The aim of this study was to investigate the reliability of calcium scoring (CS) and electron beam computed tomographic angiography (EBCTA) as a noninvasive tool in the diagnosis of coronary artery disease (CAD): 93 consecutive patients (aged 59 +/- 9 years) with symptoms suspicious for CAD underwent CS. In 87 of these subjects, an additional EBCTA investigation was performed. Using receiver-operating characteristic curve analysis, we determined a calcium score cut point providing an overall sensitivity of 80% and a specificity of 72% in detecting patients with CAD. For clinical purposes the use of cut points is difficult. We therefore determined score ranges providing >80% specificity (high score range) and >85% sensitivity (low score range) and determined the scores between these ranges as equivocal borderline scores. Calculated on a per-segment basis in assessable proximal and midcoronary segments, the sensitivity for detecting coronary stenoses >50% was 78%, and the specificity was 93%. Thus, 32 of 44 patients with significant CAD and 24 of 49 patients without CAD were correctly classified. The combination of CS and EBCTA predicted CAD in 77% (72 of 93) of patients. No or low calcium scores provided high specificity for ruling out CAD. The addition of EBCTA in those patients improved sensitivity. In patients with high calcium scores, accuracy of EBCTA was not significantly different from CS alone (72% vs 83%), whereas in patients with borderline scores it was significantly superior (80% vs 58%, p <0.03). Thus, the complementary use of CS and EBCTA appears beneficial, particularly in patients with borderline scores, and could improve sensitivity in the low score range. In the presence of high scores, no major diagnostic gain from an additional EBCTA versus CS alone could be observed.