The quest for optimal femoral torsion angle measurements: a comparative advanced 3D study defining the femoral neck axis.

Purpose: There is high variability in femoral torsion, measured on two-dimensional (2D) computed tomography (CT) scans. The aim of this study was to find a reliable three-dimensional (3D) femoral torsion measurement method, assess the influence of CAM deformity on femoral torsion measurement, and to promote awareness for the used measurement method.

Methods: 3D models of 102 dry femur specimens were divided into a CAM and non-CAM group. Femoral torsion was measured by one 2D-CT method described by Murphy et al. (method 0) and five 3D methods. The 3D methods differed in strategies to define the femoral neck axis. Method 1 is based on an elliptical least-square fit at the middle of the femoral neck. Methods 2 and 3 defined the centre of mass of the entire femoral neck and of the most cylindrical part, respectively. Methods 4 and 5 were based on the intersection of the femoral neck with a 25% and 40% enlarged best fit sphere of the femoral head.

Results: 3D methods resulted in higher femoral torsion measures than the 2D method; the mean torsion for method 0 was 8.12° ± 7.30°, compared to 9.93° ± 8.24° (p < 0.001), 13.21° ± 8.60° (p < 0.001), 8.21° ± 7.64° (p = 1.00), 9.53° ± 7.87° (p < 0.001) and 10.46° ± 7.83° (p < 0.001) for methods 1 to 5 respectively. In the presence of a CAM, torsion measured with method 4 is consistently smaller than measured with method 5.

Conclusion: 2D measurement might underestimate true femoral torsion and there is a difference up to 5°. There is a tendency for a higher mean torsion in hips with a CAM deformity. Methods 4 and 5 are the most robust techniques. However, method 4 might underestimate femoral torsion if a CAM deformity is present. Since method 5 is independent of a CAM deformity, it is the preferred technique to define expected values of torsion.