Meeting Banner
Abstract #2066

Optimization of Scan Parameters for Diffusion Kurtosis Imaging at 1.5 T

Suguru Yokosawa1, 2, Yoshitaka Bito, Yoshihisa Soutome, Kenji Ito2, Fumio Yamashita2, Kohsuke Kudo2, Makoto Sasaki2

1 Central Research Laboratory, Hitachi, Ltd., Kokubungi-shi, Tokyo, Japan; 2Division of Ultrahigh Field MRI, Institute for Biomedical Science, Iwate Medical University, Yahaba, Iwate, Japan

We investigated the effects of the b-values, number of diffusion directions, and number of signals averaged (NSA) on the accuracy of diffusion kurtosis imaging (DKI) and attempted to determine the ideal combinations of these parameters to achieve both accuracy and time efficiency at 1.5 T. We demonstrated that a b-value of 2500 s/mm2 and the NSA can improve the accuracies of DKI maps and that varying the numbers of b-values and diffusion directions was not effective for this purpose. Accurate DKI maps can be obtained with an acceptable acquisition time even at 1.5 T when the scan parameters are optimized.

Keywords

acceptable accuracies accuracy accurate achieve acquisition additional although analyses appear approximately arrangement assumed assumes attempted axial behavior biological biomedical calculations caused central channel characterize clinical coefficient coil combination combinations compartments compromised configurations contribute corporation correlation crucial dataset datasets defined degree determine developed diffusion dimensional displacement distribution division dramatically duration echelon efficiency equivalent estimating evaluated even excellent fewer field five former fourteen full generate generated gradient gradients head healthy hence house ideal identical improvement improves increasing influence institute interval intraclass investigated japan kurtosis laboratory latter least linear long maps matrix medical membranes metrics molecules motion needed neighboring noise often optimization optimizations optimized optimizing particularly planner post probability probing processing qualities quantitative radial radiology reduced reducing regardless relationships remains repeated reported require requirement restriction revealed scanners science separation setting shorten similarities slices software spatial spin squares structures substantially suggest suitable system tensor terms tissue twice ultrahigh unconstrained unknown volunteers water whereas wolfram yield