Hui Xue1, Yu Ding2, Christoph Guetter1, Andrew E. Arai3, Peter Kellman3
1Siemens Corporate Research, Princeton, NJ, United States; 2Dorothy M. Davis Heart and Lung Research Institute, The Ohio State University, Columbus, OH, United States; 3National Institutes of Health, National Heart, Lung and Blood Institute, Bethesda, MD, United States
To improve the quality of free-breathing myocardial perfusion imaging, we propose a novel reconstruction algorithm named as MOCO-SPIRiT which extends the SPIRiT reconstruction by incorporating non-rigid respiratory motion correction (MOCO). With motion correction, it is possible to employ spatial-temporal regularization for better image quality while retaining the dynamic information (i.e. preserve the fidelity of time intensity curves). In this way, the improved SNR may be used to support higher spatial resolution which is the key to minimize dark rim artifacts. Unlike k-t methods without MOCO that are susceptible to respiratory motion, the proposed method allows free-breathing. Both phantom and in-vivo tests show the new algorithm leads to higher SNR and lower artifacts, compared to TGRAPPA and L1SPIRiT.