On the measurement of the mean velocity of blood flow over the cardiac cycle using Doppler ultrasound.

A number of modern duplex scanners now have facilities for determining volumetric blood flow through intact vessels. The methods these machines use to arrive at an answer must presuppose a number of conditions which may not be met in practice. This paper examines the effect that nonuniform insonation of the target blood vessel (using continuous wave or wide-gate pulsed ultrasound) has on the mean velocity, as determined using mean, root mean square (RMS) and maximum frequency processors. It is shown that for a given beam/vessel geometry the error is dependent only on the shape of the mean component of the velocity profile which, depending on the arterial site, may be flat, parabolic or intermediate. Mean processors may overestimate the mean velocity of established flow by up to 33%, but this could be turned to advantage when it is impracticable to insonate a vessel uniformly. Maximum frequency processors are of value when either plug-flow or fully established flow of low pulsatility is present in the target vessel. In the first case the mean flow is the same as the maximum flow, while in the second the mean flow is half the time averaged maximum flow, irrespective of the size and shape of the ultrasound beam. RMS processors are probably best avoided in volumetric flow measurement applications.