Yichen Zheng1, Aiqi Sun2, Shuo Chen1, Xiaole Wang1, Chun Yuan1,3, and Rui Li1
1Center for Biomedical Imaging Research, Department of Biomedical Engineering, School of Medicine, Tsinghua University, Beijing, China, 2Neusoft Medical System, Shanghai, China, 3Vascular Imaging Laboratory, Department of Radiology, University of Washington, Seattle, WA, United States
Synopsis
Real-time 4D flow MRI, without ECG gating
and respiration control, has been developed as an effective tool to evaluate
hemodynamics. With the benefits of low rank and partial separable model, it could
be reconstructed with arbitrary acquisition duration. In this study, we investigated
the relationship between acquisition duration and image quality of real-time 4D
Flow MRI, and proposed optimized acquisition duration considering both image
quality and acquisition efficiency.
Introduction
Conventional 4D flow MRI requires
electro-cardiogram (ECG) gating and respiration control to acquire averaged
velocity maps from multiple cardiac cycles, which needs regular cardiac beat
and cannot capture the beat-by-beat flow variation1. Moreover, the
acquisition time may be prolonged because of the data rejection of irregular
cardiac beats. We have developed non ECG gating and non-respiration control real-time
4D flow MRI with certain amount of data to solve these problems. However, the
amount of data or acquisition duration, required by real-time 4D flow MRI, is
not investigated. In this study, we evaluated the image quality with different data
size for real-time 4D flow MRI and tried to propose optimized acquisition
duration.Methods
MRI
Scan: The experiment was on a 3.0 T whole-body
MR scanner (Achieva, Philips Medical System, Best, The Netherlands). The
scanning region involved all the aorta including ascending aorta, aortic arch
and descending aorta. A 32-channel cardiovascular coil was used for
acquisition. The scan
parameters were: FOV = 180mm×256mm×43mm
(FH/RL/AP), spatial resolution = 2.40mm×2.40mm×2.40mm, TR/TE = 4.4ms/2.6ms, flip angle = 5°,
encoding velocity (VENC) = 200/150/150cm/s (FH/RL/AP), temporal resolution = 8×TR = 35.2ms and total acquisition time was around 17min. The
acquisition trajectory was the same as reference2.
Neither ECG gating nor respiration control was used during the data
acquisition. Reconstruction: To
compare the image quality between different acquisition duration, different amount
of data were used for reconstruction based on the low-rank modeling2.
Comparison: The
signal-to-noise (SNR) and the velocity-to-noise (VNR)3 were calculated
for each reconstruction result. To be specific, we computed the complex
difference between 3 directions (AP, FH and RL) and reference, and obtained the
sum of square results (magnitude map) so that a MR Angiography (MRA) image for
aorta could be achieved. SNR was calculated from these MRA images. VNR was
calculated from the velocity map for three directions respectively. Because of the
enormous difference, we took logarithm to VNR.Results
Reconstructed magnitude maps (in systole) at different
acquisition duration were shown in Figure 1. Figure 2 showed the SNR
curve with the increase of acquisition duration. When the acquisition duration was short, for example, T = 1.1min, the SNR was very low, and we can’t even
distinguish the boundary of aorta. With the increase of acquisition duration,
SNR increased rapidly. However, when the acquisition duration was long enough,
for example, T > 13.6min, the SNR was converged without signification increase.
Figure 3 showed the velocity maps (in systole) at different acquisition
duration, and the VNR curves for three directions were shown in Figure 4.
Similar to the magnitude maps and SNR curve, with the growth of acquisition
duration, VNRs for three directions grew quickly and basically stayed at high
values without big change. Meanwhile, the noise and artifacts for velocity
maps gradually decreased and velocity maps gained good qualities at long
acquisition duration (T > 13.6min). Discussion
We investigated the relationship between image
quality and acquisition duration for 4D flow MRI in this study. When the
acquisition duration is too short, the low-rank modeling suffers from the ill-conditioned
problems with small number of data. As a result, magnitude and velocity maps were
contaminated by noise and artifacts. With the increase of acquisition time, the
image quality improved and converged at 13.6 minutes in aorta scan. Further study
need to be performed to investigate the optimized acquisition duration for
different acquisition duration containing various information in both spatial
and temporal domains.Conclusion
In this study, we reconstructed the data
from arbitrary acquisition duration for real-time 4D flow MRI, and assessed the
image quality with different amount of data associated with acquisition duration
for the low-rank modeling based real-time 4D flow MRI, which can provide an optimized
scan time for real-time 4D flow MRI.Acknowledgements
No acknowledgement found.References
[1] Sun A, Zhao B, Li Y, et al. Real-time
phase-contrast flow cardiovascular magnetic resonance with low-rank modeling
and parallel imaging[J]. Journal of Cardiovascular Magnetic Resonance, 2017,
19(1): 19.
[2] Sun A, Zhao B, Li R, and Yuan C.
High-Resolution 4D real-time phase-contrast flow MRI with sparse sampling. In
Proceedings of the 25th Annual Meeting of ISMRM, Honolulu, Hawaii, 2017. p.
1261.
[3] Ringgaard S, Oyre S A, Pedersen E M.
Arterial MR Imaging Phase-Contrast Flow Measurement: Improvements with Varying
Velocity Sensitivity during Cardiac Cycle 1. Radiology, 2004, 232(1): 289-294.