High PD-L1 Expression and Clinical Outcomes of First-Line Osimertinib in EGFR-Mutant NSCLC: A Multicenter Real-World Validation Study (WJOG 20724L)

Introduction Osimertinib is the standard first-line therapy for epidermal growth factor receptor (EGFR)-mutated non-small cell lung cancer (NSCLC). However, the clinical significance of high programmed death-ligand 1 (PD-L1) expression in this setting remains unclear. This study aimed to evaluate the association between high PD-L1 tumor proportion score (TPS) (≥50%) and clinical outcomes in patients treated with first-line osimertinib. Methods This multicenter retrospective study used the REAL-WIND database from 16 Japanese institutions. Among 486 patients with EGFR-mutated NSCLC treated with first-line osimertinib, 385 with known PD-L1 status formed the full cohort (Cohort B). After excluding 185 patients overlapping with the published OSI-FACT dataset, 200 formed the non-overlapping validation cohort (Cohort A). Patients were stratified by PD-L1 TPS (≥50% vs. <50%). The primary endpoint was real-world progression-free survival (rwPFS) and secondary endpoints were overall survival (OS) and time to treatment failure (TTF). Outcomes were evaluated using unadjusted analyses and propensity score–based analyses to assess the robustness of the observed associations. Results In Cohort A, high PD-L1 expression was associated with shorter rwPFS (hazard ratio [HR]=1.95; 95% confidence interval [CI]: 1.30–2.92; p = 0.001) and OS (HR=1.72; 95% CI: 1.02–3.09; p = 0.044). Similar associations were observed in Cohort B for rwPFS (HR=1.76; 95% CI: 1.30–2.37; p < 0.001) and OS (HR=1.92; 95% CI: 1.28–2.87; p = 0.002). These associations were consistent in unadjusted analyses and in propensity score–based analyses adjusting for baseline confounding, supporting the reproducibility of the findings in this real-world cohort. Conclusions High PD-L1 expression was associated with reduced clinical benefit from first-line osimertinib in EGFR-mutated NSCLC. These findings should be interpreted as a consistent clinical association observed in real-world practice rather than definitive evidence of a predictive biomarker.