3/2/18

1

PotentialRisksofMRIinDevicePatients

RedhaBoubertakh

r.boubertakh@qmul.ac.uk

Barts HealthNHSTrustQueenMaryUniversity

• MRIandcardiacimplantableelectronicdevices(CIED)• ComponentsofanMRIscanner• MRIimplantanddevicesafetyclassification• RisksassociatedwithscanningCIEDs−Mainmagneticfield− Radiofrequencywaves− Time-varyingmagneticfieldgradients

• Summary&Conclusion

Outline

• Numberofpatientsfittedwithcardiacimplantableelectronicdevices(CIED)isgrowing• LargepercentagewillrequireanMRIscanovertheirlifetime• Inthepast,CIEDshavebeenconsideredanabsolutecontraindicationtoMRI

MRIandCardiacImplantableElectronicDevices

• However,recentstudieshaveshownthat,ifstrictscreening,safetyandmonitoringproceduresarefollowed,legacynonMRConditionalCIEDscanalsobescannedwithminimalriskstothepatient

MRIandCardiacImplantableElectronicDevices

Heart,2015

HeartRhythm,2017NewEnglandJournalofMedicine,2017

NewEnglandJournalofMedicine,2017

ComponentsofanMRIScanner

StaticmainmagneticfieldB0 Radiofrequency(RF)pulses

• Magneticfields~50,000- 100000× strongerthantheearth’s

• Alwayson!

TransmitcoilBodyorT/Rlocalcoil

Perturbednetspinmagnetisation

ReceivecoilBodyorlocalcoil

Time-varyingmagneticfieldgradients

• LinearmagneticfieldgradientareusedtoencodetheMRsignal

GeneralMRISafetyRisksSource Safetyrisks

StaticMagneticfieldB0 field(1T,1.5T,3T,…)

- Projectile or“missile”effect- Displacementandtorqueeffects- Devicedisruption- Bioeffects

Radiofrequency(RF)pulsesB1 fieldAmplitude~μT,MHzfrequencyrange

- Tissueheating,Burns- Medical deviceheatinganddisruption- Interferencewithequipment(monitoringsystems)anddevices

TimevaryingGradientMagneticFieldGx,y,z gradients.Strength(mT/m)andslewrate(mT/m/ms)scannerdependent

- Peripheralnervestimulation(PNS)- Acousticnoise- Interferencewithequipments anddevices

CryogensLiquidHeliumat-269oC (4K)

- Burns- Asphyxia- Hypothermia

GadoliniumBasedContrastAgents -Nephrogenic SystemicFibrosis(NFS)

AffectCIEDs

3/2/18

2

• Anydevice/implantfallsintooneofthesethreecategories

• No CIEDisMRSafe

• AnMRConditionaldevicecanbesafelyscannedifconditionsdefinedbythemanufactureraremet

MRISafetyTerminologyforImplantsandDevices

• ThedevelopmentofMRConditionalCIEDs(2008)hasmadeMRIscanssafetouse• Conditionofuseinclude:− Thetypeofdevice

• Generator+Leadstype• Programmingmodesandparameters,• Timesinceimplantation,…

− TheMRIenvironment• Typeofmagnet,maximumvalueoffieldstrength• Spatialmagneticfieldgradient• Timevaryingmagneticfieldgradients• Typeofimagingsequence,inducedheating• Imagedbodypart• Typeofimagingcoilsused,…

MRISafetyTerminologyforImplantsandDevices

PotentialRisksofCIEDsintheMRIEnvironment

• ThestaticmagneticfieldB0(mainly1.5Tor3Tonclinicalscanners)isoneofthemainsourcesofdanger• Strongattractiveforceexertedonferromagneticobjects− “Projectileeffect”

• Additionally,adevicecanexperienceatorque(rotationalforces)toalignitwiththedirectionofB0

AssociatedRisks:• Devicemotion,vibrationanddisplacement

MainMagneticField

MagneticfieldB0

• Attractiveforcesdependontheferromagneticcontentofthedevice− Highferromagneticcontent➜ higherrisk

• Animportantparameteristhespatialmagneticfieldgradient− Rateofchangeofthestaticmagneticfieldwithinandaroundthemagnet

StaticFieldCharacteristics

7

Siemens AG Healthcare Sector 06/12 MR compatibility data sheet

Side view of the magnet

3

4

5

−5−5

−4

−3

−2

−1

0

1

2

5−4 −3 −2 −1 0 1 2 3 4

0.5 mT

1 mT3 m

T

5mT

10 mT

20 mT

40 mT

200 mT

Z Axis (m)

Vert

ical

Axi

s (m

)

Siemens AG Healthcare Sector 06/12 MR compatibility data sheet

8

Top view of the magnet

3

4

5

−4

−3

−2

−1

0

1

2

−5 −4 −3 −2 −1 0 1 2 3 4 5−5

0.5 mT

1 mT3 m

T

5 mT

10 mT

20 mT

40 mT

200 mT

Hor

izon

tal A

xis

(m)

Z Axis (m)

Siemensscanner

B0 field- Sideview B0 field- Topview

Siemens AG Healthcare Sector 06/12 MR compatibility data sheet

10

In the following we show a plot representing the 0.5 T, 1 T, and 1.5 T iso-magnetic con-tours at positions accessible to and relevant for the MR worker as far as the static magnetic field in the isocenter exceeds any of these values.

⊗: At this location, the value of the magnetic field Bo is greatest.

! The iso-magnetic contour lines are accurate to a value better than 1%. There is very little influence from the environment to the scanner. The graphic shows the cover in nominal position. On a variety of scanners, the position can vary ±5 mm in the axial direction. The maximum value therefore has a tolerance of ±10%, as small geometric deviations can cause a significant change in the value.

−0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2

1.2

1.4

1.6

1.8

0.2

0

0.2

0.4

0.6

0.8

1

1.5 T

1 T

0.5

T

1.9 T

1.5 T

Z Axis (m)

Vert

ical

(m

)

Side view

Patient table

Spatialfieldgradient- Sideview

• Attractiveforcesareproportionaltothespatialgradientofthestaticfield− Forcehighestattheboreentrance

• However,torque isproportionaltothestaticfieldstrength− Torquelargestatthecentre ofthemagnetbore

• Bothforcesdependontheferromagnetic/paramagneticcontentofthedevice

• MRconditionaldevicesaredesignedwithreducedferromagneticcontent− Stainlesssteelandtitaniumalloysused

MainMagneticField

3/2/18

3

• Evenfornon MRConditional pacemakers,theserisksarelessofanissuefordevicepost2000duetolowferromagneticcontent− Subcutaneoustissuefibrosisaroundthedevice

• Reductionindevicesizethroughouttheyearshassensiblydiminishedpotentialrisks

MainMagneticField

St.JudeMedical

Luechinger,Retal,PACE20012017HRSexpertconsensusstatement,HeartRhythm

• Fieldstrengthandspatialmagneticfieldgradient arespecifiedintheconditionsofuseofMRConditionaldevices

• Example:− Horizontalcylindricalboremagnet,clinicalMRIsystemswithastaticmagneticfieldof1.5Tesla (T)mustbeused

− Maximumspatialmagneticfieldgradient of750G/cm

MainMagneticFieldConditions

• AnMRConditionalCIEDdoesnotalwaysallowallbodypartstobescanned

• MRIconditionsusuallyalsoincludepossibleExclusionZones

• IfanexclusionzoneofanMRConditionaldeviceisimaged− Deviceisusedoff-label− FollownonMRConditionalsafetyprotocols

ImagingExclusionZones

• Strongmagneticfieldscanresultindevicemalfunction• FornonMRConditionaldevices,reedswitches havebeenshowntomalfunction• Areedswitchisusedtoprogramdevices− Asmallmagneticfieldisusedtochangedevicemodetoasynchronouspacing

ReedSwitchMalfunction

• Switchbehaviour hasbeenshowntobeunpredictabledependingonorientation− Unexpectedswitchopeningorclosure

• Electronicsolid-stateHalleffectsensors haveimprovedreliability− Behaveinapredictablemanner

ReedSwitchMalfunctionMRI�AND�PACEMAKER�REED�SWITCH

PACE,�Vol.�25,�No.�10 October�2002 1421

Figure� 2. Influence� of� reed� switch� orientation� withrespect�to�the�main�magnetic�field�B0 switch�state�in�amagnetic� field�. 200� mT.� (A)�The� light� gray� sectorsrepresent�open� switch�orientations� and� the� dark� graysectors�indicate�closed�reed�switch�orientations.� (B)�Arotation�of�180�degrees�around�the�long�axis�of�a�switchwill�invert�the�reed�switch�state.

the�reed� switch�of� the�other� device� was�closed,provided�that�both�devices�had�the�same�orienta-tion.�The�reason�for�this�observation�was�that�thereed�switches�of�the�devices�were�rotated�180�de-grees�against�each�other�along�the�long�axis�of�theswitch�(as�shown�in�the�left�two�reed�switches�inFigure�2B).

Discussion

Since� it� is� easily� possible� to� initiate�asyn-chronous�pacing�mode�by�application�of�a�smallmagnet,� one� would�expect� that�an�MRI� scannerthat�exerts�much�stronger�magnetic� fields�wouldalso� do� the� same.� Closure�of� the�reed� switch�atmagnetic� field�strengths�used�for�clinical�MRI�at0.5–1.5� T�was�generally� expected� in�previouslypublished�reports� studying�interactions�betweenpacemakers�and�MRI�scanners.6,8 However,�the�re-cently�reported�study�by�Sommer�et�al.9 suggeststhe�possibility�that�pacemakers�that�initially�con-vert�into�asynchronous�pacing�mode�may�fall�backto�programmed�pacing�mode�when�positioned�inthe�bore�of�the�magnet�for�MRI.�The�present�studyconfirms�these�findings�and�provides�an�explana-tion�for�such�observations.

The� following� physical� effects� occur� on� apacemaker,�and�thus�on�the�reed�switch,�when�ex-posed�to�an�external�magnetic�field�(B0).�The�mag-netic�field�induces�magnetic�dipoles�(~ P�and—Pas� shown� in Fig.� 3) in� the� ferromagnetic� reeds.This�magnetization�depends�on�the�magnetic�fieldin�a�highly�nonlinear�way.�The�magnetic�dipolesof� the�reeds� interact�with�each� other� (attractionforce� [FD])�of� two�different�poles�~ P�and—P’).This�attracting�force�leads�to�a�closed�reed�switchin�a�lower�magnetic�field.�In�addition,�each�of�thereeds�interacts�with�the�external�magnetic�field�B0.The�force�and� torque�effects�of�MRI�scanners�onpacemakers�and�ICDs�have�been�investigated�pre-viously�by� the�authors.10 It�can�be�expected� thatthese�effects�will�also�act�on�each�of�the�two�reeds.The�magnetic�torque�(T�$ m�@ B0,�with�m�$ mag-netic�moment�of�the�dipole)�tries�to�rotate�a�ferro-magnetic� material� parallel� to� the�magnetic� field(magnetic�torque�is�shown�in�Fig.�3�with�two�pairsof�force�vectors�% FT and�% F’T).�Furthermore,�inan� inhomogeneous�magnetic� field�a� force� (FF� $

mÑB0,�with�m�$ magnetic�moment�of�the�dipole)will�act�on�each�reed.�The�magnetic�moment�m�isequal�for�both�reeds�and�FF will�act�in�the�same�di-rection�and�does�not� influence� the�reed� switchstate.�Furthermore,�in�the�isocenter�of�the�MRI�de-vice�the�field�has�to�be�constant�(ÑB�$ 0).�There-fore,�the�two�forces�(FF)�will�be�neglected�for�thisevaluation.�In�low�external�fields�(, 50�mT),�theattracting� force� of� the� two� dipoles� is� muchstronger�than�the�forces�induced�by�the�torque�ef-

50%�of�all�tested�orientations,�the�reed�switch�re-mained�open.�Figure� 2� shows�different�orienta-tions�of�a� reed� switch�with� respect� to� the�mainmagnetic� field� and� the� observed� state.� In� theisocenter�of�all�three�scanners�the�reed�showed�thesame�paradoxical�opening.

The�observations� were� not� limited�to�posi-tions�along�the�central�axis�of�the�MRI�device.�Atother�positions�away�from�the�central�axis,�takinginto�account�the�possible�changes�in�the�orienta-tion�and�strength�of�the�magnetic�field,�no�changesin�the�described�reed�switch�behavior�could�be�ob-served.

By�evaluation�of�the�state�of�the�reed�switchmounted�in�the�pacemakers�in�the�high�magneticfield,�the�orientation�of�the�reed�switch�could�beidentified.�In�both�pacemakers� the�reed�switcheswere�mounted�in�parallel�to�the�connector�block.When� the�pacemaker� case� was�opened,� the�ex-pected�reed�switch�orientation�was�observed�indi-cating�that�the�reed�switch�state�is�not�being�influ-enced�from�the�other�metallic�components�of�thepacemaker� system.�The�reed�switches�seemed�tobe�oriented�in�the�same�way�in�both�pacemakers,but�they�showed�opposite�states.�In�other�words,when�the�reed�switch�of�one�pacemaker�was�open,

Scheidegger,Metal,PACE2002

• MRIsequencesusemagneticgradientstospatiallyencodetheimagedobject• Importantscannerparameters:− Slewrate(upto200T/m/s)− Maximumgradientstrength(suchas45mT/m)

Time-VaryingMagneticFieldGradients

40oTR TR

TE

RFpulses

Slicegradient

Phaseencodinggradient

Frequency encodinggradient

Risetime

Maxstrength

3/2/18

4

• RapidlyswitchinggradientscaninduceElectromagneticinterference(EMI)

• Electricalcurrentsandvoltagescanbeinducedinconductionwires−Deviceleads−Generator

Time-varyingMagneticFieldGradients

• InterferencescouldbeinterpretedbyaCIEDasarealormissingheartrhythmsignal(oversensing,undersensing).EMIscanleadto:

• TherapyinhibitionApacemakermaybewithholdpacing− Pacingdependentpatients

• Innapropriate shocksAnICDmayinterpretaninterferenceasrequiringanunnecessary

shock

EffectsofEMI

EffectsofEMI

Beinart Retal,Circulation2013

High-frequencynoiseandnoiseresponseventricularpacingarenotedonthebipolarventricularchannelfroma

single-chamberpacemakerduringMRI

406 ! A.N.E. ! October 2005 ! Vol. 10, No. 4 ! Gimbel, et al. ! Frequent Artifacts During MRI

Table 1. MRI Indication, Region and Scan Type, ILR Status Post MR

Scanning Event Device Indication Scan Type Post-MRI ILR Status

1 Reveal Plus Head trauma GE 1.5 T Normal function9526 Cranial Artifact: none

2 Reveal Plus Shoulder trauma GE 1.5 T Normal function9526 Shoulder Artifact: Wide and narrow

complex tachycardias3 Reveal Plus Suspected brain mass GE 1.5 T Normal function

9526 Cranial Artifact: none4 Reveal Plus Confusion GE 1.5 T Normal function

9526 Lumbar-spine Artifact: SVT5 Reveal Plus Headache GE 1.5 T Normal function

9526 Cranial Artifact: CHB, SVT6 Reveal Plus Evaluation of multiple sclerosis GE 1.5 T Normal function

9526 Cranial Artifact: SVT7 Reveal Plus Suspected brain mass GE 1.5 T Normal function

9526 Cranial Artifact: none8 Reveal Plus Weakness GE 1.5 T Normal function

9526 Cranial Artifact: SVT9 Reveal Plus Left knee pain GE 1.5 T Normal function

9526 Knee Artifact: none10 Reveal Plus Left leg pain GE 1.5 T Normal function

9526 Lumbar spine Artifact: asystole, SVT11 Reveal Plus Suspected brain mass GE 1.5 T Normal function

9526 Cranial Artifact: SVT

MRI = magnetic resonance imaging; ILR = insertable loop recorder; T = Tesla; SVT = supraventricular tachycardia; CHB =complete heart block; GE = General Electric, Corp.

DISCUSSION

Despite the incorporation of sophisticated elec-tronic filters, implantable devices used for the treat-ment or monitoring of cardiac rhythms remain sus-ceptible to performance degradation and record-

Figure 1. Artifact mimicking narrow complex tachycardia. Marked signal attenuation mimicking complete heart block.Both occurred during cranial MR scanning.

ing artifacts when exposed to various types ofelectromagnetic interference,4,7–9 and in particu-lar MRI.4,10–14 Specifically, deCock et al. demon-strated during an in vitro evaluation of the effectsof MRI on a Reveal Plus, the patient activator couldnot effectively initiate a stored event during an

Artefactmimickingnarrowcomplextachycardia

GimbelJ etal,AnnalsofNoninvasiveElectrocardiology 2005

• EMIcancauseapower-on-reset

• Thedeviceprogrammingparametersgobacktothefactorydefault− Pacingcouldbeinhibited− Activationofantitachycardia therapy

• Devicebatterystatusandlongevityaffected

• Batterydrainmayalsooccur

EffectsofEMI

• ElectromagneticinterferencecanbereducedbyfilteringtheECGsignal

• InMRConditionaldevices,generatorsandcircuitryareshieldedtominimizetheeffectsofinterferences

• Betterprotectionofthepowersupply

PreventingEMI

Nazarian Setal,Circulation2008

Figure 5.Unfiltered and filtered intracardiac electrograms obtained via the canine model inside the MRIscanner during GRE imaging. The tracings in each column are simultaneous. The electrogramgain across each row is constant. RA indicates right atrium catheter electrogram; His, Hiscatheter electrogram; RV, right ventricular catheter electrogram; A, atrial signal; H, His signal;and V, ventricular signal.

Nazarian et al. Page 14

Circulation. Author manuscript; available in PMC 2010 February 23.

NIH

-PA

Author M

anuscriptN

IH-P

A A

uthor Manuscript

NIH

-PA

Author M

anuscript

Intracardiac ECGsignalsinacaninemodelinsideanMRscanner

• VeryshortRFpulsesareusedtodisturb/tipthenetspinmagnetizationwithinatissue

• RFpulsefrequencyismatchedtotheimagednuclei(1H)atagivenstaticmagneticfield− ~64MHzat1.5T− ~128MHzat3T

• Thebodywillabsorbsomeofthisenergy➜ Resistiveheating

• HeatinggeneratedbyasequenceismeasuredbytheSpecificAbsorptionRate (SAR)inWatt/kg

Radiofrequency(RF)Pulses

3/2/18

5

SARMODEforCIEDs

NormalMode(≤2W/kg)forallCIEDsMHRA,SafetyGuidelinesforMagneticResonanceImagingEquipmentinClinicalUse,March2015

• LeadscanactasantennaeandconcentrateRFenergy− Leadlengths(40– 60cm)comparabletoMRIRFwavelengths

• EffectstronglyassociatedtoratioofleadlengthtoRFwavelengthandpresenceofloopsandleadgeometry

• Highelectricalcurrentscanbeinduced➜ heating− Resistiveeffect

LeadsHeatingEffects

• Potentialriskofthermalinjurybyohmic lossinmyocardialtissuearoundthetip

• Generatedcurrentsmayleadto:− Myocardialstimulation− Temporaryorpermanentchangesinimpedanceandthresholds− Devicemalfunctionanddamage

• Abandonedandfracturedleads,brokenleadtipsandleadloopconfigurationsmayincreaseheatingeffects− Epicardial leadtipsnotcooledbybloodflow

LeadsHeatingEffects LeadsHeatingEffectsLANGMAN, ET AL.

Figure 1. For pacemaker leads 20–60 cm in length the RF-induced pacemaker lead tip heating isshown for pacemaker-attached and abandoned capped pacemaker leads. Pacemaker leads thatare abandoned and capped heat significantly more than pacemaker-attached leads for clinicallyrelevant lead lengths (40–60 cm).

systematically compared RF-induced heating ofpacemaker-attached leads to abandoned cappedleads using an in vitro phantom. We reportedthat abandoned capped pacing leads of clinicallyrelevant lengths (40–60 cm) exhibited more leadtip heating than pacemaker-attached leads. Therelevant heating results for pacemaker-attachedand abandoned capped pacemaker leads areshown in Figure 1 as a function of lead length.While the temperature results from in vitrotesting conditions cannot be directly compared toscenarios expected in vivo, our results show thepotential for substantially greater abandoned leadtip heating. Additional research using computa-tional modeling of in vitro and in vivo scenarios isneeded to augment our understanding of the riskof both pacemaker-attached and abandoned leadsacross the wide range of clinically relevant config-urations. This research will help in the develop-ment of objective guidelines on the risks of scan-ning patients with abandoned pacemaker leads.

While the implant site of an abandoned leadis no longer required to sense or deliver pacingpulses, unintentional thermal damage to cardiactissue is never desirable. One of the first indica-tions of thermal damage in normal functioningleads is a change in the pacing capture threshold,which cannot be measured in an abandoned lead.Therefore, any thermal damage that does occur

may only be detectable if further complicationsarise. Some comfort, however, is found in the factthat the majority of abandoned lead tips are likelylodged in scar tissue. Nonetheless, the clinicalrisks associated with abandoned lead tip heatingremains undefined.

The increased risk of abandoned leads is alsoa highly relevant topic due to the recent approvalof the Medtronic MR Conditional pacemaker(Medtronic Inc., Minneapolis, MN, USA) andlead by the U.S. Food and Drug Administration(FDA).25 The MR conditional distinction meansthat the device poses no known hazards in a speci-fied MRI environment. Under the terms of the FDAapproval, the B0 static field strength must be 1.5Tesla, the gradient magnetic field (dB/dt) slew ratemust be less then 200 Tesla/m/s, and SAR must beless then 2 W/kg. Additionally, the isocenter ofthe MRI system cannot be located between the C1and T12 vertebra during imaging, which may limitimage quality in the neck and chest. As part ofthe safety considerations, a chest x-ray should beused to check for the specific radio opaque markeron the MR conditional pacemaker and leads; thisx-ray can also be used to confirm that there are noabandoned pacemaker leads present. While thisnew generation of pacemakers is still under therestriction of MR Conditional scanning, they dooffer clinicians an assurance of safety far greater

1052 September 2011 PACE, Vol. 34

Langman Detal,PACE2011

• LimittheRFpowerusedduringimaging− ReduceSARlevel

Changesinleaddesign:

• Improvementstoleadinner- andouter-coilsstructure

• Leadtipcoatingwithpolarizationresistantmaterial

• Useofheat-dissipatingfilters

ReducingLeadHeatingEffects SummaryofPotentialRisksStaticMagneticfieldB0 Time-varyinggradients RFpulses

Force,TorquePatientdiscomfort,surroundingtissue

damage,devicemalfunction✓

VibrationPatientdiscomfort,devicemalfunction ✓

Inducedcurrents/voltagesInducedVT,arrhythmia,pacing

inhibition✓ ✓

HeatingTissuedamage,impedanceand

thresholdchanges,lossofsensingand/orpacingcapture

✓

DevicemalfunctionDevicereset,modechanges,lossof

therapy,patientshocks✓ ✓ ✓

3/2/18

6

• ScanningofMRConditional cardiacdevicesissafeifspecifiedconditionsarefollowed− Relativelystraightforward− CanbedoneatanyMRIgeneralimagingunit− NoreasontodenyapatientanMRIexamination

• StrongevidencefromregistriesandclinicalstudiesshowthatnonMRConditional CIEDscanbescannedsafelyifstrictclinicalandscanningprotocolsareobserved− Specializedcentres

Conclusion

• Therestillexistsaconfusionabouttheusedterminologyanddeterminingscanningconditions− Spatialstaticfieldgradientnotalwayseasytocheck− MRsafetylabelling

Conclusion

✗

MR?Scanning Conditions Unknown.

Could be MR unsafe.

?←NonMRConditional