Ivana Išgum is Full Professor in AI for Medical Image Analysis at the Amsterdam University Medical Center and Faculty of Science of the University of Amsterdam. She leads Quantitative Healthcare Analysis (qurAI) group, an interfaculty research group embedded in the Faculties of Medicine (Biomedical Engineering and Physics) and Science (Informatics Institute). Ivana obtained her PhD in 2007 at Utrecht University, where she became an Assistant Professor in 2012, and an Associate Professor in 2015 working at the Image Sciences Institute of UMC Utrecht. In 2019 Ivana moved to Amsterdam where her current research aims at enhancing patient care by designing and enabling AI technologies in healthcare, especially in the fields of radiology and cardiology.
Ivana co-authored over 200 papers in scientific journals and conferences, she has co-organized conferences (MIDL), and workshops and challenges in conjunction with medical image analysis conferences (MICCAI, ISBI, SPIE MI), and is currently a member of the Editorial board of Medical Image Analysis.
Coronary artery disease is a leading cause of morbidity and mortality worldwide. Coronary CT is a non-invasive tool enabling analysis of the coronary arteries and providing important diagnostic information. However, current clinical analysis typically remains limited to visual evaluation of the coronary artery tree as the extraction of the detailed quantitative information requires high level of expertise and it is a time-consuming process. Besides the information about the coronary arteries, these images also contain information about the whole heart and body composition that may be valuable for the prediction of cardiovascular risk. Manual extraction of this information requires segmentation of the target structures, which is highly time-consuming rendering the analysis infeasible. Hence, such analysis is not routinely performed.
Numerous studies have shown that deep learning-based image analysis can facilitate automatic extraction of this relevant information thereby shortening the analysis times and improving reproducibility. This presentation will show our recent developments on the automatic analysis of coronary CT scans towards detection of cardiovascular disease and extraction of cardiovascular risk. The analysis includes extraction of the coronary arteries that provides the basis for the downstream assessment of the vessel morphology. This is followed by the segmentation of the artery lumen and segmentation of the atherosclerotic plaque components. The amount of plaque and stenosis of the lumen indicate the risk and provide the basis for clinical decision-making. Subsequent merging the findings from different vessels in the coronary tree allows standardized risk reporting. Moreover, deep leaning methods for the analysis of the whole heart, i.e. cardiac structures and large arteries, as well as methods for assessment of body composition will be outlined. The presentation will also showcase how the analysis can be extended to CT scans made for other purposes that visualize the heart and offer the possibility to quantify markers indicating the presence or risk of cardiovascular disease. Next to the opportunities these methods offer, the presentation will discuss current challenges towards trustworthy utilization and present the work on addressing these.
Professor David Hawkes is Emeritus Professor at the Department of Medical Physics and Biomedical Engineering at University College London (UCL). He founded the Centre for Medical Image Computing (CMIC) in 2005 and served as its director until 2015 and was the director of the Welcome EPSRC Centre for Interventional and Surgical Sciences (WEISS) 2018-2019. Previously, he was the Chair of the Division of Imaging Sciences at King's College London (2002–2004).
His career in medical imaging spans 48 years in hospital based clinical sciences as well university based research. His research has focused on image registration, data fusion, visualization, shape representation, surface geometry, modeling of tissue deformation, and disease progression modelling, with the aim of advancing medical imaging as a precise measurement tool for staging disease and for image-guided interventions. Professor Hawkes has contributed to research on various diseases, including cancer of the breast, prostate, liver and brain, lung disease and neurodegenerative diseases. He has co-authored over 500 peer-reviewed scientific publications.
He has always worked closely with industry and strongly supports entrepreneurship in medical technologies. He has assisted in the formation of a number of successful companies with a combined value of well over £1bn.
He has received numerous awards for his contributions, including Fellowships from the Academy of Medical Sciences (2011), the MICCAI Society (2009), the Royal Academy of Engineering (2002) and the Institute of Physics (1997). He received the Institute of Physics Peter Mansfield Prize for Medical Physics (2019), the MICCAI Society Enduring Impact Award (2016) the Crookshank Medal from the Royal College of Radiologists (2008), NIHR Senior Investigator (2009) and gave the Wilhelm Conrad Roentgen Honorary Lecture at the European Congress of Radiology (2006).
Reluctantly abandoning my career in rock-and-roll and inspired by the elegant mathematics of reconstruction in computed tomography I started working in medical imaging as a clinical scientist. Working in the clinical environment I appreciated how much in medicine was unknown. I became convinced that a physics and engineering approach could contribute to improving healthcare and that this must be done in close collaboration with clinical colleagues in order to be relevant. I also became convinced that real impact in healthcare engineering would only come about by working closely with industry. Only this would enable wide dissemination. Clinical translation of new technologies involves engaging with complex regulatory pathways, validation and clinical trials. Late in the day I also learnt that the patient has a voice that needs to be heard in order to better direct scarce resource to solve the problems that really matter.
This talk will describe how I was fortunate to be inspired by giants in the field. I will run through my early journey from nuclear medicine and X-ray computed tomography to medical image computing, with initial emphasis on imaging as a tool for measurement. I realized that the resources required for development of imaging hardware would be difficult to come by in the academic research environment. On the other hand computing costs were dramatically reducing and compute power was rapidly increasing. I believed that academia was the place both to develop the underpinning mathematics and algorithms, as well as the computer implementations and experimentation needed to have impact. This led to the formation of the Computational Imaging Sciences Group (CISG) at Guy’s Hospital (later KCL) and subsequently the Centre for Medical Image Computing (CMIC) at UCL. With the rapid development of MR and wide availability of other modalities probing different aspects of structure and function, we were motivated to find ways of combining information. This led to the need to tackle image registration. With integration of real-time imaging and video this led naturally to work on image guided interventions. This led to the foundation of the Welcome EPSRC Centre for Interventional and Surgical Sciences (WEISS). Close collaboration with pathologists resulting from our work with surgeons opened my eyes to the world of cellular microscopy. This led to our efforts in trying to relate the macroscopic (millimeter) scale of medical imaging to the microscopic world of histology, where cancer and other diseases are ultimately diagnosed and staged.
I retired just as machine learning and artificial intelligence were taking off and this promised a revolution as profound as that which I had witnessed in the early days of medical imaging. I will describe how I learnt from the many frustrations and failures that are inevitable in a research career in a rapidly evolving field, but also the excitement and inspiration as to what to do next. I will try to draw parallels with the promises and excitement, but also the disappointments that might lie ahead in the field of machine intelligence and AI, to inspire the next generation in computation applied to medicine.
Finally everything I have described is collaborative, involving many disciplines, and I gratefully acknowledge the many students, post-docs, visiting scientists and faculty that have made this all possible, along with the long term support of our funders. This area is truly multidisciplinary and medicine by its nature crosses all international boundaries. It has been an exciting and rewarding journey.
Prof. Dr. Dirk Wilhelm is a Professor of Surgery and Sector Lead for Medical Robotics at the Munich Institute of Robotics and Machine Intelligence (MRI) at the Technical University of Munich (TUM). He also heads the Center for Medical Robotics and Machine Intelligence as well as the research group for Minimally Invasive Therapeutic Interventions (MITI) at the Klinikum rechts der Isar. His scientific focus includes colorectal surgery, robotic surgery, artificial intelligence, and biomedical engineering.
Since 2020, Prof. Wilhelm has been an advisor for robotics to the Bavarian State Ministry of Economic Affairs, Regional Development, and Energy, and he served as President of leading organizations such as Computer Aided Radiology and Surgery (CARS). His academic achievements include numerous awards, such as the Innovation Award of Boston Scientific (2010, 2019) and the Karl Langenbuch Award of the German Society for General and Visceral Surgery (2007). Additionally, he is an editorial board member of Visceral Medicine and an active member of several scientific committees.
With several high-impact publications, Prof. Wilhelm advances the integration of artificial intelligence in surgery and develops innovative concepts such as the patient-specific surgical platform PLAFOKON.
In addition to continuously rising costs, the healthcare system is currently characterised by an increasing shortage of staff. It is currently estimated that by the year 2023 there will be a shortage of 500,000 nurses and 6,000 doctors in Germany alone. Consequences of this are delays in medical treatment, lack of operation in individual departments resulting in medical undersupply and overworked staff. In addition to the recruitment of specialists from abroad and process optimisation, the use of robotic systems is a possible approach to solving this system. For these to relieve the workload, they must have autonomous functions to be able to take over work. While this is already feasible today for logistical tasks and patient transport, for example, the limits of this approach are still unclear. A survey of AI experts currently assumes that surgical skills could be taken over by artificial intelligence by 2050-2055 already [1] - and the success story of surgical robotics and companies such as Intuitive Surg. Inc. shows that robots have already replaced humans on the patient side. But are effective robotic systems and powerful AI algorithms sufficient to reproduce surgical procedures autonomously? Do we already have the necessary components of autonomous surgery?
When revising the current literature, one might assume that this indeed is the fact and that we are about to replace humans also for very delicate tasks. One can find for example reports on autonomous suturing and knot tying systems, robotic appendectomy and retraction and many more. But is this what we understand as an intervention already and can we easily jump from these success stories to full surgeries? From a surgeons perspective and although sub-processes of an operation can already be automated, a surgical procedure however poses a much greater challenge. This not only is due to the complexity of the process, but also to the fact that operations can only be described or predicted in abstract terms, but have aspects of a highly complex system at a granular level. We assume that these can only be resolved through comprehensive standardisation of surgeries and a comprehensive approach that includes the individual patient, the underlying disease, but numerous other factors alike. The presentation will not only attempt to provide a status quo to this topic, but also approaches to how the problem can be tackled in the long term.
[1] Grace, K., Salvatier, J., Dafoe, A., Zhang, B. and Evans, O., 2018. When will AI exceed human performance? Evidence from AI experts. Journal of Artificial Intelligence Research, 62, pp.729-754.