Henrik Pedersen1, Steffen Ringgaard2, Won Yong Kim2
1Functional Imaging Unit, Glostrup Hospital, Glostrup, Denmark; 2MR Research Center, Aarhus University Hospital Skejby, Aarhus, Denmark
Correction of respiratory motion represents a major practical problem in quantitative myocardial perfusion MRI. This study presents a retrospective method of respiratory motion correction that is fully automated, insensitive to image brightness changes, and suitable for correcting respiratory motion artifacts in accelerated MRI. The proposed method uses additionally acquired respiratory navigators to predict non-rigid motion over the entire field-of-view. The underlying motion model is learned from a set of training images. Among the navigator and training configurations evaluated in this study, the optimum setup achieved an average 4-5 fold in-plane motion reduction in 32 myocardial perfusion MRI data sets.