Relationships between local geometrical features and hemodynamic flow properties

Stroke is among the leading causes of death and disability worldwide. Most strokes are ischemic, mostly caused by the blockage of a cerebral artery by a thrombotic embolus. Carotid atherosclerosis and the subsequent plaque rupture can be a major source of these emboli. It is well known that blood flow affects where atherosclerotic plaque will arise. In particular, vascular wall shear stress (WSS) has been linked to the initiation and progression of carotid plaque. However, it is difficult to measure WSS in vivo and it is time-consuming to compute WSS using computational fluid dynamics packages. The goals of this paper are (i) to identify a set of local geometric parameters that are correlated with WSS and (ii) to develop a regression model to predict WSS from the geometric parameters. We validated our regression model using the root mean squared error (RMSE), adjusted R² and Akaike information criterion (AIC). The experimental study involved six carotid arteries with the internal and external carotid arteries (ICA and ECA respectively) analyzed separately. The adjusted R²s for 9 of the 12 branches were higher than 0.8. Since the proposed local geometric parameters can be obtained efficiently, these parameters can potentially be used as carotid disease phenotypes that will allow for a much more cost-effective method to identify subjects with elevated stroke risk.