ORLANDO — CT angiography remains controversial after a report on the diagnostic accuracy of multidetector CT measured in 316 patients.
Noninvasive multidetector CT (MDCT) angiography had a 91% positive predictive value and an 83% negative predictive value, compared with conventional angiography, for identifying significant coronary stenoses in the “first large, multicenter trial of the 64-slice technology” for coronary angiography, Dr. Julie M. Miller said at the annual scientific sessions of the American Heart Association.
“This represents a high degree of diagnostic accuracy” for MDCT angiography, said Dr. Miller, a cardiologist at Johns Hopkins University, Baltimore.
But Dr. Miller's report of a high correlation between noninvasive and invasive angiography was immediately followed at the podium by strong comments from the invited discussant for the study, Dr. Michael S. Lauer. After criticizing the study's clinical relevance and questioning the technique's safety, he called for a moratorium on using MDCT angiography until its value for patients was better defined.
“This is a technology with no evidence of benefit and real concern for harm,” said Dr. Lauer, a cardiologist and director of the division of prevention and population science at the National Heart, Lung, and Blood Institute. His safety concern was that the radiation dose from repeated scans could pose a significant cancer risk.
Routine use of MDCT should stop until results from large-scale, randomized studies prove that “this test saves lives or prevents heart attacks with an acceptable margin of safety,” Dr. Lauer said.
His sharp critique contrasted with the response to Dr. Miller's report from other experts. “Dr. Miller's study is important because it shows that MDCT angiography gives you coronary anatomy outside the catheterization laboratory,” said Dr. Daniel Mark, a cardiologist and professor of medicine at Duke University, Durham, N.C.
In an interview following Dr. Lauer's podium remarks, Dr. Miller contended that his criticism mostly applied to using CT to screen people for their coronary calcium scores rather than issues of MDCT angiography. She stressed that the study did not use MDCT for screening. A reliable, noninvasive image of coronary anatomy could help many patients avoid the cost and potential complications of catheterization, Dr. Miller said.
“Until now, there was no proof that MDCT angiography was accurate for making diagnoses. We proved that MDCT works.” The next step is to compare noninvasive angiography with other noninvasive tests, she said.
On the issue of safety, the radiation dose each patient received, about 14–15 mSv, is comparable to the dose from conventional coronary angiography and less than the exposure during a nuclear perfusion scan, Dr. Miller said. The contrast volume also was similar to conventional angiography.
The Coronary Evaluation Using Multidetector Spiral Computed Tomography Angiography using 64 Detectors (CORE-64) trial enrolled patients older than 40 years (median age 59 years) who had been referred for conventional, quantitative coronary angiography at nine centers in seven countries. The study used equipment made by and was sponsored by Toshiba. Dr. Miller has received research support from Toshiba.
The patients' hearts were scanned with a 64-slice MDCT device that takes images at 0.5-mm intervals. The MDCT examination first was analyzed to get each patient's calcium score, and the study continued with 316 patients who had calcium scores of less than 600.
The images obtained from these patients were then analyzed at 19 locations throughout the coronary tree. Each location was at least 1.5 mm in diameter, and stented segments were excluded. All stenoses that blocked more than 30% of a vessel were quantified, and lesions that were 50% stenotic or greater were counted as significant. Patients underwent conventional quantitative angiography within the next 30 days.
A receiver-operator curve analysis showed that 93% of the patients with significant stenoses identified by conventional angiography were also spotted using the noninvasive method, the study's primary end point, reported Dr. Miller. The patients had a 56% prevalence of having at least one coronary artery with a significant stenosis.
A second analysis compared the noninvasive and invasive methods on a per vessel basis, with 868 individual vessels evaluated. By this measure, MDCT angiography identified 91% of the individual coronary vessels with a significant stenosis, compared with catheterization angiography, and produced a positive predictive value of 82% and a negative predictive value of 89%.
“We can define which patients need revascularization,” Dr. Miller said in an interview.