Sex-specific predictive values of biomarkers for immunotherapy efficacy in lung adenocarcinoma

Abstract

It remains a challenge to accurately predict patient responses to tumor immunotherapy, although various biomarkers have been proposed to predict patient responses to anti-PD-1 therapy. Here by integrating genomic, transcriptomic, proteomic and clinical phenotype data from three immunotherapeutic cohorts and a multiple-dimensional dataset of The Cancer Genome Atlas (TCGA) project, we uncovered a profound effect of Sex on the predictive values of conventional biomarkers in lung adenocarcinoma (LUAD): only in females were nonsynonymous mutation burden (TMB), neoantigen burden, smoking signature, KRAS mutations (especially G12C and G12V) or tumor microenvironment robustly correlated with anti-PD-1 efficacy; the correlations in males were either absent or weaker. We propose that Sex be considered in conjunction with conventional biomarkers when predicting immunotherapy efficacy, and conversely, conventional biomarkers be carefully controlled for when attempting to dissect the impact of Sex on immunotherapy efficacy.