OBJECTIVES: Immune checkpoint inhibitors (ICIs) have proven to be effective treatment for lung cancer. However, a precise predictive immuno-oncology biomarker is still under development. We investigated the associations among PD-L1 expression, tumor mutational burden (TMB), and oncogenic driver alterations in advanced non-small cell lung cancer (NSCLC) patients treated with ICIs. MATERIALS AND METHODS: This multicenter cohort study included 1017 lung cancer patients. PD-L1 expression using four IHC assays (22C3, 28-8, SP263, SP142), TMB by whole-exome sequencing and oncogenic driver alterations were analyzed comprehensively. Clinical characteristics, treatment and survival data were collected. RESULTS: The results of 22C3 and 28-8 for PD-L1 expression showed acceptable concordance (k = 0.89; 95% confidence interval [CI], 0.87-0.92), and the clinical outcomes of ICIs classified according to PD-L1 expression by both assays were also approximately the same. There was slight concordance (k = 0.16; 95% CI, 0.11-0.22) between 22C3 and SP142, and high PD-L1 expression by SP142 was correspond to very high PD-L1 expressions by other assays. Patients with both high PD-L1 expression and high TMB showed a good response to ICIs with the response rate of 64% and median progression-free survival of 9.0 months despite of small population. Common EGFR or STK11 mutations showed a lower rate of high PD-L1 expression and a worse efficacy of ICIs and KRAS mutations had no negative impact on response to ICIs. CONCLUSION: Comprehensive assessment of PD-L1 expression, TMB, and oncogenic driver alterations would help to better predict the clinical outcomes of ICIs in NSCLC patients.