Prognostic impact of immune-related adverse events in non-completers of durvalumab consolidation therapy for non-small cell lung cancer: a single-center retrospective cohort study.

Abstract

Background: Durvalumab consolidation therapy after chemoradiotherapy (CRT) is standard for unresectable, locally advanced non-small cell lung cancer (NSCLC). However, long-term outcomes, treatment discontinuation impact, and the prognostic significance of immune-related adverse events (irAEs)-controlling for immortal time bias-remain unclear. We evaluated overall survival (OS) and progression-free survival (PFS) by completion status and irAE occurrence in an exploratory single-center cohort.

Methods: We retrospectively analyzed 59 consecutive NSCLC patients receiving durvalumab post-CRT (median follow-up: 53.6 months). Completion was defined as 12 months of therapy (completers, n=25; non-completers, n=34). Survival was estimated using Kaplan-Meier analysis. Time-dependent Cox regression evaluated irAE onset to adjust for immortal time bias, with sensitivity analyses stratified by discontinuation reason. Event-free patients were censored at last follow-up.

Results: Of 59 patients (median age: 70 years), 25 (42%) completed therapy, while 34 (58%) discontinued early (disease progression: n=16, adverse events: n=15, preference: n=3). During follow-up, 28 deaths [completers: 8/25 (32%); non-completers: 20/34 (59%)] and 36 progression events [completers: 9/25 (36%); non-completers: 27/34 (79%)] occurred. Completers achieved superior median OS (70.9 vs. 39.2 months; 5-year OS: 61% vs. 22%; P<0.01) and PFS (not reached vs. 7.9 months; 5-year PFS: 52% vs. 11%; P<0.001). Among non-completers (n=34, 20 deaths), time-dependent Cox analysis showed irAE onset reduced mortality risk [hazard ratio (HR) =0.384; 95% confidence interval (CI): 0.147-0.999; P=0.0498], yielding a higher 5-year OS rate (43% vs. 0%, P=0.03). Stratified sensitivity analysis indicated a consistent directional trend, though limited by event counts (HR =0.44; 95% CI: 0.11-1.72; P=0.24).

Conclusions: One-year durvalumab completion remains the primary clinical objective. In patients discontinuing early, irAE onset was associated with a potential prognostic signal after controlling for time-varying bias; however, these findings represent cautious, hypothesis-generating observations requiring prospective multi-center validation.