Concurrent Versus Sequential PD-1 Blockade With Locoregional Radiotherapy After Chemoimmunotherapy in De Novo Metastatic Nasopharyngeal Carcinoma: A Landmark Analysis

Background Chemoimmunotherapy followed by consolidative locoregional radiotherapy has increasingly been used for selected patients with de novo metastatic nasopharyngeal carcinoma (dmNPC). However, whether PD-1 blockade should be continued concurrently during locoregional radiotherapy (LRRT) remains uncertain. Methods We retrospectively analyzed 238 patients with de novo metastatic nasopharyngeal carcinoma who received platinum-based chemotherapy plus anti–PD-1 therapy followed by definitive locoregional radiotherapy. A 12-month landmark analysis was performed to reduce immortal time bias, and inverse probability of treatment weighting was used to balance baseline and treatment-related characteristics. Recursive partitioning analysis incorporating disease burden, pretreatment lactate dehydrogenase, post-chemoimmunotherapy Epstein–Barr virus DNA clearance, and radiological response was used for risk stratification. Longitudinal absolute lymphocyte counts were evaluated after radiotherapy. Results Among 238 patients, 185 received concurrent PD-1 blockade during locoregional radiotherapy and 53 did not. After weighting and landmark adjustment, concurrent PD-1 blockade was associated with inferior progression-free survival (PFS) (weighted hazard ratio, 3.189; 95% CI, 1.352–7.524; P=0.008). Distant metastasis-free survival was also inferior in the concurrent group, whereas locoregional recurrence-free survival was similar between groups. Among the four observed treatment patterns, the sandwich strategy, defined as chemoimmunotherapy followed by LRRT without concurrent PD-1 blockade and with subsequent maintenance immunotherapy, showed the most favorable outcomes. In the landmark-adjusted comparison restricted to the full-course and sandwich strategies, the sandwich strategy remained associated with superior PFS (hazard ratio, 0.254; 95% CI, 0.078-0.821; P=0.022). Exploratory subgroup analyses suggested a less favorable outcome pattern with concurrent PD-1 blockade among high-risk patients, although individual subgroup sizes were limited. Persistent lymphopenia at 3 months after radiotherapy was associated with poor outcomes, particularly among patients receiving concurrent PD-1 blockade. Conclusion In conclusion, among patients with dmNPC treated with first-line chemoimmunotherapy followed by definitive LRRT, concurrent PD-1 blockade during LRRT was not associated with improved survival outcomes and was linked to inferior PFS and DMFS in an IPTW-adjusted landmark analysis. Exploratory analyses suggested potential heterogeneity according to baseline risk and post-radiotherapy immune recovery, but these findings were limited by small subgroup sizes. Prospective studies are warranted to define the optimal sequencing of PD-1 blockade and LRRT.