Ben Glass, MS

Head of AI, Product and Research at Roche

Computational Pathology at Scale: From Discovery to Drug Development to Global Diagnostics

Computational pathology now serves three distinct users: the scientist mining tissue for biology, the drug developer running trials across hundreds of sites, and the pathologist signing out a case in clinical practice. Each demands a different solution. Discovery favors automation and breadth: quantitative morphology extracted across indications and fused at scale with genomic, clinical, and other high-dimensional data. Drug development adds regulatory rigor and a platform that can support the operations of a global clinical trial, from site onboarding to locked data. Clinical diagnostics requires deep integration into the pathologist’s workflow and the operations of the anatomic pathology laboratory. This talk draws on a decade of building products in all three settings to show what one modeling foundation can share, where each setting forces its own design, what it takes to move from a research pipeline to AI running in laboratories worldwide, and how a shared digital infrastructure delivers long-term benefits to discovery and clinical practice alike.