Immune-related LncRNA signatures define tumor microenvironment subtypes and predict immunotherapy response in NSCLC

Long non-coding RNAs (lncRNAs) play critical roles in immune regulation and tumor microenvironment (TME) remodeling. However, their contribution to non-small cell lung cancer (NSCLC) heterogeneity and immunotherapy response remains unclear. We integrated transcriptomic data from TCGA and GEO cohorts after batch correction. Immune-related lncRNAs were identified and used for unsupervised clustering to define molecular subtypes. Survival outcomes, immune infiltration, somatic mutation profiles, and predicted drug sensitivities were compared among subtypes. Weighted gene co-expression network analysis (WGCNA) and pathway enrichment were performed to identify hub genes and biological processes. Three lncRNA-defined subtypes were identified with distinct TME characteristics: an immune-inflamed subtype enriched in B/T cells and HLA expression, an immune-escape subtype with interferon-driven MHC upregulation, and an immune-desert subtype with minimal immune infiltration. These subtypes were significantly associated with prognosis, genomic alterations, and clinical features. Cluster A (predominantly LUAD) exhibited superior overall survival and higher predicted immunotherapy sensitivity, while Cluster C (enriched in LUSC) showed higher predicted sensitivity to chemotherapy. Hub genes including SOX2, KRAS, KEAP1, and STAT1 were implicated in TME regulation. Drug sensitivity prediction suggested potential therapeutic stratification across clusters. Immune-related lncRNA signatures define novel NSCLC subtypes with distinct immune phenotypes and therapeutic responses. These findings suggest that lncRNA-based TME classification may complement PD-L1 and tumor mutational burden as predictive biomarkers for immunotherapy. Validation in prospective clinical cohorts is warranted to establish their translational utility.