Joseph O deJesus MD, Presenter: Nothing to Disclose

Levon N. Nazarian MD, Abstract Co-Author: Nothing to Disclose

Laurence Parker PhD, Abstract Co-Author: Nothing to Disclose

Andrea J. Maitino MS, Abstract Co-Author: Nothing to Disclose

The clinical usefulness of diagnostic imaging for rotator cuff tears has been evaluated in a number of studies. We conducted a meta-analysis to compare the sensitivities and specificities.

Studies were identified from a MEDLINE search of manuscripts published between 1985 and 2003 that reported sensitivity and specificity of US, MR, or MR arthrography for suspected rotator cuff pathology with surgery as the gold standard. Data from the studies were used to calculate pooled sensitivities and specificities. Chi-square tests were used to determine whether differences in sensitivities and specificities were significant.

Twenty-eight articles met the search criteria. These articles reported a total of 68 evaluations of US and/or MR as follows: 25 ultrasound, 37 MRI, and 6 MR arthrography. Most studies fell within the CI justifying the pooling method. The data were analyzed excluding 10% of the outliers (4 US findings, 16 MR, and 2 MR arthrography). The summary sensitivities and specificities for all tears (full and partial thickness): US, 90.5% (95% CI: ±1.9%), 93.0% (±1.6%); MRI, 81.3% (±3.6%), 93.3% (±1.6%); and MR arthrography, 87.8% (±9.3%), 93.5% (±5.0%). The overall sensitivities and specificities for all modalities by type of tear: full, 93.4% (±2.1%), 94.6% (±1.5%) and partial, 73.2% (±6.0%), 94.3% (±1.7%). The sensitivity of MR was lower (p<0.01) than US. The sensitivity for diagnosing a partial tear was lower (p<0.01) than that for full thickness tears. The specificity of diagnosing a partial tear was also lower (p<0.01) than that for full thickness tears.

Based on published studies, overall, US is best at diagnosing rotator cuff tears. However, MR also demonstrates high sensitivities and specificities for full thickness tears. Further work needs to be done to determine whether MR diagnosis improves in specific clinical situations.

deJesus, J, Nazarian, L, Parker, L, Maitino, A, Diagnostic Imaging of Rotator Cuff Tears: A Meta-analysis of the Accuracy of US and MR.  Radiological Society of North America 2004 Scientific Assembly and Annual Meeting, November 28 - December 3, 2004 ,Chicago IL. http://archive.rsna.org/2004/4414582.html