Risk factors and novel predictive models for metastatic neuroblastoma in children

Abstract

IntroductionNeuroblastoma (NB) with distant metastasis (DM) is a high-risk condition with a poor prognosis. Early identify the risk and prognostic differences of DM in children, which is helpful for the development of clinical diagnosis and treatment. MethodsThe study cohort included patients with NB in surveillance, epidemiological, and final outcome databases between 2010 and 2018. To identify the risk and prognostic factors for DM, both univariate and multivariate logistic and Cox regression analyses were conducted. In addition, we created and verified three online clinical prediction models. Finally, we assess the performance of the proposed predictive model. ResultsAmong the 1224 children with NB included in the study, 599 developed DM. Primary site is the most important factor affecting metastasis and prognosis. The training and validation groups of the diagnostic nomograms had area under curves (AUC) values of 0.872 and 0.824, respectively. In addition, in the training group, the AUC values at 12, 36, and 60 months were 0.68, 0.71, and 0.75 for the OS nomogram and 0.70, 0.72, and 0.75 for the CSS nomogram. In the validation group, the AUC values at 12, 36, and 60 months were 0.68, 0.72, and 0.70 for the OS nomogram and 0.67, 0.71, and 0.69 for the CSS nomogram, respectively. Calibration curve and decision curve analyses revealed good performance of the nomogram. ConclusionsThe nomogram developed in this study could appropriately predict DM and assess its prognosis in patients with NB.