부산대학교 도서관

Abstract Background Although the relationship between inflammatory response and tumor has been gradually recognized, the potential implications of of inflammatory response genes in lung adenocarcinoma (LUAD) remains poorly investigated. Methods RNA sequencing and clinical data were obtained from multiple independent datasets (GSE29013, GSE30219, GSE31210, GSE37745, GSE42127, GSE50081, GSE68465, GSE72094, TCGA and GTEx). Unsupervised clustering analysis was used to identify different tumor subtypes, and LASSO and Cox regression analysis were applied to construct a novel scoring tool. We employed multiple algorithms (ssGSEA, CIBERSORT, MCP counter, and ESTIMATE) to better characterize the LUAD tumor microenvironment (TME) and immune landscapes. GSVA and Metascape analysis were performed to investigate the biological processes and pathway activity. Furthermore, ‘pRRophetic’ R package was used to evaluate the half inhibitory concentration (IC50) of each sample to infer drug sensitivity. Results We identified three distinct tumor subtypes, which were related to different clinical outcomes, biological pathways, and immune characteristics. A scoring tool called inflammatory response gene score (IRGS) was established and well validated in multiple independent cohorts, which could well divide patients into two subgroups with significantly different prognosis. High IRGS patients, characterized by increased genomic variants and mutation burden, presented a worse prognosis, and might show a more favorable response to immunotherapy and chemotherapy. Additionally, based on the cross-talk between TNM stage, IRGS and patients clinical outcomes, we redefined the LUAD stage, which was called ‘IRGS-Stage’. The novel staging system could distinguish patients with different prognosis, with better predictive ability than the conventional TNM staging. Conclusions Inflammatory response genes present important potential value in the prognosis, immunity and drug sensitivity of LUAD. The proposed IRGS and IRGS-Stage may be promising biomarkers for estimating clinical outcomes in LUAD patients.