Effect of subject-specific head morphometry on specific absorption rate estimates in parallel-transmit MRI at 7 T

Area of Science:

• Medical Physics
• Magnetic Resonance Imaging
• Computational Electromagnetics

Background:

• Accurate estimation of specific absorption rate (SAR) is critical for the safety of 7T parallel-transmit (pTx) MRI.
• Subject-specific electromagnetic head models are needed for precise SAR calculations, but creating them is complex.
• Morphing established models to individual anatomy offers a potential solution for personalized SAR estimation.

Purpose of the Study:

• To evaluate the accuracy of morphing a reference electromagnetic head model to subject-specific morphometry.
• To assess the impact of different registration methods (linear and nonlinear) on SAR estimation accuracy in 7T pTx MRI.
• To determine the contribution of head morphometry to subject-specific SAR variations.

Main Methods:

• Synthetic T1-weighted MR images were generated from existing high-resolution electromagnetic head models.
• A reference multimodal image-based detailed anatomical (MIDA) electromagnetic model was morphed into two different subject spaces (Duke and Ella) using linear and nonlinear registration.
• Maximum 10-g averaged SAR was calculated for 5000 random RF shim sets and compared between morphed and native subject-specific models.

Main Results:

• Nonlinear registration reduced the averaged error in maximum 10-g averaged SAR estimation to 10.7% for the Duke model and 10.1% for the Ella model, down from initial errors of 17.5% and 16.7% with rigid-body registration.
• Affine linear registration also significantly improved accuracy compared to rigid-body registration.
• Head morphometry was found to account for up to half of the subject-specific differences in pTx SAR.

Conclusions:

• Morphing electromagnetic head models to target subjects improves SAR estimation accuracy by better aligning head size, morphometry, and position.
• While morphing enhances agreement, residual differences suggest that uncertainties in tissue composition also impact personalized SAR estimation.
• Future personalized SAR estimation may require separate consideration of head morphometry and tissue composition uncertainties.