Meeting Banner
Abstract #2835

Local SAR Prediction Errors with Variation of Electrical Properties in the Head at 7T

Muhammad Hassan Chishti1, Zhangwei Wang2, Desmond Yeo1

1GE Global Research Center, Niskayuna, NY, United States; 2GE Healthcare, Waukesha, WI, United States

The ability to accurately predict Specific Absorption Rate (SAR) is a key to patient safety during RF exposure in MR scans. SAR is highly dependent on accurate values of the tissues electrical properties. In this work, we investigate the errors in EM-modeling-based SAR prediction when incorrect electrical conductivity and relative permittivity values are applied. In the 7T human body model simulations performed, results show that the maximum deviation (from nominal) of the ratio of peak SAR10g to the whole head average SAR is -7.83% and occurs when both electrical conductivity and relative permittivity of the tissues are reduced by 20%.

Keywords

absorbed absorption alone analyze axial background better birdcage body brain categorized cause caused cell changing characterize clearly coil college combined concurrently conductivity converse core decrease decreases denoted density dependent deviation deviations distribution distributions divided driven electric electrical equal errors evaluating experiment exposure field fields finally future global goes ground halves head height human impacted incorrect increment induced initially investigate labeled labels largest likely literature loaded loading local magnitude maps metric mode model modeled modeling nominal normalized occurs often overestimating pass patient patients peak percentage phantom phys plots plotted ports power predicting prediction properties proportion quad quadrature quantities reduced relatively remain reported respect respectively risk robust rung seems sets shield simulations simultaneous simultaneously since slices slightly sources sphere spherical studied summarized table tendency thermal tissue tissues transmit true truth tuned types unchanged underestimated underestimating underlying variability variation variations varied vary views visible voltage volumetric white whole