Rotator cuff tendon tears: evaluation with fat-suppressed MR imaging with arthroscopic correlation in 100 patients.

Abstract

To evaluate the diagnostic accuracy of fat-suppressed magnetic resonance (MR) imaging of rotator cuff tears in a large symptomatic population. One hundred patients underwent both MR imaging and arthroscopy of the shoulder. Ninety-two patients underwent fat-suppressed conventional spin-echo MR imaging (repetition time msec/echo time msec = 2,500/20, 60), and eight patients underwent fat-suppressed, fast spin-echo MR imaging (2,000/80). With data combined for complete and partial tears of the rotator cuff (n = 31), MR imaging had an accuracy of 93%; sensitivity, 84%; and specificity, 97%. Seventeen of 20 complete tears and nine of 11 partial tears were properly identified with MR imaging. Two partial tears were not detected and three complete tears were incorrectly called partial tears at MR imaging. Of two false-positive MR imaging findings, one was called a complete tear and the other, a partial tear. Fat-suppressed MR imaging has high diagnostic accuracy in evaluating tears of the rotator cuff tendon.