Owkin's MesoNet model predicts mesothelioma survival from pathology images and could shrink Phase 3 trial sizes.

Mesothelioma lacks prognostic biomarkers and pathologist subtype classification proved inadequate for decisions. Owkin built MesoNet, a machine learning model predicting overall survival from digital pathology images and surfacing new morphological biomarkers. The page states it outperforms pathologist classification and enables stratification that reduces Phase 3 sample sizes by 6-13%.

Owkin is a French AI company building biology models and the K Pro AI Scientist, working with pharma on drug positioning, trial design and pathology diagnostics.

This sits in the Build stage of the AI use-case lifecycle: Named examples from the foundation-model labs and platforms pharma teams use to actually construct, train and generate the model or application.