Velocity measurement accuracy in optical microhemodynamics: experiment and simulation

• Published in: Physiological measurement 2012-10, Vol.33 (10), p.1585-1602
• Main Authors: Chayer, Boris, L Pitts, Katie, Cloutier, Guy, Fenech, Marianne
• Format: Article
• Language: English
• Subjects: Animals; Biomechanical Phenomena; blood; Blood Flow Velocity; Computational Biology; cross-correlation; depth of correlation; Erythrocytes - cytology; Hemodynamics; micro-particle image velicometry; microcirculation; Models, Biological; Optical Phenomena; red blood cell; Rheology - methods; Swine; velocity profile
• ISSN: 0967-3334; 1361-6579
• DOI: 10.1088/0967-3334/33/10/1585

Micro particle image velocimetry (µPIV) is a common method to assess flow behavior in blood microvessels in vitro as well as in vivo. The use of red blood cells (RBCs) as tracer particles, as generally considered in vivo, creates a large depth of correlation (DOC), even as large as the vessel itself, which decreases the accuracy of the method. The limitations of µPIV for blood flow measurements based on RBC tracking still have to be evaluated. In this study, in vitro and in silico models were used to understand the effect of the DOC on blood flow measurements using µPIV RBC tracer particles. We therefore employed a µPIV technique to assess blood flow in a 15 µm radius glass tube with a high-speed CMOS camera. The tube was perfused with a sample of 40% hematocrit blood. The flow measured by a cross-correlating speckle tracking technique was compared to the flow rate of the pump. In addition, a three-dimensional mechanical RBC-flow model was used to simulate optical moving speckle at 20% and 40% hematocrits, in 15 and 20 µm radius circular tubes, at different focus planes, flow rates and for various velocity profile shapes. The velocity profiles extracted from the simulated pictures were compared with good agreement with the corresponding velocity profiles implemented in the mechanical model. The flow rates from both the in vitro flow phantom and the mathematical model were accurately measured with less than 10% errors. Simulation results demonstrated that the hematocrit (paired t tests, p = 0.5) and the tube radius (p = 0.1) do not influence the precision of the measured flow rate, whereas the shape of the velocity profile (p < 0.001) and the location of the focus plane (p < 0.001) do, as indicated by measured errors ranging from 3% to 97%. In conclusion, the use of RBCs as tracer particles makes a large DOC and affects the image processing required to estimate the flow velocities. We found that the current µPIV method is acceptable to estimate the flow rate on the condition that the measurement takes place at the equatorial plane of the vessel. Otherwise, it is not an appropriate method to estimate the shape of the velocity profile.