MR Thermometry Near Metallic Devices Using Multispectral Imaging
This novel development enables non-invasive temperature measurement using MRI (MR thermometry), in the vicinity of implanted metallic devices such as joint replacements or fixation hardware. MR thermometry provides safety and efficacy in promising treatments such as MR-guided focused ultrasound surgery (MRgFUS), where tumor tissue is destroyed through heating with bundled ultrasound waves.
Conventional MR thermometry is based on the shift in proton resonance frequency (PRFS) that occurs with changing temperature. However, the underlying measurement technique relies on a homogeneous magnetic field and therefore fails in the presence of strong field distortions as they are induced by metallic devices such as a joint replacement. As a consequence, an increasing population of patients has to be excluded from MRgFUS treatments.
In this article, we propose MR thermometry near metallic devices by measuring the temperature dependent T1 relaxation time with multispectral imaging (MSI). MSI enables imaging in the presence of strong field inhomogeneities by splitting the measurement into sub-measurements individually adapted to the local magnetic field. We demonstrate the feasibility of this approach in a phantom experiment and compare it to the conventionally used PRFS measurement.
Weber H, Taviani V, Yoon D, Ghanouni P, Pauly KB, Hargreaves BA. MR thermometry near metallic devices using multispectral imaging. Magn Reson Med. 2017 Mar;77(3):1162-9.
Application of the proposed MR thermometry technique in a phantom and comparison with the conventionally used PRFS measurement. The phantom contains a metallic device and a heat source generating a temperature gradient throughout the phantom. The PRFS-based temperature map (b) reveals no reliable temperature information in the areas of strong field distortions, located at the top and the bottom of the device. The map is either dominated by noise or biased, with deviations from the expected temperature pattern as high as 10 °C. In contrast, the T1-based temperature map (a) yields a reasonable temperature pattern over the entire phantom, with no signs of bias around the metallic device and also no increased noise even in close proximity to the device. Although the temporal resolution of the current implementation is still limiting for most clinical applications, the precision and dynamic range may be promising for the detection of larger temperature changes.
Hans Weber and Valentina Taviani are alumni of the BMR group