Immune checkpoint inhibition alters patterns of failure in inoperable stage III non-small cell lung cancer patients treated with chemoradiotherapy

We compared failure patterns in patients with inoperable stage III non-small cell lung cancer (NSCLC) treated with chemoradiotherapy (CRT) alone versus CRT combined with sequential and/or concurrent immune checkpoint inhibitors (CRT-IO).

Retrospective real-world data from 221 patients across two German tertiary cancer centers were analyzed. Of these, 74 received CRT-IO, including sequential durvalumab (85%) and concurrent/sequential nivolumab (15%), while 148 received CRT alone. First failure site and time to failure were compared.

Between 2012 and 2022, all patients received thoracic radiotherapy (≥ 60 Gy) and at least two cycles of platinum-based chemotherapy. Induction chemotherapy was administered in 36%, and induction chemo-immunotherapy in 2%. Median follow-up was 51.7 months (95% CI 47.0–56.4). Median overall survival (OS) for the entire cohort was 37.1 months (95% CI 26.0–48.2), with OS in the CRT-IO group not reached vs. 27.1 months (95% CI 18.5–25.7) in the CRT group (p < 0.001). Median progression-free survival (PFS) was 22.8 months (95% CI 6.4–39.1) for CRT-IO versus. 9.9 months (95% CI 7.0–12.8) for CRT (p = 0.001, see Fig. 1).

Failure patterns differed significantly. CRT-IO patients had lower loco-regional progression (LRP) rates (9.5% vs. 21.8%, p = 0.023) and were more frequently alive without progression (45.9% vs. 16.3%, p < 0.001). Brain metastasis (BM) as the first failure, multifocal progression (MFP) and isolated extracranial distant metastasis (ecDM) rates were comparable between the CRT and CRT-IO subgroup. Women had a higher risk of isolated BM (17.3% vs. 6.8%, p = 0.016), whereas squamous cell carcinoma (SCC) patients had higher LRP rates (25.3% vs. 13.0%, p = 0.016). Median post-progression survival (PPS) was 19.4 months (95% CI 16.8–22.0) for CRT-IO and 9.5 months (95% CI 5.8–13.1) for CRT (p = 0.207). PPS was longer after BM (19.9 months) vs. LRP (8.5 months, p = 0.076) and significantly better in women (20.7 vs. 8.9 months, p = 0.012) and adenocarcinoma/non-otherwise-specified-carcinoma (AC/NOS) vs. SCC (p < 0.001).

CRT-IO significantly improves OS, PFS, and LRP control compared to CRT alone. Failure patterns and survival disparities by histology and gender suggest tailored surveillance and treatment strategies are needed. Further studies should optimize management of LRP and long-term outcomes in CRT-IO-treated patients.