CT angiography: first things first.

Investigators from leading centers now promote CT angiography as a major advance in diagnostics, making it possible for many patients to avoid invasive coronary angiography.2 They provide 3 lines of evidence to make their case. First, in a number of cross-sectional patient-oriented studies, the diagnostic accuracy seems to be high, with claims of sensitivity and negative predictive value approaching or exceeding 95%.3 The most recent high-quality multicenter study using the more advanced 64-slice technology yielded less sanguine findings, though, with a reported sensitivity of 85% and a negative predictive value of only 83%; the investigators themselves concluded that “CT angiography cannot replace conventional coronary angiography at present.”4 Second, preliminary investigations show that findings on CT angiography predict risk of major coronary events.5 Although not surprising, it is not clear whether prognostic information provided by CT is more or less accurate than that provided by conventional noninvasive tests, like measurement of exercise capacity or detection of myocardial scar or inducible ischemia by perfusion imaging or ultrasonography. The third line of evidence is illustrated by the careful clinical epidemiological study presented by Chow et al6 in the current issue of Circulation Imaging. The investigators analyzed the proportion of normal findings among patients undergoing invasive coronary angiography at a single large Canadian institution before and after establishment of a cardiac CT program. Before the CT program was implemented, 32% of patients had normal findings on invasive angiography. The proportion decreased by a modest amount to 27% after the CT program was started. In 3 other geographically remote centers where cardiac CT was not offered, the proportion of normal coronary angiograms remained essentially constant, at 30% to 31%, during the same calendar period. The investigation presented by Chow et al has a number of strengths. Because procedures like conventional and CT angiography are regionalized in Canada, the investigators were able to effectively perform a population-based outcomes study,7 even though they collected data from only 1 institution. Not all major centers in Canada adopted the new CT technology, making it possible to perform a properly controlled quasi-experimental analysis. This type of investigational approach has been used to assess the impact of population-based strategies like smoking bans,8 and it has particular promise for estimating effects of interventions in which randomized trial evidence is absent. The authors argue that implementation of CT angiography reduces the frequency of normal invasive coronary angiograms, a conclusion that is reasonable given the robust data they present. Putting these 3 lines of evidence together, what can we say? CT angiography has high diagnostic accuracy, but not high enough to be considered a replacement for invasive coronary angiography. CT angiography can probably predict major cardiovascular events, but there is a much more robust literature for other noninvasive methods, in particular myocardial perfusion imaging.9 And now, given the data presented by Chow et al, we can say that implementation of CT angiography within health care systems reduces the prevalence of normal coronary angiograms, but only to a modest degree. Even with a CT angiography program, over 25% of invasive angiograms were normal. On all 3 lines of evidence there are reasons for pause, yet this technology is being widely adopted, even within Canada’s controlled health care system. What is wrong with this picture? When a patient sees a physician complaining of symptoms suggestive of coronary artery disease, what is his/her main concern? Is it what an invasive coronary angiogram, if done, would look like? We would hope not, as it is now well known that the coronary angiogram often misses disease likely to cause major clinical problems.10 Is the patient worried about the proportion of coronary angiograms performed at the local hospital that show normal findings? Doubtful. Does he/she want to know the probability of suffering a major, possibly fatal, cardiovascular event if natural history were allowed to take its course? Or is he/she, working with his/her doctor, asking a more fundamental question—is there something to worry about, and if so, is there something that can be done that we can confidently say will reduce the probability of a major clinical event. In other words, will this test help? The opinions expressed in this article are not necessarily those of the editors or of the American Heart Association. The views expressed are those of the author and do not reflect the views of the NHLBI, NIH, or the US Department of Health and Human Services. From the Division of Prevention and Population Sciences, National Heart, Lung, and Blood Institute (NHLBI), Bethesda, Md. Correspondence to Michael S. Lauer, MD, FACC, FAHA, 6701 Rockledge Dr, Room 10122, Bethesda, MD 20892. E-mail [email protected] (Circ Cardiovasc Imaging. 2009;2:1-3.) © 2009 American Heart Association, Inc.