Magnetic Resonance Imaging Correlates With Computed Tomography for Glenoid Version Calculation Despite Lack of Visibility of Medial Scapula

Purpose: To assess the accuracy of measuring glenoid version on magnetic resonance imaging (MRI) in the presence of varying amounts of the medial scapula body as compared with the gold standard of glenoid version measured on computed tomography (CT) imaging, including the entire scapula in a cohort of young patients with shoulder instability and without glenohumeral arthritis.

Methods: A retrospective review was performed on instability patients with preoperative MRI and CT imaging. Measurements of available scapular width and glenoid version were performed using the Cobb angle method to measure the angle between the plane of the glenoid fossa to Friedman’s line on axial images. Intra- and interrater reliability analysis was performed using intraclass correlation coefficients to assess agreement between MRI and CT measurements. Paired t tests were used to compare measurement differences between MRI and CT.

Results: Thirty-two patients with both MRI and CT scans were assessed. Intra- and inter-rater assessment revealed strong agreement for scapular width measurement. For glenoid version measurement, intra-rater agreement was excellent and inter-rater agreement was moderate on CT and good on MRI. The mean available scapular body width was 24.7 mm longer on CT as compared with MRI (95% confidence interval 17.5-31.9, P < .0001; 109.8 ± 8.2 mm vs 85.1 ± 16.9 mm, respectively), with MRI having an average of 78.2% (±17.6%) of the CT scapular width shown on CT. No significant difference in glenoid version was found between MRI and CT (95% confidence interval -0.87 to 1.75, P = .499; MRI -2.57° vs CT -2.13°).

Conclusion: MRI provided significantly shorter available scapular widths when compared with CT imaging in a cohort of patients with glenohumeral instability and without arthritis. However, this failed to produce a significant difference of ≥5° in measured glenoid version compared with CT measurements when 75% (8 cm) of the scapular width was present on MRI. Measuring glenoid version on MRI does not appear to be significantly affected when the entirety of the medial border of the scapula is not included in the imaging field.

Level of evidence: Level III; study of diagnostic test.