Exploring Machine Learning for Predicting Peripheral and Central Precocious Puberty Through Cross-Hospital Validation

Precocious puberty, including Peripheral Precocious Puberty (PPP) and Central Precocious Puberty (CPP), presents diagnostic challenges in pediatric endocrinology, leading to delayed interventions. This study utilized machine learning models-Random Forest (RF), Gradient Boosting Machine (GBM), and Extreme Gradient Boosting (XGB)-to predict and differentiate between PPP and CPP using 12 clinical features extracted from electronic medical records (EMRs). Internal validation on TMUH data showed XGB achieving the highest sensitivity (0.88) and AUC (0.86). In external validation with WFH data, RF demonstrated superior generalizability, with a sensitivity of 0.91 and AUC of 0.89. These results highlight RF's robustness for cross-hospital implementation and the potential of machine learning to improve early diagnosis of precocious puberty.