PHASE: Personalized Head-based Automatic Simulation for Electromagnetic properties in 7T MRI.

Area of Science:

• Biomedical Engineering
• Computational Electromagnetics
• Medical Imaging

Background:

• Accurate human head models are crucial for electromagnetic (EM) simulations, especially for Specific Absorption Rate (SAR) assessments.
• Current methods rely on the Virtual Population due to limitations in public resources and manual annotation of patient data.
• There is a need for scalable, automated generation of patient-specific head models.

Purpose of the Study:

• To introduce Personalized Head-based Automatic Simulation for EM properties (PHASE), an open-source toolbox.
• To enable automated generation of high-resolution, patient-specific head models for EM simulations.
• To evaluate the performance of PHASE-generated models against gold standards.

Main Methods:

• Utilized paired T1-weighted MRI and CT scans from 15 real patients.
• Developed an automated toolbox (PHASE) for generating 14-tissue-labeled head models.
• Conducted semi-automated segmentation and EM simulations for performance evaluation.

Main Results:

• PHASE models demonstrated comparable global SAR and localized SAR (SAR-10g) values to gold standard references.
• The automated approach proved effective in generating accurate, patient-specific head models.
• The toolbox achieved results suitable for safety guideline evaluations.

Conclusions:

• PHASE is a promising tool for generating large-scale, high-quality human head models for EM simulations.
• The open-source availability of PHASE facilitates further research and development in the field.
• PHASE models can support accurate EM simulations for safety assessments and personalized medicine.