Related Experiment Video
Updated: Jul 5, 2025
Scanning Skeletal Remains for Bone Mineral Density in Forensic Contexts
Published on: January 29, 2018
Clinical validation of a deep-learning-based bone age software in healthy Korean children
Hyo-Kyoung Nam1, Winnah Wu-In Lea2, Zepa Yang3,4
1Department of Pediatrics, Korea University College of Medicine, Seoul, Korea.
Insights
Deep learning bone age software showed low accuracy in estimating chronological age in healthy Korean children, tending to underestimate bone age, especially in younger children.
Area of Science:
- Pediatric endocrinology
- Medical imaging analysis
- Artificial intelligence in healthcare
Background:
- Bone age (BA) assessment is crucial for evaluating developmental status and growth disorders in children.
- Accurate BA estimation aids in diagnosing and managing pediatric growth abnormalities.
Purpose of the Study:
- To evaluate the clinical performance of a deep learning (DL)-based software for bone age estimation in healthy Korean children.
- To compare the DL software's BA estimations with chronological age (CA) using concordance rates and Bland-Altman analysis.
Main Methods:
- Retrospective analysis of 553 left-hand radiographs from 371 healthy Korean children (aged 4-17 years).
- Evaluation using a commercial DL-based BA software (BoneAge, Vuno).
- Clinical performance assessed via concordance rate and Bland-Altman analysis against CA.
Main Results:
- A significant difference was found between CA and DL-estimated BA (P<0.001).
- Good correlation (r=0.96) was observed, but with a root mean square error of 15.4 months.
- The DL software demonstrated a 58.8% concordance rate with a 12-month cutoff and tended to underestimate BA in children under 8.3 years.
Conclusions:
- The DL-based BA software exhibited a low concordance rate in healthy Korean children.
- The software showed a tendency to underestimate bone age, particularly in younger children.
Purpose:
Bone age (BA) is needed to assess developmental status and growth disorders. We evaluated the clinical performance of a deep-learning-based BA software to estimate the chronological age (CA) of healthy Korean children.
Methods:
This retrospective study included 371 healthy children (217 boys, 154 girls), aged between 4 and 17 years, who visited the Department of Pediatrics for health check-ups between January 2017 and December 2018. A total of 553 left-hand radiographs from 371 healthy Korean children were evaluated using a commercial deep-learning-based BA software (BoneAge, Vuno, Seoul, Korea). The clinical performance of the deep learning (DL) software was determined using the concordance rate and Bland-Altman analysis via comparison with the CA.
Results:
A 2-sample t-test (P<0.001) and Fisher exact test (P=0.011) showed a significant difference between the normal CA and the BA estimated by the DL software. There was good correlation between the 2 variables (r=0.96, P<0.001); however, the root mean square error was 15.4 months. With a 12-month cutoff, the concordance rate was 58.8%. The Bland-Altman plot showed that the DL software tended to underestimate the BA compared with the CA, especially in children under the age of 8.3 years.
Conclusion:
The DL-based BA software showed a low concordance rate and a tendency to underestimate the BA in healthy Korean children.
More Related Videos
09:02Cortical Bone Assessment Using Ultrasonic Guided Waves: A Reproducibility Study in a Healthy Population
Published on: January 31, 2025
07:12Semiautomated Longitudinal Microcomputed Tomography-based Quantitative Structural Analysis of a Nude Rat Osteoporosis-related Vertebral Fracture Model
Published on: September 28, 2017