A 16-center clinical trial has put to rest lingering questions about the ability of 64-slice cardiac CT angiography to diagnose coronary artery disease for intermediate-risk patients.
The ACCURACY Trial confirmed the positive findings of numerous single-center trials that preceded it, said first author Dr. Matthew J. Budoff, an associate professor of internal medical at Harbor-UCLA Medical Center in Los Angeles.
Based on the experience of 230 chest pain patients referred for cardiac catheterization, the prospective trial found that 64-slice coronary CTA accurately detected occlusive coronary artery disease. Its sensitivity rates at 50% and 70% thresholds for stenoses were 95% and 94%, respectively. The specificity rates were 83% for both thresholds of occlusion.
The evaluations of three blinded independent readers were 99% accurate for ruling out the presence of coronary artery disease.
"Prior to coronary CTA, we have never had a noninvasive test that had a negative predictive value that high," said coauthor Dr. James Min, an associate professor of medicine in the cardiology division of Weil Cornell
Medical College in New York City. "With virtual certainty you can exclude coronary artery disease, and those patients don't need to undergo unnecessary invasive procedures."
The Journal of the American College of Cardiology published the results online Aug. 28.
The study avoided methodological problems that led to questions about the validity of previous coronary CTA trials, Min said. Unlike earlier studies, the ACCURACY Trial included patients with high body mass indexes, high coronary artery calcium scores, and high heart rates. Nearly one in four intermediate-risk patients, about half the rate identified in previously published trials, had at least one significant stenosis causing a 50% or greater blockage of a coronary artery.
The trial assessed patients who had an intermediate risk of coronary artery disease, the class of patient for whom recently published appropriateness criteria suggest coronary CTA is best suited, Budoff said.
"It performed well in the correct population," he said.
The trial designers attempted to simulate real-world clinical conditions by including all coronary artery segments for consideration, according to Budoff. Readers used multiplanar reformatting, volume rendering, and other 3D postprocessing techniques to maximize the number of segments that could be evaluated.
On a per-vessel basis, 910 vessels and 3435 segments in 229 patients were available for interpretation. Coronary CTA correctly identified 73 of 89 vessels with 50% or greater stenosis and 32 of 39 vessels with 70% or greater stenosis.
The investigators found no differences in the results for obese patients and those with more normal body weights or for patients whose heart rates were less than 65 beats per minute compared with patients with high heart rates. Calcium scores over 400 reduced specificity significantly.
Because of the study design, it is not clear whether the results apply to patients with known coronary artery disease or to those presenting in an emergency room setting.
GE Healthcare sponsored the trial. Budoff and Min serve on the GE speakers' bureau. Imaging was performed on 64-slice GE Lightspeed VCT scanners.
The PICTURE study, another GE-sponsored trial, will be discussed this month at the 2008 American Heart Association meeting. It will attempt to determine if 64-slice CTA can supplant nuclear perfusion testing for establishing the presence of obstructive coronary disease, Budoff said.
—By James Brice
