Clinical outcomes and predictors of response to PD-(L)1 blockade in patients with NSCLC without actionable genomic alterations who never used tobacco.

Abstract

Predictive biomarkers of response to immune checkpoint inhibitors (ICI) remain poorly defined in patients with non-small cell lung cancer (NSCLC) without a history of tobacco use and lacking actionable genomic alterations (AGA). We aimed to identify clinical and molecular predictors of response to ICI-based regimens in patients who have never smoked and lack AGA. We retrospectively analyzed patients with metastatic, AGA-negative NSCLC who never smoked, treated with ICI-based regimens across multiple independent cohorts. Tumor-infiltrating lymphocyte (TIL) densities were quantified using a machine learning-based algorithm, and immune cells biomarkers were assessed with multiplexed immunofluorescence. Transcriptomic correlates of ICI response were analyzed in the SU2C cohort. Among 741 patients with AGA-negative NSCLC and no history of tobacco use, objective response rate (ORR) was 23.2%, median progression-free survival (mPFS) 4.5 months, and median overall survival (mOS) 16.8 months. PD-L1≥90% and TMB≥90th percentile were independently and significantly associated with improved ORR, mPFS, and mOS (all p<0.01). PD-(L)1+CTLA-4 combinations outperformed chemo-immunotherapy and PD-(L)1 monotherapy in terms of mPFS and mOS. Transcriptomic analysis revealed enrichment of innate and adaptive immune pathways in responders, including increased MHC class I/II antigen presentation and T-cell activity. High TIL density was also associated with superior ORR and PFS. Multiplexed immunophenotyping confirmed higher immune cell infiltration in patients who experienced durable clinical benefit. We demonstrated how combination therapies may improve ICI outcomes in patients with AGA-negative NSCLC and no history of tobacco exposure. Very high PD-L1, TMB, and immune-enriched phenotypes may guide treatment personalization.