Is three-dimensional reconstruction really necessary for glenoid version measurement on magnetic resonance imaging in patients with glenohumeral instability?

Abstract

BACKGROUND: The importance of glenoid version in glenohumeral instability (GHI) has become increasingly recognized. However, it remains unclear whether three-dimensional (3-D) correction of magnetic resonance imaging (MRI) scans provides more accurate glenoid version measurement in patients with GHI. The primary hypothesis of this study was that a statistically significant difference would exist between glenoid version values measured on standard two-dimensional (2-D) axial slices and those measured on 3-D-corrected axial slices. METHODS: A retrospective analysis of glenoid version measurements was performed in 54 patients (22 females, 32 males) with glenohumeral instability. Measurements were obtained using the Friedman method on standard 2-D axial MRI slices and 3-D-corrected axial MRI slices. Two observers performed each measurement twice. Measurements obtained by each observer were compared using paired t-tests. Measurement reliability was evaluated using intraclass correlation coefficients (ICC). RESULTS: For observer 1, mean glenoid version values were-0.69 degrees +/- 3.95 degrees and-1.24 degrees +/- 3.94 degrees on standard and 3-D-corrected axial MRI slices, respectively (p=0.16). For observer 2, corresponding values were-2.97 degrees +/- 4.48 degrees and-2.81 degrees +/- 4.92 degrees, respectively (p=0.66). Interobserver reliability was good for both techniques (ICC=0.87 for 2-D and 0.83 for 3-D measurements). Intraobserver reliability was excellent for measurements performed on 2-D slices (ICC=0.95 and 0.97) and ranged from good to excellent for measurements performed on 3-D-corrected slices (ICC=0.86 and 0.98). CONCLUSION: The findings of this study suggest that 3-D correction may not be necessary for glenoid version measurements performed on MRI in patients with glenohumeral instability. Multicenter studies with larger patient populations should be conducted to enhance the clinical relevance of these findings.