Comparison between semiquantitative and quantitative methods for the assessment of knee synovitis in osteoarthritis using non-enhanced and gadolinium-enhanced MRI.

OBJECTIVE: To compare different semiquantitative and quantitative methods using both non-enhanced and gadolinium-enhanced MRI techniques for the assessment of synovitis in knee osteoarthritis (OA).

METHODS: Knees with end-stage clinical OA in patients undergoing total knee replacement surgery were included in this cross-sectional study. MRI was performed on all knees. Standard non-enhanced and gadolinium-enhanced sequences were acquired. Using non-enhanced MRI, we semiquantitatively assessed two features widely used as surrogates for synovitis: effusion-synovitis and Hoffa-synovitis. Using gadolinium-enhanced sequences, we semiquantitatively assessed synovial thickness. We quantitatively evaluated the total synovial volume on the gadolinium-enhanced sequences as well. We assessed the correlations of effusion-synovitis and Hoffa-synovitis with synovial thickness and volume, applying Spearman correlation analysis. The diagnostic performance of both synovitis features on non-enhanced MRI was assessed using synovial thickness on gadolinium-enhanced MRI as the reference.

RESULTS: A total of 104 subjects (one knee per subject) were included. Correlations of effusion-synovitis with synovial thickness and volume were r = 0.41 and r = 0.43 (P < .001) r = 0.32 and r = 0.39 (P < .0001).

CONCLUSION: Using synovial thickness assessed on gadolinium-enhanced sequences as the reference, effusion-synovitis showed superior correlations and sensitivity. Effusion-synovitis should be preferred over Hoffa-synovitis as a surrogate marker for synovial thickening, in studies of knee OA for which gadolinium-enhanced sequences are not available.