In coronary atherosclerosis, the arterial lumen size and shape can be markedly irregular, eccentric and variable. Traditional angiographic interpretation, emphasizing percent diameter stenosis, has been criticized as an inadequate descriptor of such diseased arteries. Computerized quantitative angiographic technologies, yielding a true lumen area measurement, may be superior. High-frequency epicardial echocardiography (HFEE) is a technique that allows on-line evaluation of coronary arterial wall and lumen at the time of cardiac surgery. It has been extensively validated and yields accurate measurements of normal and diseased coronary lumen areas. This study compares quantitative coronary angiography (QCA) estimates of lumen area to those obtained by HFEE to determine if the computerized angiographic method more accurately predicts residual luminal area than traditional angiographic percent diameter stenosis measurements. Although actual luminal morphology was quite variable, there was a good correlation between lumen areas determined by HFEE versus QCA: r = 0.85, n = 67, HFEE = 0.8 QCA - 0.1 (HFEE 4.0 +/- 0.30 mm2, mean +/- SEM range 0.3 to 14.0; QCA 5.1 +/- 0.40 mm2, range 0.7 to 11.8). Percent diameter stenosis determined from the angiograms did not correlate well with HFEE or QCA measurements of residual luminal area. Separation of "normal" arterial segments (defined as < 25% diameter stenosis) from "abnormal" segments (> 50% diameter stenosis) by angiography did not agree with lumen areas as defined by either HFEE or QCA. Better separation occurred when QCA-determined luminal areas were used to separate normal from abnormal arterial segments.(ABSTRACT TRUNCATED AT 250 WORDS)