Quantitation of aortic regurgitation by computer analysis of digital subtraction angiography

Abstract

Digital subtraction angiography provides the potential to determine aortic regurgitant fraction by computer analysis of time-intensity curves generated from regions of interest positioned over the aorta and left ventricle after aortography. To validate this ability, we studied six dogs instrumented with an electromagnetic flow probe on the ascending aorta. Aortic regurgitation of varying severity was produced by a basket catheter introduced through the right carotid artery. Aortograms were performed using continuous fluoroscopy at 30 frames/s and stored in digital format in a 256 x 256 pixel matrix. An image-processing computer was utilized to plot summated pixel intensity versus time for both the aortic and the left ventricular regions of interest. Regurgitant fraction was calculated from the time-intensity curves using an algorithm analogous to that employed by dye-dilution methods. Regurgitant fraction determined from digital angiography was compared with that obtained by electromagnetic flow and was found to correlate well (r = 0.94, SEE = 7.4%) over a wide range of values. Thus, these data indicate that aortic regurgitant fraction can be accurately determined from computer analysis of digitally acquired aortograms in an animal model of acute aortic regurgitation.