Acceleration of T2 prime parameter mapping and dual echo DCE/DSC MRI using compressed sensing

Abstract

MRI is widely used in many clinical applications because it is a flexible diagnostic imaging modality, and is possible to characterize many parameters in human body such as relaxation times, perfusion, diffusion, and so on. The reason why MRI becomes more valuable in clinical fields is not only these advantages as stated above but also the developments of sequences and techniques for fast and better imaging to break the limitations of temporal resolution or long scan time. Compressed Sensing is one of the promising techniques for these improvements of MRI in dynamic imaging. In this paper, accelerations of two dynamic MR applications are implemented using compressed sensing. One is T2 prime (T2`) parameter mapping which may provide better information to predict infarct growth in acute stroke patients than diffusion weighted imaging (DWI) that was commonly used. To investigate the possibility of reducing total scan time, the images are reconstructed from highly downsampled dynamic MR data using compressed sensing with patch based low rank penalty. Then, T2 maps are constructed from the reconstruction images and compared with the maps from full data. The other application is dual echo of dynamic contrast enhanced (DCE) and dynamic susceptibility contrast (DSC) MRI which has capability to extract some clinically important parameters of tumor hemodynamics for the diagnosis of cancer. In vivo accelerated acquisition is implemented in compressed sensing scheme and its econstruction result is compared to images from the full data which means not accelerated and the reconstruction images of downsampled data from the full data. The principles of MRI, compressed sensing and the two dynamic MR applications are also introduced.