A diagnostic modeling study showed that a three-variable machine learning model can distinguish central precocious puberty from benign breast development in girls using only age, baseline follicle-stimulating hormone, and the baseline luteinizing hormone-to-follicle-stimulating hormone ratio. The study included 130 girls with early pubertal development, with a nested-validation AUC of 0.813 and a pooled out-of-fold AUC of 0.805. Retrospective estimates for dual-threshold triage suggested that approximately 60.0% of girls might be able to defer stimulation testing. However, five cases of central precocious puberty were initially classified for observation, highlighting the need for multicenter prospective validation and local calibration.
Source:BIOENGINEER.ORG · Read the original article
Leave a Reply