MR diagnosis of rotator cuff tears of the shoulder: value of using T2-weighted fat-saturated images.

Abstract

The purpose of this study was to compare the use of fat-saturated T2-weighted spin-echo MR imaging with that of conventional T2-weighted spin-echo MR imaging to detect full- and partial-thickness tears of the rotator cuff of the shoulder, using arthroscopy as the standard for the diagnosis. Forty-nine consecutive patients with shoulder pain who had both MR imaging and arthroscopy of their affected shoulders were studied. The study group consisted of 10 patients with full-thickness tears of the rotator cuff, 20 with partial-thickness tears of the rotator cuff, and 19 with intact rotator cuff tendons proved arthroscopically. All patients had T1-weighted, conventional T2-weighted, and fat-saturated T2-weighted MR images obtained in a plane slightly oblique to the coronal plane along the long axis of the supraspinous tendon. The images were divided into two sets, one including both T1-weighted and conventional T2-weighted images, the other including T1-weighted and fat-saturated T2-weighted images. Two musculoskeletal radiologists reviewed each set of MR images in a blinded fashion. The reviewers were asked to decide whether the rotator cuff showed a full-thickness tear, a partial-thickness tear, or no tear on each set of images. In the case of partial-thickness tears, the reviewers were asked to indicate which surface (bursal or joint) of the rotator cuff was affected. These data were correlated with the arthroscopic findings and with each other using McNemar and kappa analysis. Detection of full-thickness tears was excellent for both reviewers using either imaging technique. However, when the fat-saturated technique was used, the sensitivity increased significantly, from 80% to 100%. Detection of partial-thickness tears was poor with conventional spin-echo MR imaging (15%). Although significantly improved when fat saturation was used (35%), detection rates were still lower than rates reported in the literature. Identification of the torn surface of the rotator cuff was correct in 50% of cases in which partial-thickness tears were successfully identified. Fat saturation decreased the specificity of identification of both partial and full-thickness tears of the rotator cuff but significantly increased the specificity with which intact rotator cuffs were identified. Use of the fat-saturation technique improved detection of both full-thickness and partial-thickness tears of the rotator cuff on MR images compared with standard spin-echo imaging techniques. Despite this improvement, detection of partial-thickness tears was poor with both techniques. Furthermore, correct identification of which surface of the rotator cuff was torn in patients with partial tears was nearly random.