Eight gene mutation-based polygenic hazard score as a potential predictor for immune checkpoint inhibitor therapy outcome in metastatic melanoma.

Abstract

Background: Immune checkpoint inhibitor (ICI) therapies have revolutionized the treatment of metastatic cutaneous melanoma, but have only benefitted a subset of them. Gene mutations were reported to impact the ICI therapy outcomes in metastatic melanoma but have not been fully investigated. Hence, we systematically analyzed the impact of cancer-related gene mutations on the clinical outcome in metastatic melanoma patients who underwent ICI therapies. Methods: Publicly available discovery and validation cohorts (312 patients and 110 patients respectively, all the patients received ICI therapies) were included in this study. Cox proportional hazards regression analysis was used to assess the association of 468 cancer-related gene mutations with overall survival (OS) in the discovery cohort, and the polygenic hazard score (PHS) was constructed subsequently, and validated in the validation cohort. The Tumor Immune Estimation Resource (TIMER) online tools, which are based on The Cancer Genome Atlas database, were used to analyze the impact of gene mutations on tumor-infiltrated immune cells in melanoma samples. Results: We found eight gene mutations that were significantly associated with the overall survival (BAP1, CARD11, IGF1R, KMT2D, PTPRD, PTPRT, ROS1, and TERT, p < 0.05, mutation frequency >0.05). The PHS, which was based on these genes, was found to effectively discriminate the subset which benefited most from ICI therapies (HR = 1·54, 95%CI, 1.25–1.95; p < 0.001). After adjusting with age, sex, ICI regimes, and tumor mutation burden (TMB), we found that PHS was an independent predictor for the outcome of ICI therapies (adjusted HR = 1.84, 95%CI, 1.22–2.79; p = 0.004). The PHS was validated in the validation cohort (log-Rank p = 0.038). Further research found that CARD11 and PTPRD mutations were significantly associated with more tumor-infiltrated immune cells in melanoma samples. Conclusion: For the first time, we have shown that PHS can independently and effectively predict the ICI therapy outcome in metastatic melanoma, which once validated by larger research, may help the decision-making process in melanoma.