Researchers at Harvard Medical School unveiled COMPASS, an AI model designed to predict patient response to immune checkpoint inhibitors (ICIs). Reported as outperforming existing approaches, COMPASS is built by Associate Professor Marinka Zitnik’s team and aims to improve selection of patients likely to benefit from ICI therapies. The project uses data-driven modeling to capture features tied to immunotherapy outcomes, with the stated goal of strengthening personalized cancer treatment decisions. For oncology programs, accurate response prediction can affect trial design, patient stratification, and real-world prescribing patterns. The work contributes to the growing portfolio of computational tools targeting biomarker gaps in immuno-oncology—where response rates remain limited and toxicity risk is clinically meaningful.