Significance of automated stenosis detection during quantitative angiography. Insights gained from intracoronary ultrasound imaging.

BACKGROUND: Automated stenosis analysis is a common feature of commercially available quantitative coronary angiography (QCA) systems, allowing automatic detection of the boundaries of the stenosis, interpolation of the expected dimensions of the coronary vessel at the point of obstruction, and angiographically derived estimation of atheromatous plaque size. However, the ultimate meaning of this type of analysis in terms of the degree of underlying atherosclerotic disease remains unclear. We investigated the relationship between stenosis analysis performed with QCA and the underlying degree of atherosclerotic disease judged by intracoronary ultrasound (ICUS) imaging. METHODS AND
RESULTS: In 40 coronary stenoses, automated identification of the sites of maximal luminal obstruction and the start of the stenosis was performed with QCA by use of curvature analysis of the obtained diameter function. Plaque size at these locations also was estimated with ICUS, with an additional ICUS measurement immediately proximal to the start of the stenosis. Crescentlike distribution of plaque, indicating an atheroma-free arc of the arterial wall, was recorded. At the site of the obstruction, total vessel area measured with ICUS was 16.65 +/- 4.04 mm2, whereas an equivalent measurement obtained from QCA-interpolated reference dimensions was 7.48 +/- 3.30 mm2 (P = .0001). Plaque area derived from QCA data was significantly less than that calculated from ICUS (6.32 +/- 3.21 and 13.29 +/- 4.22 mm2, respectively; mean difference, 6.92 +/- 4.43 mm2; P = .0001). At the start of the stenosis identified by automated analysis, ICUS plaque area was 9.38 +/- 3.17 mm2, and total vessel area was 18.77 +/- 5.19 mm2 (50 +/- 11% total vessel area stenosis). The arterial wall presented a disease-free segment in 28 proximal locations (70%) but in only 5 sites (12%) corresponding to the start of the stenosis and none at the obstruction (P = .0001). At the site of obstruction, all vessels showed a complete absence of a disease-free segment, and the atheroma presented a cufflike or all-around distribution with a variable degree of eccentricity.
CONCLUSIONS: At the site of maximal obstruction, QCA underestimated plaque size as measured with ICUS. Atherosclerotic disease was consistently present at the start of the stenosis and was used as a reference site by automated stenosis analysis. At the start of the stenosis, ICUS demonstrated a mean 50 +/- 11% total vessel area stenosis, with a characteristic loss of disease-free arcs of arterial wall present in proximal locations. Thus, the site identified by automated stenosis analysis as the start of the stenosis does not represent a disease-free site but rather the place where compensatory vessel enlargement fails to preserve luminal dimensions, a phenomenon that seems related to the observed loss of a remnant arc of normal arterial wall.