Meeting Banner
Abstract #3333

Accelerating Diffusion Kurtosis Imaging Using Model Based Denoising

Jonathan I Sperl1, Tim Sprenger1,2, Ek T Tan3, Marion I Menzel1, Christopher J Hardy3, and Luca Marinelli3

1GE Global Research, Garching, Germany, 2Technische Universit√§t M√ľnchen, Munich, Germany, 3GE Global Research, Niskayuna, NY, United States

Diffusion Kurtosis Imaging (DKI) suffers from high sensitivity to noise and therefore requires long scanning times (up to 150 diffusion weighted images, DWIs). This work proposes a model-based denoising technique to overcome this limitation: A generalized multi-shell spherical deconvolution model is formulated and DWIs are denoised by a projection into the space spanned the model. We demonstrate noise reduction for DKI metrics yielding improved image quality of kurtosis maps from as few as 30 DWIs. This corresponds to greater than four-fold reduction in scan time as compared to the widely used 140-DWI acquisitions.

This abstract and the presentation materials are available to members only; a login is required.

Join Here