Introduction: Recently, immune checkpoint inhibitors (ICIs) has been reported to achieved convincing clinical benefits and significantly prolonged the overall survival (OS) of advanced non-small cell lung cancer (NSCLC) patients. Sensitivity to immunotherapy was related to several biomarkers, such as PD-L1 expression, TMB level, MSI-H and MMR. However, novel biomarkers for the prognosis to ICIs treatment need to be further investigated, and it is an urgent demand to establish a systematic hazard model to assess the efficacy of ICIs therapy for advanced NSCLC patients.Methods: In this study, gene mutation and clinical data of NSCLC patients was obtained from the TCGA database. Then, we analyzed the detailed clinical information and mutational data of two advanced NSCLC cohorts received ICIs treatment from the cBioPortal for Cancer Genomics. The Kaplan-Meier plot method was used to perform survival analyses, selected variables were used to develop a systematic nomogram. The prognostic significance of ERBB4 in pan-cancer was analyzed by another cohort from cBioPortal for Cancer Genomics.Results: Mutation frequencies of TP53 and ERBB4 was 54% and 8% in NSCLC, respectively. Mutual exclusive analysis in cBioPortal indicated that ERBB4 does show co-occurencing mutations with TP53. Patients harbored ERBB4 mutations were confirmed to have a better prognosis for ICIs treatment, compared to ERBB4 wild type (PFS: p=0.0360; OS: p=0.0378) and only TP53 mutations (OS: p=0.021). The mutation status of ERBB4 and TP53 are tightly linked to DCB for ICIs treatment, PD-L1 expression, TMB value and TIICs. Finally, a novel nomogram was built to evaluate the efficacy of ICIs therapy.Conclusion: ERBB4 mutations could serve as a predicting biomarker for prognosis of ICIs treatment. The systematic nomogram was proven to have a great potential to evaluate the efficacy of ICIs therapy for advanced NSCLC patients.