Exploration of Novel Immunological Terms in Lung Cancer With Large Populations: Implications for Immunotherapy.

Background: Ideal biomarkers to predict the response to immunotherapy in lung cancer are still lacking. Therefore, there is a need to explore effective biomarkers in large populations. Objective: The objective of this study is to explore novel immunological classifications that are associated with immunotherapy response through the ssGSEA algorithm. Methods: Six independent lung cancer cohorts were collected for analysis including The Cancer Genome Atlas (TCGA), Gene Expression Omnibus (GEO), and the EMBL-EBI database. The ssGSEA algorithm was performed to extract immune terms. Then, TCGA samples were involved as a training group and other cohorts were used as a validation group. After LASSO and Cox regression, prognostic associated immune terms were extracted and an immune-related risk score (IRS) signature was constructed. Furthermore, the association between IRS signature and clinical data, genome features, stemness indices analysis, tumor immune microenvironment, immunotherapy efficiency, and targeted therapy response was also investigated. Results: A total of 1,997 samples were enrolled in this study including six large lung cancer cohorts. Fifty-four immune terms were calculated through the ssGSEA algorithm in combined cohorts. Then, a nine-immune-term risk score model named IRS signature was established to predict the prognosis in combined cohorts. We classified patients into high-risk and low-risk subgroups according to the cutoff point. Subsequently, analysis of clinical data and genome features indicated that the patients in the high-IRS group tend to have advanced clinical features (clinical stage and T classification), as well as a higher level of copy number variation burden, higher tumor burden mutation, and higher tumor stemness indices. Immune landscape analysis demonstrated that high-IRS groups exhibited lower immune cell infiltration and immune-suppressive state. More importantly, the predicted result of the Tumor Immune Dysfunction and Exclusion analysis showed that high-IRS groups might be more insensitive to immunotherapy. Meanwhile, we have also identified that high-IRS groups were associated with better efficiency of several targeted drugs. Conclusion: To summarize, we identified a novel IRS model based on nine immune terms, which was quantified by the ssGSEA algorithm. This model had good efficacy in predicting overall survival and immunotherapy response in non-small cell lung cancer patients, which might be an underlying biomarker.