Experimental validation of quantitative coronary arteriography for determining pressure-flow characteristics of coronary stenosis.

The applicability of classic fluid dynamic equations to tapering stenoses in vasoactive, flexible coronary arteries in vivo and the validity of quantitative coronary arteriography was tested by comparing experimentally measured and x-ray-predicted pressure gradients at equal flows for left circumflex stenoses in five dogs chronically instrumented with a balloon. occluder, flow probe, and proximal and distal catheters for.injection of contrast media or recording distal coronary pressure. Arterial borders on cut-film orthogonal arteriograms were digitized and computer processed into a three-dimensional reconstruction of the stenosis. The total pressure gradient was calculated from stenosis dimensions using classic fluid dynamic equations. Over the full range of flow, the correlation of x-ray-predicted and experimentally measured pressure gradient for 51 separate stenoses was y = 1.11x ± 0.75, r = 0.95, p < 0.001, with a standard deviation about the regression line of 9.4 mm Hg and with 95% of x-ray-predicted values falling within ± 18.5 mm Hg of the experimentally measured values (95.% confidence limits). Mean experimentally measured and x-ray-predicted pressure gradients were 10.1 7.7 mm Hg (± SD) and 10.9 ± 5.6 mmn Hg at low flow and 48.2 ± 23.1 mm Hg and 55.8 ± 28.8 mm Hg at high flow, respectively. The mean difference was 3.9 ± 4.3 mm Hg at rest flow and 11.9 ± 10.5 mm Hg at high flow. For all data over the entire range of flows, the frequency distribution of differences between x-ray-predicted and experimentally measured gradients was a bell-shaped curve with a peak, or mean difference, of ± 4 mm Hg, a standard deviation of ± 9.8 mm Hg and 95% confidence limits for individual values of ± 19.6 mm Hg. These data demonstrate the validity of applying classic fluid dynamic theory to tapering stenoses in vivo. Quantitative coronary arteriography on the average or in individual instances approximately predicts the pressure gradient-flow characteristics of coronary arterial stenoses in intact animals. However, as indicated by the above measures of variability, there is considerable scatter in x-ray predictions that limits its applicability to individual clinical cases. We believe that this scatter is a result of difficulties in visually tracing arterial borders on arteriograms and can most likely be reduced by automatic border recognition techniques.