Diagnostic performance of CT-arthrography and 1.5T MR-arthrography for the assessment of glenohumeral joint cartilage: a comparative study with arthroscopic correlation

To compare the diagnostic performance of multi-detector CT arthrography (CTA) and 1.5-T MR arthrography (MRA) in detecting hyaline cartilage lesions of the shoulder, with arthroscopic correlation. CTA and MRA prospectively obtained in 56 consecutive patients following the same arthrographic procedure were independently evaluated for glenohumeral cartilage lesions (modified Outerbridge grade a parts per thousand yen2 and grade 4) by two musculoskeletal radiologists. The cartilage surface was divided in 18 anatomical areas. Arthroscopy was taken as the reference standard. Diagnostic performance of CTA and MRA was compared using ROC analysis. Interobserver and intraobserver agreement was determined by kappa statistics. Sensitivity and specificity of CTA varied from 46.4 to 82.4 % and from 89.0 to 95.9 % respectively; sensitivity and specificity of MRA varied from 31.9 to 66.2 % and from 91.1 to 97.5 % respectively. Diagnostic performance of CTA was statistically significantly better than MRA for both readers (all p a parts per thousand currency signaEuro parts per thousand 0.04). Interobserver agreement for the evaluation of cartilage lesions was substantial with CTA (kappa = 0.63) and moderate with MRA (kappa = 0.54). Intraobserver agreement was almost perfect with both CTA (kappa = 0.94-0.95) and MRA (kappa = 0.83-0.87). The diagnostic performance of CTA and MRA for the detection of glenohumeral cartilage lesions is moderate, although statistically significantly better with CTA. CTA has moderate diagnostic performance for detecting glenohumeral cartilage substance loss. MRA has moderate diagnostic performance for detecting glenohumeral cartilage substance loss. CTA is more accurate than MRA for detecting cartilage substance loss.