Despite impressive image quality, it is unclear if noninvasive coronary angiography with multislice spiral computed tomography (CT) is powerful enough to act as a filter before invasive angiography (INV-A) in symptomatic patients.

We therefore studied 133 consecutive symptomatic patients with suspected coronary artery disease (CAD) and an indication for INV-A (chest pain and signs of ischemia in conventional stress tests). Patients with known CAD, acute coronary syndrome, or a calcium volume score >1000 were excluded. In all patients, both INV-A and multislice CT angiography (MSCT-A) (Philips MX 8000 multislice spiral CT, scan time 250 milliseconds, slice thickness 1.3 mm, 120 mL of contrast agent, 4 mL/s, retrospective gating) were directly compared by 2 independent investigators using the American Heart Association 15-segment model. Altogether, we studied 1596 segments, 74% had diagnostic image quality. Multislice CT angiography correctly identified 68 significant stenoses of the 75 stenoses seen with INV-A (sensitivity 91%). In 945 of 1185 diagnostic segments, stenosis could correctly be ruled out with MSCT-A. There were 3 times more stenoses seen with MSCT-A compared with INV-A (positive predictive value 29%) mainly because of misclassification of nonobstructive plaques as stenosis. The per-patient analysis allowed to exclude significant CAD in 42 (32%) of 133 patients. In only 6 of 53 patients, MSCT-A failed to detect significant stenosis, 4 of those were in small segments not requiring intervention. Calcium scoring alone was less suited as a filter before angiography: 25 patients (18% of study group) had a calcium score = 0, and 8 of these patients turned out to have significant stenoses.

Multislice CT angiography, but not calcium scoring alone, offers promise to reduce the number of INV-A in symptomatic patients with suspected CAD by up to one third with minimal risk for the patient.