Deep learning models developed for gastric cancer prognosis and immunotherapy response prediction showed improved risk stratification using digital pathology. The study, published in the Journal of Translational Medicine, reported models that integrate histopathology features to guide prognosis and potential immunotherapy benefit. Researchers validated models on retrospective cohorts and highlighted implications for trial enrichment and personalized treatment planning. The work underscores digital pathology and AI as decision tools in oncology workflows, while raising questions about external validation, regulatory clearance, and integration into clinical practice.