Document Type
Article
Publication Date
3-1-2011
Abstract
RATIONALE AND OBJECTIVES: To compare computer-generated interpretation of coronary computed tomography angiography (cCTA) by commercially available COR Analyzer software with expert human interpretation.
MATERIALS AND METHODS: This retrospective Health Insurance Portability and Accountability Act‑compliant study was approved by the institutional review board. Among 225 consecutive cCTA examinations, 207 were of adequate quality for automated evaluation. COR Analyzer interpretation was compared to human expert interpretation for detection of stenosis defined as ≥50% vessel diameter reduction in the left main, left anterior descending (LAD), circumflex (LCX), right coronary artery (RCA), or a branch vessel (diagonal, ramus, obtuse marginal, or posterior descending artery).
RESULTS: Among 207 cases evaluated by COR Analyzer, human expert interpretation identified 48 patients with stenosis. COR Analyzer identified 44/48 patients (sensitivity 92%) with a specificity of 70%, a negative predictive value of 97% and a positive predictive value of 48%. COR Analyzer agreed with the expert interpretation in 75% of patients. With respect to individual segments, COR Analyzer detected 9/10 left main lesions, 33/34 LAD lesions, 14/15 LCX lesions, 27/31 RCA lesions, and 8/11 branch lesions. False-positive interpretations were localized to the left main (n = 16), LAD (n = 26), LCX (n = 21), RCA (n = 21), and branch vessels (n = 23), and were related predominantly to calcified vessels, blurred vessels, misidentification of vessels and myocardial bridges.
CONCLUSIONS: Automated computer interpretation of cCTA with COR Analyzer provides high negative predictive value for the diagnosis of coronary disease in major coronary arteries as well as first-order arterial branches. False-positive automated interpretations are related to anatomic and image quality considerations.
Recommended Citation
Halpern, Ethan J and Halpern, David J, "Diagnosis of coronary stenosis with CT angiography comparison of automated computer diagnosis with expert readings." (2011). Department of Radiology Faculty Papers. Paper 12.
https://jdc.jefferson.edu/radiologyfp/12
PubMed ID
21215663
Comments
This article has been peer reviewed. It is the authors' final version prior to publication in Academic Radiology. Volume 18, Issue 3, March 2011, Pages 324-333. The published version is available at DOI: 10.1016/j.acra.2010.10.014. Copyright © Elsevier Inc.