Variability in measures of coronary lumen dimensions using quantitative coronary angiography.

OBJECTIVES: The purpose of this study was to determine the true total variability of quantitative coronary angiographic measures and their components in the clinical setting. BACKGROUND: Many studies describe quantitative coronary angiographic variability on the basis of repeated quantitative coronary angiographic measures from the same cineangiogram. Although these studies characterize well the performance of quantitative coronary angiographic analysis methods, they do not include other potentially important sources of variability in results of this procedure, such as day to day variations in patients and equipment or variability in selection of frames for analysis. METHODS: Coronary angiograms from 20 patients who underwent diagnostic angiography followed by percutaneous transluminal coronary angioplasty an average of 2.9 days later were reviewed. A total of 30 lesions well visualized in both films were analyzed multiple times using an automated first-derivative edge-detection quantitative coronary angiographic technique. RESULTS: The coefficient of variation for quantitative coronary angiographic measures of the same lesions from separate angiograms ranged from 8.11% to 14.01%. Average diameter was the least variable and percent diameter stenosis the most variable. Day to day variations in the patient, procedure and equipment accounted for an average of 30% of the total variability. Of the remaining variability, only 13.26% was due to variability in frame selection. CONCLUSIONS: These results provide useful information for planning clinical studies using quantitative coronary angiography, identify areas where additional improvements in this technology are needed and define more clearly the applicability of quantitative coronary angiography in the setting of routine clinical practice.