Meeting Banner
Abstract #1805

Direct Reconstruction of Proton Resonance Frequency-Shift Temperature Maps from K-Space Data for Highly Accelerated Thermometry

Pooja Gaur1, William A. Grissom2

1Chemical and Physical Biology, Vanderbilt University, Nashville, TN, United States; 2Biomedical Engineering, Vanderbilt University, Nashville, TN, United States

Real-time volumetric MR thermometry is desirable for several applications of MR-guided focused ultrasound. We propose a constrained temperature reconstruction method that estimates temperature changes directly from k-space data. Because we fit an image model comprising fully-sampled baseline images to the accelerated k-space data, temperature maps containing aliasing artefacts are removed from the solution space, and high acceleration factors can be achieved without temporal regularization that sacrifices temporal resolution. The method is demonstrated in 2D simulations and experiments. The method will readily extend to 3D, where it will enable accurate real-time volumetric thermometry at a high frame rate.

Abstract PDF Poster PDF

Keywords

absence accelerated acceleration accurate accurately achieved achieves acknowledgments acquisitions aliasing apart applications array arrow assistance audience authors backtracking best biology biomedical blurring bottle brain chemical clinicians coefficient collected combine comprising constrained containing contains cooling degrade descent desirable deviation direct directly domain enable engineering error errors evolution expectation experiment extend fiber finer fitting focal form frame frequency full fully generated golden guided heat heated highly hybrid in vivo increasing indicates induced inserted introduced iteratively least like limited magnitude maps matrix measured minutes model newton noise none optic parallel phantom phys physical pixels polynomial positive probe produce program projections propose proposed quadratic radial readily real receive reconstruct reconstructed reconstruction reconstructions recorded reduced reflect reflecting regularization removed reported resolution rise sacrifices sample sampling scientists search sensing several shaped simulated simulation simulations slice smith solution space spaced sparse target temperature temporal temporally thank thermometry trajectory triangle true tube uncertainty validate validation vector visible volumes volumetric whole windows