Magnetic resonance imaging conditionally safe neurostimulation leads: investigation of the maximum safe lead tip temperature.

RESEARCH—ANIMAL RESEARCH—ANIMAL

TOPIC

Magnetic Resonance Imaging Conditionally Safe Neurostimulation Leads: Investigation of the Maximum Safe Lead Tip Temperature Robert J. Coffey, MD* Ron Kalin, BS‡ James M. Olsen, BS, ME‡ *Research, Clinical, and Medical Safety, Medtronic Neuromodulation, Minneapolis, Minnesota; ‡Research, Medtronic Neuromodulation, Minneapolis, Minnesota Correspondence: Robert J. Coffey, MD, 353 Dryden St, Thousand Oaks, CA 91360. E-mail: [email protected] Received, August 5, 2013. Accepted, October 28, 2013. Published Online, October 30, 2013. Copyright © 2013 by the Congress of Neurological Surgeons

BACKGROUND: Magnetic resonance imaging (MRI) is preferred for imaging the central nervous system (CNS). An important hazard for neurostimulation patients is heating at the electrode interface induced, for example, by 64-MHz radiofrequency (RF) magnetic fields of a 1.5T scanner. OBJECTIVE: We performed studies to define the thermal dose (time and temperature) that would not cause symptomatic neurological injury. METHODS: Approaches included animal studies where leads with temperature probes were implanted in the brain or spine of sheep and exposed to RF-induced temperatures of 37C to 49C for 30 minutes. Histopathological examinations were performed 7 days after recovery. We also reviewed the threshold for RF lesions in the CNS, and for CNS injury from cancer hyperthermia. Cumulative equivalent minutes at 43C was used to normalize the data to exposure times and temperatures expected during MRI. RESULTS: Deep brain and spinal RF heating up to 43C for 30 minutes produced indistinguishable effects compared with 37C controls. Exposures greater than 43C for 30 minutes produced temperature-dependent, localized thermal damage. These results are consistent with limits on hyperthermia exposure to 41.8C for 60 minutes in patients who have cancer and with the reversibility of low-temperature and short-duration trial heating during RF lesion procedures. CONCLUSION: A safe temperature for induced lead heating is 43C for 30 minutes. MRIrelated RF heating above 43C or longer than 30 minutes may be associated with increased risk of clinically evident thermal damage to neural structures immediately surrounding implanted leads. The establishment of a thermal dose limit is a first step toward making specific neurostimulation systems conditionally safe during MRI procedures. KEY WORDS: Deep brain stimulation, Magnetic resonance imaging, Radiofrequency, Spinal cord stimulation, Thermal lesion Neurosurgery 74:215–225, 2014

DOI: 10.1227/NEU.0000000000000242

I

nteractions between magnetic resonance imaging (MRI) and implanted devices came to surgeons’ attention in the 1980s when magnetic forces upon ferromagnetic aneurysm clips and general surgical vascular clips were discovered to cause torsional movements sufficient to rupture previously secured blood

ABBREVIATIONS: ASTM, American Society for Testing and Materials; CEM, cumulative effect minutes; CNS, central nervous system; DBS, deep brain stimulation; RF, radiofrequency; SAR, specific absorption rate; SCS, spinal cord stimulation; WBHT, wholebody human hyperthermia

NEUROSURGERY

www.neurosurgery-online.com

vessels.1 Even after the phenomenon became well known, incidents of fatal hemorrhage occurred in patients who were thought to have been implanted with clips made from nonferromagnetic materials.2-9 More recently, heating of nonferromagnetic neurostimulation leads, cardiac leads, or other medical monitors, devices, or prostheses within the radiofrequency (RF) fields generated by MRI devices has been associated with injuries that varied from minor and/or temporary to serious, permanent, or fatal outcomes.10-17 The mechanism of metallic foreign body or medical implant heating involves MRI systems use of radio frequency (RF) coils to transmit and receive pulsed RF magnetic fields

VOLUME 74 | NUMBER 2 | FEBRUARY 2014 | 215

Copyright © Congress of Neurological Surgeons. Unauthorized reproduction of this article is prohibited.

COFFEY ET AL

in order to alter the spin axis of hydrogen nuclei. Pulsed RF magnetic fields also induce electric fields in body tissues and implanted metallic structures. Those induced electric fields, in turn, generate currents that dissipate at the electrode-tissue interface and cause heating of the surrounding tissue. Cases have occurred where MRI procedures caused permanent tremor suppression owing to thalamotomy effects or caused temporary dystonia in patients with fully implanted deep brain stimulation (DBS) systems.14 One reported case described permanent injury in a bilaterally implanted DBS patient who had 1 neurostimulator (the one involved in the complication) implanted in the abdominal wall—to allow participation in shooting sports—instead of in the infraclavicular region.15 A hemorrhagic lesion occurred surrounding the lead in the subthalamic nucleus. The multiseries total spinal axis MRI scan that the patient underwent used a 1.0T system at a high specific absorption rate (SAR) by using a body transmit coil—all of which constituted off-label conditions. Although product-labeling instructions exist for safe MRI of patients with neurostimulation systems, general guidelines remain unsettled. Reasons include the fact that DBS leads are in direct contact with brain tissue, whereas epidural spinal cord stimulation (SCS) leads are separated from neural structures by intervening tissues and cerebrospinal fluid (CSF); imaging centers may be unaware that a patient has a neurostimulation system other device implanted; patients may answer a questionnaire incorrectly; a variety of MRI coil arrangements are in use; physicians and staff may be unaware of contraindications, limitations, or special precautions required to image implanted patients safely; and those personnel may be unaware that different implant sites (brain, spinal canal, other) or devices from different manufacturers are subject to different precautions or contraindications depending upon the brand, field strength, coil arrangement, and imaging sequences. Local vascular anatomy and blood flow—which vary among individuals and influence heat dissipation around implanted leads—also can influence whether a clinically detectable thermal injury occurs. This multiplicity of physical, anatomic, and physiological variables makes the relationship between thermal dose and potential neurological injury probabilistic in nature. We performed experiments in an ovine model to study the histopathological effects of different thermal doses on spinal and brain tissue when administered through custom heating probes that replicated the geometry and dimensions of neurostimulation leads, and delivered RF energy at the same 64-MHz frequency as generated by 1.5T scanners in common clinical use. Our goal was to find the MRI-induced thermal dose (temperature and time duration) that the brain and spinal tissues could tolerate with an estimated zero risk of clinically detectable thermal injury. The heating probes resided in spinal or brain tissues of interest and administered lead tip temperatures from 37C to 49C for 30-minute exposures. Forty-nine degrees is within the range of temperatures (although below the maximum temperature) that has been observed in anthropomorphic phantoms when neurostimulation systems were exposed to conventional MRI pulse sequences at clinical imaging centers

216 | VOLUME 74 | NUMBER 2 | FEBRUARY 2014

(American Society for Testing and Materials [ASTM] Phantom, Zurich Med Tech AG, Zurich, Switzerland http://www.dymstec. com/img/501-QDASTM01AA-A.pdf). We then compared our experimental findings—using the cumulative effect minutes (CEM) thermal dose formula (widely used in the scientific community and applicable to this limited temperature range)—to the published experience with whole-body human hyperthermia (WBHT, cancer therapy) and to central nervous system (CNS) lesioning by using RF heating techniques.18-30 The animal studies and literature research described here focused on a 30-minute exposure session at maximum energy (maximum temperature), which is a conservative time frame to obtain adequate diagnostic images at a nominal 64-MHz RF energy at maximum power (2W/kg SAR) in normal mode on 1.5T (1.5T) MRI systems. An important limitation is that the experiments and calculations described herein do not necessarily predict or model lethal or sublethal thermal injury on a neuronal cellular level. Rather, we applied thermal dose conversion formulas used in thermal cellular pathology research to compare our experimental observations against a body of hyperthermia and RF lesion research. One underlying safety-related principle is the notion that as long as metallic wires are used to conduct electricity between a neurostimulator and electrode leads, all implanted neurostimulation leads—even MRI-compatible ones—will undergo a temperature rise in the MRI environment, albeit less than non-MRI-compatible systems. Another limitation is that we address only a limited foundational question about MRI of implanted patients—namely, to define the safe MRI-induced thermal dose that the brain and spinal tissues can tolerate with no risk of injury. Our answer— 43C for 30 minutes—leaves considerable work for device manufacturers to redesign neurostimulation systems that permit safe imaging of patients with MRI scanners in normal operating mode at 2W/kg whole-body SAR (or 3.2W/kg head SAR).

MATERIALS AND METHODS Ethics, Rationale, and Limitations The good laboratory practices studies described here were conducted in compliance with the Animal Welfare Act of 1966 and all amendments (Public Law 89-544, Animal Welfare Act of August 24, 1966). The use of the sheep model to predict injury in humans was acceptable and conservative, especially for intraspinal structures, because the smaller size of sheep spine used with human-sized leads would increase the likelihood of detecting any thermal injury effects. Physiological Research Laboratory is registered with the United States Department of Agriculture, Animal and Plant Health Services, and is accredited by the American Association for Accreditation of Laboratory Animal Care. Adult sheep of mixed breeds and sex were selected following a prescribed quarantine and were cared for according to The Guide for Care and Use of Laboratory Animals.31 All necessary veterinary care was allowed to maintain the health and comfort of study animals, including reparative surgeries and drug administration. The need to assess histological findings associated with a series of precisely specified time and temperature exposures influenced the use 64-MHz RF energy to replicate the MRI environment rather than performing experiments in the less predictable actual MRI environment—where the magnitude and



MRI CONDITIONALLY SAFE NEUROSTIMULATION LEADS

time course of lead heating could not be tightly controlled. This method facilitated compliance with laboratory practice guidelines by permitting the acquisition of multiple data points from each experimental animal, thereby limiting the number of animals required. With respect to the range of temperatures studied, MRI-induced heating of non-MRI conditionally safe chronically implanted neurostimulation leads can reach temperatures up to 70C—well beyond the 49C upper boundary of our experiments. We limited the highest temperature to 49C because hyperthermia research and CEM43 calculations already have established a 100% likelihood of neural tissue injury after 30 minutes of exposure above 49C.18-27 A related reason to limit the upper range of temperature exposures was the requirement for 7 days of survival for histological examinations at uniform intervals for each heated location. Heating any brain or spinal site to temperatures greater than required by the experimental objectives risked the causation of neurological deficits that would compel early euthanization and loss of data for the affected animal. Our goal was not to explore how much thermal damage could be induced, but to find the boundary or threshold temperature where exposure for 30 minutes would have an approximately zero probability of causing permanent neural injury.

Experimental Heating and Sensing Leads We constructed custom heating and sensing leads to achieve local temperatures within the range of heating that can occur at the lead contacts in patients implanted with DBS or SCS systems during 1.5T MRI studies of the brain or spine in a 64-Mhz RF field. The leads for each set of investigations (Figure 1) had approximately the same diameter as

currently marketed DBS or SCS leads (1.27 mm). Insulation over the outer shield of the 50 V coaxial cable (used as the lead body) was removed for a length of 1.0 cm to act as a return path for the RF current. A control system regulated the RF power to maintain the selected temperature for each heating session to within 1 standard deviation less than (6) 0.3C in the brain and in the spinal epidural space (Figure 2). Temperatures were measured using integral fluoroptic temperature probes (Luxtron Model STB, LumaSense Technologies, Inc, Santa Clara, California) positioned adjacent to and in contact with the RF-heated lead tips. During heating sessions, the exposed lead area heated surrounding tissues to the intended lead tip temperature (control, 43C, 45C, 47C, or 49C) for a radial distance of approximately 3 mm.

DBS Sheep Brain Eight female Rambouilet sheep underwent bilateral burr-hole surgery under general endotracheal anesthesia to implant heating-sensing leads into the lentiform nucleus (caudate-putamen, rostral target), thalamus (caudal target), and adjacent white matter of each hemisphere (total, 4 leads per animal). MRI guided the implants using a Stealth Treon station (Medtronic Navigation, Louisville, Colorado). After the leads were implanted, RF-induced heating was applied to achieve a stable temperature for 30 minutes at 43C, 45C, or 49C at the 3 heated sites, or left unheated at normal core temperature (37.2C-39.1C) at the fourth lead site as a control in each animal. After all locations were heated, the leads were withdrawn, the incisions were closed, and the animals were allowed to recover for at least 7 days of observation. Thereafter, the animals were

FIGURE 1. Mechanical drawings and photographs of the heating-sensing brain and spinal leads. The exposed heating tip of the simulated spinal lead (Lower Left) is 6 mm long · 1 mm diameter; the simulated DBS brain lead (Lower Right) is 3 mm long · 1 mm diameter. The maximum width of the heating tip plus temperature monitoring probe is 1.5 mm (60/100 inches). Scale in the photographs is each line = 0.5 mm. Transparent heat-shrink tubing covers all but the exposed heating-sensing tip of the leads. DBS. deep brain stimulation.

NEUROSURGERY



COFFEY ET AL

SCS Sheep Spine Five sheep of mixed varieties and both sexes underwent surgical exposure of the upper lumbar spine under general endotracheal anesthesia for dorsal epidural placement of the SCS heating-sensing lead. Intraoperative fluoroscopy was used to localize RF heat administration at 5 different nonoverlapping sites along the spinal axis in each animal, separated by at least 2 vertebral segments (variable between L1 and C5). This was to maximize the amount of data obtained from a minimal, yet adequate number of animals. The temperature at each site was an active surgical control (core temperature with no RF heating) or 43C, 45C, 47C, or 49C for 30 minutes. One cervical section above the highest lead location site in each animal was used as an additional nonsurgical control for histopathological examination. The lead was withdrawn and the incision closed in each animal, after which they recovered for 7 days before euthanization to allow a detectable tissue response to develop.

RESULTS DBS Sheep Brain Table 1 and Figure 3 summarize the histopathological findings in the DBS sheep cohorts. Only discrete microscopic changes were found at sites exposed to 43C for 30 minutes. The difference between the 37C (control) and the 43C sites was minimal, indicating that 43C for 30 minutes appeared safe from a histopathological perspective within the gray and white matter sites investigated in this study. Exposure to 45C revealed a slightly broader zone of thermal injury effects in both gray and white matter, and 49C produced unequivocal evidence of thermal effects at all sites (Figures 3 and 4). Another observation was the lack of variation with temperature of the zone of apparent cerebral tissue necrosis (gap, or absence of tissue, as interpreted by the pathologist). This was a feature that resulted from lead implantation, not thermal injury. In contrast, the zone of reactive gliosis around the heated portion of the DBS lead varied with temperature—and appeared responsible for the progressively larger width of the lead tract plus the zone of gliosis at temperatures of 45C and 49C, respectively.

FIGURE 2. Experimental setup and heating profile over 30 minutes of the experimental brain and spinal leads. Top, artist’s rendering of the equipment used in the sheep brain and experimental models. The brain and spine equipment are shown together for illustration purposes, only, because the experiments were conducted separately and each animal was used for brain or spine testing (not both). Center, the lead tip reached the programmed temperature immediately (43C in this example) and held steady for 30 minutes (1800 seconds). Bottom, the temperature in tissue surrounding the experimental lead fell off rapidly by $1C per millimeter distance away from the lead surface. RF, radiofrequency.

euthanized and gross and quantitative histopathological examinations of the brain were performed. Each set of histological sections containing the lead tract was evaluated in a blinded fashion by 2 veterinary neuropathologists and scored according to the degree of necrosis (tissue loss or gap along the lead tract), surrounding gliosis, hemorrhage, and diameter of the tract plus surrounding gliosis.


SCS Sheep Spinal Cord Table 2 and Figure 3 summarize the histopathological findings in the SCS sheep cohorts. In addition to unheated thermal control sites where the experimental SCS lead was located in the adjacent epidural space, but not heated—the SCS experiments also included pathological examination of surgical control sites in the mid and lower cervical segments, above the highest level of lead tip implantation. The thermal control group and the heated animal groups—all of which underwent lead implantation— displayed similar pathological findings in the epidural fat and vessels, including blood extravasation (minor hemorrhage) in the epidural space surrounding the leads. Absence of such findings in the surgical control group (no lead implant) indicates that the findings were caused by the lead implantation procedure, and not related to heating. No neural tissue findings different from the control groups occurred in the 43C cohort. Pathological changes




TABLE 1. Average Width of DBS Lead Tract and Gliotic Zone Affected at Each Temperaturea,b

a b

No. of Sites Heated and Examined

Temperature, °C

8 9 7 3

Control (37.2-39.1) 43 45 49

Width of Necrosis Zone, Mean (Range), mm ,0.5 ,0.5 0.5-1.0 .1.0

Mean Hemorrhage Score

Mean Gliosis Score

Mean Width of Lead Tract 1 Gliosis, mm

3.0 1.9 1.4 1.3

2.0 2.4 1.6 2.2

0.9 0.9 1.4 2.3

(0-,0.5) (0-1.0) (0.5-1.0) (all . 1.0)

DBS, deep brain stimulation. Hemorrhage and gliosis scores: 0, none; 1, minimal; 2, mild; 3, moderate; 4, extensive.

in the spinal cord, adjacent nerve roots, dorsal root ganglia, and/ or dura were observed to some degree at all spinal sites heated to between 45C and 49C, with more intense findings at higher temperatures (Figures 3 and 5).

REVIEW OF HYPERTHERMIA, RF LESIONS, THERMAL MODELS, AND PROBABILITY OF INJURY Whole-Body Hyperthermia Whole-body hyperthermia in cancer therapy exploits the higher drug sensitivity of rapidly dividing tumor cells (vs normal cells) to chemotherapeutic agents when subjected to elevated temperatures. Related research suggests that heating the mammalian CNS to 43C for 30 minutes causes no clinically detectable adverse effects, and that whole-body temperatures up to 41.8C for 60 minutes causes only minimal or clinically acceptable adverse effects, depending on the particular study.18,20-22,24-26 Although we could find no clinical WBHT data for human exposures at

FIGURE 3. Average width of DBS lead tract and gliotic zone, and percentage of SCS spinal sites affected at each temperature. The width of the experimental lead tract plus gliotic zone (gray-dashed line, left-hand vertical axis) and the percentage of spinal cord and nerve root sites (sections) that revealed thermal injury (black solid line, right-hand vertical axis) increased proportionally at temperatures greater than 43C. DBS, deep brain stimulation; SCS, spinal cord stimulation.

NEUROSURGERY

43C for exactly 30 minutes, thermal dose conversion formulas suggest that heating the entire brain (and the entire body) to 42 C for 60 minutes carries a higher likelihood of adverse neurological effects than focal heating restricted to the small volumes pertinent to MRI interactions with neurostimulation leads. For example, in small animals with brain sizes that are orders of magnitude smaller than human brains, temperatures approaching 43C for 30 minutes did not produce detectable thermal injuries or neurological sequelae. RF Thermal Lesions Radiofrequency thermal techniques are used in selected disorders to inflict therapeutic lesions in the nervous system or in the conducting system of the heart. A temporarily inserted probe applies RF energy until the desired temperature-time end point is achieved.28-30 Cases also have been reported where deliberate RF lesions were created through DBS electrodes that were scheduled to be explanted because of infection or as part of a staged lesioning study (Table 3).32-34 Determinants of RF lesion size in neural tissue—in descending order of importance—are the probe dimensions (exposed conductive length and diameter), probe temperature, and duration of heating. The physical, chemical, and electrical properties of gray and white matter, the proximity and density of blood vessels in and near the target regions (heat sink effects), and the proximity of some targets to the ventricles or subarachnoid space (additional CSF heat sink effects) can mitigate thermal injuries, and require surgeons to use relatively high temperatures (65C-70C, or higher) for 1 to 2 minutes to create permanent lesions. Limited autopsy data reveal that lasting effects required temperatures $65C for $60 seconds.30 An unpublished tradition in functional neurosurgery also involves the performance of test or trial RF-thermal heating in awake patients at 45C for 60 to 120 seconds, followed by observation for immediate therapeutic effects or side effects. Lack of side effects at successively higher temperatures indicates that it is safe to escalate stepwise to 65C to 70C (or higher) to produce a final permanent lesion. Thermal Dose Comparisons Above 43°C All thermal dose conversion methods entail limitations (addressed in the Discussion), especially with respect to accurately



COFFEY ET AL

FIGURE 4. Seven-day brain histopathology examples at each lead temperature. All temperature exposures were 30 minutes. Photomicrographs are hematoxylin and eosin stained. The blue horizontally angled line in each photomicrograph is the widest portion of the lead tract and/or zone of necrosis plus surrounding gliosis that was measured in each section. All measurements were taken at the site of the exposed heating element, between 1 and 3 mm proximal (superficial) to the lead tip and deepest extent of each lead tract. A, temperature = 37C. The unheated control tract contains a small amount of hemorrhage surrounded by a rim of gliosis. The findings are all related to lead implantation and removal; no heat was administered at this site. B, temperature = 43C. The tract contains hemorrhage at the margins surrounded by a rim of gliosis. Small foci of amorphous eosinophilic material are similar in appearance to the 37C sites, which makes this finding likely due to the insertion and removal of the experimental lead. C, temperature = 45C. The lead tract shows a focus of necrosis with the same pathological features as Figure 5A, but with a narrower width profile. D, temperature = 49 C. The deep end of the lead tract shows focal necrosis (amorphous pale eosinophilic material) surrounded by a rim of gliosis.

predicting or modeling heat and time effects on cellular, subcellular, and molecular processes and structures. We used CEM43, cumulative equivalent minutes of exposure at 43C, only as an expression to compare temperature and time exposures within a few degrees higher than body temperature. Bearing in mind that CEM43 may not accurately model or predict thermal effects at the subcellular level, use of 43C as a thermal injury break point arose empirically from cell culture clonogenicity studies.23,27,35 The CEM43 equation allows one to calculate an isoeffective thermal dose (causing the same degree of tissue damage) at temperatures other than 43C and for different durations of exposure. “R” is the number of minutes needed to compensate for a 1C temperature change either above or below the break point

of 43C. For temperatures above 43C, R is approximately 0.5 (for most tissues), “Ti” is the temperature during time interval “i” (in degrees Centigrade), and “ti” is the exposure time in minutes at temperature “Ti” in the following equation:

CEM 43∘ C 5 i

t X R 4~3Ti : i

In the example of RF lesions, trial heating at 45C to 50C for 1 minute may induce only temporary effects.29 Because 50C for 1 minute is equivalent to 45C for 32 minutes according to the CEM formula, one may estimate that a DBS lead heated to 45C

TABLE 2. SCS: Summary of 7-Day Spinal Tissue Responses at Each Temperaturea,b

Temperature, °C

Spinal Cord Lesion(s), n (Dimensions)

Nerve Root or Dorsal Root Ganglion Lesion(s): Proportion of Sections

5 6

Surgical control (no lead) Thermal control (37)

1c (1.0 · 2.0 mm) None observed

None observed None observed

5

43

None observed

None observed

6

45

None observed

1/6 (nerve root, only)

6

47

3/6 (barely detected to 0.8 · 0.5 mm)

3/6 (nerve roots, only)

6

49

3/6 (from 0.8 · 0.25 to 2.6 · 1.6 mm)

4/6 (3 roots, 1 ganglion)

No. of Sites

Epidural Hemorrhage, n, grade None observed 3/6 Grade 1-2.5 2/5 Grade 1 3/6 Grade 1.5-2 5/6 Grade 0-1.5 4/6 Grade 0-2

a

SCS, spinal cord stimulation. Epidural hemorrhage grades: 0, none; 1, minimal; 2, mild; 3, moderate; 4, extensive. c Coincidental microscopic finding in the C-6 spinal cord segment of 1 surgical control animal that had no lead implanted and no heating at this vertebral level. b





FIGURE 5. Spinal cord and nerve root histology examples at each lead temperature. All temperature exposures were 30 minutes. Photomicrographs are hematoxylin and eosin stained unless otherwise noted. A, B, temperature = 43C. A, one atypical nerve root was identified. B, one abnormal dorsal root ganglion was found with mildly shrunken neurons on the same side as the atypical peripheral nerve (Luxol fast blue stain). Both of these neuronal changes may be artifactual. C, D, temperature = 45C. C, T-3 nerve roots (peripheral nerve histology). Atypical nerve rootlets at T-3 revealed swollen myelin sheaths lacking axons and mild to moderate inflammation surrounding the epineurium in the epidural fat (boxed area). A nearby nerve rootlet appeared normal (dashed oval). D, a higher-power photomicrograph of the boxed area in E. No other abnormalities in any of the animals were found at sites heated to 45C. E, F, temperature = 47C. E, spinal cord. One small focus of necrosis (boxed area) lateral to the dorsal horn at T3 is shown at higher power in F. G, H, temperature = 49C. G, spinal cord. A zone of bilateral liquefactive necrosis that measures 2.5 mm wide · 1 mm deep is present at the superficial edge of the dorsal columns at the T10 vertebral level. A thin rim of axonal spheroids and swollen myelin sheaths surrounds the necrotic zone—consistent with a 7-dayold thermal injury. H, high-power photomicrograph of a nerve root fiber (peripheral nerve histology) adjacent to the heating site in a different animal. A thermally damaged axon is present (circle) among other preserved (normal) axons.

for 30 minutes also has a very low likelihood of causing a permanent clinical effect. MRI Testing and DBS Clinical Case Series Current labeling for DBS leads and neurostimulator systems restricts head scans to the use of transmit-receive head coils

at 0.1 W/kg SAR. Tests to support this limit determined that a maximum temperature rise of 2C would occur during a 15-minute scan at 0.1 W/kg with the leads and extensions routed in the customary manner in the original ASTM anthropomorphic phantom. Although the general rule holds true that MRI scans performed at higher SAR levels generate

TABLE 3. Comparison of Exposed Electrode Dimensions for RF Lesion Probes, Experimental Leads, and Commercial DBS Leadsa Lead Brain autopsy 130 Brain autopsy 230 Brain autopsy 330 Brain autopsy 430 Brain autopsy 530 Sheep—SCS Sheep—DBS Commercial 338732,33 Commercial 338932,33 a

Exposed Contact(s): Length · Diameter, mm

No. of Contacts and Spacing, mm

Temperature or Current · Time (°C · minutes)

Lesion Size: Length · Diameter, mm

5.0 · 1.1 3.0 · 1.2 5.0 · 1.6 10 · 1.6 10 · 1.6 6 · 1.2 3 · 1.2 1.5 · 1.2 1.5 · 1.2

1 contact 1 contact 1 contact 1 contact 1 contact 1 contact 1 contact 4 at 1.5 mm 4 at 0.5 mm

72 · 6 65 · 2 70 · 1 80-90 · 1 80 · 1 49 · 30 49 · 30 40 V, 400 V · 1 40 V, 400 V · 1

7·3 4·2 8·8 10-12 · 10 12 · 10 2.6 · 1.6 (maximum) 2.3 4 6 0.8 · 3 6 0.6 4 6 0.8 · 3 6 0.6

DBS, deep brain stimulation; RF, radiofrequency; SCS, spinal cord stimulation.

NEUROSURGERY



COFFEY ET AL

higher temperatures proportional to the increase in SAR, the IEC-ISO Joint Working Group (International Organization for Standardization and the International Electrotechnical Commission) subsequently found—owing to differences between the phantoms and human tissues—that the current SAR limits-based ASTM phantom tests probably are more conservative and have a greater safety margin than originally believed. Other transmitreceive head coil labeling for brain imaging in SCS patients restricts RF power to 1.5W/kg SAR, and to 2W/kg SAR for sacral nerve stimulation patients—searchable on manufacturers Web sites or at ,http://www.mrisafety.com/safety_info.asp.. Our sheep brain and spine experiments support the findings reported by Larson et al36 and Tagliati et al37 that neural tissue in implanted patients appears to tolerate temperature rises greater than the present limit of 2C (0.1 W/kg SAR) for 15 minutes. The results of our experiments, the oncological experience with WBHT, the neurosurgical practice of RF lesioning and test lesioning, and large series of MRI scans in implanted patients now suggest that 30 minutes of RF-induced heating of neural tissue up to 43C—a temperature rise of #6C above normal— should be a safe no-adverse-effect level for the next generation of implantable neurostimulation systems. Probability of Clinical Effects Above 43°C The experimental data and calculations presented here suggest that 43C is safely tolerated by the brain, spinal cord, dorsal root ganglia, and nerve roots for a 30-minute exposure to 64-MHz RF energy in a 1.5T MRI scanner for 30 minutes. However, errors in clinical practice, or use of non-MRI approved leads, can lead to MRI exposures greater than 30 minutes per session and may lead to RF-induced lead temperatures higher than 43C. In vitro data, autopsy data (albeit limited), human clinical RF lesion data, the animal studies reported here, and analysis of clinical case series and reports allow one to estimate—albeit in a conjectural manner—the risks that might be associated with unintentionally high thermal exposures.28-30,32-34,36,37 With respect to electrode array dimensions, the postmortem effects summarized in rows 1 to 5 of Table 3 resulted from lesions inflicted by using probes that encompass the dimensions of currently marketed DBS leads and of the experimental leads used in our sheep brain and spine studies. This allows one to reasonably apply CEM43 calculations in a strictly limited manner to compare time and temperature thermal dose exposures extracted from the reported neurosurgical RF lesion data. The RF lesions induced in our experiments were smaller than the smallest human thalamotomy lesion found at autopsy (4 mm · 2 mm) after heating to 65C for 2 minutes by the use of a 3 mm · 1.2 mm probe. The CEM43 thermalequivalent dose for that thalamotomy was higher than what can be induced in MRI conditionally safe leads at 64-MHz RF energy, 2W/kg SAR in normal mode on a 1.5T system. However, as evidenced by phantom measurements and case reports of unintentional lesions during MRI procedures, temperatures of 52C to 70C can be induced in nonconditionally safe leads.


In terms of MRI safety for brain and spinal leads, 30-minute exposures to temperatures between 43C and 45C appear to carry a relatively low likelihood of clinically apparent (symptomatic) neurological injury. Longer exposures and temperatures higher than 45C increase the risk of direct neural injury or neurovascular injury with clinical consequences dependent on lead location. Heating brain or spinal leads to 49C for 30 minutes will always cause histologically detectable injury to adjacent neural or vascular tissues. However, to predict whether MRI-induced thermal lesions will cause pain, transient symptoms, or permanent deficits in human patients, one must account for the lead contact locations relative to white matter tracts, gray matter nuclei, nerve roots, or pertinent blood vessels. Because cranial and spinal lead locations are almost universally documented by postoperative imaging studies, clinicians can estimate which neural structures or pathways are at risk if a patient was to undergo MR imaging where labeled limitations were exceeded (eg, use of higher than recommended SAR, or prolonged scanning time). However, apart from direct neural tissue injury, excessive heating can damage small blood vessels that are invisible on MRI and thereby cause intracranial bleeding. If doubts arise before imaging a particular patient, physicians or technicians can consult the device manufacturer or specific labeling documents available at online resources.

DISCUSSION Methods and Limitations Ovine models were chosen for the following reasons. Among the medium-sized animals available for study, the sheep’s neural axis is most similar to humans. Sheep are similar in size to small (60 kg) humans and have been used in other studies.23,24 A 7-day follow-up period was selected for the histological evaluation of RF thermal damage to the brain, spinal cord, nerve roots, ganglia, and epidural tissues because neuropathological findings after heating (or any insult or injury) are most apparent after that interval. Our use of 64-MHz RF energy at precisely controlled temperatures and times—instead of in the less well controlled clinical MRI environment—provided thermal dosage data of high integrity and allowed the testing of multiple sites to limit the number of animals used. Both mechanical and thermal controls were used in the SCS spinal studies to assure an accurate interpretation of the findings. The standard method for assessing hyperthermia, CEM43, is adequate as used here to normalize thermal doses (thermal histories) from different data sets between body temperature and 49C to 50C. Owing to simplifying assumptions, CEM43 does not accurately reflect the multiple and complex cellular and subcellular thermally induced biochemical alterations that can cause neuronal cell death or sublethal loss of neuronal function. Arrhenius calculations and other mathematical models are more accurately used to quantify irreversible damage or alterations to molecular processes within living cells—topics beyond the scope of our investigation.35 The experience of neurosurgeons using RF lesioning in the CNS to treat movement disorders and chronic




pain supports the notion that thermal effects can be reversible for temperatures #45C for a few minutes. Cellular and subcellular events have been investigated elsewhere, but do not address the temperature-time threshold for clinical injury from MRI-induced heating of neurostimulation leads. Animal Studies The sheep DBS lead data dovetail with information gleaned from other lines of experimental and human clinical investigations, including reports of MRI scans in DBS patients where no complications occurred, or where temporary or permanent sequelae occurred. The DBS sheep data also comport with the RF lesion experience in functional neurosurgery. Our animal findings in the brain and spine further agree with the experimental and clinical whole-body hyperthermia experience, and with CEM 43, a valid conversion model for temperature-time exposure. The sheep brain and spinal data revealed that a 30-minute exposure to RF heating of DBS or SCS leads to 43C for 30 minutes produced barely discernible microscopic effects beyond those associated with mechanical effects of implanting the leads alone. A temperature of 43C, an approximate 6C temperature rise above the normal body temperature, appears to be a reasonably conservative upper temperature limit with which to establish a criterion for zero risk of clinically observable neurological signs and symptoms. The 43C for 30 minutes limit also appears to provide an adequate safety margin in the neural tissues and surrounding spinal structures that we analyzed, and translates to a no-clinical-effect prediction when compared with WBHT and RF lesioning data. The sheep data, in light of the RF test lesion experience and CEM43 predictions, indicate that only a slight increase in thermal tissue damage occurs between 43C and 45C. In contrast, 30 minutes at 49C produced thermal injuries at all sites in the brain and spinal studies and represents a thermal dose that could cause clinically evident neural injury.

CONCLUSION The experimental and clinical evidence presented here suggest that no clinically detectable neural damage occurs in brain, spinal cord, or nerve roots at localized temperatures #43C for 30 minutes. These conclusions are based on convergent lines of research that involve brain and spinal experimental data, the use of therapeutic RF lesioning of the brain, reviews of published MRI device-related adverse events, and the clinical and experimental literature on WBHT. Disclosure The authors have no personal, financial, or institutional interest in any of the drugs, materials, or devices described in this article.

REFERENCES 1. Price RR. The AAPM/RSNA physics tutorial for residents. MR imaging safety considerations. Radiological Society of North America. Radiographics. 1999;19(6): 1641-1651.

NEUROSURGERY

2. Barrafato D, Henkelman RM. Magnetic resonance imaging and surgical clips. Can J Surg. 1984;27(5):509-510, 512. 3. Laakman RW, Kaufman B, Han JS, et al. MR imaging in patients with metallic implants. Radiology. 1985;157(3):711-714. 4. Brown MA, Carden JA, Coleman RE, McKinney R Jr, Spicer LD. Magnetic field effects on surgical ligation clips. Magn Reson Imaging. 1987;5(6):443-453. 5. Becker RL, Norfray JF, Teitelbaum GP, et al. MR imaging in patients with intracranial aneurysm clips. AJNR Am J Neuroradiol. 1988;9(5):885-889. 6. Romner B, Olsson M, Ljunggren B, et al. Magnetic resonance imaging and aneurysm clips. Magnetic properties and image artifacts. J Neurosurg. 1989;70(3):426-431. 7. Klueznik RP, Carrier DA, Pyka R, Haid RW. Placement of a ferromagnetic intracerebral aneurysm clip in a magnetic field with a fatal outcome. Radiology. 1993;187(3):855-856. 8. Kanal E, Shellock FG, Lewin JS. Aneurysm clip testing for ferromagnetic properties: clip variability issues. Radiology. 1996;200(2):576-578. 9. Dujovny M, Alp MS, Dujovny N, et al. Aneurysm clips: magnetic quantification and magnetic resonance safety. J Neurosurg. 1997;87(5):788-794. 10. Shellock FG, Slimp GL. Severe burn of the finger caused by using a pulse oximeter during MR imaging (letter). AJR Am J Roentgenol. 1989;153(5):1105. 11. Shellock FG, Kanal E. Guidelines and recommendations for MR imaging safety and patient management. III. Questionnaire for screening patients before MR procedures. The SMRI Safety Committee. J Magn Reson Imaging. 1994;4(5):749-751. 12. Shellock FG, Shellock VJ. Vascular access ports and catheters: ex vivo testing of ferromagnetism, heating, and artifacts associated with MR imaging. Magn Reson Imaging. 1996;14(4):443-447. 13. Gleason CA, Kaula NF, Hricak H, Schmidt RA, Tanagho EA. The effect of magnetic resonance imagers on implanted neurostimulators. Pacing Clin Electrophysiol. 1992;15(1):81-94. 14. Spiegel J, Fuss G, Backens M, et al. Transient dystonia following magnetic resonance imaging in a patient with deep brain stimulation electrodes for the treatment of Parkinson disease. J Neurosurg. 2003;99(4):772-774. 15. Henderson J, Henderson JM, Tkach J, Phillips M, Baker K, Shellock FG, Rezai AR. Permanent neurological deficit related to MRI in a patient with implanted DBS electrodes for Parkinson’s disease: case report. J Neurosurg. 2005; 57(5):E1063. 16. Newcombe VFJ, Hawkes RC, Harding SG, et al. Potential heating caused by intraparenchymal intracranial pressure transducers a 3-tesla magnetic resonance imaging system using a body radiofrequency resonator: assessment Codman MicroSensor Transducer Technical note. J Neurosurg. 2008;109:159-164. 17. Tanaka R, Yumoto T, Shiba N, et al. Overheated and melted intracranial pressure transducer as cause of thermal brain injury during magnetic resonance imaging Case report. J Neurosurg. 2012;117(6):1100-1109. 18. Fike JR, Gobbel GT, Satoh T, Stauffer PR. Normal brain response after interstitial microwave hyperthermia. Int J Hyperthermia. 1991;7(5):795-808. 19. Dewey WC. Arrhenius relationships from the molecule and cell to the clinic. Int J Hyperthermia. 1994;10(4):457-483. 20. Takahashi S, Tanaka R, Watanabe M, et al. Effects of whole-body hyperthermia on the canine central nervous system. Int J Hyperthermia. 1999;15(3):203-216. 21. Lee SY, Lee SH, Akuta K, Uda M, Song CW. Acute histological effects of interstitial hyperthermia on normal rat brain. Int J Hyperthermia. 2000;16(1):73-83. 22. Wust P, Hildebrandt B, Sreenivasa G, et al. Hyperthermia in combined treatment of cancer. Lancet Oncol. 2002;3(8):487-497. 23. DeWhirst MW, VigliantiI M, Lora-Michels M, Hanson M, Hoopes PJ. Basic principles of thermal dosimetry and thermal thresholds for tissue damage from hyperthermia. Int J Hyperthermia. 2002;19(3):267-294. 24. McDannold N, Vykhodtseva N, Jolesz FA, Hynynen K. MRI investigation of the threshold for thermally induced blood-brain barrier disruption and brain tissue damage in the rabbit brain. Magn Reson Med. 2004;51(5):913-923. 25. Haveman J, Sminia P, Wondergem J, van der Zee J, Hulshof MC. Effects of hyperthermia on the central nervous system: what was learnt from animal studies? Int J Hyperthermia. 2005;21(5):473-487. 26. Hildebrandt B, Hegewisch-Becker S, Kerner T, et al; German Interdisciplinary Working Group on Hyperthermia. Current status of radiant whole-body hyperthermia at temperatures .41.5C and practical guidelines for the treatment of adults. Int J Hyperthermia. 2005;21(2):169-183. 27. Pearce JA. Relationship between Arrhenius models of thermal damage and the CEM43 thermal dose. In: Ryan TP, ed. Energy-Based Treatment of Tissue and Assessment V. Bellingham, WA: Proc SPIE; 2009.



COFFEY ET AL

28. Brodkey JS, Miyazaki Y, Ervin FR, Mark VH. Reversible heat lesions, a method of stereotactic localization. J Neurosurg. 1964;21:49-53. 29. Favre J, Taha JM, Nguyen TT, Gildenberg PL, Burchiel KJ. Pallidotomy: a survey of current practice in North America. Neurosurgery. 1996;39(4):883-890. 30. Cosman ER, Cosman ER Jr. Radiofrequency Lesions. In: Lozano A, Gildenberg P, Tasker R, eds. Textbook of Stereotactic and Functional Neurosurgery, 2nd ed. Berlin, Heidelberg: Springer; 2009:1360-1382. 31. Institute of Laboratory Animal Resources, Commission on Life Sciences, National research Council. The Guide for Care and Use of Laboratory Animals. Washington, DC: National Academy Press; 1996. 32. Raoul S, Leduc D, Vegas T, et al. Deep brain stimulation electrodes used for staged lesion within the basal ganglia: experimental studies for parameter validation. J Neurosurg. 2007;107(5):1027-1035. 33. Raoul S, Leduc D, Deligny C, Lajat Y. Therapeutic lesions through chronically implanted deep stimulation electrodes. In: Lozano A, Gildenberg P, Tasker R, eds. Textbook of Stereotactic and Functional Neurosurgery, 2nd ed. Berlin, Heidelberg: Springer; 2009; 1427-1442. 34. Oh M, Hodaie M, Kim SH, Alkhani A, Lang AE, Lozano AM. Deep brain stimulator electrodes used for lesioning: proof of principle. Neurosurgery. 2001;49(2):363-369. 35. Pearce JA. Comparative analysis of mathematical models of cell death and thermal damage processes. Int J Hyperthermia. 2013;29(4):262-280. 36. Larson PS, Richardson RM, Starr PA, Martin AJ. Magnetic resonance imaging of implanted deep brain stimulators: experience in a large series. Stereotactic Funct Neurosurg. 2008;86(2):92-100. 37. Tagliati M. Safety of MRI in patients with implanted deep brain stimulation devices. Neuroimage. 2009;47(suppl 2):T53-T57.

Acknowledgments

neurosurgery, ie, test lesion until 43C and definite lesion above 65C, which is more applicable to the case of possible heating during MRI scans. The authors conclude that temperatures of 43C or lower, approximate 6C rise above the normal body temperature, appear to be a reasonable conservative upper temperature limit to establish a criterion for zero risk of clinically observable neurological signs and symptoms. The assumption that pulsed RF magnetic fields induce electric fields in body tissues and implanted metallic structures in these sites can generate currents that dissipate at the electrode-tissue interface causing heating of the tissue surrounding the electrodes needs further studies. Poor practice of MRI acquisition with patients with implantable devices led to accidents that have delayed the progress of functional neurosurgery performed inside of the MRI operating room for years.3,4 It is unfortunate that the authors did not quantify increases in temperature caused by MRI induction, information that is lacking in clinical practice. Does MRI increase in tissue temperature above the 6C limit the authors suggest? Which are the MRI sequences that do and do not do so? Which are the guidelines to perform safe MRI in patients with a neuromodulation probe? Although well performed, this study does not bring knowledge regarding MRI interaction with probes in the brain, it does confirm with a good methodology what was already known since the RF lesions establishment by functional neurosurgeons. The study is clinically useful, however, giving the information that 6C increases during up to 30 minutes are safe, because MRIs can be performed during this time frame.

John Pearce, PhD, of the Department of Electrical and Computer Engineering, University of Texas, Austin, TX, United States, provided valuable assistance and guidance during all phases of this project and graciously provided a prepublication version of his manuscript (Reference 35).

Antonio De Salles Los Angeles, California Alessandra Gorgulho Sao Paulo, Brazil

COMMENTS

T

he authors performed experiments in the sheep’s brain and spine to gain insights on the tolerance of the neural tissue to radiofrequency (RF, 64 MHz) administered by heating probes. They tried to simulate neurostimulation leads while in the magnetic resonance imaging (MRI) environment. They started from the assumption that, when metallic wires are used to conduct electricity between a neurostimulator and electrode leads, even with MRI-compatible devices, temperature rise occurs surrounding the leads. They studied a range of temperatures that theoretically could be induced by the interaction of the MRI electrical fields with the intracranial probes. Temperatures from 37C to 49C for 30-minute exposures were studied. The authors are to be commended for their initiative to try to help patients needing to have quality MRI and simply not having access to it because they harbor a neurostimulator. Proper guidelines are needed for radiologists and technicians performing MRI studies, so that patients are not simply denied to have the examination, just because it is safe to say no, depriving the patient of an important examination. The authors were able to set safe time and temperature limits with their study. Deep brain and spinal RF heating up to 43C for 30 minutes produced indistinguishable effects compared with 37C in their controls. RF heating greater than 43C for 30 minutes produced temperature-dependent, localized thermal damage. This study confirms what stereotactic surgeons performing RF test functional lesions before a definite therapeutic lesion have known in clinical practice for years.1,2 They tried to find other examples of hyperthermia in medicine, including whole-body hyperthermia of 41.8 C for 60 minutes in oncology patients undergoing radiotherapy, although this example is definitely below temperatures used in functional


1. Anzai Y, Lufkin R, De Salles AAF, et al. Radiofrequency ablation of brain tumours using MR guidance. Min Invas Ther Allied Technol. 1996;5:232-242. 2. De Salles AAF, Brekhus SD, De Souza EC, et al. Early postoperative appearance of radiofrequency lesions on magnetic resonance imaging. Neurosurgery. 1995;36(5): 932-936. 3. De Salles AA, Frighetto L, Behnke E, et al. Functional neurosurgery in the MRI environment. Minim Invasive Neurosurg. 2004;47(5):284-289. 4. Lee MW, De Salles AA, Frighetto L, Torres R, Behnke E, Bronstein J. Imaging techniques and electrode fixation methods for deep brain stimulation—Intraoperative MRI 0.2T, 1.5 T and fluoroscopy. Minimally Invas Neurosurg. 2005;48(1):1-6.

T

he authors of this article aimed to determine safe limits for temperature and duration of heating of neural tissue—as applicable to the heating observed during magnetic resonance studies of patients with implanted neurostimulation devices. In essence, they are trying to have a formal justification for the safe limits of MRI exposure in recently approved “MRI conditionally safe” neurostimulators that are used in our daily clinical practice. Their findings confirm the previously stated postulate of “,43ºC for 30 minutes” being safe, at least on a histological level. Animal data may be sufficient for drawing this conclusion—but the human data will be needed to define other, not associated with histologically proven tissue damage, effects of prolonged tissue heating. These may include pain, sensory, and motor phenomena, psychological and behavioral alterations, both at the time of thermal exposure and in subsequent minutes/hours/days. The hope, obviously, is that no such effect would occur, but without clinical evaluation, no assumption of complete safety may be made. In particular, one will have to consider that patients who have implanted devices and require MRI studies are different from healthy controls—they already have physiological




Konstantin V. Slavin Chicago, Illinois

diology: Can we safely perform MRI—scans in patients with implanted DBS or SCS systems? This question is important because we now implant DBS devices in a patient population that gradually becomes younger and will have to be exposed to MR scans for other diseases. Obviously, there are technical differences between the diverse DBS systems on the market. To the reviewer’s knowledge to date, the only commercial system with a CE mark for head MRI is the ACTIVA—series (Medtronic, Minneapolis, Minnesota). All other systems are scanned with the same clinical protocol but in neglect of the distinct technical specifications. This is study clarifies in an ovine model certain numbers that need to be known to the manufacturers of implantable DBS systems to safeguard MR scanning in patients with implanted DBS systems.

his is an experimental report that addresses a very important question in the field of stereotactic and functional neurosurgery and neurora-

Volker A. Coenen Freiburg, Germany

impairments that cause chronic pain, movement disorders, epilepsy, psychiatric disorders, etc, and may have different thresholds for temperature-related experiences. Even taking into account individual variability in sensitivity for thermal effects, there will be a common denominator for safety in the general population and I agree with the authors that their establishment of thermal dose limit is a first step in defining safety of MRI and other procedures in patients with implanted neurostimulation devices.

T

NEUROSURGERY



Safe leads and lead changes in competitive team sports.

The Kora Pacemaker is Safe and Effective for Magnetic Resonance Imaging.

Laser patterning of platinum electrodes for safe neurostimulation.

Maximum safe resection of glioblastoma multiforme.

Safe introducer technique for pacemaker lead implantation.

Investigation of Parallel Radiofrequency Transmission for the Reduction of Heating in Long Conductive Leads in 3 Tesla Magnetic Resonance Imaging.

The burden of lead: how much is safe?

How safe is repetitive Transcranial Magnetic Stimulation?

The Human Body's Own Language to be Considered for Safe and Effective Neurostimulation.

Ferromagnetic particles as magnetic resonance imaging temperature sensors.

Maximum Likelihood Reconstruction for Magnetic Resonance Fingerprinting.

Safe performance of magnetic resonance of the heart in patients with magnetic resonance conditional pacemaker systems: the safety issue of the ESTIMATE study.

Safe Handover : Safe Patients - The Electronic Handover System.

[Magnetic resonance imaging].

Pelvic magnetic resonance imaging.

Eye-safe compact Raman light detection and ranging temperature profiler.

Gynecologic magnetic resonance imaging.

Intracranial magnetic resonance imaging.

Shoulder magnetic resonance imaging.

Magnetic resonance imaging.

Manganese-enhanced magnetic resonance imaging investigation of the interferon-α model of depression in rats.

Altered executive function in the lead-exposed brain: A functional magnetic resonance imaging study.

What is the maximum safe tourniquet time in elective upper limb surgery?

Safety and outcomes of magnetic resonance imaging in patients with abandoned pacemaker and defibrillator leads.