Distribution and concordance of PD-L1 expression by routine 22C3 assays in East-Asian patients with non-small cell lung cancer

Background Currently, programmed death ligand-1 (PD-L1) expression has been widely applied in clinical trials and real-world clinical practice as a major biomarker for the efficacy of immune-checkpoint inhibitors. The purpose of this study is to reveal the distribution and concordance of PD-L1 expression in a large-scale consecutive cohort from East-Asian patients with non-small cell lung cancer (NSCLC). Methods PD-L1 testing was conducted using 22C3 assays, and cases were categorized into the high, low, and no expression of PD-L1 based on the tumor proportion score (TPS). Target-capture next-generation sequencing was used to identify molecular events. Results A total of 4550 patients and 4622 tests of PD-L1 expression were enrolled. There were 3017 (66.3%) patients with no PD-L1 expression (TPS < 1%), 1013 (22.3%) with low PD-L1 expression (TPS 1-49%), 520 (11.4%) with high PD-L1 expression (TPS >= 50%). Higher proportions of positive PD-L1 expression (TPS >= 1%) were observed in smokers, males, squamous cell carcinoma, and high-grade lung adenocarcinoma. Further analyses revealed fair agreement in primary and metastatic lesions (kappa = 0.533), poor agreement in multi-focal primary tumors (kappa = 0.045), and good agreement in biopsy and resection samples (kappa = 0.662) / two biopsy samples (kappa = 0.711). Mutational analyses revealed association between high PD-L1 expression (TPS >= 50%) and EGFR wild-type, KRAS mutation, ALK rearrangement, and TP53 mutation. Conclusion The study reveals the unique distribution pattern of PD-L1 expression in a large-scale East-Asian cohort with NSCLC, the concordance of multiple PD-L1 tests, and the association between PD-L1 expression and molecular events. The results shed a light on the optimization of PD-L1 testing in clinical practice.