Up to now, vascular indicator-dilution curves have been analyzed by numerical integration or by fitting empirical functions to the data. Here, we apply a recently developed mechanistic model with the goal to quantitatively describe flow distribution in the sinusoidal network of normal rat livers and those with high-fat emulsion-induced NASH. Single-pass outflow concentration data of sucrose were obtained from in situ perfused rat livers after impulse injection. The model fitted to the data consists of a continuous mixture of inverse Gaussian densities assuming a normal distribution of regional flow. It accounts for the fractal flow heterogeneity in the organ and has three adjustable parameters with a clear physiological interpretation. The model fitted the data well and revealed that the intrahepatic flow dispersion of 49.6 % in the control group increased significantly to 87.2 % in the NASH group (p < 0.01). In contrast to previously used empirical functions, the present model exhibits a power-law tail (~t(-2.4)), which is a signature of fractal microvascular networks. The approach offers the possibility to determine hepatic blood flow heterogeneity in perfused livers and to evaluate the functional implications.