Evaluation of the Susceptibility Artifacts and Tissue Injury Caused by Implanted Microchips in Dogs on 1.5 T Magnetic Resonance Imaging

Transcription

1 FULL PAPER Surgery Evaluation of the Susceptibility Artifacts and Tissue Injury Caused by Implanted Microchips in Dogs on 1.5 T Magnetic Resonance Imaging Miyoko SAITO 1), Shin ONO 2), Hideki KAYANUMA 2), Muneki HONNAMI 1), Makoto MUTO 1) and Yumi UNE 3) 1) Departments of Surgery II, 2) Radiology and 3) Pathology, School of Veterinary Medicine, Azabu University, Kanagawa , Japan (Received 1 September 2009/Accepted 31 December 2009/Published online in J-STAGE 20 January 2010) ABSTRACT. Performing magnetic resonance imaging (MRI) in patients with a metallic implant raises concern over the potential complications, including susceptibility artifacts, implant migration, and heat injury. The purpose of this study was to investigate these complications in dogs with implanted microchips by evaluating MR images and the histopathological changes after 1.5 Tesla (T) MRI. Five dogs underwent microchip implantation in the cervicothoracic area. One month later, the area was imaged using 1.5T MRI in three dogs. The microchips were removed surgically together with the surrounding tissue in all dogs. There was significant signal loss and image distortion over a wide range around the area where the microchip was implanted. This change was consistent with susceptibility artifacts, which rendered the affected area including the spinal cord undiagnostic. The artifact was more extensive in T2*-weighted images (gradient-echo) and less extensive in proton density-weighted images (fast spin-echo with short echo time). Histopathologically, all microchips were well-encapsulated with granulation tissue, and there were no evidence of migration of microchips. Cell debris and a moderate number of degenerated cells with fibrin were seen in the inner layer of the granulation tissue in each dog that underwent MRI. These changes were very subtle and did not seem to be clinically significant. The results of this study suggest that, in 1.5T MRI, susceptibility artifacts produced by implanted microchips can be marked, although the dogs with implants appeared to be scanned safely. KEY WORDS: heat injury, microchip, MRI, susceptibility artifacts. J. Vet. Med. Sci. 72(5): , 2010 The identification of animals has become widely recognized as being important as attention to animal welfare, secure quarantine, and wildlife protection has been promoted. Thus, more reliable and safer methods to accomplish this are in great demand. A microchip is an implantable electronic device, employed worldwide as a passive identification method. A microchip contains metal materials as its essential components. Performing magnetic resonance imaging (MRI) in the presence of a metallic device raises concern over the potential complications including implant migration, heat injury, and MR image distortion [7, 14, 16]. Given the potential widespread application of this device and increasing use of MRI, investigations on the safety and imaging artifacts produced by this device are necessary. The MRI system with a high-magnetic field has been increasingly used in veterinary medicine, and the influence of metallic devices on the patient and images can be more significant as the radiofrequency magnetic field increases. The purpose of this study was to investigate the effect of implanted microchips on MR images due to the interaction of the microchip with a strong radiofrequency magnetic field produced by a 1.5 Tesla (T) MRI scanner. We also examined a tissue change and the viability of ID data stored on the microchip after exposing it to a 1.5 T MR environment. *CORRESPONDENCE TO: SAITO, M., School of Veterinary Medicine, Azabu University, Fuchinobe, Sagamihara, Kanagawa , Japan. msaito@azabu-u.ac.jp MATERIALS AND METHODS Five purpose-bred beagles were studied. They were clinically healthy females and ranged in age from two to five years old and in body weight from 12 to 15 kg. The microchip (LIFECHIP, Digital Angel Corp, Saint Paul, MN, U.S.A.) used in this study was cylindrical and approximately 2 mm in diameter and 11 mm in length. It was implanted into the dogs using the injector following the manufacturer s directions. Briefly, the microchip was injected under the skin of the dorsal part of the cervicothoracic junction, with the device along the axial plane of the dog. The dogs were observed for one month to ensure the formation of tissue encapsulation surrounding the microchip. One month after the implantation, a radiograph was taken in order to determine the location of the microchip. Three dogs were then anesthetized with isoflurane and underwent MR imaging (MRI group). MR imaging of the cervicothoracic area was conducted using a 1.5 T superconducting magnet (Toshiba ViSART, Toshiba, Tokyo, Japan) with a surface coil. The direction of the static magnetic field (B 0 ) of this MRI was parallel to the bore axis. Transverse, sagittal, and dorsal planes were obtained with T1-weighted imaging (T1WI) [repetition time (TR) =500, echo time (TE) =15, spin-echo], T2-weighted imaging (T2WI) (TR=5000, TE=120, fast spin-echo), T2*-weighted imaging (T2*WI) (TR=450, TE=15, gradient-echo), fluid attenuated inversion recovery (FLAIR) [TR=8000, TE=120, inversion time (TI) =2300, fast spin-echo], proton density- weighted imaging (PDWI) (TR=2800, TE=12, fast spin-echo), and T1WI contrast studies using gadoteridol (ProHance, Eisai, Tokyo,

2 576 M. SAITO ET AL. Japan) at a dose of 0.2 ml/kg. The frequency-encoding direction was dorsoventral in the transverse plane and craniocaudal in the dorsal and sagittal planes. The receiver bandwidth was preset to Hz/pixel in T2WI and PDWI and Hz/pixel in other sequences. The other imaging parameters used for this study included a matrix and 5-mm slice thickness with a field of view (FOV) of cm in the transverse plane, a matrix and 5- mm slice thickness with a FOV of cm in the dorsal plane, and a matrix and 3-mm slice thickness with a FOV of cm in the sagittal plane. The FOV was set to include the entire region affected by the artifact shown in the locator image in all dogs undergoing the MRI procedure. An interslice gap was set to 20% of the slice thickness and all images were obtained using a 2D acquisition. In order to evaluate the maximum tissue injury load, the total time for MRI in individual dogs was set to approximately 90 min. This was considered to be the longest possible duration of the MRI procedure in a clinical setting for the purpose of diagnosis. Immediately after the completion of imaging, 20 mg/kg cefazolin sodium hydrate (Cefamezin, Astellas, Tokyo, Japan) was administered intravenously and a wide ellipse of the skin and cutaneous tissue fully encompassing the microchip was removed operatively under general anesthesia. A mark was made with a surgical stitch on each side of the tissue at the cranial and caudal end of the microchip to verify the implant s direction after its removal. The microchip was then taken from the harvested soft tissue specimen. The tissue was immediately fixed in 10% buffered formalin and routinely processed for histopathological evaluation. In the remaining two dogs, which did not undergo MRI (control group), the operative specimen was obtained in the same fashion as in dogs in the MRI group. This unscanned specimen served as a control for assessing histopathological changes resulting directly from the presence of the microchip versus tissue reactions arising due to exposure to an MR environment. The incision site was closed routinely. The dogs were housed individually and the surgical site was observed closely for 14 days. Twenty mg/kg cephalexin (Keflex, Shionogi, Osaka, Japan) was administered twice a day for the first 5 days. The specimens obtained from dogs in both MR and control groups were cut perpendicular to the long axis of the microchip. The tissue was processed for routine paraffin embedding, and 3 4- micron-thick sections were prepared with hematoxylin and eosin staining (H.E.). Changes in granulation tissue surrounding the microchip (i.e., encapsulation) were assessed particularly for thermal and migration damage according to the previous study [9]. Tissue evaluation was made based on the following: presence of direct thermal damage (as evidenced by presence of edema, hemorrhage, degeneration, and necrosis) and migration damage (as evidenced by presence of capsule rupture and hemorrhage in the rim surrounding the encapsulation). The removed microchip was scanned by a reader to determine the survival of its data-recording function. The MR images were analyzed to measure the magnitude of the artifact using commercially available software (Photoshop, Adobe System Incorporated, Tokyo, Japan). The borderline of artifact area was estimated by visual evaluation. The following parameters were utilized to determine the largest artifact size and clinical significance on each sequence of each plane (Fig. 1): Longest artifact (cm): measured by drawing a round or oval that enclosed the visible artifact in the dorsoventral (DV), lateral (Lat), or craniocaudal (CrCa) directions (Fig. 1A). Artifact area as a percentage of the cervical area (%): calculated as the actual visibly distorted area divided by the area of the cervical region. In the case of the sagittal or dorsal plane, the cervical region was defined as the area from the cranial end of C1 to the caudal end of C7 (Fig. 1B). Estimated entire artifact area (cm 2 ): calculated by drawing a round or oval that enclosed the visible artifact with the microchip as the central point (Fig. 1C). Presence of an artifact on spinal cord structure. All procedures in this study were conducted in accordance with the guidelines approved by the Animal Research Committee of Azabu University, and all protocols in this study were approved by this committee. RESULTS Survey radiographs confirmed that the microchip was located at the original insertion site in all dogs. Actual total time for MRI in individual dogs was 77 to 105 min. There was a significant signal loss and image distortion over a wide range around the area where the microchip was implanted on all MR images. This change was characterized as a central signal void with a high signal rim in T1W (with and without contrast), T2W, FLAIR, and PDW images. The range of the central void was huge, and a high signal rim was not evident in T2*W images. These findings were consistent with susceptibility artifacts, and these artifacts rendered the affected area undiagnostic. The typical appearance of the artifact is shown in Fig. 3. The affected area was the most extensive in T2*W images and least extensive in PDW images in all planes. Of all the images obtained, the longest artifact was 11.9 cm for the maximum (T2*WI, sagittal, craniocaudal direction), which was 10 times longer than the actual device size. The longest artifact was 3.8 cm for the minimum (PDWI, transverse, lateral direction) in all the images obtained. The artifact area as a percentage of the cervical area was 67% for the maximum (T2*WI, dorsal) and 15% for the minimum (PDWI, transverse). The estimated entire artifact area was calculated as 158 cm 2 for the maximum (T2*WI, transverse) and 32 cm 2 for the minimum (PDWI, sagittal). All sequences had at least one image where the artifact extended over the spinal cord. The effect on the cord was particularly apparent in T2*WI and the artifact rendered the spinal cord undiagnostic in most slices of transverse images in all three dogs (Fig.

3 MRI IN DOGS WITH MICROCHIPS 577 A B C Fig. 1. Artifact magnitude measurement on MR images. Longest artifact (cm): measured by drawing a round or oval that enclosed the visible artifact in the dorsoventral (DV), lateral (Lat), or craniocaudal (CrCa) directions. The artifact distance was determined by measuring the distance shown by the arrow between the dotted lines (A). Artifact area as a percentage of the cervical area (%): calculated as the actual visibly distorted area (shown as grid lines) divided by the area of the cervical region (shown as horizontal lines). In the case of the sagittal or dorsal plane, the cervical region was defined as the area from the cranial end of C1 to the caudal end of C7 (B). Estimated entire artifact area (cm 2 ): calculated by drawing a round or oval that enclosed the visible artifact with the microchip as the central point (shown as grid lines) (C). A-C: sagittal planes. C1: the first cervical vertebra. *: brain. 4). The largest artifacts and artifact presence in the spinal cord are presented in Table 1. The surgery and the postoperative course were uneventful. The sutures were removed 14 days after surgery. Grossly, all microchips were situated within the encapsulation. On histopathological examination, all microchips were well-encapsulated with granulation tissue. Rupture of granulation capsule was not evident in dogs of both MRI and control groups. There were minor edema and hemorrhage in the granulation tissue and in the soft tissue outside the encapsulation area in dogs of both MRI and control groups. Cell debris was observed in the innermost layer of the granulation tissue regarding the frontal aspect of the implant in a dog of the MRI group, dog 1 (Fig. 4A), and a moderate level of degenerated cells with fibrin was seen in the inner layer of the granulation tissue of the frontal aspect of the implant in another dog of the MRI group, dog 2 (Fig. 4B). However, these changes were minor and were not consistent findings among the MRI group (Table 2). The stored ID data were read effectively with the reader in all microchips. DISCUSSION There are a few publications describing magnetic susceptibility artifacts in MR images in veterinary medicine [2, 3, 10]. A recent scientific article described magnetic susceptibility artifacts seen on 0.2 T MRI in animals after vertebral surgery [2]. Microscopic metal fragments from the surgical burr, suction tip, or other surgical instruments were thought to be the most likely source of artifacts in this report. Magnetic susceptibility artifacts or effects occur at inter-

4 578 M. SAITO ET AL. Fig. 2. Typical appearance of the microchip artifact on MR images in each sequence. The change was characterized as a central signal void with a high signal rim in T1WI (TR=500, TE=15, spin-echo), T2WI (TR=5000, TE=120, fast spin-echo), FLAIR (TR=8000, TE=120, TI=2300, fast spin-echo), and PDWI (TR=2800, TE=12, fast spin-echo) images. The range of the central void was huge, and a high signal rim was not evident in T2*WI (TR=450, TE=15, gradient-echo). D: dorsal. V: ventral. Cr: cranial. Ca: caudal. faces between substances with different magnetic susceptibilities such as air-tissue, fat-tissue, bone-tissue, and metaltissue. Among those, the artifact produced by magnetic material such as ferromagnetic metal is also called metallic artifact. In general, the extent of the susceptibility artifact depends on the magnetic susceptibility of the material, the degree of its own magnetization, and the external magnetic field strength [11]. Therefore, ferromagnetic substances, which have a strong magnetic susceptibility as well as magnetization, produce the largest artifacts, especially when placed in a higher-level magnetic field. For example, an experimental study showed that the artifact diameter generated by a cobalt (ferromagnetic) surgical instrument was more than 10 times longer than the actual instrument in a 1.5 T magnetic field. In contrast, the artifact from pure titanium (paramagnetic) was less than twice the actual size under the same conditions; however, as the magnetic field increased to 3.0 T, this extended to 10 times longer than the actual instrument, even using pure titanium [11]. A microchip consists of four essential components, an antenna, capacitor, connector wire, and its covering. Microchips have no battery inside, and they are designed to generate electricity in the antenna by electromagnetic induction using a low-radio-frequency-signal provided by the scanners. The metal composition differs between microchip manufacturers. The type of metal and its structure have a marked influence on the level of susceptibility artifacts. For example, the microchip produced by Datamars includes a ferrite core, which consists of ferromagnetic material. We asked for detailed information on the metallic components of the microchip used in this study, but could not get this Fig. 3. Transverse T2*WI (TR=450, TE=15, gradient-echo) MR image of a dog with an implanted microchip shows the microchip artifact extending over the spinal cord. Cr: cranial. Ca: caudal. information from the manufacturer. Although we did not test the ferromagnetism of the microchip, it was strongly attached to a conventional magnet. A typical susceptibility artifact on MRI is characterized as a central signal loss and image distortion with a rim of increased signal intensity around the vicinity of the magnetic substance or near junctions between tissues of differing magnetic susceptibility [2, 3, 5, 18]. A rim of increased signal intensity is not apparent when in the presence of a stronger ferromagnetism [9, 11]. Susceptibility artifacts elongate along frequency-encoding directions in spin-echo, fast spin-echo, and gradient-echo, and extend in the phasedencoding direction in echo planar imaging [19]. Our present findings were well-consistent with these characteristics of susceptibility artifacts, it is therefore clear that the artifacts shown in this study are susceptibility artifacts generated by the microchips. Since susceptibility artifacts cannot be prevented completely in patients with metallic implants, attempting to reduce these artifacts by selecting suitable pulse sequences and imaging parameters for MRI is recommended [4]. Several methods to achieve this have been introduced in experimental settings. In general, fast spin-echo imaging has less pronounced artifacts than spin-echo or gradient-echo. This is thought to be because the receiving bandwidth is usually set wider in fast spin-echo than spin-echo, and gradientecho is markedly influenced by magnetic inhomogeneities since gradient-echo sequences depend on gradient reversals to generate the echo [16]. With regard to the optimization of parameter sets in each pulse sequence, an increasing receiver bandwidth is effective in spin-echo if it can be altered by the operator, and not only increasing the receiver bandwidth, but also choosing a short TE is recommended in

5 MRI IN DOGS WITH MICROCHIPS 579 Fig. 4. (A) Photomicrograph of cutaneous tissue including granulation surrounding the microchip in dog 1 of the MRI group. H&E staining, magnification 40, bar=1,000 m. The inset shows the cell debris at the innermost layer of the granulation tissue encapsulating the microchip. H&E staining, magnification 100, bar=500 m. (B) Photomicrograph of cutaneous tissue including granulation surrounding the microchip in dog 2 of the MRI group. Note the moderate number of degenerated cells with fibrin seen at the inner layer of the granulation tissue. H&E staining, magnification 100, bar=200 m. (C) Photomicrograph of cutaneous tissue including granulation surrounding the microchip in dog 5 of the control group. Right: H&E staining, magnification 40, bar=1,000 m. Left (inset): H&E staining, magnification 100, bar=500 m. gradient-echo. It is known that TE does not usually have a marked influence on susceptibility artifacts in spin-echo [16, 19]. For the purpose of reducing susceptibility artifacts, there are other recommendations, including decreasing the voxel size: increasing the matrix, and/or decreasing the slice thickness, and/or decreasing the FOV, and changing the frequency encoding direction [4, 15]. It was reported that reduction of the susceptibility artifact can be achieved by positioning the long axis of the implant as parallel as possible to the main magnetic field [16]. Although the number of dogs in this study was too small to draw a conclusion, T2* W images (gradient-echo) tended to be more susceptible to microchip-induced artifacts and PDW images (fast spin-echo and shorter TE, and a wider receiver bandwidth in our system) was less affected compared to other sequences under our MRI conditions. The repeated exposure of a material to an external magnetic field can make the magnetic susceptibility of the material increase. We should be aware of this because repeated MRI is often necessary in certain patients. We experienced the phenomenon whereby the size of the microchip susceptibility artifact clearly increased on the second MR scan when compared to the first time in the same dog. It also should be considered that susceptibility artifacts can appear on images even when metal is placed outside the FOV [20]. It is not the intention of this report to discourage the use of microchips for animal s identification. However, it is important for veterinarians and manufacturers to be aware of the artifacts on MR images produced by microchips and educate animal s owners appropriately. The artifact produced by microchips implanted under the skin of the dorsal part of the cervicothoracic junction may cause difficulty in interpretation of MRI in this region, which includes the cervical to cranial thoracic spinal cord. The image of the brain may also be affected when the animal is small in size. Therefore, it may be advisable to implant microchips in

6 580 M. SAITO ET AL. Table 1. The largest artifacts and artifact presence in the spinal cord on each sequence Transverse Dorsal Sagittal estimated presence estimated presene artifact area estimated presence longest artifact area entire of artifact longest artifact area entire of artifact longest as % of centire of artifact Dog artifact as % of artifact on artifact as % of artifact on artifact cervical artifact on cm cervical area spinal cm cervical area spinal cm area area spinal DV* Lat* area cm 2 cord** crca* Lat* area cm 2 cord** DV* CrCa* cm 2 cord** T1W present avarage T2W present average FLAIR present present present average PDWI present present present average T2*WI present present present present present present average * DV, dorsoventral direction; Lat, lateral direction; CrCa, craniocaudal direction. ** present, there was at least one image where the artifact was present on the spinal cord structure; space, there was no image where the artifact was present on the spinal cord structure. Table 2. Comparison of histopathological changes in granulation tissue surrounding the microchip among MR and control groups Dog Encapsulation Hemorrhage Degeneration Edema MRI group Control group Plus (+), evident; Minus ( ), not evident. other region for specific breeds known to be predisposed to cervical spinal cord disorders, and small breeds known to be relatively common in brain diseases. An investigation of how to minimize artifacts caused by microchips in MR images was not carried out in the current study because it was beyond our purpose. Given the potential widespread application of microchips and the increasing use of MR imaging, further studies are necessary to determine how to minimize this type of artifact. Important potential complications on performing MRI in patients with metallic implants are heat injury and implant migration. There are various theories regarding the origin of this heat [12]. A radiofrequency pulse produced by an MRI scanner induces an eddy current on an electric conductor such as metals (electromagnetic induction). These eddy currents heat up the implants by means of joule heating. This is thought to be a major source of heat generation in tissues near the implant [1]. There are a number of factors known to affect the degree of this heating effect. Electrical properties such as the magnetic susceptibility, electrical resistivity, and thermal conductivity of metal materials of the implant have a significant influence on heat production. Other factors include the implant s size, shape, and orientation relative to the direction of the magnetic field, the magnetic field strength, and RF pulse duration and duty cycles [6, 12]. In cases where the implant is elongated, the heat pri-

7 MRI IN DOGS WITH MICROCHIPS 581 marily concentrates at the tip of the implant, and such a heat concentration point is called a hot spot [6, 12]. It has been shown that the tip of small metallic wire causes a major temperature rise in a static magnetic field of 1.5T [6]. The extent of the temperature rise is even higher when the implant is smaller or has an antenna [6, 12]. Evaluation of the tissue effects attributable to MRI-induced heating and migration of the implant have been conducted by histological examinations involving experimental animals to determine safety in patients with implants [9, 13]. In our histopathological results, there were no considerable differences in tissues of implant sites between MR and control groups and evidence for implant migration in the body was not observed. It may be possible that the degeneration in the area adjacent to the tip of the implant seen in one dog undergoing the MRI procedure was due to local tissue heating. This finding may reflect the fact that human patients with implants which are considered to be safe on MRI often feel pain or heat sensation during the procedure [12]. While the whole body average and the partial body specific absorption rate (SAR) value is a standard indicator of the temperature rise induced by a radiofrequency pulse, the necessity of obtaining information on local SAR values is increasing. For example, in an experimental setting, local SAR values near the tip of a 1-mm-diameter metallic implant increased up to 300 W/kg [6], whereas the current radiofrequency safety standard limit of the partial body SAR is 10 W/kg [8]. Therefore, it may be possible that our subtle pathological changes were caused by a local heating effect of MRI on the microchip; however, we think that the histopathological changes were too subtle to be of clinical significance, especially in dogs under general anesthesia. The stored ID data can be altered by a strong magnetic field, but the reading performance was not affected by the 1.5T magnetic field in this study. The results of this study suggest that the susceptibility artifact produced by an implanted microchip on 1.5T MRI can be marked if the implant is close to the region of interest. This can become a significant issue especially when the animal is small in size. The dogs with implanted microchips appeared to be scanned safely, and the microchips could be read effectively after 1.5T MRI. ACKNOWLEDGMENTS. Microchips were provided by Dainippon Sumitomo Pharma Co., Ltd. We thank students of departments of Surgery II for their technical assistance and animal care. REFERENCES 1. Budinger, T.F Nuclear magnetic resonance (NMR) in vivo studies: known thresholds for health effects. J. Comput. Assist. Tomogr. 5: Freer, S.R. and Scrivani, P.V Postoperative susceptibility artifact during magnetic resonance imaging of the vertebral column in two dogs and a cat. Vet. Radiol. Ultrasound. 49: Gavin, P.R Artifacts. pp In: Practical Small Animal MRI (Gavin, P.R. and Bagley, R.S. eds.), Wiley-Blackwell, Iowa. 4. Gross, B Findings of magnetic susceptibility artifacts of the vertebral column. Vet. Radiol. Ultrasound. 49: Guermazi, A., Miaux, Y., Zaim, S., Peterfy, C.G., White, D. and Genant, H.K Metallic artefacts in MR imaging: effects of main field orientation and strength. Clin. Radiol. 58: Ho, H.S Safety of metallic implants in magnetic resonance imaging. J. Magn. Reson. Imaging. 14: Kangarlu, A. and Robitaille, P.M.L Biological effects and health implications in magnetic resonance imaging. Concept Magnetic Resonance 12: Kawamitsu, H Safety of 3T-MRI. Nippon Hoshasen Gijutsu Gakkai Zasshi 64: (in Japanese). 9. Marra, S., Leonetti, J.P., Konior, R.J. and Raslan, W Effect of magnetic resonance imaging on implantable eyelid weights. Ann. Otol. Rhinol. Laryngol. 104: Martig, S., Konar, M., Schmökel, H.G., Rytz, U., Spreng, D., Scheidegger, J., Höhl, B., Kircher, P.R., Boisclair, J. and Lang, J Low-field Mri and arthroscopy of meniscal lesions in ten dogs with experimentally induced cranial cruciate ligament insufficiency. Vet. Radiol. Ultrasound. 47: Matsuura, H., Inoue, T. and Ogasawara, K. Quantitative analysis of magnetic resonance imaging susceptibility artifacts caused by neurosurgical biomaterials: comparison of 0.5, 1.5, and 3.0 Tesla magnetic fields Neurol. Med. Chir. 45: Muranaka, H., Nakamura, O., Usui, S., Ueda, Y. and Morikawa, K Influence of implants on human body during MRI examinations: fundamental experiment using metal balls. Nippon Hoshasen Gijutsu Gakkai Zasshi 61: (in Japanese). 13. Roguin, A., Zviman, M.M., and Meininger, G.R Modern pacemaker and implantable cardioverter/defibrillator systems can be magnetic resonance imaging safe: in vitro and in vivo assessment of safety and function at 1.5T. Circulation 110: Shellock, F.G. and Crues, J.V MR procedures: biologic effects, safety, and patient care. Radiology 232: tradiotti, P., Curti, A., Castellazzi, G. and Zerbi, A Metal-related artifacts in instrumented spine. Techniques for reducing artifacts in CT and MRI: state of the art. Eur. Spine. J. 18 (Suppl. 1): Suh, J.S., Jeong, E.K., Shin, K.H., Cho, J.H., Na, J.B., Kim, D.H. and Han, C.D Minimizing artifacts caused by metallic implants at MR imaging: experimental and clinical studies. Am. J. Roentgenol. 171: Takamura, T MRI jiyuu jizai., Medical View, Tokyo (in Japanese). 18. Teitelbaum, G.P., Bradley, W.G. Jr. and Klein, B.D MR imaging artifacts, ferromagnetism, and magnetic torque of intravascular filters, stents, and coils. Radiology 166: Tsuchihashi, T Artifact of MRI. Nippon Hoshasen Gijutsu Gakkai Zasshi 59: (in Japanese). 20. Tsuchihashi, T., Fujita, I., Iwasaki, A., Mori, K. and Suzuki, T Removal of high signal artifact (marching metal artifact) by the magnetic substance. Nippon Hoshasen Gijutsu Gakkai Zasshi 58: (in Japanese).