Nick Todd1, Allison Payne2, Dennis Parker3
1Physics, University of Utah, Salt Lake City, UT, USA; 2Mechanical Engineering, University of Utah, Salt Lake City, UT, USA; 3Radiology, University of Utah, Salt Lake City, UT, USA
We present MRI temperature imaging using a 3-D gradient echo sequence that undersamples k-space and is reconstructed using a model predictive filtering (MPF) algorithm. The MPF algorithm combines information from an identified thermal model of the tissue with undersampled k-space data. The 3-D imaging was chosen for its superior spatial resolution and coverage. The technique provides temperature maps with 2mm3 isotropic spatial resolution and 6 second temporal resolution. The 3-D MPF technique was compared to the traditional 2-D PRF technique over 8 HIFU heating experiments. The standard deviation of the temperature difference between the 2 methods was 0.57 degrees C.