Virchow: A Foundation Model for Pan-Cancer Detection and Rare Cancers

A new AI-based foundation model, Virchow, has been introduced for clinical-grade computational pathology. Trained on over a million whole-slide images (WSIs), Virchow excels in detecting nine common and seven rare cancers, achieving high accuracy even with limited data. It outperforms specialized clinical models for rare cancer detection and demonstrates promise for future diagnostic applications. This model could revolutionize cancer detection by reducing reliance on expensive testing methods.