Posteroinferior Labral Cleft at Direct CT Arthrography of the Shoulder by Using Multidetector CT: Is This a Normal Variant?

Abstract

To describe a posterior labral cleft at direct computed tomographic (CT) arthrography of the shoulder by using multidetector CT and to compare this finding with a true posterior labral tear. Institutional ethics review board approval was obtained, and informed consent was waived. One hundred twenty-seven shoulders in 126 patients were examined with direct CT arthrography by using 16- or 64-section multidetector CT and arthroscopy. Two musculoskeletal radiologists retrospectively reviewed CT arthrographic images for the presence, location, and size of a posterior labral tear, defined as a detectable contrast material-filled focal discontinuity of the labrum on an axial image, proved by using arthroscopy. A posterior labral cleft was defined as a false-positive lesion at CT arthrography that was proved to be a normal finding arthroscopically. Sensitivity, specificity, accuracy, positive and negative predictive values of tears and clefts were determined; incidence according to the patient's age and sex and the laterality (right or left shoulder), location, and size of the lesion were compared. In 127 shoulders, radiologists 1 and 2 found 12 and 11 posterior labral tears, respectively, seen exclusively in male patients with posterior instability. Radiologist 1 observed 24 (18.9%) clefts, and radiologist 2 observed 20 (15.7%) clefts, seen more commonly in female patients (P = .037 for radiologist 1, P = .026 for radiologist 2) and in the inferior quadrant of the posterior labrum (along 7- to 8-o'clock positions, P < .05 for both radiologists); these clefts were shallower than labral tears (P = .005 for radiologist 1, P = .025 for radiologist 2). At direct CT arthrography, a labral cleft may be a normal variation of the posterior labrum.