Findings from the PM4Onco Workshop on Determining Patient Similarity Search Strategies for Personalized Therapy in...

The PM4Onco project advances precision oncology by developing analytical and visual methods to support data-driven therapy decisions in Molecular Tumor Boards (MTBs). An interdisciplinary workshop at Martin Luther University Halle-Wittenberg in March 2025 explored strategies for defining and integrating patient similarity into MTB workflows. Using a synthetic oncology dataset, participants with clinical and data science backgrounds defined similarity criteria and structured search strategies for patient similarity. The workshop outcomes provide structured decision frameworks that operationalize patient similarity reasoning in a reproducible manner. Their implementation supports transparent and evidence-based analysis, fostering integration between algorithmic modeling and clinical expertise.