Reproducibility of Digital Subtraction Fluoroscopic Readings for Coronary Artery Calcification

Abstract

Digital subtraction fluoroscopy, an inexpensive screening test for coronary atherosclerosis, is highly sensitive in detecting coronary calcifications. However, no previous study has reported interobserver agreement for this test. Six hundred and thirty-one subjects underwent digital subtraction fluoroscopy in the 60 degrees left anterior oblique projection. Images were acquired with pulsed fluoroscopy at 15 frames per second. An averaged mask was subtracted from successive images. These fluoroscopic images were stored on a digital disk and replayed in cine loop format. An observer, blinded to clinical information, read the fluoroscopic studies for the presence of calcium in the left main-left anterior descending artery, circumflex artery, and right coronary artery. The images were then stored on digital tape and reread by a second blinded observer. The percentages of interobserver agreement regarding the presence and absence of calcium in left main-left anterior descending, circumflex, and right coronary arteries, were 91.9%, 92.9%, and 92.2%, respectively. The overall kappa values, which are 0.85, 0.77, and 0.82 in left main-left anterior descending, circumflex, and right coronary arteries, respectively, show a highly significant level of agreement (P < .0001). Digital subtraction fluoroscopy is a reliable screening test for coronary calcifications.