Hepatocellular carcinoma with ARID1A mutation is associated with higher TMB and poor survival.

Abstract

e16667 Background: Hepatocellular carcinoma (HCC) is a heterogeneous disease that lacks molecular predictors of response to available treatments. AT-rich interactive domain 1A (ARID1A), a SWI/SNF chromatin remodeling gene, is commonly mutated in human cancers and hypothesized to be a tumor suppressor gene. In recent years, immune checkpoint blockade immunotherapies (ICIs) are promising therapies for multiple cancers including HCC, and tumor mutation burden (TMB) was an important biomarker to predict the efficacy of ICIs. Besides, previous research reported that ARID1A deficiency was associated with mismatch repair (MMR) in cancer, and may cooperated with ICIs. Methods: Whole exome sequencing data of 10336 pan-cancer patients including 353 HCC patients was obtained from the Cancer Genome Altas (TCGA). Next generation sequencing (NGS) data of 15849 pan-cancer patients including 1926 HCC patients from Chinese clinical dataset were analyzed to explore the correlation between ARID1A gene mutation and TMB. TMB was defined as total number of somatic non-synonymous mutations in coding region. Prognostic data of 136 HCC patients were obtained from the MSK-IMPACT Clinical Sequencing Cohort (MSKCC). Results: In total, 8.62% (891/10336) of pan-cancer patients in TCGA harboring ARID1A mutation and 8.47% (3188/37628) in Chinese cohort, while mutant percentage of HCC was 9.35% (33/353) in TCGA, 9.03% (174/1926) in Chinese cohort, the alternation frequency of ARID1A in two cohorts was no significant difference (P = 0.3334). The medium TMB level of ARID1A mutant group was higher than wild-type group with significant difference both in Chinese pan-cancer and HCC (medium TMB level, P < 0.0001). Results of 18 kinds of ARID1A mutated tumors patients in TCGA cohort indicated that ARID1A mutation was an independent risk factor in liver cancer affecting overall survival (OS, HR 2.43, 95% IC 1.07-5.54, P = 0.0286). Besides, Kaplan-Meier analysis was performed on HCC patients in MSKCC cohort, ARID1A statue resulted in significantly shorter OS (median, 12.2 vs 21.43 months; HR 2.5; P = 0.0092). Conclusions: The results indicated that ARID1A gene mutation was a poor prognostic factor in hepatocellular carcinoma, and these mutational states were associated with higher TMB level, which might be a potential positive biomarker of immunotherapy.