Comparisons of angiographic core laboratory analyses of phantom and clinical images: interlaboratory variability

Centralized, quantitative coronary analysis (QCA) has become the standard for determining change in coronary anatomy in clinical investigations. QCA systems and laboratory methods, however, vary among core facilities, and analysis variability among angiographic core laboratories (ACL) has not be studied. We evaluated QCA accuracy and variability among active ACL, using differing QCA systems by comparing analyses of phantom and clinical cinefilm images. Automated, unedited analyses were performed on images of 11 plexiglass phantom lumens (0.67-5.05 mm) acquired under varying radiographic conditions. Analysis differences from actual luminal diameters ranged widely (+0.42 - (-)0.45 mm) among ACL. Measurement of diameters < 1.0 mm were overestimated and diameters > 3.0 mm were underestimated. Measurements of midrange diameters (> 1.0 mm and < 3.0 mm) were most comparable among ACL (93% within +/- 0.2 mm). Clinical image analysis was performed using differing QCA systems and laboratory methodology on 11 randomly selected study films. Comparative analyses revealed significant variability between laboratories in the assessment of minimal lumen diameter (0.22 +/- 0.38 mm P < 0.05). These data describe analysis variability among ACL and demonstrate a need for establishing ACL performance standards.