Virchow / Virchow2 (Paige, MSK)
A pathology foundation model trained on millions of slides that can detect cancer and predict biomarkers from an ordinary H&E slide.
Virchow (2024, 632M parameters, 1.5M slides) and Virchow2/2G (2024, up to 1.9B parameters, 3.1M slides from MSK and global sites, mixed magnification) are the largest proprietary pathology foundation models; they underpin Paige's pan-cancer detection and biomarker products and were trained with Microsoft compute.
How it works
Self-supervised DINOv2 vision transformer pretraining on tissue tiles at several magnifications; frozen encoder plus small task heads.
- Scale and data diversity
- Strong biomarker prediction from H&E
- Proprietary weights
- Scanner and stain shift
Latest papers
topQuery for this technology: (TITLE:"Virchow / Virchow2" OR ABSTRACT:"Virchow / Virchow2" OR TITLE:"Paige, MSK" OR ABSTRACT:"Paige, MSK") AND (cancer OR tumor OR tumour OR oncology OR carcinoma OR lymphoma OR leukemia OR leukaemia OR myeloma OR sarcoma OR melanoma OR glioma). Results are unfiltered search hits about Virchow / Virchow2 (Paige, MSK), not a curated reading list.