AI in Oncology · AI Models

Successor to Virchow: ViT-H/14 pretrained on 3.1 million whole-slide images from about 225,000 patients across 45 countries, at mixed magnifications (5×–40×), with pathology-specific augmentations.

General-purpose self-supervised vision encoder for H&E histopathology tiles, pretrained on more than 100 million tiles from over 100,000 whole-slide images; the reference foundation model for pathology feature extraction.

632-million-parameter vision transformer pretrained on 1.5 million whole-slide images from about 100,000 patients — the largest pathology pretraining set at its release — and used to build a pan-cancer detection model covering 17 cancer types, including rare ones.

The first AI software for pathology authorised by the US FDA (De Novo, September 2021): flags prostate biopsy whole-slide images likely to contain cancer so the pathologist reviews the suspicious region.

Commercial AI reader for screening mammography, evaluated as an additional reader in the UK's double-reading pathway; the GEMINI prospective study in Aberdeen reported extra cancers found when Mia flagged cases the readers had cleared.

Commercial mammography AI with FDA clearance and CE marking, used as a concurrent or replacement reader in breast screening; the MASAI randomised trial in Sweden used it to triage screening reads.