AUTOMATED BONE AGE ESTIMATION USING ARTIFICIAL INTELLIGENCE - Boneage.io® - IN HEALTHY CHILDREN

Abstract

Bone age assessment is vital for diagnosing and
managing growth disorders. Traditional methods like the
Greulich & Pyle Atlas and Tanner-Whitehouse 3 (TW3) method
are either quick but less accurate or detailed but labor-intensive.
This study evaluates the accuracy and reproducibility of
Boneage.io®, a cloud-based AI solution using the TW3 method to
estimate bone age in healthy Korean children aged 6–13. A total
of 1,040 radiographs were analyzed; the results showed minimal
deviation between estimated bone age and chronological age,
with Cohen's D effect sizes of 0.021566 for boys and 0.026172 for
girls. Boneage.io® provides reliable, real-time monthly bone age
results, effectively addressing challenges of traditional methods
and demonstrates high accuracy and reproducibility for clinical
use.