3D-MR vs. 3D-CT of the shoulder in patients with glenohumeral instability.

OBJECTIVE: To determine whether 3D-MR osseous reformats of the shoulder are equivalent to 3D-CT osseous reformats in patients with glenohumeral instability.
MATERIALS AND METHODS: Patients with glenohumeral instability, who were to be imaged with both CT and MRI, were prospectively selected. CT and MR were performed within 24 h of one another on 12 shoulders. Each MR study included an axial 3D isotropic VIBE sequence. The image data from the isotropic VIBE sequence were post-processed using subtraction and 3D software. CT data were post-processed using 3D software. The following measurements were obtained for both 3D-CT and 3D-MR post-processed images: height and width of the humeral head and glenoid, Hill-Sachs size and percent humeral head loss (if present), size of glenoid bone loss and percent glenoid bone loss (if present). Paired t-tests and two one-sided tests for equivalence were used to assess the differences between imaging modalities and equivalence.
RESULTS: The measurement differences from the 3D-CT and 3D-MR post-processed images were not statistically significant. The measurement differences for humeral height, glenoid height and glenoid width were borderline statistically significant; however, using any adjustment for multiple comparisons, this failed to be significant. Using an equivalence margin of 1 mm for measurements and 1.5% for percent bone loss, the 3D-MR and 3D-CT post-processed images were equivalent.
CONCLUSION: Three-dimensional-MR osseous models of the shoulder using a 3D isotropic VIBE sequence were equivalent to 3D-CT osseous models, and the differences between modalities were not statistically significant.