Peer-Reviewed Publication
Nat Med2026July 31, 2026Journal Article
End-to-end multimodal pathology foundation model with clinical dialogue.
Eugene Vorontsov1,2, George Shaikovski3, Adam Casson3,4, Julian Viret3, Eric Zimmermann5, Neil Tenenholtz5, Yi Kan Wang3,4, Jan H Bernhard3,4, Ran A Godrich3,4, Juan A Retamero3,4, Jinru Shia6, Mithat Gonen6, Martin R Weiser6, David S Klimstra7, Razik Yousfi3,4, Nicolò Fusi5, Thomas J Fuchs3, Kristen Severson5, Siqi Liu3,4
1Paige, New York, NY, USA. eugene.vorontsov@paige.ai.
2Tempus, Chicago, IL, USA. eugene.vorontsov@paige.ai.
3Paige, New York, NY, USA.
4Tempus, Chicago, IL, USA.
5Microsoft Research, Cambridge, MA, USA.
6Memorial Sloan Kettering Cancer Center, New York, NY, USA.
7Yale University, New Haven, CT, USA.
Abstract
Recent rapid progress in the field of computational pathology has been enabled by foundation models. These models are beginning to move beyond encoding image patches toward whole-slide understanding, but their clinical utility remains limited. Here we present PRISM2, a multimodal slide-level foundation model trained on 2.3 million whole-slide images and 14 million question-answer pairs derived fro…
Create a free account to keep reading
Free members get 10 full research views every month across publications, clinical trials, FDA clearances, adverse events, and NIH grants. No credit card required.
Want unlimited research access? See Pro plans