Multidimensional Embedded MEMS Motion Detectors for Wearable Mechanocardiography and 4D Medical Imaging Mojtaba Jafari Tadi TURUN YLIOPISTON JULKAISUJA – ANNALES UNIVERSITATIS TURKUENSIS SARJA - SER. D OSA - TOM. 1404 | MEDICA - ODONTOLOGICA | TURKU 2018 30957047_Turun_yliopisto_Thesis_Mojtaba_Jafaritadi_Faculty_of_Medicine_cover_wire_18_11_26.indd 1 26.11.2018 12.18.12 MULTIDIMENSIONAL EMBEDDED MEMS MOTION DETECTORS FOR WEARABLE MECHANOCARDIOGRAPHY AND 4D MEDICAL IMAGING Mojtaba Jafari Tadi TURUN YLIOPISTON JULKAISUJA – ANNALES UNIVERSITATIS TURKUENSIS SARJA - SER. D OSA - TOM. 1404 | MEDICA - ODONTOLOGICA | TURKU 2018 30957047_Turun_yliopisto_Thesis_Mojtaba_Jafaritadi_Faculty_of_Medicine_inside_18_11_26.indd 1 26.11.2018 8.11.17 University of Turku Faculty of Medicine Medical Physics and Engineering Doctoral Program in Clinical Research Turku PET Center, Turku University Hospital Abstract Faculty of Science and Engineering Department of Future Technologies Mojtaba Jafari Tadi, M.Sc. Supervised by University of Turku, Faculty of Medicine, Medical Physics and Engineering, Doctoral Program in Clinical Research, Associate Professor. Antti Saraste, MD, Ph.D. Professor Mika Teräs, Ph.D. Turku PET Center, Turku University Hospital, Department of Medical Physics Department of Cardiology, Faculty of Science and Engineering, Department of Future Technologies Department of Biomedicine Turku Heart Center University of Turku Turku PET Center Multidimensional Embedded MEMS Motion Detectors for Wearable Mechanocardio- Turku University Hospital University of Turku graphy and 4D Medical Imaging Turku, Finland Turku University Hospital Turku, Finland Background: Cardiovascular diseases are the number one cause of death. Of these deaths, Dr. Eero Lehtonen, D.Sc. (Tech) almost 80% are due to coronary artery disease (CAD) and cerebrovascular disease. Multidi- Department of Future Technologies mensional microelectromechanical systems (MEMS) sensors allow measuring the mechanical University of Turku movement of the heart muscle offering an entirely new and innovative solution to evaluate car- Turku, Finland diac rhythm and function. Recent advances in miniaturized motion sensors present an exciting Reviewed by opportunity to study novel device-driven and functional motion detection systems in the areas of both cardiac monitoring and biomedical imaging, for example, in computed tomography (CT) Professor Omer T. Inan Professor Simo Särkkä and positron emission tomography (PET). School of Electrical and Computer Department of Electrical Engineering Methods: This Ph.D. work describes a new cardiac motion detection paradigm and mea- Engineering and Automation surement technology based on multimodal measuring tools — by tracking the heart’s kinetic Georgia Institute of Technology Aalto University activity using micro-sized MEMS sensors — and novel computational approaches — by de- Atlanta, United States of America Helsinki, Finland ploying signal processing and machine learning techniques — for detecting cardiac pathological disorders. In particular, this study focuses on the capability of joint gyrocardiography (GCG) Opponent and seismocardiography (SCG) techniques that constitute the mechanocardiography (MCG) concept representing the mechanical characteristics of the cardiac precordial surface vibrations. Professor Kouhyar Tavakolian Results: Experimental analyses showed that integrating multisource sensory data resulted Department of Electrical Engineering in precise estimation of heart rate with an accuracy of 99% (healthy, n=29), detection of heart University of North Dakota arrhythmia (n=435) with an accuracy of 95-97%, ischemic disease indication with approxi- North Dakota, United States of America mately 75% accuracy (n=22), as well as significantly improved quality of four-dimensional (4D) cardiac PET images by eliminating motion related inaccuracies using MEMS dual gating approach. Tissue Doppler imaging (TDI) analysis of GCG (healthy, n=9) showed promising The originality of this thesis has been checked in accordance with results for measuring the cardiac timing intervals and myocardial deformation changes. the University of Turku quality assurance system using the Turnitin Conclusion: The findings of this study demonstrate clinical potential of MEMS motion OriginalityCheck service. sensors in cardiology that may facilitate in time diagnosis of cardiac abnormalities. Multidi- mensional MCG can effectively contribute to detecting atrial fibrillation (AFib), myocardial ISBN 978-951-29-7509-9 (PRINT) infarction (MI), and CAD. Additionally, MEMS motion sensing improves the reliability and ISBN 978-951-29-7510-5 (PDF) quality of cardiac PET imaging. ISSN 0355-9483 (Print) ISSN 2343-3213 (Online) Keywords: MEMS motion sensors, seismocardiography, gyrocardiography, dual cardiac and Grano Oy - Turku, Finland 2018 respiratory gating, cardiovascular disease, cardiac PET/CT imaging i 30957047_Turun_yliopisto_Thesis_Mojtaba_Jafaritadi_Faculty_of_Medicine_inside_18_11_26.indd 2 26.11.2018 8.11.18 Abstract Mojtaba Jafari Tadi, M.Sc. University of Turku, Faculty of Medicine, Medical Physics and Engineering, Doctoral Program in Clinical Research, Turku PET Center, Turku University Hospital, Faculty of Science and Engineering, Department of Future Technologies Multidimensional Embedded MEMS Motion Detectors for Wearable Mechanocardio- graphy and 4D Medical Imaging Background: Cardiovascular diseases are the number one cause of death. Of these deaths, almost 80% are due to coronary artery disease (CAD) and cerebrovascular disease. Multidi- mensional microelectromechanical systems (MEMS) sensors allow measuring the mechanical movement of the heart muscle offering an entirely new and innovative solution to evaluate car- diac rhythm and function. Recent advances in miniaturized motion sensors present an exciting opportunity to study novel device-driven and functional motion detection systems in the areas of both cardiac monitoring and biomedical imaging, for example, in computed tomography (CT) and positron emission tomography (PET). Methods: This Ph.D. work describes a new cardiac motion detection paradigm and mea- surement technology based on multimodal measuring tools — by tracking the heart’s kinetic activity using micro-sized MEMS sensors — and novel computational approaches — by de- ploying signal processing and machine learning techniques — for detecting cardiac pathological disorders. In particular, this study focuses on the capability of joint gyrocardiography (GCG) and seismocardiography (SCG) techniques that constitute the mechanocardiography (MCG) concept representing the mechanical characteristics of the cardiac precordial surface vibrations. Results: Experimental analyses showed that integrating multisource sensory data resulted in precise estimation of heart rate with an accuracy of 99% (healthy, n=29), detection of heart arrhythmia (n=435) with an accuracy of 95-97%, ischemic disease indication with approxi- mately 75% accuracy (n=22), as well as significantly improved quality of four-dimensional (4D) cardiac PET images by eliminating motion related inaccuracies using MEMS dual gating approach. Tissue Doppler imaging (TDI) analysis of GCG (healthy, n=9) showed promising results for measuring the cardiac timing intervals and myocardial deformation changes. Conclusion: The findings of this study demonstrate clinical potential of MEMS motion sensors in cardiology that may facilitate in time diagnosis of cardiac abnormalities. Multidi- mensional MCG can effectively contribute to detecting atrial fibrillation (AFib), myocardial infarction (MI), and CAD. Additionally, MEMS motion sensing improves the reliability and quality of cardiac PET imaging. Keywords: MEMS motion sensors, seismocardiography, gyrocardiography, dual cardiac and respiratory gating, cardiovascular disease, cardiac PET/CT imaging i i 30957047_Turun_yliopisto_Thesis_Mojtaba_Jafaritadi_Faculty_of_Medicine_inside_18_11_26.indd 1 26.11.2018 8.11.18 Tiivistelmä Contents Mojtaba Jafari Tadi I Research Summary 1 Turun yliopisto, lääketieteellinen tiedekunta, Lääketieteellisen fysiikan ja tekniikan oppiaine, Kliinisen tutkimuksen tohtoriohjelma, 1 INTRODUCTION 2 PET-keskus, Turun yliopistollinen keskussairaala, 1.1 Emerging Technologies for Health Monitoring . 2 Luonnontieteiden ja tekniikan tiedekunta, Tulevaisuuden teknologioiden laitos 1.2 Problem Statement and Motivation of the Thesis . 3 1.3 Organization of the Dissertation . 5 Moniulotteisten sulautettujen MEMS-liiketunnistimien käyttö sydänkardiografiassa sekä lääketieteellisessä 4D-kuvantamisessa 2 Background Physiology and Underlying Cardiovascular Diseases 6 2.1CardiovascularSystem.............................. 6 Tausta: Sydän- ja verisuonitaudit ovat yleisin kuolinsyy. Näistä kuolemantapauksista lähes 2.1.1 Anatomy and Mechanical Physiology of the Heart . 6 80% johtuu sepelvaltimotaudista (CAD) ja aivoverenkierron häiriöistä. Moniulotteiset mikroelek- 2.1.2 CardiacWallMotion........................... 6 tromekaaniset järjestelmät (MEMS) mahdollistavat sydänlihaksen mekaanisen liikkeen mit- 2.1.3 Cardiac Electromechanical Conduction . 7 taamisen, mikä puolestaan tarjoaa täysin uudenlaisen ja innovatiivisen ratkaisun sydämen ryt- 2.1.4 CardiacTimeIntervals.......................... 8 min ja toiminnan arvioimiseksi. Viimeaikaiset teknologiset edistysaskeleet mahdollistavat uusien 2.2CardiovascularDiseases............................. 9 pienikokoisten liiketunnistusjärjestelmien käyttämisen sydämen toiminnan tutkimuksessa sekä 2.2.1