Regression models to predict hip joint centers in pathological hip population.

Area of Science:

• Biomechanics
• Orthopedic Surgery
• Medical Imaging

Background:

• Accurate hip joint center (HJC) localization is crucial for gait analysis and orthopedic surgery.
• Existing regression equations for HJC prediction, like Harrington's and Davis's, may have limitations in populations with hip deformities such as femoroacetabular impingement (FAI).

Purpose of the Study:

• To evaluate the accuracy of Harrington's and Davis's HJC regression equations in a cohort with cam-type FAI.
• To assess the influence of skin thickness on HJC prediction accuracy.
• To develop and validate new adult-specific HJC regression equations.

Main Methods:

• Pelvic CT imaging was performed on 67 participants (21 healthy, 46 with FAI).
• Geometric HJCs and bony landmarks were digitized; skin thickness was measured.
• Non-parametric statistical and Bland-Altman tests compared predicted vs. actual HJC.
• New adult-specific regression models were developed and validated using external datasets and leave-one-out cross-validation.

Main Results:

• Harrington's model (12.3 ± 5.9 mm error) showed significantly less error than Davis's model (25.0 ± 6.7 mm error).
• No significant differences in prediction error were found between healthy controls and the FAI group.
• Measured skin thickness (anterior: 9.7 ± 7.0 mm, posterior: 19.6 ± 10.9 mm) correlated with BMI and may reduce systematic error.
• New adult-specific equations yielded prediction errors comparable to Harrington's model.

Conclusions:

• Conventional regression models, particularly Harrington's, are valid for HJC prediction in individuals with cam-type FAI.
• Skin thickness measurements can potentially improve HJC prediction accuracy by accounting for soft tissue variability.
• The developed adult-specific regression equations did not significantly improve prediction accuracy over existing models in this study.