Variabilities in measurement of coronary arterial dimensions resulting from variations in cineframe selection

Publisher Website
Google Scholar
Abstract
To quantitatively analyze a coronary arterial segment from a cineangiogram, an enddiastolic or neighboring cineframe is usually selected, such that a possibly existing coronary lesion is visualized optimally, as judged by the cardiologist. However, different cardiologists may select different (although usually neighboring) frames, even when following the same selection criteria. It is also possible that the frames are selected from different cardiac cycles. In this study the effects of such phase shifts on the reproducibility of the quantitative measurements were studied. In a total of 38 consecutive patient films obtained at a filmspeed of 25 frames/sec, the frame O demonstrating the severity of a lesion optimally, as judged by a senior cardiologist, the three preceding frames, the three following frames and one frame exactly one cycle prior to or following frame O were selected; frame O was always chosen in the end‐diastolic phase of the cardiac cycle. In each film one coronary arterial segment with a focal lesion was analyzed quantitatively in these eight frames with the Cardiovascular Angiography Analysis System (CAAS). No significant differences were found in the mean difference and the standard deviations of the differences (variabilities) in the obstruction diameter, interpolated reference diameter, percent diameter stenosis, extent of the obstruction and area of atherosclerotic plaque obtained in the various frames with respect to frame O. Therefore, it may be concluded that the selection of a cineframe for quantitative analysis in the end‐diastolic phase of the cardiac cycle is not very critical; in other words, the obstruction measurements are not time‐dependent for frames in the end‐diastolic phase. It is argued that the quality of mixing of the contrast agent in the arterial segment is a major source of the observed variations; filling artefacts are potentially present in each of the selected frames.

This publication has 13 references indexed in Scilit: