Precision and reproducibility of quantitative coronary angiography with applications to controlled clinical trials. A sampling study.

Abstract

Most computer methods that quantify coronary artery disease from angiograms are designed to analyze frames recorded during the end-diastolic portion of the cardiac cycle. The purpose of this study was to determine if end diastole is the best portion of the cardiac cycle to sample, or if other sampling schemes produce more precise and/or reproducible estimates of coronary disease. 20 cinecoronary angiograms were selected at random from a controlled clinical trial testing the effects of plasma lipid lowering on atherosclerosis. Sampling schemes included sequential and random sampling of two to five frames within the complete cardiac cycle, systole, and diastole. Three vessel measures and percent stenosis were evaluated for each sampling scheme. From the sampling experiment, it was determined that sampling sequentially end diastole yielded the most precise estimates (i.e., exhibiting minimum variability within a cycle) of the vessel measures. With regard to reproducibility (i.e., similar values across cycles), sampling randomly within the cycle was best. Overall, the average diameter of a vessel segment was the most precise and the most reproducible of the measures. Sample size calculations are given for each of these measures under the best sampling scheme.

Evaluation of human panelists in assessing coronary atherosclerosis.

“…However, intrusive lesions are an appropriate (and possibly the prime) end-point measurement because 1) enumeration of focal lesions in coronary angiograms significantly predicts long-term prognosis 26 and 2) the extent of raised coronary lesions is the single most significant anatomic finding predictive of the presence of ischemic heart disease at autopsy. 27 The results reported here for human panel readers, taken with our published estimates of the reproducibility of image-processed lesion measurements in the same films, 10 lead to the conclusion that the capabilities of selective angiography are optimized when films are first examined by human panel readers and next with image processing.…”

“…8 " 10 The reliability of film evaluation by humans for changes in coronary status is influenced by 1) technical features, such as film quality, the conditions for film reading, the number of films a reader has recently evaluated, and order of film presentation and 2) coronary pathological and anatomic characteristics, such as of coronary status (including bypass grafts). In this situation, as in other controlled studies dealing with humans, analyses regarding the validity of the angiographic interpretation (by comparisons with a known "true" value) are not technically or ethically possible.…”

Imaging Biomarkers in Atherosclerosis Trials

“…Technical limitations to this methodology include the inherent limitations of x-ray imaging with reduced contrast and sharpness, image digitization noise, and quantization errors in the analog-to-digital conversion process. 38 Biological limitations include differences in vasomotor tone, vessel motion, myocardial blush, filling of the vasa vasorum, and insufficient blood-contrast mixing. 38 Although coronary artery lumen stenoses of 70% or more often cause myocardial ischemia and are targets for revascularization, acute coronary and cerebrovascular syndromes frequently are caused by the rupture of plaques that cause stenosis less than 50% and are not flow limiting.…”

“…38 Biological limitations include differences in vasomotor tone, vessel motion, myocardial blush, filling of the vasa vasorum, and insufficient blood-contrast mixing. 38 Although coronary artery lumen stenoses of 70% or more often cause myocardial ischemia and are targets for revascularization, acute coronary and cerebrovascular syndromes frequently are caused by the rupture of plaques that cause stenosis less than 50% and are not flow limiting. 39 These smaller culprit plaques are underestimated by angiography, which relies on measuring the size of the vessel lumen alone.…”

Brachial artery vasoreactivity is associated with cross-sectional and longitudinal anatomical measures of atherosclerosis in postmenopausal women with coronary artery disease

“…Computer software developed to track the edges of coronary segments was used to obtain a series of measurements for each pair of arterial segments from baseline and final angiograms using identical views of the coronary arteries for the baseline and end-of-trial measures [21][22][23]. Measurements were predominantly derived from diameter measurements and were averaged over three sequential frames [24]. These measurements were made at the site of lesions previously identified by a panel of angiographers, the imaging analyst, or both.…”