Qazi Saad Bin Khalid, Kamran Ellahi Memon, Samina Kausar Satti.
Diagnostic accuracy of MRI in characterization of rotator cuff tears.
Pak J Radiol Jan ;29(4):228-33.

Objectives: To assess the diagnostic accuracy of routine MRI in the detection and characterization of rotator cuff tears, by correlating the findings with arthroscopy. Study Setting and duration: Prospective study was conducted in Radiology Department Dallah Hospital Riyadh from July 2017 to December 2018. Methodology: A total of 104 patients were diagnosed with rotator cuff injury on MRI during study period, out of which 79 patients were selected for the study that underwent further evaluation with arthroscopy. The information was assessed for significant correlation between MRI analysis and arthroscopic findings utilizing kappa statistics. The sensitivity, specificity, predictive value and accuracy of MRI for the determination of full and partial thickness tears were determined utilizing arthroscopic findings as the reference standard. Results: Out of 79 patients the males were 52 and 27 females, the youngest patient included was 23 years and the oldest pateint was 76years. Commonly rotator cuff tears were seen in patients over 40 years old. The sensitivity of MRI was found to be 88.9% from the results of this study with 100% specificity with 100% positive predictive value and 84.1% negative predictive value. The sensitivity of MRI for partial thickness tear was 100% and the specificity was 87.4%. The positive predictive value was 81.4% with 100% negative predictive value. The accuracy for full and partial thickness tears noted as 94.3% and 90.7% respectively. The p-value <0.01, kappa value was 0.87 for full thickness tears while 0.82 for partial thickness tears, suggestive of excellent level of conformity between MRI and arthroscopy for the investigation of rotator cuff tears. Conclusions: To diagnose rotator cuff tears the MRI revealed high sensitivity and specificity as well as also provides useful information for all the important therapeutic and prognostic implications.

PakMediNet -Pakistan's largest Database of Pakistani Medical Journals - http://www.pakmedinet.com