Highlight
This article introduces TESERA, a novel computational pathology framework that extracts prognostically significant morphologic subtypes from hematoxylin and eosin (H&E) stained slides of hepatocellular carcinoma (HCC). TESERA identifies two distinct tumor subtypes linked to key molecular pathways and mutations, and provides a prognostic morphologic index predictive of overall and disease-free survival, validated across multiple cohorts.
TESERA’s approach utilizes routinely acquired histologic images, enabling low-cost and widespread biology-aware risk stratification, potentially guiding clinical management and therapeutic decisions in HCC.