Comparison of anatomical parameters of cam femoroacetabular impingement to evaluate hip joint models segmented from CT data

Abstract

Subject-specific anatomical finite element models obtained from three-dimensional (3D) segmentation have the potential to provide great insights into the pathomechanisms of femoroacetabular impingement (FAI). Still, the accuracy of the geometries used to construct these models needs to be evaluated. To this aim, we segmented 54 (n = 54; age = 34 ± 7 years; BMI = 26 ± 4 kg/m²) hip joint models from subject-specific computed tomography (CT) images, and measured multiple anatomical parameters (axial alpha angle, radial alpha angle, femoral head–neck offset, femoral neck–shaft angle, medial proximal femoral angle, femoral torsion, acetabular version and centre–edge angle) from both the multiplanar images and the 3D models, to assess the intraobserver, interobserver and intermethod reliabilities. We implemented a method to ensure that anatomical characteristics from segmented models were representative of original CT data. Observations from both CT data and 3D models demonstrated strong to near-perfect intraobserver, interobserver and intermethod agreements (p < 0.01). Bland–Altman plots indicated a slight discrepancy when assessing the asymptomatic FAI population, where planar CT images possibly did not capture the full depth of the cam deformity and underestimated geometric parameters. We indicated possible discrepancies to expect when segmenting hip joint models for clinical evaluation and finite element modelling, notably when observing femoral head–neck offset.

Keywords:

• Hip
• femoroacetabular impingement
• computed tomography
• segmentation
• models
• reliability

Acknowledgements

The authors wish to thank Brandon Fournier BASc, from the Department of Mechanical Engineering, for his help with CT segmentation; Andrew Adamczyk MD, from the Division of Orthopaedic Surgery, and Kawan Rakhra MD FRCPC, from the Department of Diagnostic Imaging, for their insight into CT readings.