64-Slice CT for Detecting CAD
Abstract & Commentary
By Michael H. Crawford, MD Dr. Crawford is the Professor of Medicine, Chief of Clinical Cardiology, University of California, San Francisco; and is the Editor of Clinical Cardiology Alert.
Synopsis: In 94%, interpretable images were found. Accuracy was 95% for sensitivity, 96% for specificity, 97% for positive predictive value, and 92% for negative predictive value for lesions with > 50% stenosis.
Source: Fine JJ, et al. Comparison of Accuracy of 64-Slice Cardiovascular Computed Tomography with Coronary Angiography in Patients with Suspected Coronary Artery Disease. Am J Cardiol. 2006;97:173-174.
The new 64-slice cardiovascular computed tomography (CVCT) is theoretically superior to the old 16-slice CT for detecting coronary artery disease (CAD). Thus, Fine and colleagues studied 66 consecutive patients who underwent CVCT and coronary angiography to diagnose obstructive CAD. Both procedures were performed within 30 days of each other. Patients with high heart rates were given beta-blockers to reduce their heart rate to 50-60 beats per minute for CVCT. The results of both studies were read independently by readers blinded to the other test results. A lesion of > 50% stenosis was considered significant. Also, left ventricular ejection fraction (EF) was determined by both techniques. Vessels < 1.5 mm in diameter were excluded from analysis, since CVCT is unreliable at this level of resolution. The angiography was considered the gold standard.
Results: Two hundred forty-five coronary arteries were evaluated in the 66 patients. CVCT was of diagnostic quality in 62 of the 66 (94%). In comparison to angiography, the sensitivity of CVCT was 95%, specificity 96%, positive predictive value 97%, and negative 92%. Vessel-by-vessel accuracy was 98% for the left main, 93% for the left anterior descending, 92% for the circumflex, and 92% for the right. Mean EF was identical for both techniques (59%). Fine et al concluded that CVCT is a reliable diagnostic test for suspected obstructive CAD.
Clearly, the new 64-slice CVCT is superior to the older 16-slice technique for detecting significant CAD based upon previous studies; 16-slice CT was not done in this study. The issue is whether it is good enough to forgo angiography, unless catheter-based therapy is indicated. This means how good is it at identifying normal vessels and mild disease, which hangs on the negative predictive value. Unfortunately, this has not changed with CVCT vs 16-slice CT; it is still about 92%. Is this good enough to send a chest pain patient home from the emergency department? I spoke to our seasoned, internal medicine oriented (as compared to surgical) ED doctors, and they said they were uncomfortable with this degree of accuracy until a clinically oriented trial is done. Such a trial would evaluate outcomes after CVCT to see what the myocardial infarction/death rate is with a normal or mild disease CVCT. On the other hand, a positive study for significant CAD may speed up getting the patient revascularized without the delays built into stress testing. Also, the positive predictive value is much higher at 97%, so referral to angiography is an acceptable approach.
There are a few caveats to CVCT. First, in about 5% it is of too poor a quality to interpret. Second, it requires a heart rate < 60 beats/min for the accuracy reported in this study. Patients who are hypotensive and tachycardic would not be good candidates. Third, the radiation exposure is considerable and higher than the average coronary angiogram. So is CVCT ready for prime time? Maybe not, but it is being installed or used all over the United States as you read this. It would be ideal if outcome study data were forthcoming soon.