Medical Image Analysis & Artificial Intelligence Symposium 2022 Danube Private University Wachau/Austria
Ladies and Gentlemen, I would like to express my sincere gratitude and thanks to all who contributed to the success of our first Symposium “Medical Image Analysis & Artificial Intelligence” at the Danube Private University, Faculty for Medicine, in Autumn 2022. The event was hosted by the Danube Private University and took place in their newly built conference and concert hall in Unterloiben in the Wachau, Austria. Our primary objective for this Symposium was to bring together thought leaders and scientific expertise at a time where digitalization plays such a crucial role in the diagnosis and treatment of patients. With this event we were able to showcase how multidisciplinary collaboration can lead to break through developments in science to improve the health of patients and the public. European scientists, academics and business professionals had a platform for scientific debate enabling them to exchange newly acquired knowledge in the fields of Medical Image Analysis and Artificial Intelligence. My special thanks goes out to our colleagues at the Medical Image Analysis and Artificial Intelligence (MIAAI) group at the Danube Private University. Since its establishment in Summer 2022, under the scientific direction of Univ.-Prof. Dr. Ramona Woitek and Ass.-Prof. Dr. Sepideh Hatamikia, the MIAAI Group focuses on the analysis of medical imaging data for the development of quantitative biomarkers, as well as the use of Artificial Intelligence to predict the presence of disease, its progression or treatment response. Additionally, my thanks goes out to our Key Note Speakers Professor Jean Abraham, Director of Cambridge Breast Cancer Research Unit, University of Cambridge, UK, and Professor Evis Sala, Professor of Oncological Imaging, University of Cambridge, UK. Their outstanding research work has led to significant medical advances in cancer research and we continue, with great anticipation, to see developments unfold for the future of cancer patients. I hope that the following summary of speakers, research topics and posterwalks provides you with a lasting memory of a truly successful first symposium and we look forward to further events with you in the future. Kindest regards, Robert Wagner, MA Director for Higher Education Planning, Strategic Management, Research & Development Danube Private University © Pressefoto LACKINGER
The MIAAI Symposium in the Wachau region was a unique opportunity to celebrate this year’s launch of our MIAAI research centre at DPU. Guests from home and abroad, researchers and staff from DPU as well as several other universities made this event very special to our young group. Rarely have academics from such diverse backgrounds come together to discuss topics related to imaging and AI and to critically examine projects and results at eye level in such an inspiring atmosphere. Our symposium showed very clearly how important it is to provide a platform for international multidisciplinary exchange and that knowledge and skills can only be optimally used and promoted in interdisciplinary collaboration. Researchers from radiology, oncology, urology and dentistry met colleagues from biomedical engineering, AI, data science and physics. We heard about the complexity of ovarian cancer and how medical image analysis contributes to its understanding. We were shown how multidisciplinary data sharing and collaboration can revolutionise breast cancer therapy. We saw the importance of critical thinking about radiomics to avoid drowning in a sea of low-value data and features. We heard how decision support systems simplify and speed up decision making in uro-oncology. We were presented with the history of radiology and image analysis in a nutshell – highlighted by the use of virtual reality and robotics in this field. We learned that images can also be created from soundsand that image analysis allows to analyse sound without ever hearing it. There are few international meetings where such a wide range of topics can be discussed – with all of them finding together under the topic Medical Image Analysis and AI. These presentations and discussions have given the organisers and audience plenty of food for thought for future projects and collaborations. I would like to thank our keynote speakers, Professor Evis Sala and Professor Jean Abraham, who came all the way from Cambridge and Rome to be with us for this special day and whose work and presence enriched the meeting far beyond expectations. As the MIAAI team, we look forward to collaborating with you and your research groups at the University of Cambridge and the Fondazione Policlinico Universitario Agostino Gemelli in Rome and we hope to establish a close network including the DPU’s local research partners for the development of joint projects and knowledge exchange. I would like to give special thanks to Director Robert Wagner and President Prof. h.c. Marga B. Wagner-Pischel for creating such a wonderful atmosphere at the symposium, for the opportunity for this exchange and for the trust that led to the foundation of MIAAI. I wish all of MIAAI’s members and friends success in their endeavours to conquer new fields of research and am looking forward to being part of a growing community interested in the beauty of medical imaging and in investigating how AI can help us personalise medicine for the benefit of our patients. Univ.-Prof. DDr. Ramona Woitek Head of the Research Group MIAAI (Medical Image Analysis & Artificial Intelligence) Danube Private University, Krems, Austria
Index Prof. Evis Sala, MD, PhD, FRCR Univ.-Prof. DDr. Ramona Woitek Dr. Lorena Escudero Sanchez, PhD Prof. Jean Abraham Ass.-Prof. Dr. Sepideh Hatamikia ao. Univ.-Prof. Mag. Dr. Wolfgang Birkfellner Dr. Ross King Dr. Maria Bernathova Prof. Dr. Christian Wetterauer Univ.-Prof. Dr. Constantin von See, MaHM
7 Symposium 2022 Curriculum Vitae: Prof. Sala is an academic radiologist with a special interest in Cancer Imaging. She is the Professor of Oncological Imaging at the University of Cambridge, UK. Previously, she served as Chief of Body Imaging Service at Memorial Sloan Kettering Cancer Center and Professor of Radiology, Weill Cornell Medical College in New York until December 2017. Before joining the Memorial Sloan Kettering in July 2012, she was a University Lecturer in Radiology and Specialty Teaching Director (Radiology) at the University of Cambridge, UK. She obtained her PhD from University of Cambridge, UK in 2000 and completed her training in Clinical Radiology at Cambridge, UK in August 2005. Today, Prof. Sala leads the Radiogenomics and Quantitative Imaging Group in the Department of Radiology. She is also active in many academic organizations. She is the Chair of Radiology Society of North America (RSNA) Oncologic Imaging Track, serves on the Oncologic Imaging and Therapies Task Force of RSNA and the Genitourinary Imaging Subcommittee of European Society of Radiology. She is a member of Board of Trustees of the International Society for Magnetic Resonance in Medicine (ISMRM), the International Cancer Imaging Society (ICIS) and The European Society of Urogenital Radiology (ESUR). Prof. Sala is an Editorial Board member and Head of Oncology Section of European Radiology. In recognition for her contribution to education and research in oncological imaging, she was elected as a Fellow of ICIS in 2014, a Fellow of ISMRM in 2015, a fellow of ESUR in 2018 and received the RSNA Honoured Educator Award in 2014 and 2017. Abstract: Integrated Radiogenomic for Unravelling Tumour Heterogeneity and Treatment Monitoring in Ovarian Cancer At every point along the cancer continuum of care, from early detection through to advanced disease, a patient must face decisions that may have a direct impact on their long-term outcomes. These decisions are increasingly complex as an ever-growing array of management options emerge, and highly complex data drive both individual patient care and clinical studies. Consequently, much uncertainty remains around how to optimize decisions for a given patient. At the root of the challenge is the profound clinical and genomic heterogeneity. Risk stratification based on standard clinical and pathologic features often has suboptimal performance. Moreover, tumour heterogeneity is the main driver of treatment resistance in any cancer but especially in in high grade serous ovarian cancer (HGSOC) as both primary tumours and metastatic lesions are spatially and temporally heterogeneous. They would require multiple biopsies to extract and Prof. Evis Sala, MD, PhD, FRCR Professor of Oncological Imaging University of Cambridge © HUEBL
analyse small portions of tumour tissue, which still doesn’t allow for a complete characterization of the tumour genomic landscape. Therefore, radiomics has great potential for a comprehensive evaluation of the entire tumour burden as it is non-invasive and is already often repeated during treatment in routine practice. Radiomics (on both CT and MRI) can quantify such spatial and temporal heterogeneity, potentially providing “virtual biopsies”. Integration of radiomics with other multi-omics data is crucial for precision oncology.
9 Symposium 2022 Curriculum Vitae: After completing her degree in Medicine, Prof. Woitek carried out speciality training in Radiology at the Medical University of Vienna from 2008–2014. She subsequently went on to write her PhD entitled “Fetal MRI in the (micro-) structural and functional assessment of the posterior fossa”. Between 2014 and 2017 she was Postdoctoral Fellow and Consultant Radiologist at the Department of Biomedical Imaging and Image-guided Therapy, Medical University of Vienna. She was awarded the Erwin Schrödinger Research Fellowship at the University of Cambridge in 2017 after which she was appointed Senior Research Associate and Honorary Consultant Radiologist at the Department of Radiology, University of Cambridge. Since 2022 she is Professor of Radiology and Head of the Research Group Medical Image Analysis and AI (MIAAI) at the Danube Private University (DPU) in Krems, Austria. Prof. Woitek is Author of 56 peer-reviewed scientific publications. Abstract: Get to know MIAAI The audience will be indroduced to the research group MIAAI (Medical Image Analysis and AI), its members and main areas of work and will be taken on a tour through the field of radiomics research. Radiomics allows the characterisation of tumours using large numbers of quantitative features extracted from medical images (such as CT, MRI and mammograms) and thereby strongly increases the information gained from these images when compared to clinical image interpretation by radiologists. Artificial intelligence (AI) allows the development of radiomics-based prediction tools for patient outcome, such as survival or treatment response which can support clinicians and patients in their decision making. Radiomics are gaining immense interest, but profound understanding of the methodological background and risks for bias are crucial and MIAAI is working towards improvements in both areas. Besides maximizing the utility of standard clinical imaging, MIAAI is also exploring advanced MRI techniques, such as sodium and hyperpolarised carbon-13 MRI, and their potential for very early response assessment in patients with breast cancer to spare patients from the side effects of potentially non-efficacious treatments. Novel techniques and AI-based tools in particular can be a reason for concern on the one hand, but are also celebrated for their potential to reduce our clinicians’ workload. MIAAI is interested in understanding the expectations towards AI among different members of a healthcare team and to see if improved knowledge transfer can positively affect these expectations. Univ.-Prof. DDr. Ramona Woitek Head of Research Group, MIAAI, Danube Private University, Krems, Austria © Cathrin Andel
Medica l Image Ana lysis & Ar t i f icia l Intel l igence Symposium 2022 Curriculum Vitae: Dr. Lorena Escudero Sánchez is an interdisciplinary researcher, specialised in Artificial Intelligence (AI), Machine Learning (ML), Deep Learning (DL), Data Science, Image Analysis and software development, applying the scientific vision and experience gained as a Particle Physicist to develop advanced cancer imaging techniques that will make an impact on society. Lorena is a Turing Fellow of The Alan Turing Institute, a Borysiewicz Interdisciplinary Fellow of the University of Cambridge, and a Rokos PDRA of Queens’ College Cambridge. She graduated from the University of Salamanca with a degree in Physics and a Master in Cosmology and Particle Physics. After a short time working at CERN, she then moved to Valencia with an FPU fellowship to complete her PhD in Neutrino Physics, working on the T2K experiment in Japan. In 2016 she moved to Cambridge and worked as a Research Associate at the Cavendish Laboratory, working mainly on pattern-recognition algorithms and related software development. She has been awarded numerous scholarships and awards, including the 2016 Breakthrough Prize in Fundamental Physics (as part of the T2K collaboration) and the PhD Award of Excellence by the University of Valencia. Lorena has led teams within worldwide collaborations, e.g. as a convenor of the Simulation and Reconstruction working group of the DUNE experiment(1000+ members), and managed international initiatives such as the “UK-Latin American Neutrino Initiative” (2018–2019). Abstract: Assessing the robustness of Radiomic features for generalisable prediction outcomes Radiomic image features are becoming a promising non-invasive method to obtain quantitative measurements for tumour classification and therapy response assessment in oncological research. These features include characteristics related to the distribution of the pixel values, as well as shape measurements and analysis of first and higher order textures, and convey information from imaging data that may not be visible to the human eye, thus providing valuable quantitative, reader-independent information. However, despite its increasingly established application and potential in precision oncology, there is still a need for standardisation criteria, as well as further validation of feature robustness with respect to imaging acquisition protocols, reconstruction parameters and patient demographics. In this talk, I will discuss the studies that we have carried out in our group in a variety of cancers (hepatic, renal and ovarian), in order to establish the robustness and reliability of CT-derived radiomic features. I will also present use-cases in which the impact of these studies has been further analysed in terms of generalisability of prediction outcomes in real-world datasets. Finally, I will mention the efforts to make the extraction of these features, and predictions based on them, accessible in tools used by radiologists. © HUEBL Dr. Lorena Escudero Sanchez, PhD Senior Research Associate, University of Cambridge Borysiewicz Interdisciplinary Fellow, University of Cambridge Turing Fellow, The Alan Turing Institute
11 Symposium 2022 Curriculum Vitae: Jean Abraham is Professor of Precision Breast Cancer Medicine and an Honorary Consultant in Medical Oncology at the University of Cambridge. She directs the Precision Breast Cancer Institute and co-leads the Integrated Cancer Medicine theme and the Breast Programme in the Cancer Research UK Cambridge Centre. Jean is deputy theme lead for Cancer at the Cambridge NIHR Biomedical Research Campus. She is Chief Investigator of eight national/ regional trials. She is a member of the National Clinical Studies Group for Breast Cancer and has advised NICE on breast cancer therapeutics and the House of Commons Select Committee on Genomics. She completed her undergraduate training in Pharmacology and in Medicine at the University of Liverpool. She was awarded a Cancer Research UK National Clinical Training Fellowship and completed her PhD at the University of Cambridge. Abstract: Multi-Modal Data Integration in Breast Cancer Management Breast cancers are a heterogeneous disease with multiple sub-types. Many clinical trials and translational studies have collected disparate pieces of information on patients, often siloed in different laboratories or institutes with no holistic view of the biological information on any one patient. We need to combine multiple data modalities from individual patients to develop richer disease models that be used to predict response to therapy, prognosticate and investigate breast cancer with greater precision and accuracy. However, translating and implementing this into more routine care is complicated. My research programme brings together clinical trials and translational studies of breast cancer patients to develop clinical pathways that aim ultimately to deliver multi-modal data integration in near real time. I will give examples of real-time multi-modal data integration that impacts on clinical care. Initially focusing on genomics and transcriptomics but developing the pathways for radiomics and digital pathology as well as other “omics”. We are hoping to continue to collaborate with Professor Ramona Woitek as part of this project. © HUEBL Prof. Jean Abraham Director of Cambridge Breast Cancer Research Unit, UK
Medica l Image Ana lysis & Ar t i f icia l Intel l igence Symposium 2022 Curriculum Vitae: Ass.-Prof. Dr. Sepideh Hatamikia completed her PhD and also a postdoctoral position in Biomedical Engineering at the Medical University of Vienna. Since March 2022, she has been working at the Danube Private University (DPU) where she is Head of Computational Imaging in the MIAAI research group. In addition, Ass.-Prof. Dr. Hatamikia is a researcher at the Austrian Center for Medical Innovation and Technology (ACMIT) since January 2018 and continues collaboration with this research center on different research projects. Her main research interest includes medical image processing (image reconstruction, image registration and image segmentation), artificial intelligence (machine learning and deep learning) as well as 3D printed imaging phantom development. Abstract: Computational Imaging in MIAAI research group Dr. Hatamikia will talk about previous research projects in the field of Medical Image Analysis, introduce the computational imaging group at MIAAI and present current computational projects. © HUEBL Ass.-Prof. Dr. Sepideh Hatamikia Head of Computational Imaging, MIAAI, Danube Privater University, Krems, Austria
13 Symposium 2022 Curriculum Vitae: Prof. Mag. Dr. Wolfgang Birkfellner completed his PhD project in image-guided therapy at Vienna General Hospital (AKH) in 1996. He went on to complete a Post Doctorate at the University Hospital Basle/Switzerland from 2001-2003 at the Department of Radiology. On completion he became an Associate Professor of Medical Physics at the Medical University Vienna’s Center for Medical Physics and Biomedical Engineering. His main areas of specialisation include expertise in image fusion, medical imaging physics, image processing, multi-modal imaging and image-guided therapy. Since 2017, he is director of the postgraduate course for board – certified medical physicists in Austria. Professor Birkfellner has authored or co-authored more than 140 publications, his overall h-index is 37. Abstract: “Medical Imaging and Medical Physics – 125 Years of applied research” Professor Birkfellner’s presentation takes the audience on a journey through the history of medical imaging and imaging research. Starting from the first publication of the discovery of X-rays by Wilhelm Konrad Röntgen in the Austrian newspaper “Die Presse”, over the very first attempts at cross-sectional imaging like CT and MRI – developments that lead to many Nobel Prize winners in the field - Professor Birkfellner then introduces the audience to his own prototypes of extended reality devices for medical interventions and surgeries and robots designed specifically for imaging-guided interventions. ao. Univ.-Prof. Mag. Dr. Wolfgang Birkfellner Center for Medical Physics and Biomedical Engineering, Medical University Vienna, Austria © HUEBL
Medica l Image Ana lysis & Ar t i f icia l Intel l igence Symposium 2022 Curriculum Vitae: Dr. Ross King completed his BS degree in Engineering Physics from the Colorado School of Mines in 1988. He went on to write his Ph.D degree in Physics at Stanford University. Following positions at Institute for Medium Energy Physics (IMEP) where he headed the Electronics Department, as well as Alpha Thinx Mobile Phone Services AG and uma information technology AG where he held positions as senior manager and technical director respectively, he went on to build up the Research Studio Digital Memory Engineering at Austrian Research Centers. Here he served on the Board of Directors of the EU FP-6 Integrated Projects BRICKS as well as on the Scientific Board of the EU FP-6 Integrated Project Planets. Starting at Austrian Institute of Technology GmbH (AIT) in 2008 he played an integral part in the Digital Safety and Security Information Management. Here he coordinated the FFG Research Studios Project “Digital Memory Engineering” and went on to become AIT Thematic Coordinator for Data Science and subsequently Head of the DSAI Competence Unit. Dr. Ross King is member of the Big Data Value Association (BDVA) and served as Chairman of the Board of the Open Preservation Foundation from 2013 until 2017. He is the author of numerous Computer Science Publications. Abstract: Audio Analysis for Medical Diagnostics Dr. King will present a paper describing a neural network model for the multiclass classification of coughs based on the Respiratory Sound database that was originally compiled to support the scientific challenge organized at Int. Conf. on Biomedical Health Informatics - ICBHI 2017. The paper approaches the challenge of audio analysis by converting the audio signal to an image signal using spectrogram analysis, then appliying convolutional neural networks to the results image. The resulting model combines inception networks with ensemble learning and compares well with other state-of-the art published models. At the same time, Dr. King will use this paper as an example of how machine learning is applied in general, what some of the challenges are, and what improvements are necessary in order to advance the use machine learning as a tool for medical diagnostics. Dr. Ross King Head of Competence Unit Data Science & Artificial Intelligence AIT Austrian Institute of Technology GmbH
15 Symposium 2022 Curriculum Vitae: Mag. Dr. Maria Bernathova is Consultant Breast Radiologist at Medical University of Vienna, Austria. She graduated from Komenius-University of Bratislava, Slovakia in 1996 in Medicine and went on to complete her Master’s Degree in Medical Informatic at the UMIT – The Health & Life Sciences University, Hall in Tirol, Austria in 2013. From 2008 – 2012 she was Cofounder and Director of SMG Clinic Gibraltar, Gibraltar (Europe) and since 2011 Consultant Breast Radiologist at the Medical University of Vienna. Since 2015 Dr. Bernathova has been a Consultant Breast Radiologist at Privat Klinik Döbling in Vienna as well as Founder of a private clinic for breast health “Be-Sure” in Vienna. In addition she holds the position of Domain Expert in the IDS Project (Development of a decision support for interdisciplinary tumor boards, a project hosted by Siemens). Abstract: Bridging the gap between medicine and computer science Progress report Subject of this talk is going to be an up-cycling of medical data, which have been stored in the course of medical care. Two projects with very different origin will be presented. Main focus will be on “lessons learned”, as one of the projects failed and could not be proceeded. However, this failure brought a deep insight in medical data collected on daily base within the university hospital. One could call it a teething problem in a new era of medical data up-cycling, however without proper arrangements during medical documentation a re-use of data is not going to be realistic. Beyond the medicine there are inspirational successes for re-use of Big Data collected by internet companies. Target advertising by Google or Amazon is one of those, which everyone experienced while browsing on internet. Wishful thinking in health care environment would be applications which could make an intelligent search in medical data storages possible, as it is already “new normal” while we search for some information in Google. A typical application which would be gamechanger in daily patient care are “Trial matching”, “Similar patient search”, “Guideline adherence” during medical decision process or risk assessment. All those new applications would not be possible without support of computer science. Therefore, a human-computer collaboration will be the new “normal” in health care. Impactful bridging of those very different areas are important topics for current research. © HUEBL Dr. Maria Bernathova Consultant Breast Radiologist, Privat Klinik Döbling, Vienna
Medica l Image Ana lysis & Ar t i f icia l Intel l igence Symposium 2022 Curriculum Vitae: Prof. Dr. Christian Wetterauer completed his medical studies at Albert Ludwig University in Freiburg, Germany. He completed speciality training in urology at the university hospitals of Zurich and Basel, followed by a research fellowship at the Institut Pasteur in Paris, France. He currently works as Chief Physician in Urology at University Hospital Basel where he is also Head of Center of Excellence for Robotic Assisted Prostate Cancer Diagnostics. His research focuses on minimally invasive and robotic-assisted endourologic procedures, and on digital health solutions, including AI-based decision support tools. Prior to this position he headed the department for Prostate Cancer Diagnostics & BPH Treatment. Prof. Dr. Christian Wetteraurer was appointed Professor of Urology at the Danube Private University (DPU), Krems in 2022. Abstract: Clinical application of AIPC in urology Management of prostate cancer patients is a complex process involving multi-disciplinary teams. Efficient exchange of information among these departments is crucial. However, in most organizations, relevant data is stored disparate and siloed in various IT systems. In daily clinical practice, clinicians spend a substantial amount of time collecting, integrating, and assessing patient data in order to care for their patients. Effective data integration tools that extract and combine data from multiple sources as well as adequate data represenation are required. In the last years, we have worked together with an industry partner, Siemens Healthineers, on prototypes for AI supported pathway-specific clinical decision support systems based upon best practice as defined by the EAU prostate cancer guidelines. This work consisted of both data mapping, data integration and development of front-end data representation. We evaluated the first CE-certified version of the clinical decision support software AI-Pathway Companion Prostate Cancer in the context of prostate cancer treatment decision making. © HUEBL Prof. Dr. Christian Wetterauer Professor of Urology, Danube Private University, Krems, Austria
17 Symposium 2022 Curriculum Vitae: Prof. Dr. Constantin von See was appointed as Director of the Center for CAD/CAM and Digital Technologies in Dentistry at Danube Private University (DPU) Krems in 2014. Having completed his studies in Dentistry (Dr. med. dent.), he went on to obtain his medical license at Georg-August University, Göttingen, Germany in 2001. He is an oral surgeon and completed his PhD on experimental studies in the field of “Reactive modifications of bone and soft tissue after implanting hydrogel expanders” at the Clinic for Cranio-Maxillo-Facial Surgery of Hannover Medical School in 2011. Prof. Dr. Constantin von See also holds a Master’s Degree in Health Management. Since 2016 he is Head of the Master’s Program for Esthetic Reconstructive Dental Medicine at the Danube Private University (DPU) Krems, Austria. Abstract: Artificial intelligence in dentistry - ready for everyday practice? Background: Machine Learning (ML) and Deep Learning (DL) applications have been introduced to dental radiology and dental healthcare more than two decades ago. With the digitalization in dentistry and here especially intraoral scanning new opportunites were applicable. Material and Methods: A literature review on dental applications in ML and DL was performed to have an overview on the lastest developments and detect trends in artifical intelligence in dentistry. Results: Most applications are found nowadays in orthdontic treatment applications. In recent years dental radiology has become a new field in research. Discussion: Even though more and more DL applications are available in dentistry main aspects in mathematical and ethical considerations are not fully thought of today to specify applications in dentistry © Nik Pichler Univ.-Prof. Dr. Constantin von See, MaHM Director CAD/CAM and Digital Technologies in Dentistry, Danube Private University, Krems, Austria
Medical Image Analysis & Artificial Intelligence Posterwalk Summary
Medica l Image Ana lysis & Ar t i f icia l Intel l igence Symposium 2022 Nr. 1: Maria Delgado-Ortet, PhD: Lesion-Specific 3D-Printed Moulds of Pelvic Tumours for Image-Guided Nr. 2: Laszlo Jaksa, MSc: 3D Printer Enabling Customized Anatomic Models Nr. 3: Dr. med. dent. Richard Mosch: 3D-Gedruckte Zähne – der neue Standard? Nr. 4: Ass.-Prof. Dr. Sepideh Hatamikia: 3D printed silicone phantoms with controllable MRI signal properties Nr. 5: Ass.-Prof. Dr. Sepideh Hatamikia &Dr. Geevarghese George: Feature selection strategies for improved predictions of histopathologic response to neoadjuvant chemotherapy in patients with ovarian cancer Nr. 6: Ass.-Prof. Dr. Sepideh Hatamikia: Toward on-the-fly trajectory optimization for Carm CBCT under strong kinematic constraints Nr. 7: Dr. Olgica Zaric, PhD, MSc: Tissue Sodium Concentration Quantification at 7.0-T MRI as an Early Marker for Chemotherapy Response in Breast Cancer: A Feasibility Study Nr. 8: Dr. Olgica Zaric, PhD, MSc: Quantitative sodium 23Na-MRI– an improved protocol for breast cancer treatment efficacy evaluation Nr. 9: Dipl.-Ing. Philipp Lazen, BSc: B1+ Corrected Metabolite Concentration Estimates from 7 T FID CRT MRSI Nr. 10: Dipl.-Ing. Philipp Lazen, BSc: Comparing 7 T FID CRT MRSI with Amino Acid PET Nr. 11: Stefanie Kaser, MSc: Ion CT image reconstruction with the TIGRE toolbox Nr. 12: Alexander Seper: Deep learning based diagnostic model for prostate and colon carcinoma Nr. 13: Prof. DDr. Ramona Woitek: Radiomics-based predictors of response to neoadjuvant chemotherapy in patients with high grade serous ovarian cancer Nr. 14: Prof. DDr. Ramona Woitek: Hyperpolarised carbon-13 MRI allows early response assessment in patients with breast cancer undergoing neoadjuvant treatment Nr. 15: Ass.-Prof. Dr. Sepideh Hatamikia &Dr. Florian Schwarzhans, PhD, MSc: Evaluation of adiomics robustness on mammograms using different modifications of manually drawn regions of interest Nr. 16: Dr. Florian Schwarzhans, PhD, MSc: Real time eye tracking and blinking compensation for artefact free acquisition of polarization sensitive OCT volumes Nr. 17: Dr. Amirreza Mahbod, PhD: A Two-Stage U-Net Algorithm for Segmentation of Nuclei in H&E-Stained Tissues Nr. 18: S. M. Ragib Shahriar Islam: Learning Cytoarchitectonic Structure From 3D Polarized Light Imaging Nr. 19: Dr. med. dent. Dragan Ströbele: Development of an algorithm to predict the force progression of 3d-printed orthodontic springs Nr. 20: Prof. Dr. Christoph Kleber: Accelerated Combinatorial Medical Material Development through Machine Learning from Theoretical and Experimental Data Sets Nr. 21: Lorenz Perschy: Prediction of breast cancer in a high-risk patient cohort using deep learning in MR imaging data
Medica l Image Ana lysis & Ar t i f icia l Intel l igence Symposium 2022 Nr. 1: Maria Delgado-Ortet, PhD: Lesion-Specific 3D-Printed Moulds of Pelvic Tumours for Image-Guided Tissue Sampling of High Grade Serous Ovarian Carcinoma Curriculum Vitae: Maria Delgado-Ortet graduated in Biomedical Engineering from the University of Barcelona and completed postgraduate courses at Chalmers University of Technology (Sweden) and Trinity College Dublin (Ireland). She joined the University of Cambridge as a PhD student in joint collaboration between the Department of Radiology Radiogenomics and Quantitative Imaging Group and the Department of Engineering. Maria’s doctoral research is mainly focused on understanding spatial and temporal tumour heterogeneity in High Grade Serous Ovarian Carcinoma (HGSOC) applying artificial intelligence (AI) and image analysis. She has led the technical development of image-guided tissue sampling techniques for HGSOC both in vivo –multimodal image registration for realtime ultrasound-guided biopsies using CTderived maps– and ex vivo –lesion-specific 3D-printed moulds for image-guided sampling of tumours surgically resected. Abstract: Lesion-Specific 3D-Printed Moulds of Pelvic Tumours for Image-Guided Tissue Sampling of High Grade Serous Ovarian Carcinoma High-Grade Serous Ovarian Cancer (HGSOC) is the most prevalent and lethal subtype of ovarian cancer. Integration of radiology and tissue-derived (histopathological and genomic) data can improve prediction of patient outcome and treatment response. With the aim of allowing detailed spatial correlation of imaging and molecular pathology data, we introduce a computational pipeline for 3D printing lesion-specific moulds of pelvic gynaecological lesions. The moulds are modelled and printed prior to surgical resection through an automated computational pipeline that computes the shape and size of the delineated tumour on routinely acquired CT or MR imaging. A prospective pilot study on five patients with suspected HGSOC scheduled for debulking surgery in Addenbrooke’s Cambridge University Hospital concluded the successful development of a precision tissue sampling technique to guide the sampling of resected tumours that can be integrated in the clinical routine for patients with HGSOC which involves radiology, surgery and pathology.
Medica l Image Ana lysis & Ar t i f icia l Intel l igence Symposium 2022 Nr. 2: Laszlo Jaksa, MSc: 3D Printer Enabling Customized Anatomic Models Curriculum Vitae: Laszlo Jaksa holds a BSc and an MSc in Mechatronics from the Budapest University of Technology and Economics. Currently he is a researcher at the Austrian Center for Medical Innovation and Technology, and a PhD student at the Technical University of Vienna, Institute of Lightweight Design and Structural Biomechanics. His main research focus is using a combination of thermoplastic and silicone rubber 3D-printing for various medical applications, especially anatomic models. His side activities involve medical device prototyping and ceramic additive manufacturing. Abstract: 3D Printer Enabling Customized Anatomic Models In additive manufacturing, a frequently reported application is the production of anatomical models. One emerging technology is extrusion-based silicone rubber 3D-printing, which allows the customization of mechanical and radiological properties of printed objects. In this study, the abilities of a custom-built 3D-printer were investigated. Test objects were printed with three different silicone rubbers, focusing on the sagging of unsupported overhangs and bridges as a function of material viscosity. It was observed that the silicone with the highest viscosity had the highest average sagging both in the overhang and the bridge specimens. Tuning both hardness and radiological properties through infill structuring was also investigated with rectangular blocks of various gyroid infill densities. These were subject to Shore A hardness measurement and computed tomography. Both the Hounsfield Units and the hardness values were heavily inf luenced by the infill percentage, making this method potentially useful in producing realistic anatomic models.
Medica l Image Ana lysis & Ar t i f icia l Intel l igence Symposium 2022 Nr. 3: Dr. med. dent. Richard Mosch: 3D-Gedruckte Zähne – der neue Standard? Curriculum Vitae: Ass. Prof. Dr. med. dent. Richard Mosh has been working on the further development of digital dentistry in the Digital Technologies/CADCAM team for six years, first as a research assistant and then as a dentist. He has researched the technical development of digital fabrication in all areas of dentistry, whether in guided implantology, printed crowns and bridges or in complete dentures. Since 2021 he has been working as a senior physician. Abstract: 3D printed teeth -the new standard? The case shown illustrates the possibilities of 3D printing as a solution for aesthetic, temporary and cost-effective dental restorations. The ability to print restorations in a short time enables maximum individuality in modern dentistry. Thanks to the much more efficient data transfer via the network, distances can be saved and the high use of materials can be reduced. Thanks to the further development in the field of 3D printing, a precise and cost-effective alternative to the production of first temporary and in the future also definitive prosthetic restorations is possible. It is clear that digitalisation has already found its way into dentistry and will undoubtedly be used more and more in everyday practice in the future with its diverse application possibilities.
Medica l Image Ana lysis & Ar t i f icia l Intel l igence Symposium 2022 Nr. 4: Ass.-Prof. Dr. Sepideh Hatamikia: 3D printed silicone phantoms with controllable MRI signal properties Curriculum Vitae: Ass.-Prof. Dr. Sepideh Hatamikia completed her PhD and also a postdoctoral position in biomedical engineering field at Medical University of Vienna at the Department of Medical Physics and Biomedical Engineering at the Digital Image Processing Laboratory. Since march 2022, she joined Danube Private University (DPU) and currently she is the Head of Computational Imaging in the Medical Image Analysis and Artificial Intelligence (MIAAI) research group. In addition, Ass.-Prof. Dr. Hatamikia is a researcher at Austrian Center for Medical Innovation and Technology (ACMIT) since January 2018 and continues collaboration with this research center on different research projects. Abstract: 3D printed silicone phantoms with controllable MRI signal properties Abstract: Selection of materials for development of Magnetic Resonance Imaging (MRI) phantoms is always very challenging. So far, mainly gel-based or water-filled compartments have been used for MRI phantoms. 3D printing is a promising approach which allows manufacturing potential materials with controllable MRI signal properties to mimic geometric structures and lesion heterogeneities. The aim of this study is to investigate the feasibility of using extrusion 3D printed silicone materials to mimic human tissue and its range of relaxation times on MRI.
Medica l Image Ana lysis & Ar t i f icia l Intel l igence Symposium 2022 Nr. 5: Ass.-Prof. Dr. Sepideh Hatamikia &Dr. Geevarghese George: Feature selection strategies for improved predictions of histopathologic response to neoadjuvant chemotherapy in patients with ovarian cancer Curriculum Vitae: Dr. Geevarghese George obtained his PhD in Physics from the University of Strasbourg (France) for his work at the Institut Charles Sadron (Theory and Simulation of Polymers group, CNRS), supervised by Dr. Joachim Wittmer. His thesis was on the numerical study of statistical and rheological properties of polymer films. During his PhD, he also developed an interest in machine learning and pivoted into the field through self-directed projects and coursework. His research interest at MIAAI (DPU) is in the area of machine/deep learning for medical image based diagnosis/prognosis. Curriculum Vitae: Ass.-Prof. Dr. Sepideh Hatamikia completed her PhD and also a postdoctoral position in biomedical engineering field at Medical University of Vienna at the Department of Medical Physics and Biomedical Engineering at the Digital Image Processing Laboratory. Since march 2022, she joined Danube Private University (DPU) and currently she is the Head of Computational Imaging in the Medical Image Analysis and Artificial Intelligence (MIAAI) research group. In addition, Ass.-Prof. Dr. Hatamikia is a researcher at Austrian Center for Medical Innovation and Technology (ACMIT) since January 2018 and continues collaboration with this research center on different research projects. Abstract: Feature selection strategies for improved predictions of histopathologic response to neoadjuvant chemotherapy in patients with ovarian cancer Abstract: Response to neoadjuvant chemotherapy (NACT) in patients with high grade serous ovarian carcinoma (HGSOC) is evaluated histopathologically using the chemotherapy response score (CRS) for omental deposits. Recent radiomics-based models have shown promise in predicting CRS. This study investigates how different feature selection strategies improve CRS prediction. Our results demonstrated that feature selection methods can improve the prediction of response to NACT based on CRS in terms of accuracy, sensitivity and specificity.
Medica l Image Ana lysis & Ar t i f icia l Intel l igence Symposium 2022 Nr. 6: Ass.-Prof. Dr. Sepideh Hatamikia: Toward on-the-fly trajectory optimization for Carm CBCT under strong kinematic constraints Curriculum Vitae: Ass.-Prof. Dr. Sepideh Hatamikia completed her PhD and also a postdoctoral position in biomedical engineering field at Medical University of Vienna at the Department of Medical Physics and Biomedical Engineering at the Digital Image Processing Laboratory. Since march 2022, she joined Danube Private University (DPU) and currently she is the Head of Computational Imaging in the Medical Image Analysis and Artificial Intelligence (MIAAI) research group. In addition, Ass.-Prof. Dr. Hatamikia is a researcher at Austrian Center for Medical Innovation and Technology (ACMIT) since January 2018 and continues collaboration with this research center on different research projects. Abstract: Toward on-the-f ly trajectory optimization for C-arm CBCT under strong kinematic constraints Abstract: Kinematic constraints are common while acquiring C-arm cone beam computed tomography (CBCT). Such constraints cause collisions with the imager while performing a full circular rotation and therefore eliminate the chance for having 3D imaging in CBCT-based interventions. As collisions are mainly unpredictable in the operation theater, a framework which enables a real-time trajectory optimization is of great clinical importance. In this study, we introduce a new search strategy which has the potential to optimize trajectories on-the-f ly while avoiding potential kinematic constraints.
Medica l Image Ana lysis & Ar t i f icia l Intel l igence Symposium 2022 Nr. 7: Dr. Olgica Zaric, PhD, MSc: Tissue Sodium Concentration Quantification at 7.0-T MRI as an Early Marker for Chemotherapy Response in Breast Cancer: A Feasibility Stud Curriculum Vitae: Dr. Olga Zaric has been dedicated to the development and implementation of advanced magnetic resonance imaging (MRI) techniques, such as diffusion-weighted imaging (DWI), sodium (23Na), and chemical exchange saturation transfer (CEST) imaging in breast tumor characterization and treatment efficacy monitoring. Her future research will be focused on retrospective analyses of the radiomics derived from patient data who have undergone multiparametric MRI scans and on prospective studies, such as sodium fingerprinting imaging and employment of deep learning techniques to develop predictors for breast lesions diagnosis and classification and early response to immune checkpoint inhibitors (ICI) treatments. The studies will be performed in a collaboration of DPU and High Field MR Centre, MUV as part of joint projects whose preparation is currently in progress. Abstract: Tissue Sodium Concentration Quantification at 7.0-T MRI as an Early Marker for Chemotherapy Response in Breast Cancer: A Feasibility Study In this study, we investigated the feasibility of the imaging protocol tissue sodium concentration (TSC) quantification at 7.0-T MRI as an early marker for chemotherapy response follow-up in women with breast cancer. We aimed to explore whether the early changes of TSC can better predict a pathologic complete response (pCR) in women with locally advanced breast cancer than can changes in tumor size. When comparing our quantitative results with histologic findings, we found that the area under the receiver operating characteristic curve for tissue sodium concentration (TSC) reduction can better differentiate between participants with and without a pCR after the first chemotherapy cycle than can two-dimensional tumor size reduction. We, therefore, conclude that TSC is a promising imaging marker for the identification of patients with a poor prognosis and advanced progression of breast cancer, and it could play a crucial role in improving personalized breast cancer treatment in the future.
Medica l Image Ana lysis & Ar t i f icia l Intel l igence Symposium 2022 Nr. 8: Dr. Olgica Zaric, PhD, MSc: Quantitative sodium 23Na-MRI– an improved protocol for breast cancer treatment efficacy evaluation Curriculum Vitae: Dr. Olgiac Zaric has been dedicated to the development and implementation of advanced magnetic resonance imaging (MRI) techniques, such as diffusion-weighted imaging (DWI), sodium (23Na), and chemical exchange saturation transfer (CEST) imaging in breast tumor characterization and treatment efficacy monitoring. Her future research will be focused on retrospective analyses of the radiomics derived from patient data who have undergone multiparametric MRI scans and on prospective studies, such as sodium fingerprinting imaging and employment of deep learning techniques to develop predictors for breast lesions diagnosis and classification and early response to immune checkpoint inhibitors (ICI) treatments. The studies will be performed in a collaboration of DPU and High Field MR Centre, MUV as part of joint projects whose preparation is currently in progress. Abstract: Quantitative sodium 23Na-MRI– an improved protocol for breast cancer treatment efficacy evaluation Sodium concentration changes in tissue can be considered as a reliable biomarker for cell viability and functioning, while relaxation properties of sodium in the tissue might be sensitive to tissue microstructure or changes due to different pathological conditions. This work aims to develop a clinically reliable protocol for quantitative assessment of treatment efficacy, which is non-invasive, accurate, and obtained with sodium magnetic resonance imaging (23Na-MRI) and magnetic resonance fingerprinting (MRF). An initial protocol was generated for the endocrinological disorders study and investigation of tissue sodium concentrations (TSC) of the calf muscle and modified to be applicable for breast cancer immune checkpoint inhibitor (ICI) treatment efficacy monitoring using 23Na-MRI. When the new protocol was followed, we found that the difference between uncorrected and corrected TSC values is approximately 25%. In conclusion, the accuracy of the method might be increased by including additional post-processing steps and tissue compartment segmentation in TSC calculations.
Medica l Image Ana lysis & Ar t i f icia l Intel l igence Symposium 2022 Nr. 9: Dipl.-Ing. Philipp Lazen, BSc: B1+ Corrected Metabolite Concentration Estimates from 7 T FID CRT MRSI Curriculum Vitae: Dipl.-Ing. Philipp Lazen, BSc is a PhD student in his second year at the Medical University of Vienna. He works on MR spectroscopic imaging (MRSI). For his master thesis I developed a pulse simulation model which can be used to test different excitation pulses in a numerical phantom. His PhD focuses on various approaches towards improving MRSI, such as the application of higher order shim coils as well as post processing methods like B1 correction. Additionally, he is involved in various clinical studies involving clinical patients and healthy volunteers. As such, his work unites aspects of software development and coding, data science and analysis, and biology and medicine. Abstract: B1 + Corrected Metabolite Concentration Estimates from 7 T FID CRTMRSI Magnetic resonance spectroscopic imaging (MRSI) is an imaging modality similar to MR imaging that can map distributions of metabolites in tissues, such as the brain. Conventionally, MRSI offers only qualitative measurements, but no absolute metabolite concentrations. As a remedy, metabolite ratios are commonly used to provide better comparability between subjects and MR scanners, but a more sophisticated solution is the calculation of metabolite concentration estimates (CE) based on internal water referencing. Additionally, MRSI – like all MR based modalities – can suffer from inhomogeneities of the involved magnetic fields, which reduce the reliability of the acquired results. For example, magnetic excitation pulses are naturally imperfect and result in a spatially inhomogeneous transmit field B1 +, causing the resulting signal strength in a voxel to vary with its location. In this study, we calculated metabolite CEs in a cohort of healthy volunteers and corrected the resulting values for B1 + inhomogeneities.
Medica l Image Ana lysis & Ar t i f icia l Intel l igence Symposium 2022 Nr. 10: Dipl.-Ing. Philipp Lazen, BSc: Comparing 7 T FID CRT MRSI with Amino Acid PET Curriculum Vitae: Dipl.-Ing. Philipp Lazen, BSc is a PhD student in his second year at the Medical University of Vienna. He works on MR spectroscopic imaging (MRSI). For his master thesis I developed a pulse simulation model which can be used to test different excitation pulses in a numerical phantom. His PhD focuses on various approaches towards improving MRSI, such as the application of higher order shim coils as well as post processing methods like B1 correction. Additionally, he is involved in various clinical studies involving clinical patients and healthy volunteers. As such, his work unites aspects of software development and coding, data science and analysis, and biology and medicine. Abstract: Comparing 7 T FID CRTMRSI with Amino Acid PET Magnetic resonance spectroscopic imaging (MRSI) and positron emission tomography (PET) are both imaging modalities capable of going beyond the mere structural contrasts delivered by other modalities, instead delivering information about metabolic processes in physiological and pathological tissues. To achieve this, MRSI uses the slightly different resonance frequencies that different metabolites have in an external magnetic field, and PET uses radioactively marked amino acids as tracers. Thus, both approaches can map the activity of cells, which is often elevated in tumors. Due to their similar nature, we decided to compare the results from our 7T MRSI with clinical PET scans in glioma patients. This poster is based on a publication which represented the first quantitative comparison of 7T MRSI and PET.