Deep Learning-Based Spatial Profiling Identifies MET-Associated Immune Exclusion and Therapeutic Vulnerability Underlying Racial Disparities in Head and Neck Cancer

Black/African American (BAA) patients with head and neck squamous cell carcinoma (HNSCC) experience more aggressive disease and poorer survival outcomes than White patients; however, the tumor immune microenvironmental features underlying these racial disparities remain poorly characterized. Here, we identify MET as a potential molecular link between immune exclusion and therapeutic response associated with racial disparities in HNSCC. Compared with White-derived tumors, deep learning-based spatial profiling revealed a distinct immune landscape in BAA-derived tumors, characterized by reduced CD8⁺ T cell infiltration, increased immune checkpoint expression, and elevated spatial MET expression. Across independent cohorts, higher MET expression was consistently associated with reduced CD8⁺ T cell infiltration, supporting a relationship between MET-associated tumor biology and an immune-excluded tumor microenvironment. Importantly, in a phase 2 clinical trial cohort treated with pembrolizumab plus cabozantinib, higher predicted spatial MET expression was strongly associated with treatment response and improved clinical outcomes, with the association particularly pronounced among BAA patients. This study establishes a spatially resolved framework linking tumor-intrinsic MET expression with the immune microenvironment and therapeutic response, suggesting that MET-associated biology may contribute to racial disparities in HNSCC progression and treatment outcomes.