Association Between Dinutuximab Beta Exposure and Post‐End‐of‐Treatment Survival in Neuroblastoma: A Weighted Patient‐Level Analysis of Three Clinical Studies

OBJECTIVES: The association between exposure to dinutuximab beta (DB) and event-free survival (EFS) or overall survival (OS) of neuroblastoma patients was assessed using data collected during three clinical trials (five cohorts). METHODS: A systematic review (March 2026) was conducted to identify relevant studies (prospective; registered DB indication and posology). Patient-level information on outcomes and their predictors was extracted from study reports. Because of immortal-time and reverse causation biases, survival was examined among patients who were event-free at their end-of-treatment (EOT) date (i.e., last dose + 25 days). To address selection bias, stabilised inverse-probability weights were estimated, and weighted post-EOT survival models were fitted. RESULTS: Of 665 patients, 98 had relapse or progression before EOT. The median duration of follow-up in the post-EOT cohort was 3.78 years. The total number of cycles was independently associated with both EFS (HR = 0.80, 95% CI: 0.70-0.90 per cycle, p < 0.001) and OS (HR = 0.78, 95% CI: 0.68-0.90 per cycle, p < 0.001). With each treatment cycle, post-EOT hazard of an EFS event decreased by 20% (95% CI: 10%-30%) and hazard of death decreased by 22% (95% CI: 10%-32%). The results were consistent across most sensitivity analyses (e.g., excluding IL-2 recipients, restricting to frontline maintenance immunotherapy, alternative specifications and exposure metrics), but not among patients with relapsed or refractory neuroblastoma. In a target-trial emulation (5 cycles vs. <5 cycles), associations were significant for EFS (HR = 0.46, 95% CI: 0.32-0.66, p < 0.001) and OS (HR = 0.44, 95% CI: 0.28-0.67, p < 0.001). CONCLUSIONS: Greater prior exposure to DB, particularly a higher number of treatment cycles, was associated with better post-EOT EFS and OS. The study did not estimate an on-treatment causal effect of DB, and the residual confounding cannot be fully excluded. Prior treatment exposure may help inform post-treatment risk stratification.