Identification of metabolic genes for the prediction of prognosis and tumor microenvironment infiltration in early-stage non-small cell lung cancer

Abstract

Early-stage non-small cell lung cancer (NSCLC) patients are at substantial risk of poor prognosis. We attempted to develop a reliable metabolic gene-set-based signature that can predict prognosis accurately for early-stage patients. Least absolute shrinkage and selection operator method Cox regression models were performed to filter the most useful prognostic genes, and a metabolic gene-set-based signature was constructed. Forty-two metabolism-related genes were finally identified, and with specific risk score formula, patients were classified into high-risk and low-risk groups. Overall survival was significantly different between the two groups in discovery (HR: 5.050, 95% CI: 3.368–7.574, P < 0.001), internal validation series (HR: 6.044, 95% CI: 3.918–9.322, P < 0.001), GSE30219 (HR: 2.059, 95% CI: 1.510–2.808, P < 0.001), and GSE68456 (HR: 2.448, 95% CI: 1.723–3.477, P < 0.001). Survival receiver operating characteristic curve at the 5 years suggested that the metabolic signature (area under the curve [AUC] = 0.805) had better prognostic accuracy than any other clinicopathological factors. Further analysis revealed the distinct differences in immune cell infiltration and tumor purity reflected by an immune and stromal score between high- and low-risk patients. In conclusion, the novel metabolic signature developed in our study shows robust prognostic accuracy in predicting prognosis for early-stage NSCLC patients and may function as a reliable marker for guiding more effective immunotherapy strategies.

Keywords

• lung cancer
• overall survival
• metabolism
• immune infiltration
• biomarker