64-slice CT can identify presence of significant lesions

Noninvasive imaging could be suitable means for rapid triage of patients presenting with chest pain or equivocal stress test results.

Although its accuracy in determining lesion severity was limited, the 64-slice computed tomography scanner was highly accurate in establishing the presence of coronary disease in a recent study.

“We have a dichotomy with this technology where we can either say that the patient has coronary disease, and probably identify if it’s significant, or we can very accurately say that the patient has no disease. However, if you are trying to say how diseased a given lesion is, then it’s not very accurate,” Gilbert L. Raff, MD, director of cardiac CT and MRI research at William Beaumont Hospital in Royal Oak, Mich., told Cardiology Today.

Once the presence of disease is determined, Raff said invasive angiograms are still the gold standard for determining lesion severity. “I don’t see CT replacing that without further technical improvements,” Raff said.

Gilbert L. Raff

Raff and colleagues conducted a study of 70 consecutive patients to determine the accuracy of the 64-slice CT scanner (Sensation 64 Cardiac, Siemens Medical Systems), which has improved spatial resolution of 0.4 mm and an improved rotation time of 330 ms compared to 16-slice scanners. Results were published in a recent Journal of the American College of Cardiology.

Smaller branches

All patients enrolled in the study were undergoing elective invasive coronary angiography. According to the research abstract, patients could be excluded for atrial fibrillation but not for high heart rate, coronary calcification or obesity.

Calcium score over 400 Agatston U were present in 26% of patients, 25% had heart rates >70 bpm, and 50% were obese.

Researchers analyzed all vessels, including those that were < 1.5 mm in diameter. “Previous studies in 16- slice scanners focused only on the main branches of the coronaries, and we tried to extend that to the smaller branches. We believed that the smaller branches are also important and because the resolution of the new machine is so much higher, we could also comment on those,” Raff said.

Lesions identified by multislice CT were analyzed for significant coronary stenoses by segment, by artery and by patient. Results were compared to quantitative coronary angiography.

Of 1,065 segments, 935 (88%) could be evaluated; 773 (83%) could be assessed quantitatively by both multislice CT and quantitative coronary angiography.

Diagnostic accuracy

The Spearman correlation coefficient between multislice CT and quantitative coronary angiography was 0.76 (P<.0001).

Analysis by segment showed the specificity of multislice CT was 86%, the sensitivity was 95%, the positive predictive value was 66% and the negative predictive value was 98%.

By artery, the specificity was 91%, the sensitivity was 92%, the positive predictive value was 80% and the negative predictive value was 97%.

Analyzed by patient, the specificity was 95%, the sensitivity was 90%, the positive predictive value was 93% and the negative predictive value was 93%.

The standard deviation of the difference between CT and coronary angiography was 28%, which means there is often a difference of about 30% between CT and cath readings in a given lesion.

“The 64-slice CT is much easier to read. The 16-slice modality often shows blurred images,” Raff said.

Raff said the 64-slice CT was much more expensive than current 16-slice models, but the method should eventually be cost effective “because you can evaluate patients very quickly and because you can probably avoid doing a significant number of catheterizations,” he said. – by Jeremy Moore

For more information:

• Raff GL, Gallagher MJ, O’Neill WW, Goldstein JA. Diagnostic accuracy of noninvasive coronary angiography using 64-slice spiral computed tomography. J Am Coll Cardiol. 2005;46:552-557.