Labeling Efficiency Is Critical in Pseudo-continuous Asl

INTRODUCTION: Pseudo-Continuous Arterial Spin Labeling (pCASL) (1-3) is a new ASL technique that has the potential of combining advantages of continuous ASL (large brain coverage, high SNR) and pulse ASL (no special hardware). One of the most critical parameters in pCASL is the labeling efficiency, defined as (arterial blood in the control scan – arterial blood in the label scan)/2. Unlike continuous ASL, the labeling process of pCASL is not strictly an adiabatic inversion. Thus the labeling efficiency may be dependent on B0 inhomogeneity, B1 inhomogeneity and flow velocity. As a result, the labeling efficiency may vary for different subjects, for different labeling locations in the same subject, and for different physiologic state at the same labeling location. Here, we performed three sets of studies to: 1) empirically determine the optimal labeling location to achieve best labeling efficiency; 2) experimentally estimate the labeling efficiency using phase-contrast MRI as a normalization factor; 3) demonstrate that labeling efficiency may change with physiologic state and should be estimated for each physiologic condition. METHODS: A total of 21 healthy controls (age 32±8 years) were studied on a 3T scanner (Philips). In the first set experiments (n=8), with the imaging slices in the same location, the labeling position was varied to be 49, 63, 74, 84, 94, 119, and 149mm distal to the anterior-commissure (AC) posteriorcommissure (PC) line (bottom panel, Fig. 1). The CBF-weighted signal (control-label) was calculated and compared across different labeling positions. In the second set of experiments (n=10), the labeling position was chosen to be the above determined optimal location and the imaging slices covered the whole brain (27 slices with slice thickness = 5mm). Importantly, a phase-contrast MRI was performed at the level of left/right internal carotid and left/right vertebral arteries (Fig. 2), from which we can calculate the total amount of blood entering the brain in units of ml/min. We also know the total intracranial volume and mass from a standard T1w MPRAGE image (processed with FSL software). Thus, we can determine the whole-brain-averaged CBF (ml/min/100g brain) = phase contrast (ml/min) / intracranial mass (100g). This value can then be used as a normalization factor to estimate the labeling efficiency in pCASL by solving the equation: Whole-brain-averaged CBF from phast-contrast = [1]