Assessment Model of Pediatric Bone Age Based on High-Frequency Ultrasound: A Prospective Cohort Study

Ultrasound-based assessment of bone age (BA) in pediatric populations remains underutilized and insufficiently validated. The purpose of this study was to develop a simple and accurate model for estimating BA using high-frequency ultrasonography.

Participants were prospectively stratified into pilot, modeling, and validation groups based on enrollment dates and predefined inclusion criteria. High-frequency ultrasound was employed to evaluate ossification rates (ORs) at five skeletal sites located at the wrist and knee joints. A ridge regression model was used to optimize the weighting of individual ORs and to construct the ultrasound-based BA estimation model. Radiographic BA, determined using the TW3-C-RUS method, was used as the reference standard. Correlation and agreement between ultrasound-based and radiographic BA measurements were assessed using appropriate statistical methods.

A total of 2001 children (median age, 9.5 years; interquartile range, 6.6-13.2 years; 983 boys) were included in this study. ORs at all five anatomical sites demonstrated strong positive correlations with BA. The ridge regression model yielded high predictive accuracy. In the validation cohort, the ultrasound-based BA showed excellent concordance with radiographic BA (intraclass correlation coefficient [ICC] = 0.99). The model also demonstrated high diagnostic performance for identifying abnormal BA, with sensitivity of 93.8% in boys and 100% in girls, and specificity of 100% in both sexes.

This study presents a rigorously validated ultrasound-based model for pediatric BA assessment. The model achieves high sensitivity, specificity, and agreement with established radiographic standards, supporting its clinical utility as a radiation-free alternative for BA evaluation.

ChiCTR2400080249 (retrospectively registered).