Super-resolution for real-time volumetric MR-temperature monitoring

Real time magnetic resonance (MR) thermometry provides continuous temperature mapping inside the human body and allows precise monitoring and control of thermal therapies. In practice it is still difficult to acquire images with 3D isotropic resolution. A high in-plane resolution is therefore preferred at the cost of the slice thickness, which may limit the accuracy of the thermometry. In this study, we evaluated the efficiency of a super-resolution algorithm to increase the spatial resolution in the slice encoding direction and obtain isotropic voxels during real-time hyperthermia. For that purpose, we propose to acquire 2D images while performing a dynamic spatial image shift in the slice encoding direction to obtain the undersampled images. The feasibility and the potential of super-resolution techniques to get isotropic voxels in MR temperature maps acquired in real-time was demonstrated and evaluated on an ex-vivo phantom undergoing hyperthermia treatment. It is shown that that the application of super-resolution to the undersampled temperature data improves the temperature accuracy considerably.