Objectives: Immune checkpoint inhibitors (ICIs) improved outcomes in non-small cell lung cancer (NSCLC) patients. We report the predictive utility of human leukocyte antigen class I (HLA-I)... Show moreObjectives: Immune checkpoint inhibitors (ICIs) improved outcomes in non-small cell lung cancer (NSCLC) patients. We report the predictive utility of human leukocyte antigen class I (HLA-I) diversity and tumor mutational burden (TMB) by comprehensive next-generation sequencing. Methods: 126 patients were included. TMB high was defined as >= 10 nonsynonymous mutations/Mb. Patients exhibit high HLA-I diversity if at least one locus was in the upper 15th percentile for DNA alignment scores. Results: No difference in response rate (RR; 44.4% versus 30.9%; p = 0.1741) or 6-month survival rate (SR; 75.6% versus 77.8%; p = 0.7765) was noted between HLA-I high diversity and low diversity patients. HLA-I high diversity patients did significantly more often exhibit durable clinical benefit (DCB), defined as response or stable disease lasting minimally 6 months (64.4% [29/45] versus 43.2% [35/81]; p = 0.0223).TMB high patients exhibited higher RR (49.1% versus 25.4%; p = 0.0084) and SR 6 months after start ICI (85.5% versus 70.4%; p = 0.0468) than TMB low patients. The proportion of patients with DCB, did not differ significantly between TMB high and low subgroups (60.0% [33/55] versus 42.3% [30/71]; p = 0.0755).Patients with combined dual high TMB and HLA-I diversity had higher RR (63.2% versus 22.2%; p = 0.0033), but SR at 6 months did not differ significantly (84.2% versus 64,4%; p = 0.1536). A significantly higher rate of patients experienced DCB in dual high compared to the dual low group (73.7% [14/19] versus 35.6% [16/45]; p = 0.0052). Triple positive patients (high TMB and HLA-I diversity and PD-L1 positive) had higher RR (63.6% versus 0.0%; p = 0.0047) and SR at 6 months (100% versus 66.7%; p = 0.0378) compared to triple-negative patients. Conclusion: HLA-I diversity was able to predict durable clinical benefit in ICI treated NSCLC patients, but failed to confirm as a predictor of response or survival. TMB confirmed as a predictive biomarker. Show less
