Automated Bone Age Assessment of Older Children using the Radius

Area of Science:

• Medical Imaging
• Pediatric Endocrinology
• Computer-Aided Diagnosis

Background:

• Skeletal development assessment is crucial for identifying growth pathologies.
• Current methods struggle with accurate bone age assessment in older children (16/14-19 years).
• Phalanges and carpal bones lack distinct features in fully developed skeletal structures.

Purpose of the Study:

• To develop a comprehensive Computer Aided Diagnosis (CAD) system for skeletal development assessment.
• To enhance bone age assessment accuracy across the entire pediatric age range (0-19 years).
• To integrate the radius bone into existing CAD algorithms for improved performance in older children.

Main Methods:

• Utilized a large dataset of 1,400 normal hand X-rays from children aged 0-18.
• Developed a CAD algorithm for bone age assessment using phalanges and carpal bones.
• Integrated the radius bone as a region of interest (ROI) to address limitations in older age groups.

Main Results:

• Previous CAD algorithms accurately assessed bone age for children up to 16 (male) or 14 (female) years.
• Preliminary studies indicate that incorporating the radius bone significantly improves assessment accuracy for older children.
• An integrated CAD system utilizing phalanges, carpal bones, and the radius shows robustness for ages 1-19.

Conclusions:

• The Digital Hand Atlas project aims to automate skeletal development assessment.
• Integrating the radius bone is essential for accurate bone age assessment in older children and adolescents.
• A comprehensive CAD system incorporating multiple hand bones offers a robust solution for pediatric skeletal development evaluation.