Coronary angiography by real-time MRI with adaptive averaging

Cardiac and respiratory motion present significant challenges for MR coronary angiography, which have not been completely resolved to date by either breath-holding or respiratory navigation. Adaptive averaging during real-time MRI may provide a useful alternative to these techniques. In this method, cross-correlation is used to automatically identify those real-time imaging frames in which the vessel is present, and to determine the location of the vessel within each frame. This information is then used for selective averaging of frames to increase the signal-to-noise ratio and to improve visualization of the vessel. The correlation theorem was employed to raise the speed of this algorithm by up to two orders of magnitude. Segmenting data collection and reconstruction into subimages allows the extension of this technique to higher spatial resolution. Adaptive averaging provides a robust method for coronary MRI which requires no breath-holding, navigation, or ECG gating. (C) 2000 Wiley-Liss, Inc.