Assessment of arterial stenosis in a flow model with power Doppler angiography: accuracy and observations on blood echogenicity.

The objective of the project was to study the influence of various hemodynamic and rheologic factors on the accuracy of 3-D power Doppler angiography (PDA) for quantifying the percentage of area reduction of a stenotic artery along its longitudinal axis. The study was performed with a 3-D power Doppler ultrasound (US) imaging system and an in vitro mock flow model containing a simulated artery with a stenosis of 80% area reduction. Measurements were performed under steady and pulsatile flow conditions by circulating, at different flow rates, four types of fluid (porcine whole blood, porcine whole blood with a US contrast agent, porcine blood cell suspension and porcine blood cell suspension with a US contrast agent). A total of 120 measurements were performed. Computational simulations of the fluid dynamics in the vicinity of the axisymmetrical stenosis were performed with finite-element modeling (FEM) to locate and identify the PDA signal loss due to the wall filter of the US instrument. The performance of three segmentation algorithms used to delineate the vessel lumen on the PDA images was assessed and compared. It is shown that the type of fluid flowing in the phantom affects the echoicity of PDA images and the accuracy of the segmentation algorithms. The type of flow (steady or pulsatile) and the flow rate can also influence the PDA image accuracy, whereas the use of US contrast agent has no significant effect. For the conditions that would correspond to a US scan of a common femoral artery (whole blood flowing at a mean pulsatile flow rate of 450 mL min(-1)), the errors in the percentages of area reduction were 4.3 +/- 1.2% before the stenosis, -2.0 +/- 1.0% in the stenosis, 11.5 +/- 3.1% in the recirculation zone, and 2.8 +/- 1.7% after the stenosis, respectively. Based on the simulated blood flow patterns obtained with FEM, the lower accuracy in the recirculation zone can be attributed to the effect of the wall filter that removes low flow velocities. In conclusion, the small errors reported in vitro may support the clinical use of this technique.