Both Radiologists and Surgeons Miss a Substantial Number of Subscapularis Tears on Magnetic Resonance Imaging Examination Prior to Shoulder Arthroscopy

Abstract

Purpose: To compare radiologist and surgeon magnetic resonance imaging (MRI) interpretations of subscapularis (SSC) tears against intraoperative arthroscopic examination. Methods: We conducted a retrospective review of prospectively maintained data on patients who underwent arthroscopic rotator cuff repair with and without SSC tears between 2011 and 2022. The radiologists’ assessments of SSC integrity were extracted from the MRI reports. One high-volume fellowship-trained shoulder surgeon assessed all MRI scans for the presence or absence of SSC tears. Radiologist and surgeon MRI-based assessments were compared against the diagnostic gold standard (intraoperative arthroscopic examination) and classified according to the Lafosse classification. Sensitivity, specificity, positive and negative predictive values, and accuracy were calculated. Results: A total of 1,090 patients were included for analysis, 839 with intraoperatively confirmed SSC tears (77%) and 251 without such tears (23%). Radiologists diagnosed SSC tears with a sensitivity of 56.0% (470 of 839 patients) and specificity of 67.3% (169 of 251 patients) via MRI, whereas the surgeon diagnosed SSC tears with a sensitivity of 71.4% (599 of 839 patients) and specificity of 78.1% (196 of 251 patients). Overall, radiologists only correctly identified 56% of SSC tears whereas the shoulder surgeon correctly identified approximately 71.4% of tears. When considering arthroscopically confirmed tear size, that is, intact, partial, full thickness [upper border], and complete, the radiologists’ diagnosis rates with MRI were 67.3%, 40.4%, 64.6%, and 78.3%, respectively, and the surgeon’s diagnosis rates were 78.1%, 52.7%, 81.9%, and 97.5%, respectively. Conclusions: In a community practice, radiologists miss approximately 50% of SSC tears on MRI examination. A shoulder surgeon with the benefit of clinical examination misses 28.6% of tears on MRI. Accuracy increases as tear size increases, with radiologists detecting 40.4% of partial tears and 78.3% of complete tears and a surgeon detecting 52.7% and 97.5%, respectively. Level of Evidence: Level III, diagnostic study (nonconsecutive patients).