Immune landscape-driven subtyping reveals distinct microenvironment and prognostic profiles in thymic epithelial tumors
摘要
Thymic epithelial tumors (TETs) exhibit marked heterogeneity that challenges clinical management and prognostic evaluation. By consensus clustering of immune-related gene expression from public datasets and an in-house validation cohort, we identified two immune subtypes: a lymphocyte-rich subtype (LRS, n = 86) characterized by strong T-cell activity and favorable prognosis, and a myeloid/stromal-rich subtype (MSRS, n = 33) defined by immunosuppressive features, stemness, oncogenic alterations, and poorer outcomes. An APC gene-based classification panel demonstrated robust performance in distinguishing these subtypes, as confirmed by ROC curve analysis and validated in the Ruijin cohort using FFPE-RNA sequencing and multiplex immunofluorescence. Single-cell transcriptomics further delineated subtype-specific tumor microenvironment features and revealed that MSRS tumors suppressed CD8+ T-cell function through MIF-mediated mechanisms. Collectively, these findings establish a clinically relevant immune classification of TETs that integrates molecular, immunological, and clinical features, providing insights into tumor heterogeneity and supporting the development of personalized therapeutic strategies.