CHICAGO — A second multicenter trial has shown that noninvasive CT angiography is highly accurate in assessing coronary artery disease when compared with conventional invasive angiography.

The per-vessel negative predictive value of 64-slice coronary CT angiography (CCTA) was 97% for identifying blockages greater than 50%, and 99% for blockages greater than 70%, when measured in 232 patients with typical or atypical chest pain in the Assessment by Coronary Computed Tomographic Angiography of Individuals Undergoing Invasive Coronary Angiography (ACCURACY) trial. Positive predictive values were 51% and 33%, respectively, Dr. James K. Min and his associates reported at the annual meeting of the Radiological Society of North America.

“The ACCURACY results [obtained] in a prospective, multicenter fashion definitively establish the high diagnostic accuracy and high negative predictive value of 64-detector-row CT angiography in chest pain patients with intermediate prevalence of coronary artery disease,” said Dr. Min, director of the cardiac CT laboratory at New York-Presbyterian Hospital, New York City.

The findings echo those of the recent Coronary Artery Evaluation Using 64-Row Multidetector Computed Tomography Angiography (CORE-64) trial, in which CT angiography had a 91% positive predictive value and an 83% negative predictive value for identifying significant coronary artery stenoses. CORE-64 was the first large, multicenter trial of the 64-slice technology for coronary angiography, but was criticized by some attendees at the annual scientific sessions of the American Heart Association where it was presented. Concerns were raised that the radiation dose from repeated CT scans could pose a potential cancer risk. No such concerns were raised at the radiology meeting.

To reduce the amount of radiation given to patients in the ACCURACY trial, investigators used a radiation-dose reduction algorithm called EKG modulation that reduces CT angiography radiation by about 40%, Dr. Min said in an interview. The radiation dose per patient was about 10–15 millisieverts (mSv), which is about twice that of an invasive coronary angiogram and about half that of a noninvasive thallium stress test.

Since the trial began, a new algorithm called perspective axial gating has been commercially released and is reported to reduce exposure by 90%, to about 2–4 mSv. Both algorithms work by activating the CT scanner during select parts of the cardiac cycle only, Dr. Min said. For comparison, New York City residents are exposed to about 3 mSv of radiation annually through background exposure.

Neither study used CT angiography for screening. “I believe very emphatically that the data to date don't support CT angiography as a screening tool at all,” Dr. Min said. “In asymptomatic patients, we don't have any data of what to do with the results, and if treatment benefits them.”

CT angiography is of greatest benefit for patients without known coronary disease who have low or intermediate pretest risk. “If you have a high pretest suspicion that someone has coronary artery disease, then direct progression to invasive coronary angiography or even myocardial perfusion imaging is probably a better alternative,” he said.

The ACCURACY trial was unique in that it included all coronary artery segments in its analysis and all patients irrespective of their baseline coronary calcium score. In the CORE-64 trial, stented segments were excluded, as were patients with a calcium score higher than 600. As a result, the ACCURACY findings of high diagnostic accuracy are even more impressive and representative of actual clinical usage, Dr. Min said.

Between May 2006 and January 2007, ACCURACY investigators performed CCTA prior to conventional quantitative coronary angiography (QCA) on 232 patients who had typical or atypical chest pain and had been referred for evaluation at 16 U.S. centers. The images were obtained on a GE Healthcare LightSpeed VCT CT scanner, and analyzed at 15 different locations throughout the coronary tree. The investigators used equipment made by GE Healthcare, which sponsored the study. Dr. Min is on the speakers' bureau for GE Healthcare.

Three independent radiologists interpreted the CCTA images, and one independent radiologist interpreted the QCA images. No segments were excluded based on nonagreement between readers, and only one segment was debated among readers, Dr. Min said.

The patients' mean age was 57 years (range 31–82 years); 138 were male, 203 were white, and 13 were black, and their average body mass index was 31 kg/m

QCA detected 82 blockages greater than 50% in 55 patients and 31 blockages greater than 70% in 34 patients.

For noninvasive CCTA, per-patient sensitivity was 93% and specificity was 82% for blockages greater than 50%; sensitivity was 91% and specificity was 84% for blockages greater than 70%, Dr. Min said.

Additional analyses of the ACCURACY data are being conducted, including a comparison of CT angiography to historical single-photon emission computed tomography (SPECT) imaging, cost-effectiveness versus standard of care, incremental benefit of CT angiography beyond traditional coronary calcium scoring, and interreader, interpatient, and intersegment reliability.

Accuracy of Noninvasive CT Angiography Supported by Trial