MRI-based 3D Models of the Pelvis Can Replace CT-based 3D Models for Range-of-Motion Analysis in Femoroacetabular Impingement

Purpose Although femoroacetabular impingement (FAI) describes a dynamic osseous abutment of the femur against the acetabulum, current standard imaging assessment is static. Impingement analysis based on computed tomography (CT) was recently introduced, but magnetic resonance imaging (MRI) would offer a radiation-free alternative. Thus we asked (1) what the mean distance is between surface points of three-dimensional (3D) pelvis models derived from CT/MRI; (2) whether impingement-free range of motion (ROM) correlates between CT and MRI; and (3) whether zones of impingement match for 3D models based on CT and MRI.

Methods and Materials The institutional review board approved this comparative retrospective study of 20 symptomatic hips with FAI. Three-dimensional CT scans (isovoxel: 1 mm³) of the entire pelvis and the distal femoral condyles were obtained. Preoperative MR arthrograms of the hip were obtained including 0.8 mm³ isovoxel T1 3D volumetric interpolated brain examination (VIBE) and 1 mm³ isovoxel 3D T1 VIBE Dixon sequences of the entire pelvis and the distal femoral condyles. Threshold-based manual segmentation was performed using commercial software (Amira). Both 3D models were compared with software developed in house that includes two specific algorithms for detection of the acetabular rim and for detection of the center of rotation. We calculated percentage of the surface points with < 1-mm difference between the CT- and MR-based 3D models; assessed correlation of impingement-free ROM (in flexion; extension; internal rotation 90 degrees of flexion; external rotation in 90 degrees of flexion; abduction; and adduction) between CT and MRI; and compared the location of impingement zones between CT and MRI using the clock-face system that divides the femur and acetabulum into 12-hour positions.

Results Overall, 83% and 79% of the surface points of the proximal femur, respectively, of the acetabulum differed < 1 mm between the CT-based and MRI-based 3D models. Correlation for the ROM values was excellent (Spearman ρ: 0.993; p < 0.05) between CT and MRI. Location of impingement did not differ between CT-based and MRI-based ROM analysis in 12 of 12 acetabular and 11 of 12 femoral clock-face positions.

Conclusion MRI-based 3D models of the pelvis can replace CT-based 3D models for ROM analysis in FAI. The 3D MRI-based impingement analysis of the hip is a further step toward noninvasive personalized surgical planning of FAI, especially for complex deformities such as abnormal femoral torsion.