Assessment of Femoral Torsion on Magnetic Resonance Imaging is More Reliable Using Axial-Oblique Sequences Compared With Standard Axial Slices in Patients With Femoroacetabular Impingement Syndrome

Purpose: To determine the agreeability of femoral torsion measurements on axial and oblique axial magnetic reso-nance imaging (MRI) sequences in patients with femoroacetabular impingement syndrome (FAIS). Methods: Patients who underwent primary hip arthroscopy for FAIS between January 2012 to January 2019 were identified. Inclusion criteria were all patients with an MRI scan containing the pelvis and knee imaging. MRI-based measurements of femoral torsion were performed on axial and oblique-axial slices by 2 raters, and inter-rater and intrarater reliability was assessed. Bland Altman plots were constructed to evaluate the agreeability between femoral torsion measurements performed using axial and oblique-axial slices. Bivariate correlation analyses were performed to assess the relationship between measurement methods on each respective scan. A linear regression was performed between measurements performed using axial and oblique-axial sequences. Results: A total of 164 patients were included. The mean true -axial and oblique axial femoral torsion were 12.2 degrees +/- 9.9 degrees and 11.1 degrees +/- 9.2 degrees, respectively. The intrarater reliability for axial and oblique-axial measurements were 0.993 and 0.997, respectively. The inter-rater reliability for axial and oblique-axial measurements were 0.925 and 0.965, respectively. The number of differences within the limits of agreement for axial and oblique-axial femoral torsion measurements was 58.54%. On Pearson correlation analysis, strong positive correlations were found between oblique-axial measurements at multiple time points (r = 0.994, P < .001), as well as axial measurements at multiple time points (r = 0.986, P < .001). A strong positive correlation was found between axial and oblique-axial measurements (r = 0.894, P < .001). A significant regression equation indicated that for each additional increase in axial femoral torsion, the oblique-axial femoral torsion increased 0.837 (95% confidence interval 0.772-0.901). Conclusions: Femoral torsion values measured on oblique-axial sequences are smaller than on true-axial sequences. Femoral torsion measurements on axial and oblique-axial MRI sequences exhibit poor agreement. Oblique-axial sequences demonstrated greater measurement consistency at multiple timepoints. When evaluating torsional measurements, it is important to delineate which axial sequence was used, especially in patients with suspected severe femoral antetorsion. Standardization of MRI femoral version protocols within one's practice can ensure more consistent decision-making, especially in patients with suspected femoral antetorsion.