Development and validation of a machine learning model for predicting postherpetic neuralgia risk.

Objective: Postherpetic neuralgia (PHN) is a debilitating complication of herpes zoster (HZ), and early identification of high-risk patients is crucial for timely intervention. This study aimed to develop and validate machine learning models to predict the risk of developing PHN following HZ onset. Methods: A retrospective analysis of two prospective cohorts was performed. The training cohort comprised 627 patients from the Affiliated Hospital of Putian University, and an independent external validation cohort included 219 patients from Zhangzhou Affiliated Hospital of Fujian Medical University. Least Absolute Shrinkage and Selection Operator (LASSO) regression and the Boruta algorithm were used for feature selection. Ten ML models were constructed and evaluated based on metrics including the area under the receiver operating characteristic curve (AUC), accuracy, sensitivity, F1 score, calibration, and clinical utility. The optimal model was further interpreted using SHapley Additive exPlanations (SHAP). Results: The incidence of PHN was 19.0% (119 of 627) in the training cohort and 22.8% (50 of 219) in the validation cohort. Subsequently, five key risk factors were identified. Among the 10 models, XGBoost exhibited the best comprehensive performance, with an AUC of 0.826 (95%CI: 0.786-0.866) in the training cohort and 0.840 (95%CI: 0.784-0.896) in the validation cohort. SHAP analysis revealed that age was the most important predictor, followed by timing of antiviral therapy, acute pain severity, prodromal phase pain, and diabetes. Conclusion: The XGBoost model based on five clinically accessible factors effectively predicts the risk of PHN. This tool can assist clinicians in early risk stratification and guide personalized management for patients with HZ.