Efficient prediction of MRI gradient‐induced heating for guiding safety testing of conductive implants.

Purpose: To propose an efficient numerical method to predict the temperature increase of an implantable medical device induced by any linearly polarized homogeneous magnetic field, according to the ISO 10974 methodology for testing of gradient‐induced device heating. Theory and Methods: The concepts of device‐specific power and temperature tensors are introduced to mathematically describe the electromagnetic and thermal anisotropic behavior of the device, from which the device heating for an arbitrary exposure direction can be predicted. The proposed method is compared to a brute‐force approach based on simulations, and validated by applying it to four reference orthopedic implants with a commercial simulation software. Results: The proposed method requires about 5%$$ \% $$ of the time required by the brute‐force approach, and 30%$$ \% $$ of the memory occupancy. The temperature increase predicted by the proposed method over a range of incident magnetic field exposures deviated from brute‐force direct simulations by less than ±$$ \pm $$0.3%$$ \% $$. Conclusion: The proposed method allows efficient prediction of the heating of an implantable medical device induced by any linearly polarized homogeneous magnetic field using a small fraction of the simulations required by the brute‐force approach. The results can be used to predict the worst‐case orientation of the gradient field, for subsequent experimental characterization according to the ISO 10974 standard. [ABSTRACT FROM AUTHOR]