Background <p>Despite the effectiveness of immune checkpoint blockade (ICB) therapy in cancer treatment, oncogene-addicted subsets of non-small cell lung cancer (NSCLC) typically derive limited benefits. Herein, we established and evaluated a mutation-based signature to predict ICB efficacy in patients with both non-mutant and oncogene-mutant NSCLC.</p> Methods <p>The predictive power of the signature was evaluated and validated across advanced NSCLCs cases with or without targetable oncogene alterations, using four public and one in-house immunotherapy cohort. NSCLC cohorts from The Cancer Genome Atlas were used to explore the underlying mechanisms of the signature. Herein, we defined a gene mutation number (GMN) based on the signature and classified patients into high and low GNM groups (GMN-H and GMN-L, respectively).</p> Results <p>The GMN-H group exhibited significantly improved overall survival (OS) (hazard ratio [HR], 0.23; 95% confidence interval [CI]: 0.14–0.37; <i>P</i> &lt; 0.001 in the discovery cohort) and progression-free survival (HR, 0.43; 95% CI: 0.28–0.65; <i>P</i> &lt; 0.001 in the discovery cohort; HR, 0.41; 95% CI: 0.26–0.66; <i>P</i> = 0.0011 in the validation cohort) after ICB therapy. The survival advantage of GMN-H persisted in patients with KRAS mutations, whereas only trends toward improved survival outcomes were observed in those with EGFR, HER2, or MET alterations. Moreover, in the KRAS^G12C- and KRAS^G12V-mutant subgroups, patients with GMN-H exhibited a better OS. Compared with GMN-L, GMN-H was associated with higher levels of neoantigens, enrichment of WNT and MAPK pathways, and increased intra-tumoral CD8 + T cell infiltration.</p> Conclusions <p>The mutation-based signature established herein represents a robust biomarker for identifying patients with advanced NSCLCs who are most likely to benefit from ICB therapy, including those with targetable oncogene alterations.</p>