In the case of advanced skin cancer, treatment decisions are often based on the personal experience of the treating physician due to a lack of evidence. Significant progress can be made here with a decision support system that is able to support physicians in the selection of treatment methods by analysing individual cases with artificial intelligence (AI) methods. For this purpose, previous cases form the basis of experience, which are documented in the form of treatment data in clinical information systems. Decision support systems could thus contribute to ensuring better oncological treatment and replace costly observational studies in the future.
However, despite many years of research efforts, medicine is still at the beginning in the use of AI. This is mainly due to the lack of availability of high-quality data, data protection concerns and requirements for the explainability of results.
This project makes a new contribution to the use of AI, in particular the methodology of case-based reasoning, in the care of skin cancer patients and forms the basis for the development of an assistance system for medical professionals that offers the possibility to search, find and present known treatment cases based on relevant, complex medical criteria while maintaining explainability.
To this end, using the example of the Skin Tumour Centre Münster, it will be investigated which data from the documentation systems can be evaluated by which artificial intelligence techniques. This will create an important basis for the subsequent definition of a corresponding cooperative interdisciplinary research project.
In summary, the following objectives are pursued in Pre-OnkoCase-Project:
• Data analysis and quality assessment of Hospital Information System data.
• Evaluation of semantification and information extraction techniques of Hospital Information System data using the ADT/GEKID base dataset, classifications and ontologies.
• Klinik für Hautkrankheiten Universitätsklinikum Münster
• Institut für Medizinische Informatik Münster