Deep learning-based automatic measurement of the femoral head ossification center in healthy Korean children: development of a novel radiographic growth chart

Category	Detail
Condition	Assessment of femoral head ossification centers (FHOC) in pediatric hip joints
Key Mechanisms	Deep learning-based automated segmentation and measurement of FHOC size from AP pelvic radiographs
Target Population	Healthy Korean children aged infancy through adolescence
Care Setting	Pediatric radiology and orthopedic clinical settings

FHOC appearance timing is a critical indicator for skeletal dysplasia, endocrinopathies, and developmental dysplasia of the hip (DDH).
Manual FHOC size measurements are limited by observer variability, time consumption, and reproducibility issues.
A three-stage cascaded deep learning algorithm enables accurate, reproducible, and efficient FHOC size measurement and growth chart creation.

Use AP pelvic radiographs to evaluate pediatric hip joint morphology and FHOC development relative to age.
Assess FHOC size and appearance timing as indicators for skeletal and developmental hip disorders.

Implement automated deep learning tools to standardize FHOC size measurement and reduce observer variability.
Use radiographic growth charts derived from healthy populations for precise diagnosis and follow-up.

Regularly monitor FHOC size progression using standardized imaging protocols and automated measurement algorithms.
Ensure pelvic radiographs meet criteria for pelvic rotation and tilt to maintain measurement consistency.

Be aware of potential inaccuracies in manual measurements due to ambiguous anatomical landmarks and image quality variability.
Consider radiation exposure minimization by using low-dose imaging systems such as EOS.

Healthy Korean children with normal growth parameters and no musculoskeletal or systemic bone-affecting conditions

Automated FHOC size measurement facilitates objective assessment and may guide early diagnosis and management of pediatric hip disorders.

Use standardized AP pelvic radiographs with controlled pelvic rotation and tilt indices for accurate FHOC assessment.
Apply deep learning-based segmentation and measurement algorithms to improve reproducibility and efficiency.
Reference age- and sex-specific radiographic growth charts for FHOC size to contextualize individual patient findings.

Clinical Scorecard: Automated Assessment of Femoral Head Ossification Centers in Healthy Korean Children Using Deep Learning: Creation of an Innovative Radiographic Growth Chart