Synopsis

4D flow MRI suffers from long scan times which limit maximum spatial resolutions. A promising approach is to restrict the field-of-excitation (FOX) to the region of interest and therefore reduce the field-of-view (FOV) not only in partition/slab direction, but also in phase encoding direction. In this work, we replace the slab-selective excitation of a 4D flow sequence by a 2D spatially-selective excitation with reduced FOX to enable reduced FOV imaging. We investigate the impact of the excitation on velocity encoding and demonstrate correct velocity quantification with $$$10\%$$$ reduced scan times in phantoms and in-vivo.

Introduction

4D flow MRI¹ is the only technique allowing velocity vector field quantification in-vivo. However, the 3D slab-selective acquisition of the velocity vector field requires long scan times and thus limits the spatial resolution. A promising approach is to restrict the field-of-excitation (FOX) to the region of interest and therefore allow to image a reduced field-of-view (FOV) not only in partition/slab direction, but also in phase encoding direction.

In this work, we replace the slab-selective excitation of a 4D flow sequence by a 2D spatially-selective excitation with spiral trajectory and demonstrate correct velocity quantification in a reduced FOX (rFOX) and reduced FOV (rFOV) in a phantom and intracranial vessels.

Methods

2D Selective Excitation

We designed a 2D selective RF-pulse with spiral k-space trajectory (Fig.1a) and bandwidth-time-product $$$\text{BWTP}=1.2$$$ in the small flip angle (FA) regime as proposed by Pauly². The spiral starts at $$$k_{\text{max}}=2.51\,\frac{\text{rad}}{\text{cm}}$$$ and moves in $$$T=3.5\,\text{ms}$$$ with $$$n=10$$$ turns inwards to $$$\boldsymbol{k}$$$-space center.

The actual $$$\boldsymbol{k}$$$-space trajectory was measured according to Duyn³ prior to scanning (parameters: Tab.1, line1). Two RF-pulses were calculated to excite a FOX of $$$(6\times6\times\infty)\,\text{cm}^3$$$: i) based on nominal gradient waveforms (RF_nom); ii) based on measured gradient waveforms (RF_meas).

4D Flow MRI Combined with 2D Selective Excitation

In order to quantify flow by phase-contrast imaging the phase of moving spins $$$\phi(v)=\gamma\,M_1\,v$$$ is manipulated by bipolar gradients such that the total first gradient moment between the effective center of excitation $$$t_0$$$ and echo-time $$$t_{\text{TE}}$$$ is $$$M_1=\int_{t_0}^{t_{\text{TE}}}G(s)\,s\,\text{d}s=\frac{\pi}{\gamma\,v_{\text{enc}}}$$$. While for symmetric SINC pulses $$$t_0$$$ is nearly at the center of the RF-pulse⁴, there are controversial reports about $$$t_0$$$ of 2D pulses^5,6.

Therefore, we performed Bloch-simulations of our 2D spiral excitation using moving spins with varying velocities $$$v_{x,y}\in[-20,20]\,\text{m/s}$$$ to obtain the phase as a function of velocity at the end of excitation $$$T$$$. Thereby, we obtained $$$M_1^{v_i}(T)=\frac{\phi(T)}{\gamma\,v_i}$$$, where $$$i=x,y$$$. Finally, we compared it to the respective gradient moment $$$M^{v_i}_1(T) = \int_{t_0}^TG_{i}(s)\,s\,\text{d}s$$$ for $$$t_0\in[0,T]$$$.

Based on our findings, we included the 2D excitation pulse in a gradient echo 4D flow ECG-triggered MR sequence¹.

Phantom and In-Vivo Experiments

Two phantom and one in-vivo measurement were performed at 3T (Magnetom Verio, Siemens, Erlangen, Germany). First, we included RF_nom and RF_meas in a GRE sequence to investigate the squared excitation and non-excitation region in a large FOV using three adjacent phantom bottles (parameters: Tab.1, line2&4). To eliminate proton density and receive profiles, we normalized the data with the corresponding slab-selective acquisition (parameters: Tab.1, line3&5).

Next, the excitation performance and accuracy of flow quantification was investigated non-time-resolved in a flow phantom with constant flow. For comparison, we imaged all possible four combinations of full/reduced FOV and selective/non-selective excitation with RF_meas (parameters: Tab.1, line6-9).

Finally, a volunteer was scanned according to an approved ethical protocol and with full informed consent. A flow sequence (parameters: Tab.1, line10&11) was acquired in cerebral vessels with two different excitations targeting the Circle-of-Willis: i) 1D slab-selective excitation and ii) 2D $$$(6\times6\times\infty)\,\text{cm}^3$$$-selective excitation using RF_meas.

In all flow acquisitions, read-out was along the non-selective axis ($$$k_z$$$) and both phase-encodings perpendicular to it. Acquisition parameters are summarized in Tab.1.

Results

Fig.1c-e show the excitation profile obtained with RF_nom and RF_meas. The residual excitation of RF_meas outside the excitation window of $$$(10\times10)\,\text{cm}^2$$$ in Fig.1d is $$$(0.6\pm0.6)\%$$$.

The Bloch-simulations of moving spins show nearly velocity-independent phase at the end of the RF-pulse $$$T$$$ (Fig.2), implying $$$t_0=T$$$.

Fig.3a-d show all four combinations of 2D/non-selective excitation of the flow phantom in a full/reduced FOV. For the rFOV with non-selective excitation, biased velocities are measured, as seen quantitatively in the Bland-Altman-plot (Fig.3f). This is not seen with 2D-selective excitation (Fig.3e,g). The mean velocity difference to non-selective excitation in a full FOV is for 2D-selective excitation with full FOV $$$\Delta\,v/v_{\text{ref}}=(1.1\pm5.5)\%$$$ and with rFOV $$$\Delta\,v/v_{\text{ref}}=(2.9\pm5.9)\%$$$.

Fig.4 shows the absolute velocity and its three components during systole $$$130\,\text{ms}$$$ after ECG-trigger in the Circle-of-Willis for both 1D- and 2D-selective excitation. Qualitatively no aliasing of regions outside the FOX was observed.

Quantitatively, the flow through the internal carotids and basilar artery over the cardiac cycle is compared for both different FOXs and corresponding FOVs in Fig.4a. The maximum difference is reached during systole, $$$130\,\text{ms}$$$ after ECG-trigger, with $$$\Delta\,v/v_{\text{ref}}=6.6\%$$$.

The acquisition time with 2D excitation and rFOV was reduced by $$$10\%$$$.

Discussion and Conclusion

Acknowledgements

References

1. Markl M et al. 4D flow MRI, J. Magn. Reson. Imaging 2012;36:1015-1036.

2. Pauly J et al. A k-Space Analysis of Small-Tip-Angle Excitation. J. Magn. Reson. Imaging 1989;81:43-56.

3. Duyn J H et al. Simple Correction Method for k-Space Trajectory Deviations in MRI. J. Magn. Reson. 1998;132:150-153.

4. Bernstein M A et al. Handbook of MRI pulse sequences, John Wiley & Sons, Ltd 2005

5. Hardy C J et al. Rapid NMR Cardiography with a Half-Echo M-Mode Method. J. Comput. Assist. Tomogr. 1991;15:868-874.

6. Cline H E, et al. Fast MR cardiac profiling with two-dimensional selective pulses. Magn. Res. Med. 1991;17:390-401.