Comparison of MR-arthrography and MDCT-arthrography for detection of labral and articular cartilage hip pathology.

Abstract

To compare the diagnostic ability of MR arthrography (MRa) and MDCT arthrography (CTa) in depicting surgically proven hip labral tears and articular cartilage degradation. Labral pathology and articular cartilage were prospectively evaluated with MRa and CTa in 14 hips of 10 patients. The findings were evaluated by two independent observers (a musculoskeletal fellow and one senior musculoskeletal radiologist). Sensitivity, specificity, accuracy, and positive predictive value were determined using arthroscopic and open surgery findings as the standard of reference. Interobserver agreement was recorded. All images were assessed for the presence of a labral tear (according to Czerny classification) and for cartilage erosion using a 3 point scale for both methods: 1 = complete visualization-sharp edges, 2 = blurred edges fissuring-partial defects, 3 = exposed bone. The same classification was applied surgically. Disagreement between the senior observer and the fellow observer was recorded in three cases of labral tearing with MRa and six with CTa. Disagreement was also found in four cases of cartilage erosion with both MRa and CTa. The percent sensitivity, specificity, accuracy, and positive predictive value for correctly assessing the labral tear were as follows for MRa/CTa, respectively: 100/15, 50/13, 90/14, and 90/13 (P < 0.05). The same values for cartilage assessment were 63/66, 33/40, 50/57 and 55/66 (P > 0.05). Interobserver reproducibility with MRa is very good for labral tearing assessment. MRa is better for assessing labral tears. CTa shows better, but not statistically significant, demonstration of the articular cartilage.