Afis AJALA1,2, Jiming Zhang2, Erick Buko1,2, Luning Wang3, Debra Dees2, Janie Swaab2, Benjamin Cheong2, Pei-Herng Hor4, and Raja Muthupillai2
1Department of Physics, University of Houston [Main Campus], Houston, TX, United States, 2Department of Diagnostic and Interventional Radiology, Baylor St. Luke's Medical Center, Houston, TX, United States, 3Philips Healthcare, Gainesville, FL, United States, 4Physics and Texas Center for Superconductivity, University of Houston [Main Campus], Houston, TX, United States
Synopsis
PC-MRI
based Left ventricular (LV) diastolic indices such as E/Em ratio are
conventionally estimated from two separate PC acquisitions: One sensitive to
myocardium and the other to blood velocities. This ensures optimal velocity to
noise ratio and zero velocity aliasing. A rapid single acquisition (dual echo
dual VENC PC-MRI method) sensitive to both low and high velocities was
developed. Preliminary result on 7 healthy subjects demonstrated a high
agreement of LV diastolic indices obtained
from the proposed and conventional method.
Introduction
A key parameter in the evaluation of LV
diastolic function is the ratio of peak blood flow velocity (E) to peak myocardial velocity (Em) during diastole. Em
is several fold lower than E. The traditional approach of setting the
velocity sensitivity or VENC value of PC-MRI slightly above the maximum
velocity likely to be encountered across the FOV to avoid aliasing, for example
E in this instance, diminishes the velocity to noise ratio (VNR) for estimating
Em 1.Until now, E and Em are measured in two acquisitions (or at best two TRs) with differing velocity sensitivities at the cost of
doubling acquisition time - an approach we call single-echo dual velocity
sensitivity (SEDV) technique.Purpose
The aim of this work is (1) To demonstrate the
feasibility of a rapid dual echo approach with dual velocity sensitivity (DEDV)
method2 that can simultaneously measure E
and Em without
compromising VNR in asymptomatic normal subjects, and (2) To compare the
diastolic function metrics (E/Em and
the deceleration time, DT) estimated
from the SEDV and DEDV approaches.Materials and Methods
Subjects: 7 subjects (5 males, 2 females age:
51±13 yrs) were imaged on a commercial 3T MR imager (Ingenia, Philips
Healthcare) with VCG gating, and a phased array coil for signal reception. All subjects provided written informed
consent.
MRI
acquisition: A custom software patch that allowed the prescription of different
VENC values for a dual echo scan was created (Fig. 1) for a free-breathing
segmented k-space PC-MRI sequence,
and used in the study. PC-MRI of the
short axis slice positioned below and parallel to the mitral valve plane was
acquired with the DEDV pulse sequence with velocity sensitivity for the first
and the second echoes set at: 150 cm/s and 20 cm/s; acquired voxel size: 2 x 2
x 8 mm3; temporal resolution: 20 ms. The
same slice was acquired with conventional SEDV technique with VENC values set
at 150 cm/s and 20 cm/s in two successive repetition times, while maintaining
identical FOV, and spatial resolution as that of DEDV.
Data Analysis: Mean
blood and myocardial tissue velocities were estimated from regions of interest
(ROIs) circumscribing LV blood and myocardium, from PC-MRI acquired with high
and low velocity sensitivities respectively.
Myocardial region was further subdivided into segments to
extract Em of septal,
inferior, lateral, and anterior walls. E/Em and DT were calculated for each segment using both techniques. The positive axis of velocity was defined as
tissue motion towards the apex of the heart. Linear regression and Bland-Altman
plots used to compare the E/Em
ratio and DT for the SEDV and DEDV methods.
Results
Myocardial and blood flow velocity curves
throughout the cardiac cycle extracted from the high and low VENC acquisitions
for the SEDV and DEDV acquisitions were similar (Fig. 3). E/Em
between SEDV and DEDV techniques was well-correlated (r2 = 0.88),
and had negligible bias: 0.01 (CI: -3 to 3) (Fig. 4). DT between the two
techniques was well correlated (r2 = 0.81) and Bland Altman analysis
showed a bias of -2.7 ms (CI: -45 ms to 39 ms).
The mean scan time for SEDV and DEDV techniques was: 7.2 ± 2.4 min
versus 4.5± 2.3 min, respectively.Discussion
The results from this study show that E and Em
values obtained from the DEDV method compares well with the conventional SEDV
method, and the calculated indices lie in the range reported for healthy,
asymptomatic volunteers3. It is important to note that the scan time
reduction with DEDV does not impose any constraints in spatial, and/or velocity
resolution. Furthermore, this approach
can be integrated with recently described spatio-temporal undersampling methods
such as k-t BLAST, or compressed sensing4. It will be possible to quantify the
late diastolic peak velocity (another important diastolic index) if the entire
cardiac cycle is covered using a retrospective phase reconstruction. Conclusion
In this
work, we demonstrated the feasibility of a rapid PC-MRI method that is capable
of yielding diastolic functional metrics such as E/Em ratio and DT
that are comparable to those obtained using conventional SEDV approaches.Acknowledgements
No acknowledgement found.References
[1] N. J. Pelc, R. J. Herfkens, A.
Shimakawa, and D. R. Enzmann, “Phase contrast cine magnetic resonance imaging,”
Magn. Reson. Q., vol. 7, no. 4, pp. 229–254, 1991.
[2] A. Ajala, J. Zhang, B. Cheong, P.-H.
Hor, and R. Muthupillai, “A Dual Echo Dual VENC (DEDV) phase contrast method
for simultaneous measurement of myocardial and blood flow velocities,” in Joint
ISMRM–ESMRMB, 2018, p. 0011.
[3] J. G. Delfino et al., “Comparison
of myocardial velocities obtained with magnetic resonance phase velocity
mapping and tissue doppler imaging in normal subjects and patients with left
ventricular dyssynchrony,” J. Magn. Reson. Imaging, vol. 24, no. 2, pp.
304–311, Aug. 2006.
[4] C.
Baltes, S. Kozerke, M. S. Hansen, K. P. Pruessmann, J. Tsao, and P. Boesiger,
“Accelerating cine phase-contrast flow measurements using k-t BLAST and k-t
SENSE,” Magn. Reson. Med., vol. 54, no. 6, pp. 1430–1438, Dec. 2005.