Abstract

e14636 Background: Alterations of SMARCA4 or SMARCB1 (SMARCA4/B1) proteins have been identified in various aggressive cancers. Recent studies show that tumors harboring deleterious SMARCA4/B1 alterations express PD-L1 and respond to immune checkpoint inhibitor (ICI) therapies. Current ongoing immunotherapy-based clinical trials for cancers deficient in SMARCA4/B1 include combinations of atezolizumab and the TIGIT antibody tiragolumab (Phase II), as well as nivolumab and the CTLA-4 antibody ipilimumab (Phases II & III). Here we present a comprehensive analysis of the pan-cancer landscape of SMARCA4/B1 alterations, based on real-world data from a large cohort of Chinese patients. Methods: From 2021 to 2023, we collected genomic information from 16,113 cancer patients, spanning more than 40 different cancer types. These patients had undergone targeted next-generation sequencing (NGS) through the Med1CDx panel, which targets 601 cancer-associated genes. After excluding 1,034 samples of unidentified tumor origins, 15,079 samples were analyzed further. The analysis included somatic and germline pathogenic/likely pathogenic (P/LP) single-nucleotide variants, insertions/deletions, and copy number alteration in SMARCA4/B1. Results: We identified 421/15079 (2.8%) patients with SMARCA4/B1 alterations, with a median age of 63 years (range 13-90). Cancer types with a higher frequency of SMARCA4/CB1 mutations include: endometrial cancer (9.8%), thyroid cancer (6.9%), gastric cancer (6.2%), cholangiocarcinoma (5.7%), and lung cancer (4.4%). The predominant mutation type is variants of uncertain significance (VUS) gene mutations, constituting 72% (302/421) of the spectrum, with nearly all being missense mutations. These are mainly observed in lung cancer, thyroid cancer, and colorectal cancer. The frequency of pathogenic/likely pathogenic mutations is 28% (119/421), mainly distributed in lung cancer and liver cancer. Q201L, P222L, and E1364del are the most common mutations in SMARCA4, with a mutation rate of approximately 4.1%, occurring across over five cancer types. The most common co-mutations of SMARCA4/B1 were TP53 (47.3%), EGFR (25.4%), KRAS (15.9%), STK11 (7.1%) and CDKN2A (7.4%). Clinicopathological feature of age showed no significant differences between SMARCA4/B1 mutant and wild-type groups. Tissues with SMARCA4/B1 mutations exhibited a higher tumor mutational burden (TMB) (median=8.3) than wild-type tissues (median=4.2) (P < 0.001). Conclusions: Elevated TMB levels in patients with SMARCA4/B1 mutations suggest a significant potential benefit from immunotherapy. The presence of co-mutations may also aid in predicting the efficacy of combination immunotherapy treatments. Furthermore, the high prevalence of VUS emphasizes the need for additional research to determine their clinical relevance.