The Effect of Labeling Duration and Temporal Resolution on Arterial Transit Time Estimation Accuracy in 4D ASL MRA Datasets - A Flow Phantom Study

Abstract

Medical imaging modalities, such as four-dimensional arterial spin label magnetic resonance angiography (4D ASL MRA), can acquire blood flow data of the cerebrovascular system. These datasets are useful to determine criteria of normality and diagnose, study, and follow-up on the treatment progress of cerebrovascular diseases. In particular, variations in the arterial transit time (ATT) are related to hemodynamic impairment as a consequence of vascular diseases. In order to obtain accurate ATT estimations, the acquisition parameters of the applied image modality need to be properly tuned. In case of 4D ASL MRA, two important acquisition parameters are the blood labeling duration and the temporal resolution. This paper evaluates the effect of different settings for the two mentioned parameters on the accuracy of the ATT estimation in 4D ASL MRA datasets. Six 4D ASL MRA datasets of a pipe containing a mixture of glycerine and water, circulated with constant flow rate using a pump, are acquired with different labeling duration and temporal resolution. A mathematical model is then fitted to the observed signal in order to estimate the ATT. The results indicate that the lowest average absolute error between the ground-truth and estimated ATT is achieved when the longest labeling duration of 1000 ms and the highest temporal resolution of 60 ms are used. The insight obtained from the experiments using a flow phantom, under controlled conditions, can be extended to tune acquisition parameters of 4D ASL MRA datasets of human subjects.