Local SAR compression with overestimation control to reduce maximum relative SAR overestimation and improve...

Area of Science:

• Medical Physics
• Electromagnetics
• Biomedical Engineering

Background:

• Local Specific Absorption Rate (SAR) compression algorithms often overestimate actual SAR values.
• This overestimation is not linearly dependent on actual local SAR, leading to significant relative errors at low SAR.
• Such inaccuracies can limit the performance of transmit arrays in applications like MRI.

Purpose of the Study:

• To develop novel strategies for reducing maximum relative overestimation in local SAR compression.
• To improve the efficiency and performance of parallel transmission systems by addressing SAR overestimation issues.

Main Methods:

• Two strategies were proposed: 1) employing an overestimation matrix to approximate local SAR, and 2) utilizing pre-calculated VOPs (Volume of Perturbation) as an overestimation term.
• These methods aim to decrease the maximum relative overestimation for a fixed number of VOPs.

Main Results:

• The proposed strategies demonstrated a reduction in the number of VOPs by approximately 20% for a given maximum relative overestimation compared to a previous method.
• This improvement comes at the expense of increased absolute overestimation at high actual local SAR values.

Conclusions:

• The novel SAR compression strategies presented are more effective than previously published methods.
• These strategies offer significant improvements in SAR compression, particularly in scenarios where maximum relative overestimation impacts parallel transmission performance.