Angiographic core laboratory analyses of arterial phantom images: Comparative evaluations of accuracy and precision

Abstract

Quantitative coronary arteriography (QCA) is commonly regarded as a reproducible and accurate method of assessing coronary anatomy. Centralized, quantitative core laboratory analysis of clinical study images has consequently become the standard for determining interval change in coronary anatomy. QCA systems and laboratory methods, however, are known to vary among core facilities and the effect of such differences on the variability of quantitative assessments among angiographic core laboratories (ACL) has not been studied. We evaluated QCA variability among seven active ACL, using four differing QCA systems by comparing analyses performed on a common set of phantom cinefilm images. Phantom analyses were performed in an automated, un-edited fashion on images of a plexiglass arterial phantom containing eleven precision drilled lumens (0.68-5.06 mm). Phantom images were acquired under varying radiographic conditions (5″ and 7″ image intensifier magnifications at 70 and 90 kV with uniform (scatter medium) and non-uniform (patient anatomy) backgrounds). Over the range of phantom diameters, analysis differences from actual luminal dimensions ranged widely (+ 0.42 to −0.45 mm) among ACL.