A method was developed to produce a composite image of microvascular networks with grayscales proportional to volumetric flows. Velocities in arterioles and venules were assessed with a high-resolution laser Doppler imager (LDI). The vascular structures were quantified from the micrograph with a computerized vessel detection algorithm. After registering the detected vascular network with the LDI scan, volumetric flows were calculated along the centerlines of the vessels. In vivo data were obtained from the hamster cheek pouch in 6 studies. Flow continuity of the flow map was evaluated by comparing the main flow (Q) with the sum of branch flows (Qs), averaging over the respective vessel segments incident to each bifurcation. The method was reproducible across the 6 studies with the correlation coefficient (r) between Qs and Q ranging from 0.913 to 0.986. In all, over 20,000 flow estimates from 360 vessel segments (24-160 microm in diameter) at 166 bifurcations were analyzed. With flow normalized between 0 and 1, the linear regression yielded: Qs = 1.03 Q + 0.006; r = 0.952, n = 166, P < 0.0005. The bimodality imaging method exploits a large amount of velocity and diameter data, and therefore should be useful for studying heterogeneous flows in the microvasculature.
2004 Wiley-Liss, Inc.