Influence of Maximal B-value, Fit Polynomial and Number of Diffusion Directions on the Measured Kurtosis: a Phantom Study

Introduction In contrast to free diffusion, the restrictions present in biological tissue cause the diffusion propagator to deviate from a Gaussian function. Diffusion kurtosis imaging (DKI) [1,2] extends Diffusion Tensor Imaging (DTI) in order to quantify these deviations from Gaussian behaviour yielding additional information about the tissue microstructure. To calculate the kurtosis, diffusion weighted signals with a range of b-values must be acquired. The kurtosis can then be determined by a polynomial fit of the logarithmic signal. Unfortunately, it is not trivial to perform this fit correctly and stable. Therefore, the aim of this work was to investigate the dependence of the measured kurtosis on the acquisition parameters maximum b-value, order of the fitting polynomial and number of gradient directions under well-defined conditions using recently developed diffusion phantoms [3].