Quantitative angiography during coronary angioplasty with a single angiographic view: A comparison of automated edge detection and videodensitometric techniques

Abstract

Little information is available on the reliability of coronary luminal measurements obtained from quantitative analysis of a single angiographic view, an approach that is central to the practical use of on-line quantitative angiography. In the present study we investigated the contribution of two different techniques of quantitative angiography, edge detection (ED) and videodensitometry (VD), to the application of this concept during coronary angioplasty. Forty-six balloon angioplasty procedures were included in this study, all of them performed in a stenosis located in the mid right coronary segment. This coronary location was chosen to optimize data collection on luminal morphology and to minimize the number of factors that may adversely affect quantitative analysis with both techniques. In all cases two orthogonal angiographic projections were obtained before, after balloon dilatation, and at follow-up. Correlation coefficients and differences between orthogonal measurements obtained with each technique were used to evaluate the agreement between orthogonal readings at every stage of the procedure. The obtained correlation coefficients and mean differences (MD) between orthogonal measurements were as follows; before percutaneous transluminal coronary angiography (PTCA), 0.67 (MD 0.01 ± 0.47 mm 2) and 0.57 (MD 0.05 ± 0.64 mm 2) for ED and VD, respectively (Pitman's test for SD, p < 0.05); after balloon dilatation, 0.32 (MD −0.56 ± 1.53 mm 2) and 0.53 (MD −0.15 ± 1.43 mm 2) for ED and VD, respectively (paired t test for MD, p < 0.05); and at follow-up 0.79 (MD −0.15 ± 0.97 mm 2) and 0.73 (MD 0.17 ± 1.16 mm 2) for ED and VD, respectively ( p = NS). The presence of coronary dissection did not influence the variability in measurements observed after balloon dilatation. A considerable variability between orthogonal cross-sectional area measurements obtained with ED and VD was observed at all stages of coronary angioplasty, a finding that does not support the clinical application of area measurements with ED or VD from a single view. Similar observations were made after the exclusion of angiographically evident dissections. However, after balloon dilatation the agreement between orthogonal area measurements was significantly better with VD than with ED. Our results provide new insights into the problems posed by coronary intervention with respect to the on-line angiographic assessment of such intervention and to the potential solution of these problems. With either of these two quantitative techniques, area measurements obtained from a single angiographic view should be interpreted with caution.