Impact of treatment sequence of immunotherapy and stereotactic radiotherapy on survival in non–small cell lung cancer patients with brain and bone metastases: An NCDB data analysis.

e20119 Background: Optimizing the sequence of immunotherapy (IT) and stereotactic radiotherapy (SRT) in metastatic non-small cell lung cancer (NSCLC) may enhance survival, yet site-specific efficacy remains undefined. We investigated the impact of IT followed by SRT (IT→SRT) versus SRT followed by IT (SRT→IT) sequencing on overall survival (OS) in NSCLC patients with brain metastases (BrMs) or bone metastases using National Cancer Database (NCDB) data. Methods: This retrospective cohort study included 3,158 adults with Stage IV NSCLC diagnosed between 2015 and 2022 who had either brain metastases (n = 2,475) or bone metastases (n = 683). All patients received both SRT and IT, started 10–66 days apart. Patients were grouped by treatment sequence: SRT→IT or IT→SRT. Confounding was addressed using inverse-probability-of-treatment weighting (IPTW) and overlap weighting (OW) derived from propensity scores. OS was evaluated using Kaplan–Meier estimation and Cox proportional-hazards models within each metastasis site. Missing covariates were handled with multiple imputation by chained equations (25 datasets), and hazard ratios (HRs), 95% confidence intervals (CIs), and p values were combined across imputations using Rubin’s rules. Results: In the brain-metastasis cohort, 2,061 patients received SRT→IT and 414 received IT→SRT. Median OS was 27.4 months with SRT→IT versus 23.3 months with IT→SRT. The MI-pooled unweighted Cox model yielded an HR of 1.13 (95% CI 0.98–1.30; p = 0.081) for IT→SRT versus SRT→IT, while IPTW weighting produced a similar but statistically significant association (HR 1.10, 95% CI 1.02–1.19; p = 0.012); OW estimates were directionally similar but less precise (HR 1.10, 95% CI 0.90–1.35; p = 0.341). In the bone-metastasis cohort, 520 patients received SRT→IT and 163 received IT→SRT. IT→SRT was associated with longer survival (median OS 22.7 vs 15.5 months). The MI-pooled unweighted Cox HR for IT→SRT versus SRT→IT was 0.81 (95% CI 0.64–1.01; p = 0.064), with a statistically significant effect under IPTW weighting (HR 0.79, 95% CI 0.68–0.91; p = 0.001) and a directionally consistent but less precise OW estimate (HR 0.80, 95% CI 0.58–1.10; p = 0.165). Conclusions: Treatment-sequencing effects in metastatic NSCLC appear to be site specific. In patients with brain metastases, SRT→IT was associated with modestly longer OS, with sequence effects sensitive to the weighting approach. In patients with bone metastases, IT→SRT was associated with substantially longer OS, corresponding to an approximate 20% relative reduction in mortality in IPTW-weighted analyses. These findings support consideration of metastasis-site biology when determining the sequencing of IT and SRT and highlight the need for prospective trials explicitly designed to evaluate site-tailored multimodality strategies.