3D-MRI versus 3D-CT in the evaluation of glenoid deformity in glenohumeral arthritis using Dixon 3D FLASH sequence

To compare MRI with 3D reconstructions and 3D-CT with respect to assessment of glenoid wear in osteoarthritic shoulders.

3D reconstructions were generated for CT and MR (utilizing the Dixon technique) imaging performed on 29 osteoarthritic shoulders. Two reviewers independently performed glenoid morphometric measurements and evaluated glenoid erosion. Mean differences between the two modalities were calculated. Inter-observer agreement was calculated using kappa coefficient.

The combined mean absolute difference (bias) in glenoid version between 3D-CT and 3D-MRI was 2.7° ± 1.6° (range 0.15-7.85, P value = 0.7). The combined mean absolute difference in glenoid inclination between 3D-CT and 3D-MRI was 6.8° ± 4.1° (range 0.8°-15.75°, P value = 0.17). No significant inter-reader variation in glenoid version and inclination measurements on 3D-CT and 3D-MRI was found (P > 0.05). The inter-reader reliability for both CT and MRI was high for Walch grading of glenoid bone loss (κ = 1, κ = 0.81, respectively).

3D-MRI is comparable to 3D-CT with respect to axial glenoid bone loss, as measured by glenoid version. However, for coronal bone loss estimation, measured by glenoid inclination, 3D-CT remains the gold standard. Thus, 3D-MR can be used as an alternative for preoperative assessment of glenoid version in arthritic shoulders.