Comparison of MRI and arthroscopy in diagnosis of shoulder joint injury

ZHANG Lu; ZHAO Fei; CHENG Tianming; GE Yinghui

doi:10.19300/j.2017.L5159zt

Abstract：Objective To investigate the performance of conventional MRI in the diagnosis of shoulder joint injury. Methods Retrospective analysis was performed on data from 55 patients with shoulder joint injury who had undergone conventional MRI and arthroscopy. The MRI indicators included the abnormal morphology and signal of rotator cuff and labrum, the impingement of the subacromion, and the bony injury. Using arthroscopy results as the golden standard, the sensitivity, specificity, and accuracy of MRI diagnosis were calculated. Kappa test was performed to test agreement between findings on conventional MRI and arthroscopy. Results On MRI, there were 25%showing partial rotator cuff tear, 13.3%Bankart's damage, 20% SLAP damage and 11.6% undetected subacromial impingement. The specificity of MRI for the diagnosis of traumatic lesions and the sensitivity of full-thickness rotator cuff tear (100%), bone injury (100%), subacromial impingement (88%) and Bankart's damage (87%) were high, but the sensitivity of SLAP injury (75%) and partial rotator cuff tear (75%) were low. The accuracy of MRI was highest in diagnosing full-thickness rotator cuff tear (100%) and bone injury (100%), followed by Bankart's damage (96%), accuracy was least in diagnosing SLAP lesion (89%). The results were completely consistent between MRI and arthroscopy in diagnosing full-thickness rotator cuff tear and bone injury with kappa score 1, and showed good consistency for Bankart's tear with kappa score 0.904. The kappa scores were relatively lower in diagnosis of partial rotator cuff tear, SLAP lesion and subacromial impingement, but all>0.75,the consistency was also good, respectively (P<0.05, respectively). Conclusion MRI is an effective imaging method for the diagnosis of shoulder joint injury, but the diagnostic accuracy of SLAP lesion and partial rotator cuff tear needs to be improved.