doi: 10.1161/CIRCIMAGING.108.792572
Original Articles |
Diagnostic Accuracy and Impact of Computed Tomographic Coronary Angiography on Utilization of Invasive Coronary Angiography
From the Department of Medicine (Cardiology), University of Ottawa Heart Institute (B.J.W.C., A.A., G.A.W., L.C., T.D.R., Y.Y., N.G., R.S.B.), Ottawa, Ontario, Canada; Department of Radiology, Ottawa Hospital (B.J.W.C., T.D.R., C.D., R.S.B.), Ottawa, Ontario, Canada; and Department of Cardiac Sciences, University of Calgary (P.D.G.), Calgary, Alberta, Canada.
Correspondence to Benjamin Chow, MD, FRCPC, University of Ottawa Heart Institute, 40 Ruskin Street, Ottawa, Ontario, Canada K1Y 4W7. E-mail bchow{at}ottawaheart.ca
Received May 16, 2008; accepted November 19, 2008.
Background— Computed tomographic coronary angiography (CTA), given its high negative predictive value, is a potential gatekeeper for invasive coronary angiography (ICA). Before CTA can be further accepted into clinical practice, its impact on healthcare resources needs to be better understood. We sought to determine the clinical impact of CTA on ICA referrals, CTA accuracy, and normalcy rate.
Methods and Results— To determine the impact of CTA, consecutive patients (n=7017) undergoing ICA before and after implementing a dedicated cardiac CT program were reviewed and compared with 3 other centers (n=11 508). To determine CTA accuracy, we evaluated consecutive CTA patients who underwent ICA. For normalcy rate, we identified patients with a low pretest probability for obstructive coronary artery disease. With the implementation of a cardiac CT program, the frequency of normal ICA decreased from 31.5% (1114 of 3538 patients) to 26.8% (932 of 3479 patients) (P<0.001). These findings were significantly different (P=0.003) from the 3 centers, in which normal ICAs were unchanged (30.0% [1870 of 6224 patients] to 31.0% [1642 of 5284 patients]). CTA had excellent per-patient sensitivity (99% [CI, 95% to 100%]), positive predictive value (92% [CI, 86% to 96%]) and negative predictive value (95% [CI, 72% to 100%]). Because of referral bias, specificity (64% [CI, 44% to 81%]) was low; however, the normalcy rate of CTA was 94% (CI, 90% to 97%). After adjusting for referral bias, the adjusted sensitivity was 90% (CI, 89% to 91%), and the adjusted specificity was 95% (CI, 94% to 96%), with positive and negative predictive values of 92% (CI, 91% to 93%) and 93% (CI, 92% to 94%), respectively.
Conclusion— The clinical implementation of CTA appears to positively impact ICA by reducing the frequency of normal ICA. The operating characteristics of CTA support its potential role as a tool useful in ruling out obstructive coronary artery disease.
Key Words: computed tomography • coronary angiography • clinical impact • accuracy • normalcy
CLINICAL PERSPECTIVE
Related Articles
- CT Angiography: First Things First
- Michael S. Lauer
Circ Cardiovasc Imaging 2009 2: 1-3.[Extract] [Full Text] [PDF]
- Diagnostic Accuracy and Impact of Computed Tomographic Coronary Angiography on Utilization of Invasive Coronary Angiography
- Benjamin J.W. Chow, Arun Abraham, George A. Wells, Li Chen, Terrence D. Ruddy, Yeung Yam, Nayia Govas, Phoebe Diane Galbraith, Carole Dennie, and Rob S. Beanlands
Circ Cardiovasc Imaging 2009 2: 16-23.[Abstract] [Full Text] [PDF]