Improved Subspace Estimation for Low-Rank Model-Based Accelerated Cardiac Imaging

Abstract

Sparse sampling methods have emerged as effective tools to accelerate cardiac magnetic resonance imaging (MRI). Low-rank model-based cardiac imaging uses a pre-determined temporal subspace for image reconstruction from highly under-sampled (k, t)-space data and has been demonstrated effective for high-speed cardiac MRI. The accuracy of the temporal subspace is a key factor in these methods, yet little work has been published on data acquisition strategies to improve subspace estimation. This paper investigates the use of non-Cartesian k-space trajectories to replace the Cartesian trajectories which are omnipresent but are highly sensitive to readout direction. We also propose “self-navigated” pulse sequences which collect both navigator data (for determining the temporal subspace) and imaging data after every RF pulse, allowing for even greater acceleration. We investigate subspace estimation strategies through analysis of phantom images and demonstrate in vivo cardiac imaging in rats and mice without the use of ECG or respiratory gating. The proposed methods achieved 3-D imaging of wall motion, first-pass myocardial perfusion, and late gadolinium enhancement in rats at 74 frames per second (fps), as well as 2-D imaging of wall motion in mice at 97 fps.

High‐frame‐rate full‐vocal‐tract 3D dynamic speech imaging

“…We accelerate the acquisition with a “self‐navigation” strategy. Unlike previous approaches that collect navigator data with a separate radio frequency excitation , this self‐navigation strategy combines the acquisition of both navigator and imaging data into one single repetition time (TR) using a multi‐echo readout . This combined acquisition of both datasets is particularly desirable for speech imaging applications not only because it effectively increases the imaging speed by shortening TR, but also because it avoids missing temporal components that associate with important articulatory dynamics.…”

Self-navigated low-rank MRI for MPIO-labeled immune cell imaging of the heart

“…The slice rephase and read dephase pulses are the same after every RF pulse, enabling collection of suitable data for ${\mathcal{D}}_1$. In our implementation, we additionally replace the read dephase pulse with a “music note” (♪) trajectory [9], which has the same integral as the typical read dephase pulse (i.e., it ends in the same k -space location). The 1-D trajectory of the typical read dephase pulse has a null space such that it cannot detect perpendicular translation [9], so it is preferable to use a 2-D navigator such as the music note trajectory or a spiral trajectory.…”

“…In our implementation, we additionally replace the read dephase pulse with a “music note” (♪) trajectory [9], which has the same integral as the typical read dephase pulse (i.e., it ends in the same k -space location). The 1-D trajectory of the typical read dephase pulse has a null space such that it cannot detect perpendicular translation [9], so it is preferable to use a 2-D navigator such as the music note trajectory or a spiral trajectory. We have designed and implemented the music note to: 1) traverse a high-SNR region of k -space; and 2) be less demanding of gradient hardware than spiral trajectories.…”

Accelerated High-Dimensional MR Imaging With Sparse Sampling Using Low-Rank Tensors

“… 6 The data acquisition scheme for subsets of navigator data can be implemented utilizing a variety of strategies which have been proposed for the Partially Separable (PS) model (i.e., the low-rank tensor image model with d̂ = 2) [20], [34], [46], [47]. …”