Large Language Models in medicine – researchers show how AI can improve healthcare in the future
Unique at a Medical Faculty in Germany
Unique at a Medical Faculty in Germany
The internist and scientist Jakob Nikolas Kather will fill the new Else Kröner Professorship for “Clinical Artificial Intelligence” at the EKFZ for Digital Health at Dresden University of Technology from June 1, 2022. With his expertise and his team the award-winning physician from Aachen strengthens the research on artificial intelligence (AI) in clinical practice at the University Medical Center Dresden. To ensure a direct link to everyday medical practice, Prof. Kather will work in Medical Clinic I at Dresden’s University Hospital, where he will care for patients. With the new professorship, one of the main concerns of the EKFZ for Digital Health will be implemented – research that goes beyond the medical disciplines as well as collaboration in everyday care: physicians learn programming and researchers from computer science or technical subjects learn in return to identify and solve relevant clinical problems.
The Else Kröner Fresenius Center for Digital Health is a unique institution in Germany. Here, the interaction between high-tech and medicine is institutionalized and professionalized in interdisciplinary research teams. My team and I want to contribute to making TU Dresden a leading location for AI in medicine,” says Prof. Kather about his upcoming move.
His research focuses on the application of artificial intelligence in cancer in clinical practice. With his Computational Oncology research group, he bridges the gap between different specialties. Physicians learn programming, while researchers with informatics or engineering backgrounds learn to identify and solve relevant clinical problems. The aim is to further develop the evaluation and interpretation of complex image data and thus improve diagnostic and treatment approaches, for example in tumor diseases such as colorectal cancer or gastric cancer, but also in inflammatory diseases or in transplantation medicine. An important partner will be the National Center for Tumor Diseases (NCT/UCC), which is also located on the campus of the University Hospital. At the NCT, research into cancer diseases and the care of tumor patients is linked as closely as possible. In his previous position at the University Hospital of RTWH Aachen, Prof. Kather was already able to show that it is possible to derive medical recommendations for action from routinely available data by means of “Deep Learning”.
The background to the interdisciplinary collaboration envisaged at Dresden University Medical Center is to move quickly from idea to prototype and, with the help of artificial intelligence, to improve the diagnosis and treatment of cancer. “Thanks to the comprehensive support of the Else Kröner-Fresenius-Foundation and the professorship it finances, the Faculty of Medicine and the University Hospital can set standards in patient-oriented research on digital medicine,” says Prof. Michael Albrecht, Medical Director of Dresden University Hospital. “Prof. Kather and his team form another important pillar for the first truly integrated eHealth campus on the premises of a German university hospital. With the practical relevance and patient focus practiced here by all players, the conditions are in place for playing a decisive role in shaping the future of medicine.”
In the clinical care of patients, a huge amount of data is generated that is currently only partially used for clinical decision-making. In particular, image data such as pathological or radiological images contain a great deal of information, but other types of data such as text or laboratory values are also not currently used in their entirety. Artificial intelligence can detect subtle patterns in these data and therefore make them usable. This can help physicians derive more information from available data, which can be used to make better clinical decisions. On the one hand, this concerns the diagnosis of diseases, but also the classification into subtypes or disease stages as well as the prediction of disease progression. One example is tumor diseases, whose treatment has become increasingly complex in recent years. “It is important that the various steps are optimally interlinked: The identification of clinically relevant problems, the development of new AI methods, and the ultimate clinical testing and development of a medical product. This only works in an interdisciplinary environment with short paths and a common vision,” says Prof. Kather.
It is precisely this interdisciplinary work that is essential for the digital transformation in medicine. “With Prof. Jakob Kather, we were able to attract an outstanding scientist to TU Dresden who, together with his team, lives what TU Dresden stands for: Interdisciplinarity. Medicine combined with artificial intelligence is a forward-looking field of science that we are also addressing in the new Biomedical Engineering and Medical Informatics degree programs,” said Rector Prof. Ursula M. Staudinger. “With his research, Prof. Kather has achieved outstanding scientific recognition. With his expertise and his team, he is a great gain for the EKFZ and the University Hospital,” says Prof. Jochen Hampe, scientific spokesperson of the EKFZ for Digital Health as well as director of the Medical Clinic I of the University Hospital Dresden.
Over the next few years, a team of young and creative minds is to be built up on the university medicine campus and trained in an interdisciplinary manner. “My team and I would like to contribute to the EKFZ for Digital Health and the University Hospital Dresden becoming Germany’s leading center for artificial intelligence in clinical practice. The whole spectrum of clinical expertise is to be covered: the development of prototypes, clinical validation and regulatory aspects. The Clinical AI department is involved in the entire spectrum,” concludes Prof. Kather.
Large Language Models in medicine – researchers show how AI can improve healthcare in the future
A Busy Autumn for the HybridEcho Team
Symposium on Large Language Models in Medicine