Construction and Validation of a Model for Predicting Severe Chemotherapy-Induced Peripheral Neuropathy in Pediatric Cancer Patients Based on Patient-Reported Outcomes

Abstract Background Severe chemotherapy-induced peripheral neuropathy (CIPN) significantly compromises quality of life in pediatric oncology patients and often necessitates chemotherapy dose reduction, treatment discontinuation, or leads to life-threatening outcomes. Early detection and intervention are critical to halting CIPN progression and improving clinical outcomes. In this study, we developed and internally validated a predictive nomogram for severe CIPN in pediatric oncology patients by integrating treatment-related risk factors and patient-reported outcomes. Methods A prospective study was conducted from January 2024 to October 2025 at a tertiary hospital in Shanghai. 342 pediatric cancer patients receiving vincristine-containing regimens were consecutively enrolled. A logistic regression framework identified independent determinants of severe CIPN, and a nomogram was constructed. The predictive model was evaluated via the area under the receiver operating characteristic curve (AUC), the Brier score, the calibration curve and the Hosmer-Lemeshow test. Results Severe CIPN occurred in 15.5% of participants. The nomogram was created with the independent predictive factors including CIPN history, cumulative dose, and mean scores of patient-reported "numbness/tingling in the hands/feet" and "limb weakness". The AUC of the proposed model demonstrated acceptable discriminatory ability. The Brier score was 0.108. The calibration curve showed good consistency between the predicted and actual probabilities. The Hosmer-Lemeshow test indicated that the predictive model was well-calibrated. A concordance index quantifying predictive ranking ability was derived from internal validation. Conclusion The nomogram can identify pediatric cancer patients with vincristine-containing regimens at high risk for severe CIPN and may be a rapid visual tool for nurses. This model incorporates patient-reported outcomes to help nurses identify children with cancers who are at high risk for severe CIPN and allow timely preventive action.