Johannes F.M. Schmidt1, Lukas Wissmann1, Robert Manka1, 2, Peter Boesiger1, Sebastian Kozerke1
1Institute for Biomedical Engineering, University and ETH Zurich, Zurich, Switzerland; 2Department of Cardiology, University Hospital Zurich, Zurich, Switzerland
Scan acceleration methods have proven invaluable for dynamic contrast-enhanced perfusion imaging. One practical issue, however, relates to the requirement of breathholding or shallow breathing which cannot always be ensured in clinical routine exams. An extended k-t PCA algorithm is proposed which corrects for non-rigid frame-to-frame motion based on motion operators derived from two pencil-beam navigators for feet-head and anterior-posterior motion detection and the fully sampled k-t training data. It is demonstrated that this approach successfully corrects for respiratory motion artifacts and hence enables free-breathing 3D cardiac perfusion MRI.