Overall survival after relapse in high-risk neuroblastoma: an adjusted indirect comparison of post-relapse dinutuximab beta, historical controls, and relapse after first-line dinutuximab beta maintenance.

Objective: To compare overall survival (OS) in patients with relapsed neuroblastoma treated with dinutuximab beta (DB) after relapse with that of two comparator populations: historical controls with no prior anti-disialoganglioside-2 exposure and patients who relapsed after first-line DB maintenance immunotherapy. The objectives were to evaluate the effectiveness of DB in the relapsed setting and to assess a potential legacy effect of prior first-line therapy. Methods: This unanchored indirect comparison used individual patient data from three cohorts: historical controls with no prior anti-GD2 therapy (N/N), patients who relapsed after first-line DB maintenance immunotherapy and did not receive subsequent DB (DB/N), and patients treated with DB after relapse without prior exposure (N/DB). Harmonized eligibility criteria were applied across cohorts, and follow-up was aligned using a common post-relapse index date. Adjusted comparisons were performed using propensity score reweighting. Overall survival was measured from the observed (N/DB) or imputed (N/N and DB/N) post-relapse index date and analyzed using weighted Kaplan-Meier methods and weighted Cox proportional hazards models, with hazard ratios (HR) reported. For the N/N and DB/N cohorts, the post-relapse index date was imputed using a quantile regression model estimating the interval from first relapse to initiation of DB, fitted to data from the N/DB cohort. Results: A total of 329 patients were assessed, of whom 266 were included in the weighted analyses. Baseline imbalances in age at relapse, time from diagnosis, MYCN status, and International Neuroblastoma Staging System (INSS) stage were largely resolved after weighting. No evidence of a legacy effect of first-line DB exposure was observed, as OS did not differ significantly between patients who relapsed after first-line DB maintenance immunotherapy and historical controls (HR = 1.26, 95% CI 0.70-2.26; p=0.437). In contrast, DB administered after relapse was associated with significantly improved OS compared with both historical controls (HR = 0.53, 95% CI 0.35-0.81; p=0.003) and patients who relapsed after first-line DB maintenance immunotherapy (HR = 0.42, 95% CI 0.25-0.72; p=0.001). These findings were consistent across sensitivity analyses. Conclusion and limitations: This adjusted indirect comparison demonstrated that DB administered after relapse was associated with significantly improved OS compared with both historical controls and patients who relapsed after prior first-line DB maintenance immunotherapy. No evidence of a legacy effect of first-line DB immunotherapy was identified. These findings support the clinical relevance of DB in the relapsed setting and indicate that prior first-line DB maintenance alone was not associated with a persistent post-relapse survival advantage. The study did not evaluate DB retreatment after prior first-line exposure.