Measurement of Glenoid Bone Loss With 3-Dimensional Magnetic Resonance Imaging: A Matched Computed Tomography Analysis

Abstract

To compare the measurement of glenoid bone surface area (GBSA) and glenoid bone loss (GBL) between 3-dimensional computed tomography (3D CT) and an autosegmentation approach for 3D magnetic resonance imaging (MRI) of patients with recurrent shoulder instability. Eight subjects (2 women and 6 men; age range, 15-72years [mean, 44 ± 19years]) were consecutively enrolled who had both CT and MRI of the shoulder for clinical shoulder instability. Inclusion criteria were patients with shoulder instability or other shoulder injury who had both a CT scan and MRI performed of the same shoulder. All patients underwent a 3D CT scan and a 3-Tesla 3D MRI with additional volumetric and autosegmented sequences. En face views of the glenoid for both CT and MRI were auto- and manually measured for overall GBSA and GBL using best-fit circle technique; the amount of GBL was compared with loss of GBSA and was expressed as a percentage of bone loss. There were no differences in GBL measured by 3D CT (41mm2, 6.6%) vs 3D MRI (40mm2, 6.5%, P= .852). The mean GBSA was not different among the manual- and autocalculated 3D CT (644mm2 vs 640mm2, P= .482). In addition, the manual MRI scan glenoid area was similar to the autocalculated 3D MRI (622mm2 vs 618mm2, respectively; P= .482). Overall regression analysis demonstrated excellent correlation between CT and MRI for both GBSA and GBL calculations (R2= 0.84-0.90). 3D MRI of the glenoid is nearly identical to 3D CT scans for measurement of GBSA and GBL, making 3D MRI a reliable alternative to a CT scan for a preoperative shoulder evaluation of the glenoid pathology. This study shows that a 3D MRI could be a radiation-free and reliable alternative to a preoperative CT shoulder scan. Level III, case-control study.