Single-Cell Spatial Atlas of High-Grade Serous Ovarian Cancer Uncovers MHC Class II as a Key Predictor of Spatial Tumor Ecosystems and Clinical Outcomes

The tumor microenvironment in high-grade serous ovarian carcinoma (HGSC) is a complex network of malignant-host cell interactions, yet its orchestration remains poorly understood. We present a single-cell spatial atlas of metastatic HGSC from 280 patients, integrating high-dimensional imaging and molecular profiling. Analyzing 929 single-cell maps, we identify spatial domains with diverse cell compositions and show that immune cell coinfiltration at the tumor-stroma interface affects clinical outcomes. Using Cell Feature Importance Identification by RAndom forest (CEFIIRA), we find that tumor cell MHC class II (MHCII) expression is a key predictor of prolonged survival. Validation with deconvoluted single-cell and two distinct spatial transcriptomic datasets, along with immunopeptidomic analysis, confirms that MHCII expression correlates with immune activation, antigen presentation, and T-cell receptor clonality. Using a patient-derived immuno-oncology platform, we demonstrate that tumor MHCII expression associates with increased CD8+ T-cell cytotoxicity after PD-1 blockade, whereas blocking MHCII inhibits this activation. Our atlas offers new insights into immune activation, potentially improving patient stratification in HGSC. SIGNIFICANCE: This study delivers the first large-scale single-cell spatial atlas of HGSC, revealing how tumor-immune organization shapes outcomes. We identify tumor-intrinsic MHCII as a key driver of local immune activation and immunotherapy responsiveness, providing a mechanistic biomarker that can immediately inform improved patient stratification and therapeutic decision-making in ovarian cancer. See related commentary by Conejo-Garcia and Dangaj Laniti, p. 1041.