Intravoxel incoherent motion MRI for liver fibrosis assessment: a pilot study

Background: There has been a growing need for an imaging method for the accurate diagnosis and staging of liver fibrosis as a non-invasive alternative to liver biopsy. Purpose: To evaluate the feasibility of intra-voxel incoherent motion (IVIM) imaging for classifying the severity of liver fibrosis. Material and Methods: Fifty-seven patients who underwent navigator-triggered, diffusion-weighted imaging (DWI) of the liver on a 1.5-T system using nine b-values and had a reliable reference standard for the diagnosis of liver fibrosis (histopathologic findings [n = 27] or clinical findings for normal [n = 18] or cirrhotic liver [n = 12]), were included in our study. Liver apparent diffusion coefficient (ADC), pure diffusion (D-slow), perfusion fraction (f), and perfusion-related diffusion (D-fast), and the product f . D-fast were compared with the liver fibrosis stages (F). The accuracies of these parameters in diagnosing severe liver fibrosis (F >= 3) were evaluated using the receiver-operating characteristic (ROC) curve analysis. Results: The liver fibrosis stages had the strongest negative correlation with f . D-fast (rho = -0.52). All of the parameters, except for D-slow, were significantly lower in patients with F >= 3 than in those with F <= 2 (P <= 0.001). The area under the ROC curve for diagnosing severe fibrosis was the largest for f . D-fast (0.844) with an overall accuracy of 79.0% (45/57) at the optimal cutoff value and followed by f (0.834), D-fast (0.773), ADC (0.762), and D-slow (0.656). Conclusion: IVIM imaging is a promising method for classifying the severity of liver fibrosis, with the product f . D-fast being the most accurate parameter.