A New Algorithm for Deriving Pulsatile Blood Flow Waveforms Tested Using Stimulated Dynamic Angiographic Data

In vascular pathology the assessment of disease severity and monitoring of treatment requires quantitative and reproducible measurements of arterial blood flow. We have developed a new technique for processing sequences of dynamic digital X-ray angiographic images. We have tested it using computer simulated angiographic data which includes the effect of pulsatile blood flow and X-ray quantum noise. A parametric image was formed in which the image grey-level represents dye concentration as a function of time and distance along a vessel segment. Adjacent concentration--distance profiles in the parametric image were re-registered along the vessel axis until a match occurred. A match was defined as the point where the sum of squares of the differences in the two profiles was a minimum. The distance translated per frame interval is equal to the bolus velocity. We have tested several contrast medium injection methods including constant flow and a range of discrete pulses per second. The technique proved to be robust and independent of injection technique. Average blood flow was measured for simulated pulsatile waveforms with mean flows of up to 650 ml/min (peak velocities up to 186 cm/s) in a range of diameters from 2 mm to 6 mm. The standard deviation of the error in the mean flow estimates over the whole range of velocities and vessel sizes was +/- 1.4 cm/s.