Project 1 New Dimensions in Understanding Therapy Response and Relapse for Breast Cancer and Rectal Cancer Patients using Digital Technologies Principal Investigators: Prof. Inti Zlobec (Institute of Tissue Medicine and Pathology, University of Bern) and Prof. Sven Rottenberg (Institute of Animal Pathology, University of Bern). Total funding (w/o matching): CHF 1,000,000. This project tackles residual disease and tumor regression after neoadjuvant therapy by developing spatially informed, morpho-molecular predictors that extend beyond current biomarker paradigms. By combining established 2D pathology with novel 3D volumetric histopathology, spatial transcriptomics, and AI-based analysis, we characterize epithelial–mesenchymal transition states in residual human tumors at single-cell resolution. The integrative models generated in this work will be validated in patient cohorts and lay the groundwork for spatially driven, clinically actionable relapse prediction in breast and rectal cancer. "There is a tremendous wealth of knowledge hidden inside tissue samples that to date remains unexplored. Our work will provide an example of how advanced digital technologies and AI can unlock this information and provide a better understanding of relapse prediction after cancer therapy," says Prof. Inti Zlobec.
Project 2 LLM4Humans - Empowering humans with local LLMs in research and clinical practice Principal Investigator: Prof. Benjamin Ineichen (Department of Clinical Research, University of Bern). Co-Applicants: Prof Sören Huwendiek (Institute for Medical Education, University of Bern). Total funding (w/o matching): CHF 500,000. This project develops secure, locally run language models that help researchers and clinicians work more efficiently with sensitive unstructured data, such as radiology reports and qualitative research interviews. These models support tasks like summarising clinical information, structuring large text collections, and assisting with analysis while keeping all data protected on local infrastructure. The project will also create clear guidance for responsible and effective LLM use across the University of Bern. “Our aim is simple: give people AI tools they can trust with the sensitive work they do every day,” says Prof Benjamin Ineichen. Department of Clinical Research (DCR)
Project 3 A Human-in-the-Loop approach to Clinical Decision Support in Radiation Oncology Principal Investigators: Prof. Sarah Brüningk (Center for AI in Radiation Oncology, Inselspital, University of Bern) and Prof Benjamin Ineichen (Department of Clinical Research, University of Bern). Co-Applicants: Prof. Mauricio Reyes (ARTORG Center for Biomedical Engineering Research, University of Bern) Prof. Fred Mast (Department of Psychology, University of Bern) & Prof Olgun Eliçin (Department of Radiation Oncology, Inselspital Bern). Total funding (w/o matching): CHF 500,000. This project focuses on building a clinical decision support system for radiation oncology that links predictive models, mechanistic simulations, and evidence retrieval into one interactive platform. The tool lets clinicians explore patient-specific scenarios, compare treatment options, and review the underlying evidence in a transparent way. It is designed to support clinical reasoning, not automate it, and will be tested in real-world head and neck cancer care. "We want decision support that doctors can question, probe and shape. This project aims to build AI that strengthens clinical reasoning rather than replacing it." , says Prof Sarah Brüningk. Department of Clinical Research (DCR)
