Two methods for measuring diffusional kurtosis images were compared regarding their perfomance using an anisotropic synthetic fibre phantom. Our results can be of interest for researchers investigating water diffusion properties in biological tissues.Diffusion magnetic resonance imaging is widely recognized to be a unique, powerful and non-invasive technique for fibrous tissue microstructure investigation, particularly the human brain. Usually, water diffusion in biological tissues is quantified via diffusion tensor imaging (DTI), which uses a Gaussian approximation for the probability distribution of molecular displacements. However, the displacement probability distribution can considerably deviate from the Gaussian form for increasing diffusion weightings ($$$b$$$-values). The propagation of water molecules is impeded by interaction with tissue components, confinements, anisotropy, and is modulated by interfacial interactions with the cell membranes. Due to the heterogeneity and complexity of the tissue microstructure, the various contributions to the average MRI signal in in vivo studies cannot be easily separated from each other. Estimation of the next term in the cumulant expansion of the MRI signal, proportional to the so-called kurtosis, has lately become a popular method to quantify diffusion beyond the Gaussian information offered by DTI. Diffusional mean kurtosis (MK) was suggested to be a sensitive marker for tissue pathology, although, it requires longer acquisition times compared to DTI. A robust and rapid method for estimating MK, with a protocol adapted on commercial scanners, was proposed by Hansen et al.¹, and proved to be feasible in clinical use. Several recent studies demonstrated its reproducibility, comparable to the usually obtained MK metrics, in the rat and human brain, as biomarker for glioma subtypes². The primary purpose of this study was to perform a comparative analysis of the mean kurtosis metrics, obtained with a conventional and a fast protocol for a highly anisotropic medium with a well-known structure, such as a synthetic phantom and to investigate the effect of self-induced susceptibility gradients at different orientations on of the fibres in the static magnetic field.The construction of a fibre phantom with a fibre-density gradient region is described in Ref³. Imaging was performed on a whole-body 3T Siemens MAGNETOM Trio Scanner (Siemens, Medical Systems, Erlangen, Germany) using a 12 channel phased-array receive head coil. Diffusion-weighted images were measured with the fibre axis oriented at three angles with the static magnetic field: parallel (0°), perpendicular (90°) and at 45°. The double-refocused diffusion-weighted spin-echo EPI pulse sequence was used for acquiring the signal over 64 field gradient directions. For the fast protocol only 13 diffusion-weighted images (described in Ref.¹) were acquired for three $$$b$$$-values (0, 1000 and 2500 s/mm2). The bias in the diffusion-weighted images due to the background noise was corrected using the power-images method⁴. Conventional DKI parameter maps were computed with ExploreDTI⁵ for conventional kurtosis imaging and with the help of in-house Matlab scripts (Matlab, The MathWorks) for fast DKI. The number of averages for fast kurtosis was 20 in order to equal the number of diffusion-weighted volumes in conventional DKI experiments.Fig.1 shows a comparison of the conventional MK metric and the fast mean kurtosis parameter, $$$\bar{W}$$$, in two structurally different platforms of the phantom. The phantom was positioned in such a way that the fibers were aligned at 45 degrees to the main magnetic field. One can see a similar quality for the images in both methods with minor detectable differences. Fig.1(c) presents a quantitative comparison of the data in both platforms as scatter plots. In voxel by voxel comparison we clearly observe significant correlations between the two maps. However, the spread of points around the $$$MK = \bar{W}$$$ identity line is larger than that reported for brain¹, probably due to the larger anisotropic effects of the synthetic fibers. Fig.2 demonstrates the same quantitative and qualitative comparison of both DKI methods for two platforms in the phantom. Here, the phantom was positioned perpendicularly to the direction of a static magnetic field. According to a strong dependence of absorption ability of fibers on the orientation to magnetic field, we see a lower quality of the maps computed by fast kurtosis method in this orientation.The close correspondence of new measure $$$\bar{W}$$$ of kurtosis with the conventional MK was investigated for an anisotropic synthetic phantom. Fast estimation of diffusion kurtosis metric has the potential to constitute a useful tool for kurtosis imaging and become a beneficial microstructural biomarker. We also observed a strong dependence of the image quality on the orientation of fibers with respect to the magnetic field, that is, the fast method is sensitive to the susceptibility differences arising at interfaces.