Abstract
This study analyses the relation between image quality and contrast kinetics in bolus-timed carotid magnetic resonance angiography (MRA) and interprets the findings by Fourier-based numerical modelling. One hundred patients prone to carotid stenosis were studied using contrast-enhanced carotid MRA with bolus timing. The carotid MRAs were timed to start relatively early without accounting for the injection time of the contrast medium. For interpretation different starting times were modelled, utilising the spectral information of the test bolus series. In the test bolus series the arterial time-to-peak showed a large 95% confidence interval of 12-27 s, indicating the need for individual MRA timing. All bolus-timed MRAs were of good diagnostic quality. The mean (+/-SD) arterial contrast-to-noise ratio was 53.0 (+/-12.8) and thus high, and 95% of the MRAs showed a slight venous contamination of 11.8% or less (median 5.6%). According to the Fourier-based modelling the central k-space may be acquired about 2 s before the arterial contrast peak. This results in carotid MRAs with sufficiently high arterial enhancement and little venous contamination. In conclusion, in bolus-timed carotid MRA a relatively short timing provides good arterial contrast with little venous contamination, which can be explained by Fourier-based numerical modelling of the contrast kinetics.