Sarcoma AI
Publications

Learning tumor representations designed to remain reliable across hospitals, scanners and heterogeneous acquisition protocols.

Combining MRI with molecularly anchored ground truth for clinically meaningful diagnostic modelling.

Investigating clinically meaningful representations from whole-slide pathology images of rare soft-tissue tumors.

Learning unified patient representations from imaging, pathology, molecular biomarkers and structured clinical information.

Building calibrated and explainable systems that communicate uncertainty and identify cases requiring specialist review.