Computer-aided King classification of scoliosis

Area of Science:

• Orthopedic surgery
• Medical imaging analysis
• Spinal deformities

Background:

• Scoliosis surgical planning depends on accurate spinal curve classification.
• The traditional King classification is prone to human measurement and judgment errors, leading to variability.
• Reducing classification variability is crucial for consistent surgical planning.

Purpose of the Study:

• To develop and evaluate a computer-aided method for King classification of scoliosis.
• To reduce the inherent variability associated with manual King classification.
• To assess the reliability of the proposed computerized system.

Main Methods:

• A computerized system measures vertebral endplate inclination and position from posteroanterior radiographs.
• The system automatically identifies apical and end vertebrae, calculates Cobb angles, and classifies curve types.
• Kappa statistics were used to evaluate intraobserver and interobserver reliability with five observers.

Main Results:

• The computerized system significantly improved average intraobserver kappa values from 0.75 to 0.90.
• Average interobserver kappa values improved from 0.66 to 0.86 with the aid of the system.
• The system demonstrated enhanced reliability in classifying 105 scoliotic cases.

Conclusions:

• The developed computerized system effectively assists surgeons in performing King classification for scoliosis.
• The system reduces variability in scoliosis curve classification.
• This tool has the potential to improve the consistency of surgical planning for scoliosis.