• artificial intelligence
• gene signature
• immunotherapy
• non small cell lung cancer
• predictive biomarkers

We propose a retrospective case-control study on genetic and clinical data of NSCLC patients belonging to the TCGA project, freely available on Cbioportal. Patients in the case group were treated with immunotherapy (first, second, or subsequent line of therapy), while those in the control one undertook chemotherapy. For each patient, tumor’s genomic profile was sequenced before therapy. We developed an innovative approach to find significant mutated genes to the respective network of somatic co-mutation. Therefore, we wrote an Artificial Intelligence methodology in the Python programming language for analysis interpretation and gene clustering. We compared the results between cases and controls. Results showed a predictive role of the gene signature in the immunotherapy group and a strong association between TMB score and gene signature.