Integrating and Contextualizing Multi-Modal Data for Personalized Medicine
Loading...
Author
Lamarine, Marc
Poussin, Carine
DOI
Abstract
The growing availability of heterogeneous biomedical data presents both an opportunity and a challenge for personalized medicine. This talk presents a multi-modal data integration framework designed to support translational research and clinical decision-making in complex diseases. By combining clinical variables, imaging data, molecular profiles, and functional assays within unified AI-driven models, the approach enables robust patient stratification and outcome prediction. A central case study focuses on frontotemporal dementia, where graph-based data fusion and machine learning are applied to identify predictive liquid biomarkers and disease subtypes. Emphasis is placed on interdisciplinary workflows, model validation, and deployment practices required for real-world clinical use. The presentation highlights how contextualizing multi-modal data within biologically and clinically meaningful representations can transform fragmented datasets into actionable insights, accelerating biomarker validation and supporting precision diagnostics and personalized therapeutic strategies.
Publication Reference
LSZ Impact 2025, Zurich, Switzerland, 26 May 2025
Year
2025