Background: The immune microenvironment influences tumor progression and prognosis in lung adenocarcinoma (LUAD). Although immune gene signatures are linked to immunotherapy response, their prognostic value in patients receiving conventional treatment remains unclear. Aim: To evaluate the prognostic significance of immune gene expression in LUAD treated with conventional therapy and to define immune-based molecular subgroups for risk stratification. Methods: An immune gene panel derived from a published non-small cell lung cancer dataset (GSE93157) was analyzed in The Cancer Genome Atlas LUAD cohort ( n = 517). After filtering low-expression genes, 117 immune-related genes were retained. Unsupervised hierarchical clustering identified immune expression subgroups, and T-cell, B-cell, and NK-cell signatures were assessed for associations with clinicopathologic variables and overall survival (OS). Subgroup differences were evaluated using chi-squared tests, and survival was analyzed using Kaplan–Meier and Cox proportional hazards models. Results: Four immune expression subgroups were identified. Immune subgrouping was significantly associated with pathologic tumor-node-metastasis stage ( P = 0.0023), driver mutation patterns ( P = 0.0053), smoking status ( P = 0.0394), and treatment response ( P = 0.0005). Immune-high clusters showed heterogeneous driver mutation profiles, including a predominance of EGFR/KRAS/ALK wild-type tumors in subgroup 1, whereas immune-low subgroup 4 displayed a distinct distribution of driver alterations. High immune gene panel expression predicted improved OS (median, 58.8 vs. 39.3 months; hazard ratio HR = 0.57, 95% confidence interval CI = 0.42–0.78; P = 0.0010) and remained independently prognostic after adjustment (HR = 0.66, 95% CI = 0.47–0.91; P = 0.0124). T-cell, B-cell, and NK-cell signatures similarly predicted favorable survival. Conclusion: High immune gene expression independently predicts survival in LUAD treated with conventional therapy, reflecting an active immune microenvironment and supporting immune-based risk stratification.
Liu et al. (2026) studied this question.