Diffusion kurtosis imaging with the breath-hold technique for staging hepatic fibrosis: A preliminary study

Abstract Purpose To evaluate the potential of diffusion kurtosis imaging (DKI) analysis with the breath-hold technique to assess the stage or classify hepatic fibrosis. Materials and methods Patients (n = 67) suspected of having a disease of the hepatobiliary system examined by diffusion-weighted imaging (DWI) using a 3.0-T magnetic resonance imaging unit were enrolled in this study. To evaluate hepatic fibrosis, mean kurtosis, Mean apparent diffusion (MD) and apparent diffusion coefficient (ADC) values were compared between groups with varying fibrosis; F0–F1, F2–F3, and F4. The Steel–Dwass test was used for overall comparisons. Correlations between the fibrosis stage and mean kurtosis, MD or ADC values were assessed using Spearman's rank correlation. Discriminative capacities of DKI were evaluated using receiver operating characteristic (ROC) analysis. Results There were significant differences in ADC, MD and mean kurtosis values between non-cirrhosis and cirrhosis groups. Moreover, the mean kurtosis value was statistically different between the F0–F1 and F2–F3, F0–F1 and F4, and F2–F3 and F4 groups (all P P P  > 0.05). In addition, mean kurtosis and MD values significantly correlated with the extent of hepatic fibrosis staging (Spearman's rank correlation coefficient, ρ = 0.851 and − 0.672; P P  = 0.078). According to the ROC analysis for the assessment of no fibrosis (F0), fibrosis (≥ F1), and advanced fibrosis (≥ F2) and liver cirrhosis, the DKI cut-off values were 0.923, 0.955, and 1.11, respectively. Conclusion Using the DKI method with the breath-hold technique in the liver, the stage of hepatic fibrosis can be classified into normal and early hepatic fibrosis, substantial stages, and advanced hepatic fibrosis.