Diffusion kurtosis MRI versus conventional diffusion-weighted imaging for evaluating inflammatory activity in Crohn's disease

PurposeTo assess the efficacy of diffusion kurtosis imaging (DKI) and to compare DKI-derived parameters with that of conventional diffusion-weighted imaging (DWI) for grading the inflammatory activity of Crohn's disease (CD). Materials and MethodsIn all, 38 patients with CD underwent 3T magnetic resonance enterography (MRE) with DKI (b values of 0-2000 s/mm(2)). The inflammatory activity of the bowel segments was graded by magnetic resonance index of activity (MaRIA) as inactive (<7), mild (7 and <11), or moderate-severe (11). Apparent diffusion for non-Gaussian distribution (D-app) and apparent kurtosis coefficient (K-app) on DKI as well as apparent diffusion coefficient (ADC) on DWI were compared. ResultsIn all, 86 bowel segments including inactive (20), mild (19), and moderate-severe (47) CD were analyzed. The differences in K-app, D-app, and ADC among inactive, mild, and moderate-severe CD were significant (all P < 0.05). K-app (r = 0.862), D-app (r = -0.755), and ADC (r = -0.713) correlated well with MaRIA in all segments. Stronger correlation with MaRIA in moderate-severe CD was found for K-app (r = 0.647) than that of D-app (r = -0.414) and ADC (r = -0.580). Receiver operating characteristic (ROC) curve analysis showed high accuracy of K-app, D-app, and ADC for differentiating active from inactive CD (AUC: 0.953 for K-app, 0.944 for D-app, 0.907 for ADC) as well as differentiating inactive-mild from moderate-severe CD (AUC: 0.946 for K-app, 0.887 for D-app, 0.846 for ADC). The threshold K-app of 0.731 allowed differentiation of active from inactive CD with 89.4% sensitivity and 95% specificity. ConclusionDKI of CD is clinically feasible and might be superior to conventional DWI for grading the inflammatory activity of CD. Level of Evidence: 2 Technical Efficacy: Stage 2 J. Magn. Reson. Imaging 2018;47:702-709.