Assessing the Electromagnetic Fields Generated By a Radiofrequency MRI Body Coil at 64 MHz: Defeaturing Versus...

Area of Science:

• Medical Imaging
• Computational Electromagnetics
• Biophysics

Background:

• Birdcage coils are crucial for radiofrequency (RF) delivery in Magnetic Resonance Imaging (MRI).
• Accurate computational modeling of these coils is essential for evaluating RF-induced heating.
• Experimental validation is necessary to ensure numerical models reflect physical coil behavior.

Purpose of the Study:

• To systematically evaluate five computational models of birdcage coils for MRI.
• To compare model accuracy and computational cost using defined metrics.
• To determine the suitability of different model simplifications (defeaturing) for predicting coil performance.

Main Methods:

• Implemented five computational models (specific, generic, hybrid) using identical geometry and numerical algorithms.
• Varied driving methods ('defeaturing') to create distinct model types: S2, G32, G16, H16.
• Evaluated accuracy via symmetric mean absolute percentage error (SMAPE) against measurements of frequency response, electric field (E-field), and magnetic field (B-field).

Main Results:

• All models computed B-field magnitude within 35% of measurements.
• Specific (S2), G32, and H16 models accurately predicted E-field inside the phantom (SMAPE ≤ 16%).
• Only the S2 model accurately predicted E-field outside the phantom (SMAPE < 11%).

Conclusions:

• Comparing B-field accuracy solely along the coil's central axis can be misleading.
• Generic or hybrid models suffice for accurate internal phantom fields if currents are modeled correctly.
• A specific model is required for precise E-field prediction in the space between the coil and phantom.