The study of diffusion kurtosis imaging value for assessing liver fibrosis and comparison with ultrasound elastography

Shuangshuang Xie,Qing Li,Yue Cheng,Lulu Fan,Hanxiong Qi,Lihua Chen,Lixiang Huang,Jia Li,Li Zhou,Qian Ji,Wen Shen
DOI: https://doi.org/10.3760/cma.j.issn.1005-1201.2018.11.007
2018-01-01
Abstract:Objective To explore the efficacy difference of diffusion kurtosis imaging (DKI) and ultrasound elastography (UE) in the diagnosis of liver fibrosis. Methods Thirty-five patients whose serological examination showed hepatitis B or hepatitis C virus infection, disease course≥ 1 year, and finally liver biopsy confirmed pathological fibrosis grade in Tianjin Second People's Hospital from December 2015 to April 2017 were prospectively enrolled as patient group. During the same period, twenty healthy volunteers who matched the age, sex and BMI with patient group and showed normal liver function within the last month were enrolled as control group. All of the subjects underwent DKI experiment, and subjects in patient group underwent UE experiment in addition. Liver mean apparent diffusion (MD) and mean kurtosis (MK) were obtained in all subjects and liver stiffness measurement (LSM) was obtained in patient group. The patient group was staged for hepatic fibrosis based on liver biopsy results (S0 to S4). Differences in liver MD and MK values between control and patient groups were tested using independent sample t test (normal distribution) or Mann-Whitney U test (skewed distribution). Differences in liver MD, MK, and LSM between patients with different fibrosis stages were tested using One-way ANOVA (normal distribution) or Kruskal-Wallis test (skewed distribution). The correlation between liver MD, MK and LSM values with fibrosis stages were analyzed using Pearson correlation test. The diagnostic performance in staging fibrosis was analyzed using ROC analysis. Results Liver MD in patient group was lower than that in control group, and the difference was statistically significant (P<0.01). There was no significant difference in liver MK between the two groups (P>0.05). The AUC value for the diagnosis of liver fibrosis by MD was 0.950 (95%CI:0.855 to 0.990). Of the 35 patients, 15 were S1 (mild fibrosis), 13 were S2 (moderate fibrosis), 4 were S3, 3 were S4 (S3+S4 were severe fibrosis). The difference of MD and LSM between different stages of liver fibrosis was statistically significant (P<0.05), and there was no significant difference in MK (P>0.05). Liver fibrosis stages was highly correlated with MD (r=-0.757, P<0.01), and had no correlation with MK (r=-0.010, P=0.956), and moderately correlated with LSM (r=0.440, P<0.01). The AUC values of liver MD and LSM for characterization of ≥S2 stage liver fibrosis were 0.867 and 0.800, respectively, without statistically significant difference (P=0.486). The AUC values for characterization of≥S3 stage liver fibrosis were 0.918 and 0.653, respectively, with a statistically significant difference (P=0.032). Conclusion MD derived from DKI can be used for noninvasive assessment of liver fibrosis, and it is superior to LSM in distinguishing different fibrosis stages and detecting severe fibrosis.
What problem does this paper attempt to address?