MR evaluation of rotator cuff pathology using T2-weighted fast spin-echo technique with and without fat suppression.

Abstract

This study was designed to compare MR imaging findings from T2-weighted fast spin-echo images with and without fat suppression and then to compare surgical results with those MR imaging interpretations in the evaluation of rotator cuff disease. T2-weighted fast spin-echo images--with and without fat suppression--of 177 shoulder MR imaging studies were reviewed separately in a retrospective, randomized, and blinded fashion. Both sets of images were read in conjunction with the corresponding proton density-weighted images. The diagnosis of normal tendon, tendinosis, degeneration, and partial- and full-thickness tears was made using established criteria. T2-weighted fast spin-echo techniques with and without fat suppression showed excellent agreement in the diagnosis of normal tendon (kappa=.90) amd full-thickness tears (kappa=.98), good agreement for partial tears (kappa=.70) and moderate agreement for the combined group of tendinosis and degeneration (kappa=.53). MR imaging and surgical correlation in 43 patients showed 86% specificity (95% confidence interval, 65-96%) for intact tendons, and 100% sensitivity (95% confidence interval, 82-100%) for full-thickness tears on T2-weighted fast spin-echo imaging--both without and with fat suppression. For partial tears, MR imaging showed a sensitivity of 92% (95% confidence interval, 65-99%) with fat suppression and 67% (95% confidence interval, 39-86%) without fat suppression. T2-weighted fast spin-echo imaging--with or without fat suppression--is a highly sensitive technique in the diagnosis of normal tendons and complete tears of the rotator cuff. Because of increased lesion conspicuity, fat suppression tends to perform better in the diagnosis of partial tears.