Rethinking vascular artifacts: testing the sensitivity of ASL vascular signal as a biomarker of disease

It was recognized early on in the development of ASL that the contribution of the vascular signal was a confound in measuring tissue perfusion. To this end, Alsop et al. proposed introducing a post-label delay (PLD) to allow for the blood to arrive at the microvasculature level and preferably wash-out from the macro-vasculature1. Considering, however, that the ASL signal decays during the PLD, a compromise is typically made between the need to minimize the effect of the vascular signal and optimum SNR. The choice of PLD becomes more complex in clinical applications where the arterial transit time (ATT) is expected to vary considerable across patients as well as within a patient, such as in carotid occlusive disease. In such applications, a concomitant measurement of both CBF and ATT would be ideal. However, ATT acquisition leads to longer scanning time, lower SNR, and higher motion sensitivity. To this end, Mutsaerts et al. have recently shown that spatial coefficient of variation (sCoV) of the ASL signal can be used as a proxy for the ATT measurement2. In this study, we tested whether an asymmetry in sCoV in carotid occlusive patients could predict the side of the occlusion with higher sensitivity than the CBF asymmetry. We also investigated the temporal variance of the ASL signal and tested its feasibility to detect the vascular signal in the ASL data from this patient population.