# Predictive Modeling of Central Precocious Puberty Using IGF-1 and IGFBP-3 Standard Deviation Scores

## TL;DR

This study shows that using IGF-1 and IGFBP-3 standard deviation scores in predictive models can help diagnose early puberty in children more effectively.

## Contribution

The study introduces a predictive modeling approach using IGF-1 and IGFBP-3 SDS for diagnosing central precocious puberty in Korean children.

## Key findings

- CPP diagnosis rates were higher in boys (65.6%) compared to girls (54.2%).
- FSH had the highest AUCs for CPP diagnosis using raw values with sex-specific cutoffs.
- Models using IGF-1 and IGFBP-3 SDS improved diagnostic accuracy with AUCs of 0.800 in girls and 0.920 in boys.

## Abstract

Background/Objectives: Central precocious puberty (CPP) is diagnosed via gonadotropin-releasing hormone (GnRH) stimulation testing, which can be burdensome in pediatric settings. This study evaluated the utility of baseline hormonal markers—particularly insulin-like growth fac-tor 1 (IGF-1) and IGF-binding protein 3 (IGFBP-3)—as auxiliary tools for CPP diagnosis in Korean children. Methods: We retrospectively analyzed patients who underwent GnRH stimulation testing. Baseline LH, FSH, IGF-1, and IGFBP-3 levels were assessed, along with standard deviation scores (SDS) calculated using two different reference intervals. Multivariable logistic regression was performed to improve diagnostic accuracy. Performance was evaluated using area under the curve (AUC) values from receiver operating characteristic (ROC) analyses, stratified by sex. Results: Among 2464 Korean children (2025 girls and 439 boys), CPP diagnosis rates were 54.2% in girls and 65.6% in boys. Among baseline markers, FSH showed the highest AUCs using raw values with sex-specific cutoffs (AUC = 0.767 in girls and 0.895 in boys). Although IGF-1 SDS and IGFBP-3 SDS showed AUCs < 0.7 when used alone, predictive models incorporating these SDS values yielded higher performance (AUC = 0.800 in girls and 0.920 in boys. Conclusions: SDS-based IGF-1 and IGFBP-3 enhance CPP diagnosis when used in predictive models, emphasizing the need for sex-specific interpretation and standardized reference intervals in real-world clinical practice.