You are on page 1of 6

JOURNAL OF MAGNETIC RESONANCE IMAGING 20:315–320 (2004)

Original Research

Evaluation of Specific Absorption Rate as a


Dosimeter of MRI-Related Implant Heating
Kenneth B. Baker, PhD,1* Jean A. Tkach, PhD,2 John A. Nyenhuis, PhD,3
Michael Phillips, MD,2 Frank G. Shellock, PhD,4 Jorge Gonzalez-Martinez, MD, PhD,5
and Ali R. Rezai, MD5

Purpose: To compare the magnetic resonance imaging Key Words: magnetic resonance imaging, bioeffects; mag-
(MRI)-related heating per unit of whole body averaged spe- netic resonance imaging, safety; neurostimulation sys-
cific absorption rate (SAR) of a conductive implant exposed tems; implants; deep brain stimulation; specific absorption
to two different 1.5-Tesla/64 MHz MR systems. rate
J. Magn. Reson. Imaging 2004;20:315–320.
Materials and Methods: Temperature changes at the elec-
© 2004 Wiley-Liss, Inc.
trode contacts of a deep brain stimulation lead were mea-
sured using fluoroptic thermometry. The leads were placed
in a typical surgical implant configuration within a gel-filled
phantom of the human head and torso. MRI was performed ONE OF THE PRIMARY safety concerns related to per-
using two different transmit/receive body coils on two dif- forming magnetic resonance imaging (MRI) on biologic
ferent generation 1.5-Tesla MR systems from the same tissue is heating that can occur as tissue absorbs ra-
manufacturer. Temperature changes were normalized to diofrequency (RF) energy (1,2). Specific absorption rate
whole body averaged SAR values and compared between (SAR), defined in the International Electrotechnical
the two scanners. Commission (IEC) standard (3) as the amount of RF
Results: Depending on the landmark location, the normal- power absorbed per unit of mass of an object indicated
ized temperature change for the implant was significantly in W/kg, represents the current national and interna-
higher on one MR system compared to the other (P ⬍ 0.001). tional dosimetric term used to characterize the thermo-
Conclusion: The findings revealed marked differences genic aspects of this electromagnetic field (3–5). Con-
across two MR systems in the level of radiofrequency (RF)- cerns over RF-induced heating are compounded by the
induced temperature changes per unit of whole body SAR presence of conductive implants in the body, particu-
for a conductive implant. Thus, these data suggest that larly those that are electronically activated or that have
using SAR to guide MR safety recommendations for neuro- an elongated configuration, as they can increase the
stimulation systems or other similar implants across dif- potential for RF heating in the tissue immediately sur-
ferent MR systems is unreliable and, therefore, potentially rounding the implanted device (3). For example, Achen-
dangerous. Better, more universal, measures are required
bach et al (6) reported a temperature elevation at the tip
in order to ensure patient safety.
of a pacing electrode, unattached to a cardiac pace-
maker, of up to 63.1°C during 90 seconds of MRI. More
recently, Rezai et al (7) reported temperature elevations
at the distal end of a deep brain stimulation (DBS)
electrode of 25.3°C after 15 minutes of MRI.
1
Department of Neurology, The Cleveland Clinic Foundation, Cleve- Other than encouraging the need for diligence when
land, Ohio. implants are present, neither the IEC (3,5) nor the Food
2
Department of Radiology, The Cleveland Clinic Foundation, Cleveland, and Drug Administration (FDA) document offers partic-
Ohio.
3
ular guidance with respect to performing MR proce-
Department of Computer and Electrical Engineering, Purdue Univer-
sity, West Lafayette, Indiana. dures on patients with conductive implants. As such, it
4
Keck School of Medicine, University of Southern California, Los Ange- typically falls upon the manufacturers of these devices
les, California. to demonstrate MR safety data for their devices. Al-
5
Section of Stereotactic and Functional Neurosurgery, The Department though it is possible simply to exclude patients with
of Neurosurgery, The Cleveland Clinic Foundation, Cleveland, Ohio.
Contract grant sponsor: NIH; Contract grant number: NS44575-01;
implantable devices from MRI, this approach denies
Contract grant sponsor: Medtronic, Inc., Minneapolis, MN. access to an important diagnostic imaging modality.
*Address reprint requests to: K.B.B, The Cleveland Clinic Foundation, The need to provide reliable, safe MRI in patients with
9500 Euclid Avenue, Desk S-90, Cleveland, OH 44195. implantable devices is evident.
E-mail: bakerk@ccf.org
Received November 18, 2003; Accepted April 19, 2004.
SAR has been routinely used as an indirect quantita-
DOI 10.1002/jmri.20103 tive measure of RF energy in the reporting of safety
Published online in Wiley InterScience (www.interscience.wiley.com). recommendations for clinical MRI procedures when
© 2004 Wiley-Liss, Inc. 315
316 Baker et al.

conductive implants are present (7–9). However, this passing into the skull. The excess lead that exists be-
may not be appropriate because SAR has been applied tween the burrhole and the extension wire was subse-
primarily as an index of heating in biologic tissue de- quently allowed to arrange itself in a random fashion
void of metallic or conductive implants (10). A single along the surface of the phantom skull. It was held
SAR parameter may not be a sufficiently specific pre- firmly against the surface of the phantom skull by a
dictor of device heating given the four different SAR thin piece of silicone rubber. Finally, within the phan-
measures defined in the recent IEC standard (3). Fur- tom skull the distal ends of the two DBS leads were
thermore, it appears that MR system manufacturers positioned parallel and coronally coplanar, separated
are using proprietary and evolving models of the human by a distance of 3 cm at the distal tips (contact 0). The
body upon which to base their SAR calculations. There- position and routing of all hardware components was
fore, the purpose of this study was to explore the valid- held constant throughout the experiments in order to
ity of using SAR in making safety recommendations for avoid any potential geometry-based changes in the in-
patients with implanted devices by comparing the ex teraction between the DBS hardware and RF field.
vivo MRI-related heating of an electronically-activated
implant between two different generation commercial
Thermometry System and Placement of
1.5-Tesla MR systems from the same manufacturer.
Temperature Probes
Temperatures were recorded at multiple locations us-
MATERIALS AND METHODS
ing an MR-compatible fluoroptic thermometry system
Phantom (Model 3100, Luxtron, Santa Clara, CA). Samples were
Experiments were performed using a previously de- digitized at a rate of five samples per second and trans-
scribed phantom model (7). The phantom, designed to ferred to a PC for off-line analysis. For the purpose of
approximate the size and shape of the human head and the experiments described, fluoroptic thermometry
torso, was filled with a semi-solid aqueous gel prepared probes were placed 1) 0.1 mm above contact plate zero
to simulate the electrical conductivity and thermal con- of the distal right DBS electrode (R_E0), 2) 0.1 mm
vection properties of human tissue (11,12). A plastic above the contact plates of the distal left DBS electrode
grid frame with adjustable posts was placed at the bot- (L_E0), 3) at a reference position midway between con-
tom of the phantom to ensure stable positioning and tact zero of the right and left electrode tips (1.5 cm
support of the neurostimulators, extensions, and leads, equidistant from each), and 4) at an additional refer-
as previously described (7). The phantom was placed on ence position within 1 cm of the edge of the phantom
the MR table such that its position was consistent with positioned in the region of the neck of the phantom.
that of a patient in a typical head-first supine position. Additional probes were present at other electrode con-
The mass of the filled phantom, 30 kg, was the weight tacts; however, temperature measurements from those
entered at registration to enable the MR systems to probes were redundant to the data presented here and
calculate the whole body averaged SARs. are not included.

Neurostimulation System MR Systems and Imaging Sequence


The Activa Tremor Control System (Medtronic, Inc., Ex vivo testing was performed on two different genera-
Minneapolis, MN), used for DBS, was used in the ex- tion 1.5-Tesla/64 MHz MR systems from the same
periments. For a typical bilateral procedure, this sys- manufacturer (Siemens Medical Solutions USA, Inc.,
tem includes two implantable pulse generators (IPG), Malvern, PA), as follows: MR system 1 (Symphony; soft-
two quadripolar extensions, and a pair of DBS leads. ware version, Numaris 4 VA21B) and MR system 2
The 40-cm-long DBS lead used in the present study is (Vision; software version, Numaris 3.0 VB33G). It
insulated along its entire length, except for four contact should be noted that while both are considered whole
plates at the connector end as well as four contact body MR systems, the MR system 1 is a short-bore
plates in the distal 10.5 mm of the lead. The distal four system and the MR system 2 is a long-bore system.
contacts, each 1.5 mm in length and 1.27 mm in diam- Experiments were performed using the circularly polar-
eter, are separated from one another by 1.5 mm of ized (CP) transmit/receive RF body coil.
polyurethane material. All hardware used was repre- Temperature recordings were obtained during multi-
sentative of the manufactured “finished” version of the slice fast spin echo (FSE), T1-weighted spin-echo (SE),
implant and was not altered in any manner before test- and fast spin-echo inversion recovery pulse sequences
ing. (FSE-IR). The MR landmark for the scans was placed at
The bilateral DBS leads (Model 3387-40), extensions the most distal contact (contact 0) of the DBS leads or at
(Model 7495-51), and IPGs (Model 7426) were posi- the pectorally-placed IPG. Specific acquisition parame-
tioned in a manner to emulate the clinical method ters (i.e., TR, number of slices, echo train length [ETL],
(7,13). The IPGs were positioned to represent subcuta- etc.), and/or the amplitudes of the RF pulses (i.e., ef-
neous, subclavicular placement with the attached fective flip angles) were modified to produce a range of
51-cm extension routed ipsilaterally along the neck. whole body averaged SAR values. The software reported
The connector that joins the extension to the DBS lead values for the whole body averaged SAR based upon the
was positioned to emulate the more common post-au- MR sequence parameters, the transmitter reference ad-
ricular placement. The lead was routed along a phan- justment and applied voltages, and the patient weight
tom skull and circled the burrhole cap one time before entered during registration.
SAR and MRI-Related Implant Heating 317

Data Analysis Table 1


Summary of Raw Temperature Data
Baseline temperature recording for each experimental
Whole ⌬T ⌬T
condition was initiated at least 20 seconds before the MR Run
Sequence body SAR (° C) (° C)
start of MRI. In order to control for variations in the system no.
(W/kg) R E0 L E0
duration of the individual MRI procedures, statistical
comparisons were based on the temperature increase Landmarked at contact 0
as recorded 60 seconds after the initiation of scanning. FSE-T2 1 S02 0.07 4.90 6.32
FSE-T2 1 S03 0.06 4.00 5.13
The temperature sample at that point was eight-bin and
FSE-T2 1 S04 0.04 2.88 3.89
center averaged, and thus represented an average of
FSE-T2 1 S16 0.06 4.00 5.59
the eight samples recorded between 51.2 and 61.8 sec- Fast-IR 1 S06 0.02 1.51 2.02
onds after scan initiation. This was done in order to T1-SE 1 S07 0.03 2.27 2.88
minimize the influence of any potential signal noise in T1-SE 1 S08 0.08 5.96 7.79
the temperature recordings. The averaged value then FSE-T2 2 V02 1.78 3.55 1.45
was converted to an absolute temperature rise by sub- FSE-T2 2 V03 0.88 1.69 0.74
tracting the pre-sequence baseline average. FSE-T2 2 V20 1.83 2.97 1.66
Unfortunately, direct comparisons based on the exact FSE-IR 2 V05 0.86 1.66 0.75
whole body averaged SAR value were not possible in the IR 2 V06 7.85 12.40 6.39
Landmarked at the IPG
current study as the two different MR systems could not
FSE-T2 1 S09 0.58 2.05 0.44
be made to deliver the exact same SAR for a given
FSE-T2 1 S10 0.33 1.46 0.33
imaging sequence. This inability was related to the val- T1-SE 1 S11 0.38 1.53 0.39
ues for exposed or head averaged SAR reaching the T1-SE 1 S12 0.72 3.11 0.79
maximal allowable IEC limit on the Symphony system. FSE-T2 2 V07 2.34 1.77 1.00
For this reason, results were normalized and expressed FSE-T2 2 V08 1.20 0.68 0.44
in terms of temperature rise per unit of SAR. This ap- FSE-T2 2 V11 22.35 15.40 8.96
proach is justified by the observed linear relationship T1-SE 2 V09 2.12 1.92 1.09
between implant heating and whole body averaged SAR IPG ⫽ implantable pulse generator, FSE ⫽ fast spin echo, IR ⫽
seen for each generation of MR system. Thus, for com- inversion recovery, SE ⫽ spin echo, Sag ⫽ sagittal, MR system #1
parison between the two MR systems, the absolute ⫽ symphony, MR system #2 ⫽ vision, R_E0 ⫽ right lead, contact 0,
value then was normalized to the console-reported in- L_E0 ⫽ left lead, contact 0.
dices of whole body averaged SAR. Comparisons be-
tween the two MR systems were made using Student’s
Figure 3 summarizes the normalized heating data for
two-tailed t-test or, where appropriate, the Mann-Whit-
the right and left DBS lead between the two MR systems
ney Rank Sum test.
when the landmark was placed over contact 0. When
normalized to the whole body averaged SAR, the me-
RESULTS dian values for MR system 1 and MR system 2 were 68.5
(range: 62.6 –75.6) and 1.9 (range: 1.6 –2.0) for the right
Twelve different pulse sequences were evaluated with DBS lead (P ⬍ 0.01, T ⫽ 15) and 90.4 (range: 85.6 –97.4)
the scan landmark at contact 0 of the DBS lead and and 0.8 (range: 0.8 – 0.9) for the left DBS lead (P ⬍ 0.01,
eight with the landmark at the IPG. The results of the T ⫽ 15), respectively. Similarly, Fig. 4 summarizes the
experiments are summarized in Table 1. Figure 1 pro- normalized temperature data for sequences performed
vides a sample of the raw temperature data recorded with the scanner landmark at the IPGs. The normalized
from MR system 1 (right) and MR system 2 (left), as mean values (⫾ SD) for MR system 1 and 2 were 4.08 (⫾
measured at the thermometry probe positioned over 0.39) and 0.73 (⫾ 0.14) for the right DBS lead (P ⬍ 0.01,
contact 0 of the left DBS lead. The data shown were T ⫽ 16.1, degrees of freedom [df] ⫽6) and 0.97 (⫾ 0.14)
taken with the landmark positioned at the electrode and 0.43 (⫾ 0.06) for the left DBS lead (P ⬍ 0.01, T ⫽
contacts and correspond to runs S02 and V03 listed in 6.9, df ⫽ 6), respectively.
Table 1. Temperature changes generally were extremely Finally, an additional 15 pulse sequences were eval-
rapid, with the temperature recorded at the 60-second uated with the scan landmark placed at locations other
time point typically representing 60% or more of the than the IPG and contact 0 of the DBS lead. This in-
final or maximum temperature recorded throughout cluded 1) 15 cm rostral to the IPGs (11.5 mm caudal to
the sequence. contact 0 of the DBS lead) in approximately the region
Figure 2 illustrates the relationship between RF in- of the neck (X ⫽ 65; N ⫽ 4); 2) 15 cm below the IPGs in
tensity, as indexed by whole body averaged SAR, and the region of the xyphoid process (X ⫽ 35; N ⫽ 5); and 3)
temperature change at contact 0 of the right and left 30 cm below the IPGs in the region of the umbilicus (X ⫽
DBS lead for both MR systems. The temperature 20; N ⫽ 6). Figure 5 illustrates how the magnitude of the
change recorded was highly correlated (r ⫽ 0.99, P ⬍ temperature difference normalized to whole body aver-
0.01) with the amount of RF power for both MR system aged SAR between the two MR systems varied as a
1 (N ⫽ 7) and MR system 2 (N ⫽ 5). This suggests that, function of landmark position. The difference between
within a given MR system, the primary determinant of the two scanners was maximal with the landmark
the magnitude of the temperature change for a given placed at the electrodes (X ⫽ 76.5 cm), but rapidly
lead system and configuration is the overall level of decreased as the landmark was moved caudally (lower
estimated RF power. values of X).
318 Baker et al.

Figure 1. Example of the raw temperature data showing the change in temperature as function of time, where time zero
represents the onset of MRI (i.e., start of pulsed RF). Data are shown from MR system 1 (left: Run #S02) and MR system 2 (right:
Run #V03) and represent temperature changes recorded at contact 0 of the left DBS lead. Note that the temperature increase is
initially quite rapid. Also, the temperature increase as recorded 60 seconds after the start of MRI is approximately nine times
higher on MR system 1, despite a reported whole body averaged SAR value that was less than 1/12th of what was delivered on
MR system 2 (0.07 W/kg vs. 0.88 W/kg, respectively). Both measurements were made with the landmark position at contact 0
on the DBS electrode.

DISCUSSION (i.e., a specific scanner hardware running a specific


The most significant finding of this study was that the software version) from the same manufacturer. Specif-
amount of heating per unit change in the calculated ically, the temperature change per unit whole body av-
whole body averaged SAR, as recorded at an implant- eraged SAR was as much as 90 times higher when
able conductive device, was profoundly different be- using the transmit/receive RF body coil on MR system
tween two different generation 1.5-Tesla MR systems 1 as opposed to that observed on MR system 2. These

Figure 2. Temperature change as recorded at contact 0 of the right and left DBS lead as a function of whole body averaged SAR.
The landmark was at contact 0 on the DBS electrode. Each plot shows the temperature change as recorded at contact 0 of the
left (empty circles) and right (filled circles) DBS lead. Note that, within a given MR system, the recorded temperature change is
well correlated with the amount of RF power, despite the use of different sequence types. *Note that for MR system 2, two of the
five data points for each lead partially overlap.
SAR and MRI-Related Implant Heating 319

Figure 3. Comparison of the change in temperature (T) per


unit of whole body averaged SAR (W/kg) between the two MR
systems for both the right and left DBS lead with the landmark Figure 5. The effect of the landmark position on the magni-
set at contact 0 of the DBS electrodes. Median values are tude of the difference in normalized temperature change
shown for each MR system, with the error bars representing (change in temperature [T]/whole body averaged SAR) between
the first and third quartile values. *P ⬍ 0.01. the two MR systems. The data are plotted as the log of change
in temperature (T) per unit of whole body averaged SAR. The
maximum difference is observed when the landmark position
findings strongly suggest that calculated whole body is placed at contact 0 of the DBS electrode and diminishes as
averaged SAR is not a reliable metric for RF power the landmark positioned is placed more caudally.
irradiation across different MR systems in the presence
of an electronically-activated conductive implant. Thus,
MR systems resulted from a change in the physical
safety results with regard to heating for implanted de-
relationship between the fluoroptic probes and the
vices that are determined on specific types of MR sys-
metal implant as the phantom was moved from one MR
tems relying on whole body averaged SAR as the dosi-
system to the other. Finally, it is worth noting the high
metric means of characterizing RF power deposition
degree of linearity within each system that existed de-
cannot be reliably generalized to other imaging systems
spite variations in the type of sequence used (i.e., spin
and software versions.
echo vs. fast spin echo vs. inversion recovery fast spin
As expected, the findings support a linear relation-
echo). Overall, more than 95% of the variance in tem-
ship between SAR and heating within a given type of MR
perature on a given type of MR system was accounted
system, where the observed temperature change was
for by the amount of RF energy delivered.
highly correlated with whole body averaged SAR. The
Although the source of the discrepancy between the
reliability of the data presented is supported further by
two MR systems is not clear, it is likely a combination of
the consistency of the findings across all of the elec-
hardware- and software-related factors. With respect to
trodes monitored, as this would tend to exclude the
hardware issues, the largest contribution probably
possibility that the observed difference between the two
would derive from differences in the design of the trans-
mit RF body coils. From the software perspective, any
difference in the algorithm (e.g., body modeling) used to
estimate whole body averaged SAR would play a large
role. The results presented in this manuscript most
likely reflect the combined effects of these factors,
though the relative contribution of each factor to the
observed difference reported herein cannot be defini-
tively determined. While it is well known that the elec-
tric field distribution associated with an RF resonator
will vary with coil design, this information is not avail-
able to the standard MR user, and hence, the impact
cannot be assessed. Similarly, the methodologies used
to calculate SAR are not accessible to the user. That
such differences do indeed exist between the method-
ologies employed by each of two scanners used in this
study is evidenced by the terminology used to charac-
terize the reported SAR values reported, with MR sys-
Figure 4. Comparison of the change in temperature (T) per
unit of whole body averaged SAR (W/kg) between the two MR
tem 1 reporting “Head,” “Exposed Body,” and “Whole
systems for both the right and left DBS leads with the land- Body” SAR values and system 2 reporting “Partial
mark position at the level of the IPGs. Mean values are shown Body,” “Local Body,” and “Whole Body”. This difference
for each MR system, with the error bars showing the standard in terminology also explains why no attempt was made
deviation. *P ⬍ 0.01. to normalize the temperature data to any value other
320 Baker et al.

than the reported whole body averaged SAR. Further- tem types. It is likely that some of the observed differ-
more, preliminary work in the laboratory using an ence arises from the different methods of SAR
adult, human subject has shown that while the whole calculation used across the two generations of MR sys-
body averaged SAR values reported by MR system 1 tems. A more direct measure of the RF magnetic field,
change as a function of the landmark position, those for as for example its RMS value (B1,RMS) may be a more
MR system 2 remained more or less constant. suitable measure than SAR for predicting the potential
Differences across MR systems in the calculation of for RF-induced heating near an implant. As well, im-
the whole body averaged SAR could have potentially plant manufacturers should be encouraged to continue
disastrous clinical implications if that value is used as to pursue development of MRI compatible devices, as it
the metric for defining universally safe operating con- clearly should be possible to develop lead systems that
ditions for patients with conductive implants. For ex- do not conduct current at the RF used in MRI and yet
ample, the RF body coil recommendations made by allow for the conduction of appropriate stimulation
Rezai et al (7) were based on the same neurostimulation (14). Until the manufacturers of these systems and de-
system evaluated on the same MR system (MR system vices solve these potential problems, however, the onus
2) used in the current study. In that study, the inves- for safety lies in the hands of radiologists at individual
tigators concluded that, when using the transmit/re- institutions.
ceive RF body coil, whole body SAR exposures of up to
0.9 W/kg could be performed without exceeding 2°C at
the tip of the electrodes. Although the results of the REFERENCES
present study are consistent with those recommenda-
1. Schaefer DJ. Safety aspects of radio-frequency power deposition in
tions for MR system 2 (maximum ⌬T/whole body aver- magnetic resonance. Magn Reson Imaging Clin N Am 1998;6:775–
aged SAR on MR system 2 was 2.0), generalization of 789.
those results to MR system 1 could be catastrophic. 2. Shellock FG. Radiofrequency-induced heating during MR proce-
Experiments on MR system 1 showed that 6°C of heat- dures: a review. J Magn Reson Imaging 2000;12:30 –36.
3. International Electrotechnical Commission (IEC). Medical Electri-
ing could be achieved with a whole body averaged SAR
cal Equipment. Particular requirements for the safety of magnetic
of only 0.08 W/kg, a value that is less than 1/10th of resonance equipment for medical diagnosis. International Stan-
the safety limit proposed for data collected from MR dard IEC 60601-2-33; 2002.
system 2 (7). As the maximum ⌬T/whole body averaged 4. Zaremba L. FDA guidance for MR system safety and patient expo-
SAR on MR system 1 was approximately 97, exposure sures: current status and future considerations. In: Magnetic res-
onance procedures: health effects and safety. Boca Ration, FL: CRC
to 0.9 W/kg whole body averaged SAR on that system Press; 2001. p 183–196.
would result in a theoretical temperature rise on the 5. U.S. Department of Health and Human Services, Food and Drug
order of 90°C. Certainly, the “static” conditions within Administration, Center for Devices and Radiological Health. Guid-
the phantom used in the present study result in more or ance for industry and FDA staff. Criteria for significant risk inves-
tigations of magnetic resonance diagnostic devices. July 14, 2003.
less a worst-case scenario, but it is unlikely that cere-
6. Achenbach S, Moshage W, Diem B, Bieberle T, Schibgilla V, Bach-
bral blood flow or other physiologic mechanisms could mann K. Effects of magnetic resonance imaging on cardiac pace-
effectively dissipate any heating approaching that mag- makers and electrodes. Am Heart J 1997;134:467– 473.
nitude. This is particularly true considering the rate at 7. Rezai AR, Finelli D, Nyenhuis JA, et al. Neurostimulator for deep
which the temperature increase occurs. brain stimulation: ex vivo evaluation of MRI-related heating at 1.5-
Tesla. J Magn Reson Imaging 2002;15:241–250.
It is important to note, however, that the recommen- 8. Shellock FG, Hatfield M, Simon BJ, et al. Implantable spinal fusion
dations of the neurostimulation system manufacturer stimulator: assessment of MR safety and artifacts. J Magn Reson
do not allow the use of a transmit/receive RF body coil, Imaging 2000;12:214 –223.
instead being specific to the performance of MRI using 9. Soletra™, Neurostimulator for Deep Brain Stimulation, Model
7246. Physician and hospital staff mannual. Minneapolis, MN:
a transmit/receive RF head coil (9). Therefore, in con-
Medtronic; 2001.
sideration of the present findings, until the compari- 10. Shellock FG, Schaefer DJ. Radiofrequency energy-induced heating
sons have been made between various types of MR during magnetic resonance procedures: laboratory and clinical ex-
systems with transmit/receive RF head coils, it would periences. In: Shellock FG, editor. Magnetic resonance procedures:
be reasonable to carefully consider applying those health effects and safety. Boca Raton, FL: CRC Press; 2001. p
75–95.
guidelines. Indeed, the conservative approach would be 11. American Society for Testing and Materials. ASTM International.
such that MR examinations not be performed on pa- Standard test method for measurement of radio frequency induced
tients with electronically-activated implants in the ab- heating near passive implants during magnetic resonance imaging.
sence of ex vivo experimental data specific to the type of West Conshohocken, PA: ASTM International; 2002.
12. Park SM, Nyenuis JA, Smith CD, et al. Gelled vs. non-gelled phan-
MR scanner, or at the very least the same device and
tom material for measurement of MRI-induced temperature in-
software version, on which the imaging is to be per- creases with bioimplants. IEEE Trans Magnetics 2003;39:3367–
formed. 3371.
In the meantime, manufacturers of MR systems 13. Finelli DA, Rezai AR, Ruggieri PM, et al. MR imaging related heating
should be encouraged to adopt a universally accept- of deep brain stimulation electrodes: In vitro study. AJNR Am J
Neuroradiol 2002;23:1795–1802.
able, standardized measure of RF power deposition, as 14. Greatbatch W, Miller V, Shellock FG. Magnetic resonance safety of
it is clear that SAR estimates are not reliable as a meth- a newly-developed fiber-optic cardiac pacing lead. J Magn Reson
odology of determining power delivery across MR sys- Imaging 2002;16:97–103.

You might also like