Unilateral fetal-type circle of Willis anatomy causes right-left asymmetry in cerebral blood flow with pseudo-continuous arterial spin labeling: A limitation of arterial spin labeling-based cerebral blood flow measurements?

The accuracy of cerebral blood flow measurements using pseudo-continuous arterial spin labeling can be affected by vascular factors other than cerebral blood flow, such as flow velocity and arterial transit time. We aimed to elucidate the effects of common variations in vascular anatomy of the circle of Willis on pseudo-continuous arterial spin labeling signal. In addition, we investigated whether possible differences in pseudo-continuous arterial spin labeling signal could be mediated by differences in flow velocities. Two hundred and three elderly participants underwent magnetic resonance angiography of the circle of Willis and pseudo-continuous arterial spin labeling scans. Mean pseudo-continuous arterial spin labeling-cerebral blood flow signal was calculated for the gray matter of the main cerebral flow territories. Mean cerebellar gray matter pseudo-continuous arterial spin labeling-cerebral blood flow was significantly lower in subjects having a posterior fetal circle of Willis variant with an absent P1 segment. The posterior fetal circle of Willis variants also showed a significantly higher pseudo-continuous arterial spin labeling-cerebral blood flow signal in the ipsilateral flow territory of the posterior cerebral artery. Flow velocity in the basilar artery was significantly lower in these posterior fetal circle of Willis variants. This study indicates that pseudo-continuous arterial spin labeling measurements underestimate cerebral blood flow in the posterior flow territories and cerebellum of subjects with a highly prevalent variation in circle of Willis morphology. Additionally, our data suggest that this effect is mediated by concomitant differences in flow velocity between the supplying arteries.