Fast imaging techniques allow monitoring of contrast medium (CM) first-pass kinetics in a multislice mode. Employing shorter recovery times improves cardiac coverage during first-pass conditions, but potentially flattens signal response in the myocardium. The aim of this study was therefore to compare in patients with suspected coronary artery disease (CAD) two echo-planar imaging strategies yielding either extended cardiac coverage or optimized myocardial signal response (protocol A/B, six/four slices; preparation pulse, 60 degrees /90 degrees; delay time, 10/120 msec; readout flip angle, 10 degrees /50 degrees; respectively). In phantoms and myocardium of normal volunteers (N= 10) the CM-induced signal increase was 2.5-3 times higher with protocol B (P < 0.005) than with protocol A. For the detection of individually diseased coronary arteries (> or =1 stenosis with > or =50% diameter reduction on quantitative coronary angiography (QCA)), receiver-operator characteristics of protocol B (signal upslope in 32 sectors/heart) yielded a sensitivity/specificity of 82%/73%, which was superior to protocol A (P < 0.05, N= 14). For the overall detection of CAD, the sensitivity/specificity of protocol B was 85%/81%. An adequate signal response in the myocardium is crucial for a reliable detection of perfusion deficits during first-pass conditions. The presented protocol B detects CAD with a sensitivity and specificity similar to scintigraphic techniques.