Meeting Banner
Abstract #4421

On Patient-Specific Models for Accurate SAR Estimations at 7 T

Jin Jin1, Feng Liu1, Ewald Weber1, Stuart Crozier2

1University of Queensland, St Lucia, Queensland, Australia; 2The University of Queensland, St Lucia, Queensland, Australia

To more effectively predict the patient SAR values for high-field MRI applications, a novel approach of creating patient-specific tissue model is demonstrated in this study. First, a matching model from a high-resolution image- and tissue-library is selected for the target patient. The library tissue distribution is then warped to match the patients geometry as the corresponding library image is registered to the low-resolution scans of the patient. Results from studying the models 1-gram SAR distribution using finite-different time-domain method suggest that the developed patient model can predict regions of elevated SAR within the patient with remarkable accuracy.

Keywords

absorbing absorption accuracy accurate acquisition adopted agreement algebra allow among apart applications approximated arbitrarily available averaging body bottom calculation calculations cells challenging coil coincided common commonly compensate computed condition conventionally correct counts created creates creating database denotes denoting details developed development deviation distribution distributions effective efficiency electromagnetic employed energy equivalent essentially estimations evaluate examined excitation exciting field fields flow generic good gradually gram head importance indicated indicates intensity inter intersection inverse largely largest loaded loading location locations magnitude margins matching measures median model models mutual necessary necessitate numerical optimal original overall overlap pair pairs patient pilot plot ports power predicted prediction proof quality random reduced reduction registered registering registration registrations remarkably represent require resignation resolution respect respectively safety sampled scaled segmenting significantly simulated simulations simulator source space studies subject subjects suggested symbol system systems table target template tissue today trans tuned types typically variations varied variously viable volume weber website whenever