To determine an optimal b-values sampling scheme for different non-Gaussian diffusion models in the liver, including diffusion kurtosis imaging (DKI), stretched-exponential model (SEM), intravoxel incoherent motion (IVIM), we optimized diffusion-weighting b-values sets using simulations mimicking real data and a b-value selection based on a Monte Carlo-like approach. Estimation performances were evaluated in terms of mean square error on signal and mean absolute percentage error on parameters in simulated data. The parameter estimation, with optimized set of b-values was finally applied on real data. The results showed that comparable fitting parameters and reconstructing signal can be obtained with fractional b-values.
This abstract and the presentation materials are available to members only; a login is required.