Clinicopathologic characteristics and immunotherapy outcomes in SMARCA4-mutant (mut) non-small cell lung cancer (NSCLC).
9577 Background: The catalytic unit of the SWI/SNF chromatin remodeling complex is encoded by the SMARCA4 gene, which is mutated in ~10% of NSCLCs. We sought to characterize the clinicopathologic characteristics and outcomes to immune checkpoint inhibition in SMARCA4-mutant NSCLC. Methods: We collected clinicopathologic and genomic data from patients with NSCLC that had undergone targeted next generation sequencing (NGS) by OncoPanel at the Dana-Farber Cancer Institute. SMARCA4 frameshift, nonsense, and splice-site mutations were considered pathogenic, as were missense mutations if predicted to be pathogenic by Mutation Assessor and Polyphen-2. Clinical outcomes to immune checkpoint inhibition among SMARCA4-mutant NSCLCs were retrospectively assessed. Results: Of 2690 patients with NSCLC, 8% (N = 211) harbored SMARCA4 mutations. Clinicopathological characteristics were balanced between SMARCA4 mut and SMARCA4 wild-type (wt) in terms of age, histology, and PD-L1 expression. We observed a male predominance (P = 0.03), greater use of tobacco (P < 0.001), a higher tumor mutational burden (TMB) (P < 0.001), a higher prevalence of advanced disease (P < 0.001), and a lower prevalence of concurrent targetable driver mutations (P < 0.001) in SMARCA4mut vs SMARCA4wt NSCLCs. Among 513 patients with nonsquamous NSCLC who received immune checkpoint inhibitors, 11% (N = 57) harbored SMARCA4 mutations. From the start of immunotherapy, we observed no difference in overall response rate (ORR 21.5% vs 19.3%; P = 0.3), median progression free survival (mPFS 3.2 months vs 2.1 months; P = 0.4), or median overall survival (mOS 12.0 months vs 8.2 months; P = 0.09) in SMARCA4wt vs SMARCA4mut NSCLC, respectively. However, among KRASmut NSCLC, a concurrent SMARCA4 mut conferred a significantly lower ORR (23.1% vs 0.0%; P = 0.02), a significantly shorter mPFS (4.8 months vs 1.7 month; HR: 0.31 [95% CI: 0.15-0.61]; P < 0.001), and a significantly shorter mOS (15.6 months vs 2.7 months; HR: 0.25 [95%CI: 0.12-0.49]; P < 0.001). The deleterious effect of SMARCA4 mut on immunotherapy outcomes in KRAS mut NSCLC was maintained when controlling for concurrent STK11 mut. Conclusions: SMARCA4 mutations define a genomic subset of NSCLC with unique clinicopathologic characteristics, and confer worse outcomes to immunotherapy in KRAS mut NSCLC.