Meeting Banner
Abstract #1158

About the Performance of Multi-Dimensional Radial Self-Navigation Incorporating Compressed Sensing for Free-Breathing Coronary MRI

Gabriele Bonanno1, 2, Gilles Puy3, 4, Yves Wiaux3, 5, Ruud B. van Heeswijk1, 2, Matthias Stuber1, 2

1Department of Radiology, University Hospital (CHUV) and University of Lausanne, Lausanne, Vaud, Switzerland; 2Center for Biomedical Imaging (CIBM), Lausanne, Vaud, Switzerland; 3Institute of Electrical Engineering, cole Polytechnique Fdrale de Lausanne (EPFL), Lausanne, Vaud, Switzerland; 4Institute of the Physics and Biological Systems, cole Polytechnique Fdrale de Lausanne (EPFL), Lausanne, Vaud, Switzerland; 5Institute of Bioengineering, cole Polytechnique Fdrale de Lausanne (EPFL), Lausanne, Vaud, Switzerland


A novel image-based respiratory self-navigation method was developed for free-breathing coronary MRI. Under-sampled radial sub-images are acquired on a beat-to-beat basis, non-linear reconstruction is performed, and motion parameters are extracted for direct motion correction. In a first step, the new algorithm was optimized and evaluated using a numerical simulation of a moving heart. Subsequently, the performance was quantitatively ascertained in an in vivo study that included 12 healthy adult subjects where it was objectively demonstrated that self-navigation incorporating compressed sensing is a powerful tool for motion artifact suppression in radial coronary MRI.