Imaging Informatics: from Image Management to Image Navigation

O. Ratib Department of Medical Imaging and Information Sciences, University Hospital of Geneva, Switzerland

Summary vide the necessary tools for radiologists to Introduction review them and make adequate diagnosis. Objective: To review the rapid evolution of imaging informatics In the early eighties, emergence of These early home-grown systems were dealing with issues of management and communication of digital digital imaging in medicine brought a clumsy and expensive, but they lead the way images starting from the era of simple storage and transfer of im- new challenge to computer scientists to to the concept of a filmless radiology. ages to today’s world of interactive navigation in large sets of multi- find ways of managing increasingly large dimensional data. volumes of image data in digital format. Methods: This paper will review the initial concepts of Picture The conversion of conventional Xrays The Pioneers Archiving and Communication Systems (PACS) and the early devel- from traditional films to digital detec- opments and standardization efforts that lead to the deployment of tors generated very large size images that Early adopters of this new concept de- large intra-institutional networks of image distribution allowing required adequate display technology that veloped several innovative and unprec- radiologists and physicians to access and review images digitally. was beyond the capabilities of standards edented systems that were tested in the With the deployment of PACS came along the need for advanced CRT at that time. Conversely, a rapid field and in real clinical practice. Samuel tools for image visualization and image analysis. evolution of scanners technology raised J. Dwyer lll, an electrical engineer work- Results: Review of the history of PACS and Imaging Informatics the number of images and the resolution ing on digital imaging, who was one of clearly shows that the early developments were focused on the that can be acquired resulting in a huge the co-organizer of the landmark PACS

radiologist’s requirements for diagnosis and image interpretation. growth in data volumes. This rapidly conference in Los Angeles, oversaw the These early developments lagged behind the rapid adoption of digi- growing field led to tremendous interest building of what is probably one of the tal imaging in areas outside radiology. Only recently, imaging among the scientific and research com- first PACSat the University of Kansas. informatics shifted toward the development of new tools geared munity. The earliest reference to PACS Based on connecting ultrasound machines toward the needs of other users such as surgeons, referring physi- that we know of came in 1979 when pro- and film digitizer in an Ethernet network cians and care-providers, and even for the patients themselves. Also fessor Heinz Lemke, PhD, at the Tech- connected to a handful of workstations in the recent years, the development of multimedia and communica- nical University of Berlin published a served as a demonstration system and a tion tools in the consumer market has influenced the design and paper on applied image processing and proof of concept. At the same time, Dr. strategic development of image management platforms inside and computer graphics methods in a study Steven Horii who also attended the 1982 outside healthcare institutions. of head CT scans [1]. In his paper he PACS conferenceand who was then at Conclusions: The focus of imaging informatics has clearly shifted in described all the basic components in- New York University, had been working the last decade from basic infrastructure design to complete data cluding interfaces to hospital informa- on a project to create side-by-side digi- and image navigation systems. While the challenge of storing and tion systems. Soon after, a first pivotal tal displays of nuclear medicine head managing large volumes of imaging data have slowly vanished event was a SPIE/PACS conference held scans and head CTs. Due to the difficul- with the rapid development if information technology, the new chal- in 1982 in Los Angeles under the initia- ties he encountered in reading encrypted lenge emerged from the new requirements of image manipulation tive of Dr. André Duerinckx and Samuel proprietary image formats of the CT and analysis in clinical practice. J. Dwyer III, PhD, internationally rec- studies he soon realized that a more open ognized pioneers and leaders in this standard for images would be necessary. Keywords emerging field [2]. It was at this confer- He later became instrumental in standard- Medical imaging, imaging informatics, PACS, image analysis work- ence that the acronym PACS for Picture ization efforts and one of the promoters stations Archiving and Communication Systems of the DICOM initiative (see below). At was first cited and used ever since. A UCLA another pioneering effort was Yearb Med Inform 2009:167-72 recent article published in 2003 [3] re- under way that came to fruition between ports a transcript of the original work- 1989 and 1991 under the leadership of shop of 1982. In the following years this H.K. “Bernie” Huang, DSc, FRCR, that new discipline has slowly matured had come to UCLA in the early 1980s through different early projects in the US and started a new medical imaging divi- and in Europe aimed toward the devel- sion and a graduate program in medical opment of systems capable of handling physics [4]. Prof. Huang put his gradu- large volumes of imaging data and pro- ate students, who included myself, to

IMIA Yearbook of Medical Informatics 2009 168 Ratib

work on building a homegrown PACS. J.R. Scherrer the creator of the Diogene tween academic centers and industry. It was initially deployed in pediatric ra- hospital information system [8], Dr. O. With this trend came the urge for stan- diology and used early low-resolution Ratib and his team developed their own dardization efforts allowing multi-ven- monochrome monitors. It was built image management system based on a dor systems to interconnect. around the first two CR units in the distributed architecture [9] with an inno- United States allowing direct capture of vative imaging workstation called Osiris digital Xray images. More significantly, based on interactive graphic user interface from a manufacturing source he obtained (GUI) that was developed to be portable on Emergence of Standards two computer boards that enabled him different operating systems such as Unix, The need for a standard allowing inter- to decode the digital information on the Microsoft Windows and Apple computers connectivity and exchange of images CR plates. This made it possible to dis- [10]. Other projects were also attempted in between devices from different manu- play the x-rays on the PACS monitors. other countries such as Austria, France, Italy facturers led the American College of Later, UCLA became one of the first sites and United Kingdom [11]. In the latter, some Radiology (ACR) and the National Elec- to show the use of high-resolution moni- ambitious projects were envisioned as early trical Manufacturers Association tors for PACS workstations capable of as 1982 were plans to create the first filmless (NEMA) to form a joint committee in displaying images at full resolution of hospital in the United Kingdom at St Mary’s order to create a standard method for the 2000x2500 pixels and a dynamic depth Hospital in London were submitted by Dr transmission of medical images and their of 10 bits. At the University of Florida, J.O.M.C. Craig of St Mary’s with the Re- associated information. This committee, Gainesville, Janice Honeyman-Buck was gional Scientific Officer, Harold Glass.In formed in 1983, published in 1985 the also struggling in the early 1990s to com- same time in France a very early PACS ACR-NEMA Standards Publication No. plete a limited PACS. With her group project (called SIRENE) was initiated by 300-1985s. While the initial versions of she attempted to break the code and ex- Dr J.M. Scarabin, a neurosurgeon in the ACR-NEMA effort (version 2.0 was tract CT and MR image data from the Renne’s University Hospital and Dr R. published in 1988) created standardized scanners way before DICOM standard Renoulin, an engineer working in the field terminology, an information structure, came along, to store these images in a of enterprise networks in a Telecommuni- and unsanctioned file encoding, most of central archive built on common juke- cations Research Center in Rennes. The the promise of a standard method of com- box of large optical disks. EuroPACS association (http://www. municating digital image information At the same time a small group of early europacs.org/) supported by all the early was not realized until the release of ver- adopters in Europe have initiated sev- leaders of these projects has continued to sion 3.0 of the Standard in 1993. The eral exploratory projects [5]. Professor pursue its original mission of promotion and release of version 3.0 saw a name change, A.R. Bakker at the university of Leiden education of new concepts in PACS and to Digital Imaging and Communications was one of them. Together with Profes- medical imaging and continues today to be in Medicine (DICOM), and numerous sor Heinz Lemke from Berlin and a hand- one of the leading international organiza- enhancements that delivered on the ful of others they created the first PACS tion in the field holding annual meetings promise of standardized communications society “EuroPACS” in 1984. Bakker and and educational seminars and workshops. [13]. The DICOM Standard now speci- his co-workers at BAZIS in Leiden car- The early projects suffered from sig- fied a network protocol utilizing TCP/ ried out PACS research and development nificant limitations in performance and IP, defined the operation of Service as part of hospital information systems were only able to handle a limited num- Classes beyond the simple transfer of (HIS) activities. BAZIS, a non-profit or- ber of simultaneous users and relatively data, and created a mechanism for ganization, developed and supported an small volumes of data. Complex image uniquely identifying Information Ob- integrated hospital information system management mechanisms and routing al- jects as they are acted upon across the widely in use in hospitals in the Nether- gorithms were created for pre-fetching and network. DICOM was also structured lands. In 1986, Dr. M. Osteaux and his pre-loading image files on selected work- as a multi-part document in order to team at the University Hospital of Brus- stations [12]. Users were bound to follow facilitate extension of the Standard. sels together with the Pluridisciplinary predefined distribution rules and workflow. Additionally, DICOM defined Infor- Research Institute for Medical Imaging With such complex communication mod- mation Objects not only for images but Science (PRIMIS) started one of the few els these systems had limited scalability and also for patients, studies, reports, and multi-vendor PACS installations world- could not easily adapt to different institu- other data groupings. With the enhance- wide [6]. At the same time at the Uni- tions and were soon abandoned. ments made in DICOM (Version 3.0), versity Hospital in Geneva a new PACS These early prototypes developed in the Standard was now ready to deliver architecture with distributed archive and centers in Europe and in the United States on its promise not only of permitting the interactive image visualization worksta- were soon challenged by emerging com- transfer of medical images in a multi- tion [7] was developed as an extension mercial systems. In many early imple- vendor environment, but also facilitat- to the existing hospital information sys- mentations joint developments and col- ing the development and expansion of tem (HIS). Under the leadership of Prof. laborative efforts were initiated be- picture archiving and communication

IMIA Yearbook of Medical Informatics 2009 169 Imaging Informatics: from Image Management to Image Navigation

systems (PACS) and interfacing with institutions in the United States, in Eu- versity of Pisa under the leadership of medical information systems. rope and in Asia. In 1993 Dr. Eliot Siegel, Professor. D. Caramella. In Asia, follow- In the mid nineties, all vendors started now vice chairman and professor of di- ing early pioneer work starting in 1984 to adopt DICOM standard in their prod- agnostic radiology at the University of in Osaka university and in 1988 in ucts facilitating the development of PACS Maryland was the first to take on the Hokkaido university, the first large scale and the exchange of images in digital challenge of building a totally filmless PACS projects started in the early nine- form between modalities and interpreta- hospital. At the same time Steve Horii ties. In 1994, at Osaka University, Pr. K. tion workstation. “Soft copy reading” on and his team who were heavily involved Inamura and his team installed one of computer screens slowly replaced tradi- in the development of the DICOM stan- the first clinically operational PACS tional film reading in most radiology de- dard also initiated a collaborative effort implemented in an environment where partments. With the development of with industry to implement a commer- HIS and RIS were already operational. PACS came the need for better workflow cial system at the University of Penn. After a rapid growth in number of PACS management and efficient delivery of in- Professor H.K. Huang who joined Pro- in the following years a recent review of formation and images where needed by fessor Ronald L. Arenson, chairman of the history of these installations show that the radiologists and physicians. Differ- the radiology department in the Univer- a number of projects were abandoned due ent scenarios were applied in different sity of California at San Francisco medi- to poor performance and inadequate settings resulting in very heterogeneous cal school, also helped building a clini- quality [16]. With the advent of new and often inefficient and complex sys- cal PACS while continuing innovative commercial systems and higher network tems. In order to streamline their prod- research and developments in imaging bandwidth and better workstations a rapid ucts and unify their strategies, vendors informatics [15]. Soon after, in 1998, growth in the number of institutions needed more specific guidelines and de- under the leadership of Prof. O. Ratib implementing PACS between 1997 and tailed specifications on clinical work- who was appointed as vice chair of the 2002 show an exponential growth curve. flows. This lead to the IHE initiative for department of radiology at the Univer- In Korea a rapid deployment of PACS in standardization and documentation of sity of Los Angeles Medical Center, public and private hospital was driven image management workflow in clinical UCLA migrated from two generations by a government inventive policy favor- settings [14]. Integrating the Healthcare of home-built PACS to a commercial ing institution that would implement Enterprise (IHE) is an initiative by care system with one of the first off-site long PACS systems and favor private indus- providers (including ACC, HIMSS and term archive based on an Archive Ser- try that would invest in PACS, resulting RSNA) and vendors to improve the way vice Provider(ASP) model [7]. An in- in several local small industry to invest information systems communicate to novative approach where the responsi- in numerous large full-scale PACS support patient care. IHE defines Inte- bility of long term storage of images was projects. In the recent years a large na- gration Profiles that use established data outsourced to a vendor providing large tionwide PACS and teleradiology net- standards to integrate systems for effec- capacity storage service based in data work project was launched in Hong Kong tive interoperability and efficient work- centers located remotely from the medi- by the Hon-Kong Hospital authority IT flow. IHE makes it possible to achieve cal centers. A similar approach was department in collaboration with Prof. the level of integration required in the adopted by Northwestern Medical cen- H.K. Huang team. The Hong Kong Wide era of the electronic health record. Each ter in Chicago under the leadership of Area Image Distribution/PACS Project IHE Integration Profile describes a clini- Dr. David Channin who was responsible covers 44 public hospitals and 51 spe- cal requirement for systems integration of enterprise-wide image management cialized outpatient centers. and a solution to address it. It defines and PACS with a strong commitment to functional components, called IHE Ac- IHE guidelines. In Europe, in Graz Aus- tors, by specifying in careful detail the tria, following pioneer work of Profes- Academics Pushing the Limits transactions, based on standards such as sor G. Gell on PACS and RIS integra- DICOM and HL7, each Actor must per- tion, and ambitious project for a fully While industry slowly penetrated the form. IHE provides specification for a more filmless hospital based on a commercial market with commercial PACS solution unified framework for implementation of PACS was launched in the new Danube that responded to the initial needs for integrated PACS with HIS and RIS. Hospital (formerly SMZO) in Vienna un- image management and image interpre- der the leadership of Professor Hruby tation on digital workstations it lagged chairman of the department of radiol- behind the rapid increase in demand and From Prototypes to ogy. In Italy after some pioneer work changes in clinical requirements. This strongly geared toward teleradiology at gap between commercial systems and Professional Systems the university of Trieste, and university the clinical demand was due to the lack As commercial PACS systems started to of l’Aquilla, a collaborative effort with of clear guidelines and specifications appear on the market they were soon industry for implementation of PACS and from the medical community due to adopted by some early adopter in large teleradiology was undertaken at the uni- unanticipated changes in paradigm and

IMIA Yearbook of Medical Informatics 2009 170 Ratib

accelerated growth in utilization and need where they are needed. This change advanced and innovative context-based development of imaging techniques in in paradigm was also supported by the image retrieval (CBIR) based on ontol- clinical practice. This growth and rapid rapid evolution of web-based technolo- ogy were developed [21,22] to enrich change in requirements also coincide gies allowing faster data streaming of very the core functionalities of picture with a drastic change in most industri- large data sets as well as facilitated access to archiving and communication systems alized countries were healthcare policies high performance data processing through [23]. The goal of such systems being to were heavily driven by economic con- thin-client web-based front ends. access and retrieve images by content strains and explosion of costs driving Another area where academic research rather than by alphanumeric-based indi- governments and agencies to impose new has driven new changes in design of im- ces. Although this approach was origi- management strategies driving toward age management and PACS solutions is nally developed for multimedia reposi- higher productivity at lower costs. in the area of image analysis and image tories such as those on the World Wide Research for innovative and more processing. Innovative techniques are Web, techniques for content-based ac- efficient solution was mostly carried in slowly appearing in the field of com- cess to medical image repositories are academic institutions where close inter- puter-assisted diagnosis providing quan- subject of high interest in research as well actions between clinical settings and tech- titative and automated techniques that as clinical applications. More specific ap- nical research was possible. With the rapid help radiologists in interpretation and plications of ontology and semantic mod- shift in workflows and changes in clini- analysis of images improving their ac- els have also been adopted to better describe cal practice new requirements triggered curacy and performance. But the area of the structure and content of diagnostic im- innovative and sometime disruptive so- image interpretation is not he only area aging studies. Ontological models were used lutions to emerge. Major efforts were where the rapid development of new to represent radiological and clinical knowl- invested in exploring new techniques that tools was making rapid progress. In the edge to integrate PACS other clinical infor- will allow faster transfer and wider dis- field of 3D rendering and image visual- mation systems and to support radiology tribution of images in institutions where ization, new paradigm of image display interpretation process [24] . image data were needed not only within and image processing started to emerge More specific developments for link- radiology departments but also in clini- with solution tailored more for physi- ing broader research activities and host- cal wards, specialized services, surgical cians outside radiology departments such ing collaborative efforts in medical im- suites etc. Also, because images were as surgeons, oncologists, cardiologists, aging have also emerged from academic needed more and more outside radiol- and referring physicians that need im- and scientific communities around the ogy a change in requirements of work- age data for patient management and world. An example of such coordinated station features was also driven beyond clinical decisions. The increasing num- effort is the Biomedical Informatics Re- the traditional PACS workstation intended ber of interventional procedures relying search Network (BIRN, www.nbirn.net) for image interpretation and primary di- on image guidance and the unprec- which mission is to construct the infra- agnosis by radiologists. Early work by edented evolution of minimal invasive structure for gathering, organizing and Dr. Paul Chang at the University of surgery techniques require high resolu- managing knowledge, and developing a Pittsburg and later distributed commer- tion images to be available in many clinical federated architecture for linking mul- cially by Stentor Inc., brought a com- settings outside radiology. Image access in tiple databases across sites contributing pletely new and innovative approach to operating rooms and interventional suites data and knowledge [25]. At the core of image management [17]. Focusing on the set a new level of requirements for image this knowledge organization is BIRNLex, performance and speed of image access management and communication [19]. a formally-represented ontology facili- anywhere-any-time Dr. Chang and his Professor Heinz Lemke has advocated, tating data exchange tool providing a team explored new ways of compress- for many years now, the need for special means to cast queries against specific data ing and transmitting large image sets architecture and workflow modeling in sources in the federation from storage devices to workstations operating rooms, which lead him to sug- Medical imaging is has also evolved [18]. Progressive compression and de- gest a new usage of the PACS acronym toward functional and molecular imag- compression of images associated with for “Picture Archiving and Communi- ing modalities that are able to produce advanced data streaming techniques, pro- cation in Surgery” [20]. The specific re- images that represent an in-vivo charac- vided faster and almost real-time access quirements for image management in op- terization of biological processes. to images while allowing full resolution erating suites, lead to the launch of a new Positron Emission Tomography (PET) rendering to be maintained dynamically. extension of the DICOM standard for using radio-labeled tracers represents the This pioneer work has changed the rules surgery. These new standardization ef- most common molecular imaging mo- and initiated new trends in image man- forts are carried out by the working group dality, but other techniques based on agement and communication replacing 24 of the DICOM committee under the molecular markers, are nowadays emerg- traditional pre-fetching and point to point leadership of Prof. Lemke. ing in other modalities such as MRI of- transmission of images to a more dis- In an effort to better handle image ac- fering new perspectives of functional tributed access to images when they are cess in large and complex system, some imaging. The ability to combine the func-

IMIA Yearbook of Medical Informatics 2009 171 Imaging Informatics: from Image Management to Image Navigation

tional or molecular data with the ana- developed under Open Source licensing, geles (UCLA) in 2004 and then carried tomical images adds a new dimension to providing source code to the community out in Geneva by two Swiss radiologists the images. Adding new dimensions to to share. Some ambitious Open Source Dr. Antoine Rosset and Prof. Osman the images such as time varying param- project included the Vista project (ref) a Ratib, soon joined by Joris Heuberger, a eters (often referred to as the fourth di- home-grown electronic medical record third computer scientist from Geneva mension) and functional and metabolic implemented in Veterans Affairs (VA) Hos- [26]. Together they designed and devel- parameters (referred to as the fifth di- pitals in the United States. The impact of oped this new image-processing platform mension) require special navigation tools open source is even greater in specialized specifically designed to take advantage and multidimensional rendering software areas of medicine such medical imaging. of the significant progress in perfor- that are usually not available on tradi- These vertical markets have always been mance and flexibility of 3D rendering tional PACS workstations. Such visual- a challenge for vendors and manufac- and image processing on consumer-mar- ization tools are often only available on turers due to the small size of special- ket personal computers [27]. high-end 3D rendering workstations in ized users and high expectations in terms The OSIRIX software was tailored to academic settings and specialized radi- of complexity and performance of the provided physicians unfamiliar with ad- ology departments. Ironically, where tools that users need. This has naturally vanced image processing tools with a these tools are needed the most is in clini- driven the market to high-end and high- platform allowing them to manipulate cal services and in areas where clinical cost developments and marketing strate- and visualize large sets of image data decisions and patient management gies that also try to cope with very rapid using advanced volume rendering and choices are being made. They are also evolution of computer technologies and three dimensional navigation tools. becoming important communication software developments that make most OSIRIX user interface was designed to means for radiologists to transmit their find- products obsolete in very short time in- allow physicians to rapidly become fa- ings to referring physicians. The availabil- terval, which does not allow the manu- miliar with the manipulation of 3D ob- ity of such tools are however very limited facturers to generate sustainable return jects and navigating through multi-di- outside radiology departments due to their on investment. Open Source solutions mensional images. Advanced image pro- high cost and complexity. can challenge this limited market share cessing and analysis tools are being added of large industrial vendors and provide to the program everyday. Developers more adequate and affordable solutions. from all around the world have contrib- Open Source: the New A list of recent Open Source initiatives uted to the extension of OSIRIX by add- in Medical Imaging Software can be ing innovative and specialized image pro- Paradigm in Imaging found at (www. idoimaging.com). In the cessing features. Furthermore, the Informatics scientific community, a new set of open OSIRIX software architecture allows for source libraries have emerged for the vi- separate processing modules to be added A new paradigm in computer software sualization of multidimensional data. The to the program as plug-ins. came with the success of Open Source most successful set of tools is provided projects such as Linux, Firefox, Open- by Insight Software Consortium (www. Office and other widely adopted soft- insightsoftwareconsortium.org): The Vi- Conclusions ware platforms. The rationale behind sualization Toolkit or VTK (is a well rec- open source is very simple: When pro- ognized and widely adopted software li- An amazing evolution of imaging grammers can read, redistribute, and brary that runs on multiple platforms and informatics occurred in the last couple modify the source code for a piece of has been used for numerous scientific and of decades from the early days where software, the software evolves. People medical applications so far. The recent systems for digital image management improve it, people adapt it, people fix adjunction of the Insight Toolkit or ITK, had to be developed to allow radiolo- bugs. And this can happen at a speed that, mostly funded by the US National Li- gists to review images on soft-copy if one is used to the slow pace of conven- brary of Medicine as part of the Visible workstations in replacement of traditional software development, seems as- Human Project, adds a wealth of addi- tional hard-copy films to the recent tonishing. The rapidly growing open tional rendering and image processing developments of multimodality image source community has realized that this tools for medical applications. Numer- display and analysis software platforms. rapid evolutionary process produces bet- ous Open Source software projects pro- The early adopters and pioneers in the ter software than the traditional closed viding a variety of medical imaging tools field have paved the road of today’s model, in which only a very few pro- have emerged in the recent years. PACS that are taken for granted by the grammers can see the source code and One of the most widely adopted Open newer generation of radiologists and everybody else must blindly use an Source software platform for medical physicians.While tremendous progress opaque set of software tools. This trend imaging is the Osirix software (www. were made in designing and adopting also made some progress in medical osirix-viewer.com) that was initiated at new and innovative paradigms, the pen- informatics and some early projects were the University of California in Los An- etration of imaging informatics in medi-

IMIA Yearbook of Medical Informatics 2009 172 Ratib

cal institutions is still lagging behind. integrated EMR. Comput Med Imaging Graph 26. Rosset A, Spadola L, Ratib O. OsiriX: an open-source Users are still struggling with inappro- 2003;27(2-3):207-15. software for navigating in multidimensional DICOM 8. Geissbuhler A, Lovis C, Spahni S, Appel RD, Ratib images. J Digit Imaging 2004;17(3):205-16. priate an inefficient tools while they O, Boyer C, et al. A humanist’s legacy in medical 27. Rosset C, Rosset A, Ratib O. General consumer are being challenged by the unmanage- informatics: visions and accomplishments of Pro- communication tools for improved image able quantities of data generated by fessor Jean-Raoul Scherrer. Methods Inf Med 2002; management and communication in medicine. J modern imaging modalities. 41(3): 237-42. Digit Imaging 2005;18(4):270-9. 9. Ratib O, Lemke H, Trayser G, Vurlod JF, Do H, The success of developments in Scherrer JR. Distributed Image Management and Correspondence to: PACS and large institutional projects is Hierarchical Storage in an Integrated RIS and PACS. Osman Ratib MD, PhD, FAHA in big part due to the efforts deployed In: MEDINFO’92. Geneva (Switzerland): North- Professor and Chair of Radiology by devoted advocates of standards and Holland Ed. Amsterdam; 1992 Dept. of Medical Imaging and Information Sciences their tireless work in standardization bod- 10. Ligier Y, Ratib O, Funk M, Perrier R, Girard C, Head of Nuclear Medicine Division Logean M. Portable image-manipulation software: ies. Without these standards the devel- University Hospital of Geneva what is the extra development cost? J Digit Imaging 24, rue Micheli-du-Crest opment of commercial systems would 1992;5(3):176-84. 1205 Geneva, SWITZERLAND have never been possible, and users would 11. Ratib O, Ligier Y, Scherrer JR. Digital image Tel: +41 22 372 71 45 management and communication in medicine. still be left with homegrown heteroge- Fax: +41 22 372 71 84 Comput Med Imaging Graph 1994;18(2):73-84. E-mail: [email protected] neous system that could never be inte- 12. Bakker AR, Stut WJ Jr, De Valk JP, Reijns GR. grated with other information systems. PACS costs: modelling and simulation. Med Inform Open source software is an idea (Lond) 1988;13(4):307-13. whose time has finally come. For 13. Dicom Committee. DICOM strategic document, Historic Landmark Paper Selected by twenty years it has been building mo- D.S.C.a.I.W. Groups, Editor; 2008. 14. O’Donnell K. IHE Radiology User’s Handbook Osman Ratib for the 2009 IMIA mentum in the technical cultures that 2005, ACC/HIMSS/RSNA. built the Internet and the World Wide 15. Carey B, Seshadri SB, Arenson RL. PACS: “Back Yearbook of Medical Informatics* Web. Now it’s breaking out into the to the Future”. In: Jost RG, editor: Medical Imaging The reference paper selected is a tran- commercial world and into vertical V:PACS Design and Evaluation. Proc SPIE 1991;1446:414-9. script of the first workshop held on markets such as medical applications, 16. Inamura K, Kousaka S, Yamamoto Y, Sukenobu Y, Picture Archiving and Communication and that’s changing all the rules. It is Okura Y, Matsumura Y, et al. PACS development Systems (PACS) in 1982. Although the particularly promising in advanced in Asia. Comput Med Imaging Graph 2003;27(2- proceedings of that ’82 conference are image display and analysis applications 3):121-8. 17. Morgan MB, Branstetter BFT, Lionetti DM, not available online and are hard to where the rapid increase in demand Richardson JS, Chang PJ, The radiology digital find, this paper is a re-edition of the cannot be matched by traditional ex- dashboard: effects on report turnaround time. J Di- discussions held at the workshop and pensive commercial solutions. git Imaging 2008;21(1):50-8. transcribed by Prof. André Duerinckx, 18. Morgan MB, Branstetter BFT, Mates J, Chang PJ, Flying blind: using a digital dashboard to navigate one of the organizer of that sympo- a complex PACS environment. J Digit Imaging sium. This milestone in the history of References 2006;19(1):69-75. development of PACS clearly sets up 19. Korb W, Bohn S, Burgert O, Dietz A, Jacobs S, 1. Lemke HU, Stiehl HS, Scharnweber H, Jackel D. Falk V, et al. Surgical PACS for the digital operating the concepts and vision of image man- Applications of picture processing, image analysis room. Systems engineering and specification of agement in healthcare, some of which and computer graphics techniques to cranial CT user requirements. Stud Health Technol Inform 2006; are still relevant today. Most partici- scans. In: Computer-aided analyses of radiological 119:267-72. pants mentioned in this paper were the images. Newport Beach: IEEE; 1979. 20. Lemke HU, Osteaux M. PACS and digital imaging— 2. Duerinckx AE. First international conference and new directions. Eur J Radiol 1993;17(1): 1-2. early pioneers who contributed signifi- workshop on picture archiving and communication 21. Muller H, Michoux N, Bandon D, A. Geissbuhler cantly in the development of this field. systems (PACS) for medical application. NewPort A. A review of content-based image retrieval systems It is also at this workshop that the ac- Beach: SPIE; 1982. in medical applications-clinical benefits and future ronym PACS was first cornered and 3. Duerinckx AJ. Introduction and the PACS ’82 panel directions. Int J Med Inform 2004;73(1):1-23. discussions: Panel 1—Equipment manufacturers’ 22. Muller H, Rosset A, Garcia A, Vallee JP, used ever since. view on PACS and Panel 2—The medical Geissbuhler A. Informatics in radiology (infoRAD): community’s view on PACS. 1982. J Digit Imaging benefits of content-based visual data access in Duerinckx AJ. Introduction and the PACS 2003;16(1):32-68; discussion 29-31. radiology. Radiographics 2005;25(3):849-58. ’82 panel discussions: Panel 1 - Equip- 4. Huang HK, Ratib O, Bakker AB, Witte G. Picture 23. Deserno TM, Antani S, Long R. Ontology of Gaps ment manufacturers’ view on PACS and Archiving and Communication System (PACS) in in Content-Based Image Retrieval. J Digit Imaging Medicine. NATO ASI Series. Berlin/Heidelberg: 2008;22(2):202-15 Panel 2 - The medical community’s view Springer-Verlag; 1991 24. Kahn CE Jr, Channin DS, Rubin DL. An on PACS. 1982. J Digit Imaging 2003; 5. Lemke HU. PACS developments in Europe. Comput ontology for PACS integration. J Digit Imaging 16(1):32-68; discussion 29-31. Med Imaging Graph 2003;27(2-3):111-20. 2006;19(4): 316-27. 6. Osteaux M, Van den Broeck R, Verhelle F, de Mey 25. Bug WJ, Astahkov V, Boline J, Fennema-Notestine C, J. Picture archiving and communication system Grethe J, Gupta A, et al. Data Federation in the biomedical (PACS): a progressive approach with small systems. informatics research network: tools for semantic * The complete paper can be accessed in the Yearbook’s Eur J Radiol 1996;22(3):166-74. annotation and query of distributed multiscale brain full electronic version, provided that permission 7. Ratib O, Swiernik M, McCoy JM. From PACS to data. AMIA Annu Symp Proc 2008: 1220. has been granted by the copyright holder(s)

IMIA Yearbook of Medical Informatics 2009