Objective: The objective of our study was to prospectively compare the diagnostic performance of MR arthrography and conventional MRI with surgical correlation in the same patient for detecting labrum and articular cartilage defects.
Subjects and methods: Twenty-eight patients (mean age, 31.8 years) underwent MR arthrography, conventional MRI, and subsequent hip surgery, which served as the reference standard. Labrum and cartilage defects were evaluated at MRI by two independent readers. A McNemar test and kappa statistics were used for statistical analysis.
Results: At surgery, 31 labral tears were identified. MR arthrography had an advantage over conventional MRI for detecting labral tears at the anterosuperior quadrant (sensitivity of MR arthrography, 81% and 69% for readers 1 and 2, respectively; sensitivity of conventional MRI, 50% for both readers); this difference in performance between MR arthrography and conventional MRI was statistically significant for reader 1 (p = 0.02) but not for reader 2 (p = 0.2). Interobserver agreement for labral tears was higher for MR arthrography (κ = 0.81) than for conventional MRI (κ = 0.63). Surgery showed 31 acetabular cartilage defects and nine femoral cartilage defects. MR arthrography had an advantage over conventional MRI for detecting acetabular cartilage defects (sensitivity of MR arthrography, 71% and 92% for readers 1 and 2, respectively; sensitivity of conventional MRI, 58% and 83%), whereas there was no advantage to using MR arthrography for detecting femoral cartilage defects with statistically significant difference for the acetabular cartilage or femoral cartilage. Interobserver agreement was slightly higher for MR arthrography (κ = 0.50) than for conventional MRI (κ = 0.40) for assessing the acetabular cartilage and was almost identical for the femoral cartilage (κ = 0.62 and 0.63, respectively).
Conclusion: MR arthrography was superior to conventional MRI for detecting labral tears and acetabular cartilage defects and showed a higher interobserver agreement. For femoral cartilage lesions, both modalities yielded comparable results.