Free-breathing multiphase whole-heart coronary MR angiography using image-based navigators and three-dimensional cones imaging

Abstract

Noninvasive visualization of the coronary arteries in vivo is one of the most important goals in cardiovascular imaging. Compared to other paradigms for coronary MR angiography, a free-breathing three-dimensional whole-heart iso-resolution approach simplifies prescription effort, requires less patient cooperation, reduces overall exam time, and supports retrospective reformats at arbitrary planes. However, this approach requires a long continuous acquisition and must account for respiratory and cardiac motion throughout the scan. In this work, a new free-breathing coronary MR angiography technique that reduces scan time and improves robustness to motion is developed. Data acquisition is accomplished using a three-dimensional cones non-Cartesian trajectory, which can reduce the number of readouts 3-fold or more compared to conventional three-dimensional Cartesian encoding and provides greater robustness to motion/flow effects. To further enhance robustness to motion, two-dimensional navigator images are acquired to directly track respiration-induced displacement of the heart and enable retrospective compensation of all acquired data (none discarded) for image reconstruction. In addition, multiple cardiac phases are imaged to support retrospective selection of the best phase(s) for visualizing each coronary segment. Experimental results demonstrate that whole-heart coronary angiograms can be obtained rapidly and robustly with this proposed technique.

Copyright © 2012 Wiley Periodicals, Inc.

Figures

FIG. 1

3D cones coronary MRA sequence. (a) 3D cones k-space trajectory. (b) ATR-SSFP imaging sequence. (c) Triggered pulse sequence with multi-phase acquisition. TD: cardiac trigger delay, D: preparation cycles for iNAV, iNAV: 2D navigator image, F: fat saturation, C: catalyzation cycles for IMG, IMG: 3D cones image acquisition for one cardiac phase.

FIG. 2

Motion tracking for subject B. (a) Reference leading/trailing sagittal iNAVs (rhb0 = 2) with the tracking ROIs. (b) Tracking results in S/I and A/P with rhb0 = 2. (c) Calibrated tracking results with rhb = 202. Solid line: leading iNAV, dotted line: trailing iNAV.

FIG. 3

Motion tracking for subject D. (a) Reference leading sagittal and trailing coronal iNAVs with the tracking ROIs. (b) Calibrated tracking results in S/I (leading and trailing), A/P (leading), and L/R (trailing). Solid line: leading iNAV, dotted line: trailing iNAV.

FIG. 4

Motion compensation for subject A. Thin-slab MIP reformats at the three acquired cardiac phases with no respiratory motion compensation (left), S/I and A/P compensation using only the leading iNAV (middle), and S/I and A/P compensation using both leading/trailing iNAVs (right). Compensation using the leading iNAV already improves the depiction of the RCA for all phases. By incorporating the trailing iNAV, the RCA becomes sharper in phase 3/3 and also phase 2/3. Arrows point to areas of improvement.

FIG. 5

Motion compensation for subject B. Thin-slab MIP reformats from cardiac phase 2/3 before (left) and after (right) S/I and A/P respiratory motion correction using leading/trailing iNAVs. Arrows point out structures that have sharpened.

FIG. 6

Motion compensation for subject D. Thin-slab MIP reformats from cardiac phase 2/3 with no respiratory compensation (top), S/I and A/P correction using the leading iNAVs (middle), and full 3D correction with leading/trailing iNAVs (bottom). Note the progressive improvement in image quality.

FIG. 7

Multi-phase reconstruction for subject B. Thin-slab MIP reformats from three cardiac phases. The RCA is best visualized in phase 1/3 while the LM, LAD and LCx are best visualized in phase 2/3.