assuming amplitude of is related with the temperaturedistribution un- an incoming 915 Mhz der body skin. planewave which was attenuated expo- in this paper, an integralequa- nentially by the tissue. tion was found from a simplifiedmodel that re- The appli- lates the rising temperature cator was be controlled bya temperature curve and the Ini- tial temperature distribution at one of the nodes within the tissue. under the skin. If Temperatures distributions were computed the initial temperature islinetrly distributed, it was proved that the rising at every time step to simulatean actual temperature curve could be expressed as an infinite treatment. The results indicate that series of ex- power deposition does not polential functions. On this basis, an approxi- translate mate solution method was obtained that directly into heating. Other important can esti- factors are power deposition rate, blood mate the steady value in advance. Experiment shows that the method can estimate the steady flow, surface boundary conditions, variation of tissue properties value with an error less than 0.2c by using the with data measured in 4 minutes and 0.1 depth, the presence of aircavities. c by using the data measured in 6 minutes.By prolonging the measuring period, the estimated value Ismonoton- ously increased, and tends to the actualsteady value. Moreover, the inverse problem (find the Initial temperature distribution fromthe data measured) is under study. 328 308 1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas Biomedical Engineering Scientific Papers

BE12-0.6 8E12-0.7 Unconstrained Measurement of Tempera lure and Its Modeling the Thermoregulatory System Distribution of Decubitus Ulcer Patients Using Thermal Entrainment Technique and a Kernel Method. Dr. Ahmed K. Ahmed. T. Nakano' (1), T. Tamura. T. Fuj imoto, T. Tsuj iT. Togawa(2), Faculty of Technological Studies,Kuwait (1)South Jacan Health Science Canter, (2)Inst. for Had. The system considered here is the process & Dan. Engng..Tokyo Mad. & Dan. Univ.,2-3-10, relating the thermal stimulus (input) to heart rate (output) of a man. This paper kandasurugadai ,Chi yodaku. Toicyo, Japan 10i describes Electronic Equipment developed to control air blower which produces ther mal stimulus of hot and cold air at diffe- Eacubitus ulcer is the ulceration of the skin caused by a circ- rrent frequencies to serve for the input ulatory disturbance during long ...7.t.iselaant to bed.Using sheet to the thermoregulatory system. It also describes experimental prr..:edure and iden- wIta:ning IC matrix sensors, the surface temperature in 15are- tification procedure of tlic system model as of the decubitus ulcer patient were measured without pasting using an algorithm of Kernel identification of voltterra series which greatly sensors directly upon the lkin.This data was continuously reco- reduces the computational burden and eli- rded in EPROM every 5 minutes for 10 hours.After data processi- minates the restriction of using white Gaussian input of a test signal. The cal- ng of the signal.the temperature and its distribution were ill- culation, estimation of voltterra Kernels ustrated automatically through graphs. The r4.orded distribution and formulation of the model are presented.

of the temperatures showed rare posture changes in the patients A second order model is the most appropriate for a good estimate of the system as compared ta 'taw of the norm! control. It was observed that dynamics. The model contains the linear the temperatures of the area of the decubitus ul..er part (first order Kernels) and Quadratic were mean part( second order Kernels) in parallel. 35.5C and 3'C higher than those of the normal area.lt:rewas

no difference 11 ulcer temperibire luring sleet and non- Other results and discussion will be presented. It seems that the model may sleep periods.This device is useful for prediction and prevent- give physilogical interpretation of the ion of the decubitus ulcer. dynamics of the thermoregulatory system.

BE12-0.8 BE12-P.1 Experimental Study of Power A Liposome-Based Immunosensor: Detection of Deposition and Power Reflection Tumor Markers,M.Aizawa*,S.Shimizu,H.Matsumura, for Microwave Hyperthermia H.Yokoyama and H.Kamei,Electrotechnical Antennas, Matthew Pressly* and Laboratory,1-1-4,Umezono,Tsukuba 305,Japan. Dr. J.A. Pearce, The University of Texas at Austin, Austin, Texas We have developed a new type of biosensing systems where the dynamic light scattering tech- In microwave tumor hyperthermia, the nique is applied to observe agglutination of heating is induced by an antenna liposomes caused by an antigen-antibody reaction which radiates microwave energy into on their membranes. The liposome is a small the tissue. The pattern of heating spherical body (d=20-100 nm) composed of phos- induced by different applicators has pholipid bilayers which construct a basic struc- been studied on a tissue phantom in ture in biomenbranes. By the use of our system, order to predict applicator we tried to detect one of tumor markers, pan- performance on human tissue. creatic oncofetal antigen (POA), in human sera. Minimization of power reflection has We prepared a suspension of liposomes having also been studied. polyclonalanti-POA molecules on their membranes. When a test solution with POA molecules The phantom material has electrical was added into the suspension, agglutination of properties similar to muscle tissue. the liposomes occurred through the POA antigen- Heating patterns are analyzed using antibody reaction. Fromthe dynamic light scat- thermographic imaging. The tering measurement, we found that there is a frequency-dependent reflection linear relation between the size of agglutina- coefficient has been determined for tion products and the POA concentration. On the loaded antennas using a basis of the relationship, we could estimate the Hewlett-Packard 8510B Network POA concentration in sera with a small sample Analyzer. Antennas studied include: volume (1 pl) and in a short period (20 s). Spirals, slots, bowties, resonant The minimum detectable concentration in this ring, and parasitic ring. sensing system is ca. 200 nanomol/ml under our ordinary experimental conditions. 32 1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas 309 Biomedical Engineering ScientificPapers

8E124.2 Nonoarametric Testing of Plasma insulin BE12-P.3 CharactenStics es Potential Marl Reference Regan for Oython Parametersof efsop Plasma Cortisol in aiierm of Short ftreast ati7Mr Rist ORAR.C. Hermida* and Stature?. R.C. Hermida*, L. Garcia and F. Heiberg, Uni- C. Hernandez, University of Santiago,Spain. versities of Santiago, Spain & Minnesota,USA. Among 12 plasma hormones from healthy adult 21 plasma cortisol profiles, women sampled around the clock and the sampled from Bavarian year, children every 4 hrs for 24 hrs pattern discrimination methods single across all seasons, out plasma served for the computation of insulin as a primary classifier for a so-called paradesm, a high (H) vs. low a 90% prediction region for the emplitude(A)- (L) BCR. The relation betweenplasma insulin and acropht.:03) pair. Cortisol data BCR is an indirect one in all 4seasons. These data from-32 Spanish children (20 boys and 12 girls) aged were re-analyzed by bootstrapping (BS) of circadian 6-14 yrs, of short stature, sampled at about characteristics, assuming that deviationsof each 3-hr intervals for 24 hrs across all seasons,were also analyzed by sample from 2 means. those ofthe BCR-group and linear harmonic analysis. The best the individual, are correlated alongthe 24-hr scale fitting periodis 24hrs for either sex and allsubjects. The circadian in keeping with a first orderauto-regressive model. parameters were interpreted by Results from BS validate differencesin circadian referenceto the Bavarian paradesm. Despitedifferences rhythm-adjusted mean for insulin inspring and in age, ethnicity, climate, summer between the I & H BCR groups Circadian BS diet, health and sampling conditions between bothpopulations, isa resampling procedurehere suggestedto individual estimates of the circadianAand 0 were inside the introduce inferential considerations into Bavarian paradesm for 19 boysand 9 giris.Cortisol discriminant analysis and recommendedfor broad markerrhythmometry basedon routine use in nonparametric estimation, relatively few samplesconstitutes apractical hypothesis testing and biomedical signalsimulation approach for parameter estimation of the adrenalcortisol cycle.

BE12-P.4 BE12-P.5 Throngolysis inMembrane-BoundThrombi Using Liposomal- TheRelationship Encapsulated Streptokinase, PD Nguyen..:, EA O'Rear:, Between Composition of Tumor BM Fungi, R Cells and Its Damaged: Liu AE Johnson*. :Dept. of Chemical Engineering, :Dept. of Yongge,Wang Junjiano, Chemis- Quan Lin, try, University of Oklahoma, Norman, OK 73019, USA. Dept. of Bioengineering,Huazhong Univ. of Sci. and Tech., Wuhan,Hubei, P.R.C. Experiments are designed Application of streptokinase (SK) in thrombolytictreatment has to determinehow the cell been hampered by the uncertainty of optimaldose, arising from structure isdamaged by shear stresses in Couette rheometer. immune- inactivation, complications resulting fromunderuse of We find that viabilityof EAT (Ehrlich ascites SK, and above all, fear of bleedin3. Asa result, patients with tumor) cellsat about 41°c and S180 (sarcama history of severe hypertension, recent strokes, cerebralhemor- 180) at about 430 abruptly decreased rhage, and cerebrovascular diseases are excluded fromthis kind by shear stresses 25 dynes/cm, compared with theslow and of treatment. We are currently investigatinga new delivery sys- linear loss of viabilitywith the rise of temperature tem for the transport of SK to tzrget area by theuse of lipo- from 30'c to 40°c. This somes. abrupt change is consistentwith marked cell change of The ability of liposomes to encapsulate drugs, membranemicroviscosity which is caused by proteins, and bilayer systems nucleic acids have stimulated interest intheir use as drug car- of membraneundergoing a thermo- tropic phase riers.In our study, high concentrations of SK (10,000IU/mL) change(phase transition) from the were used for encapsulation in large unilamellar vesiclesof liquid-crystalline to theliquid with therise of phosphatidylcholine. These vesicles have temperature. a homogeneous size of So the sharp change of less than 0.21unin diameter. During slow flow through viability is due to phase transition of cell membrane. membrane-bound thrombi, liposomal encapsulated This indicates that SK retained the cell membrane fluidity the ability to activate the fibrinolytic is vulnerable to shear system, presumably by lys- stresses. The larger the membrane fluidity of, ing of vesicles from shear forces. Bymonitoring the pressure the easier drops across the Nuclepore membranes and damaged the cell is.The viability of clot-dissolving time, tumor oell the preliminary treated with colchicineis raised, results show thatthecapabilityof SK- which suggests encapsulated liposomes to dissolve clots that thedeformability is limited are comparable to that by cell miorotubules. On of an equivalent amount of unencapsulatedSK in platelet-poor accountofthe high nuolear/cytoplasmio ration in many cancer cells, plasma. We .believe that. the therapeutic-efficacy would be mechanicalstresses at the cell surfaceare improved if SK or other plasminogen activatorswere encapsu- transmitted to the nuclearmembrane and contents lated in a medium such as liposomes and onlyreleased when it and even nuclear membranes are damaged while sur- has reached the blood clots, and that hemorrhagicphenomena face membranes are not in somecells. Not like might be greatly alleviated by the the special structure relative dampening of tumor cellcannot protect diffusive to convective transport. from shear kill.

310 1988 World Congress on MedicalPhysi3 csIfIliomedicalEngineeringSan Antonio, Texas Biomedical Engineering Scientific Papers

BE12-P.6 BE12-P.7 THE BIORHEOLOGICAL PROPERTIES OF HUMAN NORMAL AND Iontophoresis with insulin in dogs. PATHOLOGICAL GALLBLADDERS - G.H. Zhong*, Y.P. Wu, J.Bustelo: F.del Pozo, M.J.Medialdea and F. G.R. Wong, West China University of Medical Sci- Vilatilana.Fundacion C.G-M.Instituto de Bio- ences, Chengdu, Sichuan, P.R.C. ingenieria.Avd. dela Marina,3-5.11007 CADIZ ESPARA. According to the "thin shell theory" and some sup- positions, a two-dimensional quasi-elastic strain The objetive of this investigation is the - energy function of gallbladders based on the bii- insulin administration using tradition... - ary histology was established. The relationships techniques of iontophoresis as an alternati between the stress and strain of the gallbladders ve to the traditional treatements of Diabe- were obtained. A new non-contact system which was tes Mellitus . composed of VDA (Video-Dimensi'n-Analyzer), Pres- The differents experiments were:the possibi sure transducer and other transmissions was used lity of introduce insulin through the intac to measure the two-dimensional deformation of gall- t skin without it losing its biological and bladders. A non-linear multi-dimensional curve to determine the best parameters that per-- fitting method was applied to fit the measured re- mit the elaboration of the principles of a sults. We found that the biorheological properties non-invasive insulin pump. of gallbladders were related to their histology A study was carried out on a 50 dogs were - structures. Furthermore, we found that the strain made "in vivo", divided into four groups. - energy function and its parameters can be used to Were determinated the values of insulinemia evaluate some biliary diseases such as cholecyto- and glycemia after the insulin application pathy clinically. by means of iontophoresis during thirty minutes.

BE12P.8 BE12-P.9 A Study on 1 New Adsorptive Antidote for Artificial An Experimental Approach to the Hemocompa- Liver. Xu,Guo-Yang; Nu,Tng; Pong,Noi -Chez; PhouHan; tibility and Penn ability of Activated Char- Heo,Shu -Chu; eto., Research Institute of Biomedical coal Coated with Silastic and Its Copolymers, Engineering, Jinan University, Guangzhou, People's Liu You-chu,' Dong Guang-jiao, and Wang Xiu- Republic of China yu, Dept. of Chemistry, The Fourth Military Medical College, Xian, Shaanxi, P.R.C. This text reports a result of a staged study on a new adsorptive The methyl vinyl silastic (PVMS) and its copoly- antidote for artificial liver. This adsorptive antidote was pre- mers, PVMS-MA and PVMS-AM coated activated charcoal pared from a natural chitin raw material by wimple chemical and (AC) that prepared by 6000 radiation methods. Stu- physical treatment. It has a good biological compatibility and dies on hemocompatibilityof these coated materials have been carried out by SEM and pseudocoloring of a relatively strong adsorbability to bilirubin, endotoxin, hate - a grating encoded phase-picture method. The resu- roaromatio amino goings. lts of our in vitro studies show that the permeabi- A static test showed that 1 g of this antidote could decrease lity of silastic and its copolymer- coated charcoal the bilirubin in a 10 ml plasma with a concentration of 50 units for librium is very high. The detoxication efficacy to 20-25 units within 2 hour* and the amount adsorbed was 0.25- of the coating appeared PVMS-MA- AC>PVMS-AM-AC>PVMS 0.30 mg/g. As for endotoxin, it was decreased from 54 PG /al to -AC. The platelets on PVMS-MA copolymer were slig- 4-5 PG /al and tbe amount adsorbed vu 500-510 PG/g. The amount htly deformed. The number of adhered platelets tcx_ adsorbed of same heteroaromatic amino acids In complex amino the copolymer coating was 28% of the platelets ad- acids were as follows Phenylalanine 1.52 mg/g; Tyrosine 1.20 mg/g hered to PVMS and was only 8% of the platlets adher- Histidine 1.31 re/g. Fulling 70 g of this adsorbent into an arti- ed to AC. The results suggest that the hemocompatibi- ficial liver device, we performed a blood perfusion for 2 hours lity of the PVMS-MA copolymer induced by radiation is to experimental dogs of icterus model weighing 11-14 Kg. The better than silastic and that the copolymer coatings contents of bilizubln and glutamic pyruvic transaminase in the are likely to be used hopefully for detoxication of blood before and after perfusion were decreased respectively by artificial kidney or artificial liver. Abbreviation 53-63 and 42-4e ti The experimental dogs recovered normal ac- AM acrylamide; MA methacrylic acid. tivities and appotive soon and survived after perfusion. An experimental female dog even mated and got pregnant after perfusion. Nevertheless, experimental dogs of the same model without perfusion treatment would die within al days. 331.

1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas 311 Biomedical Engineering ScientificPapers

BE12-P.10 BE12-P.11. Low intensity constant directcurrent effect on wound healing, N.Atalay,* An experimental system forquantitative G.Aydin, S.Tokta§, Medical determination of mechanicalparameters of Fac. of Dicle Univ. Biophysics Dept. human hair. G. Nikiforidis*, D. Diyarbakir,TURKEY Tsambaos, A. Bezerianos, P. ProkopiouUniversity of The effect of low intensityconstant directcurrent Patras, 26001,Patra, Greece. on soft ti.-ue wound healing in rabbitswas studied and the healing rate of electricallystimulated) wounds A quantitativeindicator of the general status of were compared with the controls. SecondlyHydroxy- the human hair can be givenby the exact prolinecontent of granulating woundwas analyzed determinationof theprincipal parameters in electrically stimulated andcontrol groups. characterizingitsmechanical behavior.In the Current was obtained from present study an experimental procedure,which a specially designed low determines the stress-strain intensity direct currentgenerator which keeps current curve of human hair under different strain rates isanalyzed. In this constant by regulating the voltage.100/4A- 150friA paper the direct current was appliedto the wounds and max. followingisdescribed:))The output voltage was limited to 6V. mechanicalcomponentsandthesystemof Electrodes used transducers 2) The electroniccircuits by which were stainless-steel mesh ones and controlwounds the strain is applied and the were covered by dummy pads withno electrode stress as a function it.One day of in of time is measured 3) The mathematical negativecurrentwas appliedand algorithm and the _program of processcontinuedwithpositivecurrentuntilthe calculus for the healing process was completed. determination ofYoung modulus, yieldstrain, Currentapplication ultimate strain and relaxation procedurewas 8-10 hours/day and6days/week. modulus of human Hydroxyproline content of both of hair. In this framework followinga standardized the groups were protocolthestatistical determinedat7thand12nddaysbyWoessner's correlationofthe method. mechanicalbehaviorofhumanhairtothe findingsofmorphological and biochemical It was found that the healingwas accelerated with analyses as well as the clinicalexaminations, is constant direct current application and possible.Consequently quantitative content hydroxyproline determinations of electricallystimulatedwoundswere of the influence of differentfactors such as U-V higher than the control ones. radiation and pharmaceuticaltreatment, on the condition of human haircan be performed.

BE12-P.12 L.12 -P.13 A r i a i a 515 f i711176.617,M / 7 4 : i 7 O O i O l it i 5 .5 we AN APPROACH FOR SHORTENINGMEASURING TIME OF HUMAN ifiRdM hl CNA:1/W Of 51;01: L BODY TEMPERATURE, X.L. Zheng,C.L. Peng and E.X. cia*, R Heemida.T Iglesias arid. C. Ludeico. Zheng*, Chongqing University, Si_Juan,P.R.C. Hosp. fiaterno-Infontil. La Coruna & Lifliversitt: The rapid measurement of human bodytemperature is ofSaistirariO.SantiatiO (SP61A) necessary for the clinic and daily life. This paper submits the electrical model of bodytemp- insulin ,...icentrations (in 1111.1/rril)were determined erature measuring system. bu radioiromunoassay in plasma obtained at "3-hr UAder certain conditions, the modalcan be simpli- intervals for 24 hours from 11 boys and 7 girlsof fied as an R-C circuit of secondorder. The pro- short stature, aged 6-14 yrs. Circadian rhytnrocora- cess of body temperature measurement is determined by the thermal resistance and capacitance meters*were computed bu the single of human and population body and also of the sensor. According to the mean cosinor fits of a 24-hr cosine curve model, discussions have been hadfor the shorten- GenderP 11E91:A1i-se Arnplitude±se AcrophaseL±se ing of measuring time of body temperature. We considered here how to decrease the effectof ini- Boys.005303+3 0 12.7 ±2.5 -299*± 9" tial temperature of the sensor. We present the Girls05234.1+4 0 17.0 ±5.0 -251°1:13' preheating method to avoid the influenceof envi- 411 <.00132.1+2.2 13 1+2.0 -277"± 10" ronmental temperature. i Midline estimating statistic of rhythm. On the basis of preheating method,a prototype of z From midnight, with 360°=24 hours. the fast-response electronic clinicalthermometer has been constructed. Evaluation of risk by appropiately designedques- The measuring time is less than 30 seconds and the accuracy is tionnaires should next complement blood better than sampling 4-0.1°C for oral and rectaltemperature measure- forhormonaldetermination andpropersignal ments. This shows potential advantagesin clini- processing and parameter estimation in groups of cal uses. children differing in gender, ageor ethnicity, among other factors. ern

312 1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas Biomedical Engineering Scientific Papers

BE12-P.14 BE12-P.15 AN AMBULATORY SCREENING DEVICE FOR THE DETECTION OF Automated Plastic Forming System for SLEEP RELATED BREATHING DISORDERS Artificial Organ Components with a t. Penzel: G. Amend, K. Meinzer, J.H. Peter Robot, H. MATSUDAIRA*, T. TOGAWA & K. Med. Poliklinik of Philipps-Universitat Marburg TSUCHIYA, Dept. of Mech. Engng., Sch. of Sci. & Engng., Waseda Univ. Sleep-related breathing disorders (SRBD) are of high Tokyo Med. & Dent. Univ., Tokyo, JAPAN. prevalence (10%) in a male population with an age between 40 and 60 years. To diagnose these disorders extensive investigations in a sleep labora- A new plastic forming system was developed. tory are necessary. Therefore the need for an easy The principle of this system is piling up cross- to handle screening device is growing. sectional layers to form the entire figure. Each Different studies in our sleep laboratories showed layer was fabricated by extruding forming that the time course of heart rate and a frequency material intermittently in a series of locations analysis of nocturnal breathing sounds reflect using a plastic dispenser attached to a robot the structure of sleep very well. Physiological and which was controlled by a computer. As the pathological patterns can be differentiated. material, a room-temperature curing and an In colaboration with a medical electronics firm we ultra-violet curing siliconerubber which succeeded in the development of a small portable viscosities are roughly 800-1,000 Pa were used. device based on the analysis of heart rate and Inthe preliminary experiment,simple nocturnal breathing sounds. The device records the shapes such as straight tubes, bifurcating heart rate beat-to-beat with a resolution of 1s tubes, tubes having leaflets in the inside and and 6 bit in a solid state memory. Two further spiral tubes were formed. Errors in forming bit/s record loudness of breathing sounds and the process remained within 0.5 mm. The result normalized power in the frequency range between suggests a possibility of applying this method 50 cps and 800 cps which is typical for snoring. A to form plastic models which are difficult to be recording up to 18 h permits a comparison of night- formed by the casting or machining. time and day-time pattern. Data are read out via a serial interface and can be analyzed in a PC. Pattern analysis on heart rate and time intervall analysis on breathing sounds allow efficiently to differentiate normal sleep, sleep apnea and other indications for further sleep lab examinations.

BE12-P.16 BE12-P.17 The Computer Simulation for Ultra- Mapping of Pain by Electrical sonic Point Objective Image, Zhou Impedance Measurements, Couros Ghaznavi*, Roman Chrucky, University Guang-1'u* Yi di -Ping, ChengduInst. of Bridgeport, Bridgeport, CT 06601 of Radio Rng., Chengdu, PR China. Our study attempts to objectively map by means of dermic impedance measurements, the painful To simulate a B -mode ultrasonic image on site. The hypothesis on which our study rests is that in a healthy individual, dermic impedance computer -has been considered of great measurements would re'eal a symmetry of elec- significance for the obtainment of accu- trical patterns. Pain alters this symmetry. Changes in circulation take place in the painful racy and optimization in a practical de- site and this leads to a change in impedance sign work for B-mode ultrasounic system. values. The study undertaken here is to probe however, the reliability of simulation the reasons and to find an effective method of localizing painful areas. An instrument ata results depends on the analysis of a frequency of 200 KHz has been designed to locate the low impedance points on the skin. sound field. The sound field distribut- The impedance between two specific points of ion is used to simulate the practical the body is a function of the frequency of the image in a linear array B-mode ultrason- applied voltage. At high frequencies, electrical current effectively short circuits the entire ic field. cell by penetrating the high capacitive cell The image function and the pointrespon- wall. Thus, the impedance of a cell shows only incremental changes beyond a critical frequency. se function have been successfully de- This frequency is approximately 200 KHz for the rived and they are all dealt with in the skin. Consequently its assumed that the disymetry of the dermic impedance noticed is frequency domain. due.to changesin circulation.

333

1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas 313 Biomedical Engineering Scientific Papers

BE12-P.18 BE12-P.1.9 An Algorithm for DetectingParameters of Abnormal Nystagmus Beats Myoelectric Control of Practical L San*, Institude of AviationMedicine, Artificial limbs, LUO Yong-zhao; Bejing, China Shanghai Institute of Physiology, Chinese Academy of Sciences, 320 The key point in ENG data processingis identifica- Yue Yang Road, Shanghai, China tion of ENG waves and extraction ofabnormal nystagmus beats. A new algorithm has been develop- The loss of muscles, atrophyand disfunc- ed in our Lab and was called'Moving Mean Thres- tion of the muscles in the hold.' amputees' stumps interference, such as oculomotorwaves or result in a shortage of thesources of EEG blink waves, added to nystagmusbeats, and became control signals. In order tosolve this abnormal beats. Based on the fact of that the in- problem, less ENG signal patternsare used terference became verynarrow pulses relative to to control more actions ofartificial limbs time domain of slow phasewhen the ENG waves were and from less musclesmore EMG control pat- differentiated, we used themean ranging from terns are produced. One kindof our artifi- -S.D. to +S.D. as estimation ofthe slow phase cial hands only needone ENG pattern. It is velocity. Thus the interferencecan be cancelled used to control the opening ofthe fingers. notably. The characteristics of the algorithmare: The closing is automatic. Forthe arm pros- 1) Parameters of normal ENG waves can be calculated theses with three degrees offreedom only correctly. 2)Interference added to ENGcan be can- three ENG signal patternsfrom two muscles celled. 3) Because of its smoothingeffect, high are used. Whole arm prosthesescan be con- frequency interference, suchas EMG, can be filter- trolled by EMG signal fromthree muscles ed. More than 3500 of ENG records ofall kinds of around the stump. To meetthe needs of dif- abnormal beats have been processedwith this algo- ferent kinds of amputees,a series of upper rithm, and the correct rate of identification of limb prostheses has beendeveloped and used ENG parameters with this algorithmis 98.4%. The by many amputees. algorithm is very suitable forprocessing ENG records of patients.

BE12 -P.20 The Identification of Parameters HE12-P.21 in Heterometric Autoregulation Long -Term Continuous Monitoringof 'Activities of Daily Living' with of Cardiac Contractility.HanYan- Particular Reference to an: Li Shao-xiong. Zhao Xue-wei. Solitary Senile People, K.Ikeda*, A. Watanabe Kong Su-ming. Shandong Medical and M. Saito, Institute ofMedical Electronics, University. Jinan. P.R. of China University of Tokyo, Bunkyo,Tokyo 113, Japan A long -term ADL monitoring The system identification techniquewas system, especially for solitary senile people, which applied to study the dynamic processof consists of a number the heterometric autoregulation in of transmitters with vibrosensorsand a common re- ven- ceiver with data-processing tricular contraction. The experiments and data-memory units, were carried out in two groups of adult is designed for trial.The transmitters are small enough to be stuck Wistar rats.The process can be discrib- on utensils or furnishings, ed by a second order differential such as toilet door,refrigerator door, or kitchen equa- drawers. tion.Using Gauss-Newton method the par- When one of the vibrosensors ameters a.b.c and d were obtained.The detects any activity ventricular contractility is regulated of the sjsject, it transmitsits own code signal by parameters a.b.c and d. for several times running,otherwise it remains Among the quiescent. parameters c is the most importantone All the transmitters forone particular subject use same radio which reflects the relationshipbetween frequency. When the receiver detects the radio signal arterial pressure deviation andthe de- and succeeds in rivative of end diastolic pressure identifying a particularcode at least three times d(EDP) /dt. determining the in suc5ession, the code isto be stored as an event stability of in its memory unit the system and assesseing thecontrac- together with itsoccurence tility of the ventricular myocardium.If time. the value of c is less than The contents of thememory is automatically read -ad/b. the out at a fixed period, and system is unstable. The bloodpressure sent to the institution tends to drop down indicating the concerned, through the telephoneline, in order to cont- be analyzed. ractility of the myocardium is impaired. The capability and usabilityof this system are discussed tosome extent, especially from the view point of healthcare for the aged. 4

314 1988 World Congress on Medical Physics andBiomedical EngineeringSan Antonio, Texas Biomedical Engineering Scientific Papers

BE12-P .22 BE12- , Test for Ergodicity of Normal Human Blood ,...iradian Rhythms in Systolic and Diastolic Blood Pressures, F.E. Yates*, L.A. Benton, S.J. Berry and Pressures (SBP, DBP) in Normotensive Humans, L. Korenberg, Crump Inst.for Medical Eng., A. Benton*, M. J. Korenberg, S. J. Berry and F. E. UCLA, Los Angeles, CA; Dept. of Elect. Eng., Yates, Crump Institute for Medical Engineering, Queen's University, Kingston, Ontario, Canada. UCLA, Los Angeles, CA; Dept. of Elect. Eng., Queen's University, Kingston, Ontario, Canada. We hypothesized that repeated measures of 24 h ambulatory blood pressurein single normal individuals might recapitulatethe statistics of similar single records of a cross section of normal Ambulatory SBP, DBP and heart rate (HR) were measured every subjects, with respect to ultradian (< 24 h) periodic dynamics. 7.5 min for 24 h id two groups of normal subjects: younger (Y, 14 Systolic and diastolic blood pressures (SBP, DBP) were measured men and 7 women, 24-65 y) and older (0, 7 men and 10 women, every 7.5 min for 24 h. We obtained 13 records on each of 2 66-95 y). Variance structure of the -192 measurements per subject subjects (L - 35 y woman, F - 59 y man), and single records (group was examined by the Fast OrthogonalSearch method of R) on 13 men and 6 women (24-65 y).All subjects had 24 h Korenberg, modeling data as a series of sine/cos "features", which, rhythms in both blood pressures. We examined variances of each unlike the standard Fourier series, is not necessarily harmonic record for ultradian periodic structure using the fast orthogonal (comensurate). Any feature reducing variance > 5% was considered search method of Korenberg, which models data as a (not significant. (Shuffling a data set eliminated all features.) We could necessarily harmonic) series of sine/cos pairs as features.We detect features between 0.5-24 h. Results: All subjects had strong windowed the features into 3 domains:9 h, 6 h, 3 h. Results: % -24 h ( "circadian") rhythms in SBP, DBP, and HR. In Y there was occurrence rates of significant features for SBP (9, 6, 3 h) and DBP also a 9 h feature in DBP (p < 0.01), but no other ultradian (< 24 h) (9, 6, 3 h) were: F (38, 31, 15) (46, 46, 23) /L (38,38, 31) (54, 23, 15) / R (16, 26, 10) (47, 21, 16). No group had an ultradian feature periodicity in any variable. In contrast, 0 showed prominent 9 h prevalence (t) > 55%, but all 3 groups had a 9 h DBP fp > 45%. and 3 h features in both SBP and DBP, as well as a 6 h feature in R had no other fp > 25%, whereas F and L had SBP 9 h fp = 40%. DBP. HR had no ultradian features in any group. There was no L also had SBP 6 h and F DBP 6 h fp = 40%. Conclusions. (1) scx difference in these findings.Conclusion:Cardiovascular ultradian rhythmic structure of normal human BP is inconstant, (2) pressure dynamics change with age in normotensive subjects such near-periodic features are not rare (fp = 30% overall), (3) a 9 h that ultradian rhythms become prominent in both SBP and DBP DBP feature is most common, (4) individuals have distinguishing after age 65.(Supported by USPHS Grant GM 23732 and the "dynamic signatures" seen in the distribution and occurrence rates Hartford Foundation.) of their ultradian rhythms. (Supported by the Hartford Foundation.)

EE12-P .24 BE12-P.25 Digital recovery of respiratory rhythmfrom The Electro Optical Modulation of the ECG and the pulse wave, J. Colominas*, Chtochrome C, Yu Wei *, Yong-Qing J.Rosell, P. Riu and R. Pallas-Areny, Div. Hang & Ci Lin, Nanjing Institute Instrumentacion y Bioingenieria,ETSI Tele- of Technology, Nanjing,China com, Box 30002, 08080 Barcelona, Spain

Thephysical proximity of organs ofrespiratoryand Recently itis an interesting research topic in cardiovascularsystemsresults ina mechanical inter- bioelectronics to find some certain molecules and action between them and the corresponding presence of biomolecules which can be used in electronic de- interferences in the signalsretrieved from one, which vices. Two structure frames of cytochrome C are may be attributted to the other.Neural control mecha- made in our Lab. One is a thin mambrane of cyt-C nismsare alsoresponsible for anotherkind of inter- immobilized in gel of PPA, the other is the lang- action in theform of arespiratory-inducedcardiac muir Blodgett film of lipid absorb.ng cyt-C mole- rhythmvariability(respiratory sinus arrhythmia, RSA) cules. A measurementsystem with microcomputer By modelling the way in which the interaction modifies has been developed in order to obtain the electro- signalsassociatedwith one system,we canconsider optical modulation character of cyt-C. The samples theresult of the interactionas a means ofinducing of cyt-C are inserted into two electrodes which informationfrom onesystem into the signalsassoci- are light transparent and applied with alterna- ated with the other, and then try to recover it. tive voltage. The the photo transmission in cases Ourexperimentallyassessedmodels indicatethat, of different electrodes, applied voltage, irradia- aside from the RSA, there is an amplitude modulation tion frequencies and immobilizing method etc, has and a base band component associated with the respira- been recorded. The photo transmission is changed tion,both in the ECG and in the external pulse wave. according to the redox of cyt-C. Finaly,it is By digitallyprocessingthese signals,we recover the tried to exlain the mechanism of the electro-opti- respiratory rhythmfrom them withoutanyhardware cal character of cyt-C. modificationsfor existingequipment,thus increasing the amount of information obtained for the same cost.

3 3 5'

1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas 315 Biomedical Engineering ScientificPapers

BE12-P.26 BEL2-P.2/ ENZYME CATALYZED ETHANOL DETOXIFICATION OF Sensors and systemsfor the THE DOG David Whitmire*, Dept Chemical Engineering, Auburn Univ, Alabama measurement of energyflux through skin,L.Yu*,D.Y.Liu An ethanol detoxification formulation Oceanic University using a three enzyme systemwith of Qing- substte pumped cofactor recycling dao; H.Zong,QingdaoElectronic (SPCR) was used successfully in dogs. Co. Qingdao,P.R.China. When administered intraduodenallyto anesthetized dogs the formulation caused a 35 to 45 percent Tne energy or reduction neat flux tnroughoody in intravenously administeredblood skin is different ethanol over that of thecontrol. When in concept fromskin temperature. People adminstered intragastrically prior wno nave thesame to the intragastri- administrationof skin temperature ethanol, the for ation prevented may have unequalenergy greater than 96 percent of the flux. Energy flux expected may oe used toprone amount of ethanol from appearing the temperature in the vascular space. within the oodyand has some relations with the conditionsof inner organisms. So that itnay oecome a useful parameter for clinical diag- .iosis. Descriptions of the sensorsand systems that usedare given,their functions and measurements arepresented.

BE12-P.28 A data BE12-8..1. acquisitionsystem forelectrical impedance imaging,J. Rosen*, APPLICATION OF BIOMEDICALENGINEERING R. Pallas, P. TO SKIN MANAGEMENT, Riu and P.Rolfe+,Div.Instrumentation y E. Flam, Ph.D.*, Bioingenieria, NTL Associates, Inc., 29 ETSITelecom,Box30002, Ainsworth 08080 Barcelona,Spain, and Dep. of Biomed. Avenue, East Brunswick,NJ 08816 Ells. and Med. Phys., Univ.of Kee le, United Kingdom(+). In an earlier study theimportance of aeration to theproper management of the skin in prevention ofdecubitus In an impedancemeasurement system using N ulcers was discussed. points, the accessible number of independentmeasurements This current paper image reconstruction is for presents further N(N-I)/2.We have developeda research into the factorsthat systemthatincreases the speed of datapresentation contribute to the healthof the skin. by carrying out N-I Wher. skin is subjected It simultaneous measurements. to adverse consistsof analogmeasurement circuitry controlled environmental conditions, bya it microprocessor cardcommunicated to a personal experiences changes in itsphysical computer via an IEEE-488 bus.The PC is in charge of properties. Exposure to mechanical data processing. and imagepresentation. aggravations compounds the The driving signal is problem, generated from an EPROM,which and when internal factors,such as feedsa D/A convertorand is successively applied disease states are added, different electrode pairs to the combined using a conmutationstage. result is significantdowngrading of 15 voltages are simultaneouslymeasured by the skin's integrity tion and coherent amplifica- and ability to demodulation and then digitizedby a protect. These contributing factors singleA/Dconvertor. Digital signals have been analyzed. cally coupled to the following are then opti- A technique for stages. developing a weighted equationof their effects is presented to help the biomedical engineerselect the most suitable ranking andarrangement of conditions fora particular patient.

3 3 6 316 1988 World Congress on Medical Physics and BiomedicalEngineeringSan Antonio, Texas Biomedical Engineering Scientific Papers

8E12-8.2 33E12-8.3 A pocket -sized computer aselectronic memo-book INNOVATIONS FOR IMPROVED HEALTH CARE for patients. H. Hutten*and J. Bergeler, T.G. Krishna Murthy* and Swamy Lakshminarayan Johannes Gutenberg-Universit5t, D-6500 Mainz Bangalore, India (FRG) Recent advancements in Science, Technology, Many patients have to keep memo-books in order engineering and medicine coupled with the to supply the doctor with diagnostically and emergence of interdisciplinary fields and therapeutically relevant data. Typical examples closer interaction and Co-ordination have are diabetics (blood glucose, insulin injection, led to the development and clinical food, hypoglycemia, other events) and pain pa- utilisation of a number of battery tients (center of pain, extent of pain region, operated therapeutic appliances and degree of pain, pain expression, applied drug devices.ax An integrated therapeutic etc.). A special pocket-sized computer was de- approach which has the final target veloped that meets all requirements for an elec- of providing patient relief at moderate tronic memo-book. It contains a 8-bit CPU, a 64- costs provides a challenging opportunity kByte EPROM (program), a 32-kByte RAM (system to the present decade of engineers, software), a 32-kByte EEPROM (data), a quartz- scientists and technologist to evolvenew controlled clock, and a V24-interface. The sys- techniques in close interaction with tem is battery powered (6 months with usual medical specialists. Nearly 10 per cent employment) and an additional buffer battery for of the population of most nationsare the protection of program and data. It is oper- handicapped clue to one reason or the other. ated by 14 keys (0-9, Y/N, decimal point, ON). Rehabilitation engineers havea really The user program is arranged as sequential menue important and vital technosocial role to with branches. The user is asked by the system play. ureater challenge lies in via a display (2 lines, 20 characters each). All innovating adaptive techniques depending answers are displayed and can be outputted via on the socioeconomic conditions and the the interface on a monitor or printer. actual needs. The coming decade may witness the attainments of targets which can lead to the goal of HEALTH FOR ALL BY 2000 A.D.

BEI2 -U.4 BE12-R.5 Instruments fora New Direction in Surgery. Endoscopic Liquidiser and Surgical Aspirator. 3.Wickham, R.Miller, M.Coptcoat, K.Ison*, T.Mills K.Ison*, M.Coptcoat and 3.Wickham, Institute of and P.Swain, Department of Minimally Invasive Urology, London WC2H 83E, England. Surgery, Institute of Urology, London WC2H 83E, England. Surgical techniques are becoming less invasive and new instruments are being developed to The future of surgery is clear. Patients want support this trend.A need is emerging for a treatments to be less painful and mutilating, device which can remove appreciable quantities while administrators want patients to recover of tissue quickly through a small diameter, for more quickly and cost less to service. Surgical example in endoscopic prostatectomy. We report procedures are under pressure to become less the development of a machine to do this. invasive. Less cutting means faster healing and The device cuts and liquidises tissue for opens up the prospect of using more local or removal by suction through an endoscope or minimal anaesthetic regimes. The challenge to laparoscope.Its detailed design will be scientists and engineers isto develop described. An experimental system has been instruments which promote minimally invasive constructed to look at the effects of cutter surgery.. design on the mechanisms controlling the removal Endoscopic techniques are employed for diagnosis rates of different tissues. These rates are and intervention inan increasing number of comparable with those achieved using diathermy medical specialities. We review these areas of during in vitro and in vivo trans-urethral application, with particular emphasis on surgical prostatectomy (>10 g/min),but are faster in techniques and associated instrumentation. We model laparoscopic procedures (>30 g/min). describe novel endoscopic sewing, stapling and When used on soft tissue in patients, aspirated cutting machines developed locally. We also particles vary from micrometres to millimetres comment on additional devices which will foster in size. Size distributions depend on cutter and the applications of endoscopic surgery and pave tissue type. The device has also been used to the way for the formation of a new surgical fragment calculi. Progress with the device will grouping. be reviewed and its applications discussed. The feasibility of using the machine under robotic control is being studied and results from this sessment will be presented.

1938 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas 317 Biomedical Engineering Scientific Papers

BE12-G.6 BE12-a.7 Evaluation of Portable Insulin InfusionPumps, U. Bostrom*, B. Bouligand andI. Strom, Image Processing of FootPressures for Clinical Engineering Department, University Assessment of Leprosy, I. Sailaja, K. M. Patil* Hospital of Linkiiping, S-581 85Linkoping, C. Eswaran and K. S. Rao, Sweden. Biomedical Engineering Division,I. I. T, Madras -600036, India Portable insulin infusionpumps have been used in Sweden since 1981. They arenow used by about 300 patients. A The application of digitalimage processing in the prognosis has shown that in 1988 assessment of leprosy is there should be about 30 described.Amonochrome 000 portable insulin infusionpumps in use all over the analog image of the feet is world. Several designs are represented obtained by breakdown on the market. More of total internal reflectionof light on a device and more sophisticated programmablepumps are called barograph. The intensities in the imageare introduced. There is a need for technicalassessment of related to the safety and reliability before the hospital pressures under the feet ofa subj- decision of ect standing on the barograph.The scheme of image purchase. processing adopted after digitizing The basic technical principles of portable the image, for insulin infusion noise suppression and edgedetection is highly pumps have been studied and the technical safetyproblems caxoutationally effective to have been analyzed. Advantages, disadvantages be useful in process- and ing vast amount of data.Segmentation aids in edge maintenance problems for five differenttypes of pumps detection obviating the have been analyzed through testing, interviews necessity of using complex with thinning and connectivityalgorithms. The processed physicians and studies of literature. image gives (i) A test method suitable for clinical centres of pressures of bothfeet, engineering departments together and individually,useful in the determina- will be presented. The possibilitiesto make safety controls tion of the side to and to maintain the pumps will be discussed. which the subject isleaning, isopressure contours givinga picture of the distribution ofpressures under the feet and (iii) pressure under specificareas of interest. Analysis of the aboveparameters shows that the leprosy subjects develophigh peak and average pressures in specific areason the deformed and insensitive foot.

BEI2-0.8 Study of Body Movements of BE12-8.9 Alzheimer Disease and Model DTZ-1 Medial Electronic Schizophrenic Patients UsingUnconstrained Measurement. Syringes, Li Ming Liang* He Yi T. Fuj ir.oto*, T. Toyosig ima. T. Tsuj i and T. Togawa, Inst. for Xia and Tian Anmin* No.409 Hos- Med. pital, the Ministry of Electro- Engng., Tokyo Med. & Den.Univ-2-3-10, nic Industry, Shaanxi, Boaji,China Kandasur.gadai Chiyodaku Tokyo,Japan101 The Medical Electronic Syringesresolve the problem of manually delivering drugs in intravenous injection for the nursing staff The close :Plationship betweenbody movements and mental state and make the intravenous injection into is well known especially in mental the e corder where Cie4 luracte- process of the electronic instrument, dosage ristir: atrormal i ties of movement arc observed.fiemeasured these and meanwhile, make the speed of theintra- Sxly movements using an acceleration venous injection adjustable, in addition, sensor without constraints. reduce the nurses' labour tensity. The The continuous accelerationsof mist, right wrist and ankle ale syringes have widely range preset timer so rig with an electrocard: ;,^iim that the cardiotonic glucoside, calcium were 4imultaneouly measured and r- and other injection, etc. which needto be in- ecordedbya casette data rff;order.After processing the signals, jected invein slowly in clinicaltreatment, the amplitude and frequen.,of the accelerationswere plotted can be injected even move scientifically. on the graph, and quantitative changes, The medical eletronic syringesare also able coherent relations and w- to be used for delivering drugs quioky,tra- ave patter, were studied.The cl'anges in the quantitiesof move- nsfusing brood and keeping the infusionpipe a long time for the palients who ment well reflected the mentalstate of the pat ient.The graphs are ill seriously. The widely ranged preset timer for Alzheimer disease sh:ared ';or simple and sporadic changes in the the infmenoun injection has greatautonomy Invements.lhose o" the rcizophrenicshowed sporadic,repetetive in hospital or at home and makes theintra- venous injection move adaptable specifica and stereotypic patterns.This device is useful in measurik; the tions. movement of the,patient Alth i ch ..helps-to..anal yze-the-d f fering-m- t. Suitable range tor injection: tor -various. ental states of the psychiatricpatient. drugs by syringes in 50mI and 100mI maximum pushing force 500g.

318 1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas Biomedical Engineering Scientific Papers

BE12-a.10 BE12-a.11 Model Zj-1 Steem Curing Device, BEATE - A new System for Testing the Li Ming Liang'and Tian Anming, No. Function of Hospital Ventilators 409 Hospital, the Ministry of Ele- ctronic Industry, Shaanxi, Boaji, U.Boenick* China Up until now, no integrated measuring This device is used for curing common cold, system has been available for influenza as well as allergic rhinitis. determining the ventilation The cure efficiency comes up to 88.9%. This parameters volume, volume-flow and device is superior to the other similar de- ventilation pressure, as also the vice in heating without two times. The tem- gas-specific variables oxygen perature of the outlet is controlled between content, temperature and humidity. In order to overcome this deficiency, 40 °C -60 °C. The device is possesed of the the Institute for Biomedical simple structure, low cost, low power consu- Engineering at the Technical mption and perfect fuctions, so it is very University, Berlin has developed a ecnomical, practical and conventient. The portable lung simulator with an curative effect for curing common cold is integrated tester, known by the name quick and no side effects. Specifications of BEATE. Now, for the first time, a 1. Temperature measuring accuracy test system that permits the tepid (350c...... 550c) 10c and simple determination of all the ventilation parameters relevant to 2. Temperature controlling accuracy safety, can now take its place alonnide the systems already long (35°C---55°C) ± 1°C . available for the testing of 3. Cover insulation resistance 1000M electrical safety in accordance with VDE 0750. 4. Power comsumption heating condition 300W. heat preservation condition 170W 5. Power voltage 220V+ 10%.

BE12-a.13 BE12-a.12 OUTPUT PRESSURE WAVEFORM EVALUATION OF IV Technical study on the Efficiency of Heat and INFUSION PUMPS Moisture Exchangers (HME). E. Amore *, Servizio Fisica Sanitaria e P.A. Schroeder, U. Frucht, C. Ullrich, Bioene, Parma, Italy U.Boenick* and W.Kanig The hign degree of accuracy and safety rea ched by the recent intra-venous infusion In order to humidify inhaled gases in anaesthesia and intensive care disposable Heat pumps, has led to a widespread adoption of and Moisture Exchangers (HME) have been these systems in almost every field of me- introduced. dicine,especially where long term and pre- A new experimental set up was developed to cise pharmacological treatments are needed. investigate the efficiency of HME's. Nine Different techniques in delivering the flu- commercially available HME s were measured. To prevent harm to the patient, the relative id into the venous system are currently humidity of inhaled air should be at least 7o %. available (revolving roller,linear peristal Only 4 of the 9 HME's tested fulfilled this tic,volumetric). condition. With respect to the requirements for The aim of this investigation is to calcu- minmal dead space and an optimal moistening late and measure the transient component effect, only one of the HME's met the goals. of the output pressure waveform exerted in HME's might be an alternative to traditional ways of humidifying inhaled gases. the different types of pumps tested.Fur- thermore, the influence of the stroke volu me infused in each pumping cycle on the hidraulic venous impedance,i.e. resistance and compliance,is demonstrated. Results show that pressure transient component differ significantly from the volumetric or roller system to the linear peristaltic in which the infusion seems to be much smoother. This fact reflected atthe input site of the venous system implies a diffonmnt stress and hence a minor possibility of in- 3 3 9flammations or thrombosis. 1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas 319 Biomedical Engineering ScientificPapers

m12 -m.4 BE12-a. 15 A Simple Exerciser forNickle Cadmium Reuse of Batteries in Medical Indeflators, J.Hagekyriakou, Equipment, M.A.Chaudhri*, K.O'Brien,M.A.Chaudhri*, Austin Hospital, Austin Hospital, Melbourne, Australia Melbourne,Australia During coronary angioplasty, and some other The inability of medical equipment to interventional radiologic.1 exercise NiCd batteries to maintain peak procedures involving theuse of a balloon catheter, performance and good service an instrument is used to both life, has inflate meant costly replacement and measure the inflated of the pressure. batteries. In order to lower the wastage Sterilization and re-use of such an and avoid the tying up ofa largo "indeflator" had originally amount proved impossible, of laboratory equipment to cater forthe due to the fact that different capacity batteries, the contrast/saline solution usedin the we designed indeflator and constructed our own simple reacts with the brass - exercising component system. of the pressure gauge, resulting It was designed in the production of a salt using the standard which Eurocard with a DIN 41612connector, and accumulates within thegauge and six syringe. We have exercising modules are housed in a devised a method of single BOPLA Combicardcase. Each module successfully removing thissalt from the allows selection of any three indeflator prior to sterilization. constant The process current charge rates to suit involves dismantling the the gauge/syringe assembly, particularbattery and the battery washing out the is gauge several times simply connected, the module reset, and and then freeze- drying it. This automatic cycling will then takeplace. removes the salt and The "exerciser" will firstly water which is trapped withinthe gauge. constant- The gauge current-discharge the and syringe are then re- battery to the assembled minimum recommendedvoltage and sent for ethylene-oxide and then sterilization. constant-current-charge for the The gauges may recommended time. It will then repeat be used virtually the cycle twice. indefinitely provided they are not damaged whilst inuse.

HE12-.I.16 BE12-0..17 A Patient Monitoring System in the Column-Minireactor-Based Enzyme Operating Rooms, Toyoki Kugimiya*, Analyzers of Metabolites, M. Tsuzuki, K. Numata, Surgical A.Simon- Center, ian*, Ts.Avakian, YerevanPhysics The Department of Anesthesiology, Institute, Armenia, USSR. University of Tokyo Hospital, 7-3-1, Hongo, Bunkyo-ku, Tokyo, 113, Japan Enzyme analyzers basedon column minire- actors have some advantagesagainst en- A sophisticated patient monitoringsystem zyme electrodes. They allow to solve during anesthesia is indispensable ana- not lytical problems inmicrobiological indu- only for high risk patients, butfor com- stry and medicine plicated surgical procedures, such on determination of or- as ganic substances in multicomponentsolu- open heart surgery and extensivesurgery tions. The use of such systems for malignancies. To build up a system allows to control and manipulatebiotechnological which is suitable for a particularinsti- processes and simplify the analytical tution requires a careful planning based procedures in clinics. A wideras compar- upon the needs arising from its clinical ed to enzyme electrodes activities. range of determination of concentrations,a shorter time In our new operating theater, whichstart- of contact of reaction ed to function at the beginning of products with the this enzyme, little influence ofprocesses of year, a system which was designed to meet the substrate and the most demanding clinical situations reaction product was mass transfer are to be referredto the installed. The system includes cardio- advantages of flow-through vascular and respiratory mohitoring systems. We have constructeda flow-through enzyme systems and other parameter mwitoring. analyzer providing for Close cooperation between clinicians a highly specific determination of lysineconcentration in mainly anesthesiologists) and clinical the culture fluidof glucose and uric engineers was found to be essential for acid in the blood'serum or urine. The the proper and efficient managementof the system for the patients' benefit. range of determination of concentrations for lysine is 5.5-55mM, for glucose- 0.3- 16.6mM, for uric acid- 0.06-6mM. 3 4

320 1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas Biomedical Engineering Scientific Papers

BE12-a.18 BE12-8.19 HYPERALUMINEMIA OF PATIENTS WITH NEPHROPA- Viewing and Evaluating Device for Stereoradiographs THY,Y.F.Chiang: T.B.Lan, Dept. of Nephrol. J. Barsai*, B. Heiner, and T. Illes, Medicor Train- ing Centre, P.O.B. 150, H-1389 Budapest 62, Hungary Wuhan General Hospital P. R. China The device of our design and manufactured by the This paper reports the results Of measureme- Hungarian Optical Works/MOM/ is for the simple and nt of 182 serum samples of 142 patients(pits) prompt evaluation of stereoradiograph pairs. Our with nephropathy, 116 cases of hyperalumine- idea in designing the device was to relieve the mia(HA) were found by atomic absorption spe- radiologist from calculations. Our solution by ctrophotometry(40.14% of pits were ill with mechanical means actually makes use of a hyperbolic glomerular disorders(GD), 41.551 of pits we- function of four variables. The equation is re tubular interstitial disorders(TID) ).Of t=p.f:bfp, where t is the distance betwein the which 32.391 of pits were of normal renal fu- point under examination and the basic plane, p is nction(M), and 49.31chronic renal failu- the stereoscopic parallax, b is the distance bere(CRF).SAT. of normal controls were 0.63+0.0 tween the focal points, f is the SID/source-image 9 umol/L,-5 of cases with GD and TID were distance/. The pair of stereoradiographs viewed 4.42+2.16 aBiT 4.48+2.59 umol/L respectiyely in the device through a binocular eyepiece give the (P70.05). Befc-s .3rd after administration of viewer a stereoscopic effect. The problem is to AL(OH),,S Of controls showcd rignificant determine the distance of the point under examina- changefo.P+0.618 P,o.o5), while the urine tion to the basic plane. The device solves this AL was very significanfly elevated(0.725+i.49 problem fully. All points can be localized in 2-3 and2.85+2.04umol/L P'0.01).CA. of NRF wa- mm accuracy. The device can be handled also by 3.435+2.404 and of CRF was 5.4142+2.1q7umol/L persons having no faculty of stereoscopic vision. (Plo.o1).After hemodialysis and Continuous Ambulatory Peritoneal Dialysis, the SIT levels were significantly increased, Intitiw- Atflcantly decreased after Gastrointestinal

Symptoms of HA mat- .y were:anemia, muscular and bony pains, e.xephalopathy. The related factors and mechanisms of HA we re briefly discussed.

BE12-a.20 Application and Field Trials of BE12-a.21 Videophone in Health Care, P.Kaitile, New algorithms to detect the anaerobic threshold Vistacom Industries Inc., Kipparin- noninvasively, P. Gizdulich*, P. Panuccio, L. katu 2, SF-02320 Espoo, Finland Viroli, F. Cresci, Dept. Fisiopatologia Resp., Universita 1-50134 Firenze, ITALY During increasing muscle efforts, many ventilatory functions show a discontinuity which is related to Applicability of a videophone with motion the anerobic phase that replaces the aerobic state. picture was studied in different areas The anaerobic threshold is usually evaluated by of health care. Three field trials were applying a 2-segment linear model for a clinical carried out: a communication link between and diagnostical purpose. We analyzed the three a long term patient in a hospital and her most commonly used pairs of parameters: Ve in family, a consulting link between a health function of V02, VE versus VCO2, VCO2 versus V02. care station and the emergency clinic of The three functions have been studied during exer- a hospital, and remote assistance of a cise with a cycle-ergometer. In ten healthy sub- handicapped person at home. People learned jects data were collected every 15s by using a com- in a short time to operate the equipment. puterized acquisition system, until the subjects Videophone proved to be practical. X-ray were exhausted. For each function we split the images could be shown to specialists in N-data set in two series of data: the initial n and the hospital. Ditferent moves of arms and the subsequent N-n. Two linear regressions were legs could also be shown. Teleconsul- computed for every possible n. In order to select tations mainly succeeded. According to the best value of n, two new algorithms have been results, the number of patients to be proposed and compared to the least square error moved trom nealth care station to hospital method, usually applied. Reliability of the 3 could be decreased. Videophone seemed to algorithms was tested by the within subjects analy- suit well in consulting major fractures sis of variance. VCO2 in function of V02 was docu- of bones. Operational telemedicine service mented ,o be unreliable. As for Ve vs 002 and Ve may be arranged, especially for areas with vs VCO2, the new algorithms showed a greater repro- low population densities, when further ducibility than the least square method. development of certain technical features of videophone has been realised. 341

1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas 321 Biomedical Engineering Scientific Papers

BE12-U.22 BE12-Q.23 Ergonomic Analysis of Japanese PACS, R.HOSAKA* and Applied Values of Rheooncephalogram E.TAKENAKA, Dept. of Radiology,National Defense Medical College, 3-2 Mamiki, Tokorozawa-shi, ( REG ) in Clinical Practice, Tian Saitama-ken, 359JAPAN Amtlin; Dept. Pharmacology, 3aoji Me- dical and Nursers' School, Baoji, Sh- PALS( picture archiving and communication system ) aarrzA, P.R. China. is a useful equipment ina field of radiology. Japanese industries also developsame types of 1. The vaeo4aeadache may be classified acceding equipments. But, there are several problemsin to characteristics of REG (from 930 patients ), the Japanese PACS. One of them is a problem about i.e. URA ( the REG wave of high amplitude);RWLA its man-machine interface. Almost of Japanese (the REG wave of low amplitude). PACS have a poor man-machine interface. It is 2. The classifications of REG in vase- headache constructed by a chLracter key-board as a input may guide to use medicines. The effective rate device, and one or two high-resolutionmonochrome CRT as a output device. of the vasoconstrictor drug for 250 patients This type of interface with RWHA was 96.7 % and of the vasodilator drug is not useful to control the equipment. for 50 patients with RWLA was 39.5 %. In this study, present interfacesystem and new interface system are comparedfrom view point of 3. The REG may keep watch on the state of cere- handling time length. Both systems are con- bral blood and arteriosclerosis. This date(from structed on a micro-computer likeas a PACS 498 middle-old aged mental workers) show that control system. In a case of present type,a key- the cerebral blood often isinsufficientand th- board and a monochrome CRTare used as interface at the cerebral arteriosclerotic index of REG is devices. In the other case, a touch-screenand a Gradually increased to follow theincroamentcf RGB-CRT are used as interfacedevices. ago. The result of experimentsseems to indicate that the time length of newtype is shorter than present type. The difference of both timelength is approx. 10( Z ).

BE12-a. 24 A New Patient Record System Based BE12-0..25 on a Laser Experimental Measurements for Assessing the Card J.H.U. Brown! Ph D.. University of Houston. and Influence of Static Magnetic Fieldson Pace Yallbona, M.DBaylor College of Medicine, And ti makers, V. Barbaro: P. Bartolini, M. Donni- Kitesonono. CSK, Tokyo nelli, D. Ialongo, S. Margarita, Istituto Superiore di Sanity- Roma, Italy. A method has been devised of placing 2 millionbytea(700 pages of information) on a standard sized credit card so thata medical patient can carry the card with him at alltimes, Static magnetic fields can adversely affectthe pa updateit at each visit,end have itavailable durino cemaker (PMK), switching it from theoperative sta emergencies. tus D;mples are impressed in the card witha to a predefined test mode. This effect isob- tained millwattt laser beam which also servesas the reading device. by means of the excitation of the reedre- Write speeds are in excess of 2 0 kb/sec end lay read speeds are built inside the PMK. It has been realizeda about 65 kb/sec. Any single datum can be located In less system than to produce a uniform static magneticfield one second. with en error rate of 10-11The information in a 'olume enough large to containPMK. The c/..cu can be added to as desired and the information can be read llr coil of the system has been excitedby a cur- with an ordinary 18MO p.c. with no mainframe needed. 1 't generator and the Xrays and Ecgs can be placed on the card. The software produced magnetic field on a strength has been continuously and floppy disk;is designed to present medical informationin slowiy changed. The PMK was put inside the conventional format(socioeconomic data, drugs,operations, solenoid and its operating mode was monitored. When the treatment, prognosis, etc) with the additional abiliityto add test mode appea- free text as desired. red the field value was recorded.The measurements The system is under test at the Harris County Hospital were repeated with the PMK positioned along the District Clinics and at NASA. Evaluaticn of the systemunder three orthogonal axes, for eachFMK the minimum re test will be presented. corded field value (Burin) has bee taken in account. 127 PMKs by 12 manufactureswere tested. The Bain ranged from 7.5 to 54Gausswitha mean valueof20.3Gauss. These values can currently occur in some environments.

eI 41 2 322 1988 World Congress on Medical Physics and BiomedicalEngineeringSan Antonio, Texas Biomedical Engineering Scientific Papers

BE12-a.26 BE12-a.27 JXYGEN ADVI:,OR: A Computer System for Linear analysis of dynamic response in heart rate Oxygen Therapeutius D. John Doyle. MD PhD, to physical activity in daily life, Takeshi University of loronto, Toronto, CANADA Toyoshima*, Tatuo Togawa and Takayuki Tsuji, Tokyo Medical and Dental University, 2-3-10 Kanda- Interest in quantitatively assessing Surugadai, Chiyoda-ku, Tokyo 101, Japan. pulmonary function and predicting blood oxygen tension tP02) and saturation In order to analyze the dynamic response of heart following changes in inspired oxygen has rate to physical activity including posture change existedsince the advent of oxygen under unconstrained daily life,a system for therapy. Potential advantages of such quantitative evaluation and simultaneous recording methods would be to reduce the frequency of these parameters was developed. In this system, of arterial punctures and to minimize the intensity of physical activity was evaluated the incidence of hypoxemiaand oxygen by measurement of acceleration at the subject's toxicity. OXYGEN ADVISO' is a small waist and postures were discriminated as standing, hand-held computer systemwhich 1) sitting, stooping and lying, according to the computes a number of indices of inclinations of specialangular switches attached pulmonary gas exchange and oxygen at the thigh and the sternum of the subject. This transport teg, shunt fraction, arterial- recording system was applied to five healthy alveolar oxygen tension ratio), 2) subjects. Then the dynamic response of heart rate predicts P02 changes followingoxygen to physical activity was analyzed applying linear changes and 3) produces patient-specific system analysis, regarding heart rate as the oxy-hemogobindissociation curvesfrom output, and physical activity and posture change patient temperature and arterial blood as the input of a linear system. as data. The computer hardware is based The results showed that the heart rate did not on a Sharp 1500 microcomputer with a only simply follow the intensity of the physical four color graphics plotter. Its small activity, but it also responded as a first-order size andease of use makes it very linear system with a time constant of about 70 practical for use in critical care min, and that the heart rate increased and environments. decreased instantly with posture change by about 11% both from sitting to standing and sitting to lying.

BE12-a,28 BE12-11.29 Computers and Safety - A Challenge for All A Digitized §.ecorder for Ambulatory ECG S.Hashiguchi ,T.Mitomi,Y.Kondo,Yamanashi G.R. Symonds' University,Kofu,Japan, Bureau of Radiation and Medical Devices K.Tamura,and S.Komori,Yamanashi Medical Department of National Health and Welfare College,Tamaho,Yamanashi,Japan. Ottawa, Ontario, CanadaKIA OL2

The integration of digital computers into medical A digitized recorder for ambulatory ECG has been diagnostic and therapy equipment has resulted in developed. It consists of an AD converter, a major advances in patient care. The pace of this buffer memory, a magnetic recorder such as a development has been very rapid, with design portable VCR or DAT. 12-channel ECG signals, emphasis placed on innovation and functionality. sampled at the rate of 250Hz, are converted to That this process has not been without problems 12bit data and temporarily stored in the buffer is evidenced by patient injuries and deaths, with memory. When the buffer memory is fully occupied the resulting increase in attention by government regulators, standards organizations, professional by the data, the data are read out at the rate organizations and the manufacturers themselves. of 198kbytes/s and recorded on a 8mm video tape or a digital audio tape as PCM data. ECG The set of issues raised by these problems covers waveforms of 12 channels can be recorded for 31 a wide range: the multidisciplinary nature of the hours in a :20-type magnetic tape. design process, user training and qualifications, The reproduced waveformes were not distorted and the man/machine interface, and the applicability no significant differences were noticed between of safety engineering and quality assurance to the control signal and the reproduced signal. both hardware and software. Even the concept of a "single fault" will require rethinking with The fidelity of the reproduced waveforms is much regard to what constitutes a "component" in better than the conventional (analog recording) programmed systems using complex microcircuits. ambulatory ECG recorders, and good enough to apply the myocardial ischemia analysis, where The current trend in government regulation and the reproducibility of the ST-portion of the safety standards writing is one of international waveform is of the most importance. concensus. This, taken with the above, promises an interesting challenge for everyone involved. 343 1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas 323 Biomedical Engineering Scientific Papers

BE12-a.30 BE1241.31 Assessment of vascular reactivityby bWirnediWPattam thermal entrainment of heartrate, A. Lindoivist*,J. Jalonen,P. Parviainen, K. Anti la, I. Valimaki andL.A. Laitinen, Rt. Hermlda* and F. Hedberg,Univ. Santiago. Research Institute of Military Medicine, Spain & Univ. tlinnrisota. USA. Helsinki and CardiorespiratoryResearch Unit and Institute of AppliedMathe- matics,University of Turku, Finland. In order to disting, ish newbornsat low or high risk of developing a high blood Sympathetic nervoussystem transmits pressure 03P) later in thermoregulatory circulatory oscilla- life. 16 characteristics of timeseries on systolic. tions. Oscillations of the bloodpressure mean arterial and diastolic 6P and heartrate NZ) are controlled by the periodic heart rate monitored automatically at 5-hrintervals for 46 variability. We studied theusefulness of thermal entrainmentof heart rate hours. are here used for classificationby a so- variability to measure thermoregulatory called monutest. an all-subsetsvariable selection nervous function by usinga standar- technique for biomedicaldiscriminant dized water bath stimulator.Periodic analusis. thermal stimulation oflegs was per- Wherithe circadian amplitude ofdiastolic I3P and the formed to 7 healthy control subjects range ono stanaard deviation of HRwere used between 0.01 and 0.10 Hz. Fast Fourier- cohcomitantiu as classifier for 62 transform power spectra of thesystemic newborns, this skin blood flow and heartrate signal precedure yielded a 63t1; classificationequivalent were calculated. In the skinthermally to prior criteria, the latterbeing based on the induced response exceeded spontaneous presence or absence of a family oscillations (p=0. 01) . history of high 6P. Periodic heart as rate variabilitycould be entrained deter mined by questionnaire. Themonotest only occasionally (p=0.82). Sympathetic complements rhythmornetric vascular procedures, used reactivitycan bemeasured earlier to separate from the skin blood flow but notfrom groups or newborns with or the heart rate variability. without a family history of high6P. and defines a set ;.f individualized criteria forrisk assessment

BE12-4.32 BE12-a.33 A Study of RepresentationMethod of Intrapartum Fetal ECG, F. Yano, S. Objective recognition of fetal P. Ninomija*, Aoyama Gakuin behavior with computer analysis University, Tokyo, K. Idogawa, Seitoku Gakuen of fetal aatogram, K.Maeda*, College for Women, Chiba, N. M. Tatsumura, K. Maeda, Nagata,K.Nakajima,M.Tatsumura, Tottori University, Tottori,Japan Tottori UniversitySchoolof Medicine, Yonago683 Japan The condition of a fetalheart can be estimated from a fetal ECG. To diagnoseintrapartum fetal Fetal behavior including itsactive state, resting distress at childbirth, itis important to make state and fetal breathing movement (FBM)is usually detailed variation recordsof fetal heart rates analysed by human observationof real-time, B-mode with a baseline by fetalECG R wave. For the screen. Further objective technique isfetal acto- sake, gram, which is the recording of fetal it is necessary todetect the exact movement(FM) locations of the peaks of fetalR waves. We have signals obtained by analogprocessing of ultrasonic developed a procedure toconfirm the location of Doppler low frequency signals. Visual analysis of the peak of fetal ECG Rwave from leaded fetal the record revealed clearrecognition of fetal ac- ECGs involving many noisesor artifacts. tive and resting states. FBM was also recorded. The method to detect As the FM is expressed with as follows : At first, weuse spike signals, itwas introduced itno AD converter and which fetal ECGs digitalizedat the rate of 400 the digital data samples per second. Second,we compute a moving were analysed with PC98XAcomputer. Simple statistics of spike amplitude and variance by using 32 pointsand finding a point interval, frequency histograms of spike intervals, that exceeds a criteria.Third, we compute their3-Dnumber a histograms in 10 rejection levels, regression line by some movinginterval of the and envelops of spikes were obtained. Noise point. Forth, we select theabsolute maximum in level was estimated to be the lowest rejection level the positive slope and thenegative slope of the from amplitude statistics and histogram. Fetal regression line. Fifth,we define the peak point active stateshowed high amplitude and wide of fetal R waves from thepoint of intersection of interval distribution, frequent peak signals in 3-D these regression lines. histograms till highamp- As the result, the locationof the fetal R waves litude, and frequent envelopes.Fetal resting state showed signal can be confirmed. Moreover,we have developed a disappearance. FBM showedmoderate but characteristic findings. method of representing thefluctuation of fetal R- The results showed R intervals. objective and automatic recognitionof fetal beha- S 4 4 vior.

324 1988 World Congress on Medical Physicsand Biomedical EngineeringSan Antonio, Texas Biomedical Engineering Scientific Papers

13E12-9.34 13E12-R.1 Computer simulation of intrapartum 3-D Mapping ox the Body Surface and fetal distress diagnosiswithout Epicardial Potentials, A. Calderon; uterine contraction, K.Maeda* and T. Aoki, N. Yamanoi and T. Furukawa, A.Makio, Tottori University School Institute of Medical Electronics, of Medicine, Yonago 683 Japan Faculty of Medicine, University of Tokyo, Conventional diagnosis of fetal distress needed 2 7-3-1 Hongo, Bunkyo-ku, Tokyo 113 channel records of fetal heart rate (FHR) and uterine contraction (UC). Because UC is maternal Asystem has been developed in which the signal, the diagnosis only with fetal sign, FHR, human body surface potentials and the epicardial is the purpose of this study. Fetal monitor output potentials are displayed ina 3-dimensional including FHR and UC signals was analysed by the 2 format using wire-framemodels ofthehuman computer programs including UC parameter and with- torso and the heart. This work makes use of the out it in the same 65 patients who completed the tools of computer graphics to help the doctor in vaginal delivery. Quantitative FHR parameters were his understanding and evaluation of the cardiac our FHR score evaluated in 5min and fetal distress potential distributions. index (FDI) obtained in 15min. These were analysed To illustrate the applicationsof this with the programs. The results obtained with and system, several cases with abnormal ECG - right without UC in 30min immediately before childbirth bundlebranch block (RBBB), chronic myocardial were compared to umbilical arterial blood (UA). infarction (OMI), and left ventricular FHR score and FDI showed clear correlation between premature beat (LVPB) - are given as examples of the two analyses. Correlation between FUR para- deviations from a normal path; results obtained meters and UA pH was better in FDI than FHR score. from the animal experiments are also presented. The cases who had 3 points or more of FDI showed The increased realism and spatial significantly low UA pH. Sensitivity and speci- resolution capabilitiesof the system, as ficiety for the prediction of lower UA pH than compared to ordinary ECG and the2-dimensional 7.25 were 86.4% and 88.6% in the cases who were contour map alone, lead to a clear understanding analysed by the program with UC, and theywere and interpretation of the underlying electrical 86.4% and 90.7% when UC signal was not used. phenomena of the heart while making a diagnosis The possibility of FHR diagnosis without uterine or presenting experimental results. contraction will be expected from these results.

BE12-R.2 BE12-12.3 A Method fori3ecreasingtdappingiMG Electrodes INTERPRETATION OF BODY SURFACE POTENTIAL MAPS via aligner Distribution, Shiro Usui*, Sozo Nam, BY MYOELECTRICAL MODEL SIMULATION RESULTS Naohiro Tods, Junji Toyama+ andi(azuo'famada", Dan Adam* and Samuel Sideman Toyohashi Univ. of Tech., Toyohashi 440, 'Nagoya The Department of Biomedical Engineering, Univ., and 41Meitetsu Hosp., Nagoya 464 Japan. Technion, Haifa 32000, Israel

Body surface ECG mapping is known to be useful A model of theheart and torso is used to for diagnosis of heart diseases. However, it re- generate epicardial and body surface potentials from quires so many electrodes that causes problems in the myocardial electrical sources. Theelectrical sourcesare produced by the activation processes of setting electrodes and recording the signals. To the left and right ventricles which are simulated by make it for daily clinical use,It is expected to a 3-dimensional finite-elements model. The heart decrease the number of electrodes. Lux et al. pro- model allows the description of the orientationof posed a method for this by using correlation in- the myocardial fibers and theirdistributed formation between the electrodes. However, the velocities andrefractory periods. A Purkinje number of electrode:. must be given beforehand and network is also included. Point sources are defined it requires tremendous computation time to get the on the surface which separates the activated areas transfer matrix. from thenon-activated areas. A human torso model We propose here a new direct method utilizing is used for thecalculation of the body surface the sampling theorem. We first apply the aligner potentials. These potentials arepresented as distribution analysis for a given map to find the standardE.C.G., or as the more informative local spatial frequency distribution. Then by potential maps. The simulated maps have been used picking up the highest frequency at each electrode to aid the interpretationof clinically measured position for each maps in one whole beat, the maps. Specifically, simulated activation patterns local maximum-spatial-frequency-band map has been due to ischemia show an early, low amplitude peak in constructed. Applying the sampling theorem to this the mapsduring the early stages of the OR period. This phenomenon, which has not been found in map,the electrode arrangement can then be given normal maps, directly. For example, an equally spaced 91 chan- has been located in clinically measured maps. Thus, the simulated nel map, the number of electrode is enough to be maps generated by the anisotropic modelare useful in quantifying and in half. The proposed method requires less calcu- extractingclinically important features from lation than Lux's method and gives a minimum num- measured body surface potential maps. This ber of electrode and its arrangements. represents a step towards the quantitative classification of the measured maps. X45 1988 World Congress on Medical Physics and Biomedical Engineering San Antonio, Texas 325 Biomedical Engineering Scientific Papers

8E12-R.4 Computer Simulation of Human's BE12-R.5 Cardiac Activation Sequence, Computer Model of an Inhomogeneous Yang Jihai and Shen Pengliri, Human Torso, Sui Wenquan and Shen University of Science and Tech- Feng-lin,University of Science and nology of China, Hefei, Anhui, Technology of China,Hefei,Anhui, The People's Republic of China People's Republic of China Based on the atlas of sectionalhuman Based on anatomy data and bioelectric char- torso anatomy,a three dimensionalcom- acter of the human heart, simulation of puter model of human torso including activation sequence is accomplishedon a four cavities (LV,RV,LA,RA)of heart, microcomputer. The heart is composed of two lobes of lung and body surface and 5018 cell units (cuboid 5.12mmx2.26mmx a three dimensional myocardial model 2.22mm). Three different kinds ofcells has are both introduced in this paper.The been considered: cells of atrium,ventricle torso model dipicts the structures and and conduction system. The conductionspee- appropriate proportion of the relative ds are assumed to be 50 cm/s, 100cm/sand organs inside human body,especially of 150cm/s accordingly. Theexcitation start- the heart,with more than lo,000 surface points in ventricles are chosen atfour triangles.The myocardial model,with sites, corresponding to Durrer's experi- 4,632 elements including muscleof atria mental observations. The simulationresult and ventricle will be usedin the over- is qualitatively similar to Durrer's. all simulation study ofhuman cardiac Program possesses enough flexibilitz for electrical activities.Algorithmsfor the any necessary changing in the simulation. model construction,offeringthe flexi- Some abnormal excitationprocesses are bility for any change of the simulated. parameters of the models are given in detail.Metal- while the validity of the models for the purpose of simulation research of electrocardiogram is discussed.Other possible applications of the modelsare searched either.

MU-Lb An Inverse Electrocardiographic BE12-R.7 Solution by Using Boundary Element Recording and Analysis of Ven- Method,You-shan Yang:4ei-xue Lu, tricular Late Potentials on the Zhejiang University,Hangzhou,P.R.Cnina Body Surface. Shao QingyuWiang Ping, Shao Janhua, Zhu Xinglei, Chen Kong. Based on the Laplace differentialequa- tion and the realistic torso geometricmo- del,the Boundary Element Mathod(BEM)is Ventricular late potentials (LPs) were programmed to compute the electricalac- recorded from the body surface in 105 tive potentials of thehuman's heart du- patients.50% of those with myocardial ring the QRS period from potentialsmea- infarction and aneurism showed LPs.and sured on the human's body surface.The 15-16% of patients with PVCs oran- group of X-CT tomographies containingthe gina showed LPs. Spectrum analysis whole heart was used to buildtorso geo- showed there was a peak between 70 metric model which is re;ated to 3EMal- HZ and 140 HZ in frequency spectrum gorithm.The effect on the interfaceof histogram of every patient with LPs. the lung and the thoratic cavityis con- Two different bipolar precordial re- sidered in the algorithm.The bodysurface cordings were made (v4-v6,v4-v9).The potentials are collected by theappratus ECG signals was amplified, 50HZ trap- called The Body Surface PotentialMapping ped and high-pass filtered, 70HZ low- System.This appratus is consistsof IBM- frequency cutoff was chosen. The top PC computer, sample &communicationcon- of the R wave was used as a time re- troler,multi-analog amplifier andelec- ference. QRS template and QRS area trode vest.The results support thepossi- algorithms were made to reject PVCs bility of using BEM algorithmana rea- or fales QRSs. Averaging of more than listic torso model to determine thepo- 150 beats LPs could berecorded.To tentials of the heart fromtne body sur- eleminate ring effect bidirectional face potentials. digital filtering was made. After data processing 12 parameters could be gotten for analysis. 3 4

326 1988 World Congress on Medical Physics and BiomedicalEngineeringSan Antonio, Texas Biomedical Engineering Scientific Papers

BE12-S . 1_ BE12-S.2 The Application of the Kharhunen-Loeve Transform to Spectral Analysis of EEG-A Temporal Sequences of Spatial EEG Images, M. Lazar, Z. Hartley Approach, P.S. Ravindra* Koles*, Department of Applied Sciences in Medicine, and Rajesh Prabhakaran, Regional University of Alberta, Edmonton Alberta, Canada, T6G 2G3 Engineering College - Trichy, India. The Kharhunen-Loeve transform (KIT) is a useful technique in analyzing the underlying structure in a set of multivariate Non-invasive techniques for the recording and data. The transform, when applied to image sequences, yields analysis of electrical phenomena of the brain a set of orthonormal basis images and a set of coefficients using ELECTROENCEPHALOGRAPH have gained large such that the mean squared error in reconstructing the image importance. The EEG waves recorded from the sequence with fewer than the optimal number of basis images scalp vary in frequency and amplitude. Some is a minimum. From the coefficient sequence related to each of the most important inferences from basis image the normalized power represented by each basis these recordings may be drawn from their spectra. image may be computed. The basis images once created may A manual inference is prone to biasing also be used to examine other image sequences by projecting towa- the new data onto the basis functions and measuring the rds most prominent components and being obli- normalized power values. vious of other components that may carry sub- The KLT has been applied to sequences of images stantial information. Spectrum Analysars representing the spatial activity of the EEG from healthy using FFT have been developed. volunteer subjects, where a 31 electrode montage has been A system that solves the same problem used to spatially sample the scalp voltage distribution with a using HARTLEY TRANSFORM may be used to obtain temporal sampling rate of 120cps. Applying the KLT to up to the COMPRESSED SPECTRAL ARRAY of the spectra. 2400 images (20s of data) indicates that typically 5 (out of The use of this transform decreases the analysis 31) basis images and the respective coefficients are required time drastically, compared to conventional to represent over 90% of the variance in the original images spectrum analysers. This feature will expedite with an average reconstruction correlation between the the analysis of EEG recordings where amount of original and reconstructed .mages of over 0.85. Moreover, data is very large. The mathematical basis to images from the same individual in the same cognitive state justify the use of this transform is given. A not used in creating the basis images could be projected onto system which incorporates a FAST HARTLEY TRANS- the basis functions with similar results, indicating the indeed FORM PROCESSOR has been designed. It is a real thebasis images represent common scalpvoltage time microprocessor based system. distributions.

BE12-S.3 BE12-S.4 A QuantitativeAnalysis of the Amplitude Hodu- Nonlinear Alpha Wave Response under the Feedback lation in theSpindle Frequency Band of Sleep Control of Weak EM Fields, Zhong Jicun*, P.O. Box EEG in ElderlySubjects. 422, Zhejiang University, Hangzhou, P.R. China U. Earcaro, E.Eonanni, F.Denoth, L. Murri, A general structure of an adaptive control system C. Navona, A. Starita*and A. Stefanini, Pisa, using MEFP is given. By this control, a synergetic Italy system car. evolve into a more highly ordered state. We apply the theory to the induction of high level The amplitude modulation of the 12.5-14.5 Hz Alpha waves of human EEG. To this end, a microprocessor based real-time feedback control system was band activity of sleep EEG was studied by the built. computation of three parameters: the modula- Fifty seven subjects were tested and 86% showed Ln tion index (A), the root-mean-square frequency obvious augment of their Alph2 activities during of the modulation (B) and the root-mean-square feedback. frequency of the first derivative of the modu- Three features of the Alpha wave response have been lation (C). An investigation carried out on 6 found: 1) EM feedback will be effective only when the subject is quiescent, relaxing and eye-closed, elderly subjects showed that: a) for 5 out of adn the response can be easily inhibited by any them, themean values of each of the three arousal. 2) Only when the feedback signal is with- parameters were ableto discriminate between in a "window" of optimal parameters can the Alpha NREM and REM sleep; b) no significant shift waves be mostly activated. 3) Unusual subjective was fou, i betweenthe elderly groupand a feelings of relaxing were found in subjects with Control young group, for both the REM and NREM good response. mean values of each of the parameters.

347

1985 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas 327 Biomedical Engineering Scientific Papers

BE12-S.5 BE12-S.6 40 Hz (36-42 Hz) and SensoryMotor Rhythms The Stare Analysing Methodfor EEG, Luo Jian*, (SMR) (11-15 Hz) as EEGState Variables, C.C. Turbes* and G.T. Schneider, Zheng Qingliang, Inst. ofElec. Eng., Conqing Creighton University, University, Chongqing, China, Omaha, NE 68178. District People's Hospital of Shapingha, Congqing,China. There is evid,Jnce that certainbrain rhythms EEG signal directly reflectthe varying process of may signal the occurrence of differentbrain the cerebral mental andphysiological state, so the states. Support for this hypothesismay be nonstationary character ofEEG signal is closely provided by evidence that processingof information relative to the varyingprocess of the cerebral is changed during different state. brain states. During a relatively stablecerebral state The prestimulus period proceedingthe auditory EEG signal displays onapproximate stationary signal segment. evoked potential (AER) is convertedto a power While the varying of thecerebral spectral estimate. These power spectra are state occur EEG signal becomesa nonstationary sig- used to sort for the AER whencertain power nal, and the nonstationarydegree is closely rela- values for the (36-42 Hz) andthe (11-15 Hz) tive to the varying degreeof the cerebral state. frequencies are reached in theprestimulus By means of the anlysis ofthe adaptive stationary period of the sensory motor cortex (11-15 segmentation and further processingin every sta- Hz) and the amygdala (36-42 Hz). Attempts tionary segment, we obtaina method which can dis- are made to measure the performance ofthe play the varying processof cerebral state, called 40 Hz and SMR estimators. This relates to state analysing method forEEG signal can apply to the spread of their probability clinic and neurophysiological density functions research, or to over a given data sampled. The measure of thinking science andpsychometry. For example this spread is the variance ofthese estimators. while object's internalfactor and external factor These studies are made in thecerebral cortical is specified, the varyingprocess of 11-.3 cerebral and subcortical brain regions incats. Cross state will appear with regularity,by which we can correlation, coherence and phasespectral distinguish abnormal from normal. methods are used to examine theinteraction of certain brain regions.

BE12-S.7 An Application of Event Related Brain BE12-S.8 Potentials (ERP) for Testing the Abilityto Criterion for Using Adaptive Filterto Improve the Recognize Dot Patterns, H.Goto* andT.Adachi, Signal to Noise Ratio of AveragedEvoked Response Tokyo Metropolitan Police Hospital, Tokyo102, and T.Utsunomiya, Science Universityof Tokyo, Miriam Furst; Department ofElectronic Systems, Faculty of Noda 278, Engineering, Tel Aviv University, TelAviv, Israel

ERPs were investigated during judgementof match- mismatch of two successive dotpatterns. A given The amplitude of Brain EvokedResponses (BER) varies from trial comprised, sequentially,a variable number tenths of a microvolt to tenth ofmicrovolts and is embedded (2 to 4) of random dot displays(RDD),a first 5- in the ongoing ele,:troencephalogram(EEG). This response is dot- pattern (5DP), a variable number (2to 4) of usually obtained by averaginghundreds to thousands of RDDs, a second 5DP, then two RDDs. EachRDD or 5DP repeated recordings followingan identical stimulus. The was 20 ms in duration. A RDD or 5DP was presented average (ABER) is the best estimate for BERwhen the noise in a trial at a rate of oneevery one second. 144 is random of a zero mean. If BERand the noise (EEG) have trials were presented in randomspace and sequence. different spectrums then it is possibleto reduce the signal to The 2nd 5DP was identical to,or mirror-image of, noise ratio of ABER. Severalattempts were done to improve or neutral (neither identical noz mirror-image)to ABER by using a Wiener filter whoseparameters were the 1st 5DP. 5 normal subjectswere told in advance obtained from the averagedresponse. In many cases this to press the pertinent switchas fast as possible process did not improve the signal to noiseratio. We suggest after each trial to signify whether the2nd 5DP was to apply an optimum adaptive lineartime-invariant filter on identical to, or mirror-image of,or neutral to the ABER which yields OBER.The optimal filter impulse 1st 5DP. EEGs in parietal and temporalleads were response ih.ln...h0 ...hmf is derived whileassuming that averaged by a microprocessor. In examinationsof e- ABER is the expected BER, andm is the filter's order. Let yoked potentials to the 2nd 5DPs,P650was observed. define the mean square errors of ABERand OBER as al and a! The peak latency of P650 to mirror-imagewas signi- respectively, it will be shown that if thenoise is stationary ficantly larger than that of P650 toidentical and and ergodic then 4, = of, ght. It is therefore clear that the neutral 5DP. These results suggestthat it takes a signal to noise ratio of ABEltcan be improved by applying longer time to recogn.ze mirror-imagethan to judge the suggested adaptive filter only if 1.hf< I. identical and neutral patterns, and thatERPs can be useful for testing the ability to recognizedot patterns. 4-4 x.4: 8 328 1988 World Congress on Medical Physics andBiomedical EngineeringSan Antonio, Texas

, Biomedical Engineering Scientific Papers

BE12-S.9 BE12-S.10 REAL-TIME DETECTION OF EEG ARTIFACTS A Small Sxze System of AutomaticDetection for C. Munoz Gamboa* and E. Saldivar Safety Driving Based on Electroencephalogram,F. Univ. Autonoma Metropolitana - Iztapalapa Yano, H. Kodama, S. P. Ninomija*, AoyamaGakuin 09340Mexico D.F., MEXICO University, Tokyo, K. Idogawa, SeitokuGakuen College for Women, Chiba, K. Sato,Bioinformation A new algorithm for real-time detection of EEG Laboratory, Tokyo, Japan artifacts is presented. Thealgorithmwas developed through a newclasificationof EEG We developed a practical system toprevent vehicle artifacts in three groups: accidents resulting frpm driver'sdosing. On 1. Frequencydistinguishableartifacts, like limited access highways in long distancedriving muscular movements. it is difficult for the driverto maintain a 2. Artifacts due to the acquisition technique, watchful driving condition because ofthe state of such as ECG reception. monotony resulting from receiving less stimuli. 3. Artifacts having frequency components in the It is recognized that a burst ofslow alpha rhythm EEG band, and appearing in spatial patterns. occurs on the EEG when cerebrum activitygoes down Inthis group are: loose electrodes, false because of less tensionor the tendency to contacts, ocular movements, and blinking. drowsiness. Far safety driving, itis important to Artifacts belonging to the third group are the detect the drowsy condition. EEGdata were most interesting because they appear in most EEG collected from driving experimentson the highway. recordings.The others are rejected by digital The detection of slow alpha rhythmwas analyzed by filtering or by means of a goodacquisition statistical method as follows : technique. We compute regression line andmoving variance by moving least square method. The algorithm is based upon the processingof We defined slow alpha rhythm which satisfied some conditions three parameters: amplitude and time thresholds, about -lope and variance. and thespatial patterns of the events.These This method possesses stability respect artifact, as parameters, obtained and processed in real time, it were, ECG, EMG, alternating current waves and areenough to detect every artifact occurrence eye blinking. By this method, we developed online real-time from group three. Furtherprocessing, also in system by 16 bits personal computer. real time, permits the total recognition of this As the result, accurate detection type of artifacts, with a reliability of 97.62. of slow alpha rhythm on an on-line real-time methodhas become possible.

BEM-S.11 BE12 -S.12 SOME PHYSIOLOGICAL CHANGES IN INDUCED Estimation Algorithm for Spread Forms of MEDITATIVE STATE (a case study), Spindle Waves in Multi-channel EEG, V.Lepicovskal J.Mrkvicka, C.Dostalek, Yoshikazu Ueda*, Naohiro Ishii, H.Beran, Inst.of Physiol.Regul., C.S. Nagoya Institute of Technology, Showa-ku, Acad.of Sci., Prague, Czechoslovakia, Nagoya 466, Japan. S.Vishnudevananda,Phys.lab.,Sivananda Yoga Centre, Val Morin, Canada. Spindle waves are the characteristic signals in electoroencepharogram, but have been unknown in Constituents of ancient methods of their spread form. We present an algorithmfor concentration (used a.o. in rituals) estimating the spread form of spindlewaves in have been included with a desirable multi-channel data. Our algorithm is ableto cal- effects into modern medicine, especially culate within two phasic dominants and the spread in functional diseases, enabling thus direction of spindles on a head. Inour algorithm, the patients improve the quality of life the remarkable frequency componentsare extracted , :ithout side-effects. In this case study from spindles. Second, the frequency-wavenumber a very marked inhibition of pain was spectral of each component is analyzed in partial reached at. field. The last, the spread form of the spindle Measurements of EEG, ECG, blood wave is estimated by studying these partialwave- pressure and respiration was taken in numbers geometrically and synthetically in whole the course of the shivaistic ceremony field. As applying our method to EEG data,we got called Kaavadi, in which 108 spears were two results. One is that the phcsic dominant of inserted into the tongue, cheeks and spindle with a low frequency component exists in chest. The EEG de,dctivation appeared to the central field of the head and witha high be on the background of autonomic component in the parietal one. The other is that activation. Maximal increase of spectral the most of spindles spread fromrear part to power (300 percent) was measured in the frontal one on the head. alpha range. Maximal changes in theta and delta range did not reach post hyperventilation increase. 34 9

1988 World Congress on Medical Physics and Biomedical EngineeringoSan Antonio, Texas 329 Biomedical Engineering Scientific Papers

BEI2-T.1 BE12-T.2 ECGxpert, an expert system for EKG analysis, S. Jenichen Gieseler*and VERTIGO: an expert system for the classification ofdizziness A. Buizza, G. Magenes: M. Manfrin, E. Mira, R. N. Wiederhecker, Programa de Eng. Schmid, P. Za- Biomedica/COPPE,CP: 68510, Rio de nocco - University of Pavia, Italy Janeiro, Brasil, CEP 21945 A consultation model for the differentialclassification of dizziness has been developed by means of the shell The EKG is a biological signalwith a EXPERT host of information and therefore of (Weiss and Kulikowski, 1979). More than 30types of vestibular complex analysis. ECGxpert isa expert disorders characterized by the presence of dizzinessas the system designed to aid physicians, with major symptom are considered. The knowledgebase has been few experience in EKG analysis, and synthesized in about 350 production rules. Oatahave been students in training programs. Instead limited to patient history as the assessment ofthe diagnostic of acquiring and interpretingthe EKG , accuracy that can be reached by careful sdaluation of the system request from physician case those parameters which are used in history is one of the goals of the research. routine analysis by cardiologists. VERTIGO is able to collect patient data bymeans of a flexible The system, implemented in PROLOG,i.ro- questionnaire reproducing the strategy followedby a skillfull vides a menu-driven user interface and clinician in history taking. an explanation system, allowing to add The diagnostic corvlusions are proposed with certainty measures toanswers, simula- a certainty ting an inexact reasoning. The knowle- factor representing their plausibility. Thereasoning followed dge base contains the expert'sknowle- by the system to reach a certain diagnosticconclusion can be dgeabout EKG interpretation in form explained to the user in a pseudo-naturallanguage which of production rules. As final result illustrates the production rules activatedfor that conclusion. the system generates a complete report Two versions of Vertigo have been realized,one running on a about analysis including advices on future procedures. VAX computer, the other on IBM PC. The modelis being validated by human experts belonging to the internationalcommunity of ORL clinicians, who were not involved inbuilding VERTIGO.

BU' .3 BE12-T .4 An Expert System of Medical Diagnosis of Decision support system for treatment selection in Icteric Patients, J.C. Nievola, W.C. de Lima*, oncology. S. Hyodynmaa*, P. Kolari, K. Narilinen, C.I. Zanchin & W. Dantas, Universidade A. Ojala, J. Rantanen and N. Saranummi, Technical Federal de Santa Catarina, P.O. Box 476, Research Centre of Finland, Medical Engineering 88.049 Florian6polis, SC, Brazil, Laboratory, Box 316, SF-33101 Tampere, FINLAND Telex 482 240 UFSC BR The scope of this work is to implement an expert system to aid the gastroenterology experts in the Decision support can be utilized at several stages in diagnosis of icteric patients. The system is oncology. We are developing a system for three stages: composed of a knowledge *Jase andan inference therapy decision making, design of treatment protocols engine. The knowledge base is basedon themajor and analysis of therapy results. A prototypeis points in which the expert develops his work.Then, functioning for the treatment decision making in this knowledge is used foran inference engine inoperable lung cancer. The object oriented system has that wants to establish the possibility to each been implemented on a Xerox 1186 workstation with one of the diseases considered. This knowledge Interlisp-D and LOOPS. The input information required base has 74 ruler, ocuppying 17,3 Kbytes. The covers patient ID, diagnosis, symptoms, examination and inference engine, with 103 Kbytes, is implemented test results, tumor data and staging. During the in Turbo Prolog': language, in a PCcompatible reasoning the system checks the correctness of the microcomputer. It has two phases: a preliminary diagnosis, checks if there are indications for immediate diagnostic, and a final response. Thesystem has treatment and checks if the intention of therapy an initial menu, where the user chooses one of the proposed by the user is relevant. The system has been following aspects: Ceneral ,Information, Load tested with cases of lung cancer patients anda panel of Knowledge Base, Edit Knowledge Base, Consult and oncology experts has evaluated the principles of using Exit. In the consult mode, the user answers the system and the human computer interface. The final questions about the patient history, his physical aim of the work is a decision support systemfor conditions and laboratorial examinations. After oncology. The next step will be the building ofan this, a ranking of possible diseases is obtained, oncological clinical register for the system.The based on the idea of fuzzy set theory. The integration of the system components azan oncological result information network is also essential. of the consult is analised by theexpert physician before treatment is initiated.

330 1988 World Congress on Medical Physics andBiomedical EngineeringSan Antonio, tas Biomedical Engineering Scientific Papers

BE12-T.5 BE12-T.6 A Syntactic Analysis of the Disease Classification Model Based on The Auditory Brain Stem Responses, Hierarchical Evaluation of a Symptom Hironori Koyama*, *T.SEKIYA, A.WATANABE, S.KATO,M.ARUGA and Tokyo Denki University,2-2,KandaniShiki, M.KIKUTI Chiyoda, Tbkyo, Japan 101 3-2 Namiki, Tokorozawa, Saitama 359 JAPAN

A syntactic algorithm has been developed toau- It is important for the decision making in medical tomatically detect and identify the wavecompo- diagne-os and treatment to classify the diseases. nents of the auditory brain stem respanse(ABR). It is not sufficiently carried out by using merely The algorithm for the wave detection is describ- the statistical method. We reviewed what is done ed using an attribute grammar. by medical doctors in their decision making to A typical ABS:pattern is composed of seven sub- find a bearing striking resemblance to their men- patterns, namely Wave I, Wave II,..and Wave VII. tal model. We propose a new model which has the After normalization and piecewise linear approxi- flexible algorithm for compensating missing symp- mation, an ABR is represented in the form ofa toms, taking into account the degree of contribu- sequence of triplets. Each triplet consists of tion of each symptom to diagnosis. a slope segment symbol and two attribute values, This model adopts the time series representation the amplitude and latency of the slope. Thena of every symptom such as threshold element, inte- triplet sequence is parsed according to the at- gral element, differential element and so on. The tribute grammar devised for-thewave detection. concept of the symptom weight evaluations contrib- For the detected waves, a.heuristic algorithm utes to the compensation of the missing symptom based upon a priori knowledge on latencies is and provides the flexible algorithm for the clas- applied to assign each wave to one of thesev- sification of medical data. Furthermore, this en candidates. model is able to utilize to deeper reasoning, because it is corresponling in all of n physics A preliminary test has shown that 90 to 95% of model. the waves are detected and identifiedsuccess- fully by the algorithms, including thecase that some of the wave components are missing in ABR patterns.

BE12-1.1.1 BE12-U.2 FIELD THEORY OF FLECTROGLOTTOGRAPHY AND A New Method for the Vocal Cord MEASUREMENT FOR PHONATORY BIOMECHANICS Disease Diagnosis, R.C.Chen;Q.Ling, Wan Mingxi* Cheng Jinzhi Y.Z.Jing Dept of Information and Control Engineering Xi'an Jiaotong University, Xi'an, The traditional method to diagnose the vo- The People's Republic of China cal cords disease of a speaker is through Electroglottography (EGG) is more convenient to the observation with the laryngoscope or be a monitoring method of laryngeal function than making mediate dignosis according to the high-speed films, photoglottographys ultrasonic spectrogram of the speaker. But one prob- glottography and glottography by inverse filtering lem is that it is not easy to observe the acoustic or laryngeal ail flow signal, due to non- vocal cords with the laryngoscope as well invasion and low cost. However, what does the EGG as to record the results in a simple and waveform tell us about phonation is ambiguous. objective way. Another is that the spect- This study was designed to provide the interpreta- rogram is a mediate information about the tion of the EGG waveform based on electrical field vocal cords disease and not of much value theory and vocal folds activity. This paper also to diagnosis. In this paper the method of introduces a two channel EGG system with a multi- linear prediction analysis on the short- ple electrodes array, which makes simultaneous time interval for glottal inverse filter- investigation of vertical and horizontal vibration ing from the acoustic speech waveform is of vocal folds possible, and approaches the origin discussed and a new way approaching to the and sensitivity of EGG signal. vocal cords disease diagnosis is presented.

351

1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas 331 Biomedical Engineering ScientificPapers

BE12-U.3 BE12-U.4 A Speech Onset Analyzing System, DEVELOPMENT AND APPLICATION OF PROGRAMMABLE B.F. Smith; R.M. Kroll AUDITORY STIMULATOR Clarke Institute of Psychiatry, Jia Rui Lin* Lie Yi Fang Han Shan Zhu 250 College St. Toronto, Ont., Bioengineering Department Canada M5T 1R3 Huazhong University of Science andTechnology Wuhan, China Carefull study indicates that, for themost part,disfluencies in the speech of stutterers The paper introduces a programmable auditorystim- are caused by the application of abmpt,exces- ulator for human brain stem AuditoryEvoked Response (AMP) tests. sive tensioning of the vocal folds duringthe It can generate 3 modes, aboutmore initiation of phonation. than 60 kinds of waveform, accodingto require, they The device described in thispaper, a speech can be transmited to the ears Li earphonesseparately. onset analyzer (SOA), monitors theharshness The system was consisted ofa small system by of speech onset and providesinstantaneous Intel 8085A microprocessor, where thedata of vari- positive feedback when subjectsachieve their ous waveform are stored 2 chips of EPROMof system. targets. A built in pair of countersrecord The mode, frequency or period ofthe signals can be the number of attempts and the numberof cor- changed with the states ofthe switch in the front rect responses so the unit may be usedin the of panel are recognized by theprogram. A tone abscence of a therapist. The countscannot be burst is produced by a voltagecontrol amplifier altered by the subject. Voice onsetsmay be (VCA), the envolopes are supplied bythe programmable signals. monitored for single syllabic utterancesor And white noise generator was obtain- for continuous speech. ed by a cheap electronic circuit. In addition, a The V is battery onerated, smalland light slide switch of the stimulus polarityselection can weight. The use of filters gives equalresponse determine signals of the positive,negative or al- to both male and female voices. ternate polarity. The SOA eliminates humanerror in the The programmable auditory stimulatorhas been applied for the research of the human extremely difficult task of judging theade- brain stem AERP quacy of responses at several levels because and has obtained satisfying results. It is also suitable for many other applications. it is reliably calibrated to uniformstandards.

BE12-U.5 BE12-U.6 The Frequency Domain Feature Analysis of Coherence Spectral Studies of Averaged Brain Brainstem Auditory Evoked Potential, Ma Xiao-Bo,Gao Shang- Kai,Zhou Li-Gao, Evoked Potentials During the Action of Amphetamine, Dept. of E. E.,Tsinghua University, C.C. Turbes* and G.T. Schneider, Creighton Beijing,Chima,Pei Hong-En,Jiang Wei, University, Omaha, Nebraska68178. General Navy Bospital,Beijing,China. These studies are made on cats and are concerned Brainstem Auditory Evoked Potential (BAEP) which with power and coherence estimates of auditory has important significance in objective audiome- evoked potentials at cerebral, cortical and try and diagnosis of acoustic neurinoma has been subcortical regions. Certain chemical agents widely used in clinic.The parameters of BAEP,now are used to alter the background EEG activity mainly the time domain parameters, sometimesare and auditory evoked potentials. In these not sufficient to make a diagnosis. Andusually studies the d and 1 isomers of amphetamine wave IV and V in time domain are not easy to be are used. The analog data is collected and separated.lherefore we hope to extract new parameters of BAEP for clinical use. stored on FM tape for computer processing. The data is processed using power, coherence, This paper developed the features infrequency domain of the BAEP of guinea pigs which under- partial coherence and phase spectral analyses. went thesimulated nerve tumor operation by us- Comparisons are made of the pre, during, and ing autoregressive model and spectra estimation, post amphetamine data of both EEG and auditory and some results were applied to BAEP diagnosis evoked potential activity. There are changes of human patients.Experiments indicated that the in the autospectral, coherence, and partial peaks of spectra of BAEP of guinea pigshaving coherence spectra that relate to the action lesions will move toward lower frequency inter- of amphetamine.Special consideration is val ;hen that of normal guinea pigs. Thesumof phase angle differences of bilateral poles 0 and given to these data estimates and then interpreted the difference of normalizedpredictionerrors as it relates to the interaction of brain of bilateral AR models arepresented as new pa- regions during the various drug induced patterns rameters of BAEP for clinical diagnosis. of behavior.

332 1988 World Congress on MedicalPhysics and Biomedical Engineering0 San Antonio, Texas Biomedical Engineering Scientific Papers

BE12-U..7 BE:12-V.1 QUANTITATIVEANALYSIS FOR EXPERIMENTALLY INDUCED Quantative Analysis of Influ- PERILYMPhATICFISTULABY PRESSURE MONITORINGAND CONTROLLING SYSTEM, ence of Various Factors on the T.HORIUCHI*, T.DOHI, Detection of Details in Roentge- T.MASUZAWA, T.FUJIKI, I.SAKUMA, T.FUTAKI, M.HARA and Y.NOMURA, nological Image, A.M.Gurvice,Mos- cow Research Institute for Roent- The Univ.of Tokyo, 7-3Hongo, Tokyo, Japan genology a. Radiology,117837,Mos- cow,GSP-7,Profsoyuznaja,86 The cause of perilymphatic fistula, a subgroup of sudden deafness, has not been clearly understood. Calculation of details detection on the To study its etiology quantitatively, the innerear pressure monitoring systemwith image can play an important role in tur- a computer ning roentgenology into quantative scien- controlled pressurizing device and three types of low pressure monitoring sensors was constructed and ce. When calculating it's necessary to utilized in ex vivo experiments using white take into consideration the fact that the guinea threshold contrast is determined by the pigs. Hypothesizing that cerebrospinal pressure increases cause arupture of cumulative influence of the two main noi- the roundwindow se components, one of which depends on membrane, direct or indirect pressurization to cerebrospinal region and simultanious the instrumental part of the image for- measurement ming system and the other - on the eye. of innerearand cerebrospinal pressure were carried out. The passiihtlity cf this calculation can be illustrated by means of: 1) determi- The ruptures of the round windowmembranes were seen in the experiments with direct saline infusion nation of the permissible limits in increase of radiation sensitivity of the into the cochlea when the increment of inner ear pressure were above 370 mmH.0. There was no sign of image transformers; 2) comparative analy- the ruptures when squeezing"abdomen was employed sis of various methods and technical meas ans of radiodiagnostic investigation; 3) an indirect pressurization although cerebrospinal pressure increased. evaluation of the influence of morfolo- gical noise; 4) choice of physico-tech- It is conclusive that this system is feasible to analyzepressure changes in the nical conditions end image quality cont- inner ear and rol methods. The sign of optimal aligne- cerebrospinal region simultaniously and useful to study on the etiology of perilymphatic fistula. ment of informativity and dose is the balance of all the main components of noise

BE12-V.2 BE12-V.3 3-dimensional Superimposed Description of Computed Color Imaging of Gastric Thallium -201 Myocardial Scintigram and Mucosal Hemodynamics. S.kawano: S.Tsuji, Coronary Arteriogram, T.Akatsuka*, N.Sato, T.Kamada. Dept. of Med., Osaka T.Takeda#, S.Natsubori, T.Nakajima, Univ. Med. Sch., Osaka 553, Japan. Y.Tamura, and M.Akisada#, Yamagata The microcirculatory distrubance was University, Yonezawa 992 and #University emphasized to play an important role in of Tsukuba, Japan the.gastric ulceration. In this study, the compater-assisted imaging of gastric To evaluate ischemic heart disease(IHD), mucosal hemodynamics was performed. The thallium-201 myocardial scintigram(TMS) and reflectance spectrophotometry system was coronary arteriogram(CAG) are the most important applied to the human gastric mucosa under images for the diagnosis. By using the single endoscopy as reported previously(N.Sato, photon emission CT, 3-D reconstruction of left Gastroenterol. 1979). From the data at 20 ventricle can be performed, and CAG can describe points in the stomach, the computer- the structures of coronary artery clearly. To assisted color image of mucosal blood understand the details of the regional distribution was made. Moreover, to myocardial ischemiaand stenotic coronary obtain the mucosal blood distribution in artery, it must be very useful to display the 3- real time, the electronic endoscope was dimensional ofTMS and CAG. At first, the directly coupled with the image analyzer experimentwas performed on the arterial model and microcomputer. The logarithmic ratio and actual TMS, now actual TMS and CAG are in (logVr/Vg) of the videosignal in R ana G use. Twoimagesare adjustedas the left images was calculated with image analyzer descendingcoronary artery is lying 15degree and computer. The ratio at a region of right from the center of the anteriorwall on interest was proven to have a good TMS. The relation between the degree and correlation with Hb concentration of extension of myocardial ischemia and the degree mucosa measured by reflectance spectro- and position of stenotic artery might be well photometry. With the methods, the gastric assessed by this 3-D superimposed description. mucosal hemodynamic changes were investigated in patients with gastric ulcers. By the imaging of gastric mucosal hemodynamic t was possible to predict the ulcer formation and healing. 353 1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas 333 Biomedical Engineering Scientific Papers

BEI2-V.4 l3E12-V.5 AUTOMATIC DETECTION OFCORONARY ARTERY WALL CONTOURS IN PATHOLOGICSLIDES. A Digital Processing Method to Obtain SeperatedBone and Soft L Moura and R Kitney: Tissue Pictures by Using of Dual EnergyX-ray Projections Dept of Electrical Engineering, Imperial College, London SW7 ?BT, UK. Boge Wu , Datong Jiang $ Institute of Biomedical Engineering The paper describes an imageprocessing technique for Xi'an Jiaotong University Xi'an , China the detection andlabellingof arterial wallcontoursin images obtained fromcross-sectional slices.Agradient The difference of X-ray attenuation coefficient operatoris appliedto the image and generates points of between bone and soft tissue dependents on the X-ray interest at likely, edges. These pointsare then used as starting energy being discussed, based on this relationship our method called Gray Level points by a contour followeralgorithm leading to the Piece-wise Adjustment Step-wise Subtraction was developed. Perform formation of one-pixel-wide lines, calledsegments. Heuristic a specific digital ope- search is used to locate closedsetsofsegments which ration on gray level of each pixel ofa subject's two X-ray projec- minimize a cost functionbasedonthe ratio between a tion pictures of different energy,so that for every corresponding pixel pair, in the total contour's squareperimeter and the enclosedarea. This gray level, component contributed by soft function tissue can be adjusted to the isminimized by a circle so thatthealgorithm same, then by subtraction of the two tends to find circle-like structures.Constraints imposed on Pictures, all soft tissue information is removedand a mask picture that contains the information of bone segment connection avert the formation ofinconsistent shapes contents distribution only is and also speed upprogramme execution. Further processing obtained. Subtract this no(once by once from the original low includes closed energy picture which contains the most soft tissue curvefilteringusing leastsquarefitting; information until and bone image disappeared and the remainder the labellingof contours according to their positionin become a soft tissue only the picture. Using X-ray films / TV image and to the gradient direction(inward/outward) camera, experiment according to this along the contours. wothod was completed on a personalcomputer equiped with special image processing hardware. In X-rayprojection we use copper and tin filter to emphasize high and log Results have shown that the method'sstrength lies in its energy contents respectively. Some ability to correctlyfindthe wall contours even wren major successful pictures with soft tissue onlyfor bone-soft tissue gaps Phantoms have been obtained through above are present,provided that some long segmentsare operations. he hope this method can be implemented in formed. The method issuitablefor locating and labelling a X-ray, 1.I., TV system to construct round-shapedstructures whichfrequently occur in Biology, a new low cost processing system and it willbe also useful for angiography with Iodine contrast median such as blood vessel cross-sections andcell outlines. .

BE12-V.6 BE1.2.-V.7 Data Compiession forMedical PACS, V.A. A Method for Continuous Measurement Pollak *, D. Meyer-Ebrecht, V. Krybus, Univ. of Width and Area in Medical of Saskatchewan*,Saskatoon,Sank, Canada; Images, Technical Univ. of Aachen, Aachen, Germany. Wang ZhoogRui: Dept. ofScientific Instrumentation,Zhejiang University, PACS is an acronym standing for "Picture Archiving Hangzhou, China. and Communication Systems." Imaging devices representa large and increasing proportion of A method for continuousmeasurement of modern medical technology. The pictures acquired vessels diameter from TV imageis pre- bythese devices are an essential part of the sented. It is based on PDM(Pulse-Dura- medical historyof the patient and a large tion Modulation) and DPDM(Demodulation proportion of them needs to be preserved, of PDM) principles, An auto-tracking frequently together with verbal comments for window is formed on TVscreen to cover future consultation. Traditionallythe carrier the segment of vessel to bemeasured. medium for pictorial information was photographic The window signal isa series of pulses film. Storageof film especially in large with the duration proportionalto the quantities requires space and retrieval is slow width of the vessel andthe period and error prone. This is where PACS comes into equal to the interval of play. a line swee- There the pictorial information is stored ping. We use a low-passfilter to demo- electronically, preferably in digital form. For dulate it. The output widthwaveform retrieval the stored information is accessed is transfered to electronically and transmitted to the place where a microcomputer through A/D converter frame byframe. Then it is needed by standard communication facilities. the width waveform of In the majority of cases the distances involved a vessel in a are relatively frame of image or thevariant width at short, but even very large a site we interested in distances can easily be bridged. The number of may be sent to a chart recorder by D/A pictures annually acquired in a mediumsize conversion. It hospital can go into the tensof thousands. is demonstrated thatthe method may be Billions of bytes are necessary to store and to applied to measure the widthand area transmitthis volume of data. Data compression of other medicalimages of any shape seems to be the most effective way to handle this s lghtly wobbling. problem. The featuresof adata compression system specifically developed for medical images are described in brief. tit Q."

334 1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas Biomedical Engineering Scientific Papers

BE12-W.1 BE12-W.2 Signal Detecting Technique of he Fourier Analysis of Averaged Signals modynamics, D.X. Zhang: Nanchang A. Fadda, M. Neroni: IstitutoSuperiore Hemodynamics Institute, Nanchang, di Sanita', Via Regina Elena 299 00161 Jiangxi, China Roma, Italy.

To use an information, processing system When harmonic analysis of averaged with pressure and volume wave transducers signals is required the following in diagnosis is the technique, a biomedi- considerations may be useful. The cal engineering and combination of tradi- averagingprocess can be seen, in the tional chinese a western medicine, Based frequency domain, as a comb filter with on linear systematic theory and physio-a- majorlobesat frequenciesequals to natomic state, blood circulation system k/T, where k is an integer and T is the was divided into several units on each of trigger period. As thenumber of which, we established relavent mathematic averaged records increases the main lobe expression which reproduced an in vivo becomes narrower and the sidelobes waveform very closely kr..0.99). Thus the become morelittle and numerous. For hemodynamics parameter equations in con- large number of records, the averaged trast. with invasive technique were derived, signal waweform is substantially reduced the detection of flow volume, pressure, to theperiodic component. So it is modulus, resistance etc. of circulation correct anduseful to make a Fourier system resulting in r=0.88-0.975. Transformof the signal using asimple Widely using the technique in the clinic rectangularwindow and an analysis field, we studied traditional chinese me- durationas short as the trigger period dicine and hemodynamics diagnosis, guard withno loss in resolutionor leakage of emergency and predicts of many deseases. problems. Weapplied this technique in For example, when diagnosing coronary de- the field of patternelectroretinografy cease by the technique and selective coro- obtaininga fast calculation of the 2nd nary arteriography, we got the diagnostic harmonic of averaged signals. coincidence ratio as 93%.

BE12-W.3 BE12-W.4 Adaptive Analysis of Harmonic A Modified Adaptive Filter for Bio- Oscillation for Biological Signals, logical .;igna/s.I.B. StambolievX, M.Mita°,T.Onoda and M.Hirano, M.P. Mintchev,I.B. Ratchev,Higher Iwate Medical University, Institute of MechanicalZ1Elecrical 020 Morioka, Japan. Engineering,Sofia-1156,Bulgaria. A mathematical method of discrete signal processing of harmonic oscillation in a The low frequency biological signal.* pro- finite window has been found. The signal cessing to obtain an important diagnostic is sampledand analyzed in the finite information is a main problem of electro- window which if regardedas thebasic period of an oscillation. Each instantan- gastrography.A filtering of the electro- eous value of harmonics is successively gastrograraphic(EGG) signal must be done given byfixed linear transformation of without affection on its frequency range,phase and amplitude. It is also ni-,ces- the arithmetic meansof the data from frequency sampling. By using this method, s,ifry a strong suppression of the noise fundamental frequency of an oscillation, components included or nonincluckld in the its harmonics and theirmutual phase signal frequency range.An optimization of relation can be estimated precisely. the sequrity and +legibility of wellknown This method can estimate the most adapt- modified adaptive filters is made and the ive basic period T of the signal where- device is completely computerized.liesults after time window is chosen accordingly. of simulated human EGG-signal processing The frequency resolution is determined are gi%en.The capabilities of digital mo- by the adaptivefundamental frequency dified adaptive filters using Fast Fourie and by the highest frequencycomponent Transform are discribed.The parameters of involed in the original signal. the filter provide the electrical activi- This method suppliesan algorithm for ty of the human stomachprocessing but designing electronic circuits and com- changing them the device application area puting programs. may be e,:tended.

35 5

1988 World Congress on Medical Physics and Biomedical Engineering San Antonio, Texas 335 Biomedical r:ngineering Scientific Papers

BE12-W.5 BE12-W.6 Data Compreasion for ECGand a Sigh Power-Spectrum Analysis of REG with Speed Circuit of DOT,Sheng-Bei Shan Prony Method,Liou XiaoxianglZhou Shouchang,and Chu Mingjie,Theretical and Wei-Xin lie;Dept. of Electrical Electrical Research Institute, Engineering, Northwest Telecommunications Chongqing University,Sichuan,PRC

Degineering Institute, Liam, PSC Rheoencephalograph (REG) has been widely applied to clinical diagnosis in our country.But up to now, most of themedical For the spacial wave of Electrocardiagram this units still employ thebackward manual paper discusses the data compression performance medthod to analydeREGbyviewing and measuring with ruler. Since thereare many of three kinds of orthogonal trommformations(DFT, kindsof waves in timedoMain, there are many kinds of distinguishing methods which DCT and DWT), shows that DOT has the beatperfo#. are not convenient in clinical diagnosis. mance among them, We fotmd some properties of So this paper adopts power-spectrumanaly- ECG data compression. Based on the propertiesom sis to overcome wavedifferences that the ages, sexes, and measurement etc. must propose a segmental processing method and an ad- result in error of REGof patients which may makesome aptive method for ECG data compression. Usingc- pathology characteristics disappear.Since REG is a kind of thelower ur methods, the compression performance can be frequencysignal and FFT has random undu- late in 0-12H:, it is not improved and distortion reduced. We alsoprese- suitable here. This paper probes a new time-domain-iden- nt a real-time 8-point DCT using the continuous tification method--Frony method and tryto use Marple and Covar valued logic hardware which are developedrecen- algorithm in estimation of the order and determinationof the tly-in China, The DCT is with a high speed eati- parameters. After deriving time-model of efying the medicine applications, RaG, it wo, oe transformed with Fourier Tr ens form. The experiments have showedthat the newmethod is significant,. andmay worth to be furthe.. investigated.

BE12-W.7 BE12-W.8 Spectral Analysis of Vertical Acceleration of AccurateEstimation of Parameters of the Body during Jumping. S.C. Su., K. Tsuchiya. Multiple DampedSinusoids inNoise & T.Tamura. T.Toyoshima and T.Togawa, Sch. of Sci. TheirApplication to TheAnalysis of & Eng., Waseda Univ.,and Tokyo Med. & DentUniv. Biomedical Signal . TANG YU* ZHEJIANG UNIVERSITY HANGZHOU PRC(CHINA) The vertical acceleration was analyzed by FourAer transform and powerspectrum during jumping was investigated. We improve the method,which Kumaresan and Tufts (KT) presented, for estimating Vertical accelerations on the backandthe the head were measured simultaneously with strain- damping factor and frequency of gauge accelerometers. Oneaccelerometer was multipleexponentially damped sinusoids placed at the back by elastic belt and theother Inadditive noise. we usebackward was firmly hold by teeth. Sevenmalesubjects predictor to separate the signal zeros were performed different jumping cycles (20 to from otherzeros and compute the 120 times/min by every 20 times). The waveforms backward estkmated value of parameters, were analyzed with Fourier analyzer. thenuse forward predictor to compute The results showed the frequency distribu- the forward estimated -alue of tions of the back and the headwere similarbut parameters. At last we use the meanof the magnitude of power spectra at the backwas 4 backward and forward estimated values as to 6 dB greater than that of the head for all the estimate of parameters. This method subjects and cycles. In the jumping cycle of100 wasdemonstrated to have a verygood to 120, the significant harmonics of the performance by numerical simulation, in fundamental frequency which corresponded to comparison to rriginal KT method and jumping cycle could be seen up to 25 Hz, and the Prony method. The estimate of damming fundamental frequency revealed the largest factor and frequency is magnitude. However, in jumping cycle below unbiased. Tha 80, variance of the estimates approches the harmonics could be seen only 10 Hz, the and the Cramer-Rao bounds. We give some examples second harmonics revealed the largestmagnitude. (VEP & PCG) to demonstrate their The differencecaused by the differenceof application to the jumpingmodes, i.e., one is intermittent, analysis of the biomedical signal. other is continuous. Frequency around 12 to15 Hz revealed the natural frequency of thebody.

336 1988 World Cor.gress on Medical Physics and Biomedical Engineering San Antonio, Texas Biomedical Engineering Scientific Papers

BE12-X.1 BE12-X.2 Accuracy enhancement in stereometric systems Stereotaxic System for Diagnostic and based on 1-D optoeletronic transducers. The:apeutic Procedures, T. Ribbe*, M. (°) A. Cappozzo, F. Gazzani, V. Macellari*. Berjstrom, J. Bodthius and T. Greitz. Istituto Superiore di Sanit5. - (°) Universi- Dept. of Medical Engineering, Karolinska lnstitutet, S-10401 Stockholm, Sweden. ty "La Sapienza". Roma, Italy. A head fixation system has been developed The accuracy of the optoelectronic system COSTEL enabling exact transfer of positions and of the associated stereometric model was asses between neuroradiological and therapeutic sed. The hardware is based on three 1-D cameras procedures. The key item of the system is composed of linear image sensors (2048 elements a base plate that can be rigidly attached to the patient's head, either by a plastic CCDs) and toroidal lenses, and the calibration pro mould helmet or by screws fastened to the cedure is totally automatic. The laboratory physi- calvarium. The bece plate may easily and cal dimensions imposed a small distance among the accurately be attached to various dia- cameras and therefore small angles (about 23 de- gnostic and therapeutic units. Coordinates grees) between optical axes. A small (1.2 x 0.8 x of target points (X,Y,Z) in these units 0.8 m) and light, easy to manage, calibration are directly related to the base plate. The system has been used for diagnostic object was used, which covered a small portion of procedures such as plain skull radio- the measurement field. These constraints usually graphy, computer tomography, cerebral entail large errors in the reconstruction of the angiography, positron emission tomography, 3D coordinates of markers. The results we obtained magnetic resonance tomography and gamma were the following. The s.d. of the error, for mar camera examinations as well as for therapeutic procedures such as stereotaxic kers farther than 1 m from the center of the measu biopsy, stereotaxic open surgery and rement field, was lower than 3 mm along the axis radiation treatment. pointing the cameras, and lower than 1.5 mm along For surgery a new instrument has been the other axes. This accuracy is to be attributed developed giving the surgeon opportunities to an especially devised stereometric model. to be able to reach the whole brain (no dead angels) and maintain sterility.

BE12-X.3 BE12-X.4 The summary of Double J-loop Elastic Silicone Pos- Relationship of Intracranial terior Chamber Intraocular leas Implantation. and Anterior Fontanelle Pres- Zhou KaiYit' )i8 Ying Men Kou Street, Chengdu, sure in an Invitro Model, Leon Sichuan, China. Shu-Guang Biomedical Engineer..ag Bunegin*, Maurice S. Albin, Research Institute. It is ten years, since we University of Texas Health made the first anterior chamber intra=dar 1:ns Science Center, San Antonio, by using GNY 515 silicone material and implanted Texas 78284 it. In Sep, 1981 we continued to design elastic silicone posterior chamber irtraocular lens An invitromodel ofthe anterior (ESPCIOL). After we tested on animals, we got the fontanelle (AF) was developed based suitable implantation method and instr'rnents. In onanatomical measurements taken Aug. 1982, the first ESPCIOL 7,,,as implanted into from 17 babies ranging in age from 1 human eye and its result was extremely successful. to 42 days. Anterior fontanelle In more than five years, we have implanted 100 pressure (AFP) andintracranial ESPCIOL among 96 patients. The results are pretty pressure (ICP) appearto correlate good. 20-30 days after the operation, the mean closely when (a) the transducer base naked visual acuity of 100 eyes was 0.68+0.32 plate has sufficient crossectional Follow-up 5.5 years_ 0.5 year(mean 2.5 years), diameterso that it can rest on the the mean follow-up visual acuity was 0.89 + 0.23. scalp overlying the bone adjacent to For past 10 years we have operated more tan 200 the fontanelle margin,(b) applica- eyes (include ACIOL, Iris-Plane iol), which confirm tion pressure on the sensingdevice ESPCIOL being benefits of extremely biocompatibil- is sufficient to keen the transducer ity, softness (reduced in situ trauma) and top se- base plate in contact with the scalp curity. The main complications caused by implanted overlyingthebone adjacent to the ESPCIOL are: the thickening of posterior capsule AF margin,(c) a transducer attach- (23%) and forming posterior synechia (25%, espec- ment techniquebeused that mini- ially for the 2nd-stage implantation).All these mizes drift due to adhesive slippage problems can be solved easily. and material relaxation.(d) transducer sensitivity to patientmotion was ninimal, (e) a means for assessing transducer drift in situ is present.

ge ?Ai tp 1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas 337 Biomedical Engineering Scientific Papers

BE12 -X.5 BE12-X.6 Relationship of Intracranial Automatic Power Control for Cutting and Coagula- Pressure and Anterior tion in High-Frequency Surgery. K. Fastenmeier* Fontanelle Membrane and G. Flachenecker, University of the Bundeswehr Displacement in an In Vitro Munich. D-8014 Neubiberg, Fed. Rep. of Germany. Model, Leon Bunegin*, Maurice S. Albin, University of Texas In the field of high-frequency surgery, all stan- Health Science Center, San dard high-frequency units use manualpower set- Antonio, Texas 78284 ting. On this account, the average power applied to the patient is always much higher thanneces- An invitro model oftheanterior sary for the cutting and coagulation effect. This fontanelle (AF) was developed based fact increases the well-known electricalhazards on anatomical measurements taken for the patient, as deepnecrosis, burns, and from 17 babies ranging in age from 1 nerve- and muscle stimulations. to 42 days. After repeated pressur- It can be shown that the cutting effect is relat- ization of the model cranium, it was ed to an electric sparkbetween surgicalprobe noted that (a) positivesimulated and tissue. We use the nonlinear behaviourof intracranialpressure (ICP) always this spark as a criterion for the automatic con- produced an AF bulge; (b) negative trol of the output power. A high-frequency simulated unit ICP always produced a AF with very low internal harmonicdistortion depression; (c) zero has simulatedICP been developed. The harmonicsgeneratedby the resulted in the AF membrane lying spark are measured and compared to a preset val- flat, parallel tothe plane de- ue. The difference is used as a signal for the scribedby the edges of the cranial automatic output power control. By thismeans, bones thatform the AF; (d) the the extension of the arc is continuously kept extent ofthe to AF membrane dis- the lowest level sufficient for apropereffect placement is not linearlypropor- during cutting. In the coagulation mode, the out- tional to the simulated ICP; (e) the put power is pulsed. Every pulse is terminated as bony area around and adjacent to the soon as a spark is detected. AF membrane seemsto remain in a Recent measurements show a significant reduction stable orientation over a range of of applied output power during cutting and avery simulated ICPs. rapid, yet cafe coagulation effect.

BE12-X.7 BE12-X.8 Theoretical Investigation of Magnetic A Method of Creating a NewMicrowave Field Field Cautery* Nabil M. Abd1.11-1 ir. and Manor A. Ezzadeen, Electrical for Intrauterine Treatment. Ji tiongwei% Wang Liran, Tsao Supieh, Wang Xueqing,202 Engineering Department* College of hospital, Lhenyang, China(PAC). Engineering, Mosul University, Mosul, Iraq. Recently we designed a sperlAlform of microwave rad.raor Magnetic field cautery can be done by insertinga which is applicable to both treatmentof functional uterine bleeding and termination of small solid metalic cylinder inside the bodyto very early pregnancy. IL isbo reach and touch the damaged area. Magnetic field designed as to change the circularor ellipsoid pattern of can be applied from outside to cause the tempera- radiation field on meridian planeto be one of triangular form, so that the radiation ture of the cylinder to increase to therequired energy recieved by the %hole temperature for cauterization. uterine wall tends to be distributedmore evenly. This paper presents a mathematical analysis of Microwave at 2450 MHz is transmittedthrough a semirigid magnetic field diffusion and eddy current inside a cable to the radiator. An elongatedstructure of at. elastic cylinderical metals and the calculation of rise of butterfly loop was built in the radiatordesigned according temperature in these metals due to eddy current to the rule of Al2t.dipole radiat.on. loan. The energy deposition pattern of this newly-designed radiator These calculations have been made with the aid conforms well to of the triangular shape of the uterine computer for two cylinder positions with cavity. Correction of respect position and diameter of the butterfly to the applied field. The first loop can increase case consists of tb. c- .ipling strength and make the applying the field in parallel to the axis ofthe microwave energy to be transmittti directly and evenly to the cylinder (axial field)* while the secondcase is 4r cornua uteri of both sides. Radiators for different sizes applying the field perpendicularly to the axisof 01 uterine cavity were the cylinder (radial field). These twoprograms made. Manikin tests with egg-whiteshowed a triangular compute the increase of the temperature at the energy deposition pattern on the frontalsurface. The long- surface of the cylinder with time for different est upper edge of it was 4.5 cm andthe shortest, 2cm.Pain- metals and frequencies. Low frequency has boon i_ss operation was assured by overlappingof the two wings used to ensure large akin depth and good penetra- of the butterfly to reduce its diameterbelow i.M mm on en- tion in the body. tering and coming out through thenarrow cer.icli canal. This radiator has already beenapplied clinically.

8 338 1988 World Congress on Medical Physics and BiomedicalEngineeringSan Antonio, Texas Biomedical Engineering Scientific Papers

BE12-X.9 BE12-Y.1 Surgical Treatment of Acute Inflammatory Effects of Altering Core Body Temperature Diseases of Liver, Bile Ducts, Pancreas on Somatosensory and Motor Evoked with the application of New Physical Potentials, SiavashS. Haghighi: John J. Methods, V.S.Zemskov*, N.N.Korpan, Oro, Clark Watts, University of Ya.I.Khokhlich and V.A.Pavlenko, Kiev Me- Missouri-Columbia, Columbia, Missouri dical Institute, 252004 Kiev Schevchenko Bul. 13, UkSSR. In surgical treatment of 4392 patients The effects of core temperature on the motor with acute inflammatory processes of li- evoked potentials (MEPs) were explored in rats. ver, bile ducts and pancreas there were Temperature wasgradually increased from 37.5 C applied the methods of external local to 40.5 C or more, and later decreased from 37.5 hypothermia, (3919) with the help of appa- C to 28 C or less. MEPs elicited by motor ratus "Cryoelectronica-7" and intraopera- cortex stimulation were recorded over the ;,pine at tional cryosurgical action by means of baselinetemperatureand after stabilizing "Cryoelectronica-4" (334 patients) deve- the temperature at a given degree. loped at the clinics. While treating 102 patients with obturating jaundice ofnon- The D wave latency of the MEP was decreased with neoplastic etiology after the operation hyperthermia (p=0.008) and increased with the intravenous lazer irradiation of hypothermia (p=0.009).The conduction velocity blood was being conducted by Helium Neon was increased withhyperthermia and decreased Lazer, lazer apparatus AFL-1 and monofib- with hypotherAia. There wasno significant rous optical ligh;conducter. 37 patients change in amplitude with hyperthermia (p=0.55) were treated by lazer therapy, besides although there was a trend toward reduction with electro-magnetic irradiation of low in- hypothermia (p=0.09). The interpeak latencies between the D tensity was applied on the points ofacu- and I waveswerevariable but puncture. Clinical experience of appli- tended to decrease withhyperthermia and cation of physical means of treatment by increase with hypothermia. The MEP deteriorated cooling, lazer and electro-magnetic above 42°C and below 28°C. action testifies to their high efficiency in surgical treatment of patients with Core body temperature has a significant effect the given pathology. onthe MEP andshould be consideredwhen interpreting changes in this signal.

BE12rY.2 Performanceofa twelve-electrode intraneural BE12-Y.3 stimulation device in silicon technology, Effect of Dispersion on Compr -d Action Fields of Peripheral Nervs W.L.C. Rutten *, H. van Wier and J.H.M. Put, University of Twente, 7500AEEnschede, The G. Curio+, S.N. Erne *, L. Trahms Netherlands. Phys.-Techn. Bundesanstalt- Institut Berlin- FRG + Freie Universitat Berlin - FRG A multielectrode array in silicon technology has been developed to investigate one of the present Recently, we reported on the first biomagnetic de- bottlenecks inelectrical stimulation: how to tection of Compound Action Fields (C, ) of peri- interface neurons with electronics in asensitive pheral nervs in vivo. Surprisingly, the morphology stable andselectiveway. The array consists of of the signals was monophasic and not inagreement twelvc platinum surfaces (20x60 pm, at 50 pmfrom with the expectation on the basis ofa simples each other) ona 50 pm thick tip shaped silicon couple of current dipoles. We present here a mo- substrate. The tip was inserted in the peroneal del calculation to show that this morphology is nerve of the rat to stimulate the EDL muscle dur- the consequence of the dispersion due to the dif- ing acute experiments. A six channel stimulator ferent propagation velocities of theneuron popu- delivered independently controlled pulse series to lation. the electrodes to elicit isometric twitch or tetanic muscle contractions. Sensitivity was such that one single motor unit could be selected at most of the electrodes in the array. Stable s-amulation was possible over long periods, onlylimitedby the condition of the stimulated nerve bundle, i.e. over a periodofabout four hours. Recruitment c4rves of stimulation current versus muscleforce at a single electrode had a dynamic current range of about 20 dB. In order to examine selectivity we developed a test using the refractoryproperties of neuronsand simple assumptions for the volume conductor near the electrodes and the distribution of the nodes of Ranvier. Results of the test lead to the optimal choice of electrode configurations and stimulus conditions to maximize selectivity. ur,

1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas 339 Biomedical Engineering Scientific Papers

BE12Y.4 BE12-Y.5 The Effect of a Refractory Prelod on Chemical Sensing Based IPFM in the Nervous Systems. T.Matsuoka on Concentration and H.Morita: UtsunpAiya University. Effects on Neuron MembraneProperties, 2753 Ishii-machi, Utsunomiya, Japan. R.S. Skeen,*'t B.J. Van Wie,tS.J. Fungl and C.D. Barbes,t (Chem.Engr. Dept., It has been shown that in thenorvous systems the 1Dept. ofVet. and Comp. Anat., Pharm. method of signal transmission is integralpulse and Physiol., WashingtonState University, frequency modulation (1PFM) anda refractory pe- Pullman, WA99164-2710. riod exists after occurrence of pulse.The refractory period consists of absolute refractorype- The purpose of this work hasbeen to explore riod (ARP) and relative refractory period (RRP). the potential of monitoringchanges in neuronmem- In this paper. the effect of therefractory period brane propertiesas a means of chemical sensing. To on IPFM is studied analytically and throughcom- study this, the effects of puter simulation. It is found that ethanol between 0.2 M and the refractory 1.0 M, and n-butanol between0.016 M and 0.063 M period gives the following effectto 1PFM. on the repetitive firing frequency of 1. In IPFM with ARP. pulse densityincreases in neurons from the snail Limnea stagnaliswere measured. Results show proportion to the logarithmic value of inputsig- that both alcohols induced nal and is led to saturation. Due tothese two concentration dependent changes in the firingfrequency-stimulating current (F- characteristics. S/N of IPFM does not decrease 1) relationship. Normalization for additive Gaussian noise. And ARPcauses syn- of data reduces varia- chronizing pulse train with input periodic tions between individualneurons creating a common signal. basis for comparison. This The period of the input signalcan be known by allows the netons to be use of the synchronization. classified intogroups that show distinct normalized firing frequency 2. In IPFM with RN. the change of pulsedensity response profiles to alcoholconcen- with input signal is smaller than init without tration. Good reproducibilityof response within each RRP. Consequently. dynamic range ofinput signal group is shown. These findingsare discussed in rela- becomes wide. tion to the development ofa neuron-based solid-state chemical sensor.

BE12Y.6 BE12Y.7 A Suggestion about the Use of a Single Parameteric Criterion to Representing the Excitability A device for monitoringEMG spontaneous activity during tetany(*)A. Pepino, Datong Jiangs , Shaohua Mang R. Castaldo, M. Bracale,A. Stariza(***) Irmitute of Bicyclical Engineering B. Rossi (****). *Univ.of Naples Electr_ Xi'an Jiaotong University Dept. Via Claudio 21 Xi'an , China Naples (I) **Univ. of Pisa Inform.Dept. C. so Italia 40, Pisa (I)

Excitability is one of the fundamentalproperites of biological ***Univ. of PisaNeuro, Clinic Via tissue. It is greatly recognized that Intensity-DurationNoe dei Savi, Pisa (I). can properly represent the excitability. The curve is usually The automatic analysisof the EMG spontan- interpreted by Weiss formula i = a/t b , which contains two eous activity recordedduring the activa- parameters to determine the sheap of thecurve . Some scientists tion tests (ischemia, use rheobase and chronaxie to describe excitability and hypernea) supplies sore quantitative data on scientists suggest use excitability and another the extent of electro- term irritability lytic changes induced to represent excitability. All these representations involve two by this syndrome.A specially implemented parameters . In practise, if we can describe excitabilitywith a algorithm applied to single parameteric criterion instead oftwo, it will be convinent the EMG signals makesit possible to to compare the variation of excitability in different quantify the observed conditions. amount of the EMG We consider the minimum distance of i-tcurve to origin can brief- spontaneous specific activity ly represent the main characteristics of i-t that is curve . Based on this correlated to metabolicdisorders. A idea we suggest the minimum distance (R)as a single paraaeteric criterion of excitability friendly-user device whichis based on the . The validity of the idea is demonstra- previous algorithm has ted by experiments using sciate-fibularnerve of toad under diffe- been designed. It rent conditions performs an automatic . recognition of spontaneous EMG activityand provides on-line a graphical representationand ,hree measurement parameters ofthe biological response to the activationtests.

F.r. 6V 340 1988 World Congress on Medical Physicsand Biomedical EngineeringSan Antonio, Texas Biomedical Engineering Scientific Papers

13E12-Y.8 BE12-Z.1. Evablation of Chamben-Type The Theory of 'Channel and Network' and *Mictoeteettode4 LotSt-imitation 'Breath' in Traditional Chinese Medical M. M. Hill and O. Prohaska, Case Western Reserve Science and Its Relation with r_oelectri- University, Cleveland, OH 44106. city, Gu Zhimin and Yu Guorui, Res Ctr. Shanghai Safe and affective neural stimulation demands minimal damage to tissue and electrode, while A great deal of experiments on animals and maintaining sufficient current and charge human body has come to the discovery that densities to functionally stimulate.With the characteristics and effects of 'Bio- recent advances in solid-state technology, a electricity' are extremely similar to those new chamber-type microelectrode has been of 'Breath' recorded in ancient China's fabricated. Based on the geometrical design medical books. This ':,reath' doesn't mean and a differential current density theory, the the air one breathes. design meets these demands. Its substantial base Evaluation of the chamber-design was made by is the 'Bioelectricity' through an invisi- performing comparative in vitro stimulation ble system of 'channels and Networks'. experiments between Au thin-film microelectrodes In accordance with this doctrine, the with and without the chamber. The criteria of Equilibrium Detector for Channels and Breath' evaluation were light microscopy observations and Equilibrator for Channels and Networks' and potentiographic methods. Increased current have been developed. Emitting a micro-current passing capabilities of the chambered-electrode pulse, these two instruments may balance the was demonstrated; however, due to a non-uniform effect of positive and negative bioelectri- current distribution within the chamber, some concentrated edge corrosion occurred. Computer city, make the blocked 'Channels and Networks' modelling aided In optimizing the chamber- through, and, therefore, give satisfactory geometry to reduce this non-uniformity and cure effectiveness confirmed by clinic corrosion potential. applications to about 200,000 patients/times. The realization of an improved microelectrode The figures of entire recovery of some for stimulation has been shown. diagnosis as follows: Cervical vertebra 83.0% Sciatica 39.4% Antrum gastritis 87.0% Chilblain 88.5% Obliterans arteritis 80.0% Facial paralysis 92.4% BE12-Z.2 Frequency*Selection of Meridians in Chire Medicine, T. L. Hsu , W. K. Nang',Y. Y. Lin Wang and Y. BE12-Z.3 Chiang, Biophysicf Lab., Institute of Physics, Implemental System for Diagnosis, Academia Sinica, Joint appoint with Institute of Treatment, Teaching and Consultation gicmedicalElgineering, Yang-Ming Medical Collage, 'Department of Physics, National Taiwan Narnel Univer- of Chinese Traditional Medicine, Wei sity, Taipei, Taiwan, R.O.C. Jizhou*, Jiang Chunxiao, Wei Sudan, Chen Ziaoguang, and Liu Hongjian, Bethune Medical University, Xin Min Meridians, the hypothetic lines which connect groups of acu- Avenue 6, Changchun, Jilin, P.R. CHINA. puncture points, play very important roles in both Chinese medicine and Chinese Kung-Fu. The acupuncture points cm the sane meridian sere suspected to be a series of arterial trees Implemental System is built on the basis connected by sane artery and all these arterial trees on the of artificial intellectual faculties-- sane meridian might resonate with similar frequency or fre- Specialist System. Actually, it is a quencies as the multiple of heart rate. neutral (an empty) program. As long as the data of theoretical designs of Teoretical derivation using a homogenous electric circuit medicine are inputted, the electronic analog model suggested two selected resonant frequencies, one computer world automatically produce a cane from the artery and the other one cane fran the arterial specific specialist system. It is trees. It also indicates that if the arterial trees an the constituted from the knowledge stock of same meridian could be modulated to a better resonant condition, the blood pressure wave would propagate with less atte- intellectual faculties, data stock, in- nuation and larger speed.Since pressing the acupuncture ferential machine of an association, points could change the vibrational condition ao is the re- renewal of knowledge, explanation etc. sonant frequency.This mechanism also provides ecplanatinn 300 different diseases have been tested cf the medication effect of acupuncture.The releasing "Chi" and verified, all fulfilled the design effect in Chinese Kung-Fu might be also due to the modulation requirements and achieved good result. of resonant frequencies of the arterial trees by controlling the well trained wade tone around the arterial trees.Sane experimental results will be discussed. 3 6

1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas 341 Biomedical Engineering Scientific Papers

BE12-Z,4 BEI2-Z.5 Real time multifunction pulse Sigel and Information Procasing in Biomedical Engineering aidInputs diagnosis analyser, L.Yu*, oniraditional Chinese &dein, than Ping -Yam and Be Zhen-Ya, Digital D.M.Lu, J.K.Sun. Oceanic U. Signal Processing Research Group, Radio Departhenti MagniInstitute of of Qingdao. Qingdao,P.R.China Technology, girding, P.R. Chins. Pulse diagnosis is important in Chinese Rapidly developing, the signal and intonation processing has been playinga traditional medicine. Doctorstest the pulse acre and more izportant role in aedical science and biosedical engineering of the artery of wrist andpick up and is beginning a influence on traditional Millen ledicine. informations from the feeling Over the last of fingers. decade,the present Whoa, in cooperations with their colleagues,me There are pulse analysersthat can measure devoted theiselvesto .ad record pulse's the studies of new techniquesof signal and wave form and amplitude intonation processing and their applications in aedical sciences,both the but all can not be usedsimultaneously Yestern and Chinese, with the finger test, several dozen research papers have been published in so that it is not URI, IPW and other international organizations' journals, transactions very useful for clinical diagnosis. and other periodicals,read out at their workshops, conferences, Imposing and We designed a new kind ofpulse analyser that get rid of this congresses, and sate research szbievetents *re rewarded by Chinese national trouAe. Its sensor awards.This paper gives a soden perspective on current status of thearea is a multilayer thin andflexible transducer. It can be used under Doctor's and the authors' works whether they have been publishedor not.While all finger aspects of tee area are discussed, mines are given to the following: and measure and recordthe w.ve form,amplitude,static pressure t one-dinensional signal processing applied to ECG,REG,MEG, go and simultaneously with others; Doctor's finger test. Itcan measure three t radical tight; positions of the artery at thesame time, and also may be used the Chinese a.dical ideas based artificial intelligence forautosatic synchronously with detection and recognition of diseases from the signals and inges; CEG,ACG PCG,etc. So that ithas all the functions that demand by aedical expert systeas, both the Yestern and Chinese; finger test. hospital unused autnation. An introduction of this pulse analyser is References of the published wisteria's are given. provided with the recordsof several kinds of pulses,and alsosome comoined data with other kind measurements.

BE22 -Z. 6 A Fuzzy Attribute Syntatic BE12-ZZ.1 Method Classifying On the Shape of the Relaxation Curve of Remanent and Recognizing APPCM, Zhangwei-lian, Tian Zhi- fen and Wang Ai -mini Northeast Magnetic Field Produced by Magnetic Particles in University of Technology, P.R.China Cells, I. Nemoto*, K.Ogura and H. Toyotama, Recently, how to maze an objectiverecognition of Tokyo Denki University, Hatoyama, Saitama, the Arterial Pulsewave Pattern according to Japan, and Stanley Electric Co., Tsukuba, Japan. Chinese Medicine by a computer system instead of by Ferrimagnetic particles ingested by macrophages physician';, palpation has beenint,ested by many researchers. In the paper a new method of the produce a weak magnetic field around the cells whe.. APPCM recognition is presented,in which 28 kinds strongly magnetized by an external field. The field of single arterial pulse and10 kinds of unusual from the cells decays with time (relaxation) due to the pulse according to AFiCMare recognized. The fuzzy movements of the microstructures of the cell. The feelings of the physicans ofChinese medicine are measurement of relaxation is expected to provide classified using a fuzzy attribute information on the activities of intracellular motility. region and for We previously proposed a model for the phenomenon, picking out the main features ofdifferent arterial waveforms a series approachesare proposed, e.g., employing the rotational Brownian motion in which peak-unit expression, period th .rmal agitation kT was replaced by some unknown estimation, baseline ene-gy Er. According to this model, relaxation curve fluctuation correction and noisefiltering. More- over in the method proposed in the should decay as an exponential curve. However, actual paper arterial cu* yes show a large deviation from a single exponential pulses are classified into fourgroups according to the different ways of the pulsedistortion, i.e, curve. This does not necessarily mean that the model is rate cistortion group; period distortiongroup; inappropriate. The deviation may be explained by form distortion group and positiondistortion distributions of the parameters involved in the model. group. They are particle size, the size of the phagosome or A computer system for APPCMrecognition is set up lysosome containing the particles, the energy Er and in the hospital, where thedata sampled in the the apparent viscosity. Taking into account the General Railway Hospital ofShen Yang are classi- distribu on of the parameters would make the curve fied and recognized. The simulation results ob- fitting easier but more arbitrary. The question is not tained by the systemare in keeping with that by one of how precisely the curves can be fit but which the Chinese physicians in model reflects the actual process better than the others. clinical practice. As yet we can not give a definite answer but, instead, we have several models, the data they would produce, and their interpretations. 3

342 1988 World Congress on Medical Physics and BiomedicalEngineeringSan Antonio, Texas Biomedical Engineering Scientific Papers

BE12-ZZ.2 BE12-ZZ.3 Magnetophoresis for Separation of Monocytes and FeedbackControlof aCobe2997 Diagnostic Evaluation of the Phagocytic Capacity, Centrifugal Cell Separator, S. Wooten*, H.Minamitani*, T.Shinx:mura and M.Aiba, Feio Univ., J. N. Petersen,B. J. Van Wie,Chemical Hiyoshi 3-14-1, Kohoku-ku, Yokohama 223, Japan. Engineering Department,Washington State University, Pullman, WA 99164-2710 Magnetophoresis is a new method to separate monocytes phagotyzing iron dextran microsheres ant' to evalua- In order to obtain more efficient operation of a te the phagocytic capacity in patients with liver COBE blood cell centrifuge (Model 2997, SS II diseases based upon the separation eff.3,ency of the chamber), modifications were made to allow complete monocytes from mononuclear cells under magnetic computer control.The clinical cell separator was field. Mononuclear cell suspensions were prepared interfaced to a microcomputer for automatic feedback from heparinized peripheral blood in the patients control of interface position. An optical sensor was and is healthy volunteers by density centrifugation. used to indicate total cell concentration in the buffy The mononuclear cells were incubated with iron coat product stream. This cell concentration is used as dextran microspheres at the ratio of 20 particles an indicator of the chamber interface position. A to 1 cell at 37°C for 30 min and dropped in a glass feedback control scheme was used to control the chamber filled with RR4I solution. High DC magnetic interface position by automatically adjusting the red field of 6 kG was applied to one of the outer wall blood cell and plasma product rates.Using this of the chamber and a cover glass was placed in the controller, an operator can easily maintain interface corresponding inner wall. Magnetic microsphere position, and thus total cell concentration in the buffy phagocytized monocytes migrated towards the glass coat product stream. This controller has allowed us to while the other mononuclear cells such as lymphocytes investigateoptimaloperatingregionsforthe fell down onto the bottan of the chamber. Both the centrifuge. monocytes and lymphocytes were fixed, stained and observed under vital microscope to count the phagocytized monocytes and the other cells. The monocyte phagocytic capacity was calculated fran the ratio of the migrated monocytes to the others. The results showed that the phagocytic capacities of hepatcrna and liver cirrhosis became significantly lower than those oc healthy volunteers, and that the magnetophoresis ras extremely available to evaluate the phagocytosis.

BE12-ZZ.5 BE12-ZZ.4zstablishment of the method of Cequential Effects of Supivacaine Hybridoma t..ytoelectroPhoresis Assay, Combined with Hyperthermia on Tumor X.Yan*, J.dang, S.Zhang and X.Ling, Cell Surface electricity, J.lang, fianjin Cancer Institute, fianjin, X.Yan*, -.Zhantr and T.Zhang, Tianjin China (P.h.C.) Cancer Institute, Tianjin, China A novel serodiagnosis methidwith monoclo- . equential effects of bupivacaine combined nal antibodies, hybridomacytoelectrophore- with hyperthermia on the membrane .e.lectric sis assay, and its applicationfor detect- charges of human cervix carcinoma cell line ing PA3 antigen in sera oflung cancer pa- vela were investigated. the results of cell tients are described. the method wasde%e- eLectrophoresis assay indicated that the loped on the basis of the theorythat hyb- slowing rate of cell migration for the tre- ridoma cell surface electric charges can atment sequence of bupivacaine before heat be affected by the bindingreaction an- (BH sequence, 79.2.)zas sig-ificantly hi- tigen and antioody. Hybridorna(.ell line th-er than that c' reverse sequence (HB se- PA3 producing monoclonal antibodiesto lung quence, 614.0,,,, P<0.01). Comparing with HE cancer was used as thenatural "biobeads" sequence, colony formation rate of Hela carring the antibodies. Gff all163 cases, cells dropper: from 49.33% to 37.33, wheli PA3 cell slowing rate?-,35;, 'was f "und in 35 cells were treated with BH sequence, show- of 46 lung cancer patients(positive rate, ing that Bri treatment afforded more prono- 76.08%), 8 of 35 patients 4th other can- unced tnermosensitizing effect than dB tr- cers (22.86%), 3/32 forbenign diseases of eatment did. the speed of cell migration lung (9.38%.) and only 3/50 fornormal per- showed a positive relationship with colony sons (6.00%). Thesensitivity, speci ic, formation rate (r=0.876, P

1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas 343 Biomedical Engineering ScientificPapers

BE13-A.1 BE13-A.2 Structural Modeling of SoftTissueMechanics, STRESS-STRAIN RELATIONSHIP OF THE Y. Lanir: Technion-Israel Instituteof Technology, Haifa, 32000, Israel THORACIC AORTA H W Weizsacker,J Vossoughil T D Kampp, Themechanisms which give rise to the mechanical Physiologisches Inst Der Karl-Franzens- response of tissues are intimately associated with Universitat Graz their structure and composition. Austria, Catholic Univ of The most important America, Wash DC Univ of S California-LA structural elements are networks of long fibers (primarily collagen and elastin) and a fluid-like Cylinc cal matrix consisting of water, small segments of rat thoracic solutes and pro- aortas ,ere teoglycan macromolecules. subjected to simultaneous The following features axial stretch and internalpressure in the must be considered in developingstructuralmater- physiological ial range of deformation. The laws: a. The fibers at rest are generally passive elastic properties undulated in a nonuniform of the aortic manner. b. The fibers, wall were analyzed in orientation is terms of Kirchhoff generally nonunitormas well but stresses and often preferred orientations exist. Green-St Venant strainsby c. Thenega- plotting the tively charged groups on stress-strain relationship the glycosaminoglycans for several constant levelsof complemen- afford them with osmotic properties. tary strains. Under tensile deformations the From curvesso obtained, contribution of the the 1-st partial derivatives matrix is very small. Under compressive loading, of stress on the other hand, the matrix with respect to strainwere computed and contribution is of essential plotted as a function ofstrain. significance. Stochasticmodels which relate The results show that theaxial stress- microstructure to overall mechanical strain curves are roughlysimilar tothe response have been developed for fibroustissues in general and for specific tissues circumferential stress-strain curves in such as tendon, that the slopes of the skin, parenchymaand the myocardium. These models curves are finite initially, go throughan inflection and can readily account for the observed complexaniso- rise steeply at large strain tropicandnonlinear properties of tissues while levels. The S-shape of the stress-straincurves is re- offering a realistic insight intothe internal dynamic flected in the curves ofpartial deriva- structure. In addition they canserve as a useful tool by which specific tives which are concave upwarddisplaying pathologicalproces a minimum atstrain levels occurring in ses can be diagnosed based on mechanicaltests. vivo.

HE13-A.3 BE13-A.4 A New Approach to Vascular Mec'anics A Layered Cylindrical Shell Modelfor an Arte, Hilmi Demirayl.., Research Institu- J.D. Humphrey * , R.K. Strump'.+ & F.C.P. Yin+ te for Basic Sciences, Box 74,Gebze,Ko- University of Maryland aild Johns Hopkins caeli-TURKEY, and Raymond P.Vito,Georgia Medical Institutions, Baltimore, Nu U.S.A. Institute of Technology, Schoolof Mecha- nical Engineering, Atlanta, GA30332, U.S.A. Despite extensive efforts,there are no uniformly accepted stress-strainrelations for Assuming that an artery consists of blood vessels. Possiblereasons for this may an orthotropic medial and isotropic adventitial be (1) the use ofpurely phenomenological layer, a set of strain energy density functions relations, which cannot directlyaccount for for each layer is introduced and the valuesq)f material load-induced changes in thestructure, and (2) coefficients are obtained by comparing the analytical the lack of appropriateexperimental results data. of a bi-axially stretched specimen Thus, we suggest astructurally-motivated with those of 9/71proarl, to fnym"loto experimental measurements. Havingdetermined the material coefficients, functions, W, for vario,-.s vessels. The a two layered cylindrical shell model is introduced and analysis includes large deformations, the aistributions nonlinear of stresses through arterial wall and the material behavior, load-dependent inner pressure as a function of oeformation are obtained. anis.nropy, material heterogeneity,incom- In thismodel, although pressiblilty and residualstresses. Using the radial stress is thisapproach weidentify new experimental continuous at the interface, the slope ofit and other normal stress components become protocols necessary to quantify thematerial di.conti- symmetries of blood vessels anddetermine nuous. This result is due to assumptionswe have soecific functional forms of Wdirectly from made about the layered structure ofarteries. data.

344 1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas Biomedical Engineering Scientific Papers

BE.13-A. 5 BEL3-B.1 RHEOLOGICAL BEHAVIOUR OF ON VARIATION OF MECHANICAL PROPERTIES OF BIOLOGICAL POROUS MATERIALS : AORTIC TISSUE VISCOELASTICITY AND PORORELASTICITY. Olivier LE GALLO, Salah NAILI, J Vossoughil School of Engineering and Didier GEIGER*; Architec,ure, The Catholic University of UNIVERSITE PARIS VAL DE MARNE America, Washington DC 20064 Laboratoire de Mecanique Physique 94010 CRETEIL FRANCE Many composite materials of biological :nterest are made up Several factors such as age, anatomical by a solid elastic matrix embedded in an interstitial viscous location, disease type and stage, etc fluid. To study the rheological behaviour of these materials effect the mechanical properties of blood their diphasic character is to be considered. So, if a vessels. This paper systematically stu- viscoelastic approach seems adequate to describe their diesthe variation of mechanical proper- behaviour, only a diphasic model, taldng into account the fluid ties of thoracic and abdominal bovine dynamics in the solid matrix, leads to a physical interpretation aorta along the circumference as well as of the mechanical properties of these materials. along the length. Past work indicates To determine the stress and strain fields in the whole medium that the stiffness of the vascular vessel (fluid and matrix), as well as the pressure in the interstitial varies along the aortic tree but not sig- fluid, the poroelasticity theory has been developped (Blot, nificantly. In this work it is shown that the stiffness along the length of the 1941). In this study, we point out the fact that an overall description, aorta varies much more than it was re- using the viscoelasticity theory, does not have any intrinsic corded ,before. meaning, and so that this type of model is only suitable to Also it is shown that the stiffness of characterise the mechanical properties of the material in a the aorta at any given location along the given situation and for a given state of loads. lengthvariesdepending onthe location We suggest also a new method to integrate the poroelasticity along the circumference of the aorta. It equations and compare the theoretical results to experimental is shown that the aorta is stronger along the measurements. dorsal axes and weaker along the ventral axis.

/3713-B.2 BE13-B.3 Indirect Measurement of Elastic Modulus Non-invasiveevaluation ofvenous & Diastolic Pressure in Human Finger compliance and arterial resistance in Arteries by Cuff-vibration Technique the forearm, P. GIZDULICH *, Fisica K.Seki*, H.Shimazu, H.Ito, A.Kawarada, Medica- University di Firenze, M. M.Akimoto, & K.Yamakoshi, Dept Physiol, GUERRISI & A. MAGRINI, Fisica Medica - Kyorin Univ, Mitaka-shi, Tokyo; K.Yamakcshi UniversitaTor Vergata - Roma, F. is in Res Inst, Appl Electra, Sapporo-shi, IELLAMO, Fisiopatologia Medica - Hokkaido, Japan Universita Tor Vergata - Roma, Italy

Diastolic pressure (Pd) and volume elastic Dependency of forearm venous compliance and modulus (E defined as APRAV/V0) were arterial resistance on increasingvenous indirectly°)measured by vibrating a finger pressure, due to prolonged venous occlusion, was with a 10 Hz sinusoidal pressure during a investigated. Sudden cuff inflations of 30 gradual decrease in occlusive cuff pressure torn, starting from variouscuff pressure (Pc). Sinusoidal and pulsatile changes in values, were carried out.Beat-to-beat arterial volume (AV) and its ratio against the mean pressure in a finger was recorded by a arterial volume (AV/V ) were detected by an calibrated servo-plethysmo-manometer (FINAPRES). infrared photoelectriSplethysmograph. It Increases in forearm cross section area were is known that volume chant, anartery recorded by a strain-gauge plethysmograph. shows a maximum amplitude at the transmural Venouscompliancewas calculated fromthe pressure (P ) equal to 0 mmHg due to the asymptotic value of the plethysmographic signal nonlinearvisco-elastic p.operty of arte- and the cuff pressure. Blood flow was evaluated rial wall. For the same reason, the ampli- from the initial slope of the plethysmographic tude of sinusoidal volume variation reaches signal. Results show that blood flow varies its maximum at the end-diastolic phase when significantly with the venous pressure. Arterial Pcwas controlled to be exactly equal to resistance and venous compliance did not show P,. The indirect P values determined by any significant dependance on venous pressure in this method were comparedd with those simul- the investigated range of venous pressures (0-60 taneously measured. Using the vibration torr). technique systolic, mean and diastolic pressurecan be indirectly determined by means of the volume oscillometric method. 365 1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas 345 Biomedical Engineering ScientificPapers

BE13-8.4 Effect of exercise training on biochemical BE13-9.5 and bicmechanical properties of rat aorta, EFFECT OF ENZYME DEGRADATION ON THE T. Nosaka: M. Nietsuda, M.Sato, N. Chthima PARAMETERS OF A CONSTITUTIVE MODEL OF THE and H. Fukushima, University of Tsukuba, CANINE CAROTID ARTERY, Thomas R. 1-1-1 AmakUbo, Tsukuba, Japan Canfield, Ph.D.*, Loyola University of Chicago, Stritch School of Medicine, We studied the effect of prolonged physicalex- Maywood, IL. ercise on the fibrous proteins of rat aorta, associating with the aortic wall elasticity. A constitutive model of canine carotid Twelve male rats were divided into 6 sedentary has been developed. It is based upon a cortrol rats(C) and6 training rats(T) that symmetric formulation in the three princ- rnn voluntarily in a free wheel from 9 weeks to iple strains. The artery is assumed tobe 25 weeks of age. incompressible and behave as a thick The aortic elastin and collagen contentsin T walled anisotropic tube. The properties were not significantly different fromthose in can vary across the wall thickness. The C. The content of calcium in elastinwas lower parameters of the model are estimated by in T than in C. Elastin isolated from the minimizing the weighted squared error training rats showed significant decreases in between estimated and measured values of sane of polar amino acids. The incremental diameter and longitudinal distending elastic modulus values were significantlylower force as a function cT pressure and in trainingrats thanincontrols, at the longitudinal extension. The modelis extentivn ratio of 1.5 and 2.0. currently being used to evaluate material We conclude that prolonged voluntaryrunning changes induced by selective degradation from an early age provides training rat witha of thearterial wall with the enzymes extensible aorta, relatingto elastin with collagenase and elastase. The results of fewer calcium deposits anddegenerative i'ese studies will be discussed. changes.

IL 3-C.1 Anisotropy of Isolated Perfused Striated BE13-C.2 Muscle ISOTONIC During Passive and Tetanized States. AND ISOMETRICMEASUREMENTSQN CAROTID ARTERIES,W.W. von Maltzahn , R. K. Strumpf*, J. D. Humphrey, F. R.W. Calcote, S.M. Jayaraman, University C. P. Yin *. of Texas at Arlington, Box 19138, Arling- * ton, TX 76019 Johns Hopkins M dical Institutions and 'University of Maryland We built an experimental apparatus for measuring the mechanics' properties of arterial walls under There are no data comparing thestress-strain both isometric and isotonic conditions. A segment relations of passive and active striatedmuscle fromthe commoncarotid arteryof a dog was under simultaneous biaxial loading. We measured mounted on grooved pinconnectors betweentwo the biaxialstress-strain relationshipin 10 isolated, perfused canine diaphragms. stainless steel cantilever arms, and was submerged Specimens in of uniform thickness and fiber physiological salt solution at 37 °C. In the orientation were isometric mode, both pin connectors were simultaneously loaded in the fiber(F) and attached tightly to stationary cantilever arms; in the ,etpendlcular cross-fiber (La.') directions. Deformation was measureu in the central isotonic mode, one pin connector moved freely back region. aid forth, guided by a sleeve bushing. Equibiaxial stretching over thesame strain Conversion from the isometric to t!e isotonic mode range was imposed in F and CF directions both iu was easy. Isometric experiments were carried out at constant the passive state and during tetanic axialstretch ratios of 1.3, 1.5 and contraction. Representative stress-strain 1.7; while isotonic experiments took place at constant relations are shown below. These data clearly axial loads of 0.5N, 1.0N and 1.5N. One strain demonstrate anisotropy and itsdependence on energy muscle activation. density function was used to describe allexperi- mental data. The function was an exponentialwith 000 a thtse term quadratic polynomial in the exponent. N rotaxant low0,490t Ivor The independent variables were the tangential and 14:4023-h3C4 mkt. amoo 1° axial Green-St.Venant strains. For mostmeasured V'. 0 arteries a reasonable agreement between measured w wo and modeled data was achieved. The agreementwas .1'8 8 not as good as one can achieve whin modeling iso- OA 14 1.2 1.3 1 metric or isotonic data only. STUN 386

1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas AmosszemswimmailmmomMilliMMIN

Biomedical Engineering Scientific Papers

BE13-C.3 I3E13-C.4 Models of Myocardial Tissue Swelling.D.K. Bogen; THE DOUBLE-HEADED CROSSBRIDGE ATTACHMENT University of Pennsylvania, Philadelphia, PA 19104 IN SKELETAL MUSCLE FIBERS A Tozeren% Dept of Mechanical Engineering, Myocardial fluid content, already 80% of the nor- Catholic Univ of America, Wash DC 20064 mal myocardial volume, can increase when any of the fluid compartments (intravascular, intracellu- A simple model of a double-headed cross- lar, and interstitial) of the tissueisincreased. While bridge is introduced to explain the retar- the diastolic stiffness of the left ventricle generally dation of force decay following stretch in The critical increases in parallel with an increase in the fluid skeletal muscle fibers. assumption in the modelis that once one content of the heart muscle, this phenomenon has of the heads of a crossbridge, say head 1, yet to be explained. Insight into this problem can be is attached toone of the actin sites gained by considering several models for nonlinear, available for attachment,the attachment biphasic, elastic materials which undergo large scale of the second head will be restricted to a swelling. These models partition the fluid spaces of level of strain determined by the attach- The unique feature of the the tissue into one or more compartments with ment of head 1. model is the prediction that, in the pres- rod-like and membranous elastic structural ele- ence of a ligand (PPi, ADP, AMP-PNP) and ments. These models can also be conveniently for- absence of calcium, the half time of force mulated in terms of strain energy density functions decayis many timeslarger than the inwhich reflect the amount of tissue swelling and the verse rate of detachment found in solu- relative partitioning of fluid into the compart- tion. This prediction isin agreement ments, as well as the anisotropy of the tissue. When with measured values of half times of force decay infibers under similar con- this formulationisapplied to a simple thick-walled, ditions(Schoenberg andEisenberg, 1985, spherical model of the left ventricle, the swelling Biophys J 48:863-871). Our computations behavior of the diastolic ventricle can be predicted. showthattheapparent cooperativity in One interesting result of the modelisthat the ef- crossbridgebinding experiments (Brenner fects of swelling are strongly dependent on the ecal, 1986, Biophys J 50:1101-1108) can anisotropy of the tissue. be partially accounted by the double- headed crossbridge attachment.

BE13-C.5 BE14-A.1 CollagenFiber Orientation of the Oscillatory Flow of Non Newtonian Diaphragmatic Central Tendon Quantified Blood Analog Fluids in a Curved by SmallAngle Light Scat*Pring, M.S. Artery Model, D. Mann and J. M. Sacks* and C.J. Chuong, University of Tarbell*, Department of Chemical Texas at Arlington, P.O. BOX 19138, Engineering and the Bioengineering Arlington, TX 76019 Program, The Pennsylvania State University, University Park, PA 16802. We arestudying themechanics of the diaphragm to further our understanding The influence of non Newtonian theology on wall of pulmonary respiration. The shear rate in steady and oscillatory flow diaphragmatic central tendon, a through a rigid curved artery model was studied collagenous structure, transmits forces experimentally. Wall shear rate was measured by between the costal andcrural muscles. flush mounted hot film anemometry at 6 loca- Our previous work has shown mechanical tions and the following 4 fluids were investi- anisotropy in its stress-strain curves. gated under nearly identical flow conditions: To explorc the anatomic?,1 origins of the (1) aqueous glycerin (Newtonian) (2) aqueous mechanical anisotropy, we have used polyacrylamide (shear thinning, elastic) (3) HeNe laser basedsmall angle light aqueous Xanthan gum (shear thinning, inelastic) scattering device to quantify collagen (4) bovine blood. For steady flow conditions fiber orientations, since collagen there was little difference at any of the 6 fibers are the major force transmitting locations in the wall shear rate levels measured components in tendon. This technique for the 4 fluids. However, dramatic differences allows precise quantitation of the were apparent for oscillatory flows, particu- direction anddegree of orientation of larly at the inner curvature 180° from the the collagen fibers. Results show two entrance of the curved artery model. For distinct fiber layers, which are example, at that position the peak shear rate oriented perpendicular to each other, for polyacrylamide was 5-6 times higher than for throughout the tendon. Fibers in th. glycerin and 2-3 times higher than for blood. layer at the thoracic side are aligned Aqueous Xanthan gum provided the best model of predominately parallel to the bovine blood over all positions and flow levels surrounding muscle fibers. (Supported by considered. NSF EET8707665 and NIH HL35914). 367 1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas 347 Biomedical Engineering Scientific Papers

BE.14-A. 2 BE14-A.3 VISCOELASTIC STRESS-STRAIN RELATIONSIN BLOOD: The Complex Viscosity of Blood in NarrowCapillaries EFFECTS OF HEMATOCRIT,George B. Thurston; Determined Under Oscillating Flow Conditions, Department of Mechanical Engineering, TheUniversity of D. Schneditz*, V. Ribitsch, and Texas, Austin, Texas 78712, U.S.A. T. Kenner, Department of Physiology and Institute of PhysicalChemistry, Testing the viscoelasticity of wholeblood provides a University of Graz, A-8010 Graz, Austria sensitive indication of theway erythrocytes aggregate and how they deform in flow. These theologicalproperties are strongly dependent on the hematocrit, and because The viscoelastic properties of bloodand of red blood of variations in both cell (RBC) suspensions normal and pathological blood,must be accounted for in clinical are determined by means of a practice. capillary rheometer, where the fluid inthe capillary is A viscoelastic stress-strain profile inwhich both the set Into sinusoidal oscillations. Fromthe detection of viscous and elastic components of theshear stress are plottedvs the volume flow, the pressure gradient the amplitude of the oscillatory shear and the phase strain (measured at a fixed angle between both, the viscous if frequency) shows characteristics ofthe blood sample which and the elastic ti" relate directly to the microstructure of theblood. The elastic component of the complex coefficient ofviscosity Ti* of stress is proportional to the maximumenergy per unit volume the sample is calculated. The frequencyof the osciila- stored (the maximum elasticenergy density developed) during lions is set to the fixed value of 2Hz the cycle per unit of strain. The viscous and the amplitude stress is proportional to of the oscillations Is increased In the energy density lost in eachcycle per unit strain. The elastic discrete steps. The maximum wall shear-rate energy is stored in the process of stretching and orientingthe Itl ranges from 0.13-1to erythrocytes. 200s-1 (fora 1mm i.d. capillary). Making use ofnarrow In the physiologic range of hematocrits, the viscoelastic capillaries (i.d. from 0.8mm to 0.2mm)the heterogene- stress-strain profile retains its basic characteristicinflections ity of the suspension becomes of when the hematocrit is changed by importance.The reduc- adjustment using the blood's tion of the tube diameter leads to autologous plasma. However, changes inhematocrit by a marked decrease of hemodilution with Lactated Ringer'ssolution and Dextran ri' and ti" in aggregating blood samplesespecially at solution produce characteristics which cannot be accounted for low IV. The extent of viscoelasticshear-thinning on the basis of viscosity of the suspending fluid, Is and must be decreased as well. From these resultswe conclude that attributed to changes in erythrocyteaggregation tendency and deformability. RBC-aggregation facilitates oscillatingblood flowinnar- row tubes.

BE14-A.4 BE14-A.5 Influence of Physico-chemicalfactors on the Shear-Induced Platelet Trauma, individual white cell transittime. M. THAO CHAN*, B.B. GUPTA** B. L. Capes and L. F. Mockros*, Northwestern University, * Lab. Hoechst, Tour RousselHoechst, F92800 PUTEAUX Evanston, Illinois 60208, USA ** UTC, BP 649, F 60206COMPIEGNE

Knowledge on the rheological Shearstressiswell-knowntoalterplatelet properties of leu- function and kocyte are essential for theunderstanding of morphology. Quantitative relationships are their functional behaviour inhealth and disea- developed to relatevarious se, including their release fromthe bone nar- indicators of platelet stimulation/lysisto the levelof stress,T, that stress, row, motion and deformation duringphagocytosis. and duration of t. PI atelet-stimul ati on/ lysi s The Cell Transit Time Analyzer(CTA) detected isindicatedby the transit time of individual availability of Pr- -3; changes in platelet cells passing concentration; of through 5 um pores of polycarbonatefilters release serotonin, ADP, ATP andB-thrombogl obul i n; changes (OLIGOPORE) under low hydrostaticpressurec in platelet (0-10 cm of water). aggregability;and platelet loss of LDH and AP. Available data The histogram and the logarithmcurves of fre- collectedwith uniformshear quencies distribution of 2000WBC devicesarefittoafunctionalrelationship of between 0-1 s. the were expressed as well as thenumber of plugging form tra = 10b, with'a'generally falling cells during the transit. between 2 and4 and 'b' relating to the degree of Influence of differentphysico-chemical factors stimulation/lysis and generally falling between6 such as : pressure, temperature, anti-coagulants, and 16. The data include shear stressesthat range and suspending mediawere analyzed. from101 to 104dynes/cm2 andexposure timesthat range The sensibility of theapparatus was demonstra- from 10-3to 104. The developed relation- ted in the 3 populations: controls, elderly, shipscanbeappliedtothenon - uniformshear peripheral vascular diseasepatients. fields typical of medical devices. When appliedto flows through tubes, these relationshipsare shown to predict a degree of platelet traumathatis in reasonable agreement with the measuredtraumathat has been reported in the literature.

:-. . 1,'' 68

348 1988 World Congress on Medical Physics and Biomedical Engineering San Antonio, Texas Biomedical Engineering Scientific Papers

BE14-B.1 BE.14-B.2 Structure-Function Relationships of the Mechanics of the Arteriole Wall, R.S. Chadwick*, Arteriolar Wall, G.E. Sleek*, Barry and J. Ohayon, Theoretical Biomechanics Group, University, Department of Biomedical Biomedical Engineering and Instmmentation Branch, Sciences, Miami, Florida 33161. Division of Research Services, National Institutes of Health, Bethesda, MD 20892. Arterioles undergo dramatic structural changes during vasoconstriction. Little is known about how the arteriolar wall We have developed a two-dimensional mathematical model of directs the rearrangement of its the arteriole wall to help understand the relationship between cellular components as it changes mechanical equilibrium geometry, transmural pressure, diameter. We hypothesize that vasomotor tone, and elastic contains. The model is comprised intercellular junctions between the of three functional regions: an intima represented by a finite endothelium and vascular smooth Im.scle number of endothelial cells each having a finite cross sectional (i.e. myoendothelial junctions - M23s) area; a media described as a finite layer of a continuum of direct the restructuring observed during fluid, collagen, and smooth muscle cells assumed to contract vasoconstriction. An, ultrastructural in the circumferential direction, and; an adventitia considered analysis of hamster cheekpouch as a finite layer of fluid and collagen.With specified arterioles constricted to varying stress-free reference configuration, transmural pressure, elastic degrees was undertaken to test the moduli, and degree of vasomotor tone, the finite deformation hypothesis. Our data show that the of the modelarterioleisdetermined subject to frequency of MEJs in constricted vessels incompressibility contraints and lateral interactions between is almost twice that observed in dilated adjacent endothelial cells. Because the problem is nonlinear arterioles (P< 0.05). A preferential more than one equilibrium solution can be found. We look for distribution of MEJs was also observed the solution with minimal elastic strain energy. Usually this with respect to the deformations which corresponds to a state where the endothelial cell membranes occur periluminally.The data indicate have zero strain energy. Using the model we found that that MEJs may be labile entities, smaller arterioles (having agreater fraction of cross sectional forming and dissociating as the vessel area comprised of cells) are much more efficient at increasing undergoes changes in diameter. Our data their hydrodynamic resistance than larger arterioles. We found also suggest that MEJs act as foci of the model is very sensitive to small changes in media thickness force transduction between the two cell in the sense that a relatively large increase in transmural types of the arteriolar wall. pressure is required to keep the lumen area constant

BE14-B.3 BE14-B.4 v m C,. L_PROPE14 0 si. 414 lean Anatomy and Mechanical Properties 7N SKELETAL MUSCLE. J.H. Lombard , W.J. of MicrovesselsinSkeletal Muscle, G.W. Stekiel, E. Monos, and H. Eskinder; Dept. of Schmid-Sch8nbein , T.C. Skalak, F.A. Physiology; Med.Coll. Wisconsin; Milwaukee, DeLano, S. Chu, Univ. of California, San WI 53226, USA. Diego, La Jolla, CA 92093 and Univ. of Virginia, Charlottesville, VA 22908

A significant component of microvasculer resistance in Theblood vessels in themicrocirculation of the small arteries To skeletal muscle may reside in skeletal muscle are completely surrounded by these vessels. gracilia assess active responses in muscle cells. In that position thedifferent arteries (169+7 pa. o.d.) were isolated from anesthetized hierarchies of arterioles, capillaries and venules rats, cannulated with micropipettes, and mounted in an in havea distinct wall morphology. It contains to be vitro chamber which allowedtransmurel pressure endothelium, basement membrane, a tapering elasti- controllid. Active responses to increases in transmural ca intima and smooth muscle along the arterioles, 1) isometric tangential wall pressure were expressed is pericytes along a major portion of the collecting tension.2) isometrictangential wall stress.and 3). venules, collagen, isicifit cella, and specific tangential wal/ strain. Active tension and isobaric structural featuresinthe adventitia,such as active stress progressively increased in response to 20 nerves andlymphatics. Dynamically the blood mmHg increments in transmural pressure netween 0 mmHg and vessel walls are distensible and exhibit nonlinear Active strain initially 160 mmHg (myogenio activation). viscoelastic properties. Upon a step transmural increased (0-40 mmHg). and then remained fairly constant the diameter shows an instantaneous At control pressure ME of mean pressure through 160 mmHg. elastic expansion which is followed by a slower arterial pressure). active strain in the grannie artery time dependent expansion. At long times a new (-27+22) was significantly higher than that in the larger steady state diameter is reached. The vessels are (-9+22). In contrast, intestinal saphenous artery relatively stiff, and in contrast to blood vessels mesenteric arteries did notexhibit activestrain or in otherorgans, venulesexhibit the stiffest activation.The presence ofactivetoneand myogenic properties of all microvessels. These viscoelas- myogenic responses in small arteries supplying skeletal tic properties have many implications with respect muscle suggests that these vessels may have a significant to blood flow and transport in skeletal muscle. role in regulating vascular resistance inthis tissue. NIB 6HL-29567. and Established MTH 6HL-37374. Supported by NbF Grant DCB 85-16662 and in part by Investigator Award OHO from American Heart Association). USPHS Grant HL-10881.

4-) 1988World Congresson Medical Physics andfornedical EngineeringSan Antonio,Texas 349 Biomedical Engineering Scientific Papers

BE14-C.2 LENGTH-TENSION RELATIONSHIP OF MAXIMALLY Arteriolar vasomotion and changes in perfusioncon- ACTIVATED ARTERIOLES. M.J. Davis'and R.W. Gore, ditions Dept. of Medical Physiology. Texas A&M Univ.. College D.W. Slaaf: H.H.E. Oude Vrielink, G.J. Tangelder Station, TX 77843 and Dept. of Physiology. Univ. of Arizona Tucson. AZ 85724. and R.S. Reneman. Laboratory for Miarocirculation, Segmental differences in the microvascular response University of Limburg, Maastricht, TheNetherlands to physical and chemical stimuli may possibly be explained in terms of the length-tension relationship of The arteriolar diameter varies rhythmically(vaso- vascular smooth muscle at different points along the motion) even under seemingly stationaryconditions. vascular tree. To test this hypothesis. we isolated single The relation between vasomotion and arterialpres- arterioles for study in vitro so that arteriolarpressure sure reduction and vasodilation was studied in the and activation state could be precisely controlled. 1st-. rabbit tenuissimus muscle using intravitalmicro- 2nd-, and 3rd-order arterioles (IA,2A,3A: i.d.= 20.60 pm) scopy. Vasomotion has a limited affecton the re- were dissected from the hamster cheek pouch and sistance of the transport arterioles (TA),but its cannulated with micropipettes using methods described considerable effect on the resistance by Du ling and Gore (Am. J. Physlol. 241:H108, 1981). in terminal Pressure-diameter curves were recorded for maximally- arterioles dominates the flow pattern inthe TA. activated vessels in high K+ and norepinephrine solution. Reduction of perfusion pressure leadsto effective and for passive vessels in Ca++-free solutionover an vasodilation and a gradual increase in cyclelength internal pressure range of 5 to 400 cmH2O. Maximal (CL) and amplitude (A) of vasomotion.Vasodilation active wall tension varied from 600 (3A) to 1600 (IA) as a result of vascular relaxation (adenosinesu- dynes/cm. but maximal active wall stress varied only perfusion) did not Atsult ina change in CL, where- from 4 (1A; to 6 (3A) x106 dynes/cm2. The average value as A decreased. Thus, vasodilation per sfi doesnot for wall stress is very similar to that obtained in hog result in increased CL and A. Arterialpressure and carotid artery (Dillon et al.. Science 211:495. 1981). These blood cell velocity (normalizedto diamete )had results support the concept that srlooth muscle from the strongest relation with theincrease In CL. vessels of different size is mechanically similar and that Increases in Cl over more than 40% occurred longitudinal differences in arteriolar wall stress in vivo if ar- may explain the response gradients observed in the terial pressure was considerably reducedwith a microcirculation. Wall buckling in arterioles at short concomitant reduction in flowor increase in effec- radii broadens the peak of the active stress-length curve tive diameter. CL and A do not changeas a result and may extend the pressure range over which arterioles of simple scaling. The major changeseems to be a arc most sensitive to physical and chemical stimuli. delayed onset of the constriction phase,with minor Supported by NIH HL-13437. 1'1,38104 and Texas AHA adjustments of the basic vasomotionpattern. #85G-063. Supported by MEDIGON 900-517-157

BE14-C.3 BE14-D.1 Myocardial microvascular properties andthe prediction of coronary flow, Leukocyte Behavior at a Bifurcationof J.A.E. Speen.. Arteriole in the Rat Mesentery, University of Amsterdam,Department of Medical Physics. The Netherlands. Masaaki Sato*, Shin-ichi Sakurabaand Norio Ohshima, University ofTsukuba, One would liketo understand the mechanics of Tsukuba, Ibaraki 305, Japan flow in coronary arteries and veinsfrom basic properties of themicrovessels and their mechanical loading by cardiac contraction. Leukocyte bahavior at a bifurcationof arteriole in the rat mesentery was analyzed Under normal conditionsintramyocardialblood by introducing a fluorescent dye technique. volume is appr. 12% which amount is increasedby Diluted acridine orange was intravenously infused to 25% when the heart is arrested.Assuming that observe the tissue pressure falls from ventricular lumina' leukocytes in the microvesselsunder a pressure at the endocardiumto zero in the fluorescence-TV intravital microscope. Blood flow velocities and leukocyte epicardium onecalculatesanintramyocardial flux were measured compliance value of 0.1 ml/mm Hg. This is close at 36 arteriolar bifurcation sites. The experi- matul rebults ate summarized tothe valuederivedfrom the product of in the form of a capillary distensibilityand intramyocardial leukocyte distribution function whichis defined bloodvolume. Compression of this intramyocardial as the leukocyte flux as a functionof the bulk flow into the daughter vessels. compliance by heart contraction explains well the The leukocyte phasic flows in coronary arteries and veins. The distribution function was nonlinearand the partitioning of leukocytes flux intramyocardial compliance leads to over- Letween daughter estimationof zeroflow pressures if venous branches was strongly dependenton the eccentric outflow has not ceased. position of the cells at theupstream entrance to the branch. A further analysis of coronay dynamics requires Only the leukocytes flowingat a more information on myocardial microcirculation central region in theupstream parent arteriole especiallyon the non-linear pressure volume were divided between two branchesaccording to a relation of the microvessels and the distribution sigmoidal distribution functionwhich generally of pressure and flow velocities in them. expresses "preferential distribution" ofpartic- les at vessel bifurcations.

4.) Ph! *.S4 v

350 1988 World Congress on Medical Physics andBiomedical Engineering San Antonio, Texas Biomedical Engineering Scientific Papers

BE14-D.2 BE14-D.3 Wall shear rate and velocity profiles in arterio- Estimation andAnalysis of Platelet- les. D.W. Slaaf: G.J. Tangelder, T. Arts and T.S. Analogue Concentration Profiles in Blood Reneman, Laboratory for Microcirculation, Univer- Flow, J.Shane Kippenhan*, Joachim H. sity of Limburg, Maastricht, Ths Netherlands. Nagel, Eugene C. Eckstein, Dept. of Biomedical Engineering, University of Velocity profiles were determined in vivo in Miami, Coral Gables, FL 33124, USA. rabbit mesenteric arterioles and used to calculate least estimates of the actual wall shear rate in Radial concentration profiles of platelet-sized these microvessels (17-32 um diameter). Fluores- particles in blood flow through small tubes were cently labeled platelets, used as natural markers estimated from groups of observations of particle of flow, were visualized by intravital fluores- distances from tube walls. Methods based on the cence video microscopy. Velocity and mean radial kernel method of probability density estimation position of platelets flowing within a thin opti- allowed easier quantitative characterization of cal section around the median plane of the vessel, the near-wall peak that often occurs in the were determined. Data points could be obtained as distributions; such peaks were difficult to close to the wall as 0.5 m. The velocity profiles quantify with p evious hiLcogram methods. These were flattened as compared to a parabola. To sa- methods also yielded a higher spatial resolution tisfy the no-slip condition a thin layer of fluid of the distributions' features. with a steep velocity gradient near the wall has Additionally, curve fitting of the Fourier to be assumed. Least estimates of wall shear rate transform of the density estimates was used to (WSR), as calculated using the mean velocity gra- model the experimental data. This Fourier dient 4 this layer of fluid, ranged from 472 to transform represents the "characteristic function" 4712 s-1 (median 1700 s-1). These values are at that describes all moments (mean, variance, etc.) least 1.46-3.94 (median 2.12)times higher than of the data's distribution. Such analytical those calculated for a parabolic velocity profile descriptions of the distributions, based on two (WSR-P) with the same volume flow. The ratio WSR/ or three parameters, enable compact represen- WSR-P might also be used as a least estimate of tations of sets of data, thus facilitating the the apparent viscosity of blood in the arterioles study of the rc -Jnship between these parameters relative tothe viscosity of plasma. The median and rheological conditions. value of 2.12 compares well withthe value of The above methods can be applied whenever approximately 2.34 as found in cat mesentery using handling data that consists of lists of numerical measurements of pressure gradients. observations in one or more dimensions.

BE14-D.4 BE14-D.5 OPTIMUM Capillary Network in the Skeletal Muscles LIPID SOLUBILITY INCREASES THE for Oxygen Transport to Tissue, A. Kamiya: A. Oota SENSITIVITY OF INDICATORS FOR and M. Shibata, Res. Inst. Appl. Eletr. Hokkaido ENDOTHELIAL DAMAGE IN THE LUNG. University, 060, Sapporo, Japan T. R.Harris*,L.E.Olson, and A.Pou, Vanderbilt University, Nashville, TN 37232. Theoptimumcapillary network for oxygen (0,) transport to tissue was estimated using Krogh f s Tracers suchas "C-urea (U) can be usedinindi- cylinder mode) in the skeletal muscle. The tizsue cator-dilutioncurvestoassessdiffusivity-surface area mass supplied by a single capillary in the center product (D S) for the lung vascular capillary wasestimated as the area of positive 02 tension barrier. A persistent problem is the need to separately under a givencondition of capillary flowand assessdiffusivity and surface area. We have proposed tissue O consumption rate. The total mass was thattracerswhich werebothlipophilicand hydro- determinea as tl.e function of the capillary number philic might be in such separation. Recently, we and total blood flow.The energy cost to maintain proposedthat 1C-propanediol(P)couldbe usedto the vascular system with n terminals kcapillaries) eliminate thy effects of surface area by observing the was assessed by the minimum volume model by Kamiya ratio of DS for U tc,thatfor P (Harris etal.L, and Togawa (1972). The efficiency of the entire Apol. Phvsiol. 62,1852,1987). This work showed that systemwas evaluated by calculating the ratioof the square of theratio of D1 /ZS for P and U (R2) total tissue mass/ energy cost. The results of the decreases from 1.7 to1.06 when capillary permeability calculation using physiological data of muscle is increased by infusion of alloxan in isolated perfused blood flow and 0 consumption rate in man revealed dog lung preprrations. Theory developed inthis paper 2 the optimum capillary number of (5 - 10)x109 and showedthatthisimplieda5-foldincreaseintrue the capillary density of 200 - 500/mm? which coin- capillarypermeability.Thistheoryalsopredictsthat cided wellwith the physiological estimates of R2 changes frombaselineshouldbe evengreaterif these values in the human skeletal muscles. It was linidsolubility of thediolishigher. Thus, we used concluded that the capillary network system in the C-butanediol asatracer and compareditto U in skeletal muscle is constructed so as to attain the another group ofisolateddog lungstudies.Results highest efficiency in 02 transport to tissue. showthatR2 altersfrom3.5to'.3afteralloxan indicatingagaina5-foldincreaseintruecapillary permeability. Theseresultssuggestthatpartiallylipid soluble tracers when used with urea provide sensitive measures of lung vascular damage. 3 7 ji

1983 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Twos 351 Biomedical Engineering Scientific Papers

BE14-D.6 Measuremmts BE14-E.1 of Single CapillaryPermeability of Macromolecules in Skeletal Cell Poker Measurementsof the Mechanical Proper- Muscle ties of Viscoelastic M. Shibata * and A. Kamiya, Layers, M. Duszyk, B.Schwab, Research Instituteof Applied E. Elson and G. Zahalak*,Washingt,p University, Electricity, St. Louis, Missouri 63130, Hokkaido University, Sapporo060, JAPAN USA We have recently developeda new instrument, the Transport mechanism of macromolecules acrosscap- Cell Poker, for measuring illary wall has been cellular deformability, debated by severalinvestiga- and applied it tomeasure the stiffness of tors. To clarify this mechanism, many dif- we tried to es- ferent types of cellsincluding erythrocytes, tablish a new fluorescent intravital blasts, lymphocytes, fibro- method microscopic and islet cells. In order to for direct measurementof single capillary increase confidence permeability in the technique and alsoto es- of macromoleculesbased tablish a basis for "double on the interpreting cell-pokermeasure- pore theory". The rabbit ments in terms of fundamental muscle was used tenuissimus mechanical properties for themeasurement of it was deemed desirable ti-sue local to test a material whichhad concentration curve following the intra- been well characterizedby other means. venous injection of FITC-conjugated Albumin (FA67, Thin (0(1mm)) layers ofthree polydimethylsiloxanes M.W.= 67,000) and Dextran(F0155, (PDMS) were prepared from M.W.= 155,000) and tested by indentingtheir video-image of microcirculationobtained surface with pokers (0(10pm) Argon by Diameter) driven by laserexcitedfluorescent triangular displacement microscope-TV waverorms. The nominal val- system. The measured tissue ues of the viscosities of concentrationcurve the three PDMS samples was best-fitted with the theoretical curve in the were respectively 12.5, 30.0,and 60.0 Ns/m2. parabolic partial From differential equationand the measurements of the force andmagnitude of indent- diffusive(D) and convective(V)cumonents ation at the poker tipit was possible to transport of determine across the capillary wall were the relaxation functionof the PDMS. mined. deter- This function Capillary can be characterized by three permeability coefficient(P)was parameters: the short- calculated from D valueas P=0/X, where X is and long-time elastic diffusion the moduli, G, and G., andviscos- distance (length of pore). P values of ty n. Determination of q by FA67 and F1155 two different methods obtained tissue concentration yielded (14.5, 24.8,65.2) and (15.4, 34.7, curves were 9.5'15.9 and 6.8±3.1, Ns/m2, in 78.4) which values reasonably goodagreement with the nominal were well consistentwith the values. The long-time modulus, values obtained G. appears to be whole organ measurements. lated to the surface re- It was concluded that tension at the PDMS-water present method boundary. is useful to measure the local capillary permeability ofmacromolecules.

BE14-E.2 Curvature Elasticity of Mechanically-formed BE14-E.3 Lipid Membrane Tethers, Richard E. Waugh*, Electropermeabilization and Fusion of Adherent University of Rochester School of Medicine LipidVesicles:David Needham: Department of and Dentistry, Rochester, NY 14642. MechanicalEngineeringand MaterialsScience, Duke University, Durham, N.C. 27706. A tether is a thin cylinder of membrane that can be formed from the surface of a cell or bila3er Lipid bilayer membranes in close enoughcontact vesicle by pulling on a small region of the surface will mix and "fuse" under theaction of appropriate (point attachment). Early observations indicated destabilizingstimuli. Inour experiments, that close tethers exhibit elasti^ character, and that contactisestablished byattractiveforcesbetween their radius decreases as Cie axial forceon the membranes (e.g. van dcr Waals forces, the presence tetherincreases. The coefficient thatrelates of non-adsorbing polymer and positive the axial force to the tether radius is the diclectrop hores is). Membranebreakdown and membrane bending stiffness. We have developed a destabilization of contactismade to occur via an new 'mental approach to measure the force and applied d.c. pulse. We use a technique of nucropipet radius tethers formed from vesicular membranes manipulation inorder toexper::. ant on single lipid to determine the membrane bending stiffness and vesiclesorsingleadherentpairs. Thelipid test proposed equilibrium relationships between composition,mechanicalpropertiesandadhesive the membrane tension and the tether force. The interactions of these vesiclesiswell characterized. force is generated by the negative buoyancy ofa A double chamber with transfer pipetallows vesicles glass bead that has adhered to the surface of the to be selected from one chamber, transferedIntothe vesicle. The tether radius is calculated from the adjacent chamber, containingthedeciredsolution displacement of the projection of the vesicle into and placed between parallel Pt electrodes. the pipet. Preliminary measurements of force and Electropermcabilizationandfusion is assayed via tether radius fr.,m phospholipid membranes of directobservation,andgeometryandstructureo f stearoyl-oleoyl-phosphatidYlcholine indicate a thepost-fusion product are determined by bending stiffness of these membranes between 1.7x microfipetsuction. Initialresultssuggestthat 10-19 and 3.0x 10-19 " m. They also indicate internal volume iscon, rvedwhilst excess that the relationship between aspirationpressure memoraneareaistaken up as anadherent'flap' and tether force predicted on the basis of thin spanning the lumen of the sphericalfusionproduct. membrane theory is not valid. This 'flap' can be peeled apart bythe applicationof pipet suction and the membrane ispulledintothe surface.

01 P4 d 352 1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas Biomedical Engineering Scientific Papers

BE14-E.4 BE14-E.5 ANALYSIS OF THE STREIGTH OF ADHESION Detachment of Rad Blood Cells fro:, BETWEEN A PAIR OF CYTOTOXIC T-CELL Adhesive Contact with Rigid Substrates AND TARGET CELL by Hicropipet Suction: Experiment and Analysis, D. Berleand E. Evans, Aydin Tozeren: u -Li Paul Sung and Shu University of British Columbia, Chien, MechanicalEngineering Department, Vancouver, B.C. V6T 1W5. The Catholic University of America, Wash- ington DC 20064 and Department of Physi- When cells form adhesive contacts, the chemical ology, College of Physicians and Surgeons, affinity that causes the contact to spread is Columbia University, NY NY opposed by the work of cell deformation. Likewise, whyadhesive contacts are separated A micromanipulation method is used to by the action of external forces, the determine the strength )f adhesion between mechanical deformability of the cell moderates a pair of cells by measuring the energy the action of those forces and affects the per unit area that must be supplied to application of stresses to the contact region. reduce the region of contact between the Thus, the cell acts as a transducer or coupler cell pair. We show that the strength of between the applied force and the adhesive adhesion is not constantduringpeeling contact. We have recognized this feature in but increases with decreasing contact area the design and analysis of experiments to for pairs of cytotoxic T-cell(FI) and measure the physical properties of adhesion. target cell (JY). This new finding sup- In these experiments, suction pressure from a ports the notion that the crosslinking micropipet is used to detach adherent red cells protein molecules slide towards the from rigid surfaces. We employ a finite- leading edge of the separation while re- differenze computational algorithm to solve the maining attached to both cells. The equations of mechanics governir red cell driving force for this movement is the deformation and provide a complete description elastic energy stored in the crosslinks by of the detachment process. Consequently, we the membrane tensions. The strength of are able to explicitly measure the energy per adhesion of Fl-JY pair is found tobe unit area required to separate red cells three to six times larger than the corre- agglutinated by antibodies or le)ins. sponding strength of Fl-Fl pair.

BE14-F.2 Micromechanical Manipulation of Electrically Fused RBC Membrane Shear Modulus of Elasticity Derived Red Blood Cells. Donna Miles-McCollum", Dopt. of from Observations of Steadily Tank-Treading Cells, Biomedical Engineering and Robert M.Hoc,unuth, S. Sutera*, P. Pierre and G. Zahalak, Washington Dept.of Mechanical Engineering and Materials University, St. Louis, MO - 63130, USA Science, Duke University, Durham, NC 27706. An algorithm has been developed for the approximate calculation of the membrane stress distribution and A system previously developed for the fusion and the internal pressure of a steadily tank-treading r icromanipulationofredbloodcellshasbeen red cell (Tran-Son-Tay et al, 1987, Biophys. J., modified to allow a more "gentle" fusion process and 51:915). The algorithm is based on an idealized a variation of conditions forcellular manipulation. ellipscidal model of the tank-treading cell (Keller Cellsaresuspendedina200mOsmsalt-sucrose and Skalak, 1982, J. Fluid Mech., 120:27) joined solutionand flushedintoa 250gm thickchannel with experimental observations of cell dimensions lined with Pt electrodes0.75 mm apart.The cells are (projected length and width) and tank-treading fre- aligned in a AC field and fused inlargequantities quency. The resulting expressions for the membrane with four DC pulses of 3-5kV/cm height and 20-251ts stress resultants, why! applied to the two zurva- width. Several types of fusion products are obtained ture extrema of the cen.ral streamline, reduce to depending upontheextent of fusion. The most two simultaneous equations which are explicit in gentlyfusedcellsareattached by atether while pm, the membrane sh)..ar modulus of elasticit and other pairs of fusedcellsare joinedbyalarger implicit in the cytoplasmic pressure, pc. Solu- tunnel-likestructureorafullyswelledlumen. tions of these equations for a series of experi- These results suggest that the fusion process begins mentally observed motions, corresponding to graded with point lipidattachmentsandperturbationsand levels of alplied shear rate, 1, yield values of then progressestospectrininvolvement. The fused Pm and pc which increase monotonically with the cellsareflushedinto a largermanipulation applied shear rate and, thus, the overall cell chamber with a 295mOsm solution which results is a elongation induced by the shear flow. Over the partialdeflationofthecellsandtalows the range of t covered, the calculated pm varies from connectingstructurestobe viewed more clearly. 0.5 x 10-3 to about 5 x 10-3 dynes/cm, correspond- Micromanipulationisbeing usedtoinvestigate the ing to values of the principal extension ratio extentof communicationbetwe' jthecells of an from about 2 to 3. According to these findings, apparent fusedpairand tocharacterizethe based on dynamic membrane equilibrium consider- mechanical properties of the fusionsite. ations, the red cell membrane exhibits strain- stiffening elastic behavior.

19S8 World Congress on Medical PhysicsandBiomedical EngineeringSan Antonio, Texas 353 Biomedical Engineering Scientific Papers

BE1.4-F,3 BE1.4-F.4 Determination of r.b.c. Membrane Elastic Modulus from Regulation of Red Cell Membrane Deformability Viscoelasticity Measurements ona Diluter.b.c. and Suspension Stability by Skeletal Protein Newtork,N. Mohandae X.Y. Lin & D. Barthes-Biesel; Universite Department of Laboratory Medicine andCancer de Compiegne, Research Institute, University ofCalifornia, San U.A. CNRS 858, BP 649, 69206 CompiegneFrance. Francisco, California, 94143, USA.

The motion and deformation of a r.b.c.suspended in a Biophysical and biochemical evidenceaccumulated during shear flow may be qualitatively representedby means the last decade suggests thata membrane skeletal protein e a capsule model (i.e. a liquid dropletsurrounded by a network consisting of spectrin, actin,and protein 4.1 linked deformablemerrarane). There existspresently a to band 3 in the fluid lipid bilayerthrough ankyrin is complete model of the time cependentbehaviour of a responsible for various red cellmembrane material spherical capsule suspended in properties. Recently, a number ofinvestigators have a linear flow, that attempted to define the key skeletal includes the determination of the bulkrheological protein associations by relating measured changes in membranedeformability and properties of a dilute suspension of suchparticles. In membrane ir,tchanical stability particular,the loss and storage moduli to well-defined skeletal may be protein abnorn.alities in red cellsfrom congenital hemolytic expressed exactly in terms of the intrinsicproperties anemias with membrane dysfunction.From these studies, of the particles. the membrane skeletonemerges as a network of folded The aim of this work is to show how this modelmay be spectrin molecules held togetherby protein-protein used to interprete oscillating shearmeasurements on a associations which occur at bothends of the spectrin dilute (haernatocrit between 5% and 10%)suspension of beterodimer. Disruption ofeither the r.b.c.'sina viscous medium (Dextran, viscosity spectrin-actin-protein 4.1 complexat one end of the between 7 and 10 mPas). The frequency spectrin molecule, or of spectrinself-association at the other range varies end rosulted in a marked decrease from .05 to 1 Hz. The suspensions exhibita measurable in membrane mechanical viscoelasticity and a stability. On the other hand,an increase in inter- or comparisonbetweenthe intramolecular associations of the experimental and theoreticalmoduli leadstoan skeletal proteins or evaluation of the membrane surface shear increased association of integralmembrane proteins with the. elastic skeletal network resulted in profounddecreases in membrane modulus which varies between 10 and 2 10-6N/m,a deformability. These data imply that value comparable with different skeletal protein the results obtained from interactions regulate membrane deformabilityand mechanical micropipette experiments. stability.

BE14-F.5 An Adaptive Pulse-editingTechni9ue for Accurate BE14--G.J. Aperture Impedance Cell SizeAnalysis and Erythrocyte Models of Capillary Blood Flow Deformability Measurement. T.C.Fisher; A.C.Fisher. R. Skalak*, N. Ozkaya, M. Sugihara-Seki Bioengineering Unit, University ofStrathclyde, Glasgow, Bioengineering Institute, Department of G4 ONW, Scotland, U.K. Civil Engineering & EngineeringMechanics For most haematologicalapplications, cell volume Columbia University, New York, NY10027 analysis is performed by theaperture impedance (Coulter) technique. The underlying principle is that the changes Several theoretical models of capillary in resistance (detectedas voltage pulses) as cellsare blood drawn through a minuteaperture are directly proportional flow have been evaluated by computational to the cell size. This is inan approximation, and volume vestigation of the influence of various distributions obtained in thismanner are skewed by the factors. inclusion of anomalous pulses. The effects of elasticity of the redblood cell devised to deal with thisVarious methods ha'been problem, but are generally membrane are shown to leadto reduced pres- expensive and/or complex andnot applicable to standard equipment. This system comprisesan interface which sure drop at high flow velocities. Computa- digitises the height and durationA pulses output froma tion of the entrance flow intoa narrow cap- cell counter and phases this dateto a microcomputer. All illary shows that the deformability subsequent processing is performedin software. The of the cell pulse-editing technique is basedon the relationship of allows smooth entry with littlechange in mean pulse height and duration andwill be discussed in cell velocity. The stability of multifile detail. The major advantage isthat it is auto-adaptive; (double i.e. the criteria for artefactrejection are established file) flow of red blood cells (2Dmodel) is in- by statistical analysis of theincoming data, rather than vestigated numerically. Rigid cells being pre-determined. Usedwith a basic cell counter, appear to performance comparesvery favourably with expensive be neutrally stable in double fileflow, but de- haematology analysers. Anexiting spin-off is that this formable cells may have preferredconfigura- technique also providesa rapid and simple means to measure erythrocyte deformability. tion.

354 1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas Biomedical Engineering Scientific Papers

BE14-6.2 13E14-6.3 Modelling the Mechanics of White Blood Cells: A plate model for determining the elastic modulus of shear- Elongational Flow and Passage through a Stenosis X. Z. Lit,D.Barthes-Bieseli`, C. Thao Chan2, exposed endothelial cells,C. Byrd, R. M. Nerem, and L. 1 Universito de Compiegne,BP649, F60206 Compiegne, T. Wheeler*, U. of Houston, Houston, TX 77004. 2 Lab. Hoechst, 3 Tour Roussel, F92800 Puteaux Experimental studies have shown that endothelial cells A white blood cell is modelledas a sphericalcapsule, i.e. maintain a flattened and elongated shape following exposure a liquid drop surrounded by a viscoetastic membrane. A to shear stress and subsequent detachment. Their mechanical numerical technique is used to predict the tiwilcn and the properties, which are studied using micropipette deformation of the cell in twosituations: case1,when it is suspended in an unbounded axisymiaetric elongational experiments, are influenced by the level as well as the flow;case2,when it is forced to squeeze through a duration of the shear stress. In a recent study, carried out in constriction with a smaller diameter. In both case:, the collaboration with D. P. Theret, M. J. Levesque, and M. internal viscosity of the cell and that of the suspending Sato (to appear, J. Biomech. Engrg.), th the aid of an liquid are equal. axisymmetric homogeneous halfspace model suggested by In easel,the cell bursts when the capillary number the experimental observatinnc, we detennir.e an expression (ratio of the external viscous forces to the membrane for an effective Young's modulus for such cells. In the elastic forces) exceeds a critical value. The membrane viscosity only affects the characteristic response time of present discussion we report on progress toward an the cell to a change in flow conditions. axisymmetric plate model. This model uses the solution of In case2,the resulting deformed capsule shapes resemble the halfspace model as the initial step in an alternating those of w. b. c. going through Nucleopore filters. Cell scheme. Inthe second step, a traction boundary value break-up may occur, but the higher the membrane problem is used to remove the tractions from the plane viscosity, the higher the critical capillary number. The corresponding to the free surface of the plate and in the third additionnal head lose through the stenosis increases sharply with membrane viscosity. This model may be step a mixed halfspace problem is solved to correct for the reed togain aqualitative understanding of the dynamics of perturbations to the boundary conditions on the loaded face w.b.c.'s subjected to hydrodyramical forces. introduced in step two. This three step process is repeated until the desired level of accuracy is obtained.

BE14-6.4 BE14-0.5 Contractile Cortex-Liquid Core Model A Nev Method to ,%easure Ac.o..ve Cytoplas- for Blood Phagocytes: Effect of micStrain in Human Neutrophils, S. Augmented Dissipation in Cortex, Simon' and G.W.Schmid-Sch8nbein, AMES- E. Evans.and A. Yeung, University of Bioengineering, University of California British Columbia, Vancouver, B.C. V6T 1W5 San Diego, La Jolla, CA 92093 USA.

Extensive evidence shows that the active A requirement for neutrophil locomotion and contractile apparatus in blood phagocytes is phagocytosisis active cytoplasmic deformation. concentrated in a cortical layer adjacent to the To understandhow the nonuniform actin-myosin outer plasma membrane. We have shown that these contractile elements in the cytoplasm effect cells deform passively as very viscous liquid spontaneous pseudopod formation and locomotion, we suspensions of organelles and granules. A developed a new method to measure planar Lagrang- significant question is what is the viscous ian and Eulerian strains from the displacement resistance of the cortical layer compared to the history of a triad of internalized latex markers. interior cytoplasm'? We have developed a variat- The microsphere displacement histories were ional method to predict the entry of a spherical recorded on video along with the cell contour. cell into a tube where the cell structure is The triad's principle strains, orientation, modelled by a viscous liquid core couled to a stretch ratios as well as the overall cell stretch viscous-contractile outer shell. Here, we ratios and orientation were computed. The intra- present results which show that when there is no cellular strains were found to be large relative drag at the tube wall, augmented dissipation in to the overall cell shape change. Principle the cortex has little effect on the rate of entry stretch ratios (X=1/1 )were predominantly until the viscous properties of the cortex are tional inthe direction of pseudopod formation very large. Therefore, the rate of entry of (X= 2.6 ±1.5) and contractile in the perpendicular 1/41A-like spherical cells into tubes versus direction (X =.7t.4). Regional strain analysis in suction pressure provides a direct measure of the the vicinityP of the pseudopods revealed moderate effective viscosity of the interior suspension of contractile strains. The transitionalregion, cytoplasm and organelles. Data for flow of between pseudopod and cell body, exhibited large single blood granulocytes into suction shear strains. It appears that the motive strain micropipets yield values on the order of 10 3 field is small in the gelled pseudopod region and poise for the effective viscosity at 25°C. large in the cell body. Supported by USPHS grant HL-10881. 3 7 5'

1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas 355 Biomedical Engineering Scientific Papers

BE1544.1 BE15-A.2 INTRACARDIAC FLOW VECTOR MAPPING BY 2-D DOPPLER Could color Doppler LCHOCARDIOGRAPHY areasof regurgitant jets have a physical significance ? A.Kitabatake *,J.Tanouchi, K.Ishihara, K.Fujii, B. Diebold, A. Delouche, Ph. M.Uematsu, Y.Yoshida, Delouche, J.P. H.Kusuoka, M.Hori, Guglielmi, Ph. Dumee and P. Peronneau*. M.Inoue, T.Kamada The 1st Dept. of Medicine, INSERM U. 256, Nopital Broussais, Paris, France. Osaka Univ. School of Medicine, Osaka, Japan

Color Doppler of To investigate the intracardiac flow dynamics evaluations intracardiac regurgitant jets are often used for in terms of regional flow direction and absolute assessing valvular insufficiencies, but the flow velocity, we have newly develope. two-dimen- physical relationship betweenthe velocity fields and the sional (2-D)flow vector mapping by 2-D Doppler severityof the leaks hasnever been studied. echocardiography. On the basis of the equation of Our invitro modelwas designed to measure the continuity, we determined orthoc,onal velocity velocity fields in a pulsatile turbulent jet with components to the ultrasound beams from the paral- vi.rious peak velocities (v = 2.5to5.5 m/s), lel velocity components to the ultrasound beams orifice diameters (d = 5.8 to 11.3 mm) and receiving measured by 2-D Doppler echocardiography, assuming tube diameters(D = 16to 30 mm),simulating a that there is no perpendicular flow component to large variety of valvular leaks. The results the echo plane. Weapplied rhis nethod to left demonstrate, at peak velocity, the presenceof ventricular (LV) flow in normal subjects and pa- a central core within which the velocity is uniform. tients with LV dysfunction. Results were. Innor- Below a threshold velocity (Vt), this core is mal subjects, all intraventricular flows con- free : itslength (1) is independent of v,and verged towards the outflow in midsystole. In proportional to d (1 = 6.2 d 0.4 (mm),r = 0.99). patients with LV dysfunction, outflow streams Vt is due tothe receiving tube were not so converged, : Vt = 14.4 d/D hecause of remnant of + 8.3 (m/s)(r = 0.95). Thus, inflow in the absence of in midsystole, al houghoutflowtract aliasing, color Doppler measurements of the central abruptly narrowed morphologically. Absolute veloc- core of regurgitant free jetswould allow an ity and convectional acceleration were signifi- accurate evaluation of the severity of the leaks. cantly lower in patients with LC dysfunction than those in normal subjects. Flow vector mapping isa new technique to provide a quantitative flow mapping in cardiovascular system.

BE15-A.3 BE15-A.4 A Feasible Method of Measuring the Absolute Flow-rate of a Regurgitation, Shunt, and Detecting STI from Blood Flow Ve_....ity measured by Doppler Ultrasound and Stenos4d Flow by Colv Doppler, M. Sugaware, ECG, Y. Seo and H. Bongo, The Heart Institute of Wei-qi Wang*,Qi-wing Song and Qian-ring Japan,Tokyo WrIen's Medical College, Tokyo, Shao,Dept. of Elec. Engn.,Pudan Univ., Shanghai,China 162 Japan,and Medical EngineeringLaboratory, Toshiba Corporation, Otawara, 329-26 Japan. The ratio of Fe-efection Time(PEP) and LeftVen- tricular Ejec The issuing flow-rate of a jet, Q, is on Time(LVET) is called Systolic given by Time Interval(STI). We explored Q.vA, where v is the issuing velocity of the a new method for n detecting SPI only with two biosignals,ECG jet anJ A isthe cross-sectional ancl area of the blood flow velocity. orifice. Hc-.;,ver, the measurement of A ofa jet in the The method consists of two parts: :ardiovascular system, such as a I.Utilize Doppler ultrasound to detect blood regurgitation, shunt, or stenosed flow, flow is velocity and detect ECG simultaneously in difficult. On the assumption that the jet is order to obtain PEP and LVET. "free", the following relations apply between IT.The computer control whole system,recognize the diameter of the orifice D, thelength of the R wave,initial and end points of blood flow velo- laminar core of the jet L, the centerline city waveform with software. Let R velocity of the jet vat an arbitrary distance wave as a x(>L)from the orifice,and vn: L=6.8D and synchronous point,accumulate the velocity of blood flow. Display the two biosignals. v/v=L/x. From these equations Calcu- x n 2 we obtain late and print PEP,LVET,STI,HR. Q=0.u17(xv ) /v -Jets into a water baththrough The method has been used in clinic for cardiac orifices With various diameterswere studied. The issuing flow-rate wasmeasured with function examination at three hospitals in al. Shanghai. ultrasonic flowmeter mounted aroundthe tubing connected to the orifice. vn, x, and vx were measured by using various functions ofa color Dop0.er flow imaging system. There was a good linear correlation between the Q'sestimated from the above equation (Y)and measured with the flowmeter (X) irrespective ofthe diameter of orifices: Y=1.01X40.06 (r=0.96,SE=0.19).

356 1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas Biomedical Engineering Scie cific Papers

BE.15-A.5 BE15-B.1 The Pre - election- period of the Fetal An automated computer-run Dopplerprocess for Heart: Noninvasive Detection during measuring experimental cardiac output.An in-vitro Labor with Doppler Ultrasound and study. D. KALMANSONt C.VEYRAT, H.VUTHIEN,S. GAR- Phonocardiogram, B.C.Yue,Yang-ming CIN, G. PALLOT, Y. LEMERDY. FondationROTHSCHILD, Medical College, Taipei, Taiwan,R.O.C. Paris - FRANCE. CNAMTS & ARNTIC. The purpose of this study is to measure A Doppler velocimeter (ATL051)vas fed to a compu- the pre - ejection - period(PEP) of the fe- ter (Normerel AT+) anda software devised to auto- tal heart during labor. The PEP is de- matize entirely the recording andcalculation of fined as the time interval from the on- flow rate. The Doppler probewas fixed at a cer - set of ventricular depolarization tc tain distance of a pipe (f1:20 mm), whosepulsatile the opening ol the aortic velve(() AO) flow could be set and made variableby an electric The PEP is of interest because it is pump. The Doppler gatewas first positionned at related to myocardial contra -tivity. It the centre of a cross section ofthe pipe, and is measured with tho combination use of thereafter, all the processing becameautomatic. abdominal ECG and the audio output of 1/A series of points of theelliptical cross-sec- ultrasound Doppler cardiogram. BandpasE tion was determined.2/And ellipticalregression filter with cutoff frequencies at 750 then calculated the centre, diametersand the angle and 1000 Hz is designed to locate the & between the beam and directionof flov.3/ About component wave that correlates the open- 30 points withinthe cross-sectionwere explored, ing movement of the aortic valve(A0). and their mean tippler frequency shiftrecorded. Fifty pregnant woman with a gestational 4/ These frequency shiftswere averaged, with a age between 36 and 40 weeks are studied. correction for the angle 8 for eachpoint. An or- AO is detected in 84 per cent (42 of 50). thogonal drojection was computerized,and the sys- The average value of PEP is 78+12ms.This tolic flow rate in cc/systolewas yielded. value agrees well with the result measured 40 experiments were performed withvarious outputs from phonocardiogram. The factors that ranging from 45 to 110cc / systole. prolong or shorten PEP, such as hypoxe- Accuracy of the measurementswas + 15%, for an in- mia or umbilical cord occlusion, are cident angle ranging between 40 to55°.Comparison currently under investigated. with thermodiluLion showed a similaraccuracy. ONE Ultrasound Symposium.

BE15-B.1.. BEI5-B.3 Evaluation ofPulsedDoppler Measurements of Pulmonary Artery VelocityProfiles,C. Lucas, B. Velocity Profiles, R.S. Rabinovitz*. C.J. Hartley. Ha*, G. W. Henry and J. Ferreiro, The UNC School M. J.Levesque, R. M. Nerem. BaylorCollegeof of Medicine, Chapel Hill, North Carolina, USA. Medicine, Houston, Texas, and Georgia Institute of Technology, Atlanta, Georgia. Measurementof pulmonary bloodvelocitynoninva- Velocity signals measured by a pulsed ultra- sively with Doppler ultrasound has stimulatedinterest inrelationships between velocitypatternsandabnor- sonic Doppler velocimeter (PUDVM) are derived from mal pulmonary hemodynamics. Using an intralumi- convolution of a finite-size sample volume with nal 20 MHz pulsedDopplerneedleprobe,velocity the velocity field.This results in a spatial profiles have been obtainedinthemainpulmonary smoothing and lengthening of the measured velocity artery (MPA) of lambs and dogs during control condi- profiles.A computerized simulation was developed tionsand whilehemodynamicswerealteredvia to evaluate the accuracy of the velocity profile hypoxiaor MPA particle (20-500 embolization. measurements obtained by a 20 MHz PUDVM. The Intraanimalresultswereconsistentwith clinical model includes transmit burst length, acoustic observations,i.e,increased PA pressure and/or resist- attenuation, far-wall reflection, receiver gat.c ance corresponded to decreased acceleration time (AT = width, and a zero-creasing-counter with threshold time from the onset of systoletopeakvelocity)and levels.To examine the performance of the simula- increased maximal flow reversal( Voeg), thoughinter- tion. a Poiseuille velocity profile was used as an animal variability precluded using either AT or Vneg to input, and the output velocity profile was com- quantifyaccuratelypressureorresh..ance. Promi- pared to that measured by the PUDVM for fully nent backflow along the posterior wall of the MPA was developed laminar flow in a straight tube often observed duringcontrolconditions,while AT (20

1588 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas 357 Biomedical Engineering ScientificPapers

BE15-B .4 BE15-B.5 Human Regional Blood Flows DuringSimulated Microgravity: Rapid stenosis locationloxDoppler Ultrasound using anew A Noninvasive Doppler Flownlevy Study Video tracking method. ER Greene, KW Tawney, EC Johnson D.H.King*,J.Reidy,L.Robinson, P.Nesbit,M.Payne,K.Photiou,S.Kontis,M.G.Taylor and Lovelace Medical Foundation, Albwierque,NM, USA R.G.Gosling. Guy's Hospital, London Bridge, LondonSET 9RT, United Kingdom Theoa Ar,.erialcrosssectional area is inversely proportional to blood velocity inan unbranching artery. Dopplershift is proportional to blood velocity,therefore by monitoring Doppler shift it is possibleto deduce the presence of area changes within the arterial lumen.This principle has been employed in the Quickscantechnique which is described below. Method A Atit hand held CW Doppler probe isscanned along thelongitudire: axis of the artery. The locationof a LED light mounted on the probe tip is continuouslymonitored by an overhead video camera. A micro computer records thepeak Doppler shifted frequencyand calculates the probe XY position,andthen,using this data, plots cut a fixed width line, in red, representing thenormal linen, and then overprints with a yellow line, whose widthis related to frequency shift.Diagnosis of complete occlusion is made ifthe characteristic signs of collateralfumation are evident in the Doppler shift Spectra.Inthis case the yellod line is set to equal the ltren width.Thus Quickscan superimposes a notional nap of arterialcrosssectional. area over a video image ofthe patient. This image is similar in appearancetothe conventional arteriogran. However, itdiffers in threerespects:-(a)It is completely non-in asive (b) It yieldsdata comparable to multiulti planeplane viewsviews ofthe arterial systen(c) It is not affected by timing and dilution artefactswhich can degrade arteriographic image quality.

BE15-B. 6 BE15-C. 1 Application ofthe ZoomWigner The Transverse DopplerTechnique for Transform toBlood Disturbance Blood Flow Measurement, Velocity Waveforms, H Talhamiand R I Kitney, Dept of Electrical V.L. Newhouse, DrexelUniversity, Engineering, Imperial College, Philadelphia, PA 19104, USA. London SW7 2BT, UK. Existing ultrasonic The paper will discuss the systems for blood flow mea- application of the ZoomWigner surement which use the well knownDoppler transform to the analysisof blood equation, only estimate the velocitycomponent disturbance waveforms obtained in the beam direction, andare restricted to beam downstream of a stenosis. As will to flow angles less than about be shown in the paper the ZWThas 600. A recently certain advantages over boththe discovered technique will bedescribed, using FFT and modern spectral estimation the Doppler spectralbandwidth as well methods, as the particularlyin relation center frequency with which itis possible to to the frequency resolution of specific components. estimate the velocitycomponent normal to the Shortrecordsof thedisturbance beam axis as well as that inthe beam direction, velocity waveform for various and to measure flow velocities levels of occlusion, directed at all namely 40 and angles to :he beam axis,including 90 °. 74%, were transformed to obtain a Wignertime/frequency distribution This "transverse Doppler"technique makes use which showed the flow field of the fact thatany scatterer moving through the behaviour in the preselected focus of a sound beamwill have its echo frequency bandwidth. The results were amplitude modulated by thebeam side lobes in found to be of significantly such a way as to produce higher quality than thoseobtained a Doppler spectrum using other spectralestimators. whose width is proportionalto the velocity component normal to the beamaxis. 3 7 8 358 1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas Biomedical Engineering Scientific Papers

BE15 -C.2 GE15-C.3 ACCURATE MEACJRMENT OF BLOOD Transit -Title Broadening Estimation with Fourier FLOW USING ULTRASOUND AND TIME Transform and Autoregressive Techniques DOMAIN CORRELATION, I. Hein*, V. Dominique Davet*, Steven A. Jones, Don Giddens Suorsa, J. Zachary, R. Fish and Georgia Institute of Technology W. D. O'Brien, Jr., University George W. Woodruff School of of Illinois, Urbana, IL 61801. Mechanical Engineering

Spectral broadening due to transit-time ambi- Using time domain correlation, the guity from a Pulsed Doppler Ultrasound system volume fluid flow in a vessel can be was examinedwith Fouriertransform and auto- accurately and precisely estimated without regressivemethods of spectral estimation. any previous knowledge of the vessel Measurements were obtained from laminar flow at size, flow velocity profile, or transducer several velocities, and the spectral character- measurement angle. This technique istics were comparedwith a theoretical model observes the change in arrival time fortransit-timebroadening (Garbini al.; (instead of frequency) of ultrasonic pulses JFM, 1982). Data records of 4, 35, 70 ms were reflected from moving scatterers. It has processed with FFT and BurgAR (model order 4 been validated in a blood flow phantom and 32) methods, andthe spectral width was sys'2m using both a blood mimicking calculated as a function of the peak frequency, substance and porcine blood. The from the basic formula given by Garbini. All the transducer measurement angle has been curves fit a good linear regression, as predict-. determined within 5%, and pulsatile flow ed by the Garbini model. The AR model of order 4 has been measured up to 150 beats/min yielded the lowest standard deviation for short with accuracies better than 18%. records, at all velocities. The regression for Currently, this method is being the FFT techniquedeviated fromthe model, in experimentally investigated under in vivo that it did dot pass through the origin, whereas conditions, with the eventual application the AR method gave good results. It is concluded being the diagnosis of venous thrombosis that the FFT estimation technique itself contri- in humans. [Work supported by butes significantly to spectral broadening for NSF/NHLBI 1 RO1 HL 39704.] short data records, and the AR technique more accurately predicts transit-time broadening.

BEI5 -C.4 BEI5 -C.5 ADVANTAGES OF SMOOTHED SPECTRUM ESTIMATE FUR Pulsed Doppler Velocity Estimation via ANALYSIS OF DOPPLER SIGNAL SPECTRUM Autoregression and Fourier Transform M. Piechocki, A. Herment, A. Delouche, Ph. Dumee, Steven A. Jones'and Don P. Giddens P. Peronneau*. Georgia Institute of Technology

IPPT, Polish Academy of Sciences, Warsaw, Poland ; George W. Woodruff School of INSERM U. 256, Paris, France. Mechanical Engineering

Theaim of our study was to establish asimple Doppler Ultrasound data have been collected methodforthe improvement ofthe instantaneous from a physical flow model and processed via spectrum estimate definition.An analysis of the Fourier transform and autoregressiontechniques statistical properties of FFT spectrumestimate to compare the accuracy of these methods. The applied to the Doppler signal allowed us to develop measurementswere performedwith a a smoothing procedure of this instantaneous 10MHz pulsed Doppler probeat the throat of a estimate. The smoothing of the FFT spectrum estimate contoured constriction with 90%area reduction. is achieved through a low pass filteringin the Various Reynolds numbers were used, and the frequency domain. The performances of smoothed flow was laminar with a blunt velocityprofile. spectrum estimate were theoretically and Doppler signals were digitized and recordswere experimentally studied. The variance of the smoothed processed with Fourier transform and Burg spectrumestimateis similar to that ofsimple Autoregression methods to yield curvesof FFT estimate averaged over several cardiac cycles. velocity as a function of time. Velocity was The procedure establishesimproves sensiblythe estimated from both the peak and thenormalized accuracy of the relation between the shape of first moment of the spectrum. Standard the spectrum and the flow parameters. deviations of the resultingtime series data In result, more exact determination of flow were then examined. characteristics like its stability or its maximum For peak detection, lowmodel order auto- velocity even in case of low signal to noise power regression (e.g., order4) yielded the lowest ratio is possible. variance, approximately halfof that for the Fourier transfo-mmethod. The first moment timeseries was almost identical for both spectral techniques, butthe variances were larger than those derived from peak detection. 3 '7 9

1988 World Congress on lylcdical Physics and Biomedical EngineeringSan Antonio, Texas 359 Biomedical Engineering Scientific Papers

BE15-C.6 BE15-D.1 Phase Locked Loop Method for Doppler US Catheter tip imaging with ultrasound, N. Bom*, Investigation, F.Clemente, M.Cesarellie H. tcn Hoff, C.J. Slager, F.C.van Egmond, R.Castaldo, m.Bracale, Cattedra di Elet- Gussenhoven, Thoraxcenter, Rotterdamand Inter- tronica Biomedica-Nucleo di Ricerca di university Cardiology Institute ofthe Netherlands. Ingey eria Medica e della Riabilitazione Presently many recanalization methods C.N.R.- Faculty of Engineering, Via Claudio are being developed to reopen obstructed arteriesduring 21 80125 Napoli, Italy. catheterization. Such a new obstructionremoval A new method for evaluating single frequency wave- technology is the spark erosion method.An evaluation study has been carried out forms from Doppler Ultrasound signals based on the to see if ultrasound imaging on the tip of a catheter couldbe developed analogue locked loop technique has been proposed. to - eventually - guide selective sparkerosion for P.L.L. technique can be applied to follow the in- removal of atherosclerotic plaqueand other possibly stantaneous frequency of the Doppler signal. The eccentrically located arterialobstructions. average of the instantaneous frequency of the sig- Following early work with 32-elementcircular phased nal is an unbiased estimator of the mean frequency array technique a less complex mechanicalscanning method has been designed. Six designs and the average of its absolute value may be rela- were studied that fall into two categories.In the first category ted to the variance of Doppler Spectrum. an internally rotating mirror deflects theultra- sound beam over 360 degrees. No moving Therefore the first parameter shows the shape of parts are present outside the catheter. The second the mean blood flow velocity; the latter is an eati category incorporates an exteriorly roPating tipwhere the mator of the spectrum broadening. echo imaging and the spark erosionelectrode are active in almost the same arterial A prototype based on this method has been developed. cross-section. A 40 MHz ultrasound echo tip has beenbuilt which Some interesting clinical results considering young rotates at 3000 rpm. health people and arteriopathic patients show the With this prototype real-time imagingwas studied Possibility to evaluate different hemodynamic condi- in animal experiments and humanarterial specimen tions. obtained from the pathology laboratory.Results show a clear correlation between the plaque andthe echo image.

BE15-D.2 BE15-D.3 In -Vitro Assessment of the 20MHz Intracoronary Ultra- Effect of Aortic Valve Replacement sound Doppler Velocity (IDV) Catheter, D Abouelnasr,* on Coronary D Churchwel.1, D Giddens, S Jones, Carlyle Fraser Artery Flood Velocity Waveform inPatient with Aortic Valve Disease Heart Ctr/CWLHosp of Emory U; EUSM.Dept of ME, GA Inst of Tech, Atlanta, GA S.Matsuoka: S.Kanazawa, Y.Wada,M.Goto, K.Mito, Y.Ogasawara, O.Hiramatsu, K.Tsujioka,T.Fujiwara, 1DV Catheters are currently under investigation as a F.Kajiya, Katy-=aki hedical School,577 Matsushima, Kurashiki, 701-01, Japan means of estimating coronary flow by measuring coronary blood flow velocity (CBFV). Before large scale Aortic valve disease is frequently use of the catheter on human subjects, we performed accompanied by an abnormal myocardial perfusion. in vitro studies for the assessment of performance The purpose of this study was to clarify thecharacteristics of under simulated clinical conditions: 1. Sample volume the coronary artery blood flow size was measured by moving the catheter about a fix- velocity in patients with aortic valve disease. ed target and plotting the Doppler-shifted frequency rive patients with aortic regurgitation(AR), M611 power ai determined by autoregression versus dis- three with aortic stenosis(AS) and four with aortic tance.The sample volume diameter remained very stenosis dnd regurgitation(ASR) were studied close to that of the catheter was compared to laser- before and after aortic valve replacement(AVR) byour 20MHz 80 Doppler anemometer (LDA) mesurements made at the same channel pulsed Doppler ultrasound location, and compared very well. The effects of velocimeter. The velocity waveform and the near-wall location were observed by pit:zing the cath- velocity profile across the 'vessel were measured eter in a lom (taunter coronary model, and 4mm down- at the left anterior descending coronary artery(LAD). stream, the eLects of a wake may be seen in low, The velocity waveform in AR showeda large systolic fluctuating velocities. At ranges of 5 to lOmm, the flow and a relatively small diastolic DSF velocity was higher and more stable. CONCLU- flow with short duration. In AS, SIONS: 1. The sample volume size remains discrete to the early diastolic acceleration of the velocity a range of 10mm, and is small enough to be wholly was significantly attenuated and a systolic reverse contained within the coronary artery. 2. Based on *MA flow was always observed. The velocity waveform in ASRwas in measurements, the DSF is accurate. 3. The size and between of AR and AS. effect of the wake behind the catheter must be deter- Immediately after AVR, the velocity waveform returned mined. 4. Autoregression analysis of the DSF is sen- to normal diastolic predominant pattern. sitive enough to measure velocities up to a range of 10mm. 3 8 0 360 1988 World Congress on Medical Physics andBiomedical EngineeringSan Antonio, Texas Blomedicr.1 Engineering Scientific Papers

BE15-D.4 BE15-D.5 Intraoperative Evaluation of Blood Flow Velocity POSTOPERATIVE MONITORING OF REGIONAL MYOCARDIAL Waveforms in Different Types of Coronary Artery FUNCTION IN MAN. C.J. Hartley*, R.S. Rabinovitz, Bypass Grafts -Sequential Saphenous Vein Graft and L.J. Suignard, B.S. Patel, J.E. Chelly, M.O. InternalMammary Artery Graft-, S.k:anazawa, Jeroudi, G.P. Noon, and R. Bolli, T.Fujiwara, S.Matsuoka, T.Katsumura, *Y.Ogasawara, Baylor College of Medicine, Houston, TX 77030. *K.Tsujioka, *F.Kajiya, Depts. of Thoracic and Cardiovascular surgery, and *Medical Engineering, We ha.?, developed an ultrasonic method to measur Kawasaki Medical School, Kurashiki, Japan regionalsystolic thickening (ST) of the left ventricle (LV). The sensor is a silastic coated 2 The patency of the aorta-coronary bypass graft may mn 10MHz transducer suturedlooselytothe closely relate to the blood flow wa'eform and epicardinm during surgery with the leads exteriorized near the chest drain. The sensor is geometrical blood flow disturbance in the graft as well as the volume flow rate. We investigated the extracted 2 to 3 days postoperatively by pulling on blood velocities in two types of coronary artery the leads. Thickening is measured by sensing tae phase of echoes within a range-gated sample volume bypass grafts, the life span of which are known to placed in the wall near the endocardium. Each cycle be different,i.e., the sequential saphenous vein bypass graft(shor ter span) and the internal of phase corresponds to tissue displacement of X/2 or 0.075 mn. Thickening fraction a mammary artery graft(longer). The velocity was (TF), measured by dual mode(zerocross and FFT), dimensionless index of function and ischemia, is multichannel, high frequency ultrasound determined by dividing ST by the range-gate depth. In 12 coronary bypass patients, TF ranged from measurements during bypass graft surgery. In the 15% to 35% intraoperatively and remained stable or vein graft, the skew of the velocity profile and decreased slightly during the early postoperative the flow separation were always recognized at the region just proximal to the side to side period. In one patient, however, ST was replaced by thinning (usually a sign of ischemia) during this anastomosis and the charge in the direction of time. All sensors were removed without incident. skew was observed alon; the bridge portion, while a parabolic velocity profile with a narrow This technique providescontinuous monitoring of regional LV function and its recovery and spectrum was observed in the internal mammary modification by drugs during the early artery graft. Tt Ise characteristics of blood postoperative period. velocity may be attributable to thepatency of the graft.

BE15-D.6 BE15-E.1 Pulsed Ultrasound Doppler in Cardiac Output A Sonographic Technique for Monitoring, J.B. Ruo*, A.P. Yoganathan, and M.A. Reducing Beam Distortion by Colson, Georgia Estitute of Technology, Atlanta, GA Subcutaneo4 Tissues. 30332-0100 G. Kos ', D.A. Carpenter, D.E. Robinson, Ultrasnoics Doppler echocardiography provides a technique to Institute, Sydney, Australia continuously monitor cardiac output(C 0),indicating the overall cardiac status of a patient during the im mediate 111 sonographic equipment assumes that post-operative period. To develop a relationship the speed of propagation of ultrasound in between the actual CO and a Doppler - derivedCO, an in soft tissue is constant whereas in fact vitro pulsatile flow study was performed involving the it varies by up to 10% from the rio.mally use of a Vingmed SD-100 Doppler ultrasound system and assumed value of 1540 m/sec. This two different 5 MHz implantable Doppler probes. One variation is particularly significant in of the probes erritted a conical beam, while the other the subcutaneous tissues and can cause produced a wedge- shaped beam. For tilting disc, deviation and broadening of the beam and bileaflet, and porcine heart valve designs, spectra were an increase in side lobes, resul'..3.ng in obtained at 2,4, and 6 Mmin and at 2, 3, and 4cm significant degredation in sonr.s4,aphic distances from the valve ring. Peak velocity was image quality. These aberrations have measured as well as CO determined by an been evaluated on a beam generated by a electromagnetic flow probe and by mean andmaximum linear array transducer following Dopplervelocity spectra. A profile factor was then propagation through layers of various calculated as actual C 0 /Doppler derived CO. geometry and velocity. A compensating Results show a linear relationship between Doppler technique has been developed which allows COand actual C0 with correlation' coefficients ranging the possibility for correcting for these from 0.95 to 0.99. For a given valve design using the aberrations based on the control of the maximum velocity spectra for Doppler CO evaluation, timing and the amplitude of the signal theprofile factor was found to remain relatively of each element both in transmission constant, independent of actual CO and distance from and reception. the valve: approximately 0.37 for the tilting disc, 0.44 for the bileaflet, and 0.29 for the porcine.Therefore, this factor maybe used to characterize prosthetic valve types. 381 1988 World Congress on Medi: i Physics and Biomedical EngineeringSan Antonio, Texas 361 Biomedical Engineering ScientificPapers

BE15-E.2 BE15-E.3 Bubble Sizing for Cardiac Applications,P. M.Shankar; Drexel University, Ultrasonic Imaging of Blood and Its Potential Department of Applications in Studying Hemodynamics and Electrical andComputer Engineering, Philadelphia, PA 19104 Rheology, K.K. ohung*and Y. W. Yuan, Bioengineering Program, Pennsylvania State University, University Park, PA The noninvasive measurement ofpressure in 16802 differtilt chambers of the heartis very important in cardiology.Recently it has been proposed that bubbles With the advent of technology, blood which usedto may be injected be considered anechcic in ultrasound imaging into the circulatory systemand their size can changescan be monitored as they now be visualized when a high frequency scanner is pass used. through the chambers of Work in our laboratory and elsewhere has the heart. shown that blood flow patterns Ultrasound is an excellenttool to track can be readily visu- and size the bubbles. The alized in vivo or in vitro in a mock flow loop. author and his It has been found that blood echogenicity is criti- colleagues have proposed anew technique of cally dependent upon red cell aggregation which is measuring sizes called the DoubleTrequency technique.In this technique the bubbles affected by the shear rate produced by the flow and fibrinogen content of the blood. are insonified by two sound fields,a high Echo intensity correlates well with viscosity of the frequency imaging field and lowfrequency blood pumpingfield. Thesum and difference at low shear rate since viscosity ..s also shear rate dependent at low shear rates. frequencysignals generatedfrom More recently, the a region with complete cell depletion was consis- oscillations of the pumped bubblesare used tently found to occur in flowing whole blood under for size measurements. Preliminaryin-vitro experiments showed that such technique certain conditions at low shear rates. This phe- can nomenon has never been reported previously. be used to measurepressure changes in the In amLient atmosphere by measuring this paper recent experimental resultsand real- their size time ultrasonic images will be presented changes. In this paper asummary of results to illustrate these observations. on sizing microbubblesin Also to be address- circulation ed is the relevance of ultrasonic imaging using double frequency techniqueswill as a tool presented. for studying hemodynamics and blood rheology.

BE15-E.4 BE15-E.5 Speckle Reduction in Ultrasonic Imaging using VAedical ultraconic Two-dimensional Phase-insensitive Receiving rransaucers 4itk. Gaussi al Surface Arrays,* J.J. Ciesey,* C.R. Meyer, and P.L. Velocity Cistrioutio.,,jin Shuwu* Carson, University of Michiban, Dept. of 4nejiang bniversi ty,r.aneznou, (.nina Radiology, Ann Arbor, MI 48105

Specular noise, which is inherent in all The tneoretic analysts.designof tne coherent energy forms, can limit the quality of construcJ.onand measurementof the sound ultrasonic images. Speckle can be reduced by field for medical ultrasonictran..ducer adding images formed by transducers with differ- witn Gaussian surfacevelocity distriou- ent characteristics. The amount of speckle re- tion are described.Thepiezocrystal plate duction was found to be a function of the im- of the radus 6mm ismade from We, piezo- pulse response and the input pressure waveform electric ceramics.rheresonant frequency of the transducer. is 2.51'ihz.The acousticlens of the focal A pulse-echo imager which utilizes two-di- length 59mm is madeof epoxy resin. the mensional phase-insensitive receiving arrays has Gaussian transducers ha.e equivalentra- been developed to reduce speckle in ultrasonic dus 3mm. images. Four receiving arrays were constructed The focussing Gaussiantransducers are each containing 64 segmented annulus (SA) or measured by the sound field tested instru- hexagon elements with either spherical or coni- mentation and imaged by theacousto-light cal focusing. Preliminary results for a conical- diffraction device, old comparedwith the ly foc.ised SA array showed using 64 elements in- us ally plane transducers, the focussing dependently reduced speckle by a factor of 3.8" plane transducers and thefocussing over combining all elements into one subaper- Gaussian transducers of equivalent,raaus ture. The results agreed with values found from 4.8mm.The developed Gaussiantransducers the impulse response of the transducers. have- better near fieldcharacterization than others.Thereapear none of the maxi- * This work was supported in part by PHS Grant ma and minima of acousticpressure and #2-R01-HD17243-05 awarded by the National In- reduce the minor lopesso as toimprove stitutes of Child Health and Human Development. the resolution of AM modeultrasonic diagnosis instrumentations. 2

362 1988 World Congress on Medical Physics andBiomedical EngineeringSan Antonio, Texas Biomedical Engineering Scientific Papers

BE15-E.6 BE15-F.1 A Novel Method for Pulse-echoUltrasonic Effects of Ultrasound on MammalianGonads, Tomography, W.C.A. Pereira and J.C. F. Dunn; Bioacoustics Research Lab., Uni- Machado*, COPPE/UFRJ-BIOADICA, C.P. versity of Illinois, Urbar.a, IL 61801, USA 68510, 21944-Rio de Janeiro-RJ, Brazil There are circumstances in which human gonads have A new method to obtain two dimensionalmapof the opportunity to receive significant nitrasound ultrasonic wave speed is proposed. Themethod exposures during clinical diagnostic and therapeu- consists of insonifying a medium with an tic regimes. Thus, studies were conducted to ultrasonic pulse and receiving the scattered determine whethermorphological alterations of wave on the same transducer used as transmitter male and female gonads can be produced by ultra- and in another one placed aside. Each scatterer sound as intensities approach therapeut..,. levels. illuminated by the irradiated wave returns an Mouse testes and ovaries were exposed in vivo to echo to both transducers.Then by measuring the 1 MHz CW ultrasound at spatial peak intensities in time of flight for each echo arrived on both the range 2 to 100 W/cm2 fir times ranging from transducers the position of each scatrer,relat- 300 to 1 sec. The gonads were removed surgically ed to the transmitting transducer, and the wave at times ranging from immediately following to 63 speed in the medium existent between consecutive days post irradiation and preparedhistologically scatterers can be obtained. Computer simulation for microscopic analysis. of rnis method by transmission of an ultrasonic The bserved alterations in testis can be classi- sine wave burst (pulse width of 5.12usecs and fied according to the number and type ofaffected wave frequency of 2.5 MHz) through a medium cell layers in the seminiferous tubule. Leydig consisting of polystyrene spherical particles cells were unaffected. Spermatocytes are affected (diameter varying from 0.6 mm to 1.0 mm) sur- before the spermatogonia, contrary to the situa- rounded by water has provided encouraging re- tion following ionizing radiation. sults for the proposed method. In the ovary lesions are manifested by pyknosis of the cells, vacuolization of cellsand tissue, eosinophilic cytoplasm, and general cedulardis- ruption. Inc eases in the amount of PAS positive material appe,red in the interstitial tissue. Luteinizedstructuresappeared to be preferentially altered.

BE15-F.2 BE15-F.3 MCC-INVASIVE ASSESSMENT OF VESSEL WALL PROPER- Ultrasonic System for Noninvasive TIES. R.S. Reneman; T. van Merode and A.P.G. Measurement of Regional Elasticity Hoeks. Depts. of Physiology and Biophysics, of the Artery, K. Yamamoto*, T. Imura, Univ. of Limburg, Maastricht, The Netherlands. T. Mikami and H. Yasuda, Hokkaido University, N13-W8 Sapporo, Japan. Mitigate pulsed Doppler systems (MPDS) allow the on-line recording of velocity profiles in arteries To measure noninvasively the regional elastic and the relative diameter changes of these vessels properties of the artery, we have developedan during the cardiac cycle (ild/d). From the width of ultrasonic system, which comprises an ultraso- the profiles the arterial diameter (d)can be nic displacement meter, an echographic equip- assessed. From &d/d, d and brachial artery pulse ment and a personal computer. The displacement pressure the distensibility coefficient (DC) and meter is designed to track pulsatile motion of cross-sectional compliance (CC) of the common the anterior and posterior walls of the artery carotid artery can be estimated (Reneman et al, at the proximal and distal sites and also desi- Ultrasound Med. Biol. 12, 465-471, 1986). In 80 gned to enable us to monitor easilya measured volunteers, DC and CC decreased linearly with age site on the display; therefore, even a deep from the third age decade. DC and CC were found to vessel such as the aorta can be measured. From be lower in young borderline hypertensives (n=15) pulsatile diameter change, mean diameter and than in normotensives (n=13), although blood pres- for pulse pressure,the personal computer sure differences between these subjects were limi- calculates the pressure strain elastic modulus ted. The detection of local changes in arterial Ep and regional pulse wave velocity PWV. wall displacement in addition to that of flow Ep and PWV were measured on healthy subjects disturbances improved the accuracy of the diagno- along the carotid, iliac and femoral arteries sis of minor (<30% diam red) atherosclerotic le- and the abdominal aorta. These measurements sions (diag acc: 95.3%; sens:90%; spec:90%). Since showed the progressive increase with distance with the MPDS used displacement of the anterior to peripherals. Age-related increase in Ep of the abdominal aorta was also found in another and posteriorwallcannot ,e distinguished, a .echnigue based on phase tracking of the RF-sig- series of measurement on 61 subjects. Thesys- nals, generated by the arterial walls, was deve- tem demonstrates the feasibility of noninvasive loped. This method is simple and allows the sepa- detection of atherosclerotic change. rate recording of anterior and posterior wall displacement. 383 1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas 363 Biomedical Engineerin:,lcientific Papers

BE15-F.4 BE15-F.5 Stump Tissue Stiffness Measurements of Above- PRESSURE MEASUREMENT knee Amputees Using Doppler BY ANALYSIS OF Ultrasound, T. xESONANT FREQUENCY OF ELASTIC Krouskop, M. Malinauskas*, D.Winningham, HICROCAPSULES J. Williams, Baylor College K.Ishihara*, A.Kitabatake, J.Tanouchi,K.Fujii, of Medicine, Houston, M.Uematsu, Texas. Y.Yoshida, T.Nagakura, M.Hori, T.Kamada, M.Inoue,T.Tamurat,K.Chiharat,H.AbeT, The 1st Dept. of Med. OsakaUniv., Osaka 553 Many problems for above-kneeamputees stem tFac. of Eng. Sci.Osaka Univ., Osaka 560 from poor fit of their prostheses. At present TOsaka ha...onal Hospital,Osaka 553 Japan the interaction of the residuallimb with the socket is still not well understood. Charac- We have developeda pressure measurement system terization of live human tissueis difficult but using ultrasound. The principleis that resonant is possible through mathematicalmodels or noninvasive systems. frequency of elasticmicrocapsules in liquid A Doppler ultrasound shifts depending on thesurrounding pressure. system was used to measure Methods Pressure-sensitive the stiffness of stump tissuein twelve above- microcapsules were made of vinylidenedichloride-acrylonitrile knee amputees. Measurements were made at three copolymer locations on each stump: containing harmless gas. ricrocapsules,classified anterior, posterior, into several classes according and medial. Both shallow (0 to 1.5 cm) and to the diameter, were injected into an airtightcistern. Ultrasound deep (0 to 3.5 cm) readingwere taken. Pesults indicate that there are significant was projected from a transmitterdriven by differences function synthesizer. WP between shallow and deep readings, compared the changes of these frequency spectrum of the differences being especiallyapparent at the ultrasound passing through microcapsules withthe changes in pressure. anterior portion of thestump. However, no differences were detected R---"s 1)The frequencyspectra showed marked among testing sites on attenuation in specific frequency, the stump or between shallow anddeep readings indicating at posterior and medial areas. resonance of the microcapsulesaccording to the diameter. 2)The resonant General stiffness valuescan now be found with amplitude in the this system. frequency spectrum linearly Investigation into methodsto changed with the increase location and depth pressure. 3)The resonant frequencyalso increased sensitivity are with the pressure. ongoing. Conclusion We have confirmedexperimentally the potential of this newpressure measurement mthod using the resonant frequencyof microcapsules.

BE15-F.6 BE16-A.1 In Vivo Frequency Response Measurements Desigri UsingUltrasound, M. C. Hastings: fox- Auditory Georgia Institute of Technology, L. Finley; Research Triangle Atlanta, Georgia Institute, Research Triangle Park, N. C.2'709 USA A -vscem using low-power ultrasound was In the past development ofauditory pros- deg 'oped to measure the in vivo frequency thesis systems has followed response of internal largely an organs to a mechanical empirical path. Electrode designs prestimulus. The system has two focused sently available for human ultrasonic ae incorporate transducers, a transmitterand features that were dictatedby the con- receiver. The organ to be examined is placed straints of mechanical within the crossed foci construction and of the transducers and insertion the physical unitinto the coch- insonified by the transmitter. Thnphase lea. Speech processor development modulation resulting from has the response of the followed in a similar vein,being largely tissue to the externalstimulus introduces limited by the range of available sidebands about the ultrasonic technol- carrier in the ogy that would ultimately permitwearable spectrum of the received echo. The localized processors. These approaches have displacement of proven theorgan is determined from remarkably successful insome individual the ratio of the amplitudes of the carrier and patients but have failed the sidebands. miserably in The sensitivityand spatial others. This presenteti -n focusesupon resolution of the measurement primarily the basic mechanisms of neural depend on the diameter, focus, and stimulation frequency of with intracochlear electrodesand seeks to the ultrasonic transducers. Presently, 6.35 mm explain the basis for dramaticperformance transducers operating at 20 MHz are being used differences amongindividual patients to investigate the mechanics of auditoryorgans using the same prosthesis in fish. system in terms This system can detect 12 Angstrom of variations in survivalof neural ele- displacements with a spatial resolutionof 0.28 ments within the cochlea. Design princi- mm. Localized density variations withinorgans ples are discussed t:atallow a speech can alsobe determined using this dynamic processing strategy to be tailored measurement to the technique. Applications include specific anatomical andphysiological distinguishing different types of kidneystones status of an individuaa. patient. In addi- and identifying soft ti.,ue tumors. tion, considerations for designof optimal 384 electrode arrays are discussed. 364 1988 World Congress on Medical Physicsand Biomedical EngineeringSan Antonio, Texas Biomedical Engineering Scientific Papers

IE.111-A.2 BE16-A.3 A New Model of Extracochlear Prosthesis A Nonlinemr Model for Kemp Echoes for the Profoundly Deaf, R. Newcomb*, P. Gomez and V. Rodellar Y.Hirata*, T.Ifukube, J.Matsushima+ and N.Hoshimiya, Res. Inst. of Appl. Elect., University of Maryland, College Park, +Medical School, Hokkaido Univ., Japan. MD 20742 USA and Universidad Politecnica de Madrid, Madrid, Spain A new model ofan extracochlear prosthesis that stimulates viaan electrode superfici- Kemp Echoes are under investigation as ally placed on the round window membraneof a noninvasivemeans of characterizing cochlea was proposed as a less invasive al- and diagnosing damage to inaccessible ternative to insertion in the scale tympani. parts of the ear. For this character- We have developed a Pt-Ir stimulation elect- ization signal processing techniques rode coated with polyvinyl alcohol gel that are used to set up a lattice digital is proved to be biocompatible material.The synthesis filter from the Kemp Echo electrode is suitable for the extracochlear data. The parameters of the lattice are prosthesis because it Can tightly contact to also determined in terms of the mate- the round window membrane without damage it. rial properties of the ear. By compar- As for a speedh coding method, we have pro- ing the two determinations of the lat- posed a new ideathat the pitch signaland tice, structural anomolies of the inner the second formant frequency could be trans- ear can be isolated. Because second mitted simultaneously by combining those two order effects are important to the signals. In this method, an additional stim- interpretation of KempEchoes, nonli- ulation pulse is inserted among pitch pulses nearities become significant to this fromthe and the time delay of the pulse technique as a diagnostic tool. Nonli- pitch pulse is in proportional to the second nearities due to membrane and fluid formant frequency. Those time sequential stimuli were given to guinea pigsthrough interactions within the ear are incorporated in the Kemp Echo lattice model the round window membrane using the coated presented. electrode and compoundactionpotentials from cochlear nerve bundle were recorded. In this report, the efficacy of the electrode Supported by NATO Grant 0395/87 and NSF and the speech coding method is evaluated Grant M1P 85-06924. based upon the experimental results.

BE16-A.4 BE16-A.5 Model Structure and Parameters Estimates A Neuro-Synaptic Model of Pitch for the Source Derivation of Auditory Perception with Auditory Memory, Brainstem Responses, S.Kahtoh*, T.Sekiya, K. Itoh., Faclt Med, Univ Tokyo, M.Aruga, K.Yamazaki and M.Saito, Tokyo 113, Japan. * Teikyo University of Technology, 2289- 23,0tani, Uruido, Ichihara, Chiba, Japan A model of temporal information processing in the auditory system was presented To obtain significant findings on ABR on the basis of the interaction of fast generators, ABRs to sine wave stimuli,were synaptic potentials as well as slow syn- recorded in 10 normal Ss. The stimuli rang aptic potentials in order to analyze the -ed from 50 to 80 dB HL at between 0.1 and mechanism of pitch perception of complex 1.0 KHz for more 512 responses, were given sounds. to both ears simultaneously. The responses In the model, the pitch was determined within 10.24 ms, were recorded between the through five processes, i.e., the trans- vertex and earlob with common grand at the duction of acoustic signals to cochlear frontal, and represented as a continuous microphonics, to fast synaptic potentials curve in the complex plane in which sepa- and to nerve discharges in the cochlea, rate plots for phase and vector length as the summation of the discharges in the function of frequency were shown. By using cochlear nucleus, the extraction of the this result, we obtained the frequency discharge.periodicities in the inferior characteristic of ABRsand used a fortran colliculus, the temporal-to-spatial computer program to fit a linear transfer translation of the periodicities by slow function to given frequency response mag- synaptic potentials in the medial genicu- nitude and phase data. A conjugate search late body, and the determination pitches was used that minimizes the integralof the by picking local peaks in the auditory absolute value of the error squared between memory space of the integrated slow syn- the model and the data. The transfer func- aptic potentials in the auditory cortex. tion model suggested that ABRs have three The model could explain the extraction vectors as model parameters,which are pro- and cancellation process of difference portional, derivative and second order lag tones as well as synchrony suppression. element.

365 1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas Biomedical Engineering Scientific Papers

BE16-B.1 BE16-B.2 Computerised Analysis of Eye Movements, University Hospital of Wales, Cardiff, U.K. Smooth Pursuit Ey! MovementCharacteris- ticsDependingupon Attention to a Moving Target, The simplest method of eye movement analysisis by direct visual Y.Ebisawa*, H.Minamitani and M.Takase, obse_rvaticn whereby an evaluationcan be madeof gaze nystagan, Keio University, Hiyoshi 3-14-1, Yokohama, 223 Japan reftetion saccades, slow pursuit movements andvestibular nystagras indexed by positioning, thermal and -otaticnal It is considered that characteristics The degree of analysis possible however isclearly limited, e.g. of human thermally or rotationally intned vestibular eye tracking movement are changable bythe level nystagrus nay be of quantified by either timing the dtration of subject's attention toa moving target. 256 theresparre cr pieces canting the amber of nystagnic beats in of LED were arranged horizontallyin a a given time interval. line. Theposition of thetarget, only one Following sane ratable publications in the lighted 1950's, clinics LED,was moved sinusoidally at 1 Hz, became more mare of the advantages of the ± 10 technique of electro- degree by a micro-computer. Eyeposition oculography in the assessment of vestiado-ocular flnction with signal was recordedby electro-oculographic such measures as nystagnas amplitude, technic. frequency and slow -phase Target Brightness Feedback (TBF) velocity being available from the chart method, recording. The mount whichis one type of visualfeedback of information which nay be extracted, even by this method methods, was developed in orderto increasethe however, is limited by constraints of the attention. speed, accuracy and Under the TBF condition, the target endurance of the human being. It may be argued that a comes to be darker when subject cannotcorrectly canputerised system of recording and analysiswould provide not track it. He must track thetarget as accurately only a more detailed, but also a as mart. standardised evaluation possibleso that the target doesnotturn of eye payments than is possible by conventional means. darker. Under non-feedback condition,the target Because of the difficulty in programming is a computer to deal with always bright and the subject tracks the the analysis of weak and irregular nystagnic eye movements, most target as usual. canputer systems to-date incorporate The a large degree of rental characteristics of eye movement underthe decision making. The ultimate system wouldbe one vhichcould TBF condition were compared with thoseof non- analyse eye movements accu-ately without human feedback intervention. condition. Saccadic components were This paper describes a fully oapputerisedsystem with examples removedfrom total eye movementsbyour new of analysis of refixaticn saccades and responses from thermal, algorithm. The gain of smooth pursuitmovement optckinetic and rotational stimuli. to the target movement increasedand thephase delayed when the attention increasedby the TBF.

BE16 -B.3 BE16-B.4 Electro-oculographic measurement Three Dimensional Measurements of the Ocular of inter-saccadic latencyin Fundus and Evaluation of the Papillary Excavation in Glaucomatous Eyes, congenital nystagmus,G.Bosone, T.Yoshimura*, R.Nakatani R. Reccia,G. and N.Suzuki, Faculty of Eng., Kobe Univ., Rokko Roberti, P. Russo*, Nada, Kobe 657, Japan University of Naples, 1-80125 Naples, Italy. Vertical and equidistant parallel grating lines We measured the frequency distributionof are projected onto the ocular fundus from right latencies of the quick phaseof nystagmiceye side of pupil and a set of stripes reflected into movements the normal direction by the fundus is recorded in patients with congenital with a TV camera. The deformation of the detected nystagmus (CN), from electro-oculographic grating images gives the depth of the excavation (EOG) records of the oscillatory eye on the ocular fundus.In order to accurately movements in eight patients.Since it is detect the grating image, we have used a new commonly accepted that thesame non-defective method, which subtracts two fundus images before oculomotor system controlsthe generation of and after shifting the projected gratings by a half period. By using this technique, the back both (normal) saccades andquick phase eye ground noise due to blood vessels on the images movements in CN, our measurements were is removed and only the grating images have been consideredas giving the distributionof detected. From the deformation of the each inter-saccadic latencies inthe range 0.1-0.6 stripe of the accurately detected grating image, s. For each patient, we recordeda 3-4 s long we evaluate three dimensional distribution of the EOG, ocular fundus, the depth of the optic disc and repeating each session 22times. The cup-disc ratio. In glaucomatous eyes with visual average inter-saccadic latencywas computed field changes, the excavation starts in the for each record, by evaluating the peak retinal area and the corresponding rim is time-frequency of the EOG spectrumwith destroyed. We discuss that the glaucomatous digital power spectralmethods. The frequency visual field changes is dependent on the excava- distribution of theseaverage latencies is tion showed above. peaked at about 220 + 40 ms, with a gaussian-like shape. S6

366 1988 World Congress on MedicalPhysics and Biomedical EngineeringSan Antonio, Texas Biomedical Engineering Scientific Papers

BE16-B.5 BE16-B.6 Development of a micromachined transensor An approach of early glauoana detection Yin Chang*, Institute of Biomedical Eng. for monitoring of intraocular pressure National Yang Ming Medical College Yao-An Ft, Chang Gung Memorial Hospital Taipei, Taiwan, Republic of China Y Undid; B HOk, L We*, L Tenerz, Electronics Department, Institute of Technology, University of Uppsala, Box 534, S 751 21 Can glaucoma be early detected? Cblor sensation Uppsala, Sweden, B Svedbergh, Institute of Oftalmology, Academic test may provide a valuableindicationof early Hospital, S-750 14 Uppsala 14, Sweden. glaucoma spplaxawhich the glaucomasuspectsare not aware. A 4- channel, 12-bitprecisionlight and computer controlled, ananaloscope Nontraumatic, continuous monitoring of intraocular pressure Is an intensity employed Moreland'scolor match equationis Important, as yet unsolved problem, in several states of disease, built to give the color sensation test at Chang including glaucoma. We suggest for solving this problem, the GUngMemorialHospital in this study. The result shows the glaucoma patients and sate glaucoma uttlization of a passt, transensor, consisting of a resonant suspectshave "blue shift" of match point and capacitiveinductive network, in which the resonance frequency is havea wider matching range than thenormal a function of pressure. We show that such a passive transensor can observers. be integrated with an artificial lens without Impairing normal seeing and be Implanted in the eye without trauma. Monitoring of the resonance frequency can be performed by means of microelectronic circuitry encased in eye-glasses. in this paper, the technology of transensor fabrication is presented, using micromachining techniques, like photolithography, deposition, etching and anodic bonding. This results in excellent performance, high wproductbility, and low cost. Data on prototype performance compared with theoretical estimations are given.

BE16-C.1 BE16-C.2 Dynamic Analysis of Accommodation and Eye Length Measuraient by Pupil Diameter under the Natural Viewing Laser Interferamtry Condition,F.Okuyama*,T.Tokoro,K.Yana and A.F. Fercher: Universitat Wien, A-1090 WIEN M.Nagayama , Tokyo Medical and Dental University,Tokyo,Japan, If the pupil of the eye is illuminated by a Accommodation,the the focusing system in collimated laser beam the reemitted light ex- the human eye,and changes in pupil size hibits fringes similar to the well-known occur together.This is similar to a came- Newton's interference fringes. Using this pheno- ra when focusing and diaphram size menon together with an adequate interfercmetric change togrther.It is considered that technique we have been able to measure the these two functions in the eye ocuur optical length of the eye with a precision of together in order to perceive a clear .03 mm. The geometrical eye length was obtained visual image and maintain it.This dynamic from theinterferometrically measured optical relationship has been widely discussed in length by using an average refractive index the literature. according to Gullstrand. A series of in vivo We measured accommodation by infrared measurements of several human beings shows a optometer and pupil diameter by TV in- good correlation with the acoustically deter- frared pupilmeter simultaneously under mined eye lengths. This technique is noninvasive natural binocular viewing conditions an no additional means such as anesthesia are while viewing a sationary actual target. needed. The high transversal resolution of this Fluctuationsof accommodation and pupil technique may lead to additional possibilities, diameterwere analyzed by cross-corela- e.g. the measurement of the st.rface profile of tion function and cross-spectrum using an the fundus. FFT(Fast Fourier Transform) signal analyzer(SM-2100,IWATU Electric Co.).The results yielded a minus correlation btween two fluctuations.This demonstrates that pupillary constriction occurs with a decreaseof accommodation resulting in a deeper optical depth in the human eye. 877

367 1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas Biomedical Engineering Scientific Papers

BE16-C.3 BE16-C.4 A Finite Element Model of Radial Keratotomy COMPUTER AIDED VISUAL FIELD DIAGNOSIS Surgery, R.P. Vito,* J.W. Shin, Georgia G.C.Filligoi*,L.Capitanio, N. Accornero Institute of Technology, Atlanta, Georgia; B. McCarey, G. Waring, Emory University, Visual Field (V.F.) analysis is widelyused in Atlanta, Georgia. clinical practice for the important informations it provides in neurologic and ophtalmic pathology. In the last years many computerized devices have Though radial keratotomy is commonly used by been set-up in order to achieve better results, but surgeons to correct refractive errots, the their purpose is mainly concernedon collecting da- mechanical basis for the surgery is not well ta and presenting a map of the results. understood. The immediate surgical outcome is In our project we developed firsta system for determined only by mechanical factors while, in collecting V.F. data using a standard IBM-XTCOMM- the long term, other effects, such as wound ter with a ',_lack and white monitor for presentation healing, are important. Accordingly, we have of the light stimuli. For the analysis of the field used a finite element based model of the cornea we developed a dedicated software using techniques to simulate the effects of various surgical and of image process'ng for extraction of scotoma'sfea physiological parameters as well as surgical tures (i.e., shape, size, position, etc. of V. F. errors on corneal curvature. losses) and recognition of other typical fieldde- Results indicate that incision length and fects. depth are important determinants of surgical Afterwards, a procedure for the analysisof tile outcome; a fact consistent with clinical logical sequence used by the neurologiststo set a observations.The effects of surgical errors, diagnosis on the basis of the V.F. datasavebeen such as incisions of unequal depth or length, implemented. were surprisingly small. The decisicn tree which models this strategypa rallely drives the acquisition of theproper features automatically extracted from the data. Our preliminary results confirm both validity and reliability of the parameters used in V.F.auto matic analysis.

BE16-C.5 BE16-C.6 Use of a Programmable Remapper The Limits of The Man-Machine Interface With as a Potiental Aid for Patients with Respect to the Visual System. Visual Field Defects,David S. Loshin*, H. Pichlert F. Pavuza, G. Beszedics, Tu Wien, University of Houston, College of A-1040 Vienna Austria Optometry and Richard D. Juday, Tracking and Communications Division, NASA In the field of medical instrumentationa state Johnson Space Center, Houston, TX of the art man-machine-interface includes, in most cases, a cathode ray tube (CRT). Alphanumerical A Programmable Remapper which is a digital image displays, combined with picture processing offe: a processing machine of unique design and function powerful information system to the user andare has been developed at NASA, Johnson Space commonly applied in the field of measurement Center.The Remapper is designed to perform systems, system operation and state control. coordinate transformations in real-time on video The visual part of the cognitive system offersa images. The original intent of the Remapper was to measurable overall modulation transfer function. simplify some aspects of pattern recognition in video (MTF), which is a combination of both the MTF of imagery. NASA's interest arises from problems in the CRT and the human eye. The MTF of the CRT's automated tracking and docking of spacecraft and in screen depends on the type of screen (monochrome autonomous landings on Mars. The Remapper's or color, the latter one characterized by the potential application to human low vision became type of implemented shadow-mask).Both functions apparent early in its development. In cases of ocular must be intelligently combined to obtain an pathology that result in retinal field defects such as optimized MTF for the overall system, or to adapt the screen parameters to the characteristics retinitis pigmentosa and macular degeneration, of the human eye. portions of the visual space are lost. The Remapper The paper describes the MTF of a typical human can warp displayed images so that they fall onto eye, its approximation, and the MTF of the viable retina and thus provide information that would different CRTS In addition, the recommended otherwise not be perceived. We have developed and parameters for displays working with commonly will report on alogithms for several types visual used TV and HDTV-systems are given. remappings using an image processor (with static images) to demonstrate the concept.

368 1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas Biomedical Engineering Scientific Papers

13E17.1 BE17.2 Biodynamic Response Determination to Transitory CINEMA7_C AND DYNAMIC GAIT Acceleration, Paul H. Frisch, Frank Lambert, Applied ANALYSIS Physics Inc, 31 Highview Avenue, Nanuet NY 10954 Pelisse F.1, Geiger D 2', Marie F.1 (1) Centre d'Etudes et de Recherches de Transitory acceleration profilesexperienced by military l'Appareil Locomoteur - 94460 VALENTON pilots during air combat maneuvers (ACM), emergency (2) Laboratoire de Mdcanique Physique, egress,andcrash/impacteventshavedemonstrated Universite Paris XII - 94010 CRETEIL significantperformancelimitations,during ACM, and (FRANCE) injurypotentialsranging from mild traumatodeath. Research efforts addressing methods of enhancing pilot Computation of the forces applied to the hip, knee and ankle physiologicalacceptability and limitinginjury joints during human locomotion requires the knowledge of mechanisms,have focused on dynamic simulationof ground reaction and kinematics. This study presents a method pilot/crewstationinteractionduringtransitoryand which provides a graphic representation of the load applied to the Simulation via human analogues, lower limb joint during the stance phase of the walk. sustained acceleration. The use of a force-plate (0,8 x 0,3 meters), on line with a suchasaanthropomorphic manikinsdynamically and computer allows the quantification of the resultant force of the mechanically simulate human response, while loading the pressure exerced between foot and ground. In order to recordthe ejectionseatand crew stationaswould the human spacial-temporel evolution of the geometrical center of the joint, a counterpart. Dynamic manikin testing provided detailed small incandescent lamp is externally located on the virtual axis of insighttopotentialinjury mechanisms and provide a flexion of the joint. The position of the lamp relative to the force criteriaforevaluationandcomparisonofcomplex p'.ate (external boundaries) is measured by means of two linear aircraft escape systems and crew stations.The ability to C.C.D. cameras.(1728 cells - 13x13 pm). Camera lens of 50 mm measure the three dimensional biodynamic response of a and 135 mm are used for measurement in sagittal and frontal human surrogate, predictsthe focus and G-loading the plane respectively. An additional hemi- cylindrical lens is set ofthis inside the body camera in order to magnify the size of the vertical pilotwillexperience. Detailed knowledge field. The force-plate and kinematic data are recorded every 0,01 environmentenables new protectivedevices(G-suits, second during the stance phase of the gait cycle. During the recliningseats)andmodifiedcrewstationstobe computation of analysis of the joint force transmission. a parallax developed reducing the physiological stress imposed on a correction is performed. pilot. Thepresentationwillfoc, 3onmanikin A mathematical model similar to Harrington's one is used for development simulating human biofidelity and response, simulation of the biological structure of the knee. and instrumentation techniques enabling the measure of the six degree of freeuom response of the manikin.

BE17.3 BE17 .4 PORTABLE GAIT LABORATORY: PRELIMINARY RESULTS, D. VAT - Vestibular Autorotatiun test in the Aronson, H. Kang*, V. Musselman, and A. King, De- clinic and in weightlessness. partment of Orthopaedic Surgery, Wayne State Univ. A.Manssor* and S. Vesterhauge School of Medicine, Detroit, Michigan, U.S.A. University Hospital Copenhagen, Denmark.

The gait analysis laboratory has evolved from a Vestibular °Lula' reflex tVOR) both vertical head research tool to a clinically accepted modality and horizontal is tested during voluntary Testing the vestibular function for studying children with neuromuscular disor- oscillations. in the frequencyrange 2-6 Hz ispossible ders. Current gait analysis laboratories are ex- without the neeo for expensive turntables. pensive and require considerable space. This Head movements are recorded via anangular study was undertaken to compare the early results tatesensoi. Tt :orresponding compensatory eye of our portable gait lab with existing gait anal- movements are corded via superficial elec- ysis laboratories. Five children between 7 and trodes. Anticowpensatory eye movements arid vi- 10 years of age were studied with the portable sual interference are avoicted by using stimulus gait lab. The lab consists of shoe-borne load frequencies above 2 Hz. Auto calibration of the cells, portable data acquisition system, SP2000 eye signal levelis performed using 6.5 Hz head motion analysis system, and a personal computer. rotationwith the eyes fixating a distant The shoe-borne load cells were used to collect The VOR transfer function is determined three-dimensional ground reaction force data. The point. spectra! analysis. Head SP2000 system was interfaced with a personal com- using crosspower a; e guided b/ a frequency modulated puter to record three-dimensional kinematic data .atations from the hip, knee, ankle and foot. The results sound7ignal. revealed that cadence, stride length, and walking test have been used both in the tlinic to velocity were slightly less than that previously The research related published by conventional gait laboratories. By evaluate patients andIn Space parabolic flight experiments, latest in contrast, the ground reaction forces, joint range NASA's Transfer function parameters of motion, and EMG's correlated very closely with KC-135 aircraft. are stable and reproducible and the utilization previously published data. The portable gait lab test inotoneurological patients is provides ground reactive data from multiple steps, of the promising. which is a major advantage over conventional laboratories.

1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas 369 Biomedical Engineering Scientific Papers

BE17.5 13E17.6 THE VARIABILITY OF A SIMPLELEARNED MOVEMENT Analysis of Self-Induced Hand Oscillation as a Model of Physiological Tremor R.L.G. Kirsnee& E. Byrne. St.Vincent's A.Watanabe*, K.Goto, and M.Saito, Institute Hospital, Fitzroy, Victoria.3065. Australia. of Medical Electronics, University of Tokyo, Tokyo 113, Japan. Learned ballistic movements executed in response to visual cues d1splay considerable variability. Tremor is expected to provide neurological infor- Far from being an artifact, the variability is a mation if its mechanism is fully understood. fundamental part of the response and provides The a objective of the present study is to establish measure of the central control of muscleactivity. the mathematical model of physiologicaltremor Step shifts in the position of based on the oscillation characteristics of the a vertical line on hand, a video monitor were corrected by rotation of a The two-reflex-loop model is proposed, which is subject's forearm. Rotation,angular acceleration, biceps and triceps composed of the mechanical system and the spinal emg wererecorded on a computer. and supraspinal reflex loops. The most important The acceleration waveform has been parameter in the model is the gain ratio of the found to provide a very sensitiveand quantifiable measure of response variability. two reflex loops.Two types of measurements are performed: slow relaxing movement andposture maintenance. In the former experiment reflex oscilla- Limited patient studies have shownthat in cere- tion characterized by the double-line spectrum is bellar dysfunction, the acceleration amplitude is induced separately from mechanical oscillation. considerably decreased while the variability and time In the latter the relationship between theaccel- course of the movement are much largerthan normal. eration and EMG is investigated by the Fourier analysis. The physiological parameters estimated agree well with those reported in the literature. From a clinical viewpoint, this technique can The diversity of frequency change of physiological provide considerable insight intoabnormal central tremor can be attributed mainly to the change in integration of ballistic movements. It may also suggest the loop gain ratio. treatment strategies useful in rehabilitation as well as providing an objective assessment of recovery.

BE1S-A.1 BE18 -A.2 Fiber-wrapped Chambers for Computation of Theoretical modeling of the coronary circulation, Three-dimensional Blood Flow in the Heart. * C.Guiot D.M.McQueen* & C.S.Peskin, Courant Institute ,C.Cancelli,T.J.Pedley and P.G.PiantA, of Mathematical Sciences, New York, USA. Dip. Anatomia e Fisiologia Umana dell'Universita di Torino, Corso Raffaello 30,10125 Torino, Italy.

The goal of our project is the computer prediction of three-dimensional flow patterns of blood A model of the left coronary circulation as a net- in the chambers of the heart. Our computational approach is based on the fiber-oriented method work of collapsible vessels mimicking the microva- which we have employed success- -fully in two di- scular compartment can satisfactorily reproduce mensions. Detailed dissections of the paths of the coronary flow rate, provided a distributed, muscle fibers in animal hearts (by other resear- one dimensional, separation free, viscous flow chers) have revealed a complex pattern of trajectories on nested toroidal shells contained in theory is assumed for each vessel. a basket-like sheath. Before attempting to con- Any simpler representation, amounting to a lumped struct a model of the heart, we have computed model of essel resistance and compliance, as some flows in some simpler three-dimensional structures. We have constructed a torus by means of workers have proposed, cannot reproduce the obser- a network of closed elastic fibers which make ved waveforms. The reason is that the scale for several turns about the torus lumen while making time variations of the transmural pressure in ear- one circumferential turn. In this setting we have studied the flow resulting from the passive ly and late systole is too short in comparison relaxation of an initial bulge in the torus and with the time scale defined by the system elasti- the peristaltic flow resulting from a wave of city for wave propagation effects to be ignored. muscular contraction circulating around the Both a distribuited and a lumped model have been torus. The results from this latter study compare well with a simple linear theory. worked out in detail, showing that for a given We are now constructing a whole heart model range of physiologically sensible parameter values employing fibers which follow geodesic paths on the first one can realistically simulate the main a number of relatively simple geometric shapes. features of the coronary flow waveform, while the latter exhitrits unobserved artifacts and disturbances. Q ¢.1) t1V

370 1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas Biomedical Engineering Scientific Papers

11E18-A.3 BE18-A.4 Model Prediction for the Interactive Forces Betveen Coronary pressure-flowrelation after Contractilig MyocardiumandCoronaryVessels, transition prom blood to Tyrode Y. Lanir and E. Nevo, Technion-Israel Institute perfusion in the maximallydilated of Technology, Haifa, 32000, Israel diastolic cat heart, L.C. van Dijk, R. Krams, P. Sipkemaand N. Westerhof". Mechanical interactions between contracting Vrije UniversiteitAmsterdam, PO Box myocardial tissue and intramyocardial vessels has a 7161, 1007 MC Amsterdam, The Netherlands dominant effect on the coronary flow. Previously of the coronary circulation have proposed models In 6 isolated, diastolicarrested, maximally considered the isotropic intramyocardial tissue vasodilated cat hearts, westudied changes in pressure (IMP) as representing these interaction coronary pressure-flow (P-F) relations (with zero forces. Morphological studies have shownhowever, flow pressure intercept, Pint) after a change from that intramyocardial vessels are attached to neigh- blood to Tyrode. The pressure-flow relations were boring myocytes by a densenetworkof collagen fittedwith a mathematical model: P0(1-exp(- Hence the force exerted on each vessel is struts. F/F0)+k.F+Pint; Po, Fo, k and Pint are model hydro- the sum of contributions from the isotropic parameters. Pint (in kPa, mean +SD) changed from pressure of the iutemtitial fluid and from static 2.0 +0.94 during blood perfusion, to 2.5 (+0.55), the anisotropic solid stress of the collagen. The 2.6 (+0.68), 2.5 (+0.94), and 2.7(+1.34), during contribution depends on the specific orien- latter Tyrode perfusion for 2:15, 5:30, 10:30, and 25:00 tation of each vessel. minutes, respectively. Resistance (definedas the Experimental evaluation of intramyocardialsolid ratio of perfusion pressureand flow at 10 kPa stress is not possible, but it can be evaluated perfusion pressure) steadily rose to about 170% of from theoreticalstress analysis. Recently a the value during blood perfusion. The observation structurally based myocardial material law has been that the apparent intercept pressure is developed in our laboratory and applied in amodel maintained, when a particle free (Newtonian) of the left ventricle. Simulationof the whole perfusate is used, may indicate that this cardiac cycle resulted in histograms of stressand intercept is not a result of blood rheology alone. pressure fields which lead to evaluation of the The increase in resistance with decreased forces exerted on each vessel as a function ofits perfusate viscosity, suggests an effect of edema, location in the wall and orientation. These which increases interstitial volume at the expense results will be presented and discussed in terms of of intravascular volume. their relevance to the coronary circulation.

BE18-A.5 BE18-A.6 Model of ReactiveHyperaemia, TRANSCORONARY DENSITY TRANSPORT AMathematical + S. W. Smye*,M. I. G. Bloor, Medical Physics Dept FUNCTIONS OBTAINED BY ITERATING St. James Hospital, Leeds LS9 7TF, UK. +Applied THE CONVOLUTION INTEGRAL Maths Dept. Leeds University, Leeds. UK. E.Gallasch*,T.Renner,Universitat Graz,A-8010 Graz,Austria Reactive hyperaemia, the increase in blood flow after a period of circulatory occlusion, involves During transition of a bolus of indica- a variety of vascular control processes including tor fluid through a network of arteries, metabolic, myogenic and passive factors. capillaries and veins,a series of events The role of a pressure-related change in vascular influences the variation of the time resistance, due to the myogenicresponse, has course of the concentration of the indi- been considered in detail using a tube model of cator bolus. the limb arterial system. The equations A usual descriptor for the transition of describing the flow, local pressure p, and an indicator through a vacular bed is cross-sectional area of thetube S, include a the transport function h(t).Because its relationship between S and p (myogenic response) nearly imposible to produce an ideal im- the form of which is constrained given that p pulse injection on the upstream side we shouldevolve in a stable manner. Independent have develloped a modeladjustment te01- experimental justification is available. nique for the calculation of the trans- The solutions describe flow which rises, within port functions from the up- and down- Ss of occlusion release, to a peak of stream curves.The square error (differ- approximately 10 times resting flow, decreasing ence between model and measured output) to normalover the following 30-60s. These is minimised iterative by the steepest resultsare consistent with venous occlusion decent technique. plethysmography andisotope clearancemethods. As a result we show that this method ex- The modelpredicts that the flow val..ation is hibits some advantages (stability against principally due to the changing ,,ressure-gradient observational errors and aditional noize) and not to the relatively small change in S. over usual methods via fourier transform Themodel clarifies the role that myoqenic or numerical deconvolution. responses may play in reactive hyperaemia.

331 1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas 371 Biomedical Engineering Scientific Papers

BE18-8.4 BE18-8.2 EvaluatiA of blood velocities in ventricular and Skewness of Profiles of Velocity in the Coronary atrial coronary circulations by three different Arteries of Dogs, David R. Bell*, Hanl N. Sabbah means of access using a laser Doppler method, F. and Paul Kaj lye, Department of Medical D. Stein, Henry Ford Heart and Vascular Engineering, Inst, 2799 W. Grand Blvd, Detroit, MI Kawasaki Medical School, Kurashiki,Japan Multiple range gate A laser Doppler velocimeter(LDV) with Doppler velocimetry was used an to examine the optical fiber has been demonstrated velocity profile within the lumen to be useful of for measurements of the coronary the left circumflex (L. circ) and left anterior circulation. descending (LAD) Using LDV, we took three means of coronary arteries of 10 access to the anesthetized open-chest dogs at vessels according to the difference in rest, after measuring propranolol andafter intracoronary adenosine. objects: (1)Large epicardialcoronary artery and Profiles of velocity duringpeak diastolic flow vein. t:e inserted the fiber from the sidewall into the vessel with an aid of were skewed toward the outer walls of the vessels a cuff. (P<0.007). Theextent ofskewnesswas not (2)Relatively small artery and vein of which wall affected by propranolol or adenosine. is transparent to laser beam:we measured Shear rates the weresignificantly lower along the inner wall in velocity noninvasively to the vessel byplacing the fiber tip on the surface of comparison to the outer wall under all conditions. the vessel. both in the L. circ andLAD coronary arteries (3)Intramyocardial artery and vein:we inserted the (P<0.017) fiber into the vessel from thevascular portion just penetrating into myocardium.The velocity -1 profiles across the vessels SHEAR RATE (sec ) as well as the L. Circ velocity waveforms near the central axial LAD region Inner Outer Inner Outer in the circumflex coronaryartery and the great Wall Wall Wall Wall cardiac vein were measured by thefirst access. The small artery and vein velocitieson the left Control 428+70 697-99 354+65 640+61 atrium were measured by the secondaccess. The Propranolol 2b6+39 596+108 306,59 419#.81 velocity waveforms in the deep septalartery and Adenosine 1150+330 19904526 1131+287 17814.318 in a small intramyocardial veinwere evaluated by the third access. The characteristic of the In conclusion, irrespective of the levels of flow, velocity in each vessel will be discussed as well the velocity profiles were skewed toward the outer as possible causes of the velocity waveform. walls of the coronary arteries. Consequently, the shear rate was lower along the inner wall.

BE18-8.3 BE1.8-9.4 Fluid Dynamic Differences litf,tween Smoothand Ir- Wall Shear Stress Calculations regular Coronary Stenoses: Their ImplicationsIn for Lesirn Progression, A Churchmen*, DAbouelnasr, Human Coronary Arteries,H.M. Olsmats, D Giddens, S Jones, Carlyle Fraser Hrt Ctr/ R. Holenstein, W.A. Seed andR.M. Nerem, Crawford Long Hosp of Emory U,EUSH, Dpt of HE GA Georgia Institute of Technology, Inst of Tech, Atlanta, GA Atlanta, GA., 30332-0405, U.S.A. The pathology of a coronary intimalplaque is caus- ally related to the rate of itsprogression to Estimates of wall shearstress in the human left total occlusion. Complex plaques are more likely epicardial coronary arterieshave been carried out than simple plaques to occlude rapidly. On angio- usinga computer model of the graphy, complexity corresponds to luminalirregu- coronary larity (LI). circulation.Pressure and flow The purpose of this pork is toassess employing are calculated by fluid dynamic factors of stenosesxhich might theone-dimensionalcontinuityand explain these observed differences. Models of 90% momentum equations for flowwith friction. The cross-sectional area reduction stenoseswere con- geometry of the left main (LM),the left anterior structed based on angiographic measurementsof descending (LAD), and the coronary arteries. Four models xere made simula- left circumflex (LCFX) ting lesions with and xithout LI. These models coronary arteries was taken frommeasurements on xere placed in a steady floe circuit. Reynolds human post-mortem casts,and the inputpressure numbers (Re) from 25-200 (simulatingvarious points waveforms were digitizedrecordings of aortic and in the cardiac cycle) were utilized. Flog visuali- left ventricular zation was attained by illuminationwith laser pressure from humans. Shear stress light. as a function of time Laser' Doppler anemometer (LDA) measurements method was calculated, using the were made at various locations in each model. of Bensonetal.(CardiovascularRes., Conclusions: Flow through all fourmodels remained 10:568-576,1980),forthe LM, the LAD-LCFX laminar through the range of Retested. There were bifurcation,andatvariouspositions no major differences in the primaryflow features LAD and the LCFX. alongthe between the models with and without LI. However, Both fully developed flow in the irregular lesions, smallrecirculating zones and entrance effectswere taken into account. The occurred in the concave portionsof the lesion general range of wall shearstress was found to be downstream of the stenosis throat. This is a from 0 to 50 dynes/cm, with peak shear stresses common location of platelet deposition incomplex occurring intimal plaques. intheentranceregionoftheLM. Negativeshearstress valueswere particularly in the distal observed, of the LCFX. portions of the LAD and 332

372 1988 World Congress on Medical Physics and BiomedicalEngineeringSan Antonio, Texas Biomedical Engineerint Scientific Papers

BE18 -B.5 BE18-B.6 Cotonaky Cincueation (index High So:stabled A noninvasive method for the three dimensional image of Gtavity, S. H. Zhou, E. H. Collins an coronary artery,Tetsuo Okasural*,Naoki Suzuki1,Zenji R. Collins*, Technical University of Nova Nakano2,Michihiko Matsui3,Tsuneo Suzuki4,Sachio Ikawa4,ME Scotia, P. 0. Box 1000, Halifax, Nova Scotia Research unit' Dept. of Pediatrics2Dept. of Cardiosugery3 (Canada B3J 2X4). Dept. of Laboratory medicine' Jikei Univ. School of Med. 105 Minato-ku Tokyo,Japan When pilots of jet combataircraftexecute operational manoeuvers, large angular We developed a method for the three dimensional(3D) visua- accelerations are produced which cause blood lization of coronary artries using ultrasonic two dimen- to drain from the brain and heart toward the sional(2D) echogram.The system for this method consisted feet. We wish to develop a mathematical model of an ultrasonic 2D echograph to get the images of an for the coronary circulation in such an aimed vessel and a computer graphics system which recon- accelerative environment, with a view to stracts 30 image of the vessel.In order to collect trans- improvinp our understanding of the cardiac performankt of jet pilots. The governing versed images of the coronary artery we applied ECG equations of motion will contain not only terms triggering and image processing technique.Because the related to these accelerative forces, but also coronaly artery moves with the pulsation of the heart and incorporate active neural control of the its location from the surface is too far to get a clear rate. A quasi vascular calibers and heart image.A probe was placed in the 2nd or 3rd intercostal one-dimensional unsteady flow model willbe space.Successive images are taken by moving the probe to presented along with preliminary calculations of the pressure and flow fields under a variety the direction of distal part of coronary artery. In our of operating conditions. experiments, we case to the conclusion that the 3D model using coloring and shadowing gave a clear view of ihe shape and the structure of vessels. Especially in case of aneurysmal dilatation, the change in wall thickness and distribution of plaques in the vessel can be visualize, where the angiographic method can not provide us with these informations.

BE18-C.1 BE18-C.2 Quantitative Assessment of Blood Evaluation of Human Coronary Flow Dynamics in Velocity Characteristics with Health and Disease by 80 Channel 20 MHz Ultrasound Doppler Ultrasound Pulsed Doppler Velocimeter, Y. Ogasawara*, K. Mito, S. Matsuoka, S. Kanazawa, K. Tsujioka and F. Kajiya, D.P. Giddens*, R.I. Kitney & S.A. Jones Kawasaki Medical School, 577 Matsushima, Kurashiki, Georgia Institute of Technology and Japan Imperial College of Science Human coronary blood flow velocities (the left and Technology anterior descending coronary artery (LAD)) were Doppler ultrasoundmeasurements of blood measured by the 80 channel 20 MHz ultrasound velocity must bederivedfrom the frequency pulsed Doppler velocimeter developedin our content of the Doppler signal. However, laboratory. The depth resolution of our system is 0.2 several factors contributeto a smearing or mm and the sample volume for each sampling point is broadening of the spectral content, so that nx0.52x0.2mm3. Doppler signals from 80 channels are velocity estimation can be difficult under detected by a zero-cross method and a signal from an certain circumstances. Spatial gradients, optional channel is analyzed by a fast-Fourier temporal changes, turbulence and the method transform method, both in real time. With this ofspectral estimation can all result in method, we have been able to analyze detailed blood uncertainty in velocity estimation. In velocity profiles across vessels, as well as velocity particular, the requirement of obtaining good distribution (spectrum) in a sample volume. Two type frequency response in the measurement of of probe was designed to measure different objects: A temporal variations in velocity is directly cuff type probe was used for velocity measurements of atoddswith the requirementof securing a bypass graft and a plate-type probe holder was used velocity resolution. Experimental data will for the measurements of the native LAD. In the post- be presented which illustrate these points, stenotic region of the LAD, the velocity waveform was and the application of autoregressive/maximum systolic predominant and the velocity profile was ir- entropymethods to velocityand spectral regular with wide spectrum, while the velocity by way of patterns became normal, i.e., diastolic predominant estimation will be discussed immediately after graft operation. The coronary artery examples taken from both laboratory velocity also changed significantly after aortic valve measurements and human studies. replacement in the patient with the valve disease, i.e., from systolic to diastolic predominant. 3 9 3 1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas 373 Biomedical Engineering Scientific Papers

BE18-C.3 BE1S-C.4 FREQUENCY DOMAIN DECOMPOSITION OF FLOW Measurements of Blood Flow byMRI VARIABLES UNDER PULSATILE FLOW CONDITIONS John Gore Baruch B. Lieber: State University of New York Yale University Medical School at Buffalo, Department of Mechanical and New Haven, CT Aerospace Engineering, Buffalo NY 14260. Developments in Nuclear Magnetic Resonance Concernover turbulence-induced hemolysis has led imaging have providedpowerful new methods for numerous hemodynamicists to measure the effects of probing vascular anatomyand detecting various fluctuating stresses on rad blood cells. Although types of blood flowphenomena. Methods are many agree that turbulence may cause hemolysis and also underdevelopment forquantitation reducedurability of red blood cells, there is of no blood flow in major vessels, in steady as well agreement upon neither the magnitude of turbulence as pulsatile conditions, and for estimating stress levels nor the duration of their applica- perfusion at the capillary level. A variety of tion required for cellular damage. Since pulsatile techniques are being evaluated. flows in Vascular hemodynamies occurat fairly moderate architecture may be studied in cross-sections Reynoldsnumbers, flow patterns may wax and wane at various angles using the multiplanar andmay include coherent structures and oscilla- capacityof MRI, whileprojection images tions of distinct frequenciesin additionto obtained by subtractiontechniques give images turbulence, resulting in nonstationary phenomena. much like conventional angiograms. Flowing The frequency domain approach, based on the trans- materials give rise to signalalterations in MR formation of the unsteady Navier-Stokes equations images that depend on their velocity, into the frequency spectrum, provides acceleration a method to or higher moments, as well as discriminate between their precise small scale (cellular) tur- spatial andtemporal patterns. bulent stresses and largescale (inertial) Paramagneticcontrast agents orother blood stresses induced bycoherent fluctuations. Flow pool labels can also be usedto study perfusion variables areseparated effects. into a pulsatile trend, Various methods have associated coherent problems or fluctuations and randomvariations. artifacts that are under active Decompositionof a variable into its constituents investigation, but the potential nonetheless is achieved using pattern exists to recognition and fre- develop relativelyprecise flow quency domainfiltering for the pulsatile trend, measurements non-invasively in deep structures matched filtering is than applied to discriminate using standard conventionalMRI equipment. between turbulence and coherent structures.

BE16-C.5 BE18-D.1 3D Solid Modelling of Arterial Blood Flow in the Heart and Large Vessels, Structures Using Ultrasound. M. Sugawara; The Heart Institute of Japan, R I Kitney, Departmentof Electrical Tokyo Women's Medical College,8-1 Kawada-cho, Engineering, ImperialCollege, Shinjuku-ku, Tokyo, 162 Japan London SW7 2BT, UK. Various systems for measuring thepressure and velocity of the blood have been developed, There is now considerableinterest in and a the use of three dimensional considerable amount of experimental and clinical modelling techniquesfor the study of data are now available.The introduction ofthe hot-film arterial structure. The primary anemometer with its high spatial motivation resolution and frequency response has greatly of thiswork is the development of new diagnostic extended our knowleage of the blood flowin the aorta. techniques for arterialdisease. The Here, we describe the developmentof majority flow and transition to turbulence of 'modelling methods in the aorta developed in dogs.Turbulence in the cardiovascular todate have useda wire system can also modelling approach. The workwhich be detectedwith clinicallyused will devices, such as the be presented in thepaperuses multi-sensor catheter. full voxel space to create3D solid The so-called "anacrotic notch" is analyzedin models of high quality. relation to transition toturbulence in cases with aortic or pulmonary stenosis. The paper will alsodescribe the Blood flow application of thesemethods in the left ventricle can also be measuredwith to a ultrasonic imaging of arteries. The multi-sensor catheter. The interaction results of clinical trials will be between the swirling motion generated bythe presented in which 3Dsolid models of inflow from the atrium and an aortic arterial structuresare compared with regurgitation during diastole is analyzed. pathological slides of the same Compared with the diastolic slow- pattern, the structures. systolic flow pattern in the left ventricle is rather simple. However,prior tothe major ejection wave of the blood velocity and posterior to the P wave of the ECG, thereis a spiky velocity wave.The cause of this spike is fl analyzed.

374 1988 World Congress on Medical Physics and BiomedicalEngineeringSan Antonio, Texas Biomedical Engineering Scientific Papers

BE18-D.2 BE18-D.3 Analysis of the flow field in a 3D -mode, of the Numerical Analysis of Pulsatile carotid artery bifurcation, C.C.M.Rindt ,A.A.v. Wall Shear Stress in an Arterial Steenhoven, A.Segal, R.S.Reneman and J.D.Janssen, Bifurcation, K. Perktold* and Eindhoven University of Technology, Postbus 513, D. Hilbert, Techn. Univ. Graz, 5600 MB Eindhoven, The Netherlands. A-8010 Graz, Austria

To :,:Main insight into the complex flow patterns in the human carotid artery bifurcation, steady The study if fluid dynamic phenomena and flow in a rigid 3D-model was analysed by finite flow-induced mechanical stresses in hu- element. calculations based upon a standard man blood vessels is of great interest Galerkin-formulation of the complete Navier-Stokes in order to clarify the contribution of and continuity equation. The-c calculations were hemodynamic factors on the pathogenic performed at physiological A,ynolds numbers and mechanism of atherosclerosis and throm- validated with laser-Doppler measurements. In the bosis. A hemodynamic factor of consider- bulb high axial velocities were found near the able physiological importance is the wall divider wall and a region with reversed axial flow shear stress. occurred opposite to the flow divider which occu- Here pulsatile flow through an arterial pied about 25% of the local cross-sectional area. bifurcation model is analysed numerical- At the end of the carotid sinus an axial velocity ly. The computer simulation of the time- plateau was observed near the non-divider wall. dependent, three-dimensional Navier- Secondary flow at the entrance of the carotid Stokes equations is based on a special sinus was completely directed towards the divider finite element method. wall but halfway the bulb it shows a vortex, quite The numerical results show the behavior similar to the vortex found in a curved pipe. At of the time-dependent wall shear stress the end of the bulb secondary flow was highly In- vector field and identify the zones of fluenced by the tapering geometry at this site. high and low wall shear stress values From the data obtained it is concluded that secon- during the cardiac cycle. dary flow has a pronounced effect on axial flow. From a numerical 2D-simulation it was found that the main differences between 2D- and 3D-flow are due to the absence of secondary flow in the 2D-situation.

BE18-D.4 BE18-D.5 A Computational Approach to Biomedical Electrical model of the circulation without Fluid Dynamics "blood flow through the vessel wall" J.R. Tsitlik', H.R. Halperin, and F.C.P.Yin, A. MazHer , D. Giddens, J. Ekaterinaris Georgia Institute of Technology The Johns Hopkins Medical Institutions, George W. Woodruff School of Baltimore, MD 21205 Mechanical Engineering In electrical models of the circulation, the Besides biochemical factors, fluid mechanical compliance of a vessel is represented by a factors may play an important role in athero- capacitor.Inside a blood vessel,a change in the volume of blood causes an equal magnitude but genesis. The local wall shear stress, and the opposite sign change in the volume outside the local velocity distributions are of primary interest in the hemodynamical aspect of athero- vessel. In models, these changes are represented by equal magnitude but opposite sign changes in sclerosis. Detailed information about the flow the charge on the plates of the capacitor. The field can be obtained from mathematical model capacitance is generally connected in parallel studies. Hence flow patterns are investigated here in a simplified three-dimensional arterial with a resistor representing the viscous friction. Therefore, charge can move from one model. Theblood is considered as an incompressible Newtonian fluidobeying the Navier- plate of the capacitor to the second via this Stokes equation of the fluid flow and a com- resistor. Such a current is equivalent to a putational approach isemployed to solve these volume of blood moving from the inside of the vessel to the outside, i.e. "flow through the equations. Acomputerprogram is written utilizing the spacemarching method to inte- vessel wall". To prevent such "flow ", electrical models of the circulation should have the second grate the Navier-Stokes for steady blood flair plte,of the capacitor connected not to the for different geometries. The program is used resistor, but to a terminal representing the to solvemany problems that ofinterest to pressure in the space outside the vessel,e.g. hemodynamic studies. Results for slow through curved stenosis and through curved aneurysm are the intrathoracic space. This space, in turn, should be connected to the ambient space via obtained and discussed. These results show the another capacitor representing the compliance of ability of the numerical methods to calculate the surrounding structure, e.g. the thoracic the velocity fields with separationand cage. Thus, proper connection of these capacitors secondary flows and wall shearstress for 3D can prevent erroneous model predictions, and may configurations. lean to better understanding of the circulation. 995 1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas 375 Biomedical Engineering Scientific Papers

BE18-D.6 13C113-E.1 Modelling Competitive InterferenceBetween A method for frequency stabilization of Adjacent Terminal Lymphatics, W.A. Hyman* mtdtimode He-Ne lasers in laser-Doppler Bioengineering Program, Texas A&M Univ., flowmetry, P.A. oberg, Dept of Biomedical College Station, Texas, USA. Engineering, University of Linkoping, S-581 85 Linklipi, g, Sweden.

A simple model of the fluid dynamicsofthe In most commercial laser-Doppler instruments for blood terminallymphatic collection system suggests flvw measurements unstabilized He-Ne lasers that are used as increased flow throughonevessel in light source. These low power, low-prized lasersare often response to micro localized edema,or other impaired by frequency instability to an extent whichcan be tissuepressureelevation inthesupplying interpreted as blood flow variations. The mechanism behind tissue,will cause a decrease in flow in the frequency variations are changes of length of the laser adjacentlymphvessels which meetatdistal cavity due to thermal expansion. branch points. The effect of this decrzased Such instabilities can be efficiently reduced bymeans of a flow is toretard fluid collectionthrough feedback temper...are control system sensing the intensity of theseadjacent vessels and therefore tocause two orthogonally polarized modes. The difference between thetissueregion of their originto also the two mode intensities is used as an error signal accumulate fluid and tend to becomeedematous. controlling the thermal expansion of the laser cavity. The result of this analysis is thatedema could The stabilization device has been evaluated in models and appearto propogate at themicrocirculatory comparative blood flow measurements. Stabilization of this level. The fluii mechanics of this phenomenon type is recommended to be used in laser-Doppler flowmeters issimilar to effects in thearterial system especially in applications where low perfusion levels are of called steal where changes in relativearterial particular interest. branch resistance can cause, inthe presence of Acknowledgements, Hakan Rolunan, Hans Pettersson and significant proximal resistance, a marked Hfikan Debl6n contributed substantially. National Swedish decrease in branch flov when thecontralateral Board for Technical Development (STU) suppmed the resistance is reduced. work with grant 86-4217.

BE3.8-E. 2 BE1B-E.) New Laser Doppler Velocimeter with a Dual-Fiber Local blood perfusion In human anterior tibialmuscle Pickup examined using single fiber laser Doppler flowmetry M.Goto *, O.Hiramatsu, M.Kagiyama, K.Mito, E. Goran Salerudt and Knut Kvenie5o2 Y Wada,S. Tadaoka, S.Matsuoka, Y.Ogasawara, Dept Biomedical Engineering, University Hospital,Linkoping, D.Kilpatrick, K.Tsujioka, F.Kajiya Sweden' and Akers University Hospital.Oslo, Norway2. Kawasaki Medical School, Kurashiki, Japan Laser Dapple- flowmetry has successfullybeen applied to To measure the blood flow velocity accurately by superficial tissues. Local perfusion deep withintissues has not excluding lower blood %elocities stagnated at thevicinity been studied extensively. Single fiber laserDoppler flowmetry of the fiber Up, we developedalaserDoppler offers a possibility to measure the perfusioncontinuously velocimeter (LDV) with a dual fiber pickup. The pickup within a tissue and with a minimum oftrauma. consists of two thincladfibers(agraded-index In twelve healthy volunteers, local perfusionin the anterior multimode, clad. 62.5um and core. 50p.m) whichare tibial skin and muscle tissueswas evaluated. Skin perfusion placed side by side. Thus, the measured spotwas was recorded with a standard probe, while in muscle the extended away from tne fiber tip. The He-Ne laser beam single fiber technique was used. Normalresting values and (632.8nm, 5mW) was focused on the entrance of responses to induced occlusive ischemia of 0.5, 1, 3 and 5 one minutes were studied. fiber and introduced into the blood stream witha spreading angle of 5% The backscattered lightwas All recordings were found to be stable andpermitted collected by another fiber with a receiving angle of 17.3'. evaluation. At rest, muscle perfusicnmeasurements lacked Doppler signals were analyzed by a spectrum analyzer. vasomotion, while it was observed inmost skin recordings. The duration and magnitude of the muscle The spectrum of the Doppler shift frequency showeda postischemic response increased with occlusion periods that sharp peaked pattern. The frequency correspondingto were shorter the peak power showed an excellent linearity with the than 3 minutes. At longer occlusion periodsduration alone known blood velocities. increased. In skin the increasewas seen up to 5 minutes. The In the poststenotic region of onset of the response was instantaneous for both tissues. the canine coronary artery, broad spectrumwas The single fiber method has obtained, whereas narrow spectrum was obtained in proven a great potential in normal coronary artery. evaluation of deep tissue perfusion. Improvedtechnique and This indicates the utility of our protocols will make possible the investigation system to the analysis of the disturbed flow region. of deep tissue perfusion, not only in muscle tissue, butalso in tissues like t). kidney and brain. 43' v

376 1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas Biomedical Engineering Scientific Papers

BE18-E.5 BE18-E.4 Pressure Drop and Velocity Measurements Development of LDV Equipments for Blood In a 90°-T-Junction, D.W. Liepsch*, K.Homma*,T.Eiju, Flow Measurement, R. Schwimmbeck and T. Austin, Eisenhower K.Matsuda,_J.Ohtsubot and K.Shimizu, MedicalCenter, Rancho Mirage, CA, USA. Mechanical Engineering Laboratory, MITI

, 1-2, and theShizuoka University' Hemodynamic atherogenic factors are determined by Namiki, Tsukuba Science City,Ibaraki the interaction between geometry, flow 305 JAPAN pulsatility, elastic wall motion and the non- flow behavior ofblood. Flow Doppler Newtonian The equipments of LDV (Laser pulsatility and non-Newtonian effects were Velocimetry), that is, a LDV microscope and a studied in 90°-T-junctions simulating the first LDV fundus camera,based on the differential septal branch of theleft decendingcoronary method have been developed.This method has some artery. Different types of fluids were compared: advantages that blood flow can be measured in a Newtonian, shear thinning, time varying small observing area with high S/N ratio, etc.. polyacrylamide solutions, and two-phase fluids, The LDV microscope is mainly consisted of the e.g. Tri(hydroxymethyl) aminomethanewith 40% differential LDV system, and of the microscopic ghost cells (red cells without hemoglobin). The observation system for blood flow and tissue viscosities of these fluids were measured in a beams are adjusted to structure. Two laser Couette viscometer and compared with human blood. produce fringes and focused on a measuring area The pressure drops across the bifurcation were in blood (minimum value : 15 pm in diameter) measured in the straight tube and in the branch. flow. A frequency shifter included in the LDV The loss coefficients were calculated from the optical system(shifted frequency : 160Hzto measurements. Velocity profiles were measured 7kHz) separates signal component from pedestal with a laser-Doppler-anemometer. The non- one, and the flow speed is measured in a range Newtonian fluids showed additional and different On the other hand, the LDV of 0 to 14mm/sec. flow separationzones, comparedwith the fundus camera is used for the measurement of Newtonian fluids. Their secondary flows were human retinal blood flow in-vivo with the same also different. The pressure drop was smaller Fourier and Hadamard transforms. LDV method. than for the Newtonian fluid. It was found that and autocorrelation are used for analysis of the pressure loss coefficier' is afunction of blood flow from the signals detected by a photon the flow rate ratio and of the absolutevalue of counter. the entry flow rate.

BE1B-E.6 BE18-F.1 Angioscopic Observation of Blood Flow Magnitude Magnetic Resonance Imaging of a Stenosis, D.N. Ku*, Y. Elgendy, R.I. Y.Uchidal T.Tomaru, F.Nakamura,T.Sugi- Pettigrew, Georgia Institute of moto, University of Tokyo,Tokyo,Japan Technology, Atlanta, GA and Emory University, Atlanta, GA The changes in blood streaminstenotic artery and their relationship tothrombo- Magnetic resonance imaging (MRI) offers the sis were examined by angioscopyin dogs. possibility of producing non-invasive The left common iliac artery wasperfused angiograms. The goal of this study was tobetter with K-H solution and backflow ofblood understand the physical basis of flow patterns clinical from a lumbar artery wasdrained into the seen by MRI and their likely perfusion circuit. Before constriction, significance. A fast gradient echo was used on a the blood streams in the artery werecom- Philips 1.5 Tesla system. The images were gated posed of a coaxial main streamand the for the recording of multiple images and streams beside the coaxial stream.After combined into amovie to visualize the dynamic A more than 50% constriction,the streams blood flow patterns during the cardiac cycle. beside the coaxial stream wereseparated, non-magnetic system of pulsatile flow through a and produced curles aroundthe outlet of stenosis was developed. the stenotic segment. Also,the outer MRI magnitude imaging provided a high lAyer of the coaxial stream wasseparated resolution image of thelumen of the stenosis. and produced longitudinal curles. Throm- Flow through the stenosis showed acceleration of bi appeared at the center ofboth types velocity in the throat as a brightening of the the throat, black of curles. Similar changes in blood st- intensity. Justdistal to ream and thrombosis wereobserved in the circles are formed. The jet formed by the common carotid arteryduring vasosp-sm. stenosis appears light with a dark surround The results indicate usefulness ofangio- giving the impression of a flame thrower The scopy in serial observationof blood st- black areas appear to correspond to regions of ream and thrombosis. steep velocitygradients, vortex shedding, and turbulence. MRI can provide both an anatomic and physiologic evaluation of hemodynamically significant stenoses. 37 377 1988 World Congress on Medical Physics and BiomedicalEngineeringSan Antonio, Texas Biomedical Engineering Scientific Papers

BE18 -F. 2 BE18-F.3 VELOCITY PROFILES IN STENOSED TUBE Blood Flow Images by Magnetic Resonance Imaging MODELS USING MAGNETIC RESONANCE IMAGING S.Watanabe*. H.Imai. Y.Masuda and Y.Inagaki (MRI), S. Rittgers , D. Fei, K. Kraft The Third Dent.of Internal Medicine. ChibaUniv. P. and Fatouros, TheUniversity of School of Medicine. 1.8 -1 Inohana. Chiba, Japan Akron, Akron, Ohio Slow blood flow images are seen in Only a limited number of techniques some kinds of the exist which can provide direct knowledge cardiovascular diseases by MRI.In this studywe will about blood velocity and flowrate in the present a clinical application of slow blood flow human cardiovascularsystem withthe images in the aortic dissection and mitralstenosis. majority ofthese requiring invasive The false lumen is clearly distinguished fromthe Procedures. A time-of-flight MRI true lumen since the signal intensity of false lumen velocity measurement technique is is higher due to slow blood flow than thatof true evaluated in a constrictedtube model lumen in MRI. But a location of the entryis not ne- having stenosesof 0 to 80% at cessarily detected. We obtained the result thatthe physiologicflowrates. Results from false lumen near by the entry showedvery low signal study this show thatMR displacement intensity due to the turbulent flow froman experi- images can: 1) be obtained within both ment of the continuous flow model of aorticdissec- laminaI and disturbed jets (maximum tion. The location of an entry uiagnosedon the base stenoticRe ! 4,200), 2) accurately of this result was coincident with measuremean angiographic fin- jet velocityup to 172 dings in the majority of cases (95%). cm/s, 3) detect low forward and reverse In mitrel stenosis it is well-known that velocity regions near the tube wall just the thrombi often exist in the left atrium which downstreamof the stenosis (LID < 2), causes general and, 4) demonstrate return to laminar thromboembolism. But in our study about 13patients Profiles well downstream (L/D = 10) of with thromboembolism. the thrombiwere detected only the stenosis. Regions between the jet 'n seven patients (54%) byx-ray CT. echocardiogram termination point andre-establishment or MRI, although slow blood flow images in theleft of la inar flow (Re > 1500, > 1000 and > atrium were seen in 11 patients (85%) by MRI.It is 110 downstream of 40%, 60% and 80% supposed that a detection of slow blood flowimages stenosis, respectively) can nQ in the left atrium is clinically useful. presently be detected by this technique.

BE18-F.4 Qualitative and quantitative Investigation BE18-F.5 The Velocity Profile of Blood Flow in of the Blood Flow Behavior by MRI. Presence of Intensive Magnetic Fields, M.Bucciolini*, GC.Casolo, L.Ciraolo and Chen and S. Saha,Departments of R.Renzi, University of Florence, Italy. Physiology and Biophysics, and Orthopaedic Surgery, Louisiana State University Medical MR Imaging is a Center, P.O. Box 33932, Shrevepert, Louisiana flow sensitive technique 71130. which has the potential to give qualitative and quantitative information Due to the increasing use of largemagnetic about blood circulation.The possibilityof fields in magnetic resonance imaging(MRI), it studyingthe behavior of blood flow by is important to investiage the effectof such Magnetic ResonanceImaging looks intensive magnetic fields on the bloodflow in as an our body. attractive goal,but quantification of In this study, we have analyzed the flow general case of blood flow for velocity has yet to be fully accomplished. a wide range of Hartmann's numbers (0.1 to 2.5) bysolving the In order to relate the complex MR signal magnetohydrodynamic equation for viscous blood with proton flow velocity, a calibration flow. The result indicated that the velocity in phantom has been carriedout by means profiles of blood flow were significantlydiffer- of an artificial flow circuit.Invivo ECG ent for Hartmann's number 0.5 and beyondcom- gated images pared with velocity profiles in theabsence of were then acquired by a magnetic field. employing both conventional Spin These results will be of Echo and interest in the development of future MRITech- Multiphase FastImaging sequences.Curves nology for diagnostic purposes. of signal from thoracic vesselsversus the phase of the cardiac cycle were obtained.Quantitative preliminaryresults concerning velocity trendover time are deduced and differences between normal subjects and patients with heartfailures are pointed out.

378 1988 World Congress on Medical Physics andBiomedical EngineeringaSan Antonio, Texas Biomedical Engineering Scientific Papers

BE18-F.6 BE18-G.1 Quantitative Measurement of Plug Like Limitations and Validation of the Modified and Laminar Flow in MRI, R. J. Kriz*, Bernoulli Equation for Velocimetric Estimation G. Dobben and A. Horowitz of Pressure Gradients, S.Hanya*, S.Nishiyama, Y.Kondou, A.Ishihara, and M.Sugawara**, MRI signal intensity has been previously found *Kitasato University, Sagamihara, **Tokyo to vary with flow in various ways depending on Women's Medical College, Tokyo, Japan the pulse sequence used. We have measured the For patients with various types of stenosis, the average signal change for flowing vs static modified Bernoulli equation has been used to yield water as a function of velocity for several an accurate estimate of transstenotic gradients. commonly used pulse sequences in different Theoretically, the equation is only applicable for diameter tubes on axial images. Differences a stenosis downstream from which the cross-sectional in average intensity and standard deviation in area enlarges suddenly (orifice-shaped stenosis), the ROI's are measured for plug like flow vs however no previous study had systematically con- laminar flow. Plug flow occurs when fluid firmed this. enters a tube from a static reservoir under We measured, calculated and compared pressure gra- steady conditions. The transition to dients in 40 patients with stenosis (34 pulmonary laminar flow causes the average MRI numbers to and 6 aortic) divided into two groups (27 orifice- decrease, thus fluid having the same average shaped stenosis and 13 hourglass-shaped stenosis). flow but different velocity profiles at Peak jet velocity and pressure gradient were different locations along the tube, have measured simultaneously using a multisensorcatheter. different signal intensities. The correlationof simultaneous velocimetric and manometric pressure gradients for the patients with orifice-shaped stenosis demonstrated a highly significant relationship (Y=0.83X+2.012, r=0.942, P<0.001). In patients with hourglass-shaped stenosis, there was no significant correlation (r=0.229,n.s.). We conclude that the use of the modified Bernoulli equation for estimating pressure gradients should be limited to patients with orifice-shaped stenosis.

BE1B-G.2 BE18-G.3 STEADY AND PULSATILE FLOW VISUALIZATION Pulsatile Flow Through Models of Compliant IN THE ABDOMINAL AORTA Stenoses, Donald F. Young* and Nikolaos Moore, J.E. Jr.,* and Ku, D.N. Stergiopulos, Department of Engineering Georgia Institute of Technology, Atlanta, Ga Science and Mechanics, Iowa State University, Ames, Iowa50011, USA Carotid atherosclerosis develops in regions of low and oscillatory shear stresses. Similar hemodynamic factors may influence plaque The pressure-flow relationship for a stenosis is formation in the abdominal aorta. Steady and of importance since this relationship is re- pulsatile flow visualizationwas performed in a quired to determine the effect of the stenosis glass tube model of the aorta. Dimensions for on blood flow. Host past studies have con- the model were taken from bi-planar angiograms sidered stenoses to have rigid, fixed geome- and cadaveric aortas. tries, but stenoses are often eccentric so that Under rest conditions, two regionsof a portion of the arterial wall remains compli- relative stagnation appearedon the posterior ant. wall of the aorta. Secondaryflow patterns In the present study the influence of a appeared in the abdominalaorta, but not in the compliant wall on pulsatile flow through in thoracic aorta. Under exercise conditions, one vitro models of both subcritical and critical thin stagnation region appeared and no secondary stenoses was investigated. For some tests the flow patternswere visible. With pulsatile geometry of the stenosis changed only a small flow, the stagnation regions exhibited amount during the flow cycle, but for critical oscillatory motion. The secondary flow patterns stenoses this small change affected the pressure became more complicated, but still appeared only drop across the stenosis. In other tests the vessel containing the stenosis collapsed because in the abdominal aorta. These vortices of the low distal pressure created by the propagated downstream during systole, with stenosis. Pressure-flow relationships are additional vortices forming in early diastole. described for both the uncollapsed and collapsed For pulastile exercise flow conditions, modes of behavior. virtually nosecondary flow patterns appeared. Results from this experiment show that complicated flow patterns exist locally in the abdominal aorta which may influence the development of plaque at this site. 3 9 9

379 1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas Biomedical Engineering Scientific Papers

BE18-6.4 BE18-G.5 Fluid Dynamics and Mass Transfer to and from the The Effect of Minimal Atherosclerotic Plaque Vessel Wall in Models of Arterial Stenoses, X. on the Spatial Distribution of Wall Shear Deng and T. Karino*, Montreal General Hospital, Montreal, Quebec, Canada. Stress A ModelStudy -T. Yamaguchi* and S.Hanai, Vascular Pathophysiology Lab., National Car- diovascular Center, Suita, Osaka 565, Japan. To investigate the effects of disturbed flows n' the genesis anddevelopment of atherosclerotic Low wall shear stress (WSS) regions are specific lesions in man, fluid dynamics and mass transfer arterial locations prone to atherosclerosis. However. to and from the vessel wall at sites of arterial it is likely that the WSS distribution pattern changes stenoses were studied theoretically using computa- markedly when atherosclerotic wall irregularities tional methods. Flow patterns and various other affect arterial geometry. Very few studies on exact fluid mechanical properties for the steady laminar flow of an incompressible fluid through axisym- shear stress distribution patterns following the metricbell-shaped stenoseswereobtained for appearance of an atherosclerotic lesion have been Reynolds numbers up to 100, by numerically solv- reported. We measured model WSS distribution ing the Navier-Stokes equation. Mass transfer to pattern using an electrochemical method. There was and from the permeable vessel wall of the stenoses a striking and complex change in the WSS distri- was also studied theoretically by solving a diffu- bution around the small model plaque, even when sion equation in combination with a penetration the reduction in area produced by the obstacle was theory. It was found that mass transfer from the only an order of magnitude of several percent. The vessel wall to the flowing fluid is highest at the throat of the stenosis, lowest at the separation WSS value increased to twice or greater than twice pointand moderately elevated within the vortex the Poiseuille flow value. Our results suggest that formed downstream of the stenosis. Analyses of the changes in WSS induced by the appearance of masstrans'er from the fluidto the permeable an early atherosclerotic lesion may markedly alter vessel wall showed that an accumulation of species the course of the disease. such as lipoprJteins on the vessel wall is lowest at the separation point and highest at the reattachment point. There was a minor secondary peak at the leading edge of the stenosis proximal to the point of maximum constriction.

BE18-G.6 BE19.1 Arterial Function in Postural Change, T. Yaginuma*, T. Noda, H. Komatsu, AUXILIARY FROCEMOR H. Sekiguchi and S. Hosoda, Jichi FM FAST MaiMICAL SICNAL. ANALYSIS Medical School, Tochigi, Japan. .J.Azpiraz. Leebank. A. Deluca Panaccia ThiversidedAuttinana Metropolitans-Iztapalapa The effects of body position on the aortic flow Av. Purlsima y Michoacan. Mfr_xlco. 09340. D.F. velocity and pressure signals have been studied ITD(ICO in adult male subjects by the aid of multisensor catheters. For each patients input impedance tie have designed an auAlliary processing was calculated from 10 harmonics of aortic system capableof providing real-time signal pressure and flow. The time domain wave form of processing within an IBM-PC compatible=water, forward and backward flows and pressures was where the architecture of this coprocessorwas obtained by computing the inverse Fourier- developedspecificallytoacceleratethe transform. exermtion timeofcharacteristicsdhrouttnes The ascending aortic pressure wave form revealed used for analysis of biceiedical signals suchas a late systolic peak during decelerating flow the fast Fburiertransform(FM,convolution velocity in supine position. A large change and squared. magnitu1e. from supine to standing was to have reflections minimal or more diffuse in spite of increasing For example, during the calcination of the input impedance. This was observed similar in FFT most of the processing time is used for the shape from the position of 60° tilting bed. multiplicationsinthe"tutterfly". The Wave shapes of pressure from the straight rAlprocessor wasdesigned to executethese position of a patient supported by his hip on operations efficiently, and uses a 16 bit X16 the edge of the bed, revealed a intermediate bit nsultiplierchip (MPY01EFIW 11W) capable of form between supine and standing. executing a mult.iplicaticnin 10Cns,together These results suggests that the magnitude and with other Schottky -TM elementsin order to timing of wave reflections from upper and lower operate concurrently with the host body sites are markedly changed not only by the microcanputer.This system is able to trans-ice-in gravitutional effect but by physiological a general purpose mirrocanputer into a powerful modification. real-time blomedica3 signal analysis system. C 0

380 1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas

L Biomedical Engineering Scientific Papers

BE.1.9.3 BE19.2 ADAPTIVE RECURRENT FILTER FOR ECGSIGNALS "A New Phase Linearization Method for the Common N. V. Thakor, Z. Yi-Sheng, J. Wang, Recursive Digital Filters Used in Medical Signal Johns Hopkins Processing", P. Jarvilehto* and K-P. Estola, Biomedical Engineering, Technical Research Centre of Finland, POB 316, University, Baltimore, MD21205, USA 33101 Tampere, Finland. In computer-aided analysis ofElectro- cardiogram (ECG) signals two problems are Medical signal processing typically utilizes removal of uncorrelated the sharp edged filter structures to extract important: 1) interference, and 2) classification of QRS - frequency bands of interest, e.g. in ECG and EEG An adaptive recurrent signals. In real time systems where common filter complex morphology. filter (ARF) is presented that learnsthe blocks are allowed to have only short delay time each and the minimal intersample processing time, the shape of the QRS complex only at recurrence. Uncorrelated interference is recursive filters are exploited and the phase beats, and response cannot thus be kept exactly linear. cancelled fromsubsequent Unlinearity in the phase response causes signal abnormal QRS morphology is detected. distortion due to different group velocities of The ARP is constructed as a transversal filter with adjustable weights.Adaptation the signal components. of QRS In this paper, we introduce a method for the is started ateach occurrence phase linearization of thecommon IIR filters. complexonly. The filteralgorithm The phase linearization is based on the stable minimizes the mean-squared error between implementation of the maximum phase equivalent the filter weight and the incoming QRS of the original filter. The infiniteimpulse sample with the aid of a least-mean-square response is truncated to a finite one, the (LMS) recursive estimation algorithm. data is reversed in time domain and fed back to After the filter has converged (convergence similar filter again. This implies ausage of rate is determined by the LMSalgorithm), shape of block processing. After another time reversal the ARP weights have "learned" the and a overlap-and-add section linear phase data the normal QRS complex. Uncorrelated noise from is achieved. In somecases the time reversal or ectopic beats remain astheresidue method can be replaced by the maximum phase FIR the filter. beat equivalent of the original filter, thus enabling The ARF has applications in ectopic usage of the stream processing instead. detection, pacemaker analysis, P-wave and atrial arrhythmia analysis, and noise removal.

BE19.4 BE19.5 w0APIIVE INRESNOLO PREDICIION INREAL-11HE Digital Filters for Tracking Electra- ECG DEfECIION physiological Chaos, Bernard Saltzberg: thou Song-kal,Wang Jian-tapauJun-rong* University of Texas Medical School, Shanghai Jiao long Univercity Houston, Texas 77030

We compare EGG signal processedwith a threshlod but arrhy- to decide if any ORS-complex exits. Current investigations of the chaotic fea- non-regularly and non-stable. thmia ECG changes tures of electrophysiological activity are So threshold adJustmert must be adjusting adap- mainly focused on deriving the Hausdorff real-time tively to fit the variance of EGG in dimension of short samples of normal and detetion. the paper discusses the properitiesof Abnormal EEG patterns. Because of the com- thresholds with linear estimation which need [- putational demands for deriving the Haus- thresholds. order priori statistical knowledge of dorff dimension, as well as the incompati- although it has good proper- It is nt practical bility of epoch length and stationarity inspira- ties in theory. but it gives us some reqpixements, this approach is severely prediction. tions to lead to adaptive threshold limited. Filters based on nonlinear models prediction Then the paper discusses threshold which exhibit chaotic behavior offer an forms adaptive with 1-order predicting filter, alternative method of chaotic analysis principle of threshold prediction formula on which overcomes the above problems.These minimum square error, and explains frequency digital filters can provide high resolution respones of adaptive threshold prediction. We continuous tracking of a chaotic feature of assembly realize the real-time program with the electrophysiological activity in essen- introduces language in Intel-8085. the paper tially real time. In addition these filters predicting threshold prediction with 2-order can be implemented on a standard laboratory in threshold pre- filter, to the application computer, as opposed to the requirement for dictions, explains and simplifies the formula. supercomputer resources to comoute the Haus- results can be abstained than The such better dorff dimension for comparable amounts of others. data. lo compare a variety of thresholdpredicting compares 4 methods quanitatively, the paper kinds of threshold prediction methods with Mean square-error (SS!) and mean-error(SS2). 401 1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas 381 Biomedical Engineering Scientific Papers

BE19.6 BE20-A.A. Infracardiacspectral components in cardio- The use of dual simultaneous excitation currents vascular signals,R. Pallas-Areny*, J. Colo- in Bioelectric Impedance Analysis, D.N. Smith; minas,J.Rosell, and P. Riu, Div. Instru- Medical Physics and Medical Engineering, Western mentacion yBioingenieria,ETSI Telecom, General Hospital, Edinburgh, EH4 2XU, Scotland. Box 30002, Barcelona, Spain Tetrapolar impedance measurements have been widely The spectral analysis of ECG and blood pressuresig- advocated for determining body composition. In nalsis a meansof studyingthe intact cardiovascular this method a single excitation current passes control system.In thisanalysis,thepresenceof between two extremities, usually right hand and infracardiacrhythms is usuallyassociated with respi- right foot. Body composition values are derived ratoryarrhythmia,thermoregulationandspontaneous using multiple regression relationships incorpor- vasomotor activity.Thereis noinformation available ating Height /Resistance as the principal variable. aboutanyother mechanisms which result inspectral This is taken to represent the volume of the body components in the same bands.But there is,however, but will be incorrect due to variability of cross experimentalevidenceof respiratoryinformationin section and internal resistivity. Greater thespectrumof thesesignals even whenrespiratory precision can be achieved by making separate meas- influence on them via neural control has been purpose- urements of impedance for the limbs and trunk and ly suppressed. This fact remains, so far, unexplained. combining these with appropriate weighting factors. We show that, evenin the absence of respiratory sinus However, the use of independent serial measurements arrhythmia,an amplitudemodulation of cardiovascular results in changes of current distribution within signals by respiratorymovementsresults in a spectrum the body and will not eliminate the end effects showing components in the same bands considered when of each section. The use of two currents of equal analyzing the intact cardiovascular control system. magnitude, frequency and phase in parallel paths Experimentalresults in a pacedpatient confirmthis on the right and left sides automatically reduces hypothesis. We show too that ifthe respiratory rhythm the effect of the spreading resistances at shoul- is not constant,some infracardiacspectral components can be attributed to its variability. der and groin producing a more uniform current distribution in the trrk.An appropriately weighted sum of Height / Resistance for all segments can then be produced which is a more accurate representation of body volume.

BE20-A.2 BE20-A.3 Development of an Electrical Impedance TOmJgraphy Study of the New Temperature Change Measurement System, L. Bertenstaml B. Bled, G. Gdransson, Method by Impedance CT, K.Sakamoto, H.Kanai U. Lindblad and R.B.R. Persson, RadiationPhysics Sophia University, Chiyoda-ku Tokyo, Japan 102 Department, Lund University, S-221 85 Lund, Sweden. In the hyperthermia therapy, one of the most important problem is how to measure the temperature of It has been shown, both in theory and in practice, cancer tissue located deeply in the body noninva- that it is possible to obtain information about sively. Electrical impedance of living tissue the distribution of impedance within an object by mainly consists of intra- and extra-fluid impedance driving current trough it and measure the volt- whose characteristics are almost the same as that ages that arise on the boundary of the object. of saline solution and cell membrane which is elec- We have developed a system for making suchmea- trical insulator at radio frequency. Therefore, the surements in vivo on humans. The system utilizes temperature characteristics of tissue impedance is 16 electrodes, which are used for alternatively the same as that of saline solution which is about driving currents and measure voltages. The equip- -2.5%/°C. Cancer tissue is heated up to around 43°C ment is controlled by a PC/AT compatible computer. during hyperthermia therapy, so impedance of caner The computer is also used for data collectionand tissue decr.ases about 18%. We tried to measure the image reconstruction and presentation. impedance change by impedance CT and estimated the temperature distribution in the object. And also The present algorithm relies on an enhanced back we discussed the measurement errors caused by the projection technique, using 15 curved fields and various kind of noises and geometrical problem by 16 projections. The program automatically compen- computer simulation and phantom experiment. To take sates for offset and gain errors in the hardware. an example, the temperature distribution obtainby With this system it is possible to aquire data impedanceCT agrees well that measured actually and present a two-dimensional picture of impedance within 0.5°C differences in one of the phantom distribution in a 32x32 matrix. experiments. The errors in this experiment could be The future reconstruction algorithm willuse an fathomed from computer simulation results. Even iterative forward technique, based on the Finite though we got good results in the phantom experi- Element Method (FEM). ment and computer simulation,since geometrical problem still exists, we continue to study the feasibility of the temperature measurement method by impedance CT. 402 382 1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas Biomedical Engineering Scientific Papers

BE20-A.4 BE20-A.5 Skin Electrical Parameters Obtained ELECMIC IMPEDANCE FOR EARLY DETERMINATION OF from Low-frequency Admittance Locus BRAIN DEATH AND FOR PROM= IN COMA. Plots, Z. -G. Qiao' and L. Morkrid, S.A.Radvan-Ziemnowice, Cerebral Clinical Center Dikemark Hospital, N-1385 Splberg, Bethesda,Maryland 20814,U.S.A. Norway. Rheoencephalcgraphy (REG) uses noninvasive electric impedance plethysmography for cerebral circu- The electrical impedance of biological tissue tory studies. It was markedly improved by author's can be presented by a Cole-Cole locus plot in "regional" REG (RREG). In prolonged and intermitte- the complex plane, usually in a circular form nt surveys, this method proved to be safe,more sen- with parameter (R0, fl,, 7;:nd ot). For the in- sitive and reliable,than other indicators for brain vestigation of human electrodermal activity viability,e.g. EEG and contrast or isotope arterial which vary in time with sympathetic outflow perfusion. For experimental testing was also used there are difficulties in obtaining these pa- author's "topischemia"-the transient focal cortical rameters. The theory developed in this paper hypoxia and asphyctic anoxia. shows that an admittance locus plot also In clinical tests for coma of various etiology, a part of a circle and its low-frequency re- the flat tracings of RREG have been more dependable, gion (41000 Hz) is to a very good approxima- than the iscelectric EEG, absence of clinical signs tion straight line. of cerebral functions, or any of the invasive,expe- Furthermore from simultaneous measurements at nsive and complicated blood flow tests (1980). two different frequencies with three electrod- Tracings of impedance fluctuations caused by ar- es, two luck-in amplifiers and a computerised terial pulsations,are monitored simultaneously fran recording system, it is p,ssible to estimate bilateral pairs of scalp electrodes in repeated se- in real time the parameters (1/120,7, and -k). quences: frontal,temporal,occipital and occasional- 1/R0 corresponding to zero frequency conduct- ly also parietal. Rheograph measures the impedance ance extrapolated from the low-frequency part changes by floating potential i. advanced implemen- of the admittance plot. This new method of tation of double Wheatstone bridges bipolar circuit. measuring and characterising electrode-tissue Two signal oscillators have 20-80 kHz and up to 10 interface has been used to study spontaneous mA. Amplification is in the range of EKG,about 50 x activity during baseline conditions and evoked lower that for EEG. Computerization of RREG,called responses of these parameters in psychophysio- now CRREG brings a new modality of neuroimaging for logical and neuroendocrinological research. being developed by this author, advanced studies on human cerebral dynamics.

BE20-A.6 BE20-B.1 Impedance Measurement ofChinese Meridian Intracardiac Impedance Plethysmography, System. Maw-Chang Lee*, Wen-Chang Lee, Yin Chang G. Boheim* and M. Schaldach, Zentralinstitut and ChiehChung, NationalYang-MingMedical fur Biomedizinische Technik, UniversitUt College, Taipei, Taiwan, R.O.C. Erlangen, D-8520 Erlangen, FR of Germany

Accordingto the traditional Chinese medicine, Intracardiac impedance plethysmography is gaining the vitalenergy 'chi' circulatesaroundthe increasing interest, in both diagnostic and ther- by along the channels called 'ching-lo' apeutic fields. As a diagnostic tool, its main (meridians). The acupuncture points (acu-points) application is to record pressure-volume dia- are located on the meridians.Stimulating acu- grams. Therapeutically, it is used as a means to pointsby needles or other means is torestore provide information of the pumping efficiency of the normalflow ofchialong thespecific the heart. This method of control finds applica- meridian, andthus hasspecifictherapeutic tion for implantable pacemakers and defibrilla- effects. However,thephysical substrate of tors. Impedance plethysmography has been shown to meridian is yet to be identified. be an effective physiological parameter for The purpose ofthis study is to examinethe closed loop control of pacing rate in pacemaker electrical properties of the meridian. applications, based on its direct relationship to Acupuncture needles were inserted into the acu- stoke volume. pointsofpatients. Sinusoidalcurrentsat Investigations with respect to the accuracy of variousfrequencies were passed through a pair this method, or its optimization, have mainly of needles and voltage drops were measured. been established either by performing in vivo or Impedance was calculatedfran the voltage- in vitro experiments, or using computer simula- current relations. Preliminary results show that tion to solve underlying differential equations the impedance of meridian did not differ under special conditions. significantly fran that of nearby tissue.More This paper contributes to the volumetric applica- rigorous studies are currently under way. tion of impedance measurements in providing solu- (Supported by NationalScienceCouncil of tions of the Laplace equation under simplified R.O.C.) geometrical conditions. Theoretical, experimental, and clinical results confirm the electrode arrangement needed to meet the required accuracy

, for the various applications and summarizes limi- 4 0 d tations of the method.

1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas 383 Biomedical Engineering Scientific Papers

BE20=B.2 BE20-B.3 Right ventricular volume estimation by conducti- Ventricular Volumetry by Electrical Conductance vity dilution, P.H.Rehak*, H.Metzler and W.Stenzl Measurement : the Importance of Catheter Position, University of Graz, A 9036 GRAZ, Austria C.D. Bertram* and J.C. Woodard, University of New South Wales, and B.S. Gow, University of To evaluate the conductivity dilution methodin Sydney, Australia. right ventricular volume estimationan experimental study was carried out in 7 pigs. Variations We have recorded measurements with an eight-pole of cardiac output(CO. and ventricular volumes catheter in the right ventricle (RV) of anaesthe- over a wide range were performed by applying com- tised greyhounds, and compared them with simul- binations of continuous positivepressure venti- taneous recordings of e.c.g., RV pressure, lation (PEEP), Dopamine infusion and volumeover- pulmonary arterial (PA) flow-rate and tracheal air load. Hypertonic saline was injecteddirectly flow. The catheter was introduced in four differ- into the right ventricle and indicatorconcen- ent ways: tip at apex from PA or jugular vein, and tration was measured in the pulmonary artery by tip near PA from jugular or femoral vein. Position a catheter-mounted 4-wire conductivity cell. in the ventricle was determined by videofluoro- CO, enddiastolic (EDV) and endsystolic volumes scopy allied to contrast injection and location of (ESV) were calculated from the unprocesseddilu- the tricuspid valve by pressure-sensing radio- tion curves, ''tom filtered curves and byexponen- opaque catheter. Our results show that usable tial curve fitting. These values were compared volume signals correlating well with integrated with the pulmonary artery flow as measured by ejected flow-rate are obtainable from all four an electromagnetic flowmeter and with myocardial positions if the catheter is maintained away from segment lengthsof the right ventricular free where the free wall meets the septum. In vitro wall, respectively. experiments and simulations confirmed that appar- Both CO and ventricular volumes showed thebest ent volume is markedly altered when electrodes correlation with the reference values when they approach a boundary, owing to reorientation of were calculated from the unprocessed curves (CO: flux lines. Hence, while isovolumic shape changes r=0.92, EDV: r=0.87, ESV r=0.71). causing blood shifts parallel to the catheterare The accuracy of the volume estimation - referred well compensated by a multi-electrode technique, to thesegment lengths - wasnot affected by shifts orthogonal to the catheter are transduced PEEP,Dopamine orvolume loading, but allof as apparent volume changes. The effects on appar- these factors showed a significant higher influ- ent volume of mechanical or natural lung ventilat- ence on the estimated CO than on pulmonary flow. ion generally accord with physiological expectation.

BE20-B.4 BE20-B.5 Electrical Impedance-Cuff Technique: Data Reduction and Modeling of Impedance A New Plethysmography, Cardiograms by Linear Prediction. H.Ito , H.Shimazu, A.Kawarada, H.Kobayashi, M.Cadene,T.Aljama,S.Charleston,S.Carrasco, J.Masuda & K.Yamakoshi, Dept Physiol, Kyo- R.Camarena.Universidad Autonoma Metropoli rin Univ, Mitaka-shi, Toxyo; K.Yamakoshi is tana-Iztapalapa.E.E Department,09340,Mexi in Res Inst, Appl Electr, Sapporo-shi, co 13,D.F.,Mexico. Hokkaido, Japan Impedance cardiography is a noninvasive technique A new plethysmograph, called theelectric for estimation of stroke volume and cardiac output impedance-cuff, was designed for the inin patients and normal subjects at rest andexer- direct measurement of blood flow, arterial cise.Controversial results has been reported by pressure and compliance in human limb arte- several investigators when they have studied car- ries. The device comprises acompression diac abnormal patients using Kubicek's model corre chamberfilled with electrolytesolution lated against Fick's and dye methods. anda tetrapolar electricimpedance ple- The purpose of this work is to describe a new ap- thysmograph whose electrodesareplaced proach to characterize impedance cardiography sig- inside the chamber; the former for control- nal by use of an only-pole 1_near prediction algo- ling transmural arterial pressure (P,) and rithm.The resulting model coefficients are usedas thelatter for detecting limb volum4 (V,), a data reduction feature extraction technique. mean arterial volume (V) and its pulsatile This model has been tested on twenty subjects dur- variation (AV ). Systglic (P ) and mean ing upright cycle exercise at 0,50,100,150 and .00 (P ) arterialapressure were dgiermined by watts estimates of cardiac output by CO2 re- daTcting respectively theappearanceand breathing method.Also,it has been tested in 10pa- maximum point of the pulsatile impedance tients with mitral valve insufficiency,determined variations during the gradual increasein by B-scan ultrasound images,with estimates ofcar- the chamber pressure using the volume os- diac output by Doppler ultrasound method. cillometric technique. Diastolic pressure First results suggest this approach can be used andpulse as pressure (AP) were calculated a very sensitive statistical technique to process fromthese pressuresusing conventional cardiography signals. equations. Compliance defined as AV/AP was obtained at various Plevels controlled by compression pressure.t

4 384 1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas Biomedical Engineering Scientific Papers

BE20-B.6 BE21-A.1. IMPROVED SIGNAL PROCESSING TECHNIQUES FOR Alternating Motor Unit Activation in THE DETERMINATION OF CARDIAC PARAMETERS Electrically Stimulated Muscles,Z, -P. Fang* USING THORACIC BIOIMPEDANCE andJ.T. Mortimer, Applied Neural Control Lab, Sridhar P. Reddy Uang Y. Shyu, Joachim Nagel, Ph.D., Dept. Biomedical Engineering, Case Western Nei Schneiderman, Ph.D. Reserve University, Cleveland, Ohio 44106 Noninvasive determination of stroke volume. cardiac ouptut and the systolic time intervals though of great interest in both diagnostic A new method has been developed to activate alternately and medical research, still lacks the necessary reliability and precision two motor unit groups within a muscle by using a nerve mainly due to severe shortcomings of the commonly used signal processing techniques. Due to the poor results, even the principle of electrode. With alternating stimulation, each of the muscle cardiac output determination based on impedance cardiography groups should be required to fire at only half of its remains controversial. fusion frequencytogenerateafused whole muscle High levels of noise and motion artifacts in the signals call for contraction.This stimulation scheme mimics the sophisticated signal processing techniques in order to provide asynchronous firing pattern of motor units during natural beat-tobeat detection of the single cardiac events, especially in case of excitation, and should therefore reduce the muscle fatigue exercising patients. So far, simple bandpass fitters and threshold associated with conventional high frequency stimulation. detectors have been used for event detection at the beginning and A tripolar spiral cuff electrode was placed on the medial end of electric and mechanical systole. Ensemble averaging of cardiac gastrocnemius nerve of the cat. A stimulator was designed parameters, increases the reliability but only at the expense of togenerate narrow rectangular pulses (which activate information about transient responses, short-time fluctuations and largeaxons along) and wider quasitrapesoidalpulses respiratory modulation. In order to overcome these problems we have used beattobeat (which activate selectively small axons by differential detection techniques, such as adaptive matched filtering and block ofthelargefibers)alternatelyatadesired phase-sensitive event detection. The phasesensitive method is switching rate. Both nerve action potential and muscle based on the assumption that sudden events within a biological signal twitch recording indicated that two different motor unit source, such as the onsets of slowly changing signals or small local groups could be activated alternatelyby the scheme. A extrema (as the opening of the aortic valve), lead to a sharp change of smooth muscle contraction was demonstrated when each signal spectrum. The analysis of the momentary frequency or phase muscle group was stimulated at about half of it fusion has resulted in obtaining a reliable marker for each event. As a result of frequency. Finally, a comparison test showed that muscles adaptive matched filter and phase-sensitive event detection, both the exhibitedlessfatigue when driven by the alternating precision have been substantially improved. In order to reliabilityand scheme. This method may have applications in process the data at sufficiently low cost and to gain clinical acceptance physiological study and neural prosthesis design. without sacrificing precision, we have developed analysis to be run on a personal computer (PC-AT).

BE21-A.2 BE21-A.3 Stabilization of FES System for Muscle Tension Time - variation of Quadriceps Muscle Force And Control, N.Itakura: K.Fujita, K.KUbo, Y.Iguchi Related EMG During Uniform Electrical and H.Minamitani, Reio University, StimulationInParaplegics. J.Mizrahi*, M. Hiyoshi 3-14-1, Yokohama 223, Japan Levy, Y. Steinvil, Z. Susak, P.Solzi. Technion- Isr ael Institute of Technology,Haifa, and Loewenstein Rehabilitation Hospital, Raanana, We have tried to control muscle tensionby using I sr ael. functional electrical stimulation(FES)with PID, PI, I, I-PD or I-P controller which wasdesigned A dynamic model of the leg, which enables to by the Kitamori's controller design method. And calculate the force produced in vivointhe gastrocne- stimulated Quadriceps muscles of spastic experiment have been carried out on par aplegics was developed.The measureo mius of human lower extremities. The system em- kinematics and external dynamics ofthe ploying PID, PI or I-PD controller showedstable activated leg were :neasuredin the sitting and response to the step ccumandgenerating rela- the supine lying positions on specially designed tively low tension, but non-linear characteris- apparatus,to serve as input data forthis tics of the controlledprocess might cause un- model. External stimulation of the Quadriceps stable response for higher tension. Soit should was used for tetanic contraction, with constant could be at- stimulus parameters. Simultaneously, surface EMG be clarified how the stable control activity of the stimulated Quadriceps muscles, tained all over various strengths of the muscle as well as their antagonist Hamstrings muscles, tension. was recorded.Conduction velociy of the Femoral In this study, we identifiedtransfer function and Sciatic nerves was measured before and after of the muscle tension of gastrocnemius. Using each experiment,and was found unaffected by the data, the control system responsesto the fatigue. step command, generating about1/4, 1/2, 3/4of An attempt to correlate EMG and mechanical the maximum tension, were simulatedin order to measurements suggests that the decaying measured applied forceinthe fatigued Quadriceps muscle is discuss how the Kitamori's method can accompanied byacorresponding decrease inEMG to the controlsystem including non-linear amplitudes.The latencies between stimulus process.From the results, it wasfound that artifact and onsetofEMG were 1 and 2 ms, for using the largest gain of the transferfunctions the Quadriceps and Hamstrings muscles, without os- respectively, indicating that stimulation is was the best way to controlstably direct from the electrodes to the muscles. The cillation all over the strengths. PI orRFD con- force induced in the Hamstrings has negligibly troller showed higher robustness againstthe non low compared to that of the Quadriceps. linearity than that I-PD or I-P controllerdid. 405 1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas 385 Biomedical Engineering Scientific Papers

;3E21-A.4 BEn-A.5 An ImplantableiI'rve Stimulation System, Development of the Multichannel FES System Technical Solution !.id Applications,H.G. for the Paralyzed Upper Extremities. Stoehr*,W.Mayr and H.Thom,4, 2nd Surgical N.Hoshimiyeand Y.Handa+, Clinic, Universit Vienna, Austria. Research Institute of Applied Electricity, HokkaidoUniversity,Japan, Functional electrostimulation offers a way +ShinshuUniversity, School of Medicine, for r,emobilization after parapieTi.e and Matsumoto, Japan reinnervation of the n..phrenici in cases of central paralysis of the breath MultichannelFunctional Electrical muscles. Stimulation We developed multichannel im- iFES) is very efficient for the restoration plants (up to 40 electrodes) using thin- of the paralyzed upper extremities. We developed film and thickfilm hybrid technology and several microcomputer-based FES systems, and custom designed integrated circuits is applied them to clinical usage. In this paper, well as sophisticated extracorporal :on- recent progress in the development of the multi- trol units. All stimulation paramet,:rs can chantelFES systems will bereported. (1) A be controlled preceding each stimulation master-slave t:Tehigh performance FESsystem impulse. Approach to the biological system with48 channel outputs for the restoration of is done via stainless steel electrodes the hand function of the paralyzed upper extre- which are sutured to the epineurum of the mities of the quadriplegic patient is described, nerves to be stimulated. In 1984 first which can control the finger, wrist and elbow, clinical applications in 4 paraplegic sub- and the stimulation patterns of whichcan be jects proofed the principle function of modified by the patient through the the method. They could stand up from the interactive capabilities. (2) A wheelchair and walk short distances on portable 16channel FES system which is suitable for flat ground. Meanwhile the "lung clinical usage, which will be commercially available in the near pacemaker" is produced commercially and future. Creation method of the multichannel allows fatigueless stimulation of the stimulation waveforms phrenic nerve 24 hours a day. Several high is identical for both systems, and is based on the EMG activities of tetraplegic patients could be discharged normal subjects. from the ICU and are now living at home for more than two years.

BE21-13.1. BE21-B.2 Quantification of Wound Healing after Electrical Stimulation for Functional and Electrical Stimulation, L.Vodovnik*, Therapeutic Use in the Disabled Patient A.Stefanovska, D.Kriiaj, H.Benko+, Y. Mande* and N. Hoshimiya+, R. Turk+, M. Malelid++, Faculty of Dept. of Anatomy, Shinshv Univ. School of Medicine, Electrical Engineering, +University Matsumoto 390, Japan. +:Res. Inst. Rehabilitation Institute, ++Joief Stefan of Applied Institute, Edvard Kardelj University, Electricity, Hokkaidt, Univ. Sapporo 060, Japan. Ljubljana, Yugoslavia Electrical stimulation by using percutaneously indwelling intramuscular electrodes in order to improve motor function of paralyzed and paretic extremities in The treatment of indolent wounds presentsa the major medical, social and financial problem. spinal cord injury (SCI) and stroke patients is des- During the past decades severalpapers have cribed. Functional electrical stimulation (FES) by reported about accelerated healing of chronic using EMG-based stimulation patterns whichwere obtained wounds using various modalities of electrical from normal subjects easily provided the severely disab- or electromagnetic stimulation. Due to lack of led patient reliable and reproducible control of the quantification, ignorance about the healing paralyzed extremities. Not only locomotive function in mechanisms and inadequate controls, stimula- paraplegics but also upper extremity function in high tion has as yet not been widely accepted in SCI patients such as C4 quadriplegics could be restored clinical practice. In our studywe primarily by this method. Therapeuticelectrical stimulation addressed the quantification problem but also (TES) wasapplied to paretic extremities ofSCI and investigated possible mechanisms and controls. stroke patients for relieving spasticity and restriction of range of motion (ROM) and improving volitional acti- Decubitus ulcers on ten spinal cord injured vities ofhill living (ADL). Since safe, easy and patients were daily stimulated with tetanizing steady stimulationwas currents through PALSTM skin electrodes. achieved by the percutaneous The wounds were photographed weekly and intramuscular electrode, long- :rm and frequent stimula- slides projected on a digitizing tablet. The tion could be realized by patients themselves. They stimulated their paretic extremities over6 times/day areas computed by a CAD program were whereone time was consisted of 30 min of cyclic stimu- plotted as a function of time. It was found lation. that an exponential decay of the woundsur- Such intermittent stimulation method in aday face is a good approximation. The healing provided apparent and sustained reduction of spasticity time constants were found to be about 2.4 becauseof prolongation of the carry over phenomenon weeks for trochanter wounds and 4.7 weeks after electrical stimulation. Therefore, an increase for wounds on the sacrum. in active ROM and improvement volitional APL were observed within a month. Long-term TES over one year acquired permanentimprovement of volitional activitiesof the 4C6 patient.

386 1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas Biomedical Engineering Scientific Papers

11E21-8.3 BE21-B.4 Automatic FES Controlofthe ParalyzedExtremitiesby SuppressionofVolitionaland Activity In Humans by Transcranial Cortical Percutaneous Electrodes. Stimulation, H.R. Dimitrijevic*, D. Y.Handd, A.Naito and N.Hoshimiya+ Moffit, C. Van der Linden, Division of Dept. of Anatomy, Shinshu Univ. School of Medicine, Restorative Neurology and Human Neurobi- ology, Baylor College of Medicine, 7000 Matsumoto 390, Japan Fannin Suite 2140, Houston TX 77030 +Res. Inst. of AppliedElectricity, Hokkaido Univ., Sapporo 060, Japan Transcranial cortical stimulation has been previously shown to elicit motor evoked poten- Restorationof motor function in the paralyzed extremi- tials (MEPs) inintact human subjects. In tiesof spinal cord injury(SCI) patients by FES is addition to the excitatory activity, the stimula- described. Helically-coiled percutaneous electrodes mak. tion can markedly suppress ongoing activity. To demonstrate this, weused two different ex- byTeflon-coated 19 strand stainless steel wires were perimental paradigms in a group of 10 healthy used for intramuscular nerve stimulation. Multichannel subjects. Anodal cortical stimulation was standard stimulation patterns for controllingparalyzed applied 1),while making the fastest possible extremities were obtained from trapezoidal approximation contraction in response to a visual cue, recording surface EMG activity from the thenar (contr- of integrated EMG during motion detected from normal acted) muscle, and 2), while recordingEMG subjects. Input of the threshold and maximum stimulating activity from the leg muscles during postural voltages of each muscle of the patient automatically disturbances of subjects standing on a moveable (rocker). Inboth conditions, we created stimulation data of FES for each patient. This platform demonstrated a silent period. Theseresults method easily realized versatile control ofmulti-joint illustrate that cortical stimulation, in addition system of the extremities. Therefore, not only locomo- to its excitatoryinfluence also hasanin- tiveactivitiesof the paralyzed lower extremities in hibitory effecton motor activity established through volitional or postural motor tasks. paraplegicsbut also upper extremity function ofhigh SCIpatients such as C4 quadriplegics could be restored by our multi-channel FES system.

BE21-C.1 BE21-B.5 for Magnetic Stimulation of Localized Area Apractical walking system employing the LSU Orthosis in the Human Cortex, S.Uenol T. Tashiro, paraplegics electrical stimulation of and K. Harada, Department ofElectronics, and reciprocal Kyushu University, Fukuoka, 812, Japan the thighmuscles. M. Sulomonowl R. Baratta, H. Shoji, R. Douglas, M. Ichie and S. Hwang. Department of Orthopaedics, LSU Medical Center, New Orleans, LA, USA. A method of localized stimulationof the human cortex is proposed. The basic idea is to concen- GaitOrthosis (RGO), vhich trate induced eddy currentslocally in the vicinity The LSU Reciprocal assume independent upright of a target in the cortex by a pairof coils which allows paraplegics to with the aid of only are positioned outside thehead so that time- posture and reciprocal gait varying magnetic fields pass throughthe head in a walker or crutches,was sucessfully fitted and the last 10 the opposite directions around a target. The eddy evaluated in over 112 patients in further currents induced at the target areexpected to flow years. This report describes addition of a four- together, which results in anincreased current- developments centered on the stimulation unit which flow at the target. Spatial distributions of eddy channel electrical contractsthe currents induced in the head arecalculated.The reciprocally and voluntarily current density at the targetmakes a peak which is contralateral quadriceps and hamstring to provide reciprocal gait which higher by 2-3 times than currentdensities at non- effective andcosmetic energyconsumption of the target regions.The validity of the proposed reduces the metabolic as compared to reciprocal method is demonstrated by experimentsusing frog patient by 377 to 52% butwithout the benefit of nerve-muscle preparations.A U-shaped core is used. gait with the RGO be presented to qua.stify The core is driven by discharge ofcapacitor bank stimulation. Data will atrophyof disuse by FES through a thyristor switch. When the nerve is the reversal of muscle positioned in the area between magneticpoles, the over six weeks, the increase in muscle bulk, hip rate andPO2 nerve can be stimulated by acapacitor voltage 400 flexion-extension torque, heart V, an" muscle contraction is observed.In contrast, profiles during the therapy and during reciprocal will be when the nerve bundle is positionedoutside the gait withand without stimulation this magnetic poles, nerves cannot bestimulated even by compared to highlight the benefit of the three times of capacitor voltage1200 V. portable systemas a practical rehabilitation modality. 40 1988 World Congress on Medical Physics and Biomedical Engineering ^ San Antonio, Texas 387 Biomedical Engineering Scientific Papers

BE21-C.2 BE2i-C,3 A Biamechanical Study of two Walking Braces for Development and Assessment of Neural Orthoses the Paralysed, M W Whittle* and R J Jefferson, Paraplegicsfor Oxford Orthopaedic Engineering Centre, A. Cliquet Jr. *, S. E. Solomonidis,R.H.Baxendale, Headington, Oxford 0X3 7LD, England. B. J. Andrews and J. P. Paul Bioengineering Unit, University of Strathclyde Two designs of reciprocal walking brace for and Institute of Physiology,University of Glasgow, paralysed people have recently emerged as Glasgow, G4 ONW, U.K. practical systems - the hip guidance orthosis (HGO) from Oswestry, England, and the Several types of orthoses have been designedto assist reciprocating gait orthosis (RGO) from New the locomotic , of paraplegic subjects. These vary Orleans, USA. from pure mechanical devices that work throughupper A trial has been conducted to compare the two limb mediated posture and gravitational action (e.g. braces. 22 paraplegic subjects in matched pairs hip-knee-ankle-footorthoses),to surfacefunctional were able to try each brace for 4 months, in a neuromuscular stimulation on its own whici., although crossover study. At the end of the trial, the more cosmetic than the former, may lead to muscle subject was able to keep whichever brace he or fatigue and consequent loss of function. A hybrid she preferred. The following assessments were approach can be used in order to obtain thebest of performed at appropriate intervals: clinical, both systems and thereforesome hybrid sytems that psychological, ergonomic, economic and use different control strategies have been developed. biomechanical. Since the energy required to rise fromthe sitting to The biomechanical assessment made use of two the standing position as well as to achieve "aidedgait" types of gait analysis: videotape was used to is provided by the patient's own muscles,analyses of determine the general gait parameters and the upper limb effort and metabolic cost, were carried quality of walking, and the Vicon television/ out. The metabolicassessment used indirect force platform/computer system was used to make calorimetry and the biomechanical evaluationwas done detailed kinetic and kinematic measurements. by incorporating force measuring transducers in The results show that the differences between the the walking aid and recording groundreactionforces. subjects are much greater than the differences Repeated measurements for each subject between the braces. fitted with Nonetheless, differences different types of orthoses allowed directcomparison were observed between the braces, particularly in of the relative "efficiency" of eachsystem. the patterns of movement.

BE21-C.4 BE21-C.5 Motorized Exoskelton orthosis, Y. Sakurai; A portable EMG Training Device for Amputees with H. Miyamoto, K. Tokimura and Y. Shimazaki, Automatic Data Storage. Tokyo Women's Medical College, 8-1 Kawada- U.Boenick*, W. RoOdeutscher and V. Matute, cho, Shinjuku-ku, Tokyo, Japan Technische Universitet Berlin, D low Berlin 10, Germany The handicapped persons with paralyzed lower extremities desire strongly to stand and walk on Best orthopedic management of amputees call be their own feet. In this paper, motorized achieved by myoelectrically controlled exoskelton orthosis system which assists the prostheses, which use the EMG signals of the walking of handicapped persons will be reported. stump musculature as signal sources. The exoskelton frame is fitted to the waist and However, in many amputees the EMG after legs. Hip and knee joint of this frame are powered amputation and longer inactivity periodsis so with electro-hydraulic digital actuator controlled weak that the necessary level can not be with micro-computer. Through potentiometers achieved. The major part of these patientscan mounted. coaxially on each joint, changes of joints increase their signal level aftera systematic angles are continuously detected. A pair of foot training. switch sensor made of pressure-sensitive rubber detect the instant of heel-strike and toe-off. Training devices used for this purposemeasure Inclination or three dimensional posture of the and display the muscle potentials after body is detected by gyroscopes. The communication amplification. A documentation of the time and interaction between man and machine are carried dependent training progress is not possible. To out with crutches provided with contact sensor and overcome this problem a two channel measuring command switch. Patients can walk, stop, turn to and training device was developed which stores left or right and go up and down stairs with this all features and may help the physician to System. decide more precisely whether or not a successful prosthetic management of theamputee is to be expected. Experiences from practical applications are briefly reviewed.

4 c.

388 1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas Biomedical Engineering Scientific Papers

BE21-D.1 BE21-D.2 Adaptation of consumer mIcroremputer-bases technology to create AN INTEGRATED SYSTEM OE* RELATION WITH THE affordable environmental control for the profoundly ::ntor-disabled ENVIRONMENT FOR DISABLED PEOPLE David A Soontaier" and Michael Kleviansky, U.Cape Town L.Gardeazabal, G. Buldain, A.Arruabarrena, Medical School & Croote Schuur Hospital. Observatory 7925. S.Africa E.G6mez, J.Gonzalez*, Informatika Pakultatea The combination of a variety of inputs, conputers, software and out- EHU. 649 pk 20080 DONOSTIA (SPAIN, puts against a background of disabled-users' individual needs and There are a great variety of diseases that capabilities, has created a morass of choices which the average therapist is ill-equipped to handle. This is particularly true of a produce deep problems in the movement (cerebral developing country with a relatively larger number of rural techno- palsy, hemiplegia, etc.). People who suffer from logically under-prepared people. Though there are special schools them see greatly affected their independency and for middle-level disability, eg. Cerebral Palsy, there are few facilities for the profoundly disabled, eg. high-level quadriplegics. It their posibility of relationship with other therefore became the responsibility of hospital-based therapists, people. We have designed an integrated system to and parents of disabled people to take the initiative. solve this lack, improving autonomy and Four years ago, a country-wide self-help organisation, INTERFACE: Computers for the Disabled, was established to act as an inform- communication with people. ation/expertise broker.A close liaison has developed between Inter- This system is divided in three blocks: face and the U.Cape Town Rehabilitation Technology Croup, leading to * An infrared-based environment controller that many collaborative projects. Cocoon consumer products such as TV remote controllers often contain allows to switch on and off electrical devices, embedded microprocessors. Depending on the manufacturer's openness pick-up the telephone, etc. for technical access, these devices can often be modified to accept * A communication tool controlled by a switch alternative inputs for use by disabled people.This is particularly important since so-called luxury" devices for the general population for non-vocal handicaped people. Messages arc become necessitites if disabled wish to do normal tasks, cg. opining built letter by lette, using some edition the door to a visitor, or making a phone call.However, several device' specific hand-held controllers often need to be manipulated functions and sham i.i a LC. display. from the user's customised interface such as joystick, sip-puff, eye- All the messages c...n sert to a printer via position etc. He have developed a "smart" .Microprocessor -based infra- a parallel conexioa red link to offer a wireless, 2 .way communication path between a disabled tat= and the "world" at Lags. For adaptability we chose RS232C H A keyboard emulator QZ a canputer allowing to protocol so that rot only control information but also mwssages, eg. access to a P.C. by a switch. Thus, the user notification that to instructint "open the door" had been executed. can works with most of the programs availables Device drivers, bet.% input- and output- are loaded to EPROM from a master file by the eranisP who is setting up the system to match in the market. The conexion of the keyboard the disabled user's capabilities and needs.To achieve universality emulator with the canputer is via an infrared each commercial device will need a standard plug to enable it to be emitter, avoiding the use of multiple wires. ettatched to the host microcomputer. The prototype system we have built has already given a great deal of This integrated tool is being used by a group of independence to a severely disabled client. people who suffered from cerebral palsy in a local asociation of C.P. ASPACE

BE21:-13.3 BE21-D.4 Using a Microcomputer toRehabi- VIRTUAL SOFTWARE INTERFACING IN litatc Communication Ability of REHABILITATION ENGINEERING, B. Rowley*, Disabled Man, Liuda Chu* and C. Davis, Dept. Biomedical Engr., Institute of Electronics Academia O. Amos, Special Ed. Program, Wright State Sinica, Beijing, China Unive.sity, Dayton, OHio 45435USA

One of the challenges rehr 1itation engineers A special microcomputer system has been face is interfacing disabled persons to computers developed for a severe physically disabled so they can communicate and become productive in man 'ho may move hi:. foot, a joint or head the work place. The number of options are to press two button., of the system to limited requiring compromises between needs and control the computer to achieve communica- current technoiogy. We have developed a virtual tion ability with community. The author interface method (VIM) which overcomes many limi- was involved in this project and developed tations to the use of softw.:e such as word pro- a program to reach the above goal. cessors, spread sheets, data base managers, and !The program isetwo-button controlled programming languages b;' those with severe neuro- screen editor which has following func- motor handicaps. The VIM is a resident memory tions: automatically to access disk file program providing a gateway between augmentative and save file, button controlled text communications software which interfaces the input, text modification, text print, disabled with the computer and comzierical soft- and alarm function available to secure the ware. As a result, the disabled person has out of accident. A simulated auto- access to the same software that able-boalf.d scan screen keyboard on display terminal persons use. guides the user to choose alphabets or When the VIM is invoked, a special window is function symbols to write a letter, a note opened on the monitor. Through this window the or a file to be saved or printed. Thus, user has control over commercial software using for a disabled, this syntemwill be augmentative keyboard techniques. An unlimited only a great tool to commanicatewith number of linked customized Andows are available others. for operation of augmentative software or with This system may be used both in English forty positions for commands and macros. and Chinese characters. 409

1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas 389 Biomedical Engineering Scientific Papers

BE21-D.5 BE2I-E.1 MODIFIED TACTILE VISUAL SUBSTITUTION DEVICE FOR The Use of the Eyes by Severely Disabled REPRESENTATION OF TWO- AND THREE DIMENSIONAL People for Communication and Control PATTERNS, T.Dohis, T.Horluchi, H.Narahara, I.Sakuma, T.Masuzawa, P A GRIFFITHS: E A DYMOND, R POTTER The Univ. of Tokyo, 7-3 Bongo, Tokyo, Japan Many severely disabled peopleare restricted to Topromote out-door activity for theblind and eye movement as their sole means of communication. minimize their overcausion against the subjects in This function has been harnessed in different a far distance, an visual substitutional system waya e.g. signals arising from the corneo-retinal being able to provide agreater variety of potential, the differential reflectionof infra environmental information for the blind is red light or monitoring of theeye by a tele- necessary. For this purpose, three dimensional vision camera. space recognition training system was developed and tested with normal subjects. Work at Lincoln over several years withpeople The system is consisted of a tactile (tablet with having congenital or acquired di.ordera has piezo-electricvibratory pin matrix), two- and allowed eye movement as a control functionto three-dimensional environmental models and a visual be explored in a variety of differentcircum- axis detector using position sensitive device(PSD). stances. This has led to the definition ofthe Theseenvironmental models are programmed in a important design features of eneye movement microprocessor and geographyeally represented on system. the tactile. PSD can detect the orientation of the subject atd represent the shape of model from a These will be described and diezutmedtogether certain angle on the tactile via micro with the approach which has been takento use processlrtvisual tracking method). eye movement in a daily living context so that Fivedifferent shapes of twodimensional figure the user may write letters, control domestic were utilized in the tests with ninenormal equipment, partake in recreation and through subjects. Over80X of discrimination ability was synthetic speech even make telephonecalls. observed on the average when visual tracking method was employed. Preliminary evaluationshowed that this system could be utilized for the environmental recognition training on the congenital blind.

BE21-E.2 Airflow Resistance of a New Voice Prosthesis, Computer Aided Rehabilitationof Aphasia, A. Ben Jebria*, C. Henry, N. Gioux,F. Devars J. Mild.", Technical and L. Traissac, Penn State University, ResearchCentre of Finland, Tampere, Finland, University Park, PA 16802 and Universit6 L. Salonen, de Helsinki University Central Hospital, Bordeaux-II-33076 Bordeaux, France. Helsinki, Fit land. This Performance of a new voice prosthesis, aphasia rehabilitationsystem is used to built upon a rehabilitation restore speech in patients followingtotal lar- method called yngectomy, was evaluated by measuringin-vitro, Language Enrichment Therapy(LET). Based on systems approach, LET ass well as in-vivo airflow resistances. The pros- is a comprehensive method, thesis is made of a relatively softpolyurethane thegoal of which is to restore all forms of material, unlike other devicesreported have been language processin& made of silicone. It has the form of biflanged capabilitiesby includingthem into an tube of 0.35 cm diameter andvariable length that interactive system. Various linguistic can be adjusted to fit the morphological subfunctions aretreated in a hierarch- structure ical manner of the part situated betweenthe trachea and the in reciprocal relations esophagus of each patient. The fixed flange at rather than separately. The the esophageal side of the tubeis provided with computer-aided LET rehabilitation a one-way-valve system thatopens during expir- system has been produced for the IBM AT personal ation (for voice production) andimmediately computer usinggraphics and digitized speech. closes to prevent leakage of liquidand food. Clinical Total device resistanceswere found in the range evaluations started in spring the of 35 to 46 cm H2O/L/S when theflaw rates ranged 1987 using a prototype of the from 0.05 to 0.35 L/S. Therefore, our voice therapy system. Evenseverely aphasic patients prosthesis operates in arange of airflow resis- learn to use the system and most of them find tance comparable to that of intacthuman larynx working with the computer motivating. during vowel production. In addition, the aerodynamic performance of thisnew prosthesis compares favorably with the various devicesreviewed, because it offers the leastopposition to airflow. 410 390 1988 World Congress on Medical Physics and Biomedical Engineering San Antonio, Texas Biomedical Engineering Scientific Papers

BE21-E.4 BE21.-E.5 Design and construction of a valve for endoesophageal Aspecialized electronic game for tubes. G. Michanetzis and Y. Missirlis*, Laboratory restoration of function after hand Z7-11-5Medical Engineering, University of Patras, Pat- surgery, (*)A.Pepino4,4" (**)M.Iocco, ras, Greece. (*)M.Bracale, (**)N.Misasi In certain cases of cancer of the esophagus, where * Univ. of Naples, Electr. .Dept. Via surgery is not admissible, feeding the patient is Claudio 21, Naples (I) done through a permanent endoesophageal tube. Many ** Univ. of Naples Orthoped. Clin. clinical studies have indicated, however, that re- Via Pansini, 5 Naples (I) gurgitation of stomach contents takes place when the patient is inclined, after being fed, leading to Froma general point of view biofeedback choking and subsequent death. is a form of psychophysiological training The purpose of this study was to construct a bio- where sensorial stimuli representing compatible passive valve to be mounted on a Celestin biological functions, are presented to tube accoding to the suggestion of Dr. Lekakos. The the organism which produces these events. operation of the valve would be such as to remain A specialized electronic game can be open while the patient is being fed and to close and considered like a sort of biofeedback rot allow the stomach contents to regurgitate into device to improve motivation and for the esophageal tube when the patient is inclined. monitoring the rehabilitation trend. The final design is that of a hinged disc plate of A video-graphic system, together with a magnetic material (like the "toilet seat type of an special support for the arm, has been artificial heart valve) covered with polyurethane. designed for rehabilitation after hand At the vertical or semi-inclined position (45°-90°) surgery. Aclinical evaluation has been the valve remains open due to gravity, while at the carried out on ten patients for three horizontal position (0°-30°) the disc is attracted months, andwe observed an increased by a steel ring covered with polyurethane and the motivation and a correlation between the valve closes. improvement of finger joint movement, and In-vitro testing was performed to verify the oper- the score obtained in the training. ation of the valve and results will be shown on the valve's function in dogs. This work was financially supported by MPI 40%. and C.N.R. "Nucleo di Ingegneria Medica e della Riabilitazione"

Bir21-E.6 Basic and Clinical Studies on BE22A.1 Electronic Walk Helper,Ying-bing A FAST CORRELATION TECHNIQUE FOR Liu*,Lian-sheng Liu, etc., Surgery MULT1MODAIJ1Y IMAGE MATCHING Institute,Daping, Chongqing,China. Anthony ApicelleJoachim Nagel, Ph D. Advanced nuclear medicine imaging modalities such as single Walking photon emission tomography and positron emission tomography EWH-1 belongs to one of the FES provide valuable functional or physiologic inforn cation. As this Helpers with t!ie following main features: information is complementary to anatomic imaT,s obtained via Nuclear Simple construction, small andfine Magnetic Resonance, the diagnostic potential of the Nuclear Medicine dimension, portibility and cheapness, images may be augmented by objective, accurate matchingof the application to both walkhelping and different data sets. treatment with satisfactory efficiency. Different techniques, such as correlation analysis, edge dete:tion, Experiments on animals and healthy people or fiducial landmarks have been proposed for image matching."i;;ey all demonstrated that 0.2ms stimulating pulse suffer from severe shortcomings with regard to either reliability and precision or mathematical complexity resulting in unreasonable at 30-40Hz and 20-80V could attain an calculation times. optimal, effect for the contraction of the and no discomfort to We developed a new correlation analysis to compute the three paralyzed muscles, dimensional translational and rotational shifts of the image planes the clinical-patients. necessary to combine or overlay the different images. Current Clinical analyses on more than 4000 cases correlation analysis approaches are limited, as the coupling of the indicated that EWH-1 could achieve the registration variables requires an iterative, computationally expensive pedal dorsal flex just in those who could algorithm. The approach described here improves upon past work by dorsally flex their ankles regardless of first decoupling the translational and rotational components, thus the course, but had no curing effects in eliminating the iterative part of the algorithm and significantly reducing the tetanic hoofbound and severe spasm. the computational expense. Earlier applications could more favor the The input Images are first transformed into images which are translation invariant. These translation invariant images are converted functional restoration of the paralyzed EWH-1 could be also used in the to polar coordinates and the rotation angles calculated. The rotational muscles. corrections are applied to the original images, and then the translation walk aiding of the incomplete paraplegia, distances are calculated. Given the matching differences between the cerebral palsy and sequelae of the two data sets, the NMR images are transformed so that they are aligned myelytis. Out of the clinical limits for with the PET images. EWH-1, the parameters and multichannels for motor dysfunctions at multiple joints over the knee are undertaken in our lab., so as to meet the critical needs of much more paralytics. 411 1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas 391 Biomedical Engineering Scientific Papers

BE22A. 2 BE22A.3 The Characterization and Quantization of Soft Tissue in MR Images Enhancement of Radiographic Images by MTF and Fractal by a Multiple Pixel Intensity, Region Growing Algorithm Dimension Evaluation, F. Arduini, S. Dellepiane, C. Regazzoni, F. Sturaro, G. Vernazza: DIBE University of Jacob M Agit* and Rui J.P. de Figueiredo, Fellow IEEE Genoa, Via All'Opera Pia 11/A Genova Italy. Dept of Electrical and Computer Engineering, Rice University In medical images human operator focuses his attention on P.O. Box 1892; Houston, Texas 77251-1892 tel. (713)527-4020 some particular regions or structures of interest: *Dept of Radiology, Baylor College of Medicine purpose of this paper is to describe a newmethod suggested for local enhancement of textured areas and to The., characterization and quantization of soft tissue in Magnetic present the related preliminary results. After the computation of the MTF (Modulation Transfer Function) of Resonance [MR] imaging is complicated by the nonspecific nature of the whole acquisition system by a bar pattern, a 'global the imaging modality. It is difficult to determine the composition of restoration' of the image is performed by processing the original image with an appropriate linear shift invariant the tissue volume corresponding to a particular pixel based solely on T1 filter (complementary to MTF curve) keeping attention to and T2 ratios since there are significant partial voluming effects and reduce noise effect at high frequencies. Then, in order since T1 and T2 ratios may not uniquely characterize a material that is a to detect areas of specific texture properties, the fractal dimension D has been calculated by a *fractal chemical mixture. Additional information such as the spatial distance preserving method', in order to preserve thin anatomical in the image plane or T1 and T2 spectra must be considered. Region structures. In fact according to this approach, some prefixedmasks are positioned around each pixel and D is growing algorithms such as those presented by Gonzalez and Wintz take calculated by averaging onthe most homogeneously into account both the image pixel intensity as well as spatial texturedmask (i.e. minimum variance). The fractal dimension D has been evaluated by the 'blanket approach'. information. The error function that assigns a pixel to the region is Using the 'fractal image' it is possible to localize the based on a single intensity value for each pixel and is defined as presence of specific textured areas and to select appropriately local filters, i.e. with specific E = II(XioY0)-1(XiY)1 frequeciesbehaviour, to be applied to the restored image. In fact, we recall a presence of a direct relation where I(X0Yo) 8e intensity at pixel x,y and c is error function. between D values and spatial frequencies behaviour, in particular low D (e.g. around 2.0) denote quite spatial However, there are two intensities for each pixel corresponding to the homogeneous region. Consequentlyfrequencies filters T1 and T2 ratios. poles have been calculated by taking intoaccount D We have proposed an error function defined as the vector sum of the values. This approach has been verified on some chest radiographic images and the obtained results arequite intensity differences in each image plane. The intensity I(xo,y0) used promising. to characterize the region is classically defined as the intensity of a connected pixel known to be in the region. However this can lead to propagation of a region throughout the image planes. We used an intensity value that is a function of all the intensities known to be in the region of interest. Implementation of an algorithm that uses both spatial as well as the above mentioned, multiple intensity criterion has shown to be a robust method in MR.

BE22A.4 BE22A.5 Digital Subtraction Angiography on a Constanr False Alum date Focus Detection by 2-D Noncausal and tionstationary 01112 Modeling,Eau Ping-Tai and Be Zhen-Ya, Digital Signal Processing Research Personal Computer, M.Cesarelli., M.Bracale, Group, Radio Department, Nanjing Institute of Technology, Nanjing, P.R. China. Cattedra di Elettronica Biomedica-Nucleo A I-ray photograph or tosogras can be viewed as a sample of a two.diiensional 12-Di di Ricerca di Ingegneria Medica e della random field represented by a white-noise-driven representation (10iDR). Previously, Riabilitazione C.N.R., Faculty of Engi- onlycausaland stationary1INDRs were fully used for the lackofaefficient neering-University of Naples, Via Claudio parameter estimation technique to estimate the parameters of the generalnoncausal MR whichrepresents the gray level at one point as a linear combination ofits 21, 80123 Naples, Italy. neighbors in all directions and a white noise, and only unsatisfactory results yea observed.In this paper, we study constant false alarm rate ICFARI focus detection The acquisition and the elaboration of images are by means of 2-D noncausal 1111DR modeling based on an asymptotically consistentUS assuming an important role in the evolution of the computationallyefficient edified least-squares (LS) estimation methodrecently medical diagnosis. An image elaboration system developed by the authors, for identifying parameters of such a noncausal model (cf. needs to store a great number of information in a Preprints, 8th IFACsymposiumonI.S.P.B., toappear). Furthermore,the nonstationarityis alsoassumed in the modeling as the image statisticsofthe Short time. focusesandthe normal parts of organs are different,Becausethe LS codified Up to now, this job is performed by a minicomputer- sethodisforstationary case,we firstdevelop,inthispaper, anadaptive size system. The technological progress of the fast companion ofthe method by carefully exploiting the specialstructuresofthe Dorsal equation set of that sethod,The adaptive tethod using only local pixels in electronic and the improvement of the operative sys- a sliding data window is a finite memory algoritis and is spatially recursivein tem allow the use of microcomputers for this aim. nature, and it estimatesthe parameters of the noncausal and nonstationary 1811)8 consistently and computationallyefficiently. Boththeoreticalanalyses and we are developing a DSA (Digital Subtraction Angio- cosputer situlations show that the estimates produced by this method are about10 graphy) system based on personal computer PC AT tisesas accurate as those produced by conventional LS, whilethecosputational with an image board subsystem. Digitized 512x512x8 coaplexity of the forcer is only the square root of that of the latter. Developing bit images from an external sow:ce are stored on a probability density function (PDF) for the photograph based on the noncausaland nonstationary VND2 sodeling, sating significance testing under the CFAR constraint, main memory, at 1image/sec. rate and a subtraction weconstructeda CFAR focus detector by programing PC VI81011isage processing is obtained in about 30 seconds. systeswith10112A1171 and 8088 usably language.The CFAR detector caneasily detect calculusesinurinary organs, pulsonary diseases and other diseases fro the Considering the price and the performance, the sys- corresponding I-ray photographs, while detectors based on other nodding techniques tem is useful in the routine work of the neuroand cannot,Thedetailsaboutthe structure of the CFAR detectorand ofsose angioradiology lab. erperiaents are given. 412 392 1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas Biomedical Engineering Scientific Papers

BE22-B.1 BE22-B.2 3D RECONSTRUCTION OF A WHOLE HEART Three-Dimensional Processing of Microscopic USING MRI DATA, M. KUWAHARA*, Osaka Images, A. Takahashi* and I. Nemoto, Tokyo Sangyo University, 3-1-1, Nakagaito, Denki University, Hatoyama, Saitama, Japan Daito-shi, Osaka 574, Japan, and S. Eiho, Automation Research Laboratory, In the conventional light microscopy, images of out-of Kyoto University focus regions of the object blur the image of the in-focus plane. We have been investigating methods to get of higher resolution in the optical axis of the microscope The gated MRI method gives us several sets by processing images of an object obtained with various cross sectional images ontraasverse,coronal distances between the objective lens and the object. We and sagittal planes of a ,heart in a cardiac have employed the assumption of linearity for image cycle. In this paper, we propose a method of formation and used the parameters of the lens to reconstructing 3D shapes of each part of the estimate the transfer function of the optical system. heart, i.e., left ventricle, left atrium, Firstly, under the assumption of incoherent illumi- right ventricle, right atrium, aorta and nation, the inverse problem was solved by essentially pulmonary artery, in a voxel space using applying the inverse of the image formation matrix to three pairs of the above cross sectional the observed images of a cell containing magnetic images taken at the same cardiac phase: two particles. The results show considerable improvement transverse, two coronal and twosagittal in three-dimensional resolution. The other method images. We also can observe the 3D heart used in this study, which is based on the recently composed putting together these six parts proposed model of three-dimensional image formation, superimp6sing on the original cross sectional does not require the assumption of incoherent illumi- images. nation. We have been investigating the possibility of utilizing this method to improve the three - dimensional quality of phase-contrast micro-scopy. We have shown that this three-dimensional theory can be derived from the wellknown two-dimensional image formation. Based on the theoretical work, we are studying the image formation and recovery in the phase-contrast microscopy by simulation and by using a simple sample.

nr92-11.3 BE22-B.4 Processing of "in-vivo light microscopy" images Thermal Image Enhancement by His- to model arterial thrombosis. togram Specification, Dong-Ho Lee* and J. A. Pearce, The University of Texas at Austin, Dept. of Elec- Marc Nyssen; Erik Blocked, Rene Bourgain trical and Computer Engineering, Medical Informatics Dept. Vrije Universiteit Brussel Austin, Texas 78712 B-I090 Jute, Belgium To make reproducible and accurate measurements of the forma- Thermal Imagery is typically of low cont- tion and evolution of arterial thrombi during in-vivo experiments rast since thermal gradients encountered with small laboratory animals, a measurement and processing in biological studies are quitesubtle. chain has been set up. Light transmission is used to observe the It is difficult to find targets in a ther- small (.3 mm) arteries, without disturbing the processes mal image. Contrast enhancement by histo- involved. The transmission microscope images, captured by a gram modification has been studied. Histogram equalization is commonly used chalnicon camera, are digitized in real-time by a special proces- as a method for enhancing images. However, sor, connected to the local UNIX network. the image resulting from this transforma- Unfortunately, blood is a highly scattering medium for light in tion is frequently so distorted that land- the visible wavelengths used (700nm ), so that the interpreta- mark identification is not made easier. tion of the raw images is not straightforward. We have investigated more general transformations to obtain specified output his- A model for the light transmission through flowing blood was tograms. realized and tested. This model takes into account multiple One-to-onehistogrammapping schemes were scattering and the geometry of allfactors involved, except compared including asingle ramp and dual second order effects, due to flow disturbance by the presence of ramps with differingslopes and intercepts. the thrombus itself. Examplesare shown which illustrate what The impressive amount of numeric processing, necessary to resultsare obtained from such transforma- apply the model, is handled by the network of UNIX workstations. tions, which also provide all utilities to visualize the results in different formats: as functional images, as graphs or in numeric formats.

1 3 393 1988World Congress on Medical Physics and Biomemcar EngineeringSan Antonio, Texas Biomedical Engineering Scientific Papers

BE22-B.5 Computer-Assisted Delineation of Vascular BE22-B.6 Patterns in Thcrmographic Images Microwave Active Imaging forBiomedical Applications, G.GABORIAUO,J.C. ROSENWALD, J:Ch. BOLOMEY*,P. BERTHAUD**. E. K.Y. Chan* and J. A. Pearce Institut Curie, 75u05Paris, Biomedical Engineering Program *CNRS-ESE,Plateau du Moulon,91190 The University of Texas at Austin GIF-sur-Yvette **SATIMO, Plateaudu Moulon,91190 GIF/sur/Yvette Abnormal vascular patterns revealed by thermography may be indicative of an underlying pathology.The goal of A planar microwavecamera operating at 2.45 this study was to identify and delineate vascular patterns GHz has been realized for in thermograms of human limbs by using various edge rapid imaging of biological structures.Images detecticn strategies. This would aid in the non-invasive can be formed by analysing the transmittedor the reflected diagnosis of occlusive diseases such as vasoconstrictive wave. With multiview proceedings, Raynaud's Disease. tomographic The same methodology may be used in reconstruction can be achieved. the delineation of breast vascular patterns for cancer Microwave images prescreening. In this study, human forearms were are related to the dielectric properties of tissues and,hence,to any physi- precooled by the application of ice.This created a larger cal or physiological temperature gradient between the skin surface and the factor whose these properties depend(temperature,water content, regions ofthe skinlocated over major vessels. blood flow,...) Precooling was followed by a period of ten minutes for Priliminary devotedto non-invasive thermal thermal equilibration before data acquisition. The initial control during hyperthermia treatments,the use thermal images were prefiltered with a 9 X 9 square of the camera is medianfilteringscheme, soasto now extended to situations improvethe exhibiting higher contrasts: performance of both thresholding and zero-crossing edge water lung content monitoring,oedema diagnostic ofexcessive detectors. Various edge detection templates were then irradiated patients, convolved with the filtered images. Results indicated a or organs defreezing control,constitutesome representative cases trade-off between noise reduction and the retention of whose preliminary results edge information.The boundary detection method by Frei will be presented. and Chen gave the most reproducible results.

BE23-A.1 Acupuncture Energies:ASomewhat Medical BE23-A.2 Model; Joseph M. Helms; M.D., American Acupuncture Stimulation ofEndogenous Opioids Academy of Medical Acupuncture. and the Effects on the ImmuneSystem, C.C. Moss, M.D., La Jolla CA. The presentation of the Classical Acupuncture Conceptualization of Energy Circulation Networks Acupuncture research in the Westhas focused and then to develop a gross anatomical model of primarily on pain therapy. The Chinese literature electroionic circulation consistent with the has suggested far reachingeffects on the immune classical information. It will illustrate how the system using acupuncture. In a review of the different levels of the acupuncture networks are English literature, two studiesexist that show manipulatedwith needles to treat various enhancement of cehl mediatedimmunity with disorders, and how one can understand the acupuncture therapy. A large number of studies therapeutic input in light of the anatomical have shown that acupuncturestimulates endogenous model. opioid production in theCSF and blood. An The hybrid paradigm allows an explanation of equally impressive body ofresearch shows that the the impact of the needle-as-electrode on the opioids have far reachingin-vivo and in-vitro electrical and electroionic system of the body, effects on the immune system. and an academically comfortable organization of This revies is to suggestthat the effects of acupuncture theoretical and clinical information. acupuncture on the immunesystem could be of great It will serve well as the classical reference for clinical benefit through themanipulation of the the more neurophysiological presentations. endorphin system and thatfurther resarch in this area should be undertaken.

394 1988World Congress on Medical Physics and BiomedicalEngineeringSan Antonio, Texas Biomedical Engineering Scientific Papers

BE23-A.3 ELEMO-PZUPONIZICURE (EA) ANALGESIA: EVIDENCE FCIR EMMEN= OPIOID FACIIMAIORY AND INHIBITURY EFFECTSD.J. Mayer' Dept. ofPhysiology, Medical College of Virginia Richmond, VA 23298.

Mlles been shownto increasepain thresholds in many species including man. Our laboratory has been involved in investigating the role of endogenous opiates in this form of analgesia in man and rats. In man,we have chaLleiged the analgesia produced by EA with the narcoticantagonist, naloxone. Wefound that naloxone almost completelyantagonizes the increased pain threshold. LawfrequencyEA analgesia in rats has been thought to be mediatedbyendogenous opioids ' -cause naloxoneantagonized2 Hz EAanalgesia partially reversed 2-15 Hz EAanalgesia(1 an.. 20 mg/kg respectively,Han et al. Acta Phvs_al. Sinica 38(5):475 1986). We conducted experiments with members of Han'slaboratoryin the U.S. and in their laboratory in China using the exact methods of Han et al. In no experi- vent was opiateantagonistreversal or reduction ofEA observed. Rather, we found Naloxone potentiatedanalgesia in most experiments. Additional studies have elucidated the neural circuitryinvolved in this opiateanti-analgesia system. These results will be mowed to the neural circuitry involved in other forms of analgesia. Supported by H.H.S. award DA 06576

BE23-B.2 BE23-B.1 Quantification of Passive Acupuncture Points Acupuncture: A Neuromodulation Technique for the Prognosis in Pain Management, H. C. For Pain Control, Lorenz K.Y. Ng: Dung: Dept. of Cellular & Structural Biology, Washington DC University of Texas Health Science Center at San Antonio, San Antonio, Texas. Multiple analgesic systems, opioid and non- opioid, are known to exist within the brain and A major *pediment in pain management is the spinal cord. The fact that these systems can be availability of a reliable and objective method activated by sensory afferent stimulation is the for an accurate and convenient measurement of basis for various neuromodulation techniques of pain in our body. Existence of passive pain control. Acupuncture merits consideration acupuncture points has been used as a tool for a as a form of neuromodulation since its effects routine quantification of pain.Apassive are produced by afferent sensory stimulation. acupuncture point (PAP) is clinically defined as a Acupuncture analgesia results from interactions tender area which can be detected relatively easy at different levels of the CNS between nerve on the skin surface by finger-tip pressuref less impulses arising at the origin of pain and than two kilograms. The appearance of PAP is afferent impulses from the stimulation site. known to follow a fairly predictable sequence. Evidence will be reviewed to show that acupunct- Thus, it is possible to mathematically quantitate ure stimulation activates certain somatosensory the numbers of PAP in the entire body surface. pathways, including central neurohumoral and The number of PAP is then used linearly to express endorphinergic systems which play a role in the degree of pain. Ahigher degree to pain due modulating nociceptive impulses in the CNS. to more numbers of PAP denotes a bigger quantity Neuromodulation techniques play an increasing of pain, and vice versa. There are various role in the tr.:ailment of chronic pain. As more factors that determine the number of PAP in a knowledge is gained about peripheral control of patient suffering from pain. Two of the most the central nervous processes, it is likely that important factors are the type of diseases and acupuncture will gain greater acceptance and duration of injuries.In general, the importance in treating pain of different manageability of a pain condition decreases as the etiologies. degree of pain increases. Such clinical manifestation provides a useful criterion to predict the outcome in using acupuncture as a therapeutic modality in managing pain conditions. 41 1989 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas 395 Biomedical Engineering Scientific Papers

BE23-B.3 BE23-B.4 The Effectiveness of Acupuncture in Managing Clinical Observations of Using "Lo -Tzu" in Pain. Joseph M. Helms; American Academy of the Treatment of Chronic Tension Headache. Medical Acupuncture. Jong-Rern Chen*, Hamm-Ming Sheu**, Chung-Chieng Wu * * *, and Jia-Show Liu*. The effectiveness of acupuncture in managingthe Dept. of Internal Medicine*, Dermatology**, pain of primary dysmenorrhea was investigated ina Nuclear Medicine***, and Neurology#, randomized and controlled prospective clinical Kaohsiung Medical College, Kaohsiung, Taiwan, study. Forty-three women were followed forone year in one of four groups; the Real Acupuncture Importance of a classic method of acupunc- group was given appropriate acupuncture and the ture, "Lo-Tzu" (intermeridian puncture), for the Placebo Acupuncture group was gives; random point treatment of chronic tension headache has received acupuncture on a weekly basis for three menstrual more attention recently. Patients suffered from cycles; the Standard Control group was followed such headache have tendency to display so- called without medical or acupuncture intervention; the "Shuieh Lo" (blood intermeridian). "Lo-Tzu" at Visitation Control group had monthly "Shuieh Lo is found to produce substantial benefit nonacupuncture visits with the project physician in relieving the headache. Skin biopsy showed for three cycles. In the Real Acupuncture group, that "Shuieh -Lo" used for "Lo-Tzu" turns out to be 10 or 11 (90.9%) women showed improvement; in the an abnormally enlarged and dilated venule. Veno- Placebo Acupuncture group. 4 of 11 (36.4%); in the scintgraphy of lower limbs with "Shuieh Lo" all Standard Control group, 2 of 11 (18.2%); and in show functional stenosis or occlusion of deep the Visitation Control group 1 of 10 (10%). There vein. Chronic tension headache may be considered was a 41% reduction of analgesic medication used as a predisposed factor for the formation of by the women in the Real Acupuncture group after "Shuieh Lo" in the lower limb. It is speculated their treatment series, and no change or increased that the formation of "Shuieh Lo" increase muscle use of medication seen in the other groups. spasm in the neck and head regions, and the muscle spasm, in turn, creates more frequent severe headache. It is believed that the autonomic nerve system plays a role in mediating such a mechanism of interaction reflex. By puncturing the "Shuieh Lo" as in the procedure of "Lo Tzu", the viscious cycle of this interaction reflex may be interrup- ted, and thus the headache suffering will be improved.

BE23-B.5 BE23-B.6 Stress of Electric Stimulation in Veterinary Laser Acupuncture for the Treatment of Acupuncture. Y-C Hwang; School of Veterinary Chronic Back Pain in Horses.B.B. Martin*, A.M. Medicine Tuskegee University, Tuskegee, Al Klide University of Pennsylvania, School of 36088, USA. Veterinary Medicine, Philadelphia, PA 19104

Veterinary electroacupuncture (EA may be Fifteen horses with chronic back pain from 4 stressful to the animals, because of its manual months to 48 months, who could not perform atan restraint, needling, and electric stimulation acceptable level, who did not obtain lasting which are well known stressors. To evaluate improvement from other treatments were treated by stress of EA in the animals, the following 2 using a low-powered infrared laser (300 studies were conducted.Plasma cortisol microwatts) (904 114). The laser was used to concentrations (PCC) of 24 Nubian milkinggoats stimulate 9 acupuncture points.Treatments were were measured before, during and after EA, sham performed weekly. The treatments were performed EA, or restraint alone.Mean PCC during and just for 2 minutes per point.The pulse frequency used after termination of the treatmentwas was 360 pulses per second.After completion of a significantly higher in the EA treated goats mean of 11.9 treatments, performance was improved (53.6+22.0; 31.4+18.1 ng /ml) than in thegoats in 11 of 14 horses, 4 of 11 won races, 1was lost with restraint only (32.6+10.5; 17.4+9.0 ng/ml). to follow -up and there was no improvement in 3 Another group of 8 goats were treated with horses. identical EA.Their cardiovascular and nociceptive responses were compared before and during EA.The results showed very little changes in the cardiovascular parameters. Although signs of stress were noticed during handling andat the beginning of electric stimulation.Animals were characterized by drowsiness during EA, andtheir nociceptive threshold were significantly elevated. The process of EA, therefore,may not be excessive stressor to the animals. The highly elevated PCC of EA treated goats might be due to concomitant release of ACTHand beta- endorphin from the hypophysis. 4

396 1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas Biomedical Engineering Scientific Papers

BE23-B.7 BE23-C.1 Effect of Low and High Frequency TENS on Electrical and Acupunctural Stimu- Provoked Hypertension, P.J. Pontinen* and lation Therapies for Pain Control: M.A.K.Mattila, Acupuncture Research Past Achievements and New Direc- Project, Kuopio University, SF-70211 Kuopio, tions, C. Richard Chapman*, Depart- Finland ment of Anesthesiology, University of Washington, Seattle, Washington The analgetic effects of both low and high frequency TENS have been well documented.The circulatory effects of TENS have remained in a Transcutaneous electrical stimulation and of some acupuncture are counter-irritational more controversial stage inspite which are well-conducted studies showing positive responses therapiesfor pain control in peripheral circulatory disturbances as well as thought to involve central nervous system The precisenature of these in coronary artery disease.We have studied the mechanisms. effects is a hotly debated issue. It is circulatory effects of TENS during provoked stimula- hypertension in normal healthy adults and in clear, however, thatphysical Hypertension was induced by tion, like pharmacological stimulation, potent hypertensives. as prophylaxis against pain a simple hand dynamometer(Vigorimeter).The can be used subjects compressed a rubber ball 4 minutes with in certainmedical settings, to modulate Systolic and acute clinical pain, and to help control 1/3 of their maximal crasping force. will diastolic pressures and pulse rate were taken at chronic pain. This presentation opposite hand. TENS criticallyreview the state of knowledge one minute intervals from the possible mecha- (Cefar) was then given for 30 minutes either with in this area, consider nisms of effect, and highlightrecent 2 Hz or 100 Hz frequencies, electrodes placed on Specificexamples involving both sides of upper thoracic spine and the findings. above. human laboratoryresearchwill be pro- compression of the rubber ball repeated as in research strategy in The results after 2 and 100 Hz TENS were compared vided. Problems research will be reviewed and with the control values using Student t-test. past suggested priorities forfuture research will be offered.

BE23-C.2 BE23-C.3 Analgesia from acupuncture in rats and a new Rationaliseof Transcutaneous Electrical TENS device in humans. B.Pomeranz: Depts Stimulation for Pain Control, D. EL Long*, Zoology & Physiology, University ofToronto, Baltimore, Maryland, USA Toronto, M5S1A1, Canada. Transcutanems electrical stimulation (TENS) in its We studied spina) cord endorphins in modern form was introduced to medicine in 1970. efficacy in electroacupunture (EA) analgesia in rats, using Broadly based Phase II trialssuggested low frequency/high intensity EA. In addition we pain control beyond the placebo response. developed a transcutaneous electrical nerve Controlled studies identified subgroups for which stimulation (TENS) device to deliver acupuncture the effect was greater, acute musculoskeletal pain, like TENS (AL-TENS) using similar parameters (low pain of peripheral nerve injury origin, and post- frequency/high intensity), with antihabituation operative pain were all shown to be benefited in a high percentage. TENS is now standard therapy for features. We tested this AL-TENS device against conventional EA in patients with chronic pain. many kinds of pain. Rats were anesthetized with continuous infusion of Rational use requires 1) a known diagnosis of a pentobarbital. Low frequency (4 Hz) high problem likely to respond, a complaint of pain intensity (25x threshold) EA was applied for 15 means many things and is not an indication for TENS min via needles to LI 4 in order to suppress tad alone; 2) a committed therapist willing to educate flick reflexes elicited by radiant heat. the patient; 3) a patient with the capacity and the Naltrexone (endorphin antagonist) applied to interest in spending the time to learn the use of spinal cord via an intrathecal cannula, prevented the device; 4) a competently made reliable device In the human study, patients with EA analgesia. with all the necessary software and 5) adequate chronic pain (N =171) were randomly assigned to one support to maintain the use of the device of 2 groups: one receiving AL-TENS treatments long-term. twice a week for 6 wks, the other receiving EA In the short term AL-TENS was as (with needles). Mechanisms ofaction are still unknown. good as EA, but in long term followups AL-TENS Possibilitiesinclude blockade of pain through outperformed EA, AL-TENS produces superior activation ofnon-pain carrying sensory sysbams, analgesia because it delivers low frequency/high activation ofdescendingpainsuppression systems, intensity codes and has antihabituation features production ofendogenous pain blocking neuro- which overcomes suppression of monotonous ,transmitters,and peripheral blockade of injured stimulation. 41yrervf: Noneare proven.

1988 World Congress on Medical Physics and Biomedical Engineering 0 San Antonio, Texas 397 Biomedical Engineering Scientific Papers

BE23-C.4 BE23-D.1 Parameter selection in transcutaneous electrical A Review of Transcranial Electrotherapy Stimula- nerve stimulation - experimental and clinical re- tion Techniques, Mechanisms and Application, R. sults. Bengt H. Sjolund: The Pain Clinic, Depart- Smith; Private Research and FDA Consultant, Wash- ment of Anaesthesiology, University of Lund, ton, D.C. lialmb, Sweden.

With transcutaneous electrical nerve stimulation, The two major types of CES stimulators will be many different afferent and efferent fibre systems discussed, as will the two major modes of entrance may be activated transcutaneously. In theory, any into the brain of the electric current, currently biological mechanism that is influenced by afferent postulated by manufacturers and clinicians. Find- and/or efferent nervous activity may then be acti- ings from the literature will be cited. vatld or inhibited. This is illustrated by recent The neuronal mechanisms via which CES exerts its reports not only of pain treatment but of the enfluence will be discussed, along with clinical successful treatment of ulcus cruris with this findings which support current neurohormonal technique. theories of action. The routs of application of CES currents will also be given for the various To optimize the pain control from TENS we have devices currently being utilized or researched. systematically studied which stimulation para- A short history of the use of CES in the world and meters give the most pronounced suppression of C- in America will be described, plus current U.S. fibre mediated nociceptive transmission inexperi- Government position on its use, and certain clin- mental animals and of clinical pain in man. Our ical constraints that have limited its popularity previous results are supported in that several among members of the medical profession in the mechanisms seem to operate, each activated by spe- U.S. cifif afferents at specific stimulation frequen- ces.

BE23-8.2 BE23-D.3 Effects of Transcranial Electrical TRANSCRANIAL ELECTROSTIMULATION (TCET) REDUCES PAIN Stimulation Upon Neurochemicals. SENSITIVITY AND OPIATE ABSTINENCE SIGNS. C. Norman Shealy; Roger K. Cady, J.R. Lake*, D. Brubaker, J.B. Murray, G.P. Crucian, Robert G. Wilkie, and Richard Cox, R.E. Cook, and D.H. Malin. University of Houston - Shealy Institute, 3525 S. National Clear Lake, Houston, TX 77058, USA. Springfield, MO 65807. These experiments studied the effects of extremely small pulsed currents applied bilaterally to lowim- Using a specific type of electrical pedance points on the ear, through gold-plated stain stimulator, of which the output is less steel electrodes in series with 200 k ohm. The approximately 49% below 1,000 Hz and 51% optimal amplitude for inducing analgesia was 10 pA, is 10,000 to 20,000 Hz,we have measured far below the flinch/startle threshold of 25-30 A. a variety of neurochemicals. Significant Sixty rats received 30 min. of either sham stimu- abnormalities have been found in chronic lation or 10 Hz, 10 pA, 2 msec biphasic charge-bal- pain patients in serotonin, norepi- anced rectangular pulses. Rats were injected s.c. nephrine, beta endorphins, and cholines- with 3 mg/kg naloxone or with saline and were tested terase. Following transcranial elec- for pain sensitivity (50 C wet tail flick) trical neurostimulation done daily for40 before and after stimulation. TCET significantlyin- minutes over a 2 week period, these creased tail flick latency, p<.01, in saline- neurochemicals often are normalized. The injected rats, but not in naloxone-injected rats. data in over 100 patients will be Since TCET appears to have endogenous opioid ef- presented. Relationship between chemi- fects, it might modulate opiate abstinence syndrome. cals, that is norepinephrine/beta Rats were infused s.c. for 7 days with 0.9 mg/kg/hr endorphin,etc. may bemore important morphine sulfate via Alzet osmotic minipump. A day than absolute values. after pump r,moval, rats received either TCET or sham treatment and were then observed "blind" for 15 min. for abstinence signs. TCET-treated rats showed significantly fewer signs than sham controls, p<.005. However, TCET had no effect in rats receiving 3 mg/kg naloxone. Further experiments determined optimal pulse width and optimal stimulus pauses for relieving abstinence. TCET also signfi- 4[8 cantly reduced the milder abstinence syndrome following 0.5 mg/kg/hr morphine infusion.

1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas Biomedical Engineering Scientific Papers

BE23-D.4' BE23-D.5 Psychotherapeutic Effect of Clinical Tests of Electrostimulation In Electrostimulation to the Cutaneous Branches Relief of Chronic Pain and Smoking of the 10th Cranial Nerve: Y.T. Kusumil M.D. Cessation, M. Skolnick M. Blacker, L. Raleigh, North Carolina, U.S.A. Skolnick, J. Langone, S. Sands and D. Smith The author will present the psychotherapeutic effect of microelectric stimulation (MES) of the cutaneous branches of the 10th cranial nerve. Transcranial Electrostimulation Therapy Compared to the traditional electroconvulsive (TCET) has been clinically investigated therapy (ECT), MES is unccmparably safe and does to determine if the stimulus' observed effect not produce negative CNS effects, such as amnesia, on neurotransmitters assisted subjects' or other physical complications, such as bone nicotine detoxification and withdrawal and fractures or cardiac arrhythmias. The author will provided analgesic relief of chronic pain. discuss the indications of MES, the technical Both studies utilized double blind designs aspects of MES, and other issues associated with and objective end points. Smoking cessation this therapy. was defined by decreases :In znbjects' reported cigarette usage, craving, withdrawal symptoms and salivary cotinine (a nicotine metabolite). Analgesic relief of chronic pain was indicated by the increase of pre- to post-treatment subject muskulo-skeletal performance measured by a computerized iso-kinetic exercise device (Kin-Com).The ability to conduct double blind trials stems from refinement of the electrical parameters of the TCET waveform to induce significant effects even though current and voltage are below the sensory level. These study results are consistent with clinical EEG findings using the same electrical parameters to induce relaxation responses in subjects.

BE23-E.1 BE23-D.6 TRANSCRANIAL ELECTROSTIMULATION PROMOTES A Legislative And Medical Applicability DROWSINESS IN HUMANS. S. Sands, F.C. Schweitzerr Review of Acupuncture, Charles R. Neblett, M. Skolnick; Univ. of Texas, Grad. Sch. of Biomed. M.D.*, 6560 Fannin, Houston, Texas. Scie., Houston, Texas 77030. During the course of a large clinical Acupuncture -- how is it viewed by "organized trial of low-current transcranial electrostimulation treatment (TCET) in humans, it has medicine"?Medical doctors possess vastly vary- been noted that treated subjects appear to be more ing views concerning its applicability for pa- likely to sleep than their untreated sham counter- tient care. When biases are placed aside, the medical community must ask whether acupuncture parts. Consequently, we recorded EEG topographic maps from 17 subjects (14 males and 3females, can fairly be critiqued and judged by the usual mean age 32) during the administration of3 levels strict standards of scientific investigation. Is of TCET (0, 20, and 30uA) while they rested there sufficient data to justify its utilization? comfortably with their eyes closed. Each subject Should a different standard be applied to assess the potential of acupuncture? received 44 minutes of TECT applied transcranially through small stainless steel electrodes applied Organized medicine has estaboished its position to the earlobes. EEG segments were obtained during a 3-second pause between continuous trains in the legislative arena. What acupuncturists of positive 2 msec biphasic charge-balanced rect- desire and what medicine recommends is much less variable. angular pulses delivered at 10 Hz. EEG spectra These views will be reviewed. were computed for 30 electrode scalplocations and mean frequency (MF) computed at one-minute intervals spaced across the recording session. The results confirmed our observations of increased drowsiness with the application of TCET. A current versus MF relationship was observed with the largest response at 20uA.The twenty uA's showed a significant (p<.01) decrease in mean frequency when compared to each subject's Sham (0 uA) session. 4 1 9

1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas 399 Biomedical Engineering Scientific Papers

BE23-E.2 BE23-E.3 Regulation and Control of Electro- Regulation of AcupuncturePractices in the stimulation Devices by the FDA State f Texas, H. C. Dung:Department of Michael Cole: J.D., Washington DC Cellular and StructuralBiology, the University of Texas HealthScience Center at The passage of the Medical Devq Amendments San Antonio, San Antonio,Texas 78284. of 1976 to the Food, Drug and Cosmetics Act, and changes in the reimbursement formulas applicable Though the Texas StateBoard of Medical Examiners assumed responsibility to Medicare and Medicaid coverage, have hada t regulate profound effect on the development, testing and acupuncture, and promulgatedtwo separate policy approval of electronic devices for medicaluse. statements with certain differencesin the version The. FDA now has substantial authority toregulate on the practice of acupuncture, thereis no law of choice of materials, design of circuitry, output statutory power that dictatesthe use of levels, software programs and any related aspects acupuncture in Texas.According to the conclusion of safety, efficacy and reliability. The FDA can given by the federal judge inHarris County of the also dictate the scope and duration of human tests State, one of the statementspromulgated by the and the quality and quantity of data needed for Medical Examiners is "unconstitutionaland device approval. The Agency can also closely therefore invalid and withoutforce."Such regulate advertising and label claims, and restrict decision by this federal-courtjudge creates an distribution to classes of physicians and effect which, stated boldly, "isto allow technicians. After approval, FDA regulates all acupuncture to be practiced inTexas by any phases of manufacturing and distribution. The person, traitad or not, who calls himselfor Agency has several substantial enforcement mech- herself an acupuncturistand offers services to anisms to ensure compliance, including criminal the public."The consequence is very unfortunate actions, seizures, injunctions and fines. and the medical establishmentsin the State should The movement to prospective reimbursementcan be blamed for parts of the faultand cripple product development if amounts paid for responsibility.Acupuncture can be easily specific procedures are not sufficient to fully incorporated into establishedmedical practice. cover the costs of using expensive equipment. However, no established medical community . or Several case studies will be presented to institute is willing to acceptit, even after illustrate the discussion of this topic. urging by the Senate Committeeon Health and Human Resources, after the conclusion oftwo public hearings on the matter.

BEV-A.1 BE24-A.2 Microelectronics in Implant Technology, Miniaturized Chamber-Type OxygenSensors M. Schaldach: Zentralinstitut fUr for Brain Research, O.J. ProhrkaIx + Biomedizinische Technik, Universitat , F. 1(8111 , F. Olcarug , A. JRChimowicz Erlangen, D-8520 Erlangen, FR of Germany , K. Parker , W. Chu , M. Patil , J. LaManna . Case Western Reserve yniversity, Cleveland, The application of electronics in implanted ther- Ohio 44106, USA; apeu.ic devices requires small size, high Technical University, opera- Vienna, A-1040 Vienna, Austria; tional lifetime and extremely high reliability. Otto Sensors Corporation, A -1020 Vienna? Austria Continued progress in VLSI-technology andnew materials for packaging devices have led to The A multiple site, multiple parameter development of a large variety of medical im- probe was developed for brain research. plants such as: Thin-film and solid-state techniques were utilized to design Impulse circuits for cardiac pacemakers, defi- miniaturized chaffiber-type oxygen sensors and brillators and neuro-stimulators; to combine them with temperature and electrical DC circuits for bonegrowth and nerve potential sensors re-genera- on a needle-shaped glass substrate. tion; and, The sensors are arranged at an interdistance of control circuits for drug pumps and artificial 300 pm. A thin silicon nitride insulation layer organs. forms a chamber which covers the AuAg/AgC1 A survey of the state of the art for implantable two-electrode electrochemical systems and is filled with microelectronics must include reliability consi- extracellular saline solutior. Holes in the chambers provide the derations. Custom LSI and hybrids used in im- contact between the brain tissueand the electrodes plantable medical products are affected bya va- and define the recordingareas. riety of failure mechanisms. This paper discusses The response time of the oxygen sensor dependsstrongly on the sensor the underlying causes and delineates design design and is in the guidelines that minimize their occurance. range between 0.3 sec and 7 sec. The chamber dimensions In addition, quality and reliability considera- are in the range of pm to 3 pm in height, 20 pm to 40pm in width and tions are discussed for procurement ofcompo- 100 pm to 140 pm in length; theholes have diameters nents, fabrication, screening, and testing of of 10 pm to 20 pm and the working circuits. The paper is concluded with an example electrode L.:ameters are between 6 pm and 20 of the design of a sophisticated dual chamber pm. Good resolution, low noise, lowoxygen consumption and greatly pacemaker to illustrate the steps taken in the reduced interferences between development of a high reliability hybrid design. sensor and tissue are the main advantages of the chamberdesign. )

400 1988 World Congress on Medical Physics and BiomedicalEngineeringSan Antonio, Texas Biomedical Engineering Scientific Papers

BE24-A.4 BE24-A.3 by Biosensors Based on Chemically Sensitive Drift-free Blood Na + Monitoring Field-Effect Transistors, Gary F. Blackburn . a SemipermeableMembrane and IGEN Incorporated, Rockville, MD an ISFET. T.Tamura*, K . Suematsu and T. Togawa, Tokyo Chemically sensitivefieldeffect transistors(Chem- Med. & Dent. Univ. Tokyo101 FET) were first reportedinthe early1970's and have received considerable attention since that time, Japan both in the popular press and in scientific literature. Biosensors which employthe ChemFET todetect An on-line system for continuousmonitoring biochemical events on the surface of the transistor of blood Na+ based onindirect measurement wereenvisionedforthemeasurement of enzyme was developed. Anintermediatesolution in substrate concentrations and for the measurement of a semipermeable antigen and antibody concentrations. contact with blood via While the development of enzyme-based FET's has membrane was used and itsNa+ concentration beenrelativelysuccessful,theimmunochemically- was adjusted by aservo control system so sensitive FET (IMFET) has seen little progress.The that it has the same Na+concentrationas that theory upon which the IMFET is founded is based on the ability of the FET to measure changes in the net of blood. The Na+concentration of the a chargeatthesensor-solutioninterface,ifthat intermediatesolution was measured with interfaceiselectrochemically ideallypolarizable. Na-ISFET because of itsfast response and Experimentally,thetheory has proven difficultto small size. The Na+ concentrationobtained verify because no material investigated to date has with the monitoringsystem wascompared maintained suchanidealpolarizabilityafter more analyzer. The thana few minutesinsolution. Electrochemical with that obtained by the ion measurements of the interfacial properties of response time to a stepchange of thesystem hydrophobicpolymers,inertmetals,andmetals was about 5 min.In long-term monitoring, coated with organic Langmuir-Blodgett films will be the error was within 2 mmioIII. The stability discussed. of this system does not rely uponthat of the sensor. Besidesthis, due to the use of the semipermeable membrane, itcanoperate without blood loss and b: )od contamination.

BE24-A.5 BE24-B. 1 A Dynamic Model for Transcutaneous Oxygen Fiber-Optic Chemical SensorDevelopment, Tension Measurement JohnI.Peterson,* Biomedical Engineering A. Talbot-Pedersen*, M. R. Neuman and Division of G. M. Seidel and Instrumentation Branch, Research Services,NationalInstitutesof Health, Bethesda MD A dynamic model of oxygen transport through the outer skin layers and a polarographic sensor was developed to understand the factors that determine the A general overview of theprinciples of fiber optic relationship between tcP02andarterial chemical sensors will bepresented. with some oxygen tension. The model includes the historical background. The emphasisofthe electrolyte, contact fluid, membrane, dead trendsin and viable epidermis, capillary loop and discussionwillbe on the current tissue of the papillary dermis. development and the approachesthat are being Simulations show the importance of stratum taken for sensor construction. papillare blood flow, viable epidermis and stratumpapillaremetabolic oxygen comparethe oxygen permeability of the An attemptwillbemadeto consumption, optic skin related tothat of the sensor advantages anddisadvantagesoffiber membrane and electrolyte, temperature and sensors relative toelectrodes. the oxyhemoglobin dissociation curve in determining the tcP02. Simulations were of my current found to be consistent with clinical data In addition, a brief description from tenhealthy adults; however, the oxygen sensorinstrumentation and its application importance of having an independent and an will be given. accurate method for determination of skin perfusion was demonstrated. The model showed that estimation of the metabolic oxygen consumption of the viable superficial skin tissue is only possible if the oxygen permeabilities of the skin and sensor are accurately known. 421 401 1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas Biomedical Engineering Scientific Papers

BE2A-B.2 BE24-B.3 An inthava4cataA oxygen 4en4o/L with a DevelopmentofFiber OpticOxygen andGlucose biocompatate di.610.4ion memb/tane, Sensors, R.V. Shah*, S.C. Margcrum and M.N. Gold, M. Martin*, K. Yamakoshi, A. Luthra, Crump Institute for Medical Engineering, UCLA, Los R. Wolton and P.Rolfe, University' of Angeles, CA. Oxford, U. K . The phenomenon of fluorescence q4enchine was used to design A catheter-type oxygen sensor that can be and fabricate fiber optic sensors sensitive to oxygen and glucose dip-coated with various polymer based concentrations. The prototype oxygen sensor was fabricated by diffusion membranes has been developed. coaxially grafting a hydrogcl, poly(2-hydroxyohylmethacrylate) Elrctrochemical techniques, such as potent- (PHEMA), containing a fluorescent dye, 9,10-diplkzylanthracene ial step and time response experiments (9,10-13), to fused silica optical fibers. Grafting was doge through have been utilised to assess diffusion the use of silane coupling agents. The resultingsensor was membrane characteristics of a number of sensitivetosolution oxygenconcentrations due to oxygen membranes, both modified and unmodified by quenching of 9,10-D fluorescence. Dye fluorescence was quenched haemocomnatible coatings. Scanning Electron 20% when the sensor went from an oxygen-free toan oxygen- Microscopy has been used to assess the saturated environment. Transient response times of thesensor were morphology of the same membrane electodes. reduced when the PHEMA graft thickness was reduced. Modeling In vitro and in vivo assessment on sub- of the transient data gave a diffusion coefficient ofoxygen in sequent catheter sensors which employ a PHEMA of 2.15*10.6 cm2/sec. Incorporation ofan enzymatic polyurethane membrane has been carried out. reaction which consumes or produces oxygen through redox In vivo studies show linear correlation catalysis within the oxygen sensing phase makes feasible the with a commercial blood gas analyser over development of sensors specific for the enzyme substrate. We the physiological range of oxygen tensions. incorporated glucose oxidase into the PHEMA/9,10-D matrix in Catheter oxygen sensors so fabricated may order to develop a sensor for glucose. Glucose oxidase catalyzes be used continuously in vivo over a 24 the reaction between glucose and oxygen. In the presence of hour period without any appreciable loss glucose, the matrix oxygen concentration is lowered, resulting inan of signal. Response times of 15 -20s increased fluorescent signal, the increase being proportionalto the in vitro and 40-50s in vivo have been concentration of glucose. It was determined that sensors could be obtained. made to measure glucose in the physiological range when the Thiele Moduli were between 5 and 25 (under conditions where glucose. concentrations » oxygen concentrations).

BE24-B. 4 BE24-B.5 A Transqutaneous Monitoringof Blood Glucose by A Thermoelectric Enzyme Sensor for measuring Suction Effusion Fluid and IonSensitive Field Glucose in-Blood, M. Muehibauer*, E. Guilbeau Effect Transistor Glucose-Sensors, and B. Towe, Arizona State University, M. Kikuchi*, T. Arai, S. Kayashima,N. Negishi, Department of Chemical, Bio, and Materials 0. Takatani, N. Itoh1,J. Kimura!, Engineering, Tempe, Arizona, 85287. Kuriyamal, and A. Kaneyoshil,National Defense Medical College, Saitama,Japan. 1NEC The thermoelectric enzyme sensor isa novel Corporation, Kanagawa, Japan. calorimetric approach to measuring glucose in blood. The sensor consists of a thin film ther- mopile constructed on a thin insulatingsupport, A transqutaneous monitoring of bloodglucose rolled and mounted at the tip ofa 3-mm diameter without blood sampling issuccessfully demon- catheter. It is rendered sensitive to glucose strated.A suction effusion fluid whichis by immobilizing the enzyme glucoseoxidase over obtained transcutaneously fromakin surface by the active thermoelectric junctionsof the ther- suction is employed asa sample material. This mopile. When exposed to glucose in solution, fluid contains glucose andnon protein nitrogens the enzyme becomes heated by virtue of theen- in same concentration comparing withblood. thalpy change of the enzymatic reaction. a effusion rate of the suction effusion Since A tem- only 0.4u1/min/cm2(skin), fluid is perature difference between the active andrefer- we use micro ISFET ence junctions of the 'Iermopile results and a glucose-sensor of which requiredsample volume is 5-10u1. In animal experiment, adult-white-rabbit glucose-proportional Se beck voltage is induced. skin eliminated stratum When operated in blood in vitro, thesensor corneum is used. By responded to the plasma concentration ofglucose. glucose injection via a femoralvenus blood No ambient temperature controlwas necessary. glucose concentration is changedas step-wise. The Dependent on the enzyme loading, the sensitivity blood glucose change ismesured both by this was between 20 and 60 nV/mg % and the response transqutaneous method anda conventional blood was linear up to 250 mg %. The response time was glucose analyser simultaneously.The transquta- typically less than 10 seconds. neous monitoring by the suction effusionfluid and When operated in vivo in an anesthesizeddog, the ISFET glucose sensor revealcomplete detection the sensor responded appropriately of blood glucose change within to changes in 15min delay. The glucose induced by injections of insulin. The authors have also been examinedthis method to response matched that found at the same temper- clinical oral glucose tolerance test. ature and flow rate in vitro. 42 402 1988 World Congress on Medical Physics and BiomedicalEngineeringSan Antonio, Texas Biomedical EnsinboringScientificPapers

BE24-C.2 BE24-C.1 Laser-Doppler Stethoscope for Loadfree Temperature Sensing with Fiberoptic Recording of Skin Vibration , B Hok",* Probes for Biomedical Applications, Electronics Department, Institute of D.A.Christensen*, Dept. of Technology, University of Uppsala, Bioengineering, University of Utah, Box 534, S751 21 Uppsala, Sweden, Salt Lake City, Utah 84112 USA P A eberg, Department of Biomedical Engineering, University Hospital of Linkoping, Applications are increasing forfiberoptic S-581 85 Linkoping, Sweden. temperature sensing in the clinical environment, including tissue monitoring during The widespread use of the conventional stethoscope points electromagnetically-induced hyperthermia(for at the richness of skin vibrations as a source of clinical cancer therapy),blood gas monitoring with information. Attempts in objectifying these signals have, multi-purpose fiberoptic catheters, and patient however been quite disappointing. Phonography, including monitoring during magretic resonance imaging phonocardiography and phonopulmography, is still not a (MRI). Due to their use of nonconducting glass quantitative, or even calibrated method. or plastic optical transmission fibers, these We suggest the use of a Laser-Doppler system as a sensors possess uniqueadvantages over reference method in phonography.It is shown that, by conventional metallic temperature devices, such using a system based on a low power helium-neon laser as electrical insulation, immunity from and single-side-band techniques including a Bragg cell, it is crosstalk, and nonperturbation in possible to record and quantify the vibration of skin, loaded electromagnetic fields. only by a flexible, thin retroreflecting tape. Fiberoptic temperature sensors may be Possible optical configurations are discussed, together with classified into three general categories: 1) estimations of accuracy, bandwidth limitations and noise. Intensity, 2) Wavelength, and 3) Time-domain. Experimental recoraings will be shown with parallel Examples of each of these sensor types will be conventional microphone signals. Also, the loading effect presented, including the liquid crystal tip of applied mass on the skin surface is demonstrated. (intensity), fluorescence decay (time-domain), Fabry-Perot cavity (wavelength), and the GaAs semiconductor sensor (wavelength).

BE24-C.3 BE24-C.4 Optical Fiber SensorFor Joint LASER-BASED FIBER-OPTIC SENSORS, Frank V. Bright Angle Measurement Thomas A. Betts, Gino C. Catena, Kevin S. M.M. Patil* and O.J.Prohaska Litwiler, and David P. Paterniti, Department of Case Western ReserveUniversity, Cleveland. Chemistry, State University of New York at Buffalo, Buffalo, New York 14214 USA. Optical fibers were used to design a joint angle measurement device with the intention of using it Over the last decade, there has been a near for controlling finger position in the functional exponential increase in the number of neuromuscular stimulation systems. Light trans- fluorimetrically-based fiber-optic sensors. mitted through a joint of two fibers with their During this time, fiber-optic sensors (FOS) have polished endfaces normally held perfectly aligned been developed for the determination of everything by silastic tubing, undergoes transmission losses from birmoleculcs to metal ions.However, one of when there is an angular ti't between the fibers. the cratinuing problems with FOS arises from These losses modify the int.asity of light, which matrix interferences.To P.lp these interference was detected by a photodiode and wasrelated to problems our group has been exploiting the the angle between the two fibers. The fibers were inherent multidimensional nature of fluorescence placed such that they formed a joint beside coupled to fiber-optic sensing. Specifically, (laterally) the phalangeal joint and were fixed our efforts are aimed at polarization, lifetime, on either side of the joint. Hence detectionof and rotational dynamics information measured the angle between the fibers leads to the deter- through long lengths of fiber optic. In this mination of the finger joint angle. presentation, selected results from our labor- Sensitivity (between 2 mV/0 to 12.2 mV/°) and atory will be presented. In addition, it will be linearity of this device were found to be depen- shown that it is now possible to perform all dent on the position of the fibers in the measurements that were once restricted to a silastic tubing, the tubing itself, as well as cuvette through a simple fiber optic probe. the different types, diameters and endface angles of the fibers. Investigations so far indicate that neither hysteresis nor temperature changes significantly affect the angle measurements. 423

403 1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas Rimme.41.1 Sete -"tic Papers

BE24-C.5 BE24 -D.1 Evaluation of POET: Combined Pulse Disposable Pressure Transducers for Research Oximetry and End-Tidal CO2 monitor, E.Y. and for Clinical use, N.B. Jones*, G.L. Cheng*, K.A. Stommel, Medical College of Papageorgiou I V.C. Redding, University Wisc.-Twin, Milwaukee, WI cf Leicester an- The Regional Cardio The POET (Criticare Systems, Inc.) is the first Thoracic Unit, Leicester, U.K. monitor to combine both the pulse oximeter and side-stream infrared end-tidal CO2 analyzer into It is known that the dynamic fluctuations of art- a single unit. In this study we tested the erialblood pressure with time are significant accuracy of the POET in 15 healthy nonsmoking diagnostic andprognostic indicators, particul- adult volunteers. Each patient was subjected to arly when related to flowmeasurements taken at inspired 02 fractions of 1.0, 0.21, 0.12, 0.10, the same point. Work in this area andon many and 0.08, and then CO2 concentrations of 0, 3,6 otherbiological systems is revealing a require- and 7.5%. Subjects reequilibrated with room air mentfor pressure transducerswhichdo not for 5 minutes between each gas mixture. produce serious flow disturbances and can be Pulse oximetry sensors were placed on the index introducedinto very small vessels and cavities. firlar (SpfO2) and the forehead (Sph02). For It is also required that in vivo transducers be each Oa concentration tested, a 30 second robust and safe and have a widebandwidth and steady-state of Spf02 and Sph02 was neededbe- good signal to noise ratio. If procedures devel- foredrawing arterial blood for Sa02 determina- oped as a result of research are to be adopted tion. A sampling line for PetCO2 was placed into routine practice transducers shouldbe easy just distal to the one-way exhalation valve. to use. Disposability of the transducers is also ABG were drawn and analyzed for PaCO2 after of considerable significance particularly where PetCO2 stabilized for 30 seconds for each HIV infection is a possible ftctor. This paper concentration of CO2. reviews the situation in pressure transducer Linear regression was used for data analysis. technology from this viewpoint and shows that Correlation coefficients if 0.95, 0.98 and 0.95 opportunities exist in a new class of fibre optic were found when comparing Sph02 to Sa02, SpfO2 transducer which have already been madedown to to Sa02 and PetCO2 to PaCO2 respectively. Dur- 400 pm in diameter, having a bandwidth ing this study the range of Sa02 was between 65- in excess of 1000hz and a signal to noiseratio of at 100% and PaCO2 ranged between 35-55 mmHg. We least 60 dB. conclude that the POET provides an accurate, and convenient single unit monitor for following Sp02 and PetCO2.

BE24 -D. 2 BE24 -D.3 Guiciewire-mounted sensor for biomedical pressure Analysis of PedobarogramsObtained by a measurements L. Tenere.B. McElectronics Department. Computerized Platform. A.Starita, Institute of Technology. Box 534. S-751 21 Uppsala. Sweden. T. P.Dario, M Bergamasco and A.Fiorillo, Engstram,0. Hammarstram,R. Lonc. Radisensor AB. Dipartimento di Informatica; Scuola Xungsgatan 70. S-753 21 Uppsala. Sweden. Superiore S. Anna; Centro "E.Piaggio", University of Pisa, Pisa,Italy. We have developed an ultraminiature (0.5 nun diameter) pressure sensor and mounted it at the tip of a thin guidewirt used Quantitative measurements of foot-ground for guiding the sensor to the measuring site. The sensor uses a pressure distribution are becoming increasingly micro-optical pressure sensing element made by silicon important in such areas as orthopaedics, micromachining and fiberoptic signal transmission through a rehabilitation, neurology, diabetology, single-strained hard-clad optical fiber. The measurement biomechanics and sports medicine. properties of the sensor (sensitivity. resolution, linearity. Using a recently available commercialsystem stability and frequency response) have been found fully (Ornomat, Polysens S.p.A.,Florence, Italy) adequate forclinicaluse. In animal experiments, the which consists of a platform incorporating1024 measurement properties in vivo. and the guiding properties of piezoelectric polymer-based sensors, an the assembled guidewire sensors have been verified. interface unit and a personal computer, Potential applications of the sensor include left ventricular pedobarographic images have been obtained in manometry. coronary artery investigations. manometry in different experimental conditions(normal and small animals, catheter-tip localization. oesophageal motility pathological gait, load transfer). The images measurment. Results will be shown from urodynamic corresponding to the time evolution of pressure measurements, arterypressurerecordings with waveform distribution have been processed in order to extracta numberof potentially significant analysis. venous pressure monitoringpressure measurements parameters (contact area, center of pressure, in cardiac muscle tissue and in the kidney tissue. maximum load points).

404 1988World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas Biomedical Engineering Scientific Papers

BE24-D.4 BE24 -D.5 The Design of an Ambulatory Blood USE OF A PIEZOFILM TRANSDUCER TO Pressure Monitor, R. Branstetterand VERIFY PROSTHETIC LOADING, A. Levii, J. Jaeb, Southwest Research Institute, J. Stoessel, ANCO Engineers,Inc., San Antonio, Texas and L. Mader, David M. Grant Medical Center. Daily blood pressure measurements are not adequate Often, for some ambulatory, hypertensive patients. A need exists for a capability to accurately like severalmeasurements a cardiologists would measure loads applied on prosthetic limbs by day. Additionally, measurements taken at particu- patients. This is critical in determining lar events during the day, such as when the patient proper loading of prosthetics and in pre- has taken medication, eaten a meal, or had an argu- venting noncompliance due to patient dis- ment with a spouse, might be very useful. comfort and tissue damage. Classic trans- A group ofscientists and engineers have develdped ducers have been too bulky to be of clinical an ambulatory blood pressure monitor(ABPM). Mea- utility. A prototype pressure sensor package suring 3x7x14 cm and weighing only 800 g, this com- consisting of a piezoelectric polymer PVDP fortable device has several important features. The (tradename KYNAR) has been constructed that results from 800 blood pressure measurements can be IS thin and flexible enough to fit between solid-statememory stored in a small, removable, the patient and prosthetic and thus provide cartridge. Programmable keys on the face of the in- the Clinician with sufficient information on special events strument allow the patient to mark the loading contour to permit a more com- port pro- during the day. A serial communication fortable fit between patient and prosthetic. any com- vides easy down-loading of stored data to The piezoelectric film is durable, and flex- puter. A rechargable battery provides 24 hours of ible enough to conform to any required inter- operation. face between patient and prosthetic. For valving A CO2 cylinder andspecially-designed this effort, instrumentation was installed on yields a precise, constant cuff inflation or defla- an interface cup for below-knee amputations. tion rate. An important benefit of this gas oper- The cup was then placed between the patient ated, cuff-inflation system is its very quie' oper- and the prosthetic. Based upon a series of ation, as compared to similar electrical pump sys- loading situations, a loading contour chart tems. EEGrecords show that operation during was generated which could be used by the sleeping hours can be attainedwithout disturbing clinician for modifications to the fit. the patient. This monitor is the forerunner of the Nippon Colin ABPM-630.

BE24-D.6 BENt-E.1. Apparatus for Measurement of Skin Development of Medical Telemeter Elasticity in Systemic Sclerosis using Indirect Light Transmission, M. R. Neuman*, S. P. Ballou, A. Mackiewicz K. Shimizu*, N. Kudo°, H. Murakami, and A. Lysikiewicz T. Mikami and G.Matsumoto, Hokkaido

. University, Sapporo Japan, °Toshiba Anextensometer for measuring! Corp., Ootawara, Japan viscoelastic properties of the skin on the left volar forearm has been developed and There have been some reports on the opti- applied to discriminate between patients sex and race cal telemeters using a direct light with scleroderma and age, transmission. Although their advantages matched controls. The device stretches over the conventional radio telemetry the skin from 20 to 28.5 mm at a constant have been verified, its application in while continuously rate of 0.4 mm/s practice has been limited due to the measuring the associated force. The sensor was held in contact with the skin problems in maintaining the propagation using a strong double sided adhesive tape path of the light. To solve this problem which could be pulled away if the skin a technique using an indirect Ugh* tension became too great. A microcomputer transmission is proposed. The distribu- collects the data and determines the tion of the indirect light in a room was elastance from the slope of the force- calculated, and the feasiblity of this elongation curve. A reasure related to technique was examined in various condi- plasticity is an exponential curve fit to tions. To verify the possibility shown in the force rclaxation at maximum extension. the theoretical analysis, an optical The device was tested on thirteen patients telemeter using the indirect light trans- with clinically documented scleroderma and mission was developed. The signals of 3 thirteen matched controls. The ECG's and a body temperature are multi- scleroderma patients had a mean elastance plexed in time domain and modulated into a PIM pulse sequence. With the developed of 7.1 ± 3.3 (SD) N/mmwhith was significantly different from contzols at telemeter the optical biotelemetry from a 3.5 ± 0.63(SD) N/mm (p < 0.005). This freely moving subject was carried out. A instrument may,therefore, be of use in stable data acquisition was demonstrated objectively evaluating patients with early even from the subject in exercise. stages of scleroderma. 425 1988World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas 405 Biomedical Engineering Scientific Papers

BE24-E.2 BE24 -E.3 Telemetry System for Intraoral Pressure A MICRO TELEMETRY SYSTEM FOR NEONATAL Measurement, Y. Kato*, T. Kuroda and MONITORING, Kenji Ikeda: Institute of T. Togawa, 2nd Dept. of Orthodontics, Medical Electronics, University of and Inst. for Med. and Dent. Eng., Tokyo Tokyo, Bunkyo, Tokyo 113, Japan Medical and Dental University, Tokyo, Japan. Immature babies, who need itnersivecare and are protected in incubators, are in almost allcases A telemetry system which can monitor the under fluorescent lightning for the treatments of intraoral muscle pressures in subjects under un- jaundic,. ,f the newborn. In such occasion, the constrained conditions was developed. The system light energy can be utilized as thepower source consists of an intraoral unit and an extraoral of the transmitter for wireless monitoring of one. The external unit is about 2.6 kg including respiration or heart rate of the newborn. batteries and can be carried by the subject. A very simple, micro transmitter, which contains Electromagneticcoupling between an intraoral no battery but opperates using the light energy, coil and an extraoral one is used to supply the foL measuring small displacement of abdominal skin surface due to the respiration, is developed and power to the intraoral unit and to receive the made for trial. The size is 12x12x6mm3, and the signal corresponding to the pressure of the up- weight is approx. 1g. The skin surface displace- per central incisor region. The cantilever type ment is measured by the small change of the air- pressure sensor, which is 6 mm in diameter and 1 gap capaciatnce between the skin and the trans- mm in thickness, was placed on a tooth. mitter itself. The transmitter, being entirely This system was applied in four adult male sub- solid without lead wires or connectors, iseasy jects. The averaged labial pressure of the upper to sterilize and handle. central ncisor was 1.5 g/cm2, 10.9 g/cm2, and VHF(70,,80MHz) radio wave is modurated in FM. As 5.0 g/cm during rest,swallowing and speech, the receiving antenna is installed juston the respectively. The averaged lingual pressure at upper cover of the incuvator, only very weak radio the same kocation was 4.0 g/cm2, 19.1 g/cm2 and wave is enough for telemetering. This can avoid 13.9 g/ce during rest, swallowing and speech, jamming when several systemsare used simultane- respectively. It wasproved that this telemetry ously in the same room. Some results obtained system is applicable for unconstrained, long- from practical tests at a maternity hospital will term' monitoring of intraoral pressures. be shown for discussions.

BE24-E.4 Free-space Optical Data-link for 3E25.1 Biomedical Telemetry, H.Odagiri, Interaction of Weak ELF MagneticFields K. Shimizu*, and G. Matsumoto, with Hydrated ProteinChannels. A.R. Hokkaido Institute of Technology, Liboff* and M.D. Sevilla,Oakland Univer- Sapporo, *Hokkaido University, sity, Rochester, Michigan48063. Sapporo, Japan. Experiments by differentgroups designed to test forion cyclotron resonance A real-time transmission of biomedical (ICR) interactions signals has an increasing significance in indicate that it doesnot occur in all tissues. the field of biotelemetry particularly in However, the Ca2+-dependentmotility of diatoms is telemedicine. With a view toward the a sharp function of ICR tuningas is the uptake of 45Ca2+ in human multi-variable analysis of biological lymphocytes. Both systemsnot only parameters under various environmental undergo resonant changesas a function of magnetic fieldfrequency but also conditions, a system was developed which vary according tothe enables us to obtain the information ofa ionic charge-to-mass ratio. If the interaction patient from a remote place. The system site lies within the ion channel,and the resonant consists of two stations connected witha frequency is determinedby the dehydrated ionic mass, free-space optical link. At the data this may imply that theICR transport mech- acquisition site, biological signalsare anism occurs in a sLrudturedaqueous lining within the channel, obtained, pre-yrocessed by a microcompu- and further that channelsnot having such well-ordered hydration ter, multiplexed with color images and structures may not exhibit ICR. sounds, and transmitted in a beam of We have formulateda helical hydration light. At the data-analyzing site, the triodel that conformsto the interior symmetry and light is received, and the different dimensions of the Gram Adimer channel. This model kinds of information are retrieved. Using provides for the transmissionof ions without loss this system, the data obtained at experi- of hydration shell througha periodic structure mental fields or in a remote buildingcan that is responsive, inprinciple, to Ia. Supported by DOE: be fed to a large computer in another offices of Energy Storageand building which performs the advanced Distribution (ARL)and Health and Environmental data-processing. The function of an auto- Research (MDS). matic repeat request prevent the data loss in a short time interruption occured :t L; in the data-link.

406 1988 World Congress on Medical Physics and BiomedicalEngineeringSan Antonio, Texas Biomedical Engineering Scientific Papers

BE25.2 BE25.3 Influence of the Magnetic Field on Ion Magnetic Field Influence on Living Tissues M.S. Markovt Department of Biophysics, Sofia Movement near Biological Macromolecules: University, Sofia, Bulgaria Theory and Experiments, A. Chiabrera *, M. Figueiredo, B. Bianco, The influence of static magnetic fields as well as J.J. Kaufman and A.A. Pilla, Univ. of Genoa, magnetic fields produced by AC current has been Italy and Mount Sinai School of Medicine, studied. The range of field strength from background to 350 mT has been carefully studied. New York, U.S.A. The reactions of living tissues has been estimated by number of biophysical parameters including Most bioelectrochemical processes depend on ion electrophoretic mobility of cells, passive elec- transport or adsorption. The ion velocity change, trical properties of membranes, light scattering, due to an electric field whose amplitude is propor- oxygen consumption, transmembrane potential, stability of cells and cell membranes to factors tional to the time variation of a periodic exogen- provoking cell lysis. ous magnetic induction field, is computed, taking Ekistance of specific levels of optimal effects into account both the dc and ac magnetic Lorentz at different tissues has been found.The "permitted" forces. The ac magnetic induction is very effective levels of reactivity are similar to the "biological provided that its amplitude is related to the ion windows" supposed by W.R. Adey. cyclotron frequency, or to its multiple values, and In a series of experiments has been observed that that the local viscosity is low. The 1.1aromenon oc- magnetic field may act as a factor of preventing curs at intensities of the dc magnetic induction action of other physical and chemical factors, that are also related to multiple values of the ion including ionizing radiation. cyclotrkul frequency. The predicted maximal effectiveness offers a consistent explanation of some experimental results so far obtained affecting the swimmilg direction of paramecia under electromagnetic exposure.

BE25.4 BE26-A.1 Mechanism of Microwave Radiation Effect on Biological MeabutementA oti Regionat Stood Ftow Objects, 5.1. Kirkilevsky, Ya. I Khokhlich, V.A. the Foot bymeans oti Manzient The mat Pavlenkot N.N. Korpan, Kiev Medical Institute, CleaftaneeMethod. M. Nitzan*, Y. Mahler 252004 Kiev, USSR. and Z. Abramowitz, Jerusalem College of Technology, Jerusalem 91160, Israel Theoretically grounded model has been worked out, suggesting that enzyme molecules can be the The non-invasive transient thermal clearance me- application point for the microwave electromagnetic thod provides a means for the assessment of region- field of nonthermal intensity. Frequency resonance al blood flow (RBF), in unit of ml blood per 100g of electromagnetic field with the inherent oscillationper minute. The ratio of the blood pressure in the frequency of the enzyme molecule leads to alteration artery supplying the tissue and the RBF in it is in kinetics and direction of fermentation. the specific microvascular resistance (SMR). RBF and arterial blood pressure were measured on The experimental research has proved the elaborated the dorsum of the foot, for diabetic and non- theoretical pattern, which allowed to apply patho- diabetic patients. For most of the non-diabetic genetic approach in treatment and prevention of patients, RBF and arterial blood pressure were some diseases by using extremely high frequency mutually proportional, indicating that SMR is (microwave) radiation procedures. The research approximately constant. Some of the patients, has resulted in a conclusion that full variety of however, showed different values of SMR, either biological and medical effects induced by microwave below or above the range of normal SKR.As for EBY radiation can be explained by this hypothetic the diabetic patients, no correlation between RBF mechanism, which is of great importance for biology and arteriol blood pressure was found. The mean in general, and medicine, and social hygiene. value of SKR for the diabetic patients was significantly lower than that of the non-diabetic patients.

42 7 1933 World Congress on Medical PhysicsandBiomedica! EngineeringSan Antonio, Tears 407 Biomedical Engineering Scientific Papers

I3E26-A. 2 BE26-A.3 THERMAL CAMERA IMAGING TO MEASURE ComputedThermographySystemapplied PERFUSION FROM THE TISSUE SURFACE to DynamicAnalysisofSkinSurfaceTemperature, J.C. Chan*, J.W. Valvano, Ph.D., J.A.Pearce, Ph.D., I.Fujimasa',' T.Chinzei, K. Mabuchi, Y. Abe, and K.Atsumi, L.J. Hayes, Ph.D., S.A. Prahl RCAST and IME, University of Tokyo,4.6-1 Komaba, University of Texas, Austin, Texas Tokyo,153 Japan

A thermal washout technique is being developedwhich The computed thermography system(CTS) has been applied measures perfusion from the tissue surface. This perfusion mainly to temperature abnormality in steadystate caused by measurement is similar to other indicator washout techniques. sympathetic irregular stimuli and peripheral blood flow The basic approach is to applya heat source to the surface of ob:,tructicn. The fundamental analyzing procedures are the a tissue for exactly 60 seconds. The heatsource is then asymmetry detection of temperature distribution and image removed and the temperature recovery is monitored with a subtraction between two thermograms.Thermatome and blood calibrated thermal camera. Blood flow significantlyaffects flow distribution are two main expessions of the analysis. local heat transfer making this technique quitesensitive. Recently, thermographic instruments can produce dynamic Highly perfused tissue will absorbmore heat and, hence will thermal images using the SPRITE detector or CCD elements. In have a smaller temperature increase anda faster recovery back thermograpic laboratories, doctors want to analyze dynamic to baseline. skin temperature change which is caused by thermalor The finite element numerical method isused to model mechanical stress. For the rapid acquisition of dynamic the heat transfer problem with realisticgeometries and thermal data, we have been developed a.high speed 1/0 interface boundary conditions. The relationship betweenthe actual and a large RAM memory system on personal computers of perfusion and the measuredtemperature response is PC9801(NEC) and PC/AT(IBM).The algorithms of the software determined using both analytical and numericaltechniques. It system consist of two parts: transient analysis and spectral is important that the boundary and initial conditionsbe care- analysis. The principle of the transient analysis in thermal fully controlled and monitored during both thewarming and phenomena under strong external load is depended cooling phases. In vivo experiments with upon the alcohol fixed detection of the gradient of temperature change.The pattern canine kidneys demonstrate the feasibility of the technique. expression is named secondary differentiated thermograms. As The disturbing factor; include: uncertainty ofthe temperature the spectral analysis, cycles of fluctuation of skin surface measurement, uncertainty of the time measurement, baseline temperature are calculated by FFT.The expressed pattern is temperature gradients, and perfusion gradients. called the thermal rhythm spectrography These images are possible to explain dynamic sympathetic function activityto skin arterioles and are useful to the pathophysiological analysis of autonomic nerve disorder.

BE26-A.4 BE26-A.5 Thermal Conductance and the Pattern of Distri- Measurements of Tissue Thermal Conductivity and bution of Blood flow. E. Raman*`, V. Vanhuyse, Perfusion using Sinusoidally Self-Heated Thermistors M. Maes and J. Janssens.Univ. of Antwerp, G. Anderson*, J. Valvano, J. Pearce, L. Hayes R.U.C.A., B-2020 Antwerp, Belgium. University of Texas, Austin, Texas

To examine the active elements of the thermal A microcomputer based instrument is being regu- developed which simultaneously measures intrinsic thermal lation in extremities many researchers measurethe conductivity and perfusion. In this technique,the physiological thermal conductance of perfused tis- combination of a steady state and sinusoidal voltageis sue. The results are oftenin contradiction with applied to heat a miniature thermistor probe inserted the findings from other bloodflow methods invasively into the tissue of interest. A second thermistoris (e.g. used to measure the baseline tissue temperature. The radioactive clearance rate). We studiedthe rate of applied electrical power and the resulting temperature riseare blood flow, the distribution of flow and theheat measured by the microcomputer based instrument. loss bymeans of ananalytical circulation model An approximation is used to lump the heat transfer and compared the results with the thermal conduc- due to perfusion and conduction intoan effective tance parameter. The model is fitted by experimen- conductivity term. Perfusion is then linearly relatedto the difference between effective and intrinsic conductivity tal blood flow (venous occlusion plethysmography) measurements. The probe sensitivity increases with its size. and heat loss (calorimetry) measurements.Results The steady state response is sensitive to both conductionand presented show that thermal conductance is apoor perfusion. The excitation frequency is chosen suchthat the parameter to determine rate of blood flow (systemic sinusoidal response is sensitive to intrinsic thermal errors of 50%) specially when the distribution of conductivity but insensitive to perfusion. Thisprocedure gives a simple but effective method for separating flow changes. conductive and convective transport without the necessity ofinterrupting Our conclusion is that only analytical studiessup- blood flow, as is required in previous techniqles. ply the correct chara^teristics of the activether- The accuracy of the thermalconductivity mal regulation, such as: cutaneous blood flow,pat- measurement is 2% over the range of 3 to 6 mW/cm-IC. tern of distribution, counter current Measurements in alcohol-fixed canine kidneysdemonstrate heat exchan- that the instrument can simultaneously ge, blood temperature. measure intrinsic thermal conductivity and effective thermalconductivity in perfused tissue. 4 28

408 1988 World Congress on Medical Physics and BiomedicalEngineeringSan Antonio, Texas Biomedical Engineering Scientific Papers

BE26-A.6 BE26-B. 1. Non-contact Skin Emissivity Measurement by A Coupled Electrical and Thermat Finite Element Modelof Switching Two Shades of Different Temperatures, Self-Heated Thermistors T.Togawa: Tokyo Medical and Dental University, J. Valvano*, G. Anderson, J. Pearce, L. Hayes, C. White 2-3-10, Kanda Surugadai, Chiyoda-ku, Tokyo University of Texas, Austin, Texas

Although skin emissivity data in far-infrared Thermistors are used as both active and passive region is required for accurate thermographic transducers in a variety of medical instruments. Previous studies, no convenient method to measure skin thermal models treated the thermistor as either a single point emissivity has been provided. A difficulty in the or as a lumped sphere of homogeneousproperties. When conventional methods is to determine actual radia- the thermistor is self-heated, these simple models do not tion source temperature at the skin. To overcome accurately predict the relationship between theapplied this, we employed a simple reflectance measure- electrical power and the resulting temperature rise. ment, which is essentially an emissivity measure- A closed form analytic temperature distribution can ment. Two shades of 10 cm in diameter were used, only be found if the self-heated thermistor is spherical and and one was kept in room temperature, and the the power is dissipated uniformly in all regions of the bead. other about 20 C above that. The surface inside Unfortunately, both assumptions are false. A nonuniform within the the shades was coated with black paint. The object power deposition results because the hottest areas skin surface was kept just under a shade, and the thermistor also have the greatest electrical conductivity. radiation from the skin was observed by a radiome- Transient 2-D finite element electrical and thermal ter having sensitivity in 8-14um, through a hole models are used to examine the complex interaction between of the shade. Then the shade was switched to the applied electrical power and resulting temperature rise. other by a slide mechanism, and observe the change Realistic geometries can be entered into this program. in the radiation energy. The reflectance, and thus Initially, the elvmrical properties and temperature are the emissivity can be calculated only from the uniform. At each time step, the electric field is calculated radiation energies before and after switching, and from the electrical properties generated in the previous step. temperatures of these shades, which determine the Then, the nonuniform applied power is calculated from the ambient radiation temperatures under the shades. electric field. Then, a new temperature distribution is By this method, each measurement can be made wit- lculated from the applied power, the thermal properties, hin one minute, and because of non-contact fea- and the tissue blood flow. Lastly, new electrical properties ture, this method is safely applicable for measu- are calculated based on the temperaturefield. This process rements in the pathological skin. iterates back and forth until steady state is reached and final power and temperature distributions are obtained.

BE26-B. 2 BE26-B.3 Whole Body Heat Balance during the ON THE THERMAL CONTACT BETWEEN BONE Human ThoracicHyperthermia, AND ARTIFICIAL JOINTS Zheng Lou*, Wen-jei Yang, the * University of Michigan, Department of L. S. Fletcher , and G. P. Peterson Mechanical Engineering and Applied Department of Mechanical Engineering Mechanics, Ann Arbor, MI48109-2125, Texas AO University "SA. College Station, Texas 77843 To predict the temperature fields or to derive the Wherever there is an imperfect mechanical contact and a temperature gradient, a temperature heating strategy according tothedesired discontinuity will occur. This temperature temperature range before tumor treatment, most of discontinuity is normally described in terms of the investigators usetheso-called bio-heat the thermal contact conductance, defined as the equation.In the perfusion term of the equation, heat flux divided by the apparent interface area. incomingbloodtemperatureisgenerally Understanding of this temperature discontinuity considered to be constant.But due to the large along with the associated thermal condu:tivity of intensity of energy influx, especially in the bone and cement materials is important in order treatment of the human thoracic tumor, the body to better understand the thermal phenomena that core temperature may be raised. occur as a result of the heat generated during curing of the cement and also the dissipation of A whole body heat balance model during localized heating caused by friction. hyperthermia is developed.In the model, local Presented here are the results of an temperature is calculated using a finite element experimental investigation in which the effective model. The perfusing blood, along with its energy, thermal contact conductance occurring at the is circulated to the rest of the body, where the heat interface between bone and stainless steel were dissipation is calculated using lumped models. measured. Joints utilizing four types of Foly- With this model the effect of the power dosage on methylmethacrylate (PMMA) based bone cements were the body core temperature and thus on the tested and measurements of the thermal contact temperature fields is analyzed for the human conductance were made along with measurements of thoracic hyperthermia. the thermal conductivity of the bone material for 4 2 th saturated and dry conditions.

1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas 409 Biomedical Engineering Scientific Papers

BE26 -B.4 BE26-B.5 Quantitative Analysis of Tissue Coagula- ThermalMonitoringwithanIngestible tion by Electrosurgical Current, Choh-Fei Rechargeable Capsule, J.C. Lesho*, A.F. Hogrefe, Yeap* and John A. Pearce, Dept. of Electrical & R.C. Eberhart, W.E. Radford and H.K. Charles, Computer Engineering, UT-Austin, Austin, Texas. Johns Hopkins University Applied Physics Laboratory Johns Hopkins Road Laurel, Md. 20707 Electrosurgery has been in wide-spread clinical use for over 60 years, yet physical data on the generation of cali- Rechargeable ingestible thermal monitoring capsules brated lesions are not available. have been designed, fabricated and tested. Two versions have Electrosurgical current (500KHz) has been applied been attempted. The first continuously telemeters data, through a circular electrode (5mm diameter) to canine heart while the second telemeters data only when commanded. muscle in vitro. Rms voltage, current and surface tempera- Both versions use an inductive link for data transmission and ture, measured by a thermocouple at the electrode center, recharging. The commandable version also uses the inductive were recorded. The surface spatial temperature distribution link for commands. was also determined using a video7scanning thermal camera. A battery powered receiver has also been designed, Preliminary experimental results suggest that the electrical fabricated and tested.Data may be logged manually or conductivity of tissue decreases markedly when coagulation automatically by an IBM PC. initiates at about 65 °C. The ingestible capsule is eight millimeters in diameter The thermodynamics of tissue coagulation by electro- and 21 millimeters long. The electronics are encapsulated in surgery is modeled by nonlinearly-coupled Laplace and Bio- epoxy then coated in silicone rubber.The capsule is Heat equations including the nonlinear temperature depen- rechargable without having to take it apart. The system has dence of electrical and thermal conductivity. The 2-D Finite a resolution of 0.1 °C. Readings from the continuous pill can be Element and Newton-Raphson methods were utilized to pre- taken once per second. The commandable pill can be queried dict power density and temperature as a function of both time four tim. es per minute. and position. Arrhenius reaction rate function was used to Results of the first tests of the capsule are summarized in estimate severity of tissue damage. Good agreement was the paper. The ongoing program for the development of a four obtained between computed and experimental temperatures channel system to measure temperature, pH, heart rate and prior to structural and properties changes due to coagulation. pressure is also described. The support of the NASA Goddard Space Flight Center is gratefully acknowledged.

BE26-B.6 CRYOPRESERVATION OF ORGANS USING OPTIMIZATION METHODS, G.S. Dulikravich*, J.V. Madison, PennState Univ., StateCollege, PA 16803, and L.J. Hayes, Univ. of Texas, Austin, TX 78712, USA.

The survivability of living tissues andorgans is a strong function of the local coolingrates in each part ofthe organ. When preservingan organ, a maximum survival rate can be achieved by maintaining optimal local coolingrates. The optimal cooling rates for each type of the tissuecan be experimentally determined. We will demonstrate that the optimal local cooling ratescan be enforced by unsteady temperature variationon the walls of a container in which the cryoprotectivefluid and the organ are located. An optimaltime-space variation of the container wall temperaturecan be determined computationally usingan inverse formulation. This represents an entirelynew scientific approach to living tissue banking that is basedon sound physical and mathematical grounds. Thecon- cept of optimized distributed cooling has beenre- cently mathematically formulated and computationally proven by the lead authors-for thetwo-dimensional case of pure heat conduction ina kidney and the cryoprotective fluid. Themathematical model included three regions with different heat diffusivities, but did not account forlatent heat release during the phase change. The latent heat release and variable diffusivities will be included. ^

410 1988 World Congress on Medical Physics and Biomedical EngineeringSan Antonio, Texas Author Index

Aalbers, A.H.L., 118 Andreo, P., 142 Baratta, R., 387 Barbaro, V., 322 Abbot, J., 11 Andreoni, A., 17,24 Ababa, A.B., 20 Andreucci, L., 140 Harbin', P., 218 Abdul-Kadir, N.H., 338 Andrew, J.W., 51 Sarum, U., 327 Baretich, H., 186 Abe, H., 180,303,364 Andrews, B.J., 388 Barker, A.T., 170 Abe, H., 137 Angelides, K.J., 252 Barnes, C.D., 340 Abe, R., 113,114 Ann's, H., 35 Barnes, G.T., 69,70 Abe, Y., 408 Anno, I., 99 Barnes, W.H., 122 Abe, Z., 112 Ansel, H.J., 88 Abouelnasr, D.H.,360,372 Antich, P.P., 122,128,143 Barnett, C., 153 Abramowitz, Z., 407 Antila, K., 324 Baron, J.J., 40 Antolak, J., 44 Barr, R.C., 169 Abrons, H., 226 Barrionuevo, 0.J., 188,197 Aubry, F., 111 Antolini, R., 171 Barsai, J., 321 Accornero, N., 368 Antonuk, L., 57 Barta, E., 231 Acevedo, H.F., 82 Aoki, T., 231,325 Acharya, B.L., 24,28 Aoyagi, T., 281 Barth, K., 150 Adachi, A., 33,61 Apicella, A., 391 Barth, N.H., 8 Barthes-Biesel, D., 354,355 Adachi, T., 328 Arai, T., 189,402 Bartolini, P., 322 Adam, D., 325 Arai, N., 35 Baselli, G., 256 Adams, R., 19 Arakawa, K., 189 Baskaran, S.V., 64 Adelstein, S.J., 67 Arakawa, H., 146,266 Basmadjian, G., 68 Aebischer, P., 237 Archer, B.R., 135 Affeld, K.,241, 280 Archie, J.P.,Jr., 287 Bassano, D.A., 128 Bassett, L., 34 Agarwal, G.B., 24,28 Arcovito, G., 45,140,152 Bassingthwaighte, J.B., 181,227,230 Agarwal, Y.C., 80 Arduini, F., 392 Batallan, O.R., 188,197 Agatenzi, L., 66,129 Arioa, H., 148 Bates, J.H.T., 217 Aggarwal, S.J., 281,308 Arioka, T., 47 Battista, J.J., 21,44,107 Agris, J.H., 392 Aronson, D.D., 369 Baudoin, G., 172 Ahmed, A.K., 309 Aronson, R.H., 215 Bautro, N., 38 Aiba, H., 343 Arruabarrena, A., 389 Baxendale, R.H., 388 Aine, C.J., 32 Arthur, D.L., 32 Beack, S.H., 235 Aizawa, H., 309 Arts, T., 178,179,351 Beddar, A.S., 56 Aizawa, Y., 304 Aruga, H., 330,365 Begnozzi, L., 48,151 Akamatsu, T., 247 Asakura, T., 251 Behrman, R.H., 86,146 Akatsuka, T., 333 Ashayeri, E., 32,48 Beil, C., 92 Akber, S.F., 63 Ask, P., 204 Bell, D.R., 372 Akihiko, U., 299 Asp, L., 121 Bellet, D., 284' Akimoto, H., 345 Astrahan, M.A., 84 Belletti, S., 71,115 Akisada, M., 99,104,150,333 Atalay, N., 312 Bellotti, J., 41 Akitsu, T., 148 Atsumi, K., 189,245,246,408 Belmin, J., 249 Akiyama, Y., 114 Aubert, B., 111 Ben Said, R., 301 Akkoc, C.C., 303 Aubert, N., 111 Ben-Haim, S.A., 221 Akutsu, T., 248,264 Auer, L.H., 289 Benassi, H., 45,48,151 Alastie, H., 124 Auge, J.H., 266 Bencomo, J.A., 70,117 Albert, S., 109 Augustijn, C., 179 Benedict, S.H., 75 Albin, H.S., 213,337,338 Aukburg, S.J., 220 Beneventi, S., 38 Albrechtsen, 0., 239 Auler, J.O.C.,Jr., 27 Benitez-Read, E., 14 Albright, J.A., 208,210 Aunon, J.I., 166 Benke, T.A., 188 Aldea, G., 180 Ausman, J.I., 18 Benko, H., 386 AIdman, 8., 213 Austin, R., 180 Benton, L.A., 315 Aldrich, J.E., 51 Austin, T., 377 Aletti, P., 49 Avakian, Ts.H., 320 Beran, H., 329 Bernell, S.A., 249 Ali, H., 66 Avanzolini, G.O., 218 Bergamasco, H., 404 Aljama, T., 296,384 Avidor, J.H., 241,262 Bergeler, J., 317 Allen, B.,Jr., 210 Aydin, G., 312 Berger, R.D., 257 Allen, P.D., 147 Ayton, V.T., 15,80 Bergeron, P., 264 Allessie, M., 171 Ayyangar, K., 57,84,120,135 Bergsland, P., 222 Almasi, J.J., 31,85,99 Azario, L., 45,140 Azhari, H., 182 Bergstrom, H., 337 Almeida, D., 277 Azpiroz, J.L., 277 Berk, D., 353 Almond, P.R., 54 Bermudez, N.H., 136 Alpen, E.L., 29 Berne, V., 71 Altemeier, W.A., 232 Babiloni, F., 173 Baca", C., 45,140 Bernabei, R., 140 Altobelli, S.A., 80 Hachler, G., 288 Bernabeu, P., 306 Alves, R.N., 30 Backlund, T., 277 Bernardi, T., 72 Amend, G., 313 Backonja, H., 60 Berner, B., 18 Amols, H.I., 51 Bader, R., 93 Bernini, U., 17 Amor", E., 319 Badiello, R., 72 Bernstein, H.S., 186 Amos, 0.E., 389 Gagne, F.R., 39,42,51,52,123,136 Berry, J.A., 51 Ampil, F., 41 Berry, S.J., 315 Anantha, N., 68 Bally, N.A., 145 Bajzer, Z., 234 Bersani, F., 145 Anderhuber, F., 289 Bertenstam, L., 382 Andersen, 0.S., 157 Baker, R.S., 58 Baldwin, B.C., 51 Berthaud, P., 394 Anderson, G., 408,409 Bertocohi, H., 256 Anderson, L.L., 128,152 Baldzer, K., 256 Bertram, C.D., 259 Anderson, M.P., 106 Ball, H.E., 19 Bertram, C.D., 262 Ando, F., 146 Ball, W., 14 Ballou, S.P., 405 Bertram, C.D., 384 Ando, J., 253 Banci, G., 48 Beszedics, G., 368 Ando, H., 99,104,150 Betts, T.A., 403 Andrabi, W.H., 118 Banci, H., 66 Beuchler, D., 79 Andre, L., 43,56 Bao, S.H., 201 Bews, J., 144 Andreasi-Bassi, F., 152 Bench, J.P., 232 4 3 411 Bey, P., 49 Brinker, R.A., 2% Castellano, A., 100 Boyar, R., 178,182 Bromley, J., 302 Casterline, J., 203 Bezerianos, A., 300,312 Brondo, J., 20,25 Catena, G.C., 403 Bhagyalakshmi, A., 253 Bronikowski, T.A., 224 Cavicohioli, C., 25 Bhandare, H., 122 Bronn, D.G., 42,52 Cecatti, E.R., 95 Matta, A.L., 140 Brooks, D.H., 169 Cedarholme J.C., 103 Bianoiardi, L., 45 Brooks, L.J., 236 Collier, F.E., 228 Bianco, B., 407 Brown, J.H.U., 200,322 Cerra, F., 88 Bidani, A., 217 Brubaker, D., 398 Cerutti, S., 174,256 Bielajew, A.F., 5,22,142 Brummer, H.E., 94 Cesarelli, H., 158,360,392 Biggs, P., 39,52,59 Brunia, C., 34 Cesareo, R., 105,143 Bignall, S., 259 Brunner, H.C., 94 Cevenini, G., 218 Bijhold, J., 45,86 Brusic, V., 201 Chabbani, H.A., 265 Billings, E.M., 78 Bryer, J.B., 83 Chadwick, R.S., 179,349 Bindal, C., 206 Bucciolini, H., 142,378 Chat, H., 285 Binns, W.R., 46 Buchanan, D.S., 61 Chakraborty, D.P., 69,70 Bittner, U., 150 Buckles, D.S., 171 Chambers, !LP., 316 Sivas, A., 76 Budai, R., 160 Chan, E.K.Y., 394 Bizios, R., 249 Buffaloe, G., 237 Chan, H., '16,117,150 Bjarngard, B., 90,119,134,135,136 Buhle, L., 45 Chan, J.I.S., 227 Bjorkholm, P.J., 35 Buist, R., 62 Chan, J.C., 408 Blackburn, i.F., 401 Buizza, A., 330 Chan, H.T., 348 Blacker, H., 399 Bujnowski, S.W., 108 Chandra, I., 80 Blackwell, C.R., 99 Bukovitz, A.G., 46 Blad, B., 111,382 Chandran, K.B., 240,243 Bukowski, B.A., 195 Blake, S.W., 132 Chaney, E.L., 55,79 Buldain, G., 389 Chang, S.C., 49 Bleier, A.R., 106 Bunegin, L., 213,337,338 Chang, S.Z., 201 Blight, A.R., 159 Bungo, H.W., 265 Chang, W.H., 300 Bloch, P., 45,74,133 Burhop, K., 238 Chang, Y., 367 Block, J., 207,301 Burman, C., 91 Chang, Y., 383 Block, R., 16,276 Burns, C.P., 281 Chang, Y.L., 268 Blookeel, E, 393 Burns, P., 10 Chang, Y.L., 306 Bloor, H.I.G., 371 Bush, E., 219 Chaouche, H., 172 Bocci, N., 264 Bustelo, J., 145,311 Chapman, C.R., 397 Bodnar, D.R., 162 Butler, K., 62 Charles, H.K., 410 Boenick, U., 187,319,3i,c Butter, D., 263 Charles, J.B., 265 Boethius, J., 337 Buxton, R.B., 7,92 Charles, S., 275 Boezeman, E.H.J.F., 168 Byrd, C., 355 Charleston, S., 296,384 Bogen, D.K., 247 Byrne, F., 370 Chaudhri, H.A., 74,133,147,148,298,320 Boheim, G., 383 Chavantes, C., 18 Bollea, H., 23 Cadena, H., 296 Chelly, LE., 361 Bolli, R., 361 Cadene, T., 218 Chen, A., 300,305 Bolomey, J.Ch., 394 Cady, R.K., 384,398 Chen, C., 36 Holz, A., 279 Calcote, R.W., 346 Chen, C.S., 54,120 Bom, N., 360 Calderon, A., 325 Chen, D., 26 Bonamini, R., 171 Caldwell, C.B., 89 Chen, G.T.Y., 91 Bonanni, E., 327 Camarena, R., 384 Chen, H., 207,301 Bonincontro, A., 29 Cameron, J., 70 Chen, H., 236 Boone, J.H., 15,63 Cameron, J.R., 60 Chen, H., 246 Boone, J.H., 36,72,87 Cametti, C., 29 Chen, H.H., 53,54 Boonzaier, D.A., 389 Caminal, P., 233,266 Chen, H.Q., 267,284 Boote, E., 75 Cancelli, C., 370 Chen, LLB., 378 Borasi, G., 101 Canfield, T.R., 48,308,346 Chen, J., 396 Borgquist, L., 193 Canilang, E.P., 203 Chen, K., 326 Born, E.J., 55,105 Cantero, A., 269 Chen, L., 208,210 Borovetz, H., 247,249,263 Cantons, H.C., 25,26 Chen, B.C., 29,275,331 Bosone, G., 365 Capalucci, J., 52 Chen, S., 284 Bostrom, U., 318 Cape, E.G., 242,279 Chen, T.C., 53 Boter, J., 193 Capello, A., 218 Chen, X., 276,341 Bouligand, B., 318 Capes, B.L., 348 Chang, C.W., 90 Bourgain, R., 393 Capitanio, L., 368 Chong, E., 12,13 Bourland, J.D., 55,79,170 Cappa, P., 198 Chang, E.Y., 404 Boutillon, 44 Cappozzo, A., 337 Cheng, J., 331 Bovendeerd, P., 178 Caprani, A., 253,306 Chang, H., 276,308 Bowerman, R., 76 Caprihan, A., 14 Chong, P., 119 Boyer, A.L., 22,135 Capriotte, C., 25,77 Chang, P., 207 Boyers, A.S., 103 Carbajal, V.H., 304 Chang, P., 301 Bozovic, Z., 201 Cargill, R.S., 249 Chang, Y.A., 46 Braoale, H., 158,340,360,391,392 Carley, D., 215 Cheong, W.F., 18 Brady, TO., 7,92 Carley, D.W., 214 Cherniack, U.S., 216 Brahmavar, S.M., 17,23,109 Carlile, P.V.,Jr., 68 Chew, P., 183 Brahma, A., 73,142 Caro, C.C., 255 Chiabrera, A., 407 Branca, F.P., 198 Carollo, J.J., 162 Chiang, E., 76 Brandt, H.E., 175 Carpenter, D.A., 361 Chiang, Y., 341 Branstetter, R., 4D. Carrasoo-Sosa, S., 297,304,304 Chiang, Y.F., 321 Bravar, D., 196 Carson, P., 11,76 Chien, D., 7,92 Broitman, C.P., 188,197 Carson, P.L., 362 Chien, S., 353 Brewster, L., 13 Casnooha, S.A., 285 Chihara, K., 364 Brien, B., 236 Casolo, G.C., 378 Chin, L.H., 8,118,122,135 Briggs, W.S., 100 Cassot, F., 278,289 Chinzei, T., 408 Bright, F.V., 403 Castaldo, R., 158,340,360 Chiou, R.K., 79

412 4 '20 Dillon, A.R., 316 Chisolm, G.M., 249 Cunningham, I.A., 16 Dillon, S.M., 167 ChiuTsao, S.T., 128 Cunningham, J.L., 206,209 Dimitrijevic, H.R., 163,387 Chiyotani, K., 304 Curio, G., 339 Din, J., 275 Cho, C., 132 Curry, F., 248 Ding, J., 294 Chougule, A.A., 24 Cygler, J., 138 Ding, X., 343 Christensen, D.A., 403 Ding, Z., 284 Chrucky, R., 313 D'Amato, A., 63 Dinner, U., 182 Chu, J., 13 D'Angelo, L., 151 Dionne, K.E., 243 Chu, J., 39 D'Angelo, S., 140 Disertori, H., 171 Chu, L., 27,389 D'Ermo, G., 48 Dixon, L.B., 89 Chu, M., 336 D'Ottavio, G., 143 Djakovic, D., 201 Chu, R.Y.L., 68 da Costa, R., 281 Dobben, G., 108,379 Chu, S., 349 Dadok, J., 82 Dobelbower, R.R.,Jr., 123 Chu, W., 400 Daftari, I., 120,132 Dohi, T., 333,390 Chu, W.K., 109 Dahl, R.A., 50,99 Doi, K, 89,116,117,150 Chuang, 84 Dale, S., 103 Doiron, A.R., 3 Chubaci, J.F.D., 69 Dankelmen, J., 267 Dolber, P.C., 160 Chui, C., 21,22 Dantas, W., 330 Dominguez, J.M., 233 Chung, C., 383 Dario, P., 404 Dong, G., 311 Chung, H., 132 Das, B.S., 68 Doniwa, K., 202 Chung, N., 182 Das, I.J., 22,95,121 Donninelli, M., 322 Chung, T.U., 283 Das, K.C., 69 Doppke, K.P., 13,123 Chuong, C.J., 347 Deism, K., 246 Dorson, W., 236 Churchwell, D.,372 Datta, N.R., 125 Dostalek, C., 329 Churchwell, A.L., 372 Datta, R., 24,41 Douglas, R., 387 Cinbs, C., 95,130,136 Datta, S., 24,41 Douniwa, K., 231 Cintron, 0., 12 Davet, D., 258,359 Downey, J.M., 181 Cioni, B., 163 David, Y.B., 190,193 Doyle, D.J., 323 Ciraolo, L., 142,378 Davis, C.A., 389 Doyle, L., 37 Civardi, S., 256 Davis, M.J., 350 Drexler, G., 21 Clark, J.A., 140 Davis, P., 248 Drost, D.J., 103 Clark, J.W., Jr., 159,188,230 Davis, R.B., 269 Dawson, C.A., 224 Drzymala, 13,130 Clark, R.E., 239,240 DuclaSoares, E., 33 Clarke, G.D., 81 De Agostini, A., 115 Dujovny, M., 18,40,86,127,276 Clayton, P.A., 49 de Almeida, C.E., xx,30,44,95 Dulikravich, G.S., 410 Cleeland, C., 60 de Almeida, R.I., 194 Dumee, Ph., 356,359 Clement, P., 276 De Backer, G., 68 Dunajski, Z., 34 Clemente, F., 158,360 De Buyzere, M., 234,295 Dunbar, B.J., xix Cliquet, A.,Jr., 388 de Callatay, A., 227 Dung, H.C., 395,400 Cluyse, L., 295 de Carvalho, M., 194 Dunn, D., 103 Coates, G., 225 De, D.K., 64 Dunn, F., 363 Coatrieux, J.L., 277 de Figueiredo, R.J.P., 392 Dunn, R.L., 175,176,177 Coffey, C.W., 118 de Lima, W.C., 330 Dunscombe, P.B., 40,86 Coggins, D., 180 De Scheerder, I., 295 DuPuy, H., 76 Cohen, Ma., 229 De Via, L., 68 Durham, S.J., 249 Cohen, R.J., 229,257 Deane, C.R., 269 Duszyk, M., 352 Colasenti, A., 24 Debande, B., 228 Debourg, C., 193 Dvorak, P., 73 Cole, M., 400 Dwyer, S.J.,III, 6 Collier, J.M., 74 deGuaman, A.F., 43 Dykes, F., 258 Collins, E.M., 373 Deibel, F.C., 39,107 del Pozo, F., 202,274,311 Dymond, E.A., 390 Collins, J.M., 223 Dyro, J.F., 194 Collins, R., 373 Delanghe, J., 234,295 Delano, F.A., 349 Colomines, J., 257,315,382 Eagan, P., 129 Colson, M.A., 361 deLemos, R.A., 223 Delgado, C., 160 Earle, M.A., 127 Colton, C.K., 243 Eberhert, R.C., 410 Contento, G., 160 Dellepiane, S., 392 Ebisawa, Y., 366 Conturo, T., 93 Delouche, A., 356,359 Echsel, H., 63 Conway, B., 88 Delouche, Ph., 256 Demiray, H., 344 Echt, D.S., 172 Conway, J., 37,83 Eckstein, E.C., 351 Cook, L.T., 6 Demoment, G., 76 Dempsey, J.A., 214 Eddy, C.A., 176 Cook, R.E., 398 Edelman, N.H., 215 Cooke, C.D., 100 Deng, X., 289,380 Edholm, P., 103 Cooke, R., 56 Dennett, J.C., 57 Denoth, F., 329 Edmonds, H.L., 170 Cooney, P., 116 DePascale, M.P., 140 Edoliya, T.N., 80 Cooperman, A., 206 DeRoo, T.R., 94 Edwards, E.H., 252 Coptcoat, M., 317 DeSapio, E., 45 Eichmeler, J., 204 Cornhill, J.F., 251 Deslauriers, R., 62 Sicker, B., 150 Courtier, R., 264 Detorie, N.A., 78 Eidbo, E.E., 239,240 Courtney, S.P., 210 Deutsch, M., 46 Eiho, S., 393 Cowin, S.C., 205 Devars, F., 390 Eiju, T., 377 Cox, R., 398 Einav, S., 241,262 Crapo, J.D., 221 Devaru, S., 68 Dhananjaya, M.H., 185 Eisenbarth, J., 89 Craven, J.D., 145 Eisenberg, R.S., 157 Creixells, C., 266 Dhingra, U., 83 Eisner, ILL., 94,103 Cresol, F., 321 Di Hallo, A.M., 48,150 Diamond, S., 253 Ejdesjo, L., 367 Crespo, G.G., 30 Ekaterinaris, J.A., 288,375 Crosthwaite, M.H., 66 Diao, Y., 278 Eklund, B., 199 Crucian, G.P., 398 Dickie, D.L., 206 Diebold, B., 356 Ekstrand, K.E., 122 Cugell, D.W., 260 ElKhatib, E., 15,37 Cui, C., 294 Digby, W., 7 Elad, D., 241,261,262,273 Cui, J., 208 Diller, K.R., 281,308

433 413 Elgendy, Y., 377 Flynn, A., 180 ElHefnaway, A.M., 213 Gao, D., 293 Flynn, E.R., 32 Gao, R., 299 Elias, P., 34 Flynn, N.J., 16,82,206,276 Ellingbaus, G., 228 Gao, S., 332 Fogelson, L., 166 Ellis, K.J., 20,65 Gao, W., 73 Fonstelien, E., 273,295 Elson, E., 352 Gao, Z.H., 275 Fontenla, D.P., 43 Elson, M.K., 79 Gaona, E., 54 Forester, C.V., 62 Emami, B., 13 Garber, J.W., 243 Foster, F.S., 6,9 Endo, M., 75 Garces, R., 53 Foster, W.H., 226 Engelhorn, M.W., 83 Garcia, E.V., 100 Fouke, J.H., 223,303 England, P.J., 81 Garcia, L., 310,312 Foulquier, J.N., 90 English, L., 272 Garcia, P., 264 Foulkes, B., 11 Engstrom, T., 404 Garcin, S., 357 Fox, K., 86 Eppink, R.T., 252 Gardeazabal, L., 389 Fraass, B.A, 12 Epstein, J.W., 46 Gardner, M.J., 60 Frampton, S., 281 Garlick, P.B., 80 Epstein, H.A.F., 218 Franchi, F., 66 Ericsson, A., 110 Gaspar Rosas, A., 283 Franconi, C., 48,75 Erne, S.N., 33,339 Gayer, D.P.,III, 221 Frangos, J.A., 253 Ernst, E.A., 178 Gavriely, N., 260 Frank, H., 111 Gazzani, F., 337 Eskin, S.C., 252,253,285 Fransson, A., 110 Eskinder, H., 349 Ge, W.S., 134 Frazier, 0.H., 204 Estoesta, E., 101 Gehrig, J., 93 Fredberg, J., 223 Estola, K-P., 381 Geiger, D., 290,345,369 Freeman, W.J., 164 Eswaran, C., 218 Geise, R.A., 88 Freeston, I.T., 170 Gen, H., 14 Ettinger, K.V., 20,26,75,98 Frey, C.D., 114,116 Evander, E., 225 Genant, H.K., 207,301 Friedman, M.N., 251 Evans, A.C., 92 Geng, C., 280 Fringuelli, F., 66 Evans, E., 353,355 Geng, J., 207 Frisch, P.H., 369 Evans, M., 209 Gentles, W.H., 71,87 Frize, H., 184,199 Ezquerra, N.F., 100 George, J.S., 32 Frost, J.D.,Jr., 168 George, R.E., 118 Ezzadeen, H.A., 338 Frucht, U., 319 Ezzel, G.A., 128 Gerbi, B.J., 39,94 Fryar, J., 16,104 Gerig, L.H., 3 Fu, C., 28 Gerson, R.K., 100 Fadda, A., 335 Fu, M.K., 87 Gerster, T., 96 Fadell, L.A., 51,52 Fu, W., 294 Gerstmann, D., 223 Th.J.C., 257 Fu, Y., 367 Gfirtner, H., 96 Failinger, A.L., 110 Fuchs, H., 10 Ghaznavi, C., 313 Fajman, W.A., 94 Fujii, K., 254,303,356,364 Ghidoni, J., 189 Falcao, R.C., 107 Fujii, H., 49 Ghosh, A., 98 Fallon, T., 302 Fujiki, T., 333 Giacco, C., 45 Fallone, B.G., 144 Fujimasa, I., 246 Fan, X., 296 Giani, U., 24 Fujimasa, I., 408 Giddens, D.P., 258,288,359,360,372,373,375 Fang, L.Y., 332 Fujimoto, T., 309,318 Giersiepen, H., 240,277 Fang, Z., 292,294 Fujimura, Y., 87 Giesekint, E., 279 Fang, Z., 385 Fujita, H., 117 Gieseler, S.J., 330 Farison, J,B., 20 Fujito, K., 385 Giesey, J.J., 362 Farman, A.G., 122 Fujita, M., 115 Fastenmeier, K., 338 Gigante, G.E., 101,117 Fujiwara, T., 290,360,361 Gilbert, I.A., 303 Fatemi, p., 240 Fukuda, H., 69 Giles, H.R., 82 Fatouros, P.P., 15,93,101,378 Flikui, T., 140 Giles, W.R., 230 Faust, U., 302 Fukui, Y., 231,275 Favella, A., 66 Gillette, P.C., 171 Fukushima, E., 14,80 Fear, P.I., 128 Gillies, B.A., 55 Fukushima, H., 346 Gillin, M.T., 38,130 Fei, D., 93 Fukushima, H., 303 Fei, D,, 378 Gimmon, Z., 307 Fullerton, C.D., xx Giorgio, T.D., 238 Feinberg, B., 184 PUmero, R., 20 Feldxamp, L.A., 82 Gioux, M., 390 kung, B.H., 310 Giugni, U., 71 Feldmeier, J., 47 Fling, S.J., 340 Gizdulich, P., 121,345 Feldmier, J.. 129 Ming, Y,C., 220,225 Felice, C.J., 188,197 Glasgow, G.P., Furlan, L., 256 Glass, D.S., 274 Fennigkoh, L., 184 Furst, H., 328 Fenster, A., 16 Glass, P.S.A., 175,176,177 Furcia, S.S., 172,296 Fenster, A., 45,100 Glaze, D.G., 168 Furukawa, T., 87,146,231,345 Fercher, A.F., 367 Gluer, C., 207,301 Fuse, H., 281 Gobbi, G., 38 Ferguson, A.S., 162 Fuse, T.,291 Gobel, W., 282 Ferreiro, J., 225,357 Futaki, T., 333 Fewell, 88 Gooer, A., 103 Field, '.'Jvadhard, W., 286 C., 21 Gaboriaud, G., 394 Figueiredo, H., 407 Goitein, H., 44 Gagne, R.M., 105,106 Gold, M.N., 402 Filligoi, G.C., 368 Gagnon, D., 149 Finley, C., 364 Gold, R., 34 Galbraith, D.M., 46,124 Fiorillo, A., 404 Ooldson, A.L., 32,48 Galioto, F.H.,Jr,, 190 Fish, R., 359 Goldstein, S.A., 82,206 Gallasch, E., 371 Fisher, A.C., 354 Gomez, E., 389 Gallini, R., 71 Gomez, P., 365 Fisher, T.C., 354 Galloway, G., 97 Fitzpatrick, J.H., 35 Gomez, R., 84 Galloway, R.L.,Jr., 110 Fitzpatrick, P.J., 77 Gomez-Aguilera, E., 202,274 Galvin, J.H., 43,91 Gomez-Cia, T., 269 Flacheneoker, G., 338 Gambarini, G., 25,26 Plum, E., 316 Gonzalez, D.L., 230 Gamboa, C.H., 329 Fleisohman, R.C., 31 Gonzalez, J., 389 Ganesan, S., 18,24,25 Good, W.F., 84 Fletoher, L.S., 409 Ganesh, T., 125 Goodman, G.R., 193

414 r. A COodsitt, M.M., 82 Hallen, 3., 98 Herbst, C.P., 113 Goransson, G., 382 Halperin, H.R., 183,375 Herderiok, E.E., 251 Gordon, D., 258 Halter, J.A., 159 Hering, E., 60 Core, J., 374 Hammarstrom, 0., 404 Hermanne, A., 68 Core, J.C., 6 Han, H., 211,213 Herment, A., 76,359 Gore, R.W., 350 Han, Y., 248,314 Hermida, R.C., 310,312,324 Gormley, L., 34 Hanai, 3., 380 Hernandez, C., 310 Gosling, R.G., 15,358 Hanashima, K.; 25 Herron, J.M., 84 Goss, D., 269 Hancock, S.S., 134,143 Hesse, B., 78 Gotman, J., 168 Handa, Y., 386,387 Heukelom, S., 44 Coto, H., 328 Hangartner, T.N., 105 Higgenbottam, T., 222 Goto, K., 370 Hanks, G.E., 39 Higgins, P.D., 56,123 Coto, M., 290,360,376 Hansen, Ch., 25,26 High, M., 89 Goto, T., 202,231 Hansen, J.3., 263 Hilbert, D., 375 Gotoh, K., 305 Hansen, J.W., 28,99 Hill, J., 60 Gotsman. M., 268 Hansen, T., 190 Hill, M.R.S., 341 Cow, B.S., 384 Hansen, V., 143 Hill, 3.E., 103 Graoovetsky, 3, 301. Hansenkam, J.M., 239 Hinshaw, L.B., 283 Graham, M.M., 181 Hanson, G.P., 30 Hirakawa, A., 201 Gramajo, H.D., 188,197 Hanson, W.F., 135 Hirakawa, S., 146,305 Granda, C.O., 50 Hanya, S., 379 Hiramatsu, 0., 290,360,376 Green, L.S., 232 Hara, M., 333 Hirano, M., 335 Greene, E.R., 358 Hara, S., 146,231 Hirao, Y., 75 Greene, R., 103 Harada, K., 34 Hiraoka, T., 131 Greenfield, M.A., 145 Harada, K., 387 Hirata, Y., 364 Greening, J.R., 97 Harasaki, H., 245 Hirayama, H., 286 Grefenstette, J., 35 Hardesty, R., 263 Hitohings, D., 220 Gregorio, M., 162 Harding, G., 104 Ho, W., 71,87 Gregory, R.C., 59 Harms, W.B., 13 Hoohmuth, R.M., 353 Greitz, T., 337 Harold, M.E., 171 Hocini, B., 96 Griffey, B., 63 Harris, J.D., 209 Hodson, D., 40 Criffey, R., 63 Harris, T.R., 351 Hoekje, P.L., 216 Griffith, B., 263 Harrison, D.R., 81 Hooks, A.P.G., 363 Griffiths, D.J., 231 Hartley, C.J., 357,361 Hofer, E., 294 Griffiths, P.A., 390 Harumi, K., 164 Hoffman, A.S., 239 Grigsby, P.W., 126 Hasegawa, S., 99,104,150 Hoffman, E., 222 Grotberg, J.B., 221,260 Hasegawa, T., 33,61 Hoffman, J.I.E., 180 Grund, F.M., 79 Hasenkam, J.M., 239,277,290 Hoffman, M., 292 Grupen, M., 17 Hashida, R., 252 Hogan, H.A., 206,209 Gu, Y., 211,213 Hashiguohi, S., 323 Hogrefe, A.F., 410 Guan, D.X., 298 Hashimoto, K., 304 Hogstrom, K.R., 54,97,108 Gu, Y., 210 Haslock, I., 302 Hok, B., 367,403,404 Guerrisi, M., 263,264,345 Hastings, M.C., 364 Hoke, S., 42,52,136 Guevara, M.R., 229 Hatazawa, J., 69 Holdsworth, D.W., 100 Guglielmi, J.P., 356 Hausman, R.E., 29 Holenstein, R., 372 Guilbeau, E., 402 Hawkins, E.D., 177 Holgate, R.C., 116 Guiot, C., 370 Hawkins, W.G., 83 Holleran, L.A., 249 Gulrajani, R.M., 232 Hayama, K., 64 Holmes, R.B., 35 Gunasekar, J.S.B., 41 Hayasaki, K., 100 Holowaoz, J., 75 Guntermann, P., 280 Hayashi, K., 248 Holt, R.T., 170 Gupta, B.B., 348 Hayashi, R., 76 Holtmann, P., 280 Gupta, M.M., 29 Haydock, C., 67 Honer, M.J., 86 Gur, D., 46,84 Hayes, L.J., 408,409,410 Homma, K., 110,377 Gurvich, A.M., 333 Hazra, D.K., 80 Hong, J.H., 53 Gussenhoven, W.J., 360 He, B.Q., 77 Hongo, H., 356 Gutierrez, G., 191 He, M., 299,300 Horaoek, B.M., 60,61 Gutierrez, M.A., 296 He, S., 311 Hori, K., 14,268 Guttmann, G.D., 29,147 He, X., 283 Hori, M., 180,254,303,356,364,256 Guzzetti, A., 256 He, Y.X., 318 Horiba, I., 34,100 Gvozdanovic, D., 66 He, Z., 342,392 Horiuohi, K., 189,247 Gysemans, M., 68 Hearshen, D.O., 151 Horiuohi, T., 333,390 Heidlage, J.F., 160 Horowitz, A., 108,379 Ha, B., 225,357 Heimans, J.J., 257 Hosaka, R., 140,150,322 Haaoke, E.M., 216 Hein, I., 359 Hoshimiya, N., 365,386,387 Hacerer, S.A., 278 Heiner, B., 321 Hoshino, K., 131 Haddad, G.G., 258 Heinilx, J., 139 Hosoda, S., 380 Hagekyriakou, J., 320 Helie, J., 301 Hosooka, K., 278 Bagg, R., 124 Hellstrom, L.G, 219. Hossain, F., 78 Haggag, Y., 287 Hel.ums, J.D., 238 Houi, K., 267 Haggman, VOL, 235 Helms, J.M., 394,396 Hover, K.H., 78 Haghighi, S.S., 339 Hendee, W.R., xx Howell, R.W., 7,19,67 Hahn, P., 13 Henderson, S., 74 Howells, M.R., 29 Hakim, A.M., 92 Hendrickson, F., 78 Howland, E., 60 Halama, J.R., 67 Henkelman, R.M., 4 Hrachovy, R.A., 168 Halberg, F., 202,310,324 Henkelmann, G.A.; 92 Hsu, K., 202 Hale, J.D., 255 Nankin, R.E., 67 Hsu, T.L., 341 Hale, M.E., 51 Henry, C., 390 Hsu, W.L., 97 Hall, C.W., 192 Henry, G.W., 225,357 Hu, J., 26 Hall, T., 75 Henry, L.A., 19,65 Hu, Y., 14,138,292 Hallbruoker, C., 269 Henry, P., 189 Huang, D., 127,153 Hellen, B., 219 Herbert, D.E., 6,23 Huang, D.J., 201 435 415 Huang, L., 298 Isogai, Y., 267 Jones, N.B., 158,404 Huang, P.H., 124 Isojima, K., 189 Jones, R., 165 Huang, 3., 300 Ison, K., 317 Jones, S., 258,360 Huang, W., 248 Issakov, A., 183 Jones, S.A., 359,372,373 Huang, X., 304 Itakura, N., 385 Jonson, B., 225 Huang, Y., 173,294,296 Ito, H., 202,203,254,345,384,286 Joseph, J., 60 Huberty, S., 228 It0, X., 247 Joshi, R.C., 125 Huda, W., 31,70,107,117 Ito, 14,, 69 Joyner, 8.w., 160 Hudgel, D.W., 214 Ito, S., 113 Juan, L., 249 Hughes, A., 20 Itoh, K., 365 Juday, R.D., 368 Huizenga, H., 106 Itoh, N., 402 Judy, P.F., 106 Hulse, D.A., 209 Itou, H., 90 Jung, B., 110 Humphrey, J.D., 183,344,346 Itou, H., 267 Juttner, F., 297 Hung, B., 202 boson, R.D., 249 Hung, T., 247 Iwamoto, K.S., 42 Kaohel, V., 204 Hung, T., 263 Iwao, K., 202,231 Kagan, A.R., 41 Hunt, J.W.; Iwata, A., 34,100,231 Kagawa, K., 266 Hunt, H., 13 Iwata, S., 325 Kagiyama, H., 376 Hunter, D.H., 35 Izzo, G., 65,66,143 Kahtoh, S., 365 Hunter, H., 364 Kaitila, P., 321 Hunter, H., 302 Jaohimowioz, A., 400 Kajiya, F., 181,268,290,360,361,372,373,376 Huntoon, C.L., xix Jackson, A.C., 217 Kakihana, H., 99 Huo, J.W., 298,308 Jaokson, P.C., 144 Kalender, W., 35 Hussein, H., 17 Jacobs, J.R., 175,176,177 Kalli, S., 235 Hussein', W., 180 Jaeb, J., 405 Kallman, P., 73 Hussey, D.H., 119 Jaffrin, 14.T., 244,305 Kalmanaon, D., 357 Hutohins, K.M., 127 Jain, Y.K., 68 !Canada, T., 180,254,256,303,333,356,364 Hutohinson, G., 295 Jalife, J., 160 Kambara, K., 146 Hutt, D.A., 215 Jalinous, R., 170 Kamei, H., 63,309 Hutten, H., 244,307,312 Jalonen, J., 324 Kamiya, A., 212,245,247,253,351,352 Huxley, V.H., 217 James, A.E., 93 Kamiya, K., 202,231 Hwang, N.H.C., 241,242 Jamzad, H., 104 Kamm, R.D., 220,223,261 Hwang, S., 387 Janozek, J., 88 Kampp, T.D., 344 Hwang, Y., 315 Jane, R., 233,266 Kan, C., 299 Hwang, Y., 396 Jani, S.K., 119 Kan, W.C., 129 Hyman, W., 376 Jannett, T.C., 175,177 Kanai, H., 49 Hyodo, K., 99,104,150 Jansen, B.H., 166,175 Kanai, H., 285 Hyodynmaa, S., 9,137,139,330 Janssen, J.D., 375 Kanai, H., 382 Janssens, J., 408 Kanai, T., 75 Ialongo, D., 322 JaPP, B., 77 Kanamori, H., 88,101 Ioenogle, W.V., 80 Jarmillo, B., 82 Kanazawa, S., 290,360,361,373 Ichie, H., 387 Jaros, G.G., 197 Kanemoto, H., 281 Idogawa, X., 324,329 Jarvilehto, P., 381 Kanesaka, S., 164 Iellamo, F., 345 Jassawal/a, J., 246 Kaneta, K., 84 Ifukube, T., 365 Joy, H.T., 250 Kaneyoshi, A., 402 Iga, K., 14 Jayapal, V., 25 Kang, H., 369 Iglesias, T, 312. Jayaraman, G., 250 Kang, Z., 241 Iguohi, Y., 385 Jayaraman, S,M., 346 Kano, T., 191 Ihara, T., 169 Jebria, A.B., 390 Kao, H., 97 Ikawa, S., 373 Jefferson, L.S., 190 Kaplan, D.T, 229. Ikeda, H., 49 Jefferson, R.J., 388 Kaplan, R., 55 Ikeda, K., 29,314,406 Jenkins, D., 78 Kapoor, S.L., 139 Ikoro, N.C., 122,128,149 Jenkins, M.O., 122 Karemaker, J.H., 256 Illes, T., 321 Jensen, 0.E., 262 Karhumaki, L., 171,232 Im, J.J., 263 Jeroudi, M.O., 361 Karino, T., 251,289,380 Imaohi, K., 246 Jette, D., 135,137 Karlsson, H., 79 Imai, H., 378 Ji, H., 338 Karlsson, S., 277 Imura, T., 363 Jian, K., 208,210 Kartha, P., 78 Inada, H., 271,290 Jiang, C., 276,341 Kary, H., 301 Inagaki, Y., 378 Jiang, D., 255,334,340 Kase, K., 134 Indovina, M.A., 29 Jiang, P., 326 Kase, K.R., 95,121 Indovina, P.L., 29 Jiang, W., 332 Kass, D., 179 Inglis, R., 200 Jimenez, LA., 219 Kassis, A.I,, 67 Inoue, H., 180,200,254,271,303,356,364,256 Jimoh, A., 276 Katanik, D., 40 Iooco, H., 391 Jin, J., 293,294 Katayama, H., 191 Ip, T.K., 237 Jin, S., 291,362 Katayama, Y., 63 Iramina, K., 34 Jing, Y.Z., 29,331 Katila, T., 32 Irler, W.J., 171 Johansen, A., 273 Kato, S., 331 Isabey, D., 219,223 Johnson. A.E., 310 Kato, Y., 406 Ishida, Y., 180 Johnson, E.C., 358 Katsumura, T., 290,361 Ishida, Y., 256 Johnson, 170 Katsuragawa, S., 116 Iahihara, A., 379 Johnson, K.N., 131 Katz, D., 262 Iahihara, K., 254,303,356,364 Johnson, R.J., 238 Kaufman, J.J., 407 Ishii, I., 298 Johnson, T.K., 79 Kaufman, L., 255 Ishii, N., 235,329 Johnston, G.I., 186,271 Kaul, S.K., 118 Ishijima, H., 203 Jones, C.J.H., 255 Kavadias, M., 163 Ishikaws, K., 271 Jones, D.T.L., 132,133 Kawaohi, K., 75 IShimi, Y., 87 Jones, J., 165 Kawamura, K., 69 Isik, B., 312 Jones, H., 239 Kawamura, K., 250 Islam, ILK., 12,120,124 Jones, H., 240 Kawamura, K., 254 Isobe, H., 147,148 Jones, H.T., 103 Kawano, S., 333 4' 416 Kawarada, A., 202,203,345,384,288 Kodama, H., 329 Kushner, D.C., 82 Kawasaki, S., 148 Kodama, R., 18 Kuske, R., 13 Kawashima, K., 131 Kodera, Y., 87 Kusumi, Y.T., 399 Kay, GA., 175,177 Kodera, Y., 115 Kusuoka, H., 180,256,303,356 Kayashima, S., 402 Koehler, A.H., 131 Kutcher, G.J., 91 Kazda, I., 105 Koeppe, R.E.,III, 157 Kuwahara, H., 393 Kearfott, K.J., 17,102,103 Koh, K., 132 Kvernebo, K., 376 Keddy, R.J., 141 Kohl, F., 400 Kwong, K.K., 7,92 Keenan, R.L., 15 Kohler, J., 242,280 Kehoe, T.M., 42,96,97 Koji, H., 299 Lacey, L., 16 Keith, H.W., 161 Kojima, S., 266 Lacombe, C., 147 Kellaway, P., 168 Kolari, P., 9,330 Laoy, J.L., 19 Keller, B.H., 73 Kolas, Z., 327 LaForge, D., 246 Kelley, B.S., 177 Kolff, W.J., 242 LaFrance, H., 109 Kelly, R., 267 Kolitai, Z., 89 Lagana, A., 143 Kennedy, A., 64 Komatsu, H., 380 Lage, S.G., 296 Kenner, T., 283,297,348,371 Komori, S., 323 Lego, P.J., 158 Kenwright, J., 206 Kondo, Y., 323 Lahtinen, T., 57 Kenyon, C.H., 222 Kondou, Y., 379 Lai, TA., 258 Kepski, R., 292 Kondraske, G.V., 162 Laitinen, L.A., 324 Kesavan, S.K., 224 Kong, S., 314 Lake, J.R., 398 Keshaviah, P.R., 236 Kong, W., 280 Lakshmanan, A.V, 64. Kett, H., 96 Konig, W., 319 Lakshminarayan, S., 317 Key, H., 144 Konisi, K., 99 Lakshminarayanan, A.V., 110,111 Khan, F.H., 39,94,95,105,121 Konishi, K., 104 Lam, C., 114 Khan, F.R., 59,127,152 Kontis, S., 358 Lam, K., 117,150 Khodadadi, J.H., 287 Koornneef, F., 272 Lam, S., 141 Khokhlich, Y.I., 407,339 Kootsey, J.H., 157 LaHanna, J., 400 Khoo, H.C.K., 215 Kooy, H.H., 8 Lambert, F., 369 Kiguohi, M., 87 Kopans, D.B., 35 Lambert, R.K., 261 Kikuchi, M., 47,49,189,194,244,402 Korenberg, H.J., 315 Lammers, W., 171 Kikuti, H., 331 Kormos, R., 263 Lan, F.L., 308 Kilcoyne, R.F., 82 Kornreioh, F., 163 Lan, R.H., 54 Kilpatrick, D., 376 Korpan, N.N., 336,407 Lan, T.B., 321 Kim, T.G., 226 Korrubol, J., 60 Lancaster, J.L., xx Kim, T.H., 39 Kosanetiky, J., 104 Lane, R., 51 Kim, Y.J., 64 Kossoff, G., 11,361 Lang, F., 269 KimmeSmith, C., 34 Kotani, H., 33,51,304 Lange, J., Kimura, J., 402 Kothiyal, K.P., 166 Lange, PA., 79 Kimura, K., 256 Koul, R.K., 118 Langewouters, G., 286 Kimura, S., 69 Kowe, R., 220 Langone, J., 299 King, A.I., 369 Koyama, H., 331 Lanir, Y., 344,371 King, C., 58 Koyano, T., 306 Lanson, H., 44 King, D.H., 358 Kozmann, Gy., 232 Lanting, P., 257 King, G., 32 Kozo, S., 299 Lanzl, L., 78 King, G.A., 128 Kozuka, T., 256 Larsson, B., 5 King, G.C., 48 Kraft, D., 185 Larsson, H., 304 King, R.B., 181,2300 Kraft, K., 378 Latini, P., 38 Kippenhan, J.S., 351 Kraft, K.A., 93 Laura, R., 269 Kirby, T.H., 7 Krajnic, J., 162 Laufe, L.E., 176 Kirohner, H., 171 Krams, R., 371 Laughlin, T.J., 175 Kirkilevsky, S.I., 407 Krasin, F., 60 Laughlin, T.J., 176 Kirkwood, J.R., 109 Kratky, R., 250 Laugier, A., 90 Kirsner, R.L.G., 370 Krawoiw, N., 215 Laurent, J., 188 Kiryu, S., 291 Kremkau, F.W., 77 Lavelle, W.G., 303 Kishore, P.R.S., 15,101 Krishna Hurthy,, T.G.,185,195,272,317 Lawrence, N.B., 238 Kisslinger, A., 24 Krishna Swamy, N., 272 Laxton, J., 51 Kitabatake, A., 180 Kriz, R.J., 108 Lay, J., 263 Kitabatake, A., 254,256,303,356,364 Kriz, R.J., 379 Layne, W.V., 78 Kitakaze, H., 180 Krizaj, D., 386 Lazar, H., 327 Kitasonono, H., 322 Kroll, R.H., 332 Le Gallo, 0., 345 Kitney, R.I., 255,259,265,285,334,358,373,374 Krouskop, T., 364 Leavitt, D.D., 12 Klarmann, J., 46 Krybus, W., 334 LeCarpentier, G.L., 17 Klassea, H.K., 141 Ku, D.N., 377,379 Lee, CA., 268,306 Kleck, J., 7,132 Kubo, H., 98,119 Lee, D., 393 Klein, EJE., 122 Kubo, K., 385 Lee, G.C., 291 Klein, J.L., 100 Kubodera, S., 189 Lee, J., 302 Klein, K.E., xix Kubota, H., 88,101 Lee, J.H., 235, Kleinstreuer, C., 287 Kubota, I., 26 Lee, J.H., 278 Miami°, G., 61 Kudo, N., 405 Lee, J.S., 224,264 Kleviansky, H., 389 Kugimiya, T., 320 Lee, L.P., 224,264,302 Klide, A.M., 396 Kuipers, B.J., 226 Lee, H., 383 Klose, U., 93 Kumar, V.D., 41 Lee, T.Y., 65,85,114 Klotz, L., 18 Kume, Y., 89 Lee, W., 383 Kluksdahl, E.H., 102 Kung, R., 205 Leeds, N.E., 109 Knapp°, P., 164 Kurita, A., 189 Leehan, J.A., 380 Knoche, J., 70 Kuriyama, H., 402 Leetz, H.-K., 58 Knott, E., 240 Kuroda, T., 406 Lefevre, J., 228,230 Knudsen, V.E., 286,290 Kurosawa, H., 212 Lefkopoult, D., 90 Kobayashi, H., 288,384 Kurup, R.G., 52 Lehmann, D., 166 Koch, R.F., 59,127,152 Kusakabe, H., 247 Lehnert, S., 15

431 417 Lei, T., 105 Lizuain, ?LC., 95,130,136 Mahtaboally, S.Q.G., 105 Lai, X., 280 Lo, W.H., 201 Mahyoub, F., 37 Leiohner, P., 83K. Lobriguito, A., 72 Halsey, H.N., 15,80 Leiohter, I., 76,207 Lockwood, G.R., Hajahalme, S., 235 Leightner, D.J., 79 Lodeiro, C., 312 Hajka, S., 109 Lelievre, J.-C., 147 Lofstedt, L., 199,273 Hakila, J., 390 Lemerdy, Y., 357 Logan, S.A., 252 Makino, H., 297,298 Lemieux, R., 161 Lombard, J.H., 349 Hakim), K., 293 Lenner, K.A., 303 Lono, Rog 404 Hakio, A., 325 Lenz, J., 248 Long, D.H., 397 Halaaut, C., 107 Leogrande, M.P., 38 Long,K.,55 Halohesky, P.S., 237 Lepioovska, V., 329 Longobardo, 316G.S. Halezic, H., 162,386 Lash°, 4.G., 410 Lopata, H., 315 Malik, G., 18 Lessard, U.S., 263 Lopes da Silva 165 F.H. Malik, G., 40,86,127 Leszozynsky, K., 46,85 Lopes, J.M.A., 194 Haling D.H., 398 Lever, H.J., 250,255 Lopez, A.C., 219 Malinauskas, H., 364 Levesque, H.J., 254,357 Lorange, H., 232 Kansan, H.R., 160 Levi, A., 405 Lorenz, K.T., 395 Malliarou, 3., 33 Levin,H.,187 Lorenz, W.J., 78,93 Halmgren, L., 111 Levin, S.H., 211 Loring, S.G., 222 Halmivuo, J., 171,232 Levine, R.A., 242,279 Loshek, D.D., 23 Malone, J.F., 116,272 Levy, M., 385 Loshin, D.S., 368 Hanoini, A., 17 Levy, S., 301 Lotter, M.G., 113 Hanoini, A., 24 Lewis, J., 301 Lou, Z., 409 Handava, V., 35 Lewis, P., 32 Lovaund, P., 213 Mandelkern, H., 7 Lewkowioz, H., 179 Low, L., 51 Hanfrin, H., 330 Lewkowitz, S., 183 Low, L.R., 20 Hangiardi, L., 171 Li, J.K., 179 Lowinger, T., 115 Manifold, 1.H., 83 Li, L., 267,284 Lowry, M., 80 Hanley, M.T., 209 Li, H.L., 297.318,319 Loyd, H., 51 Hann, D., 247 Li, S., 314 Lu, D.H., 342 Hanoharan, S.R., 24 Li, W., 284 Lu, J., 291 Hanolis, J., 13 Li, X.Z., 355 Lu, R., 310 Harmon, A., 369 Liao, F., 284 Lu, S., 248 Harohal, C., 49 Liberati, D., 174 Lu, M., 326 Harm, H.L., 182,183 Liboff, A.R., 406 Lunn, Z., 14 Margarita, S., 322 Liohter, A.S., 59 Lucas, C., 225,357 Hargerum, S.C., 402 Lieber, B.B., 374 Lulu, B.A., 126,139 Margulies, S., 222 Liebovitoh, L.S., 229 Lunkenheimer, P., 219 Harhoff, P., 150 Liepsoh, D.W., 377 Luo, J., 328 Marie, F., 369 Ligae, J.R., 218 Luo, Y., 314 Markov, M.S., 407 Lilly, M.B., 176 Lutohen, K.R., 218 Harrett, S., 92 Lin, C., 315 Luthra, A., 402 Marsden, D.S., 115 Lin, D., 30 Lutz, W., 53 Harsh, Lai., 117 Lin, F.4,, 53,54,120 Lux, R.L., 163,232 Martin, 8.B., 396 Lin, J.R., 332 Luxton, G., 84 Martin, L.J., 19 Lin, H., 138 Lysaght, H.J., 243 Martin, H., 12 Lin, P.P., 109 Lysikiewioz, A., 405 Martin, H., 402 Lin, T., 76 Martin, S., 129 Lin, X.Y., 354 Ha, 0., led Martin, S.E., 103 Lin, Y., 202 Ha, S., 27 Martinez, A., 219 Lind, B.K., 22,73 Ha, X., 332 HartinezColl, A., 86 Lindahl, 0., 277 Ma, X.S., 291 Haruhashi, A., 104 Lindberg, B., 197 Haayan, Ch., 214 Masoaro, L., 115 Lindblad, U., 382 Habuohi, K., 408 Hasilamani, V., 18,24,25 Lindell, S.E., 187 Mabuohi, S., 251 Hassaldi, H.A., 307 Linden, Y., 367 Macellari, V., 337 Haseen, C., 34 Lindner, G., 142 Haohado, J.3.198,363 Hastandrea, J., 123 Lindqviat, A., 324 Haoiunes, R.J., 110 Hastrooinque, H., 17 Linehan, J.H., 224 Haokie, T.R., 121,130,133 Hastroolnquel H., 24 Ling, C.C., 92,126,127,153 Haokiewioz, A., 405 Hasuda, H., 254 Ling, Q., 331 HaoMahon, H., 116 Hasuda, J., 384,288 Linnarsson, D., 219,304 Hader, L., 405 Hasuda, Y., 378 Liou, X., 336 Hadhvanath, U., 30,125 Hasuya, R., 87 Lipka, J.H., 210 Madison, J.V., 410 Hasuzawa, T., 333,390 Lipsohultz, A., 187 Madsen, E., 75 Mates, R.E., 180 Litwiler, K.S., 403 Madsen, H.T., 19 )1atheny, H.L., 141 Liu, D.Y., 316 Maeda, K., 64,324,325 Mather, B., 267 Liu, H., 276,341 Maas, H., 408 Hathisen, H.L., 79 Liu, J., 83 Hagalhaes, R.A., 198 Matsuda, K., 377 Liu, J., 396 Mamas, G., 330 Matsuda, H., 345 Liu, L., 391 Hegrini, A., 66,143,263,264.345 Matsuda, T., 47,245 Liu, R., 212 Haguire, G.Q.,Jr., 116 Matsudaira, H., 313 Liu, S.H., 229 Rah, E., 44 Matsui, H., 373 Liu, S.S., 270 Hat, K., 91 Hataukawa, H., 291 Liu, X., 27 Hahaohek, A., 198,271 Matsumoto, G., 405,406 Liu, 7.1 279 Hahady, J., 272 Matsumoto, H., 88,101 Liu, Y., 283,310 Maher, K.P., 116 Matsumoto, T., 259 Liu, Y., 311 Mahler, Y., 268,307,407 Matsumura, H., 309 Liu, Y., 391 Mahood, S., 141 Matnumura, T., 14 Liu, Z., 212 Mahtaboally, M.S., 143 Matsuo, H., 34,100

418 Matsuoka, S., 34,290,360,361,373 Hino, H., 233 Husselman, V.H, 369. Hatsuoka, T., 376 Hintchev, H.P., 335 Hustafa, A.A., 88 Matsushima, J., 365 Hiola, U.J., 26 Hyers, A., 234 Hatsuzaki, Y., 259 Mira, E., 330 Hyers, L.T., 7,75,132 Matsuzawa, T., 69 Hisasi, N., 391 Hyrianthopoulos, L.C., 91 Matthews, F.L., 220 Hishra, P., 92 Matthews, J.W., 13 Hi3sirlis, Y., 391 Naccarelli, G.V., 230 Hattila, M.A.K., 397 Hita, H., 335 Nadeau, R., 169 Hatute, V., 388 Mitamura, Y., 245,278,297 Nagakura, T., 364 Haughan, W.L., 179 Mitchell, T.P., 79 Nagano, T., 266 Maxey, H.N., 192 Hitev, G., 12,120,124 Nagasaka, E., 274 Mayer, D.J., 395 Hito, K., 290,360,373,376 Nagashima, K., 305 Mayr, W., 386 Mitomi, T., 323 Nagata, N., 324 Mazal, D.A., 40 Mitsui, K., 231 Nagata, Y., 137 HazHer, A.K., 288,375 Mitsumata, H., 250 Nagayama, H., 367 McCall, A., 48 Mittelman, A., 76 Nagel, J., 385,391 McCarey, B., 368 Mitcelstadm, S.E., 76 Nagel, J.H., 351 McCarthy, K.J., 16,104 HLura, T., 200 Nahum, A.E., 124,142 HoCawley, H., 226 Hiwa, H., 266 Naili, S., 345 McCormick, B., 13 Miyake, Y., 274 Naito, A., 387 McCraw, F.J., 117 Miyamoto, A., 189 Nakagaki, 5., 87 McCullough, E.C., 50 Aiyamoto, H., 388 Nakagawa, H.. 189 McDaniel, W., 172 Miyamoto, S., 267 Nakajima, K., 324 McFadden, E.R.,Jr., 303 Hiyasaka, K., 191,281 Nakajima, T., 333 McGee, H., 204 Miyazaki, H., 146 Nakamori, N., 116 HoGinlek, P., 55 Hiyazawa, T., 279 Nakamura, F., 377 MoGowan, N.D., 77 Hizicko, H.K., 47 Nakamura, H., 189 MoIntire, L.V., 238,252,253,285 Hizrahi, J., 385 Nakamura, K., 26 HoKetty, M.H., 72 Hizuno, K., 189 Nakamura, T., 248 McMillan, S.T., 240 Hochio, S., 267 Nakano, T., 309 MeHurry, H.L., 55,79 Mockros, L.F., 348 Nakano, Z., 373 HoNeice, G., 210 Moeller, J.H., 12 Nakatani, H., 266 HoParland, B.J., 57 Moeller, R., 43 Nakatani, T., 248 McQueen, D.H., 370 Hoffit, D., 387 Nakaya, S., 81 Mahan, D.L., 12,59,138 Mohan, R., 21 Nakazawa, H., 84 McVittie, R., 55 Mohan, R., 22 Nam, T.L., 141 Hedialdea, H.J., 145,311 Hohandas, N., 354 Nandedkar, S.D, 158. Medina, B.V., 277,297 Hohl, W., 286 Narahara, H., 390 Medina, I., 173 Hohnhaupt, A., 280 Narayanan, S., 125 Hedvick, P.A., 32 Hok, E., 46 Nariainen, K., 330 Meek, A., 85, Holho, N., 25,26 Natan, T.E.,Jr., 247 Heartens, H.,45,86 Moller, D.P.F., 228 Nath, R., 153 Heglio, H., 163 Monos, E., 249 Nath, R., 58 Mehta, B., 40 Montaye, T.J., 61 Natsubori, S., 333 Mehta, H.P., 130 Hontevecchi, F.H., 247 Naughton, D., 272 Heigooni, A., 153 Moore, J.E.,Jr., 379 Navarro, L.T., 252 Heinzer, K., 313 Mora, F., 277 Navona, C.: 327 Mali, J., 153 Mora, P., 70 Kayak, C., 9 Hellenberg, D.E., 50,99 Horan, P.R., 64,100 Nazemi, H., 287 Helo, C.P., 27,296 Moret.ti, R., 115 Neblett, C.R., 399 Meeizabal, J.A., 83 Mori, H., 69 Needham, D., 352 He.,eguetti, J.C., 27 Morita, H., 340 Negishi, N., 49,402 Mercer, C.R., 221 Moriya, T., 291 Nelson, C.E., 56 Herletti, R., 196 Morkrid, L., 383 Nemoto, I., 342 Herrick, H.W., 123 Morris, E.D., 249 Nemoto, I., 393 Merritt, R., 60 Morris, P.G., 81 Nerem, R.H., 252,254,355,357 Messinger-Rapport, B.J., 169 Morris, R.L., 186,281 Nerem, R.N., 372 Hetzler, H., 384 Mortimer, J.T., 162,385 Nerhed, C., 199,273 Meyer, C.R., 33,76,362 Horvan, D., 278,305,289 Neroni, H., 335 Heyer-Ebrecht, D., 334 Moser, E., 63 Nesbit, P., 358 Hichanetzis, G., 391 Moser, H., 283,297 Neuenschwander, H., 43,56,106 Michel, J.B., 249 Hoslehi, F., 218 Neufeld, G.N., 220 Hijnheer, B.J., 44,118,122 Moss, C.A., 394 Neuman, H.R., 401,405 Hikami, T., 278,297,363,405 Hossery, H., 268 Neumann, F., 286 Miklos, G., 22 Hota, H.C., 123 Nevo, E., 371 Hiksicek, A., 187 Houra, L., 334 Newcomb, C.H., 8 Mikulecky, D.C., 235 Houttahar, A., 180 Newcomb, R., 365 Milano, F., 142 Hrkvicka, J., 329 Newhouse, V.L., 358 Hiles, 'R., 164 Muehlbauer, H., 402 Nguyen, P.D., 310 Miles-HeCollum, D., 353 Hukai, C., xix Nibhanupudy, J.R., 32 Hiller, C., 17 Huller, R.G., 37 Nibhanupudy, R., 48 Hiller, C., 23 Munro, P., 45 Nichelatti, H., 145 Hiller, D.A., 208 Hunzenrider, J.E., 13,131 Nickerson, A.L., 199 Miller, P., 246 Hura, A., 198 Niederer, P., 219 Hiller, R., 319 Hurakami, H., 405 Niemierko, A., 44 Miller, R.A., 116 Murata, H., 104 Nievola, J.C., 330 Hills, T., 317 Murphey, C.R., 230 Nikawa, Y., 47,49 Minamitani, H., 279,343,366,386 Murphy, B., 272 Nikias, C.L., 169 Hineo, C., 252 Murray, J.B., 398 Nikiforidis, G.; 300,312 Mini, R., 71 Hurri, L., 327 Nilsson, F., 273 Hinnaar, P.C., 113 Husha, T., 164 Ninomija, S.P., 324,329

439 419 Ninomiya, S., 69 Onoda, T., 335 Payne, D.G., 77 Niroomand-Rad, A., 38,130 Ooneda, G., 250 Pa; Nishidai, T., 137 H., 358 Oota, A., 351 Pearce, 3A., 309,393,394,408,409,410 Nishimura, K., 99,104,150 Ophir, J., 76 Nishimura, T., 258 B trlson, G.D., 83 Orem, J., 215 riarson, J.D., 178 Nishiyama, G., 164 Oro, J.J., 339 Nishiyama, S., 379 Peckham, P.H., 161 Orphanoudakis, S.C., 82 Pedersen, E.H., 239,286,290 Nishizawa, K., 31 Ortengren, T., 213 Hitzan, H., 268,307.407 Pedley, T.J., 222,262,370 Orton, C.G., 6,152 Pedraza, L.J., 105,143 Niwa, M., 285 Oshima, N., 346 Nizin, P., 134 Peifer, J.W., 100 Osian, A.D., 152 Pei, H., 332 Nobre, F.F., 198 Ostergaard, J.H., 239,286,290 Noda, H., 248 Pelisse, F., 369 Ostroy, F., 176 Pelissier, R., 264,278,289 Noda, T., 266,380 Ota, Y., 69 Nolan, S.P., 252 Peng, C.L., 312 Otaki, K., 278 Peng, W., 311 Nollert, M.O., 253 Ott, S.M., 82 Penkoske, P.A., 165,167 Nomura, Y., 333 Otte, V.A., 97 Noon, G.P., 361 Pennington, E.C., 119 Oude Vrielink H.H.E., 350 Norman, A., 41 Penzel, T., 313 Overton, W., 32 Northrop, R.8., 234 Pepino, A., 340,391 Oxland, T., 210 Pepmiller, P.L., 8 Norton, R.L., 303 Oyama, T., 250 Peralta, A.P., 138 Nosaka, T., 346 Ozkaya, N., 354 Nose, K., 223 Pereira, C.A., 278 Nose, Y., 237,245 Pereira, M.A., 363 Pack, A.I., 214 Nouet, P., 90 Perez, C.A., 126 PaesLeme, P.J., 107 Noz, M.E., 116 Perkins, B.H., 176 Pagani, A., 256 Perkkio, J., 282 Numata, K., 320 Pagani, H., 256 Nyenhuis, J., 170 Perktold, K., 375 Page, P., 169 Peronneau, P., 76,253,306,356,359 Nygaard, H., 239,277,290 Paliwal, B., 79,121,133 Nyssen, H., 276,393 Perroche Roux, A.M., 44 PallasAreny, R., 185,257,315,316,382 Nystrom, H., 79 Perry, D.J., 107 Pallikarakis, N., 21,89 Persson, B.R.R., 111 Pallone, T.L., 243 Persson, J., 191,193 O'Brien, K., 320 Pallot, G., 357 O'Brien, P., 55 Persson, R.B.R., 382 Pallotti, C., 145 Peskin, C.S., 370 O'Brien, W.D.,Jr., 359 Pallotti, G., 145 O'Connor, S.A.O., 81,274,281 Peter, J.H., 313 Paloheimo, M.P.J., 170 Peters, T., 90 O'Connor, T.P., 136 Palta, J.R., 50,57,120,135 O'Neill, G., 224,366 Peters, T.M., 140 Pan, X., 255 Peters, V., 143 O'Rear, E.A., 283,310 Pan, Y., 304 Peterson, G.P., 409 O'Rourke, M., 267 Panaccia, A.D., 380 Oamura, T., 112 Peterson, J.I., 401 Panayiotakis, G., 89 Peterson, J.N., 343 Obcemea, C.H., 136 Panerai, R.8., 194,198 Obel, I.W.P., 293 Pettazzoni, P., 145 Pang, T., 296 Pettit P.L., 74 Oberg, P.A., 375,403 Pannike, A. , 200 Pettigrew, R.I., 377 Odagiri, H., 406 Panuccio, P., 321 Oddou, C., 180,290 Pettit, G., 189 Papageorgiou, GA, 404. Pfaff, 3.14,, 71 Ogasawara, Y., 268,290,360,361,373,376 Pareek, S., 149 Pham, T.T., 84 Ogawa, H., 75 Parekh, A., 158 Ogawa, M., 115 Philip, P.C., 59,127,152 Parisi, R.A., 215 Phillips, W., 67 Ogura, K., 342 Park, C.H., 58 Ohayon, J., 179,349 Phillips, H.M., 240 Park, S.H., 235 Photiou, K., 358 Ohlhaber, 7.2., 88 Parker, D.L., 27 Ohnuma, H., 203 Pianta, P.G., 370 Parker, K.H., 255 Pichler, H., 368 Ohshima, N., 254,346,350 Parnis, S., 204 Pickens, D.R., 35,93 Ohsumi, Y., 305 Partain, C.L., 93 Ohta, K., 231 Piechocki, H., 359 Parthasaradhi, K., 53 Ohta, Z., 273 Pierce, W.S., 246 Parviainen, P., 324 Piermattei, A., 45,140,152 Ohtsubo, J.. 377 Pasenl, F., 22 Oikawa, S., 148 Pierrakeas, Ch., 21 Paschoa, A.S., 107 Pierre, P., 353 Ojala, A., 9,330 Pascual, J.M.S., 27 Okada, E., 279 Pike, 8., 90 Passariello, G., 277 Okada, F., 49 Pike, G.B., 140 Passerini, G., 72 Pilla, A.A., 407 Okada, Y., 201 Pasveer, F., 228 Okai, 0., 293 Pillarella, H.R., 237 Pate, B.L., 178 Ping, L. , 87 Okajima, H., 202,231 Patel, B.S., 361 Okamoto, Y., 164,189 Pino, C., 130,136 Paterniti, D.P., 403 Okamura, T., 90,267,373 Pirker, K., 400 Pathak, C.M., 69 Pirola, L., 25,26 Okumura, K.H., 199 Patil, K.M., 318 Okuyama, F., 367 Pisani, M.A., 218 Patti, M.M.400,403 Okuyama, N., 61 Pitchen, I.C., 301 Patterson, H.S., 5,143 Olcaytug, F., 400 Pitluck, S., 74 Patterson, R.E., 103 Olch, A., 41 Piva, R.H.V., 27 Patterson, R.P., 271 Pizziconi, V., 236 Oliveira, M., 114A. Paul, J.M., 59,127,152 Olivier, A., 90,140 Pla, C., 90,56 Paul, J.P., 388 Pla, N., 56 Olsen, K.J., 28,99 Paulsen, P.K., 239,277,290 Olsmats, H.M., 372 Plaszczynsk, R.J., 187 Paulson, D., 61 Plazas, H.C., 42 Olson, A., 89 Pave, A., 227 Olson, L.E., 351 Pleat, D., 17 Pavesi, G., 174 Onai, Y., 84 Plenkovich, D., 9,151 Pavlenko, V.A., 339,407 Onal, E., 215 Podgorsak, E.B., 56,90 Pavlin, C.J., 77 Ono, N., 196 Poirier, V., 246 Pavuza, F., 368 Pola, T., 111

420 440 Pollak, V.A., 334 Rawlinson, J.A., 45 Rolfe, P., 203,204,316,402 Poluta, M.A., 197,293 Raymond, C.J., 262 Romanelli, L., 230 Pomeranz, B., 397 Raymondi, C., 38 Romani, G.L., 4 Ponnalagarsamy, R., 289 Rebelo, H.F.S., 113,114 Romano, H., 211 Pontinen, P.J., 397 Reccia, R., 366 Rompelman, 0., 257 Poopathi, V., 129 Rechenmann, F., 227 Rose, D.F., 33 Popescu, G.F., 36,108,129 Redding, V.C., 404 Rosell, J., 257,315,316,382 Poplawska, W., 292 Reddy, N.P., 203,224 Roselli,;R.J., 216 Porges, S.W., 258 Reddy, S., 78 Rosen, B:R., 7,92 Porter, D., 80 Reddy, S.P., 385 Rosenberg, G., 246 Portner, P., 246 Redpath, A.T., 13 Rosenberg, N., 262 Potter, A.G., 177 Reed, R., 71,87 Rosenman, J.G., 79 Potter, G., 76 Raft, C.S., 96,119 Rosenow, U.F., 12,42,120,124 Potter, R., 390 Regazzoni, C., 392 Rosenthal, H.S., 19,25,46,65,77,111 Potvin, W.J., 20,21 Reggi, H., 264 Rosenwald, J.C., 40,120,394 Pou, A., 351 Rehak, P.H., 384 Ross, C.K., 141 Powlis, W., 45 Rehani, H.H., 112 Ross, J., 138 Pozniak, H., 75 Rehr, R.B., 81 Rossi, B., 340 Prabhakaran, R., 327 Rehwald, W., 259 Rotger, V.I., 188,197 Prahl, S.A., 17,408 Reid, D., 81 Roth, B.J., 62,165 Prasad, G.N.S., 18,24,25,28,64 Reidy, J., 358 Rothenberg, L.N., 31 Roucou, D., 228 Prasad, S.C., 128 Reif, T.H., 242 Roudis, S., 300 Prasad, S.G., 53 Reimann, D.A., 8,151 Rouzaud, H., 90 Prato, F.S., 103 Reinstein, L.E., 51,85 Rowe, R.W., 102 Predergast, F.G., 234 Reneman, R.S., 178,179,350,351,363,375 Pressly, H., 309 Renner, T., 74 Rowlands, LA., 35 Rowley, B.A., 389 Pretorius, P.H., 113 Renner, W.D., 136 Price, R.R., 93 Renzi, R., 142,378 Roy, J.N., 92 Roy, 0.Z., 62 Prieto, F., 218,233 Reul, H., 240,277 Rozenthal, J.J., 30 Prieto, P.S., 89 Reves, J.G,, 176,177 Prinzen, F.W., 178,'79 Reyn, J.W., 260 Rudy, Y., 169 Prohaska, 0.J., 341,400,403 Reynolds, K.L., 79 Rueger, J., 200 Ruiz, E., 173 Proimos, B., 89,300 Ribbe, T., 337 Prokopiou, P., 312 Ribitsch, G., 302 Rumberger, J.A., 182,183 Pronk, R.A.F., 168 Ribitsch, V., 302,348 Runge, V.R., 93 Ruo, J.B., 361 Pucher, R.K., 288,285 Ricci, S., 38 'gliesi, V., 264 Rice, R.K., 8,122,135 Ruperal, R., 53 Russell, H., 52 Purcell, C., 61 Riohardon, P., 190 Russo, P., 17 Purdy, J.L., 4,13,50,134 Richards, M.A., 80 Russo, P., 24,366 Put, J.H.H., 159,339 Richardson, H.L., 82 Puurunen, H., 50,57,139 Richardson, P.D., 236 Rustgi, S.N., 36,108,129 Richter, N., 197 Rutledge, J., 248 Rutten, W.L.C., 159,339 Qi, Y., 304 Riddle, H., 17 Qian, Z., 278 Riddle, M., 23 Ruzyllo, W.,,292 Rylander, H.G.,III, 16 Qiao, Z., 383 Riddle, W., 93 Qin, C., 292 Rideout, V.C., 177 Rymel, J., 37 Qin, J., 285 Rieu, R., 264,289 Sabbah, H.R., 372 Qiu, Y., 188 Rijnders, A., 97 Sabeti, V., 116 Quan, L., 310 Rikner, G., 93 Sachan, V.K., 80 Quarto, E., 17 Riml, H., 151 Sacks, M.S., 347 Quill, T.J., 175 Rindt, C.C.H., 375 Sagawa, K., 179,183 Qing, T., 255 Ritchie, R.G., 178 Saha, P., 192 Quintero, F.J., 153 Pitman, E.L., 182 Saha, S., 192,205,210,378 Rabin, H.S.Z., 131 Rittgers, S., 378 Sahar, H., 262 Rabinovitz, R.S., 357,361 Riu, P., 257,315,316,382 Saidel, G.H., 221,249,401 Rachev, A.I., 268 Rivers, H.L., 146 Sailaja, I., 318 Radford, W.E., 410 Rivers, R.P.A., 259 Sailer, S., 123 Radvan-Ziemnowicz, S.A., 383 Roach, H.R., 250 Saito, H., 73,244,281,314,365,370 Ragan, D.P., 5 Robdeutscher, W., 388 Saitoh, Y., 297,298 Raganella, L., 48 Roberge, F.A., 161 Sakai, K., 306 Raisis, S.K., 107 Roberti, G., 24,368 Sakai, S., 271 Rajaratnam, S., 12,120,124 Roberts, V.C., 185,196,269 Sakakibara, H., 274 Raju, D.V.S., 272 Roberts, W.A., 12 Sakakura, K., 231 Rakfal, 56 Robertson, G.R.J., 73 Sakamoto, K., 49,382 Ramachandran, G., 190 Robilotta, C.C., 27,113,114 Sakata, N., 250 Raman, E., 408 Robinson, D., 123 Sakata, S., 114 Ramaswamy, J.S., 135 Robinson, D.E., 361 Sakthival, S., 28 Rana, H.B., 203 Robinson, L., 358 Sakuma, I., 333,390 Ranger, N.T., 102 Robinson, R.C., 7 Sakuraba, S., 350 Rantanen, J., 9,330 Rocha, H.P.O., 95 Sakurai, Y., 203,388 Rao, D.V., 7,19,67 Rodarte, J., 222 Sala, P., 26 Rao, K.S., 318 Rodellar, V., 365 Saldivar, E., 329 Rapp, J., 255 Roden, D.H., 232 Salerud, E.G., 376 Rashid, H., 8,136 Rodgers, J.E., 36,108,129 Salinari, S., 173 Rastegar, S., 17 Rodrigues, L.N., 95 Salonen, L., 390 Ratchev, I.B., 335 Rodriguez, C.A., 274 Saltzberg, B., 381 Rath, G.K., 125 Rodriguez, J., 173 Samake, G., 290 Rath, R., 125 Rodriguez, L.V., 138 Samaratunga, R.C., 14 Ravelli, F., 171 Roebuck, J., 16 Samsami, N., 39,52,123,136 Ravichandran, R., 68,139 Rogers, D.W.O., 5,142 San, L., 314 Ravindra, P.S., 327

421 441 Sanders, I., 161 Segawa, T., 146 Sanders, H., 118 Shuman, W.P., 82 Seibert, J.A., 15,36,72 Shung, K.K., 362 Sandison, G.A., 31,70,107,117 Seino, H., 69 Sands, S., 399 Shypailo, R., 65 Seki, K., 345 Syu, L.Y., 385 Sanford, H.K., 54 Seki, J., 248 Santello, J.L., 296 Sibata, C.H., 56,123 Sekiguchi, H., 380 Siddon, R.L., 8,90 Santiago, T.V., 215 Sekii, S., 191 Santini, H.T., 28 Sideman, S., 178,182,231,325 Sekiya, T., 331,365 Sapoznikov, D., 268 Siegel, J.A., 3 Selvidge, R., 301 Siegel, J.H., 169 Saranummi, N., 9,330 Semmlow, J., 183 Sarfaraz, H., 38 Sievanen, H., 171,232 Seo, Y., 356 Silage, D.A., 214 Sartori, H., 189 Sepulveda, N.G., 165,172 Sasane, J.S., 41 Silverstein, E.A., 65,113 Settels, J.J., 256,263 Sashin, D., 149 Simizu, K., 1123 Seuntjens, J., 95,234 Sasse, F.J., 177 Stmkin, A., 207 Sevilla, M.D., 406 Simmons, D.J., 205 Sastry, K.S.R., 7,19,67 Sewchand, N., 38,105 Sato, E., 147,148 Simon, S., 355 Shaber, G.S., 15 Sato, K., 191,329 Simonian, A.L., 320 Shackleton, D., 60 Sato, H.,254,346,350 Simonian, M., 120 Shackleton, H.O., 67 Sato, N., 333 Simons, A.J.R., 168 Shaffer, H.D., 184,270 Sato, S., 33 Simons, D., 279 Shaffer, M.J., 184,270 Satomura, K., 189 Simpson, E.R., 77 Shah, R.V., 402 Sim son, L D., 108,141 Sauerbrey, R., 189 Shahabi, S., 121,130,133 Saul, J.P., 257 Simms, C., 37 Shalev, S., 13,46,85,144 Saumarez, R.C., 221 Sims, C., 83 Shan, H., 248 Singh, H., 190 Savard, P., 169 Shang, T., 299 Savilonis, 8., 223 Singh, H., 205 Shankar, B., 166 Sevin°, G.V., 230 Sipkema, P., 371 Shankar, P.H., 362 Savoie, D., 107 Sisterson, J.H., 131 Shannon, D.C., 214 Saw, C.B., 153 Siwek, R.A., 18 Sawada, T., 34 Sha°, J., 326 Sjolund, B.H., 398 Shao, Q., 304,326 Skalak, R., 354 Saxberg, B.E.H., 229 Shao, Q., 356 Sayeg, J.A., 59 Skalak, T.C., 349 Shapiro, A.H., 223,261 Skeen, R.S., 340 Saylor, T.K., 47,79 Sharma, A.K., 139 Skelton, J.P., 100 Scarborough, E.C., 122,128,143 Sharma, R.R., 69 Schaldach, H., 279,294,383,400 Skolnick, L., 399 Sharma, S.C., 122 Skolnick, M., 399 Scheffer, G.J., 257 Sharma, V.K., 129 Skoufis, E.G., 242 Schell, H., 127,153 Snarp, P.F., 20 Schenone, A., 140 Sleet, D.N., 350,351 Shealy, C.N., 398 Slager, C.J., 360 Scherer, P.W., 220 Shee, T.R., 260 Schichi, K., 241,280 Slasky, B.S., 149 Sheehan, M., 127 Slater, C.S., 67 Schils, G.F., 177 Shehab, S.A., 188 Sleek, G.E., 349 Schlienger, M., 90 Shen, C., 202 Slessinger, E.D., 126 Schmekel, B., 225 Shen, F., 326 Sloane, E.B., 195 Schmid, E., 219 Shen, Q., 284 Slump, C.H., 150 Schmid, R., 330 Shen, S., 336 Smarra, N., 142 Schmid-Schonbein, G.H., 349,355 Shenasa, M., 169 Smathers, J., 7,132 Schmid-Schonbein, H., 282 Sheng, J., 211 Saathers, J.B., 75 Schmidt, R.A., 89 Sheng, Z., 381 Schneditz, D., 283,297,302,348 Smith,4., 43 Shepherd, A.P., 282 Smith, Schneider, G.T., 328,332 Sheppard, L.C., 175,176,177,178 Br., 332 Smith,B., 399 Schneider, S., 188 Sher, P.B., 152 Smith,D.N., 382 Schneiderman, N., 385 Sherman, C., 246 Schoephoerster, R., 243 Smith,E.B., 248 Sherouse, G.N., 55,79,140 Smith,G., 93 Schoeppels, S.L., 126 Sherwood, A.H., 163 Schoffa, G., 293 Smith,G.D., 141 Shettigar, U.R., 238 Smith,I., 62 Schreiner, L.J., 82 Sheu, H., 396 Smith,J.H., 229 Schreiner, V., 286 Shi, D., 278 Smith,J.R., 168 Schroeder, P.H., 187,319 Shibata, H., 351,352 Schroter, R.C., 220 Smith,M.A., 15,80,283 Shibata, N., 231 Schuder, J., 172 Smith, H.D., 206 Shibuya, T., 189 Smith, R., 398 Schueler, P., 33 Shields, C.B., 170 Schulz, R.J., 43 Smith, R.B., 91 Shimazaki, Y., 388 Smith, V., 81,153 Schurizek, B.A., 239 Shimazu, H., 202,203,288,345,384 Schuster, J., 63 Smye, S.N., 371 Shimizu, K., 377,405,406 Schuy, St., 289 Smyles, J., 48 Shimizu, H., 260 Schwab, B., 352 Snyder, A.J., 246 Shimizu, S., 309 Schwardt, J.D., 220 Soderberg, T., 277 Shimomura, T., 343 Soltn, L., 13 Schwartz, C.J., 252 Shimooka, T., 245 Schwartzberg, D.G., 94 Solomonidis, S.E., 388 Schweitzer, F.C., 399 Shin, J-14.1 368 Solomonow, H., 387 Shinoda, A., 279 Solzi, P., 385 Schwimmbeok, R., 377 Shipley, W., 39 Scott, P., 210 Sones, R.A., 36 Shiu, A.S., 97,108 Song, Q., 356 Scrimger, J.N., 44,123 Shoji, H., 387 Scurf, S., 247 Sontag, H., 12,13 Shortt, K.R., 141 Sealy, R., 60 Soubra, H., 40 Showalter, D.P., 249 Seed, W.A., 372 Span, J.A.E., 181,267,350 Shozawa, T., 254 Seeman, T., 213 Spach, H.S., 160 Shrivastava, P.N., 47,79,86 Spahn, D., 219 Sefton, N.V., 178 Shu, H.C.S., 241 Segaert, 0., 95,234 Spanne, P., 146 Shukla, A.K., 80 Segal, A., 375 Spelbring, D.R., 91 Shukla, S., 145 Spicer, K.H., 114 422 442 Spiegel, P.K., 93 Sun, J.K., 342 Tanaka, S., 203,247 Spiegelberg, A., 280 Sun, W., 294 Tanaka, T., 266 Spies, S.H., 65,113 Sun, Y., 188 Tanaka, Y., 81 Spies, W.G., 113 Sung, N., 240,303 Tang, J., 28 Sposito, S., 66 Sung, K.P., 353 Tang, L, 293 Sprague, E.A., 252 Suntharalingam, N., 50,57,84,120,135,153 Tang, Y.Z., 336 Spriggins, A.J., 206 Suorsa, V., 259 Tangelder, G.J., 350,351 Srdanovic, V. 201 Supe, S.J., 41 Tani, A., 256 Srigley, J., 18 Suresh, I., 28 Tanishita, K., 306 Srinivasan, R., 265,295 Suresh, S., 28 Tanksale, S.K., 249 Stabile, L., 128 Susak, Z., 385 Tanouchi, J., 254,303,356,364 Staderini, E.H., 101,117 Susanszky, H., 197 Tarbell, J.H., 347 Stafford, P.M., 39 Sutera, S., 353 Tarnawski, H., 15 Stamboliev, I.B., 335 Suwa, A., 104 Tashiro, T., 387 Stancil, P.E., 55 Suzuki, H., 306 Tatsumura, H., 324 Stanley, L.D., 55 Suzuki, H., 275 Taube, D., 269 Stannard, C., 60 Suzuki, N., 90,113,267,366,373 Taveras, J.H., 35 Starita, A., 327,340,404 Suzuki, S., 202 Tawney, K.W., 358 Stark, G., 294 Suzuki, T., 373 Taylor, A.F., 223 Starkschall, G., 92,108 Suzumura, N., 34,100,231 Taylor, G.H., 15 Starr, C., 114 Svedbergh, B., 367 Taylor, H.G., 358 Steckner, H., 103 Svensson, G.K., 119 Tecson, H.R., 101 Steenhoven, A.A.v., 375 Svensson, N., 79 Tedgui, A., 249 Steeves, R., 79 Svensson, H., 213 Ten Haken, R.K., 57,59,127 Stefanini, A., 327 Svensson, 0., 174 ten Hoff, N., 360 Steianovska, A., 386 Swain, P., 317 c.n Voorde, B.J., 257 Steffan, N., 288 Swanson, R., 18 enerz, L., 367 Steiger, P., 92,207,301 Sweeney, J.D., 152 Tenerz, L., 404 Steiger, S., 301 Swensson, R.G., 105 Tepper, J.E., 55 Stein, P.D., 372 Swett, N.A., 10 Terakawa, T., 47 Steinhaus, B.H., 160 Swissa, A., 207 Terashima, Y., 305 Steinke, J.H., 282 Sydanmaa, H., 235 Terio, H., 204 Steinseifer, U., 240 Syed, I.B., 66 Tesmann, G., 144 Steinvil, Y., 385 Symonds, G.R., 73,323 Thakor, N.V., 166,173,174,381 Stekiel, N.J., 349 Szabo, J.J., 54 ThaoChan, C.,355 Stenberg, J., 79 Thatcher, H., 41 Stenzl, W., 384 Tachinardi, U., 172,296 Therrien, P., 13 Stephenson, L., 161 Tacker, W.A., 170 Thews, 0., 307 Stergiopulos, N., 379 Tada, S., 267 Thiagarajan, S, 25.P. Sternick, E.S., 57 Tadaoka, S., 290,376 Thierens, H., 95,234,295 Sternick, E.S., 146 Taenaka, Y., 264 Thoma, H., 386 Stewart, B., 132 Taiganides, P.A., 269 Thomadsen, B.R., 121,130,133 Stewart, D.J., 298 Takahashi, A., 393 Thomas, S.E., 3,14 Stoeckle, N., 172 Takahashi, J., 148 Thompson, C.J., 92,102 Stoehr, N.G., 386 Takahashi, H., 137 Thomsen, S., 18 Stoessel, J., 405 Takahashi, T., 201 Thornilev, H.S., 204 Stojiljkovic, Z., 201 Takaku, Y., 131 Thrope, G.B., 1,51 Stokes, B., 118 Takano, H., 248,264 Thubrikar, H.J., 252 Stokes, S.N., 129 Takano, N., 174 Thull, R., 294 Stolero, D., 241 Takano, T., 252 Thurston, G.B., 283,348 Stommel, K.A., 404 Takao, Y., 100 Thwaites, Da., 13,38,42,96,97,118,124,125 Storchi, P.R.H., 106 Takase, H., 366 Tian, A., 297,318,319,322 Stortzum, R., 108 Takata, T., 304 Tian, W., 280 Strand, S.E., 4 Takatani, 0., 402 Tian, Z., 342 Straton, L., 295 Takatani, S., 264 Tiberio, C.A., 48 Strobel, J.D.,176 Takatsu, H., 305 Tie, B., 183 Stroh, K., 161 Takaya, T., 266 Tiederman, W.G., 240 Strohl, K.P., 216 Takebayashi, S., 250 Tien, A., 83 Stroink, G., 60,61 Takeda, H., 200,271 Titomir, L.I., 164 Strom, I., 318 Takeda, T., 99,333 Tittel, F.K., 189 Strumpf, R.K., 344,346 Takeda, Y., 203 Tkacik, H., 78 Sturaro, F., 392 Takenaka, E., 104,110,140,150,322 Tobias, C.A., 31,85 Su, S.C., 336 Takeuchi, K., 131,189 Toda, N., 325 Su, W.W., 201 Talbot, L., 241 Togawa, T., 191,266,305,309,313,318,323, Sur Y., 203 TalbotPedersen, A., 401 336,401,406,409 Subramaniam, S., 25 Talhami, N., 358 Tohda, K., 254 Sud, V.K., 265 Talusan, I., 101 Tokimura, K., 388 Suematsu, K., 401 Tamamura, T., 14 Tokoro, T., 367 Sugahara, T., 265 Tawas, N.J., 103 Toktas, S., 312 Sugawara, H., 356,374,379 Tamegai, T., 69 Tolley, H.S., 224 SugiharaSeki, H., 354 Tamisier, L., 253 Tomaru, T., 84 Sugimoto, N., 265 Tamura, K., 69,323 Tomaru, T., 377 Sugimoto, T., 377 Tamura, H., 306 Tominaga, H., 304 Sugiyama, Y., 281 Tamura, T., 191,306,309,336,364,40 Tomita, Y., 49 Sul, W., 326 Tamura, Y., 333 Tomonaga, G., 14,268 Suignard, L.J., 361 Tan, R., 188 Tonietto, G., 278 Sullivan, G.S., 32 Tan, S., 62 Tooley, H.A., 274 Sullivan, P.A., 263 Tan, Y., 188 Torres, J.N., 18,189 Suma, K., 266 Tanabe, T., 34 Torrisi, J., 36 Sun, C., 208 Tanaha3hi, Y., 291 Tortal, P., 32 Sun, J., 207 Tanaka, K., 112 Tovar, V.H., 54 443 423 Towe, B., 402 van Eijnsbergen, B., 195 Toyama, H., 104 Wang, H.T., 49 van Hark, H., 45,86 Toyama, J., 325 Wang, P.D., 270 van Hulsteyn, D., 32 Toyofuku, F., 104,150 Wang, S., 52 van /erode, T., 363 Toyoshima, T., 318,323,336 Wang, S., 210 van Nimwegen, Chr., 193 Toyotama, H., 342 Wang, S., 293 Van Wie, B.J., 340,343 Tozerea, A., 347,353 Wang, S., 300,305 van Wier, H., 339 Trahms, L., 33,339 Wang, W., 356 van Wijk van Brievingh, R.P., 228,272 Traissac, L., 390 Wang, W.K., 265,341 Vanhuyse, V., 408 Wang, X., 211 Trajkovic, H., 2C1 Vanryckeghem, W., 68 Traub, R.D., 164 Wang, X., 248 Varia, H.A., 55 Treharne, R.N., 209 Wang, X., 311 Vartiainen, E., 57 Wang, X., 338 Treiber, J., 285 Vasilionkaitis, V.V., 188 Tritthart, H.A., 294 Wang, Y., 212 Vassilakos, P.J., 21 Trontelj, Z., 33 Wang, Y.Y.L., 341 Vaz, C., 174 Truemper, E.J., 190 Wang, Z., 281 Velmurugan, V., 18 Tsai, J., 119 Wang, Z., 334 Velocci, V.V., 71,87 Tsambaos, D., 312 Waring, G., 368 Vercellone, M.I., 188,197 Tsang, N.H., 120 Warty, V.S., 249 Vergroesen, I., 267 Tsao, S., 338 Watanabe, A., 314,370 Verhey, L.J., 73 Tsitlik, J.E., 183,375 Watanabe, S., 69,99,198,331,378 Vernazza, G., 392 Tsuchiva, K., 313,336 Watkins, N., 255 Versnick, F., 276 Tsuda, A., 223 Watson, J.T., xxi Vessella, R.L., 79 Tsuji, S., 333 Watson, P.D., 216 Vesterhauge, S., 369 Tsuji, T., 266,306,309,318,323 Watts, C., 339 Veyrat, C., 357 Waugh, R.E., 352 Tsujimoto, F., 267 Vieira, C.R.S., 198 Tsujioka, K., 181,268,290,360,361,373,37 Weaver, J.B., 93 Vieth, J., 33 Tsukamoto, T., 305 Weaver, K., 127,153 Viggars, D., 13 Tsuzuki, H., 320 Webster, P.M., 260 Viitanen, J., 139 Tu, C.P., 54,120 Weed, H.R,, 269 Vilaplana, F., 145,311 Tucker, D.H., 170,176 Weeks, K.J., 126 Vine, A.K., 59 Tupchong, L., 84 Wei, H., 34 Vinelli, F., 198 Turbes, C.C., 328,332 Wei, J., 276 Vinitski, S., 63 Wei, J., 341 TUrjanmaa, V., 235 Viroli, L., 321 Turk, R., 386 Wei, S., 341 Vishnudevananda, S., 329 Wei, X.Y., 307 Twelves, C.J., 80 Vito, R.P., 344,368 Tyagi, H., 28 Wei, Y., 291,315 nachos, n.s., 287 Weinberg, W.S., 16 Vodovnik, L., 386 Uchida, I., 84 Weinhous, H.S., 47,50 von Eschenbach, A.C., 18 Weinreb, A., 76 Uchida, Y., 377 von Klitzing, L., 144 Uchikawa, Y., 33,61 Weiser, J.C., 46 von Haltzahn, W., 346 Weiss, H., 43 Uchiyama, A., 04 Vossoughi, J., 344,345 Weiss, W., 246 Ucisik, A.H., 206 Vuori, I., 171,232 Ueda, Y., 329 Weizsacher, H.W., 344 Vuthien, H., 357 Welch, A.J., 16,17,18,189 Uematsu, M., 254,303,356,364 Vynckie-, S., 96,132 Ueng, Y.S., 390 Welkowitz, W., 183 Wells, P.N.T., 144 Ueno, S., 34 Wade, S., 34,64 Ueno, S., 387 Wen, X., 275 Wada, T., 115 Wen, Z., 284 Ulin, K., 57 Wada, T., 235 Werner, B.L., 22 Ullrich, C., 319 Wade, Y., 140,290,360,376 Ulmer, W., 80 Werner, E., 25,26 Waggener, R., 129 Werner, I.D., 67 Unsworth, A., 302 Wagner, L.K., 88 Urbano, A., 173 Wesseling, K.H., 256,263,286 Wakai, R.T., 60 Wessels, B.W., 79 Urbina-Hedal, E.G., 304 Walcerz, D.B., 308 West, D.J., 15 Urie, H., 131 Walczak, F., 292 Westerhof, N., 371 Urrutia, T., 12 Wald, A., 186 Usha, T., 289 Westphal, D., 277 Walding, D.L., 190 Heyman, A.E., 279 Usui, S., 274,325 Walker, J.D., 240 Utsumi, A., 189 Whelpton, D., 273 Walker, P., 241 Wheeler, L. 355 Utsunomiya, T., 328 Walker, S., 135 Uyama, C., 264,265 Whiting, J.S., 71 Wallace, R.E., 74,133 Whitmire, D.R., 316 Wallace, R.E., 75 v.d.Voorden, E., 34 Whitsett, T.L., 283 Walravens, N., 68 v.Harsdorf, S., 33 Whittington, R., 12 Walsh, C., 260,263 Whittle, H.W., 388 Vaananen, A., 50,57 Wambersie, A., 96 Wible, B., 252 Valdezco, E., 72 Wan, X., 213 Wickham, J., 317 Valentinuzzi, H.E., xx,173,188,197 Wang, A., 342 Valeri, L., 66 Wickramasinghe, Y.A.B.D., 204 Wang, B., 194,196 Wieck, B.-0., 197 Valimaki, I., 324 Wang, C., 280 Vallbona, C., 200,322 Wiederhecker, N.G., 198,330 Wang, C.Y., 227 Wieme, R., 295 Vallverdu, H., 266 Wang, H., 77 Valvano, J.W., 408,409 Wier, W.G., 157 Wang, H., 276 Wiesen, G., 58 van Aswegan, A., 113 Wang, H., 284,308 Wigod.uy, R., 192 van Beek, J.H.G.H., 230 Wang, J., 183 Van Dam, J., 97 Wijesinghe, R.S., 159 Wang, J., 283,292,310,343 Wikswo, J.P.,Jr., 61,62,159,165,172,232 Van den Winkel, P., 68 Wang, J., 381 Wilcox, B., 225 Van Der Linden, C., '387 Wang, L., 22,108 Van der Plaetsen, A., 95 R.G., 398 Wang, L., 234 Wilkinson, D.A., 47,79,82 van Dijk, L.C., 371 Wang, L., 338 Van Dyk, J., 8,91 Wilkinson, D.J., 274 Wang, H., 26 Willerson, J.T., 182 van Egmond, F.C., 360 Wang, H., 331 Willett, C., 39 424 44 Williams, G., 21 Yamashita, T., 13): Zabel, H.J., 93 Williams, J., 363 Yamato, J., 298 Zachary, J., 359 Williams, J.R., 118 Yamauchi, K., 200 Zagzebski, J., 10,75 Williamson, J.F., 53,58,153 Yamazaki, K., 365 Zahalak, G., 352,353 Williamson, S.J., 61 Yan, J.Z., 298 Zaki, M., 212 Willis, C., 70 Yan, P.Y., 134 Zamenhof, R.G, 86. Nilson, D.L., 37,121 Yan, X., 343 Zamorano, L., 18,40,86,127,276 Nilson, T.A., 222 Yana, K., 174,233,367 Zanchin, C.I., 330 Wilson, T.A., 261 Yanagihara, Y., 265 Zanocco, P., 330 Windedahl, L., 141 Yanagisawa, T., 147,148 Zborowski, M., 237 Windham, J.P., 8 Yanez, S.O., 297 Zeggwagh, G., 384 Windham, J.P., 20 Yang, F.S., 234 Zeigler, B.P., 226 Windham, J.P., 151 Yang, J., 299,326 Zeindler, C., 250 Winfield, J.A., 128 Yang, N., 83 Zeltser, R., 262 Winningham, D., 364 Yang, R.F., 307 Zeman, V. , 301 Winter, D., 223,267 Yang, S., 391 Zemskov, V.S., 339 Wisoff, P.J., 188 Yang, T.S., 306 Zhang, C., 14,138 Wist, A.O., 15,100 Yang, W., 409 Zhang, C., 293 Wit, A.L., 167 Yang, Y., 16 Zhang, D., 335 Witkowski, F.X., 165,167 Yang, Y., 77 Zhang, H., 138 Wittkamper, F04., 118 Yang, Y., 326 Zhang, H., 210 Wolf, N.B., 216 Yang, Z., 26 Zhang, L.R., 267 Wolin, A.D., 303 Yang, Z., 248 Zhang, S., 14 Woll, J.M., 243 Yano, F., 324,329 Zhang, S., 340,343 Nollin, M., 41 Yao, J., 275 Zhang, T., 343 Nolimer, P., 225 Yarmush, H.L., 243 Zhang, W., 342 Wainer, E., 286 Yasuda, H., 286,363 Zhang, Y., 293 Wolton, R., 402 Yasuda, N., 81 Zhang, Y., 297,299,300 Wong, G., 40 Yasuda, Y., 148 Zhao, H., 68,112 Wong, G.R., 311 Yasui, S., 26 Zhao, P., 342,392 Wong, J.W., 13,23,46,85,134,143 Yates, F.E., 315 Zhao, W., 299 Wong, R.K.S., 164 Ye, G., 285 Zhao, X., 314 Hong, W., 102 Ye, Q., 264 Zhao, Y., 284 Wood, L.M., 281 Yeap, C., 410 Zhen, D.S., 201 Wood, M.L., 93 Yen, M.R.T., 225 Zhen, Y., 201 Woodard, J.C., 384 Yester, M.V., 69 Zheng, E.X., 312 Woodruff, E.A., 234 Yeung, A., 355 Zheng, Q., 328 Wooten, S., 343 Yeung, I.W.T., 65,114 Zheng, X.L., 312 Wright, A.B., 62 Yeung, K.T.D., 84 Zhong, G.H., 284,311 Wu, A., 60 Yi, J., 313 Zhong, J., 327 Wu, B., 334 Yie, C., 263 Zhong, Z., 304 Wu, C., 396 Yin, F.C.P., 183,344,346,375 Zhou, G., 298,313 Wu, J., 27 Ying, H., 176 Zhou, H., 311 Wu, J.S.S., 300 Ying, X., 85 Zhou, K., 337 Wu, S.G., 291 Yip, M., 25,77 Zhou, L., 332 Wu, Y., 117 Yip, Y.P.L., 25,77 Zhou, S., 300 Wu, Y., 311 Yoganathan, A.P., 240,242,276,303,361 Zhou, S., 336 Wu, 307,311 Yokoi, A., 202 Zhou, S.H., 373 Wurzinger, L.J., 282 Yokoyama, H., 63,309 Zhou, X., 284 Wyman, D., 143 Yonas, H., 84 Zhu, A., 248 Yoo, S., 132 Zhu, D., 208,210 Xi, B., 279 Yoon, T.S., 235 Zhu, H.S., 332 Xi, T., 280 Yorke, E., 126 Zhu, J., 68 Xia, Z., 278 Yoshida, H., 290 Zhu, M., 336 Xian, R., 299 Yoshida, T., 47 Zhu, W., 293 Xie, J., 27 Yoshida, Y., 256 Zhu, X., 279 Xie, W., 336 Yoshida, Y., 250,254,303,356,364 Zhu, X., 326 Xiong, M., 27 Yoshimura, T., 366 Zhu, Y., 173 Xu, G., 311 Yoshine, K., 236 Zhu, Y.P., 22 Xu, J., 381 Yoshioka, Y., 81 Zimmer, A.M., 65,113 Xu, N., 234 Young, D.F., 379 Zingg, W., 178 Xu, X., 173 Young, S.T., 265 Zong, H., 316 Xu, X., 89 YoungLee, 53 Zoreda, J.L., 202 Xu, Y., 208,210 Yu, B.C., 357 Zue, L.W., Xue, Z.N., 267,284 C.X., 23,134,143 Zydney, A.L., 237 Yu, G., 341 Yaffe, M.J., 89 Yu, G.R., 201 Yagi, Y., 305 Yu, J., 298 Yaginuma, T., 330 Yu, L., 242 Yakar, D., 40,127 Yu, L., 316 Yamada, K., 325 Yu, L., 342 Yamada, Y., 112 Yu, T., 339 Yamaguchi, M., 266 Yu, Z., 305 Yamaguchi, T., 245 Yuan, A., 112 Yamaguchi, T., 250,380 Yuan, Y.W., 362 Yamaki, M., 26 Yuasa, H., 87 Yamakoshi, K., 202,203,212,247,345,384,402,288 Yudelev, M., 132,133 Yamamoto, K., 363 Yui, N., 114 Yamane, T., 250 Yuichi, K., 299 Yamanoi, H., 325 Yuta, T., 278 Yamanoi, T., 212

445 425