Trends in conditional survival and predictors of late death in neuroblastoma

Purpose Significant advances in the treatment of neuroblastoma have been made in the past several decades. There are scant data examining how these improvements have changed over time and differentially affected conditional survival among high‐risk and non‐high‐risk patient groups. Methods We conducted a retrospective cohort study using the Surveillance, Epidemiology, and End Results database. We analyzed clinical characteristics and survival outcomes for 4717 neuroblastoma patients. Kaplan‐Meier methods were used to estimate overall survival (OS) and conditional overall survival (COS) with estimates compared between groups using log‐rank tests. Results Five‐year OS was 41.46% (95% CI 38.77‐44.13) for the high‐risk group and 91.13% (95% CI 89.49‐92.53) for the non‐high‐risk group. Both groups saw significant improvements in OS by decade ( P  < .001). Five‐year COS among 1‐year survivors was 52.69% (CI 49.54‐55.73) for the high‐risk group and 96.75% (95% CI 95.57‐97.62) for the non‐high‐risk group. One‐year survivors in the high‐risk group showed a statistically significant improvement in COS over time. No difference in COS was observed among 5‐year high‐risk survivors. In the high‐risk and non‐high‐risk groups, 82% and 32% of late deaths were attributable to cancer, respectively. Statistically significant adverse prognostic factors for late death were age ≥ 1 year at diagnosis, metastatic disease, and nonthoracic primary site ( P  = .001). Conclusions Improvements in COS over time have largely benefited high‐risk patients, though they are still at higher risk for late death due to cancer when compared to non‐high‐risk patients. Age, stage, and primary site, but not treatment decade, influence outcomes among 5‐year survivors.