Myeloid cell differentiation-related gene signature for predicting clinical outcome, immune microenvironment, and treatment response in lung adenocarcinoma

Abstract

Considering the key role of myeloid cell differentiation—related genes in the tumor microenvironment (TME), we aimed to build a prognostic risk model using these genes for Lung adenocarcinoma (LUAD). The mRNA gene expression profiles of LUAD patients from The Cancer Genome Atlas (TCGA) and Gene Expression Omnibus (GEO) databases were downloaded as the training and validation sets. Then, “edgeR” R package was applied to screen out the differentially expressed genes (DEGs) and univariate cox regression, backward stepwise selection analyses were performed to construct a prognostic model for LUAD. ESTIMATE, TIMER, XCELL, CIBERSORT abs, QUANTISEQ, MCPCOUNTER, EPIC, and CIBERSORT algorithms were conducted to access the association of risk levels with the stromal and immune cell infiltration levels in LUAD. Six genes (F2RL1, PRKDC, TNFSF11, INHA, PLA2G3 and TUBB1) were utilized to construct the prognostic model. The risk model showed excellent prognostic performance for LUAD in both TCGA and GEO datasets. Also, compared to the low-risk patients, the high-risk patients had higher expression of immune checkpoint molecules and showed a lower IC50 value to the chemotherapy agents. Our findings provided a myeloid cell differentiation—related gene signature that could effectively predict prognosis and guide treatment strategies for LUAD patients.