A perturbation approach for ultrafast calculation of RF field enhancements near medical implants in MRI

Peter R S Stijnman,Alexander J E Raaijmakers,Bart R Steensma,Cornelis A T Van Den Berg

doi:10.1038/s41598-022-08004-7

Peter R S Stijnman, Alexander J E Raaijmakers + Show 2 more

Open Access

https://doi.org/10.1038/s41598-022-08004-7

Copy DOI

Abstract

Patients with medical implants often are deprived of magnetic resonance imaging examination because of safety risks. One specific risk is the enhancement of the radiofrequency fields around the medical implant potentially resulting in significant tissue heating and damage. The assessment of this enhancement is a computationally demanding task, with simulations taking hours or days to converge. Conventionally the source of the radiofrequency fields, patient anatomy, and the medical implant are simulated concurrently. To alleviate the computational burden, we reformulate a fast simulation method that views the medical implant as a small perturbation of the simulation domain without the medical implant and calculates the radiofrequency fields associated with this perturbation. Previously, this method required an extensive offline stage where the result is intractable for large simulation domains. Currently, this offline stage is no longer required and the method is completely online. The proposed method results in comparable radiofrequency fields but is orders of magnitude faster compared to standard simulation technique; the finite-difference time-domain, the finite-sums, and the finite element methods. This acceleration could enable patient-specific and potentially online radiofrequency safety assessment.

Highlights

Magnetic resonance imaging (MRI) is one of the main medical imaging modalities and has become indispensable for diagnosis, treatment planning, and patient follow-up studies
We have shown that it is possible to significantly accelerate the calculation of the RF field enhancement that is associated with a medical implant in MRI, without extensive precomputations or excessive storage requirements
This is achieved by using a matrix-free minimization scheme where we use a surrogate function based on the volume integral equation method (VIE) method to replace the intractable so-called library matrix

Summary

Introduction

Magnetic resonance imaging (MRI) is one of the main medical imaging modalities and has become indispensable for diagnosis, treatment planning, and patient follow-up studies. Based on the principle of nuclear magnetic resonance, MRI requires three types of magnetic fields: a permanent strong magnetic field (1.5 or 3 T), rapidly switching magnetic gradient fields for spatially encoding the signal, and radiofrequency (RF) fields generated by an RF coil or antenna to excite the atomic nuclei to create the signal This multitude of magnetic fields poses a severe safety risk for patients with medical implants. When a medical implant is labeled, MRI conditional limitations are placed on the MRI sequence parameters (e.g. maximum RF power) These limitations decrease the image quality and/or increase the required scan time. This enhancement creates locations where the temperature is drastically increased, potentially causing tissue burns inside the patient or even excessive brain d amage[17]

Methods

Results

Conclusion