Comparison of algorithms for estimating blood flow velocities in cerebral arteries based on the transport information of contrast agent: An in silico study

Abstract

While many algorithms have been proposed to estimate blood flow velocities based on the transport information of contrast agent acquired by digital subtraction angiography (DSA), most relevant studies focused on a single vessel, leaving a question open as to whether the algorithms would be suitable for estimating blood flow velocities in arterial systems with complex topological structures. In this study, a one-dimensional (1-D) modeling method was developed to simulate the transport of contrast agent in cerebral arterial networks with various anatomical variations or having occlusive disease, thereby generating an in silico database for examining the accuracies of some typical algorithms (i.e., time-of-center of gravity (TCG), shifted least-squares (SLS), and cross correlation (CC) algorithms) that estimate blood flow velocity based on the concentration-time curves (CTCs) of contrast agent. The results showed that the TCG algorithm had the best performance in estimating blood flow velocities in most cerebral arteries, with the accuracy being only mildly affected by anatomical variations of the cerebral arterial network. Nevertheless, the presence of a stenosis of moderate to high severity in the internal carotid artery could considerably impair the accuracy of the TCG algorithm in estimating blood flow velocities in some cerebral arteries where the transport of contrast agent was disturbed by strong collateral flows. In summary, the study suggests that the TCG algorithm may offer a promising means for estimating blood flow velocities based on CTCs of contrast agent monitored in cerebral arteries, provided that the shapes of CTCs are not highly distorted by collateral flows.