Ultrasonography of the rotator cuff: a comparison with arthroscopy in one-hundred-and-ninety consecutive cases

PURPOSE: To evaluate the accuracy of high-resolution ultrasonography compared to arthroscopy in the detection of rotator cuff tears. MATERIAL AND METHODS: Preoperative ultrasonography (US) with a 10-MHz commercially available linear-array transducer and a standardized study protocol was performed in 190 consecutive shoulders in 185 patients with a history of shoulder pain for more than 3 months. The findings at US were classified into intact cuff, partial-thickness, and full-thickness rotator cuff tears, and correlated with findings at shoulder arthroscopy. RESULTS: US correctly depicted 118 of 124 rotator cuff tears with sensitivity 95%, specificity 94%, PPV 97%, NPV 91% and accuracy 95%, all 94 full-thickness tears (sensitivity 100%, specificity 91%, PPV 91%, NPV 100%, and accuracy 95%), 24 of 30 partial-thickness tears (sensitivity 80%, specificity 98%, PPV 86%, NPV 96%, and accuracy 95%). CONCLUSION: US is a highly accurate diagnostic method for detecting full-thickness rotator cuff tears, but is less sensitive in detecting partial-thickness rotator cuff tears.