Non-small cell lung cancer (NSCLC) remains a leading cause of cancer-related mortality, with chemoimmunotherapy (CIT) as the first-line standard for advanced NSCLC without driver mutations. However, predictive biomarkers for CIT response are limited. Tertiary lymphoid structures (TLS) play a critical role in antitumor immunity and may serve as potential biomarkers. This study aimed to screen and validate a TLS-derived gene signature to predict responses to first-line CIT in advanced NSCLC. Data from three randomized trials (ORIENT-11, OAK, POPLAR) and The Cancer Genome Atlas-NSCLC were analyzed. TLS scores were computed via ssGSEA based on 17 TLS-related gene signatures. Patients were stratified into TLS-high and TLS-low groups. The predictive value was assessed by survival analysis, nomograms, and receiver operating characteristic curve. Correlations with programmed cell death 1 ligand (PD-L1) and ImmuneScore were evaluated. TLS signature 3 was identified as a predictive biomarker. In ORIENT-11, high signature 3 scores correlated with longer progression-free survival (PFS) (9.92 vs 6.77 months, p = 0.001) and overall survival (OS) (not reached vs 17.60 months, p < 0.001). Multivariate analysis confirmed signature 3 as an independent predictor for both outcomes (PFS HR = 2.14, p = 0.006; OS HR = 2.24, p = 0.002). A nomogram integrating signature 3 and clinicopathologic factors showed strong discriminative power. Signature 3 also correlated with PD-L1 expression and 'hot' immune phenotypes, enhancing prediction in PD-L1-negative subsets. TLS signature 3 predicts CIT response independently of PD-L1, improving outcomes in PD-L1-negative patients and complementing PD-L1 testing. Integration into clinical practice may refine treatment decisions, warranting further mechanistic and clinical validation.
Du et al. (Sat,) studied this question.