Meeting Banner
Abstract #3719

Accelerated MRSI Using Randomly Undersampled Spiral-Based k-Space Trajectories

Itthi Chatnuntawech 1 , Borjan Gagoski 2 , Berkin Bilgic 3 , Stephen Cauley 3 , Ellen Grant 2 , Kawin Setsompop 3,4 , and Elfar Adalsteinsson 1,5

1 MIT, Cambridge, MA, United States, 2 Fetal-Neonatal Neuroimaging & Developmental Science Center, MA, United States, 3 A. A. Martinos Center for Biomedical Imaging, MA, United States, 4 Harvard Medical School, MA, United States, 5 Harvard-MIT Health Sciences and Technology, Institute of Medical Engineering & Science, MA, United States

In this work, an acquisition-reconstruction method (named random SENSE+TV) to accelerate MRSI acquisition through undersampling of spectral-spatial k-space data, while maintaining acceptable reconstruction quality as measured by a normalized root-mean-square error (RMSE) of the reconstructed metabolite maps is proposed. Random SENSE+TV uses spiral trajectories with different radii to randomly undersample the data in (kx, ky, kf) space, which leads to less-structured artifacts, and recovers the underlying data using the sensitivity encoding technique (SENSE) with a total variation (TV) regularization.

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

Join Here