Construction and validation of a novel prognostic model for lung squamous cell cancer based on N6-methyladenosine-related genes

Abstract

Abstract Background N6-methyladenosine (m6A) is the most prevalent modification in mRNA in biological processes and associated with various malignant tumor initiation and progression. The present study aimed to construct a prognostic risk model based on m6A-related genes (the downstream genes influenced by m6A modulators) for LUSC. Methods Based on TCGA, we stratified LUSC patients with and without genetic alteration of m6A modulators into altered and unaltered groups. Using univariate Cox and Lasso regression analyses, we identified prognostic m6A-related genes to construct a prognostic risk model. We then applied a multivariate Cox proportional regression model and the survival analysis to evaluate the risk model. Moreover, we performed the Receiver operating characteristic curve to assess the efficiency of the prognostic model based on TCGA and GSE43131. We analyzed the characteristics of tumor-associated immune cell infiltration in LUSC through the CIBERSORT method. Results Three m6A-related genes (FAM71F1, MT1E, and MYEOV) were identified as prognostic genes for LUSC. A novel prognostic risk model based on the three m6A-related genes was constructed. The multivariate Cox analysis showed that the prognostic risk model was an independent risk factor (HR = 2.44, 95% CI = 1.21~3.56, p = 0.029). Patients with a high-risk group had worse overall survival both in TCGA (p = 0.018) and GSE43131 (p = 0.00017). The 1, 2, and 3-year AUC value in TCGA was 0.662, 0.662, and 0.655, respectively; The 1, 2, and 3-year AUC value in GSE43131 was 0.724, 0.724, and 0.722, respectively. The proportion of infiltrated neutrophils in the high-risk group was higher than that in the low-risk group (p = 0.028), whereas that of resting NK cells (p = 0.002) was lower. Conclusion A novel prognostic risk model based on three m6A-related genes for LUSC was generated in this study.