Pseudocontinuous arterial spin labeling perfusion magnetic resonance imaging--a normative study of reproducibility in the human brain.

Pseudocontinuous arterial spin labeling (PCASL), a newly proposed perfusion magnetic resonance imaging (MRI) technique, has the potential of providing a better balance between labeling efficiency and signal-to-noise ratio than conventional ASL methods. The advantage will not be exploitable until adequate reproducibility can be achieved. In this study, we investigated the reproducibility of PCASL on twelve healthy volunteers by taking into account the inclusion of correction for coil sensitivity (CS) and labeling efficiency (α). Each subject underwent two identical sessions scheduled two weeks apart. For comparison, pulsed ASL (PASL) imaging was also conducted using the flow-sensitive alternating inversion recovery method. MR imaging was performed on a 3T whole body system using a 12-channel phased array for reception. Blood flow was quantified for gray matter and assessed globally as well as in the flow territories of the internal carotid arteries and posterior cerebral arteries. Experimental results showed that CS accounted for 35% and 60% of inter-session variability in PCASL and PASL, respectively. With CS correction alone, coefficient of variance (CV) remained larger with PCASL than PASL (p=0.02). CV was 12% with CS-corrected PASL and 11% with PCASL when CS and α were corrected for in conjunction. We concluded that CS correction is necessary for ASL imaging when using phased array receiver and that after CS correction, PCASL requires α correction to provide reproducibility comparable to PASL.