Meeting Banner
Abstract #0613

A Method for Efficient & Robust Estimation of Low Noise, High Dynamic Range B0 Maps

Joseph Dagher1, Ali Bilgin2, Timothy Reese3, Georges El Fakhri1

1Nuclear Medicine and Molecular Imaging, Department of Radiology, Massachusetts General Hospital, Harvard Medical School, Boston, MA, United States; 2Departments of Biomedical Engineering and, Electrical and Computer Engineering, University of Arizona, Tucson, AZ, United States; 3Athinoula A. Martinos Center for Biomedical Imaging, Department of Radiology, Massachusetts General Hospital, Harvard Medical School, Boston, MA, United States

We propose a joint acquisition-processing solution to the problem of field map estimation. Our optimizer carefully chooses three echo times that guarantee robust and accurate field map estimation using the reconstruction algorithm over an arbitrary spectral range of inhomogeneity values. We show that our method is not subject to the traditional accuracy-robustness trade-off. The resulting implications include: improved robustness by removing the bound on the shortest echo time difference, enhanced spectral estimation over a large dynamic range of inhomogeneity values, and eliminating the need for phase unwrapping. Phantom experiments confirm these conclusions.

Abstract PDF Presentation Video

Keywords

able account accuracy accurate acquisition adaptive ambiguity annealing applications approximates arbitrary attempt basis biomedical capability challenge chosen claim clear clearly collected combine compute computed confidence conjugate consideration contribution correctly corresponds critical despite detectable determine determined distortion domain dynamic echoes efficient efficiently enables engineering estimating estimation exists expected extending fact field forces frank function fundamental furthermore general grant grid grids guarantee hospital implies importance improvement incorrectly inhomogeneity inside instead integer interval larger lengths literature location long longer maps medical minimized momentarily noise novel nuclear numerical object obviously overlap overlapping pair performance phantom pixel post practical prince prior problem processing programming proposed quality radiology reconstruction reduces removes removing report represent representative require requires reside resolved respectively robust robustly robustness routine running scanner school seek segment segments separated separation short simulated situation slices smooth solution spaced spectral spectroscopy squared statistics summarize theory third trade uniformly unwrapping variable variance vast visible water worth written yield