Purpose: Serious tissue heating may occur at the tips of elongated metallic structures used in MRI-guided interventions, such as vascular guidewires, catheters, biopsy needles, and brachytherapy needles. This heating is due to resonating electromagnetic radiofrequency (RF) waves along the structure. Since it is hard to predict the exact length at which resonance occurs under in vivo conditions, there is a need for methods to monitor this resonance behavior. In this study, the authors propose a method based on the RF induced image artefacts and demonstrate its applicability in two phantom experiments.
Methods: The authors developed an analytical model that describes the RF induced image artefacts as a function of the induced current in an elongated metallic structure placed parallel to the static magnetic field. It describes the total RF field as a sum of the RF fields produced by the transmit coil of the MR scanner and by the elongated metallic structure. Several spoiled gradient echo images with different nominal flip angle settings were acquired to map the B1+ field, which is a quantitative measure for the RF distortion around the structure. From this map, the current was extracted by fitting the analytical model. To investigate the sensitivity of our method we performed two phantom experiments with different setup parameters: One that mimics a brachytherapy needle insertion and one that resembles a guidewire intervention. In the first experiment, a short needle was placed centrally in the MR bore to ensure that the induced currents would be small. In the second experiment, a longer wire was placed in an off-center position to mimic a worst case scenario for the patient. In both experiments, a Luxtron (Santa Clara, CA) fiberoptic temperature sensor was positioned at the structure tip to record the temperature.
Results: In the first experiment, no significant temperature increases were measured, while the RF image artefacts and the induced currents in the needle increased with the applied insertion depth. The maximum induced current in the needle was 44 mA. Furthermore, a standing wave pattern became clearly visible for larger insertion depths. In the second experiment, significant temperature increases up to 2.4 degrees C in 1 min were recorded during the image acquisitions. The maximum current value was 1.4 A. In both experiments, a proper estimation of the current in the metallic structure could be made using our analytical model.
Conclusions: The authors have developed a method to quantitatively determine the induced current in an elongated metallic structure from its RF distortion. This creates a powerful and sensitive method to investigate the resonant behavior of RF waves along elongated metallic structures used for MRI-guided interventions, for example, to monitor the RF safety or to inspect the influence of coating on the resonance length. Principally, it can be applied under in vivo conditions and for noncylindrical metallic structures such as hip implants by taking their geometry into account.