PediatricBody / AbdominalAI / InformaticsResearch
SVM model distinguishes central precocious puberty with AUC 0.83 in multicenter study
Radiology AI literature (PubMed)2w ago
A support vector machine model using eight clinical and ultrasound predictors diagnosed central precocious puberty in girls with AUCs of 0.85 (internal) and 0.83 (external) in a retrospective multicenter study of 2,148 patients.
- Retrospective design across 4 Chinese centers; 2,148 girls with precocious puberty included.
- Key predictors: bone age minus chronological age, uterine volume, basal luteinizing hormone, among others.
- Retrospective; external validation still limited to Chinese children; prospective validation needed.
RadPigeon summaries are original and for information only. They are not clinical advice.