Purpose

To evaluate the feasibility of intravoxel incoherent motion (IVIM) magnetic resonance imaging (MRI) for noninvasively staging liver fibrosis and monitoring its therapeutic response to losartan.

Methods

Forty-one rats were induced by carbon tetrachloride (CCl4) for 0, 2, 4, and 6 weeks. Among them, twenty-four rats were simultaneously treated with losartan for 4 and 6 weeks. IVIM diffusion weighted imaging was performed with 3.0T MRI unit by using thirteen b-values from 0 to 1000 sec/mm2. True diffusion coefficient (D), perfusion fraction (f), pseudodiffusion coefficient (D*), and apparent diffusion coefficient (ADC) at different time point were compared among individual fibrotic groups by analysis of variants and between anti-fibrotic and fibrotic group by student’s t test. The correlation between histological features (fibrosis, inflammation, steatosis, and hepatocellular ballooning) and each parameter was evaluated by Spearman’s ranked correlation, and diagnostic performance for staging liver fibrosis were determined by receiver operation curves (ROC) analysis.

Results

Although D*, f, ADC, and D were significant lower in fibrotic liver than normal liver, D*, f, and ADC showed moderate correlation with progression of liver fibrosis (ρ=-0.719, ρ=-0.698, ρ=-0.559, P<0.001) (Table 1). D* could be superior to detect the early and moderate liver fibrosis (AUC=0.950, 0.930, respectively). f might be beneficial to predict the presence of advanced liver fibrosis and cirrhosis (AUC=0.877)(Table 2). In losartan treated rats, D* and f showed statistical significance compared with CCl4 alone (P<0.05) (Figure 1 and 2).

Discussion

Perfusion related parameters (D* and f) in the rat model of liver fibrosis were decreased along with its progression. It is worth noting that D* was more sensitive than f in prediction of early and moderate stage of liver fibrosis. This difference could be attribute to some factors. First, in concept of IVIM theory, D* is related to the velocity of blood flow, and f represents the blood volume 1. Second, as previous study reported that hepatic artery was dilated to compensate the decreased blood volume in the early stage of liver fibrosis, known as arterial buffer effect 2. In other words, the hepatic blood volume was maintained until late stage of liver fibrosis, whereas blood flow may decrease in the early stage. In our study, f shows superior diagnostic performance than D* to predict advanced stage of liver fibrosis and cirrhosis. This could be explained by the fact that with the further progression of liver fibrosis, the increase in hepatic artery blood flow is insufficient to compensate the decreased blood volume, and eventually the entire hepatic blood volume was significant decreased 3. In our study, ADC and D show significantly lower value in normal liver than in moderate liver fibrosis (Stage 2) or advanced liver fibrosis and cirrhosis (Stage 3-4), but no significant differences in ADC and D among individual liver fibrosis, but D* shows significant decrease in early liver fibrosis in comparison with moderate liver fibrosis or advance liver fibrosis. Note that D reflects pure water molecular diffusion and its decrease during progression of liver fibrosis was affected by excessive collagen deposition, intracellular fat droplets, and hepatocellular ballooning which restricted the extracellular water diffusivity, whereas necrosis and apoptosis affect D measurement in the opposite direction and may result in the weak correlation between D and progression of liver fibrosis. Therefore, the reduction in microcirculation perfusion is likely more predominant pathophysiological change than restricted water diffusion in progression of liver fibrosis, suggesting potential of D* as a sensitive biomarker in staging liver fibrosis and monitoring anti-fibrotic response 4,5. Our results showed that perfusion related parameters (D* and f) could successfully allow to non-invasively identify the regression of liver fibrosis in rats model treated with losartan. After losartan treatment for both 4 and 6 weeks, D* and f showed statistical difference in comparison with rats treated with CCl4 alone for the corresponding time point. The alteration of microcirculation perfusion and water diffusion after losartan treatment reveal by IVIM DW imaging can be explained by histologic changes. The antifibrotic effect of losartan on liver fibrosis model of rat induced by CCl4 was pathologically confirmed, the METAVIR score was attenuated to 1.7±0.67 for week 4 rats and to 2.56±0.53 for week 6 rats, respectively. Similarly, losartan also reduced the inflammatory cells infiltration and the grade of steatosis and hepatocellular ballooning6. However, ADC does not show statistical significance in rats treated with and without losartan for 6 weeks, indicating that ADC is likely less sensitive than D* and f in detecting antifibrotic efficacy.

Conclsion

Perfusion related parameters derived from IVIM show potential to non-invasively stage liver fibrosis and assess antifibrotic response to losartan.

Acknowledgements

No acknowledgement found.