An ENE-related transcriptomic signature predicts distant metastasis risk and immune microenvironment features in nasopharyngeal carcinoma

Extracapsular nodal extension (ENE) is an adverse prognostic factor in nasopharyngeal carcinoma (NPC). In this study, we analyzed RNA-seq data from 159 NPC patients stratified by ENE status, identified differentially expressed genes, and developed an eight-gene machine-learning signature for distant metastasis prediction. The signature showed favorable predictive performance in the in-house cohort (C-index, 0.819; AUC, 0.813) and in an external validation cohort (AUC, 0.762). Risk stratification based on this signature was associated with distinct immune microenvironment features. Low-risk tumors exhibited higher immune scores, increased inferred CD8⁺ T-cell and B-cell infiltration, and elevated expression of immune checkpoint genes, including CTLA4 and PDCD1, suggesting a more immune-active phenotype. Exploratory analysis in the IMvigor210 cohort showed consistent associations between the signature and immune checkpoint blockade response; however, validation in NPC immunotherapy cohorts is still required. Among the signature genes, C19orf18 was identified as a key candidate gene. Functional assays showed that C19orf18 knockdown reduced the migratory and invasive capacities of NPC cells, while immunohistochemistry revealed higher C19orf18 expression in tumor and metastatic tissues. Overall, this ENE-related gene signature may help stratify distant metastasis risk and characterize immune microenvironment heterogeneity in NPC, while C19orf18 warrants further investigation as a potential metastasis-associated biomarker.