Razik Yousfi

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 from 700,000 pathology reports.

Discover how AI is revolutionizing cancer care through precision diagnosis, open-source innovation, and enhanced patient outcomes. Many of us know someone who has been affected by cancer. Groundbreaking AI technology is already here that can transform a cancer diagnosis—where AI detects even the tiniest and rarest cancers invisible to the human eye, maps their extent with pinpoint accuracy, and enables oncologists and pharmaceutical teams to deliver highly targeted, precision treatment.

Pathology is undergoing a significant shift. While there’s enormous potential to improve efficiency with the use of artificial intelligence (AI), the field also faces a critical challenge: the workforce is struggling to meet the demand. The pathology workforce is shrinking, with too few trainees entering the field to replace those retiring or transitioning out.