RESUMO
OBJECTIVE: To investigate the accuracy of conventional magnetic resonance imaging (MRI) in determining the severity of glenoid bone loss in patients with anterior shoulder dislocation by comparing the results with arthroscopic measurements. SUBJECTS AND METHODS: Institutional review board approval and written consent from all patients were obtained. Thirty-six consecutive patients (29 men, seven women; mean age, 34.5 [range, 18-55] years) with recurrent anterior shoulder dislocation (≥3 dislocations; mean, 37.9; range, 3-200) and suspected glenoid bone loss underwent shoulder MRI before arthroscopy (mean interval, 28.5 [range, 9-73] days). Assessments of glenoid bone loss by MRI (using the best-fit circle area method) and arthroscopy were compared. Inter- and intrareader reproducibility of MRI-derived measurements was evaluated using arthroscopy as a comparative standard. RESULTS: Glenoid bone loss was evident on MRI and during arthroscopy in all patients. Inter- and intrareader correlations of MRI-derived measurements were excellent (intraclass correlation coefficient = 0.80-0.82; r = 0.81-0.86). The first and second observers' measurements showed strong (r = 0.76) and moderate (r = 0.69) interreader correlation, respectively, with arthroscopic measurements. CONCLUSIONS: Conventional MRI can be used to measure glenoid bone loss, particularly when employed by an experienced musculoskeletal radiologist.