Merlin J Fair1,2, Peter D Gatehouse1,2, Liyong Chen3,4, Ricardo Wage2, Edward VR DiBella5, and David N Firmin1,2
1NHLI, Imperial College London, London, United Kingdom, 2NIHR Cardiovascular BRU, Royal Brompton Hospital, London, United Kingdom, 3UC Berkeley, Berkeley, CA, United States, 4Advanced MRI Technologies, Sebastopol, CA, United States, 5UCAIR, University of Utah, Salt Lake City, UT, United States
Synopsis
An alternate-cycle acquisition strategy is proposed for 3D whole-heart first-pass perfusion, capturing two separate datasets from the same first-pass, each pushing separate boundaries of currently achievable parameters whilst maintaining clinically feasible acquisition times. It is postulated this approach may also confer an advantage with regard to artefact detection.Introduction
Simultaneously
optimising parameters such as left ventricular coverage, image resolution and
contrast sensitivity is difficult in first-pass perfusion (FPP). 3D FPP shows
potential to improve coverage1, but “whole-heart” coverage demands
high acceleration forcing compromises such as loss of spatial resolution. 2D
FPP shows high diagnostic ability with more slices distributed over
alternate-RR cycles2, relaxing acceleration requirements.
This work
proposes that 3D FPP could interleave two strategies with different 3D
parameters in alternate cycles, in sum approaching the full set of desired FPP
properties. This work also aims to improve specificity against artefacts by
imaging the same myocardium with two different interleaved 3D scans (distinct
from reformatting a single 3D FPP scan). Similar confirmation strategies are
often used in cardiovascular magnetic resonance, such as repetition with
swapped phase-encode direction in late-enhancement imaging.
Methods
Spoiled
gradient-echo 3D radial ‘stack-of-stars’ imaging3 was developed for
independent FOV, resolution and position on alternate cardiac cycles. Imaging
was performed on a Siemens (Erlangen, Germany) Skyra 3 Tesla system, with an 18
channel anterior chest array and 12 channel posterior array.
After
extensive volunteer optimisation, rest perfusion was acquired in 10 patients
with short-axis (SAX) images of higher in-plane resolution acquired on odd
cardiac cycles and lower, but more isotropic, resolution long-axis (LAX) images
on even cycles (Figure 1). The sequence was highly optimised for speed, e.g.
through use of high undersampling, custom-tailored RF pulses and asymmetric
readouts (75%).
The SAX cycles
collected 98 rays with a TR of 2.0ms, enabling an acquisition time of 196ms – comparable
favourably with previous 3D FPP literature. With the LAX cycles acquiring kz
across one of the short-axes of the left ventricle, during systole, the FOV required in this
direction was minimised and therefore the number of kz partitions required for
higher through-plane resolution was lowered. The lower in-plane resolution in
these acquisitions helped reduce TR to 1.8ms which, with the 112-124 rays
collected depending on cardiac size, gave still acceptable acquisition times
of 205-228ms.
After additional
zero padding in-plane (for SAX) and through-plane (for LAX), SAX cycles obtained
6-8 wrap-free slices reconstructed as 1.1 x 1.1 x 10.0mm voxels and the LAX
cycles obtained 20-24 usable reconstructed slices at 3.1 x 3.1 x 3.1mm. A total
variation temporally constrained algorithm was used for reconstruction4
with temporal weighting = 0.7 and 50 iterations, performed in MATLAB (Mathworks,
Natick, USA).
Results & Discussion
All
patients were successfully imaged by this method at higher in-plane and
through-plane resolutions than previously achieved by 3D FPP for the SAX and
LAX acquisitions respectively (e.g. Figure 2), in realistic acquisition times.
The SAX
shots produced higher quality images, similar to 2D FPP but with contiguous
coverage and showing all slices at the same cardiac phase. An example dataset
for a patient with dilated cardiomyopathy (Figure 3) shows potential importance
for the higher resolution in distinguishing the myocardium.
The LAX images give the reader a second
viewpoint with altered motion and resolution characteristics aiding artefact
judgement, although they could also be reformatted into any plane (Figure 4).
Image quality in the LAX acquisitions was reduced, but the improved through-plane resolution moved towards 3D imaging reconstructed
with isotropic resolution, potentially adding information to the more
conventional SAX acquisition.
Conclusions
The feasibility of alternate-RR separate 3D acquisitions during a single
FPP is presented, capable of acquiring datasets that stretch current limits of
both in-plane and through-plane resolutions while delivering two independently
optimised acquisitions of the same first-pass for potentially improving
clinical confidence.
Acknowledgements
This work was supported by the NIHR Cardiovascular Biomedical Research Unit of Royal Brompton and Harefield NHS Foundation Trust and Imperial College London, UK.
MJF is funded by a British Heart Foundation (BHF) PhD Studentship Grant - FS/13/21/30143.
References
1. Fair MJ, Gatehouse PD, DiBella EVR, Firmin DN. A review of 3D first-pass, whole-heart, myocardial perfusion cardiovascular magnetic resonance, J Cardiovasc Magn Reson. 2015;17(1):68.
2. Bertschinger KM, Nanz D, Buechi M, et al. J Magn Reson Imaging. Magnetic resonance myocardial first-pass perfusion imaging: Parameter optimization for signal response and cardiac coverage. J Magn Reson Imaging. 2001;14(5):556-562.
3. Chen L, Adluru G, Schabel MC, et al. Myocardial perfusion MRI with an undersampled 3D stack-of-stars sequence. Med Phys. 2012;39(8):5204-5211.
4. Adluru G, Awate SP, Tasdizen T, et al. Temporally constrained reconstruction of dynamic cardiac perfusion MRI. Magn Reson Med. 2007;57(6):1027-1036.