组内相关
医学
超声波
射线照相术
四分位间距
一致性
队列
前瞻性队列研究
超声学家
核医学
放射科
内科学
临床心理学
心理测量学
作者
Shujun Chen,Yijun Xiong,Junjie Liu,Wei Li,Siqi Tang,Haixia Zhang,Yingping Zhang,Yong Luo,Pan Li
标识
DOI:10.1016/j.ultrasmedbio.2025.07.020
摘要
Abstract Objectives This study aims to develop an accurate and straightforward bone age (BA) assessment model using high-frequency ultrasound. It evaluates the ossification rates of five target bones at the wrist and knee joints, reallocating weighting values to enhance assessment precision. Methods This observational study (July 2023–March 2024) involved children aged 0–19 years, who were divided into pilot, modelling, and validation groups. High-frequency ultrasound assessed ossification rates at the wrist and knee joints, with radiographic BAs determined via the TW3-C-RUS method. Statistical analyses evaluated the correlation and accuracy between ultrasound-based and radiographic BAs. Results The study included 2,049 children (median age 9 years) divided into pilot, modelling, and validation groups. The ossification rates of the five target sites were strongly positively correlated with the BA. Using ridge regression, a robust ultrasound-based BA model was established, which demonstrated high accuracy with an intraclass correlation coefficients of 0.99, compared with radiographic BA. In the validation group, ultrasound showed high sensitivity and specificity for diagnosing abnormal BAs, with strong inter- and intra-rater reliability. Conclusion This study successfully generated an accurate ultrasound-based BA model that demonstrates high sensitivity and specificity, showing strong consistency with radiographic bone age assessments. Trial registration NO. ChiCTR2400080249 (retrospectively registered).
科研通智能强力驱动
Strongly Powered by AbleSci AI