PURPOSE

To develop and evaluate an M1-space (gradient first moment) under-sampling phase-contrast MRI (PC-MRI) technique with hybrid one- and two-sided flow-encoding (FE) and velocity spectrum separation (HOTSPA) for accelerated blood flow and velocity measurement.

METHODS

The temporal sampling period of the conventional reference 4D flow PC-MRI is 4*TR*views-per-segment (VPS). In the HOTSPA technique, the temporal sampling period is shortened by 50% to 2*TR*VPS. This is achieved by applying two-sided FE in the z-direction (i.e. +FE_z and -FE_z) and by alternating two orthogonal in-plane FE directions in the X-Y plane (i.e. X+Y and X-Y directions) in two consecutive cardiac phases as shown in Fig. 1. Specifically, the two-sided FE_z data includes a temporal modulated velocity waveform $$$\phi_{0}(n)\pm\phi_{v,z}(n)=\phi_{0}(n)+e^{i*\pi*n}*\phi_{v,z}(n)$$$, where $$$\phi_{0}(n)$$$ is the the flow-compensated (FC) background phase, $$$\phi_{v,z}(n)$$$ is the signal phase associated with z-directional velocity and n is cardiac phase number. The hybrid one-sided FE_x and two-sided FE_y are used to generate a $$$\phi_{0}(n)+\phi_{v,x}(n)\pm\phi_{v,y}(n)=\phi_{0}(n)+\phi_{v,x}(n)+e^{i*\pi*n}*\phi_{v,y}(n)$$$ waveform in the two remaining in-plane directions. $$$\phi_{v,x/y}(n)$$$ is the signal phase associated with x/y-directional velocity. If one applies a temporal Fourier transform of HOTSPA 4D flow data with alternating FE directions, the Fourier velocity spectra for the three orthogonal flow directions are shifted away from each other and can be easily separated using a filter. Such temporal modulated FE strategy allows for separations of the Fourier velocity spectrum into components for the FC and the three-directional (3D) velocity waveforms, accelerating PC-MRI by encoding 3D velocities using only 2TRs (±Z and X±Y) instead of 4TRs (four M1-space encoding steps including FC and three-directional FE) as in reference four-point PC-MRI. The theory and workflow are shown in Fig. 2. More details about the technique are recently published in Ref. 1. Six volunteers were scanned at the common carotid arteries (CCAs) using four sequences: 1) 2D reference four-point PC-MRI with three-directional FE plus the FC data, i.e. 2D FC/3FE; 2) our HOTSPA strategy but with 2D spatial encoding, i.e. 2D HOTSPA. The 2D sequences parameters are: TE/TR=3.9/6.3 ms, Flip Angle=20°, VENC=100-110 cm/s, FOV=256*176 cm², Acquired matrix size=256*176, slice thickness=7 mm, VPS=1 and 2 for FC/3FE, 2 for HOTSPA. 3) 4D flow FC/3FE; 4) 4D HOTSPA. The 4D sequences parameters are: TE/TR=3.6-3.9/6.2-6.4 ms, Flip Angle=20°, VENC=100-110 cm/s, FOV=256*176*20 cm³, Acquired matrix size=256*176*8, VPS=4.

RESULTS

Fig. 3a-b show examples of through-plane mean velocity and peak velocity measurement (average and maximum within the vessel lumen, respectively) of the CCA comparing three different measurements: the 1-VPS 2D FC/3FE (grey), the 2-VPS 2D HOTSPA (blue), the 2-VPS 2D FC/3FE PC-MRI (red). The velocity measurements were similar between the 2-VPS 2D HOTSPA and the 1-VPS 2D FC/3FE, although HOTSPA acquisition time was reduced by 50%. However, 2-VPS 2D FC/3FE underestimated the peak velocity by approximately 15% due to its long temporal footprint and temporal sampling period. Across the six subjects, using the 2D 1-VPS FC/3FE as the reference, the bias of 2D HOTSPA was -0.005 mL (-0.1% relative bias error) with 95% Confidential Interval (CI) [-0.3, 0.3] mL for total volumetric flow and 1.21 cm/s (1.14% relative bias error) with 95% CI [-5.1, 7.5] cm/s for peak velocity (Bland-Altman plots of Fig. 3c-d). The peak velocity of all 12 measurements based on the 2-VPS HOTSPA and 1-VPS 2D FC/3FE PC-MRI were significantly higher than the 2-VPS 2D FC/3FE measurements (P<0.05, one-side paired t-test). Fig. 4a-b show examples of through-plane mean velocity and peak velocity measurement in one slice of the 4D flow data comparing two different measurements: the 4-VPS 4D FC/3FE (red), and the 4-VPS 4D HOTSPA (blue). The 4D HOTSPA technique generated up to 40% higher peak velocity compared with standard 4D FC/3FE. As shown in Fig. 4c, the total volumetric flow measurements agree well between 4D HOTSPA and 4D FC/3FE (-0.02 mL bias with [-0.3 0.3] mL 95% CI). The peak velocity from 4-VPS HOTSPA was significantly higher than the 4-VPS standard 4D FC/3FE (P<0.05, one-side paired t-test).

DISCUSSION

In this work, we propose a new and more efficient flow encoding and velocity calculation strategy for PC-MRI using a temporal modulation technique to double the temporal resolution or reduce the scan time by 50%. This is the first study to examine a temporal modulation strategy for an under-sampled M1-space. Our strategy can be combined with conventional acceleration techniques, i.e. parallel imaging and compressed sensing, to further shorten the scan time of PC-MRI.

CONCLUSION

The proposed HOTSPA technique achieves nearly two-fold acceleration of PC-MRI while maintaining accuracy for total volumetric flow and peak velocity quantification.

Acknowledgements

NIH R01HL127153

References

1. Wang D., et.al. MRM 2016 Aug 9. doi: 10.1002/mrm.26366. [Epub ahead of print]