A signature of tumor immune microenvironment genes associated with the prognosis of non‑small cell lung cancer.

Establishing a prognostic genetic signature closely related to the tumor immune microenvironment (TIME) to predict clinical outcomes is necessary. Using the Gene Expression Omnibus (GEO) database of a non‑small cell lung cancer (NSCLC) cohort and the immune score derived from the Estimation of Stromal and Immune cells in Malignant Tumours using Expression data (ESTIMATE) algorithm, we applied the least absolute shrinkage and selection operator (LASSO) Cox regression model to screen a 10‑gene signature among the 448 differentially expressed genes and found that the risk prediction models constructed by 10 genes could be more sensitive to prognosis than TNM (Tumor, Lymph node and Metastasis) stage (P=0.006). The CIBERSORT method was applied to quantify the relative levels of different immune cell types. It was found that the ratio of eosinophils, mast cells (MCs) resting and CD4 T cells memory activated in the low‑risk group was higher than that in the high‑risk group, and the difference was statistically significant (P=0.003, P=0.014 and P=0.018, respectively). Inconsistently, the ratio of resting natural killer (NK) cells and activated plasma cells in the low‑risk group was significantly lower than that in the high‑risk group (P=0.05 and P=0.009, respectively). Kaplan‑Meier survival results showed that patients of the high‑risk group had significantly shorter overall survival (OS) than those of the low‑risk group in the training set (P<0.001). Furthermore, Kaplan‑Meier survival showed that patients of the high‑risk group had significantly shorter OS than those of the low‑risk group (P=0.0025 and P=0.0157, respectively) in the validation set [GSE31210 and TCGA (The Cancer Genome Atlas)]. The 10‑gene signature was found to be an independent risk factor for prognosis in univariate and multivariate Cox proportional hazard regression analyses (P<0.001). In addition, it was found that the risk model constructed by the 10‑gene signature was related to the clinical related factors in logistic regression analysis. The genetic signature closely related to the immune microenvironment was found to be able to predict differences in the proportion of immune cells (eosinophils, resting MCs, memory activated CD4 T cells, resting NK cells and plasma cells) in the risk model. Our findings suggest that the genetic signature closely related to TIME could predict the prognosis of NSCLC patients, and provide some reference for immunotherapy.