Territorial arterial spin labeling (tASL) [1,2] offers vessel specific flow information that was only feasible via DSA. However, tASL based perfusion faces a practical challenge of cerebral blood flow (CBF) quantification that is largely affected by its non-ideal labeling profile [3]. The actual labeling efficiency is further complicated by the misalignment between the labeling center and target artery, due to operational error or patient movement between localizing MRA scan and tASL. In this work, a rapid tASL MRA is used as calibration scan to derive the actual labeling efficiency, and hence allows the CBF to be corrected with respect to conventional ASL. The CBF quantification in tASL may be related to that obtained using conventional non-selective ASL, which is considered to be accurate, by a scaling factor as determined by the labeling efficiency. This means if we may have a conventional ASL as well as a tASL on hand, the labeling efficiency may be derived on the run. However, it is difficult to define a proper ROI in tASL perfusion map due to the irregular shape of the cerebral perfusion region, as well as the fact that many regions are fed by multiple arteries that making direct comparison to conventional ASL incorrect (Fig.1). On the other hand, MRA features have well constrained and simple region of interest that is within the vessel wall, hence it is much easier to derive the labeling efficiency using tASL and conventional MRA. In addition, a much shorter labeling time can be used to reduce the scan time, blood flow prior to M1 segment in the internal carotid artery (ICA) suffices for such purpose (illustrated in Fig.1). Territorial ASL was implemented for perfusion and MRA (Fig. 2a). The simulated labeling profile for an elliptical labeling region (long/short axis: 30/20mm) is shown in Fig.2b, it can be seen that the labeling efficiency drop rapidly as moving away from the center, hence it is impossible to estimatethe actual labeling efficiency in practice. 3D spiral read out was used in perfusion while 3D radial readout was used in MRA. A healthy volunteer without known cerebral vascular disease was recruited for this study, and consent form was obtained prior to scan. The right internal artery (RICA) was labeled (Fig. 3b) in both tASL perfusion and tASL MRA, and whole brain coverage acquisition was made. Conventional non-selective ASL perfusion and ASL MRA were also acquired for reference. An axial slice perfusion map from tASL is overlaid on the full labeling perfusion map (Fig. 3c, labeling time/PLD = 1000/1525ms), whereas the MIP from tASL with a short labeling time is overlaid on the full MIP MRA (Fig. 3d, labeling time/PLD = 200/0ms). In order to verify the consistency of MRA signal to that of perfusion with varying labeling consistency, the center of the labeling region was intentionally shifted away from center of the vessel, and signal variations were observed and compared. The ROI on the perfusion map was placed on the edge of the right frontal lobe to ensure the sole supply by RICA, ROI on MRA was selected as regions within the vessel wall in several consecutive axial slices.The calculated scaling factor between non-selective ASL perfusion to tASL perfusion and that between non-selective ASL MRA to tASL MRA were 2.23 and 2.38 respectively. The tASL perfusion and MIP tASL MRA images acquired with varying offsets are shown in Fig.4, it can be seen that signal dropped rapidly with an increasing offset from a vessel center. With a labeling radius of 30mm, little signal was left with a spatial offset of 18mm. The labeling efficiencies derived from perfusion map and from MRA with increasing offsets are plotted on the right top, consistent variations between the two were observed. A practical difficulty in tASL perfusion, given its obvious advantage, is the perfusion signal quantification such as CBF, which is hassled by the unknown labeling efficiency. In this proof of concept study, this issue is tackled by using tASL MRA images to derive the actual labeling efficiency. Instead of using the perfusion maps, MRA images features well defined and sole vessel supplied ROI, as well as fast acquisition. From the preliminary results, it is observed that very similar labeling efficiency can be derived from perfusion and MRA. Hence the labeling efficiency derived from tASL MRA may be translated to tASL perfusion for correction of the CBF quantification. With further optimization of the labeling time and acquisition strategy, it might be possible to acquire needed MRA images within tens of seconds.