A machine learning model for identifying idiopathic central precocious puberty in girls based on medical images and clinical multi-parameters

Early identification of precocious puberty (PP) is important to guarantee the growth and development of children. The aim of this study was to propose a robust machine learning model that incorporate information from pituitary MRI images, carpal bone age, gonadal ultrasound, baseline sex hormone tests, and clinical information to identify idiopathic central precocious puberty (ICPP). The experiments show that the AUCs are 0.860, 0.862, and 0.866 respectively for the training set, internal validation sets, and external validation sets. The performance suggests that the presented model could be an alternative clinical approach.