Magnetic resonance arthrography including ABER view in diagnosing partial-thickness tears of the rotator cuff: accuracy, and inter- and intra-observer agreements.

BACKGROUND

Partial-thickness tear of the rotator cuff is a common cause of shoulder pain. Magnetic resonance (MR) arthrography has been described as a useful measure to diagnose rotator cuff abnormalities.

PURPOSE

To determine the reliability and accuracy of MR arthrography with abduction and external rotation (ABER) view for the diagnosis of partial-thickness tears of the rotator cuff.

MATERIAL AND METHODS

Among patients who underwent MR arthrographies, 22 patients (12 men, 10 women; mean age 45 years) who had either partial-thickness tear or normal tendon on arthroscopy were included. MR images were independently scored by two observers for partial-thickness tears of the rotator cuff. Interobserver and intraobserver agreements for detection of partial-thickness tears of the rotator cuff were calculated by using kappa coefficients. The differences in areas under the receiver operating characteristic (ROC) curves were assessed with a univariate Z-score test. Differences in sensitivity and specificity for interpretations based on different imaging series were tested for significance using the McNemar statistic.

RESULTS

Sensitivity, specificity, and accuracy of each reader on MR imaging without ABER view were 83%, 90%, and 86%, and 83%, 80%, and 82%, respectively, whereas on overall interpretation including ABER view, the sensitivity, specificity, and accuracy of each reader were 92%, 70%, and 82%, and 92%, 80%, and 86%, respectively. Including ABER view, interobserver agreement for partial-thickness tear increased from kappa=0.55 to kappa=0.68. Likewise, intraobserver agreements increased from kappa=0.79 and 0.53 to kappa=0.81 and 0.70 for each reader, respectively. The areas under the ROC curves for each reader were 0.96 and 0.90, which were not significantly different.

CONCLUSION

Including ABER view in routine sequences of MR arthrography increases the sensitivity, and inter- and intraobserver agreements for detecting partial-thickness tear of rotator cuff tendon.