Abstract: Currently, the actual diagnostics and definition of major diseases like cancer is done through medical imaging, especially through the microscopic analysis of tissue sections and cells. For over a decade, pathology has been refraining from digital transformation, making critical treatment decisions is still dependent on the availability of high quality pathology reading. In a broader sense, the lack of an objective tissue image data collection, which can be used as basis for comparison, limits our understanding of diseases, diagnostics, therapy development and clinical trials across the board. Through its ability of learning highly complex patterns in pre-defined network structures, deep learning is expected to become the “killer application” for digital pathology. In interplay with conventional image analysis and cloud computing it will open the door for exploiting tissue as a huge and critical biomedical data source. From the perspective of computer science, such data will enhance computational tissue modelling as well as lead to diagnostic digital assays enabling medical decisions of globally standardized quality. In the talk, experiences from studying the tissue microenvironment in wound healing and cancer using digital pathology are shown, results and challenges discussed. Finally, current results from a deep learning based cervical cancer screening system under development with the US-NCI are presented.
Biography: Prof. Dr. Niels Grabe is the Scientific Director of the Hamamatsu Tissue Imaging and Analysis Center (TIGA) at the University of Heidelberg and scientific group leader at the National Center for Tumor Diseases. He is working in biomedical informatics and systems biology at the crossroads of cancer diagnostics and therapy, digital pathology, machine learning and computational modelling. The central activity of the TIGA Center is the high-throughput profiling of biomarkers in research projects and clinical trials with a focus on immunotherapy. He has published over 90 papers in international journals and started one of the first conference series on Digital Pathology in Europe in 2009. He is collaborating closely with the US National Cancer Institute and several industrial companies. In 2013 he founded the Steinbeis Center for Medical Systems Biology (STC-MSB) as a spin-off of the University of Heidelberg for translating his research into industrial applications.