Abstract 19466: Optimal Number of Readers for Identification of Significant Coronary Artery Disease by Coronary Computed Tomographic Angiography
Background: While coronary computed tomographic angiography (CCTA) possesses generally high diagnostic accuracy to detect obstructive coronary artery disease (CAD) as compared to quantitative coronary angiography (QCA), prior studies have reported differences in diagnostic performance. As prior trials have employed different numbers of core laboratory readers, the optimal number of readers to detect CAD is unknown. We compared the diagnostic performance of 1, 2, 3 and 5 randomly-selected readers to detect obstructive and high-risk CAD.
Methods: We evaluated 50 patients randomly selected from 2 multicenter studies who underwent both 64-detector CCTA and QCA. Five blinded, experienced readers independently interpreted CCTA studies and assessed for the presence of any obstructive disease (>=50% stenosis) and any high-risk disease (left main, n=2; proximal left anterior descending, n=4; or 3-vessel disease, n=2). A blinded core lab performed QCA. For each patient, different random combinations of readers were selected, and comparisons were performed for diagnostic sensitivity and specificity.
Results: Mean age was 57 ±14 years; 62% were male. Using randomly-selected combinations of 1, 2, 3, or 5 readers (Table), there were no significant differences in sensitivity and specificity for detection of high-risk disease or any per-patient obstructive disease. Significant differences were noted between groups for detection of obstructive disease on a per-segment level, with the highest diagnostic performance observed with 3 or 5 readers.
Conclusion: The diagnostic performance of CCTA to detect obstructive or high-risk CAD is generally high although there are significant variations between alternate reader methodologies. Combinations of 3 or more readers provide the highest diagnostic performance.
- © 2010 by American Heart Association, Inc.