Personalized Medicine through AI: Integrating Multi-Modal Data for Revolutionizing Healthcare
Loading...
Author
Poussin, Carine
DOI
Abstract
Personalized medicine relies on the effective integration of heterogeneous biomedical data to support clinical decision-making in complex diseases. This talk introduces a multi-modal AI platform developed at CSEM for patient stratification and outcome prediction, combining clinical variables, molecular profiles, imaging-derived features, and functional assay data. A central use case focuses on multiple myeloma, where patient similarity networks and graph-based representation learning are employed to predict individualized drug responses using pharmacoscopy and molecular data. The approach demonstrates how multi-modal data fusion enables robust prediction of therapeutic efficacy and adverse effects, even in heterogeneous patient cohorts. Emphasis is placed on adaptability, explainability, and translational readiness, highlighting the platform’s applicability across disease areas. The work illustrates how AI-driven multi-modal integration can transform complex biomedical data into actionable insights, supporting precision diagnostics and personalized treatment strategies.
Publication Reference
AI in Healthcare and Life Sciences Conference, 10 June 2025, Prague, Czech Republic
Year
2025