Zero TE (zTE) silent MRA is a novel technique based on continuous arterial spin labeling (ASL) and zTE acquisition and has shown prominent depiction of arteries in regions affected by turbulent flow and stent, where conventional MRA, like TOF, usually fails [1]. A long readout train with repeated small flip angle excitation was used in zTE acquisition. With such sampling scheme, hollow vessel is likely to appear in the inferior end of internal carotid arteries (ICA) in high resolution imaging. Additional pulsed ASL module has been reported addressing this issue via generating continued inflow of inverted blood during acquisition, and similar technique was termed hybrid ASL (hASL) [2]. Here, a silent version of hybrid ASL was implemented in zTE MRA and different kinds of hybrid ASL implementations were compared.The formation of hollowing artifacts is illustrated in Fig.1a: continuous ASL labels all the blood flowing through a labeling plane (Fig.2a), labeled blood (red arrows) flow into the imaging volume and produces contrast; however within the length of data readout after labeling, fresh blood may flow into the imaging region and lead to signal void. Typically, this occurs in the center of the vessel as blood near the vessel wall tends to travel at a slower speed. In hybrid labeling (Fig.2b and Fig.2c), either (STAR) or FAIR labeling strategy may be incorporated: straight after the continuous RF pulse, a short inversion RF pulse is added followed by an inversion time [1, 2]. The inversion pulse inverts a large volume beneath the imaging region so that within the length of data readout all the blood flowing into the imaging region will be labeled: if not by the continuous ASL, is still labeled by the pulsed ASL to eliminate the signal void. The difference between STAR and FAIR is the fact that STAR inverts the volume beneath the imaging volume (Fig.1b) whereas FAIR inverts the entire volume in control and the imaging region in labeling (Fig.1c). In addition, the slice selecting gradient and spoiler associated with the pulsed ASL can be heavily derated by utilizing the sufficiently long inversion period. In this way, the hybrid ASL with zTE readout remains silent essentially. The three methods as illustrated in Fig.2 (cASL, STAR-hASL, FAIR-hASL) were implemented on a 3.0T whole body scanner (GE Discovery MR750w). A healthy volunteer was recruited for the comparison study; consent form was obtained prior to the san. For fair comparison, the total length of the ASL labeling module was fixed among the three methods (1500 ms) whereas the inversion pulse and inversion time were respectively 1380 ms and 100 ms. Identical zTE acquisition was made for the three cases: FOV/Matrix 180mm/180; spokes per segment 512; flip angle 3o; and bandwidth 31.25 kHz; total scan time 344 second. The qualities of the resulting angiograms were judged by qualitative inspection as well as contrast to noise ratio measures (CNR).The coronal plane MIP ASL images using different labeling strategies are shown in Fig.3, the axial plane intersections at the inferior end of right ICA (blue circle in Fig.3) are shown in the right top corner. It can be seen the cASL (Fig.3a) showed heterogeneous signal variation within the vessel resembling the geometry of a hollow tube, whereas angiograms using both pASL (Fig.3b and 3c) were free of the artifacts. The measured CNR of right BA, ICA, middle cerebral artery (MCA) M1 and M2 segment (red circle in Fig.3) are shown in the bar chart in Fig.3d. It can be seen that both hASL schemes led to similarly high CNR in BA, ICA, and M1 as compared to cASL, whereas for M2 and distal arteries, CNRs using all three schemes were similar. Comparing STAR-hASL and FAIR-hASL, STAR-hASL led to slightly higher CNR and more clear depiction of some vessels (arrowed), but such subtle advantage needs to be further verified by a larger population data. In this work, we incorporated and compared two pASL strategies with cASL to eliminate the hollowing artifacts. With the same length of ASL preparation, both STAR and FAIR hASL prepared zTE MRA provide improved depiction of cerebral arteries up to the M1 segment of MCA as compared to cASL. Flow void effect was eliminated in hASL via continuous inflow of tagged blood during readout. The inversion time in pASL was utilized to minimize gradient slew rate so that the silent nature of zTE MRA may be maintained. No conclusion can be drawn on which of the two hASL method is more advantageous based on this preliminary data.