Magnetic Resonance Imaging Compatibility and Safety of the SOUNDTEC Direct System

Objective/Hypothesis: The purpose of this study was to evaluate magnetic resonance imaging (MRI) compatibility and safety of an electromagnetic implanted hearing device (the SOUNDTEC Direct System; SOUNDTEC, Inc., Oklahoma City, OK) implant during a 0.3‐Tesla open MRI imaging examination of the head and neck and to develop an MRI protocol that maximizes patient safety while minimizing the need for implant removal. The current literature regarding MRI compatibility of implantable hearing devices was reviewed. Study Design: Linear and torsional forces, heating, and implant magnetization were evaluated in vitro. Implanted fresh‐frozen human temporal bones were used to evaluate image distortion. A prospective study of 11 volunteers previously implanted with the SOUNDTEC Direct System was conducted to evaluate MRI compatibility and safety. A MEDLINE search of the literature between 1980 and July 2005 was reviewed to summarize MRI compatibility testing of implantable hearing devices. Methods: Torsional and linear forces experienced by eight implant magnets were measured using calibrated neurologic Von Frey Hairs and compared with finite element analysis predictions as well as forces required to separate the incudostapedial joints of 12 fresh‐frozen human temporal bones. Implant heating was determined by measuring the temperature change of eight implant vials compared with saline controls immediately after a head MRI scan. Implant magnetization was evaluated after repeated exposure to a 0.3‐Tesla magnetic field. An 11‐patient prospective study was performed to evaluate MRI compatibility in a 0.3‐Tesla open MRI environment using adult volunteers previously implanted with the SOUNDTEC Direct System. A modified MRI protocol was developed to maximize patient safety. Each individual underwent an audiometric and otologic examination immediately before and after MRI. Results: Peak linear force at the MRI entry measured 0.5 g ± 0.2 standard deviation (SD). Maximum torque occurred at isocenter and measured 11.4 g‐cm ± 1.2 SD. The mean torque required to separate the incudostapedial joint was 33.8 g‐cm ± 20.4 SD. The average increase in temperature of the eight implant vials was 0.45°C ± 0.11 SD, whereas the increase in temperature of the three saline controls measured 0.47°C ± 0.11 SD. The average change in magnetic flux density of the 14 implant magnets tested was 22.0 gauss. Maximum image distortion occurred during the gradient echo sequence and measured 8.6 cm in diame