Noninvasive prediction of coronary atherosclerosis by quantification of coronary artery calcification using electron beam computed tomography: comparison with electrocardiographic and thallium exercise stress test results.

OBJECTIVES: This study was designed to compare the usefulness of electron beam computed tomography for prediction of coronary stenosis with that of electrocardiographic (ECG) and thallium exercise tests.
BACKGROUND: Electron beam computed tomography can quantify coronary calcifications; however, its clinical value has yet to be established.
METHODS: Using the volume mode of electron beam computed tomography, we studied 251 consecutive patients who underwent elective coronary angiography because of suspected coronary artery disease and compared the results with those of ECG and thallium exercise tests. The total coronary calcification score was calculated by multiplying the area ( > or = 2 pixels) of calcification (peak density > or = 130 Hounsfield units) by an arbitrarily weighted density score (0 to 4) based on its peak density. The mean of two scans was log transformed.
RESULTS: Calcification was first noted in women in the 4th decade of life, approximately 10 years later than its occurrence in men. Among patients with advanced atherosclerosis (two- and three-vessel disease), calcification scores were uniformly high in women but ranged widely in men. Nine percent of patients with significant stenoses ( > or = 75% by densitometry) had no calcification. The calcification scores of patients with significant stenosis in at least one vessel were significantly higher than those of patients without significant stenosis in the study group as a whole and in most patient subgroups classified according to age and gender. A cutoff calcification score for prediction of significant stenosis, determined by receiver operating characteristic curve analysis, showed high sensitivity (0.77) and specificity (0.86) in all study patients; sensitivity was similarly high even in older patients ( > or = 70 years) and was enhanced in middle-aged patients (40 to < or = 60 years). The difference in specificity between calcification scores and ECG exercise test results had borderline significance (p = 0.058) and that between calcification scores and thallium test results was significant (p = 0.001). The latter difference became small but remained significant (p = 0.01) even after the reevaluation of thallium test results in light of each subject's clinical data.
CONCLUSIONS: Quantification of coronary artery calcification with electron beam computed tomography noninvasively predicted angiographically confirmed coronary stenosis. Results obtained with this method were at least as useful and potentially better in some patient groups than those obtained with thallium and ECG exercise testing.