I Review
Accessory Equipment Considerations
with Respect to MRI Compatibility
Elaine K. Keeler, PhD Francis X. Casey Hans Engels Elizabeth Lauder Catherine Anne Pirto
Theodore Reisker James Rogers Daniel J. Schaefer, PhD Thomas Tynes
The MR Section of The National Electrical ManufacturersAs-
sociation (NEMA). in response to a request from the Food &
Drug Administration (FDA),recently issued a position paper
to address generic issues related to the compatibility of ac-
cessory equipment produced by third party equipment man-
ufacturers or MR equipment users and intended to be used
in conjunction with MR equipment or within the MR scan-
ning room. The recommendations concern scanning acces-
sories, such as radiofrequency (RF) coils, patient monitoring
equipment and injectors, as well as patient comfort acces-
sories and positioning devices. The following issues related
to safety performance are discussed (a)the interaction of
the equipment with the MR scanner, (b) interactions of the
MR scanner with the equipment, and (c)potential safety haz-
ards for patients and staff that can be posed by accessory
equipment in the MR scan environment. The recommenda-
tions are based on combined input from NEMA member com-
panies who manufacture MR systems and MR accessories
and are presented for consideration in the design of MR ac-
cessory products and incorporation of these concepts into
testing plans to ensure MR compatibility of third party de-
vices.
Index terms: MR compatibility - M K safety. MK equipment
JMRJ 1998: &:12-18
Abbreviations: CHT = cathode ray tube. DC = direct current. EMC = electromag-
netic compatil,ility. FDA = Fond & Drug Administration. LCD :. liquid crystal display.
LED = light r'mitting diode. NEMA = National Electncal Manufiicturers Association.
OE:M = original equipment m;iriulacturer. RF = radiolrequency. SAR = sperifir ab-
soiption rat<' T/R = transmitJrrceive.
From I he Maqnetic Resonance Section, Diagnostic Imaging and ThPrapy Systems Di-
vision. Na1ioii.il Electrical Manufacturers Association. 1300 Norlh 17th Street. Suite
1847. K o s s l ~ nVA 22209. E-mail ekkQmr.pirker.com. Received September 19. 1996:
revision rrqiirsted May 20. 1997: accepted May 28. Presented a t the Workshop on
Advances in MR Safety and Compatibility. Macinan. VA. June $1-1 1. 1996. Address
reprint requests t o E K.K. at Picker International. 595 Miner Road. Cleveland. OH
44143.
ISMRM. l < l Y 4
12
THIS DOCUMENT IS intended to address considera-
tions with regard to the design and testing of third
party equipment for use with a n MR system. It also
provides information that may be of use to the end user
of that equipment. The information provided herein is
generic with respect to MRI systems. Designers of third
party equipment for use with MR systems, and users
of MR equipment building their own coils and acces-
sories, should be aware that MR is a n evolving tech-
nology and specifications and performance
characteristics of specific MR systems continually
change. Products from third party sources should be
tested under simulated or actual conditions of use. Be-
cause upgrades to the MR system to achieve higher
performance levels may affect the characteristics of the
equipment, manufacturers of accessory devices are ad-
vised that today's testing may not be sufficient to en-
sure the safety and effectiveness of their device with
tomorrow's MR system.
Additional information with regard to system-specificre-
quirements should be obtained directly from the MR mar-
ufacturer.
DeBnition of M R Compatibility
A device can be considered a s MR compatible if it can
be used safely, a s indicated, within the MR scan room
and during the scan procedure without adversely affect-
ing the performance of either the MR system or the device
itself. A device may be MR compatible for certain types of
uses but not for others. The device labeling must clearly
and explicitly define the conditions for safe and effective
use.
Scope
The document covers the following areas: (a)interac-
tion of the equipment with the MR system, (b)interaction
of the MR system with the equipment, (c) potential safety
hazards resulting from the use of the equipment in an
MR scanning environment, (d) special considerations for
surface coils, insert gradient coils, and cervical fixation
devices.
It provides guidance that is intended for use during the
initial design and subsequent installation of accessory
equipment obtained from third parties. Third party
equipment is that produced by someone other than an
original equipment manufacturer (OEM) of MR systems
and intended for use a s an accessory in the MR environ-
ment.
Because the number of possible devices that could be
useful accessories to MR scanners is very large and the
number actually used in the MR environment grows
daily, we have chosen those in most common use as ex-
amples to illustrate the variety of interactions that are of
concern. These include pulse oximeters, ECG monitors,
respirators, anesthetic apparatus, contrast injectors, and
infusion pumps. It also includes scanning accessories
such as surface coils and patient positioning devices and
patient comfort systems such a s audio and video devices.
Although implanted metallic, electronic, or mechanical
devices a s well a s surgical intervention devices are not
explicitly included in this document, the principles stated
herein can be applied to such devices with appropriate
attention being paid to the unique circumstances being
set up by the devices proximity to specific tissue.
0 THE MR ENVIRONMENT
The MR environment is the term used to describe the
general environment in the vicinity of the MR scanner (ie,
magnet). It includes the region from the center of the
magnet bore out to the 5-G line (.0005 T). Beyond the 5-G
line. the levels of magnetic field are generally considered
to be safe, although some effect may still be produced
on devices such as cathode ray tube (CRT) displays and
image intensifiers.
Characteristics of the M R Environment
The MR environment is characterized by the following:
1. The static magnetic field (most commonly in the range
of .2to 2.0 T, although field strengths a s low a s .06 T
and greater than 4 T are also in use) and the associ-
ated spatial field gradient.
2. The rapidly changing magnetic fields (gradients)that
are used during imaging (- 1 Hz).
3. Radiofrequency (RF) pulses (on the order of tens to
hundreds of MHz).
0 INTERACTION OF EQUIPMENT WITH THE MR
SYSTEM
External equipment can interact with the MR system
either through the generation of R F interference or dis-
turbance of the homogeneity of the magnetic field. Either
of these may adversely affect scanner performance and
reduce its diagnostic effectiveness.
RF Interference Considerations
In general, third party equipment should be designed
to meet the requirements of IEC 601-1-2. This standard
defines field levels for RF emissions and electromagnetic
compatibility (EMC).Related standards and tests can be
found in CISPR- 11 and IEC 80 1- 1 through 4. It should
be noted that none of the current standards are specifi-
cally concerned with MR compatibility and additional
steps actually are required. Some general guidance re-
lated to design practices for MRI accessories is given be-
low.
If RF shielding is inadequate, equipment that includes
display screens and/or microprocessors, such a s patient
monitoring equipment, may produce RF interference that
affects image quality. Interference sources include micro-
processor clocks and displays that can generate a wide
range of frequencies. Light emitting diode (LED), liquid
crystal display (LCD),or CRT displays and indicators may
appear visibly steady but are, in reality, being electrically
turned on and off at a high frequency a s a requirement
of the display technology or to conserve power. The re-
sulting clock and display interference can be radiated di-
rectly into the RF room environment or conducted
These documents can be obtained from the ANSI Sales Office.
New York. NY 10036. (212) 642-4900
through the patient interface connections (ECG leads or
pulse pickup) into the imaging area.
To prevent this RF interference from degrading the MR
image, appropriate measures should be used to provide
RF shielding, such a s housing the display in a n RF tight
enclosure or locating the unit outside the RF room. In
either case, the connections between the display unit and
the patient interface should pass through low pass filters
that significantly attenuate (minimum of - 100 dB) fre-
quencies around the Larmor frequency of the MR system
(eg, 21.3 MHz for a 0.5T magnet, or 64 MHz for a 1.5-T
magnet). For patient contact devices, such a s ECG mon-
itors, safety guidelines for patient isolation and leakage
currents must be ensured per the requirements of IEC
60 1- 1.2
Sensors and detectors such as those used in pulse
oximetry equipment contain a n additional interference
source. Pulse oximeters measure the absorption ratio of
a red LED and infrared LED to determine the patients
pulse and blood saturated oxygen content. Units that cy-
cle the LEDS on and off at a rate of several kilohertz can
be a source of radiated RF interference, which can be
picked up easily by the MR receive coil and result in de-
graded image quality, because the sensor is located very
near the receiving coil of the MR system. Shielding of the
sensor is recommended to prevent this noise from reach-
ing the MR system.
Similar RF considerations should be made in the selec-
tion of patient audio and video equipment. The tape deck,
tuner, or CD player for a patient audio system usually is
located near the operators console and a speaker or
headphone set is located within the scan room. MR man-
ufacturers require that signals between the audio equip-
ment and the speaker must penetrate the RF room
through appropriate filters or make use of a nonelectri-
cally conducting medium, eg, fiber optics or infrared
transmission, to prevent RF interference outside the
room from being conducted into the MRI environment.
Specific information about these requirements can be ob-
tained from the Siting Guides, available from most OEM
MR system manufacturers.
Simple solutions may not be available with video equip-
ment such as patient TV. As with audio equipment, con-
trol signals must be filtered at the RF room wall.
Additionally, consideration must be given to the interfer-
ing frequencies that could be generated by the TV display
itself. If a flat panel display is used, it should be properly
enclosed and RF should be shielded to prevent interfer-
ence with the MR system.
Magnetic Field Considerations
Significant masses of ferromagnetic material in prox-
imity to the magnet can disturb the homogeneity of the
static MR magnetic field. Distorted images could result if
the homogeneity of the main magnetic field is degraded
significantly (1). Respirators and anesthesia equipment,
carts, or gas cylinders should be designed to use nonfer-
romagnetic materials. Patient monitoring equipment lo-
cated within the scan room also could have an effect on
the magnetic field homogeneity. Power supply transform-
ers or the frame of the unit could contain sufficient mag-
netic material to result in a distortion of the field.
Other conductive, but nonferromagnetic, materials
may exhibit induced RF currents that can lead to image
distortion. Artifacts caused by nonferromagnetic materi-
1 I West 42nd Street, This document can be obtained from the ANSI Sales Oflice. 1 1 West 42nd Strert
New York. NY 10036. (212) 642-4900
Volume 8 Number 1 JMRl 13
als result from eddy currents that can be generated in
the objects by gradient magnetic fields used for MRI that,
in turn, disrupt the local magnetic field and distort the
image (2.3).
0 INTERACTION OF MR SYSTEM WITH
EQUIPMENT
Operation, reliability, and accuracy of accessory equip-
ment, if not designed correctly, can be affected by the MR
system. These could have serious effects on the safety
and effectiveness of the device. Potential problems that
must be addressed are interactions of the static magnetic
field, FW fields, and gradient fields with the accessory
equipment.
Static Field Effects
Magnets used for MR systems in routine clinical oper-
ation range in field strength from a s low as .06 T (600G)
to 2.0 T (20, 000 G). A number of scanners currently in
clinical investigation operate in the range of 3 to 4.6 T.
Superconductive, cryogenic magnets (>.5T) can only be
turned off through a controlled ramp down or by a n
expensive quench process. Lower field-resistive mag-
nets are easily powered down, whereas low field perma-
nent magnets are always turned on. I t should be noted
that small MR systems, designed for specific applica-
tions, such as extremity imaging, may have very well con-
tained magnetic fields (eg, 5 G is within 6 inches of the
magnet). These particular systems do not represent a
worst case, and the discussion that follows may not ap-
ply. The inherent differences in magnet types must be
well understood before any MR compatible equipment is
designed. These facts are basic to understanding how the
safety and effectiveness of a device can be influenced by
the static field environment produced by the MR system.
Mention must be made of the serious hazard caused by
the potential missile effect that exists when ferromag-
netic materials are brought near the magnet. The strong
magnetic field generated by the MR system extends
throughout the scan room and, in some cases, even be-
yond. The attractive forces increase rapidly as one nears
the magnet, producing a force that can displace even
large objects and pull them to the magnet. Objects such
as oxygen bottles or equipment carts, when pulled to the
magnet, can produce costly damage to the MR equipment
as well a s serious injury to patients and/or workers. De-
vices should be scrutinized carefully to remove ferromag-
netic parts. These may work loose over time and become
a serious problem.
Ferromagnetic objects within the magnetic field are
also subject to forces that try to bring them into align-
ment with the magnetic field. These strong torque effects
can cause metallic objects to twist. For this reason, MRI
is contraindicated for a patient with a n implanted ferro-
magnetic aneurysm clip or similar implant. Device com-
ponents subject to torque effects can break loose,
causing the device to fail or possibly injure a patient or
worker.
Respirators, infusion pumps, contrast injectors, and
anesthesia equipment may contain pumps, electric mo-
tors, or gauges that can be directly affected by the mag-
netic field. The magnets in motors may become saturated
by the fringe field of the MR magnet. Saturation can ei-
ther result in lack of motor operation, a slowing of the
operation, or a n increase in operating current of the mo-
tor, which could ultimately b u m the motor out. Similarly,
electronic circuitry may contain components that are af-
fected by static magnetic fields, such a s transformers,
femte cores, relays, and switches.
14 JMRl JanuarylFebruary 1998
Accuracy of analog style gauges may be affected by the
MR magnet fringe field because the movement mecha-
nism in some types of gauges makes use of a small mag-
net and a coil for operation. Digital gauges are unlikely
to be affected by the magnetic field, but a potential FW
noise source may arise from the numeric displays. The
effect of the magnetic field on magnetic media, such as
digital tape or floppy disks, may cause damage or erasure
of data, so devices used in the magnet room should not
incorporate these types of media for operation or storage
of patient data.
Manufacturers of third party equipment should have
appropriate test data to demonstrate routine operation of
their equipment without compromising its effectiveness
at the maximum magnetic field level within which it may
be placed. For equipment that is located immediately at
the periphery of the magnet covers, the magnetic field
level can be as high a s (or in some cases higher than) the
central field. Within the magnet room, field levels can
range from <.5mT to full system field strength. It is rec-
ommended that equipment intended for use within the
magnet room be clearly labeled with the maximum field
level within it which it can operate effectively. Instruction
must be provided regarding the effects on the equipment
of exceeding that field limit. Critical devices, the failure
of which could cause serious injury to a patient, must be
secured or permanently located in a safe position for op-
eration.
R F Field Effects
Physiologic measurements, using electrical transduc-
ers or equipment, for example patient ECG and pulse ox-
imetry equipment, may be affected by the RF transmit
field. Because the sense leads for this equipment are sit-
uated within the imaging environment, R F energy from
the transmit RF field can easily couple into the leads and
the input amplifiers of the equipment causing either sat-
uration of the amplifier, which could result in momentary
loss of the monitored signal, or permanent damage if ad-
equate protection is not provided.
The amount of FW interference which is coupled into
the leads can be controlled by appropriate design meas-
ures (for example, by selecting materials that are used for
the ECG leads and by incorporating low pass filtering on
the input of the amplifier circuitry).
Gradient Field Effects
Equipment with cables or leads in the magnet bore or
attached to the patient, such a s ECG equipment, also can
be affected by the gradient magnetic fields used in im-
aging. When the ECG leads are connected to the patient,
there is a n electrical loop created by the leads, the pa-
tient, and the ECG conditioning amplifier. The pulsing
gradient fields generate a circulating current within this
loop, resulting in a voltage across the conditioning am-
plifier. This voltage can be several times the magnitude
of the ECG R wave and can result in saturation of the
ECG conditioning amplifier, incorrect triggering of the
ECG equipment, or both.
0 POTENTIAL PATIENT SAFETY HAZARDS
Accessory equipment design should take into account
potential patient safety hazards. These include magnetic
components being attracted to the MR magnet, changes in
ground leakage currents, and possible patient bums
caused by the heating of material placed within the im-
aging area.
Leakage Currents
Consideration must be made when using filters to sup-
press RF interference from ECG and pulse oximetry mon-
itoring equipment. The input amplifier for this type of
equipment should, by design, be isolated from earth
ground to prevent the flow of unacceptable levels of 60-
Hz currents within the patient. If RF filters are placed
directly in series with the ECG leads, there is the poten-
tial for increasing the 60-Hz currents above acceptable
levels because the RF filters are referenced to the RF
room and power supply grounds. To prevent this prob-
lem, a separate battery or isolated powered preamplifier
should be located within the RF room for direct connec-
tion to the ECG leads. The output signals can then be
filtered safely when exiting the R F room.
Heating
ECG leads and pulse pickups placed within the RF
transmit field can provide a path for flow of significant
levels of RF currents through the patient or along the
shield of the cable to ground. If this RF current is not
properly limited, patient b u m s can result at the ECG
electrode location or a t body parts that are located next
to a n insufficiently insulated RF shield. The method of
cable routing within the bore of the magnet can be im-
portant in reducing these effects. In addition, use of con-
ductive leads can be replaced by alternate technology,
such a s fiber optic, carbon fiber, or high impedance
leads, minimizing the possibility of bums.
Equipment that is of primary concern for RF heating is
that which can be positioned directly in the RF transmit
field. A device must not have the potential to become hot
enough to cause injury during scanning. This includes
ECG electrodes and leads, pulse pickup and pulse ox-
imetry sensors, as well a s metallic components, connect-
ors, cables, and surface coils (4-8).
There is very little research available indicating accept-
able levels of RF currents, which can safely flow within a
patient at MRI frequencies. Measurements made during
electrosurgery indicate that RF current flow within the
ECG leads must be limited to less than 100 mA to prevent
RF bums of the patient (9).To provide the same level of
current limitation during MRI, it has been found that
graphite ECG leads must be used during MR studies. Ad-
ditionally, it also may be prudent to locate supplementary
RF impedance in the leads near their connection to the
ECG electrode pads. Metal conductor ECG lead wires are
definitely unacceptable for this purpose.
Pulse pickup and pulse oximetq sensors provide two
possible areas of concern. The first results from the ground-
ing of the shield incorporated into the sensor cable. If this
cable is placed too close to the RF transmit field, significant
RF currents can flow down the shield. The RF currents can
result in thermal heating of the cable and body parts lo-
cated next to the shield and may result in bums.
A second pulse pickup consideration is coupling of the
transmitted RF field with the LEDs and photo detectors
used in pulse sensors. It is possible for the LEDs and
photo diodes to rectify the transmitted RF and dissipate
a large amount of power in a short period of time. This
can result in localized heating of the sensor and has been
reported to cause bums in some cases. Not only must
shielding be provided to prevent noise generated by the
sensor from degrading the image but also to prevent cou-
pling of the RF transmit field with the sensor.
Magnetic Components
As previously stated, forces causing displacement
("missile effect") or torque effects on ferromagnetic ma-
terials may produce a serious hazard. Equipment in-
tended for placement close to the MR magnet should
contain a minimum of ferromagnetic material. Manufac-
turers should understand the need to minimize or elim-
inate the use of magnetic material in their products to
avoid equipment being attracted to the magnet. The force
can be significant and could result in injury of personnel
trying to restrain the equipment or being trapped be-
tween the equipment and the magnet. Additional precau-
tions should be taken during the course of product
design to prevent the inclusion of small components that
could be pulled loose and attracted to the magnet bore.
For example, connector covers that are ferromagnetic
must be attached to the equipment by a chain. Remova-
ble equipment covers should use captive hardware
(screws and fasteners). Battery-powered equipment may
seem to be nonmagnetic, but it is very important to use
batteries that are totally nonmagnetic. Battery-powered
equipment should be tested at its intended location and
maximum field strength to ensure that there is no sig-
nificant attraction.
0 SPECIAL PURPOSE RF COILS
The design of special purpose coils and associated pa-
tient positioning devices for use in MRI requires special
attention. Guidance, based on the collective experience
of MR manufacturers, is given below to allow avoidance
of safety, image quality, and reliability issues (private
communications). It is hoped that this will be helpful for
those producing RF coils for their own use (research) or
as after-market MR accessories.
Magnetic Effects
All components that are used in RF coils should be pre-
screened to ensure that they do not affect image quality.
Electronic components should be specified a s nonmag-
netic. However, care must be taken because, apparently,
even nonmagnetic components may have some magnetic
content, for example, housings and leads with nickel
flash. Brass bolts, nuts, and screws used to fasten the
coil onto the housing can contain up to a 1% contami-
nation with iron or nickel. The associated magnetic dis-
tortions are most severe in gradient-recalled echoes with
a high flip angle. By affixing sample components to the
exterior of a n imaging phantom and running such a pulse
sequence, penetration of any artifacts into the imaging
region of interest can be determined. A simple procedure
for evaluating image distortion can be found in National
Electrical Manufacturers Association (NEMA)Standards
Publication No. M S 2, Determination of Two-Dimensional
GeometricDistortion in Diagnostic MagneticResonancelm-
ages." An extensive discussion of the techniques for eval-
uation of various materials with respect to the distortion
of MR images has been done by Schenck (1). The per-
formance specification for image distortion should be ob-
tained from the MR manufacturer.
Proton Signal
Special purpose coils are normally housed in plastic
packaging. Materials used range from Plexiglas to ure-
thane. Care must be taken to ensure that the materials,
including adhesives, do not generate a n MR signal, be-
cause, depending on field of view and phase-encoding di-
rection, this signal may end up in the region of interest
and produce image artifacts. Signal arises from protons
i This standard can be obtained from NEMA Customer Semces. 1301) 841-3200.
Volume 8 Number 1 JMRl - 15
in the plastic molecules. Although the signal decay, char-
acterized by T2 relaxation time, usually is very short,
with progress in MR pulse sequence development, TEs
have become shorter and shorter and signal from the coil
housing may be visible in the image. The same is true of
molded foam pads that are used for patient positioning.
Any materials that are to be used should be evaluated
with the shortest TE pulse sequence used in clinical prac-
tice to determine whether signal from the material could
cause image problems.
Gradient Eddy Currents
Every coil or cervical fixation device that contains a
continuous loop or copper plane that is not interrupted,
for example, by capacitors, will create gradient eddy cur-
rents, adversely affecting the rise and fall times of the
gradient pulses, thus potentially degrading image qual-
ity.
Cable Resonances
Special purpose coils have a cable that is grounded
where the cable connects to the MR equipment. Interac-
tion with the RF electric field of the body coil can create
high voltages on the cable shield. If these are not properly
taken care of via balanced/unbalanced circuits and a de-
vice that prevents the cable from looping, the signal-to-
noise ratio of the surface coil can be adversely affected
and, in case of a cable loop, arcing can occur between the
cables, which may be dangerous to the patient.
System Concerns
Improper surface coil design can affect the performance
and reliability of the MR system. The output channel of
special purpose coils should be protected against over-
voltage with a diode network or damage to the input stage
of the system preamp may result.
Transmit/receive (T/R) coils exhibit mutual coupling
between the body coil and the T/R coil, which is a func-
tion of relative size and proximity. T/R special purpose
coils do not contain any decoupling and, therefore, mu-
tual coupling may be severe, in which case the body coil
may become detuned and mismatched, causing power re-
flection from the body coil to its interface circuitry. De-
pending on the amount of power, the transmit/receive
switch in the coil interface may be damaged.
T/R special purpose coils not designed and tested in
conjunction with particular MR systems may cause the
power optimization phase of the scan to fail, resulting in
incorrect setting of the flip angle. There is particular con-
cern with T/R coils of small dimensions compared with
the system standard coils. For example, the system may
optimize for a 270" flip angle or even a 450" flip instead
of 90". With the incorrect flip angle, image quality (con-
trast and uniformity) will be affected. In addition, esti-
mation of specific absorption rate (SAR) levels for RF
exposure may be incorrect, a potential safety hazard that
may result in localized tissue heating. T/R special purpose
coils should contain an attenuator or other means for en-
suring that the system flip angle is correctly set.
Receive-only special purpose coils are decoupled from
the body coil during the transmit pulse via a n internal
electronic circuit. Electrically decoupled coils use either
passive decoupling, in which the decoupling is achieved
via a parallel tank that contains fast switching diodes
switched by the induced RF voltage during the transmit
pulse, or active decoupling, in which the diodes are ac-
tively biased via a direct current (DC) control pulse.
16 JMRl January/February 1998
Care must be taken with all passively decoupled coils to
ensure that components are correctly rated to withstand
RF currents induced from the body coil RF field. For ex-
ample, if the diode chosen cannot handle the current in
the decoupling circuit during the transmit pulse, the diode
will blow and may cause an open circuit. The result will
be that the coil is no longer decoupled from the body coil,
potentially affecting performance of both coils. If an induc-
tor in the tank circuit does not have the proper gauge wire
to handle the current flowing through it, it may become
very hot, in which case the housing of the coil can melt or
even catch fire. In the extreme case, the inductor will un-
solder itself, creating coupling with the body coil.
Special purpose coils must be designed to work with
pulse sequences with low flip angles. If not correctly de-
signed, the decoupling circuits may not operate because
the associated transmit power may be insufficient to bias
the diodes. Slice profile also may be affected because the
amplitude of the side lobes of the RF pulse is insufficient
to bias the decoupling diodes.
Care must be taken with the design of actively decou-
pled coils to ensure protection of devices if the surface
coil is present inside the body coil but is not plugged in.
In this case, the diodes are not DC biased. The conse-
quences depend on the type of diode used, but diode fail-
ure is a potential outcome.
0 INSERT GRADIENT COILS
Removable gradient coils, designed to be inserted into
the magnet bore, can be used to produce very steep gra-
dient magnetic fields for rapid imaging techniques. Re-
search groups or manufacturers designing these
accessory devices may find the following information
helpful.
Torque
A gradient coil may or may not be torque compensated.
In a torque-compensated coil, torque on one side of the
coil is of equal but opposite sign to that on the other side
of the coil. An uncompensated coil in a DC magnetic field
will rotate when currents are applied to the transverse
gradients (xand y for a DC field in the z direction) and
will tend to shift when current is applied to the parallel
gradient (z gradient). A compensated coil may do the
same if a fault condition occurs where the coil is damaged
internally (eg, a short). The resulting torque and forces
depend on the amplitude of the gradient pulse and/or the
gradient field strength and on the amplitude of the DC
magnetic field. It should be noted that torque also can
occur when the gradient coil is located outside of the sys-
tem, where the magnetic field is nonuniform and nonlin-
ear. Although there have been no reported incidents of
patient injury related to the use of insert gradient coils,
the torque and forces can be of sufficient intensity to cre-
ate a safety concern. For this reason, all insert gradient
coils should be surrounded with a frame such that the
coil is essentially locked in place. The frame may be de-
signed to withstand the high forces and torque created.
It also should be capable of withstanding high accelera-
tion forces because the torque usually is created within
100 ps, depending on the rise time of the gradient.
Again, severe torques also will occur if the operator
uses the coil in any area outside the high homogeneity
sphere of the magnet (ie, in an area where the DC mag-
netic field is inhomogeneous). This situation should be
prevented by mechanical or electrical (microswitch)
means so that the coil can only be used in the isocenter
of the magnet.
Acoustic Noise
An insert gradient coil is much closer to the ears of the
patient than a whole body gradient coil. For this reason,
ear protection should be required for the patient. This
should be indicated on the insert coil housing in clear
view of the operator (10-17).
dB/d t
Because of the low inductances and fast rise times,
small insert gradient coils may produce dB/dt levels that
exceed recommended safety limits. Insert gradient man-
ufacturers should calculate the maximum dB/dt values
and limit a patients exposure either by software (limiting
the gradient waveforms allowed peak and slew values) or
by hardware (limiting the gradient amplifiers output).Al-
ternatively, the housing of the gradient coil could be de-
signed so that sufficient distance exists between the
localized dB/dt peaks at the coils edge and the patient.
Heating
Insert gradient coils often become quite hot because of
the high current densities. If the temperature becomes
too high, the gradient coil material may lose its physical
strength, which again is potentially dangerous due to the
high torques and forces. These limitations should be well
understood and the data provided by the coil manufac-
turer should describe operating conditions that cause the
temperature of the coil to rise above 40C. In addition,
there are limits to the temperature of parts that can be
in contact with the patient.
Because of these dangers, methods of eliminating haz-
ards due to excess heating must be provided, either by
providing a cooling mechanism (air, water, or both) or by
providing hardware feedback to the MRI system that lim-
its the output of the gradient amplifiers to safe levels.
In either case, thermocouple or heat-sensitive snap
switches should be mounted at critical spots of the gra-
dient coil so the gradient amplifiers may be switched off
when a certain temperature is exceeded.
The material being used for the actual gradient coil
should be of a high flame rating (UL-5V) and should
maintain its strength at high temperatures.
Interface
The gradient coil should be connected to the system via
cables that can handle the applied current amplitudes
and duty cycles. Cables for forward and return currents
should be self canceled or secured such that the cables
do not move in the magnetic field upon application of a
gradient pulse.
A connector interface box should be mounted on the
system with different connectors for the body gradient and
the insert gradient. This box should be interlocked in such
a way that no one can touch live parts without switching
off the gradient amplifiers first. The connectors should
only fit one way, such that reversing the gradient field is
impossible.
Eddy Currents/Load Compensation
When using a n unshielded body gradient coil, eddy
currents are created in the cryostat screen and other
nearby metal parts. These eddy currents increase the rise
time of the coil and have undesired effects on the phase
of the NMR signal. An eddy current compensation circuit
in the gradient amplifier typically is used to correct for
this effect.
Because insert gradients usually are much farther
away from the cryostat screen, the eddy currents may be
negligible. Therefore, upon insertion of the local gradient,
the eddy current compensation may be changed via hard-
ware or software means, if required.
The insert gradient coil will typically present a different
load to the amplifier than the body gradient coil and,
thus, a load-compensating network may be required
when switching between the whole body and insert gra-
dient coils. Failure to provide proper compensation may
result in damaged coils or gradient amplifiers.
Magnetism/Proton Signal
No magnetic parts should be included in any device
that is inserted in the bore of the magnet, not only for
safety reasons but also to maintain a highly homogene-
ous DC magnetic field.
Gradient coil material in the vicinity of the RF coil
should not give off any proton signal. This signal could
fold back into the image and be mistaken for pathology.
CERVICAL FIXATION DEVICES
Some special concerns are presented by cervical fixa-
tion devices or halos. The design of certain types of ha-
los include a continuous metallic ring with no
interruptions of a capacitive nature. In this case, there is
a n interaction with the RF coils and with the gradient
coils. The halo will couple with the RF field from transmit
a s well a s receive coils. Although the ring is not resonant,
RF currents will be induced during the transmit pulse
but with a lower amplitude than in a resonant ring. De-
pending on the amplitude of the mutual inductive cou-
pling, these currents can create RF magnetic fields that
locally exceed SAR limits in the patients tissue. The halo
will couple to the gradient coil and adversely affect the
rise time of the gradient pulses that will lead to phase
errors in the image. Each halo device should be evaluated
carefully to determine whether diagnostic images can be
produced despite these imaging artifacts. Although there
is no published evidence to document this, there is con-
cern that if metallic screws are used to secure the halo
to the patients skull, a possibility exists that the screws
may heat up during scanning due to a buildup of F W cur-
rent density and may cause injury to the patient. Tests
should be designed and performed to verify that heating
during scans does not occur (18-20).
CONCLUSION
It is the hope of NEMA member companies that this
document will be helpful in the design and selection of
accessory MR equipment. MR is a continuously evolving
technology with new developments and clinical applica-
tions being introduced at a rapid rate. We anticipate that
this document will be updated periodically to keep pace
with this rate of development.
References
1. Schenck JF. Review article: role of magnetic susceptibility in
MRI. J Med Phys 1996: 23(6);815-850.
2. Bellon EM, Haacke EM, Coleman PE, el al. MR artifacts: a re-
view. Magn Reson Imaging 1984; 2:41-52.
3. Pusey E. Lukin RB, Brown RKJ, et al. Magnetic resonance
imaging artifacts: mechanism and clinical significance. Radi-
ographics 1986: 6:891-911.
4. Fisher DM. Lilt W, Cote CJ. Use of oximetry during MR imag-
ing of pediatric patients. Radiology 199 1: 178:891-892.
5. Kana1 E, Shellock FG. Patient monitoring during clinical MR
imaging. Radiology 1992: 185:623-629.
Volume 8 Number 1 JMRl 17
 6. Shellock FG, Myers SM, Kimble K. Monitoring heart rate and
oxygen saturation during MRI with a fiber-optic pulse oximeter.
Am J Roentgenol 1991; 158:663-664.
7. Shellock FG. Slimp G. Severe b u m of the finger caused by us-
ing a pulse oximeter during MRl (letter).Am J Roentgenol 1989:
153:1105.
8. Shellock FG. Kanal E. B u m s associated with the use of mon-
itoring equipment during MR procedures. JMRl 1995: 6 2 7 1 .
9. DeRosd J F , Gadsby PD. Radio-frequency heating under RF
electrodes. Med Instrum 1979; 13(5):273-276.
10. Brummett RE, Talbot JM, Charuhas P. Potential hearing loss
resulting from MR imaging. Radiology 1988; 169:539-540.
11. Goldman AM, Gossman WE, Friedllander PC. Reduction of
sound levels with antinoise in MR imaging. Radiology 1989:
1731549-550.
12. Hurwitz R, Lane SR. BeU RA, Brant-Zawadzh MN. Acoustic
analysis of gradient-coil noise in MR imaging. Radiology 1989:
173:545-548.
13. Shellock FG. Morisoli SM, Ziarti M. Measurement of acoustic
noise during MR imagmg: evaluation of six Worst case pulse se-
quences. Radiology 1994; 191:91-93.
14. McJury M, Blug A, Joerger C, Condon B, Wyper D. Acoustic
noise levels during magnetic resonance imaging scanning a t
1.5T. f3r J Radiol 1994: 67:413-415.
15. McJury M J . Acoustic noise levels generated during high field
MR imaging. Clin Radiol 1995: 30:331-334.
16. Miller LE, Keller AM. Regulation of occupational noise. In: Har-
ris CM, ed. Handbook of noise control. New York: McGraw-Hill,
1979: 2.
17. Melnick W. Hearing loss from noise exposure. In: H a m s CM,
ed. Handbook of noise control. New York: McGraw-Hill, 1979:
1-16.
18. Ballock FC. Hajed PC, Byme TP, et al. The quality of magnetic
resonance imaging, as affected by the composition of the halo
orthosis. J Bone Joint Surg 1989: 71A:431-443.
19. Clayman DA, Murakami ME, Vines FS. Compatibility of cer-
vical spine braces with MR imaging: a study of nine nonferrous
devices. Am J Neuroradiol 1990: 11:385390.
20. Shellock FG, Slimp G. Halo vest for cervical spine fixation dur-
ing MR imaging. Am J Roentgenol 1990; 154:631-632.
Suggested Readings
1. Occupational Safety and Health Administration. Occupationalnoise
exposure. 29CFR Part 1910.95. Washington, DC: Occupational
Safety and Health Administration, Department of Labor, 1997.
2. Adair ER. Berglund LG. On the thermoregulatoly consequences of
NMR imaging. Magn Reson Imaging 1986: (4):321-333.
3. Athey TW. Current FDA guidance for MR patient exposure and con-
siderations for the future. Ann N Y Acad Sci 1992: 649242-257.
4. Barnett GH, Roper MID. Johnson AK. Physiological support and
monitoring of critically ill patients during magnetic resonance imag-
ing. J Neurosurg 1988: 68:246-250.
5. Boesigner P. Buchli R. Saner M. Meier D. An overview of electro-
magnetic safety considerations associated with magnetic resonance
imaging. Ann N Y Acad Sci 1992: 649: 160-165.
6. Budinger TF. Nuclear magnetic resonance technologies:health and
safety. Ann N Y Acad Sci 1992: 649:l-19.
7. Dunn V~ Coffman CE. McGowan JE. Ehrardt JC. Mechanical ven-
tilation during magnetic resonance imaging. Magn Reson Imaging
1985; 3:169-172.
18 JMRl - January/February 1998
8. Food & Drug Administration. Guidance for content and review of a
magnetic resonance diagnostic device 5 10(k) application: safety
parameter action levels. Rockville. MD: Center for Devlces and Radi-
ological Health Report, 1988.
9. Kagetsu N J . Litt AW. Important considerations in measurement of
attractive force on metallic implants in MR imagers. Radiolo~g1991;
179:505-508.
10. Hubbard A. Markowitz R, Kimmel B. Kroger M. Bartko M. Sedation
for pediatric patients undergoing CT and MRI. J Comput Assist Tom-
ogr 1992; 16:333-336.
11. Kanal E, Shellock FG. MR imaging of patients with intracranial an-
eurysm clips. Radiology 1993: 187:612-614.
12. Kanal E. An overview of electromagnetic safety considerations as-
sociated with magnetic resonance imaging. Ann N Y Acad Sci 1992;
649:204-224.
13. Lewin JS. Duerk JL, Jain VR. et al. Needle localization in MR-
guided biopsy and aspiration: effects of field strength, sequence de-
sign, and magnetic field orientation. Am J Roentgenol 1996: 166:
1337-1345.
14. Lufiin R, Jordan S. Lylyck P. et al. MR imaging with topographic
EEG electrod% in place. Am J Neuroradiol 1988: 9:953-954.
15. McGowan JE. Erenberg A. Mechanical ventilation of the neonate
during magnetic resonance imaging. Magn Reson Imaging 1989: 7:
145- 148.
16. Moscatel M, Shellock FG. Monsoli S. Biopsy needles and devices:
assessment of ferromagnetism and artifacts during exposure to a 1.5
Tesla RM system. JMRI 1995: 5:369-372.
17. Person B. ed. Potential health hazards and safety aspects of clinical
NMR examinations. Semin Biomed Appl NMR. Lund Radiation Physics
Department, 1984.
18. Reilly JP. Cardiac sensitivity to electncal stimulation. Report MT.
Washington, DC: Food & Drug Administration, 1989: 89-101.
19. Reilly JP. Peripheral nerve stimulation by induced electric currents:
exposure to time-varying magnetic fields. Med Biol Eng Comput
1989; 27:lOl-110.
20. Schaefer D J . Safety aspects of magnetic resonance imaging. In:
Wehrli FW,Shaw D. Kneeland JB, eds. Biomedical magnetic reso-
nance imaging 1988: 13:553-578.
21. Schaefer DJ. Dosimetry and effects of MR exposure to RF and
switched magnetic fields. Ann N Y Acad Sci 1992: 649225-236.
22. Selden H. DeChateau P. Ekman G, et al. Circulatory monitoring of
children during anesthesia in low-field magnetic resonance imaging.
Acta Anesthesiol 1990: 34:41-43.
23. Shellock F, Kanal E. Magnetic resonance: bioeffects, safety and pa-
tient management. New York: Raven Press, 1994.
24. Shellock FG, Kanal E. Magnetic resonance: bioeffects. safety. and
patient management, 2nd ed. New York: Raven Press, 1996.
25. Shellock FG. Pocket guide to MR procedures and metallic objects:
update 1994. New York Raven Press, 1995.
26. National Electrical Manufacturers Association. Determination of
two-dimensional geometric distortion in diagnostic magnetic reso-
nance images. Standards Publication No. MS 2. Rosslyn. VA:
National Electrical Manufacturers Association. 1990.
27. National Electrical Manufacturers Association. Characterization of
special purpose coils for diagnostic magnetic resonance images.
Standards Publication No. M S 6. Rosslyn, VA: National Electrical
Manufacturers Association, 1991.
28. National Electrical Manufacturers Association. Characterization of
the specific absorption rate for magnetic resonance imaging systems.
Standards Publication/No. M S 8. Rosslyn. VA: National Electrical
Manufacturers Association, 1993.
29. Villa M, Mustarelli P. Caprotti M. Mini review: biological effects of
magnetic fields. Life Sci 1991: 49:85-92.