Clinical outcomes and predictors of response to PD-(L)1 blockade in patients with oncogene-driver negative NSCLC who have never smoked.

Abstract

8563 Background: Non-small cell lung cancer (NSCLC) in patients (pts) who have never smoked is associated with poor response to immune checkpoint inhibitors (ICI). However, most studies have focused on pts with actionable oncogenes (e.g., EGFR, ALK, ROS1, RET, MET), and it remains unclear whether specific clinicopathologic and genomic features can predict ICI response in those without these actionable drivers. Methods: Clinicopathologic characteristics and outcomes data were collected from pts with metastatic, oncogene driver-negative, NSCLC who received ICI across 5 academic cancer centers in US and EU. Single sample gene set enrichment analysis was performed on NSCLC samples from the Stand Up To Cancer (SU2C) cohort to characterize transcriptomic correlates of response to ICI monotherapy in responders and non-responders. Results: Of 5639 pts with metastatic NSCLC analyzed, 708 (12.6%) tested negative for actionable oncogene drivers and had never smoked. Among these, median age was 64 years, 59.5% were women, 65.2% had ECOG PS 0/1, and 83.8% had adenocarcinoma. At a median follow-up of 36.9 months (mo), objective response rate (ORR) was 21.8%, median progression-free survival (mPFS) was 4.5 mo, and median overall survival (mOS) was 16.9 mo in this patient population. Pts with PD-L1 TPS ≥1% had significantly higher ORR (31.1% vs. 16.1%, p<0.01), and longer mPFS (HR 0.74, p<0.01) compared to those with PD-L1 <1%. Similarly, pts with very high TMB (≥90 th percentile) had higher ORR (52.2% vs 22.8%, p<0.01), longer mPFS (HR 0.50, p<0.01), and mOS (HR 0.40, p<0.01) compared to those with a TMB <90 th percentile. Pts with positive PD-L1 expression ≥1% and very high TMB had the highest ORR and the longest mPFS and mOS compared to pts with either one of these biomarkers alone. Treatment outcomes also varied by regimen: pts receiving dual PD-(L)1+CTLA-4 blockade or PD-(L)1 blockade + chemotherapy had higher ORR compared to those receiving PD-(L)1 monotherapy (30.0% vs 36.5% vs 10.3%, respectively, p<0.01). Dual PD-(L)1+CTLA4 inhibition was also associated with significantly longer mPFS (9.4 vs 6.9 vs 2.9 mo, p<0.01) and mOS (47.5 vs 19.7 vs 14.7 mo, p<0.01) compared to PD-(L)1 + chemotherapy and PD-(L)1 monotherapy, respectively. These differences were validated in pts receiving these regimens as first-line therapy. In NSCLC samples from pts who had never smoked without oncogenic driver mutations in the SU2C cohort, responders to ICI showed upregulation of innate and adaptive immune responses pathways, including enhanced MHC I/II antigen presentation, as well as increased T-cell activation, proliferation, and chemotaxis. Conclusions: These results emphasize how PD-L1≥1%, very high TMB, and use of dual checkpoint blockade are associated with improved outcomes in pts who have never smoked with oncogene-driver negative NSCLC, aiding personalized ICI use in this neglected population .