The value of many non-gaussian distribution models diffusion weighted imaging in evaluating the model of hepatic ischemia reperfusion injury model in rats

Abstract

Objective To explore the value of multiple non-gaussian distribution models DWI in evaluating hepatic ischemia reperfusion injury (HIRI) model in rats. Methods Forty-two SD rats were divided into 7 groups by random numeric table method. Each group had 6 rats. The 7 groups were 6 h, 12 h, 1 d, 3 d, 7 d, 14 d group after HIRI and control group, respectively. The experimental groups underwent right hepatic portal vein and hepatic artery ligation, and received reperfusion 3 h after operation. MRI scanning (T1WI, T2WI, single b-values DWI and 15 b-values DWI) was performed at 6 h, 12 h, 1 d, 3 d, 7 d and 14 d after reperfusion. The control group underwent sham operation and MRI scanning. According to monoexponential model, biexponential model, threxponential model, stretched-exponential model DWI and diffusion kurtosis imaging, many parametres were obtained and their dynamic changes at each time point were observed. The parameters included standard apparent diffusion coefficient (ADCs), pure diffusion coefficients (D), pseudodiffusion coefficients (D*), perfusion fraction (f), ultra-high apparent diffusion coefficient (ADCu), distributed diffusion coefficient (DDC), water diffusion heterogeneity index (α), mean diffusion coefficient (MD) and mean diffusion kurtosis (MK). One way ANOVA was used to compare the differences of parameters among different groups. Results MRI examination and pathological examination were successfully completed in all rats. The right hepatic lobe in the experimental groups appeared hypointense on T1WI and hyperintense on T2WI relative to control group from 6 h after operation. The infarcted liver lobe in the experimental groups became significantly smaller at 1 week after HIRI and almost disappeared at 2 weeks after HIRI. The findings of DWI at different b-values were consistent with those at T2WI. There were significant differences in parameters among 6 h, 12 h, 1 d, 3 d, 7 d, 14 d groups after HIRI and control group (P<0.05). There were significant differences among 12 h after HIRI, 1 d after HIRI, 3 d after HIRI and 6 h after HIRI, 7 d after HIRI, 14 d after HIRI, control group concerning ADCs values respectively (P<0.05). There were significant differences among 6 h after HIRI, 12 h after HIRI, l d after HIRI, 3 d after HIRI, 7 d after HIRI and 14 d after HIRI, control group concerning D, D*, f, ADCu, α, DDC, MK values respectively (P<0.05). In addition, there were significant differences among 1 d after HIRI, 3 d after HIRI and 6 h after HIRI, 12 h after HIRI, 7 d after HIRI and 14 d after HIRI, control group concerning D values respectively (P<0.05). There were significant differences among 6 h after HIRI, 12 h after HIRI, 1 d after HIRI, 3 d after HIRI and 7 d after HIRI, 14 d after HIRI, control group concerning f values respectively (P<0.05). There were significant differences among 12 h after HIRI and 14 d after HIRI, control group concerning MD values respectively (P<0.05). There were significant differences between 1 d after HIRI and 14 d after HIRI concerning MD values (P<0.05). Conclusion Multiple non-gaussian distribution models DWI is superior to conventional DWI in evaluating HIRI model in rats. Key words: Liver; Ischemia reperfusion injury; Non-gaussian distribution model; Multi b-values; Magnetic resonance imaging