Recently, arterial spin labeling (ASL)-based 3D time-resolved MRA (4DASL-MRA) has been proposed [1], where the single-tag with multiple-delay (TI) method is becoming major due to its time efficiency. 4DASL-MRA enables to provide information of blood-vessel hemodynamics. In contrast, black-blood (BB) MRA with preserving stationary tissue contrast is important for vessel-wall or plaque imaging and several methods have been proposed. Among them, a phase-sensitive inversion-recovery (PSIR) method is receiving a significant attention due to lumen-to-background contrast can be increased by negative blood signal, even though the phase correction is required using additional acquisition [2]. We proposed an alternative PSIR-BB technique using TOF image for phase correction [3]. However, the former methods were applied with single TI. The purpose of this study was to propose and assess a simultaneous time-encoding for bipolar MRA technique (STEP-MRA), combining both 4DASL-MRA and PSIR BB-MRA, which allows simultaneously providing the same acquired data as the standard 4DASL-MRA.

Theory

There are roughly two different methods of STAR and FAIR in pulsed ASL where the flowing blood is inverted commonly while stationary tissues are not inverted for STAR but inverted for FAIR. The tag and control are commonly defined as the upper blood stream outside of imaging slab is inverted and not inverted. An ASTAR technique [4] is a modified version of STAR. Figure 1 shows the schematic diagram of STEP-MRA. Although there are several variations for the combination of tag and/or control images in each of N phases as shown in Table 1, the three kinds of combinations are available in STEP-MRA. In STEP-MRA, the background phase correction for PSIR-BB is performed using the control image to obtain ASL-MRA where the longitudinal magnetization (Mz) of tissues in every phase is positive. Every phase map in N control images for “ASTAR_N-N” method is theoretically the same. The arbitrary phase image in N control images is available to be commonly used to correct each different phase tag image. If the Mz of tagged blood is almost fully recovered in the final phase of tagged images, the image can be commonly used for phase correction for the tagged images of TI<TI(N) even without paired control images. The “ASTAR_N-1” method is proposed based on this assumption, where the acquisition time is becoming the half of the ASTAR_N-N method (Fig. 1). Process flow for STEP-MRA is shown in Fig. 2.

Experiments

Here methods of ASTAR_N-N and ASTAR_N-1 were performed and assessed for healthy volunteer head images acquired on a Vantage Titan 3T with 14ch Atlas SPEEDER Head coil after obtaining informed consent. Pulse sequence was: 3D fast-field-echo (FFE3D), TR/TE/FA=4.7ms/1.3ms/8°, BW=651Hz/pixel, ISCE=1:3, FOV=18x23cm, 2mm x 55slice, # of segment=2, PI=x2(PE), ΔTI=201ms, # of phase (N)=7 (TI=150-1356ms), repeat time=1800ms, 3:32/tag or control. For tag, tag thickness=20cm, tag-imaging gap=2cm.

Figure 2 shows the imaging results for ASTAR_N-N and N-1 mode. Regarding the ASL-MRA for 3 methods, the peripheral vessels on ASTAR_N-N were better depicted than those on ASTAR_N-1 at longer TI, because the inflow blood Mz in Stag[TI(N)] has not recovered as in Scont[TI(N)] due to insufficient TI(N) (=1356 ms); this was clarified by the case that Scont[TI(N)] was used for the control instead of Stag[TI(N)] (3rd row in Fig. 3). If the TI(N) is further longer (~2000ms) for Stag[TI(N)], this problem will be reduced. In contrast for PSIR-BB, BB vessels gradually moved to peripheral portions with increasing TI but almost disappeared at late TI (>1000ms). BB images were almost common between N-N and N-1 because both the phase maps using correction had positive polarities except on air portions. In contrast for ASL-MRA, peripheral vessels were better depicted than BB at late phase for every method. Regarding the results of synthesizing multiple TI images (Fig. 2 tMIP, tminIP), the SNRs were increased and the arteries from proximal to peripheral were better visualized than each of the single TI images in both the ASL-MRA and PSIR-BB. By combining ASL and PSIR-BB, comparing with the stand-alone synthesis of PSIR-BB, vessel-to-backgrounds were further increased nevertheless SNR of stationary tissues were slightly reduced. These quantitative comparisons were shown in Fig. 4.We proposed and assessed a new STEP-MRA technique, combining both the 4DASL-MRA of nulling background signals and PSIR BB-MRA with preserving Mz polarities and background stationary signals, which allows simultaneously providing using the same acquired data as the standard 4DASL-MRA. In addition, vessel visualization and SNR were improved by synthesizing multiple TI data. Although further parameter optimization and clinical evaluation are required, our proposed method has a potential to provide various information with limited acquisition data.