Development and Validation of a Diagnostic Nomogram to Predict the Anthracycline-Induced Early Cardiotoxicity in Children with Hematological Tumors

Abstract

This study aimed to establish and validate an effective nomogram to predict the risk of cardiotoxicity in children after each anthracycline treatment. According to the inclusion and exclusion criteria, the eligible children were randomly divided into the training cohort (75%) and the validation cohort (25%). Least absolute shrinkage and selection operator (LASSO) regression was used to select the predictors and a nomogram was developed. Then, concordance index (C-index), the area under the curve (AUC), Hosmer–Lemeshow (H–L) test, and decision curve analysis (DCA) were employed to evaluate the performance and clinical utility of nomogram. Internal validation was processed to inspect the stability of the model. A total of 796 eligible children were included in this study and divided into a training set (n = 597) and a validation set (n = 199). LASSO regression analysis revealed that cumulative anthracycline dose, ejection fractions, NT-proBNP, and diastolic dysfunction were effective predictors of cardiotoxicity. The nomogram was established based on these variables. The C-index and the AUC of the predicting nomogram were 0.818 in the training cohort and 0.773 in the validation cohort, suggesting that the nomogram had good discrimination. The calibration curve of the nomogram presented no significant deviation from the reference line, and the P-value of the H–L test was 0.283, implying a preferable degree of calibration. The threshold of DCA also reflects that the nomogram is clinically useful. A nomogram was developed to predict anthracycline chemotherapy-induced cardiotoxicity in children with hematological tumors. The nomogram has a good prediction effect and can provide a reference for clinicians’ diagnosis and treatment.