The most common and available method for hemodynamic MR velocity imaging on commercial clinical scanners is 2D phase contrast (PC) imaging using a bipolar gradient pulse [1] to encode the signal phase with velocity in cine gradient echo (GRE) sequences with low flip angle RF pulses, and cardiac gating (cine-PC) through the cardiac cycle [2,3]. Cine-PC 2D imaging measures flow velocity in a single slice plane and scan time is increased proportional to the number of repeated slice planes through arteries or CSF passageways. The purpose of this work is to develop and evaluate a cine-PC technique with simultaneous multi-slice (SMS) image acquisition to measure velocity in several slice planes simultaneously.A 2D FLASH sequence with cine acquisition and prospective cardiac gating was modified to excite with multiband (MB) rf pulses, Fig. 1, and bipolar velocity gradient. Controlled aliasing [4] with FOV/3 shift was performed with phase cycling the MB pulse, for 3 slices (0,0,0), (0,120,240), (0,240,120), on the k-space PE axis. Imaging was performed in 6 normal subjects with both single-slice cine-GRE PC and SMS cine-GRE PC for blood velocity in 32ch head coil and CSF velocity imaging in 12ch coil. Imaging parameters for cine-PC 2D SMS=1-4 and GRE cine-PC: SMS=1-4, TR=8.4ms, TE=4.65ms, BW=280 Hz/pixel, FA=15°, IPAT=2, FOV=192×192mm2, matrix=(128x128) or (256x256), in-plane resolution (res)=(0.75x0.75) or (1.5x1.5) mm2, venc=80cm/s, with 16mm or 32 mm slice spacing and slice thickness= 4 mm. Imaging by combining head-neck-spine coil array with head (12ch), neck (12ch) spine (4ch), was performed with SMS-3, res=1.25mm, slice=5mm, venc=100 cm/s, TR=9.55ms, TE=5.8ms, FA=15°. The preparation time for spatial sensitivity data, single band and dummy scans was short, about 180ms. Through plane velocities were measured with G-venc on G-slice axis, Fig.1. The velocity phase shifts are positive (white) and negative (black) corresponding to superior cranial and inferior caudal directions, respectively.Fig. 1 shows a reduction in slice cross-talk artifact achieved with use of controlled aliasing as seen by comparing the averaged phase map of cine-data (arteries-white, vein-black). The total scan time for 2 to 4 slices using SMS2-SMS4 was 2.2 minutes and similar for conventional single-slice (sms1). There was no loss in spatial resolution using SMS acquisitions. Figure 2 shows velocity measurements compared in internal carotid (ICA) and vertebral arteries and jugular veins, and very similar velocity curves using SMS PC and single-slice PC. Figure 3A shows comparison of velocity curves of ICA in 6 subjects in both scatter plots with correlation coefficients (R) and Bland-Altman plots, in good agreement. Figures 3B-C show comparison of tSNR and spatial SNR with little change between SMS-1 (single slice) and SMS-2, and 25% drop in tSNR in SMS-4, as expected for g-factor image reconstruction noise. Figure 4 shows simultaneous velocity imaging at 3 levels using SMS-3; the ascending and descending aorta above the heart, shoulder level and upper neck. Figure 5 shows SMS-2 of velocity in two CSF passageways at different levels, the aqueduct and foramen magnum.

Comparison of different cine acquisitions was to some degree affected by changes in heart rate between acquisitions, nevertheless there was high reproducibility between SMS and non-SMS velocity curves. Simultaneous acquisition of different slice levels eliminates velocity variability due to changes in heart rate, breathing rate and blood pressure that can occur between single-slice scans made sequentially at different levels, and is potentially important for combining dynamic information at different vessel cross-sections. This improved hemodynamic information from several regions could be used to derive hemodynamic metrics such as velocity drops across vessel stenosis or critical flow and sheer patterns in aneurysms. Acquisition of 3 velocity directions with SMS in the future may be faster than 4D PC imaging [5] for 3D spatial coverage of vessel lumen and for computational fluid dynamics analysis of flow.

An enormous gain in signal encoding efficiency with SMS cine-PC stems from the fact that there is no inherent SNR loss, as occurs with in-plane parallel imaging accelerations. Consequently N slices can be acquired in the time of conventionally acquired single-slice cine-PC without SNR reduction other than reconstruction noise (g-factor). Hence, these novel SMS velocity phase images improve SNR efficiency by as much as √N over single slice acquisition with no spatial resolution or artifact penalties.

The new technique, SMS cine-PC overcomes scan time limitations in acquiring multiple slices with 2D cine-PC imaging. Providing several simultaneous cross-sectional measurements of hemodynamics in vessels or CSF in passageways, SMS cine-PC may have increased utility for medical diagnosis of hydrocephalus and cardiovascular disease.