ORIGINAL RESEARCH

MRI Interactions of a Fully Implantable Pressure Monitoring Device Ellyce F. Stehlin, BE,1* Daniel McCormick, PhD,1,2 Simon C. Malpas, PhD,1,2,3 Beau P. Pontr e, PhD,4,5 Peter A. Heppner, FRACS,6 and David M. Budgett, PhD1,2 Purpose: To investigate the potential patient risk and interactions between a prototype implantable pressure monitoring device and a 3T clinical magnetic resonance imaging (MRI) machine to guide device design towards MR Conditional safety approval. Materials and Methods: The pressure monitor device contained a catheter-mounted piezo-resistive pressure sensor, rechargeable battery, wireless communication system, and inductive pickup coil. Standard testing methods were used to guide experiments to investigate static field induced force and torque, radiofrequency (RF)-induced heating, image artifacts, and the MR’s effect on device function. The specific clinical application of intracranial pressure monitoring was considered. RF-induced heating experiments were supported by numerical modeling of the RF body coil, the device, and experimental phantom. Results: Sensing catheter lead length and configuration was an important component of the device design. A short 150 mm length catheter produced a heating effect of less than 2 C and a long 420 mm length catheter caused heating of 7.2 C. Static magnetic field interactions were below standard safety risk levels and the MR did not interfere with device function; however, artifacts have the potential to interfere with image quality. Conclusion: Investigation of MR interactions at the prototype stage provides useful implantable device design guidance and confidence that an implantable pressure monitor may be able to achieve MR Conditional safety approval. J. MAGN. RESON. IMAGING 2015;00:000–000.

A

n implantable long-term pressure monitor has many potential applications including the monitoring of intracranial pressure (ICP) of hydrocephalus patients, intraocular pressure of glaucoma patients, bladder pressure of neurogenic bladder dysfunction patients, intraarterial pressure of hypertension patients, and ventricular or aorta pressure of heart failure patients. A prototype device was constructed which contains significant components of a long-term implantable pressure monitor to investigate potential MRI interactions with the device, patient, and 3T MRI machine. The device operates off a 2F piezo-resistive based sensor catheter (manufactured by Millar, Houston, TX) with a drift of less than 2 mmHg over a 100-day period.1 This sensor has a sealed vacuum reference pressure, suitable for full implantation. Widely available on chip radio communication systems (eg, nRF24LE1 from Nordic, The Nether-

lands) provide high bandwidth (100 Hz) data for continuous pressure measurement when supplied with power on the order of 10 mW.2 These systems have a size of 10 3 10 3 2 mm, suitable for implantation.3 An inductive power transfer (IPT) pickup coil on the device allows power to be safely transferred from outside the body to the implant without electrical contact for on-demand use, or to recharge an internal energy source. These technologies provide the basis for the creation of an implantable monitoring device. Consideration of potential applications of such a device will focus on the hydrocephalus application due to the clear need for a chronic solution to provide insight into patient health and shunt or alternative treatment efficacy.4 The demand for magnetic resonance imaging (MRI) techniques in diagnostic medicine is on the rise due to its lack of ionizing radiation and ongoing improvements with

View this article online at wileyonlinelibrary.com. DOI: 10.1002/jmri.24909 Received Dec 1, 2014, Accepted for publication Mar 23, 2015. *Address reprint requests to: E.S., UniServices House, Room 107, 70 Symonds St., Auckland, New Zealand. E-mail: [email protected] From the 1Auckland Bioengineering Institute, University of Auckland, Auckland, New Zealand; 2Millar Inc., Auckland, New Zealand; 3Department of Physiology, University of Auckland, Auckland, New Zealand; 4Centre for Advanced MRI, University of Auckland, Auckland, New Zealand; 5Department of Anatomy with Radiology, Faculty of Medical and Health Sciences, University of Auckland, Auckland, New Zealand; and 6Department of Neurosurgery, Auckland City Hospital, Auckland, New Zealand

C 2015 Wiley Periodicals, Inc. V 1

Journal of Magnetic Resonance Imaging

soft-tissue contrast and resolution. It is important that a new clinical device does not exclude potential patients from these existing diagnostic technologies, particularly for the hydrocephalus population, where patient management is reliant on medical imaging of ventricle size. The static magnetic, radiofrequency (RF), and gradient switching fields in MRI can cause interactions with conductive material in implantable devices, inducing forces and torques, tissue heating, image artifacts, and potential electromagnetic interference.5–8 The purpose of this investigation was to determine the potential for patient risk through MRI interactions with a prototype medical device in a 3T clinical MRI. The prototype contained key technologies for a chronic pressuremonitoring role with flexibility for design variation of critical features, such as the sensing catheter length. Combinations of numerical and experimental test procedures were used, guided by safety standard methods.

MATERIALS AND METHODS To investigate MRI interactions, a prototype device was assembled that contained the critical elements of a long-term implantable pressure monitor: sensor, catheter, and electronic module (sensor interfacing, radio, power transfer). This device was used in numerical models and experimental scans and tests in the MRI machine. Test methods investigated the induced force, torque, RF heating, artifact, and interference with device function. Methods were developed with guidance from the American Society for Testing and Materials (ASTM standards F2502, F2213, F2182, F2119 9–12 ), the International Organization for Standardization (ISO) technical specification on active implants in MRI (ISO/TS 10974 13 ), and from a defined qualitative assessment of torque, developed and used by Shellock et al.6,14,15

Implantable Pressure Monitor The newly developed implantable pressure sensing device, while still in its prototype stage, consists of significant components of a clinical device that have the potential to cause risk when in the MR environment. The 2F piezo-resistive sensor runs wires through a catheter to the electronic unit. The electronic unit consists of a rechargeable lithium ion battery, an inductive power coil for transcutaneous energy transfer, a 2.4 GHz wireless data antenna link, and associated signal conditioning and amplification circuitry. For the ICP application, the sensor would sit inside the skull 20 mm into the brain parenchyma, with the electronics unit outside the skull underneath the skin. A rechargeable battery is included in the prototype to allow for potential periods of prolonged monitoring, overnight or during daily activities. At its current development stage, the device is encapsulated in a silicone jacket to allow for RF transparency; however, more rigid, standard medical casings would be implemented for a final device such as titanium, glass, or ceramics, which are widely used in active implants such as pacemakers, neurostimulators, and cochlear implants. Early-stage numerical modeling results indicated the sensing catheter length would be a key determinant of resulting RF2

induced heating, which is consistent with published work on leads for pacemakers, cardio defibrillators, and neurostimulators.16–18 Therefore, two versions of the device were investigated for RFinduced heating, identical other than in catheter length: 1) A short catheter (SC) device, with the sensor 150 mm from the electronic unit, and 2) a long catheter (LC) device, with the sensor 420 mm from the electronic unit. These configurations represent potential implantation locations should the device be fully implanted in the head of the patient for the ICP application, if the electronic unit was required to be positioned further distal from the head, or if an alternative long-term pressure application was addressed.

Magnetic Field Interactions The device’s translational attraction and torque was assessed against standardized test methods in a 3T clinical MRI environment (Magnetom Skyra 3T, Siemens, Munich, Germany). TRANSLATIONAL ATTRACTION. The translational forces owing to the static magnetic field are proportional to the implant’s volume and geometry of ferromagnetic material, overall geometry and mass, and the strength and spatial gradients of the applied static field. For investigating safety of implanted devices, the deflection angle test was used as recommended in ASTM F2052.10 The test specifies that the translational force experienced by a device that is deflected by less than 45 from vertical when placed in the region of the highest spatial magnetic field gradient is less than the attractive force owing to gravity. Therefore, any risk associated with force induced by the 3T magnetic field would be no worse than that experienced by normal daily activity. The test fixture was a rigid structure with a protractor at the top attachment point with 1-degree graduated markings. A string weighing less than 1% of device weight was used to suspend the prototype device from the 0-degree (vertical) marking on the protractor.10 The fixture and device were placed at an off-axis position at the edge of the magnet bore where it was empirically determined to cause maximum deflection under a static magnetic gradient of 720 gauss/cm, according to manufacturer gauss line plots.19 With the scanner inactive and just the static magnetic force of the magnetic bore influencing the device, the maximum deflection angle from the vertical direction was measured for each version of the device three times and the mean deflection calculated. TORQUE. The device may experience a static field-induced torque due to magnetization of ferromagnetic materials which, if large enough, will attempt to align the long axis of the dominant ferromagnetic component with the field.5,6 A qualitative assessment of torque, as developed and extensively used by Shellock et al,6,14,15 was used to assess the device safety. The device, in both the SC and LC configurations, was positioned on a flat plastic sheet with a millimeter grid, oriented at a 45 angle to the static field, and placed in the center of the bore where the field is uniform and at its largest, ensuring that the worst-case effect is investigated. The device was visually inspected to determine whether any rotation owing to the presence of the magnetic field occurred. The device’s orientation with respect to the magnetic field was increased in 45 increments to encompass 360 of rotation and repeated for each three primary device axis to fit in the bore for the short and long device. A qualitative scale was used to assess the torque where

Volume 00, No. 00

Stehlin et al.: MRI Interactions of Pressure Implant

0 5 no torque; 11 5 low torque, the device shifted slightly but did not align to the magnetic field; 12 5 moderate torque, the device gradually aligned to the magnetic field; 13 5 strong torque, the device showed rapid and forceful alignment to the magnetic field; 14 5 very strong torque, the device showed very rapid and very forceful alignment to the magnetic field, as defined by Shellock.6

TABLE 1. Parameters Used to Expose the ICP Device to MRI RF Energy

MRI parameter TR

5760 msec

RF-Induced Heating

TE

124 msec

For the long-term pressure device, heating was assessed for both the short and long configuration at 128 MHz/3T. Numerical simulations were used to investigate electric field (E-field) and specific absorption rate (SAR) distributions in the phantom.13,20 A preliminary numerical investigation varying the length of the catheter tip revealed a temperature peak for the LC configuration and a general trend of increasing temperature with catheter length (unpublished work). The SC configuration represents an overestimation of the length required for the ICP application where the sensor is implanted 20 mm in the brain parenchyma. Results from these simulations determined the subsequent experimental test protocols.

Turbo factor

23

Plane

Transversal

Flip angle

180

Bandwidth

130 Hz/Px

Field of view

400 x 400 mm

Matrix

512 x 512

Section thickness

10 mm

Total slices

40

Scan time

15:00

SEMCAD X v14.8.5, a commercially available software package based on the finite-difference timedomain method, was used for electromagnetic (EM) simulations of the prototype device in the MR environment.21 The model required creating the 128 MHz RF transmit body coil, simulating the ASTM head-torso phantom and the implantable device. Analysis of the electromagnetic results revealed localized E-fields, which indicate potential high temperature concentrations. The body coil was modeled as it exposes the full phantom to RF energy and has the potential to induce high fields and/or heating with a whole body average SAR limit of 4 W/kg in first level operating mode.9,13 Following methods outlined by Liu et al,22 a nonphysical coil was modeled, as previously demonstrated to replicate physical E-field distribution while avoiding long computational times associated with modeling specific geometries, tuning capacitors, and excitation modes. For the nonphysical model, current sources were enforced along the individual coil rungs joined by tuning capacitors in the end rings. The rung currents are forced with progressive phase delay to ensure that a circularly polarized electric field is generated. Using methods outlined by SEMCAD and previously used by Liu et al, the RF coil was tuned to 128 MHz when loaded with the ASTM phantom gel.21,22 The iterative process uses a series of broadband simulations with one source active to adjust the ring capacitors until resonance occurs at the desired 128 MHz for a 3T MR system. The empty coil model was verified by inspection of the centro-symmetric E-field and central uniform B-field. SAR and E-field distributions inside the ASTM phantom are well established and further verified the model before including the short and long devices.9,13,23,24 In order to overcome the complications of a large overall body coil simulation environment with the sub-mm detailed implant model, a Huygen’s box source driven, two-stage EM simulation method was used. The Huygen’s box method uses a primary simulation with a relatively coarse grid, and associated large time step, to establish the incident field distributed about a smaller volume defined by the Huygen’s box. The resulting E- and H-fields distributed about the Huygen’s box become the source of a secondary simulation within the small defined volume, containing a fine grid and requiring a small time step. This NUMERICAL MODEL.

Month 2015

method is ideal for the implant-phantom-MR RF field computational problem, as the primary simulation can be used to solve for the EM fields in the phantom about the region of interest without the implant present, before adding the detailed implant model to the secondary refined Huygen’s box simulation. EM simulation results determined the conditions of the physical scans to be carried out and were used to drive SEMCAD-based thermal simulations. The SEMCAD thermal solver utilized Penne’s Bio heat equation for which the explicit solution was applied. Material property approximations were used for the highly conductive components of the implant to overcome prohibitively long simulation times associated with the complex model requiring a high resolution grid and high conductivity enforcing a small time step. Thermal simulation results predicted peak temperature distribution for the simulated device configurations providing guidance for temperature sensor placement. Following experimental results in the physical scanner, the simulations were normalized to the measured SAR. MR SYSTEM CONDITIONS. Physical MR experiments were carried out in a Siemens SKYRA 3T using the system’s body coil for both RF transmit and receive. The body coil exposes the largest area to RF power for high SAR exposure testing. A turbo spin echo sequence was used to generate a large level of RF energy deposition. The scan parameters are listed in Table 1. Landmark positioning was constant for all scans, at the centerline of the phantom torso and section locations were set to encompass the length of the device.

Opsens MPK4 fiber optic temperature sensors and TempSens signal conditioner (TempSens, Opsens, Quebec, Canada) were used to measure changes in gel temperature while in the scanner. The 0.51 mm diameter fiber optic temperature sensors were calibrated against a high precision fluke digital thermometer (Model 1504, Fluke, Everett, WA). Based on results from the numerical simulations, three temperature

TEMPERATURE MEASUREMENT.

3

Journal of Magnetic Resonance Imaging

FIGURE 1: The implant locations tested for the prototype ICP device showing device and catheter layout (L-R: LC device, SC device, coiled LC device) with associated fiber optic temperature sensor positioning (1).

probes were positioned around the device at key areas of interest, with one temperature probe acting as a reference on the opposite side of the ASTM phantom. WHOLE-BODY SAR. The calorimetry method outlined in ASTM F2182 was used to quantify the RF energy deposited in the phantom.9,20,25 The ASTM torso-head phantom container,9 constructed from poly-methyl methacrylate (PMMA), was lined on all sides with 25-mm thick polystyrene insulating sheet, ensuring conductance of less than 0.029 W/m
K to the external environment. The phantom was filled to a 9 cm depth with 2.5 g/L NaCl dissolved in deionized water, with a measured conductivity of 0.47 S/m. The saline phantom was allowed to equilibrate in the scan room for 2 hours prior to scan time. With the fiber optic temperature probes positioned in the center of the phantom, the phantom was positioned and the scan initiated. Upon completion, the phantom was quickly removed from the MR bore, saline stirred without removing the insulation layers, and the peak temperature change recorded. Temperature measurement was continuous throughout the scan allowing for the peak temperature to be recorded once stirring diffused the temperature concentrations throughout the phantom. Whole-body average SAR was calculated from this temperature change, SAR5C DT Dt , where c is the specific heat capacity of the phantom saline (4150 J/kg C), DT the change in temperature ( C), and Dt the change in time (seconds).9

The ASTM head torso phantom was filled with a gelled saline solution (1.32 g/L NaCl, 10 g/L polyacrylic acid) which had a verified conductivity of 0.47 S/m. The gel ensures that no temperature convection currents can be established, providing a likely overestimate of the expected local heating for the implanted condition due to lack of perfusion in the phantom. The gelled saline phantom was allowed to equilibrate in the scan room for 2 hours prior to scan time. The device was positioned in the phantom as determined by numerical model results; three experiments based on catheter configurations were investigated: 1) LC device phantom torso, 2) SC device phantom torso, 3) LC device coiled in torso (Fig. 1). Implantation location 1 positions the LC device in the area of highest SAR/heating as demonstrated by numerical models. The SC device, which represents the configuration expected to be most

MRI HEATING PROTOCOL.

4

commonly used for the ICP application, was also tested in this location (Implant location 2). Experiment 3 replicates conditions from Codman on ensuring safety for patients undergoing MRI scans with an acute ICP sensing catheter implanted. The long lead of the acute ICP catheter is specified to be coiled in a circle 5 cm in diameter after leaving a straight segment of 8 cm from the sensor tip.26 This coiled LC setup was tested with 8 cm from sensor tip to coil, 2.25 turns, and 3 cm from coil to electronic unit. For all configurations a fiber optic temperature sensor was located at the catheter lead tip, as numerical models predicted this point to be where maximum temperature rise would occur. A temperature sensor was also positioned at the opposite, telemetry unit end of the device, and at the junction between device and catheter or a sharp bend in the catheter. For all heating experiments, the landmark positioning was consistent with the SAR calorimetry method, in the center of the phantom torso. The fiber optic probes were positioned using PMMA clamps and guided with vitrotubes 1.8 mm glass round capillaries (Vitrotubes, VitroCom, Mountain Lakes, NJ) to allow for accurate and repeatable positioning through the gelled saline. Baseline temperatures were recorded for 2 minutes, before recording throughout the scan, and for 2 minutes postscan. The highest temperature change seen was recorded. Temperature sensor positioning was inspected postscan to ensure no movement had occurred. Each experiment was repeated with the implant removed and the temperature sensors still in position to record background temperature changes for the given scan setup.

Artifacts Device-induced artifacts were evaluated for the prototype device and the pressure sensing catheter alone following methods from ASTM F2119.12 The device and catheter were positioned in separate rectangular PMMA containers containing a PMMA grid. Nylon cylinders were included around the device/catheter to act as reference objects in the scans. The containers were filled with a 1.5 g/L copper sulfate solution. The transmit-receive head coil was used to investigate the possible occurrence of artifacts in the case where the ICP device has the potential to interfere with brain imaging. Although extensive variation and testing of scan parameters to define the absolute worst-case scan sequence was not carried out, the sequences used to investigate this effect (below) have been previously used in the assessment of many similar devices to demonstrate maximum artifact while minimizing geometric distortion (14,15,27–29). 1. T1-weighted spin echo sequence: TR 500 msec, TE 20 msec, 10 mm slice thickness, 280 mm field of view, 256 3 256 matrix size, 2 averages, 250 Hz/Px bandwidth. 2. Gradient echo pulse sequence: TR 100 msec, TE 15 msec, 30 flip angle, 10 mm slice thickness, 280 mm field of view, 256 3 256 matrix size, 2 averages, 200 Hz/Px bandwidth. Imaging planes were orientated perpendicular and parallel to the short and long axis of the device. Multiple section locations were tested to determine those which showed the largest artifacts. The area of the artifact was determined by analysis of the reference objects and use of Siemens scanner image processing software (Syngo FastView, Siemens, Munich, Germany). Since it is Volume 00, No. 00

Stehlin et al.: MRI Interactions of Pressure Implant

TABLE 2. Eight Scan Sequences Used to Expose the ICP Device to a Range of 3T MRI Conditions

Sequence

T1-SE

T2-SE

T1-FSE

T2-FSE

GRE, 3D FGRE, 3D GRE, MTC EPI

TR (msec)

700

3000

700

5010

20

9

628

4460

TE (msec)

10

100

12

105

5

3.67

10

364

Flip angle

90

90

180

180

25

25

25

Field of view (mm)

300

300

300

300

300

300

300

Matrix size

256 3 256 256 3 256 256 3 256 256 3 256 256 3 256 256 3 256 256 3 256

300 256 3 256

Slice thickness (mm) 10

10

10

10

3

3

10

10

Distance factor

10%

10%

10%

10%

20%

20%

10%

10%

Imaging plane

Transverse Transverse Transverse Transverse Volume

Volume

Transverse

Transverse

Exposure time

1:00

1:00

1:00

1:00

1:00

1:00

1:00

1:00

T1-SE, T1-weighted spin echo; T2-SE, T2-weighted spin echo; T1-FSE, T1-weighted fast spin echo; T2-FSE, T2-weighted fast spin echo; GRE, gradient echo; FGRE, fast gradient echo; MTC, magnetization transfer contrast; EPI, echo planar imaging; GRE, gradient echo; SE, spin echo; SAR, specific absorption rate.

acknowledged that there is potential for a more severe artifact to be demonstrated through further manipulation of the MRI parameters, the sequences used in this investigation were used to allow for comparison against other implants that have reported artifact using similar conditions (14,15,27–29).

MRI Effect on Device Function The device was finally subjected to a wide variety of scan parameters to determine if the MR machine could interfere with the device in any way. A series of eight scans were selected based on previously reported results on effects of MRI on function.27 The device was also exposed to the static field multiple times in each orthogonal axis while passing in and out of the bore 10 times. Testing of the basic function and calibration of the device through exposure to a known pressure load was carried out pre- and postexposure to the range of MR Conditions outlined in Table 2.

RESULTS Magnetic Field Interactions For both the SC and LC device configurations the maximum deflection angle was 15 and the torque was qualitatively assessed as 0, indicating that the device experienced no torque while in the magnetic field environment. MR-Induced Heating NUMERICAL MODEL. Results from the numerical model support the standard asymmetrical SAR distribution for the ASTM phantom, where one side of the phantom absorbs more RF energy. The asymmetric SAR distribution in the phantom occurs due to the rotational field of the RF coil. A longer implantation lead was associated with a higher overall temperature. For the long catheter device, peak temperatures were seen at the tip of the 420 mm sensing catheter, reaching a maximum temperature of 7.8 C after 900 seconds of Month 2015

exposure. Model results showed the sensing lead as the dominant feature of the device’s RF heating. For the SC and coiled LC devices, peak E-fields and temperatures were of a much lower magnitude than the LC device, with minor localized hot spots occurring at the device tip. The model predicted less than 2 C for the SC and coiled LC configurations. The SEMCAD model and temperature distribution for the SC and LC configurations are shown in Figure 2. WHOLE-BODY SAR. The MR system reported whole body (WB) SAR for the phantom was 3.5 W/kg, where the calorimetry determined value was 3.2 W/kg. MR HEATING. The highest temperature change seen was for the LC torso experiment, consistent with the predictions from the modeling results. The temperature change at the sensor tip was 7.2 C, where the SC device caused a temperature change of 1.7 C, after the 15-minute scan. Reference change in temperature measurements from baseline ranged from 0.7–0.8 C, indicating that for all three tested conditions the presence of the device caused an increase in phantom heating (Table 3).

Artifacts The long-term implantable ICP device caused large artifacts relative to its size (Fig. 3). The gradient echo pulse sequence caused the most severe artifacts, with a circular signal void in the transverse orientation about the device’s telemetry unit with a diameter of over 100 mm and a loss of signal intensity and geometric distortions seen extending over 110 mm from the unit. The pressure sensing catheter and tip caused minor distortions of a maximum 5 mm. Figure 3 shows the largest artifact slice observed for both the gradient and spin echo sequences. 5

Journal of Magnetic Resonance Imaging

FIGURE 2: SEMCAD model setup (left), thermal simulation temperature distribution at 900 seconds for the SC (150 mm catheter, center) and the LC (420 mm catheter, right) devices. Distribution is displayed as a log scale (dB) normalized for the maximum temperature for each device (1.8 C and 7.8 C for SC and LC, respectively).

MRI Effect on Device Function No difference in device operation or sensor characterization was seen between pre- and postfunction exposure scans. The sensors response to a step increase in pressure remained within 61 mmHg.

DISCUSSION An implantable pressure monitoring device was tested in a prototype stage to investigate potential MR risks. The implant consisted of all components determined to be of potential MR safety concern, or critical to the device’s function—the sensing catheter, the implanted battery, wireless communication components, and inductive power transfer coil. The low ferromagnetic material in the device resulted in low angular deflection well under the ASTM safety standard limit and no observable torque. The results are consistent with similar devices tested in the 1.5–3T magnet range and indicate that the device will not pose any safety concerns owing to the presence of static magnetic fields of 3T or lower.5,15,27 The MRI system reported a slightly higher SAR measurement than determined using calorimetry. This discrepancy has been previously reported, with the scanner proprietary methods overestimating the SAR when compared to calorimetry results.9,25 The calorimetry calculated

SAR values are consistent with the MRI requesting to operate in First Level Mode (required for RF power deposition of greater than 2 W/kg) in order to run the specified sequence. The RF heating results agreed with numerical modeling predictions for the three tested implant locations with the relative temperature distributions across the four fiber optic temperature sensing points matched the heating distribution and intensity predicted by simulation. The straight LC device caused localized heating at the tip of the pressure sensing catheter 10 times that which would occur without the device present. This large temperature rise of 7.2 C is not as severe as those seen with exposed lead tips, such as deep brain stimulation electrodes, which can induce heating of over 25 C, due to the elongated metallic structure being fully encapsulated at the tip without an exposed electrode.30,31 The localized heating at the tip of the elongated catheter is not surprising, as it has been previously demonstrated that elongated metallic implants concentrate the RF currents induced in the body during MRI. This increased current density and associated SAR causes greatest heating at the implant ends where energy is deposited.32 The LC configuration is important to consider for potential applications outside the hydrocephalus ICP solution and provides information for working towards standardized assessment.9,13 The heating observed with the coiled LC device

TABLE 3. Maximum Temperature Change for the Four Device Configurations

Experiment

6

Model predicted heating ( C)

Maximum heating measured ( C)

Device present

Device present

Empty phantom

Long device torso

7.8

7.2

0.7

Short device torso

1.8

1.7

0.8

Coiled long device torso

1.2

0.8

0.8

Volume 00, No. 00

Stehlin et al.: MRI Interactions of Pressure Implant

FIGURE 3: Artifact test setup (left) with the SC device in the phantom and worst artifact slice observed for the (left to right) gradient echo coronal plane, gradient echo transverse plane, spin echo coronal plane, spin echo transverse plane. Artifacts are large relative to device size with a signal void of over 100 mm in the transverse orientation and distortion of signal extending over 110 mm in the coronal orientation.

was within 1 C of that occurring without the device present. The effect of coiling the long ICP catheter reduces heating by altering the effective length of the elongated catheter. This elimination may be due to the introduction of an RF choke, as investigated by Newcombe et al 33 for the acute Codman MicroSensor Transducer. Tanaka et al 26 present the devastating effect of not following the conditions of safety for MRI with an acute ICP catheter where a patient suffered thermal brain injury at the sensor tip when undergoing MRI with the sensor lead left straight. Should a long chronic ICP device lead be required, further investigation into the coiled configuration and limitations on the catheter length for the device would have to be considered. With the short, 150 mm lead length device implanted in the phantom torso in the area of high SAR concentration, a peak temperature was seen less than 1 C above the heating for the empty phantom. The dependence of RF-induced heating on device configuration and dimensions is well documented and demonstrates the need for thorough investigation into safety conditions for implant devices in the MR environment.18,34,35 These experiments assess the heating in the ASTM phantom exposed to high (>3 W/kg) whole-body averaged SAR producing scans. Based on these results, it is anticipated that the final form of the implantable pressure sensing device will cause low heating risks provided elongated catheter lengths are avoided. Large artifacts were produced by the prototype relative to the device size. In particular, large signal voids were observed when using a gradient echo pulse sequence. The sequence and parameters chosen in this instance allow for comparison against multiple published previously investigated devices. The artifact is consistent with similar active implants at 3T 29 and could potentially inhibit visualization of key areas of interest in the brain in the ICP application. However, these results demonstrate images produced under scans designed to maximize artifacts and appropriate paramMonth 2015

eter optimization and pulse sequence selection may help in minimizing the impact on image quality where metallic implants are present.36,37 Ongoing exposure to the 3T static field and multiple scans did not result in any interference with the prototype device’s function. These results were as expected, considering the device’s small size and lack of exposed leads or magnetically sensitive components, limiting the potential for MRIdevice interference through the gradient, RF, and static fields. However, it is recommended that basic functional characteristics of any medical device are confirmed pre- and post-MR scanning. This investigation was limited by factors considering the prototype nature of the device and preliminary device– MRI interaction testing. The device encapsulation method used a silicone polymer, not suitable for long-term implantation due to its inability to provide an impermeable barrier.38 Firmly established implantable active device encapsulation materials and methods such as titanium cans, ceramic casings, and associated feedthroughs are expected to be implemented with the final device. It is anticipated the varying properties such as dielectric permittivity and conductivity will subsequently influence the RF heating through energy absorbed and dispersed. However, the wide implementation of such encapsulation techniques on a variety of similar MR Conditional active implants give confidence that the final encapsulation will not inherently cause MR safety concerns. The presence of image artifacts and the effect on image quality may be more dependent on the encapsulation method used.15 The heating analysis for the device focused on the RF field-induced heating, although there is also potential for the gradient switching fields of the scanner to induce heating, vibration, and nerve stimulation. Gradient field-induced effects are of particular concern for devices containing exposed electrode leads or a large surface area of highly conductive material and were not specifically 7

Journal of Magnetic Resonance Imaging

investigated for this device.39 The lack of observed vibrations and the ability of the numerical RF heating model to explain the observed heating in the scanner support the assumption that the gradient field’s would cause little interference. However, consideration of the gradient fields will be required for the final device to achieve MR Conditional approval. The heating analysis was limited to numerical and experimental testing of the prototype in the homogeneous ASTM phantom exposed to large SAR generated by the RF body coil. For the final device design, anatomically heterogeneous body models such as the SPEAG/IS’TS Virtual Family can be used to refine the numerical approach.40 This would allow specific, anatomically equivalent implant positions to be investigated and RF coils specific to the target application. For the ICP application, it is expected heating in this refined model would be less than that demonstrated in this study with consideration of tissue-specific perfusion and the stringent head-specific exposure restrictions.13 It is expected that the heating demonstrated in this study is an overestimation of that which would occur in the implanted clinical implementation. In conclusion, MRI interactions of the prototype implantable pressure sensing device demonstrate low safety risk for the patient under expected clinical conditions and device configurations for the intracranial pressure application. The results demonstrate the importance of investigating MR heating interactions during the development stage for a new medical device to guide design parameters towards minimizing heating risk during MRI scans. The device implantation location and prospective MR scan parameters will have to be considered to avoid adverse effects on image quality resulting from the presence of artifacts. It is expected an implantable pressure monitoring device consisting of a sensing catheter lead, rechargeable battery, communication antenna, and IPT coil could qualify as MR Conditional within 3T scanners in accordance with current safety standards.

Acknowledgment Contract grant sponsor: New Zealand MBIE Smart Ideas project “Saviour Battery.”

REFERENCES 1.

Stehlin E, Malpas SC, Budgett DM, et al. Chronic measurement of left ventricular pressure in freely moving rats. J Appl Physiol 2013;115: 1672–1682.

2.

Nordic Semiconductor, nRF24LE1 Product Specification v1.6. 2010.

3.

Russell DM, Garry EM, Taberner AJ, et al. A fully implantable telemetry system for the chronic monitoring of brain tissue oxygen in freely moving rats. J Neurosci Methods 2012;204:242–248.

8

4.

Czosnyka M, Pickard JD. Monitoring and interpretation of intracranial pressure, J Neurol Neurosurg Psychiatry 2004;75:813–821.

5.

Nyenhuis JA, Sung-Min P, Kamondetdacha R, et al. MRI and implanted medical devices: basic interactions with an emphasis on heating. Device Mater Reliab IEEE Trans 2005;5:467–480.

6.

Shellock FG. Biomedical implants and devices: assessment of magnetic field interactions with a 3.0-Tesla MR system. J Magn Reson Imaging 2002;16:721–732.

7.

Dempsey MF, Condon B, Hadley DM. Investigation of the factors responsible for burns during MRI. J Magn Reson Imaging 2001;13: 627–631.

8.

Smith CD, Kildishev AV, Nyenhuis JA, et al. Interactions of magnetic resonance imaging radio frequency magnetic fields with elongated medical implants. J Appl Phys 2000;87:6188.

9.

American Society for Testing and Materials (ASTM) International, ASTM Standard F2182: test method for measurement of radio frequency induced heating on or near passive implants during magnetic resonance imaging. West Conshohocken, PA, 2011.

10.

American Society for Testing and Materials (ASTM) International, ASTM Standard F2052: test method for measurement of magnetically induced displacement force on medical devices in the magnetic resonance environment. West Conschohocken, PA, 2006.

11.

American Society for Testing and Materials (ASTM) International, ASTM Standard F2213: test method for measurement of magnetically induced torque on medical devices in the magnetic resonance environment. West Conshohocken, PA, 2006.

12.

American Society for Testing and Materials (ASTM) International, ASTM Standard F2119: test method for evaluation of mr image artifacts from passive implants. West Conshohocken, PA, 2007.

13.

ISO/IEC, ISO/TS 10974: Assessment of the safety of magnetic resonance imaging for patients with an active implantable medical device. 2012.

14.

Shellock FG, Valencerina S. In vitro evaluation of MR imaging issues at 3T for aneurysm clips made from MP35N: findings and information applied to 155 additional aneurysm clips. Am J Neuroradiol 2009;31: 615–619.

15.

Shellock FG, Bedwinek A, Oliver-Allen M, Wilson SF. Assessment of MRI Issues for a 3-T ’immune’ programmable CSF shunt valve. Am J Roentgenol 2011;197:202–207.

16.

Armenean C, Perrin E, Armenean M, et al. RF-induced temperature elevation along metallic wires in clinical magnetic resonance imaging: influence of diameter and length. Magn Reson Med 2004;52:1200– 1206.

17.

Mattei E, Calcagnini G, Censi F, et al. Role of the lead structure in MRI-induced heating: In vitro measurements on 30 commercial pacemaker/defibrillator leads. Magn Reson Med 2012;67:925–935.

18.

Gray RW, Bibens WT, Shellock FG. Simple design changes to wires to substantially reduce MRI-induced heating at 1.5 T: implications for implanted leads. Magn Reson Imaging 2005;23:887–891.

19.

Siemens AG, System Owner Manual: MAGNETOM Skyra. p 5–11, 2012.

20.

Neufeld E, K€ uhn S, Szekely G, Kuster N. Measurement, simulation and uncertainty assessment of implant heating during MRI. Phys Med Biol 2009;54:4151–4169.

21.

.Schmid & Partner Engineering AG, SEMCAD X Reference Guide. Zurich, Switzerland.

22.

Liu Y, Chen J, Shellock FG, Kainz W. Computational and experimental studies of an orthopedic implant: MRI-related heating at 1.5-T/64MHz and 3-T/128-MHz. J Magn Reson Imaging 2013;37:491–497.

23.

Amjad A, Kamondetdacha R, Kildishev AV, et al. Power deposition inside a phantom for testing of MRI heating. In: Magnetics Conference, 2005. INTERMAG Asia 2005. Digests IEEE Int 2005;1153–1154.

24.

Nordbeck P, Fidler F, Weiss I, et al. Spatial distribution of RF-induced E-fields and implant heating in MRI. Magn Reson Med 2008;60: 312–319.

Volume 00, No. 00

Stehlin et al.: MRI Interactions of Pressure Implant 25.

Baker KB, Tkach JA, Nyenhuis JA, et al. Evaluation of specific absorption rate as a dosimeter of MRI-related implant heating. J Magn Reson Imaging 2004;20:315–320.

26.

33.

Tanaka R, Yumoto T, Shiba N, et al. Overheated and melted intracranial pressure transducer as cause of thermal brain injury during magnetic resonance imaging. J Neurosurg 2012;117:1100–1109.

Newcombe VF, Hawkes RC, Harding SG, et al. Potential heating caused by intraparenchymal intracranial pressure transducers in a 3Tesla magnetic resonance imaging system using a body radiofrequency resonator: assessment of the Codman MicroSensor Transducer. J Neurosurg 2008;109:159–164.

34.

27.

Shellock FG, Knebel J, Prat AD. Evaluation of MRI issues for a new neurological implant, the sensor reservoir. Magn Reson Imaging 2013; 31:1245–1250.

Nordbeck P, Weiss I, Ehses P, et al. Measuring RF-induced currents inside implants: impact of device configuration on MRI safety of cardiac pacemaker leads. Magn Reson Med 2009;61:570–578.

35.

28.

Weiland JD, Faraji B, Greenberg RJ, et al. Assessment of MRI issues for the Argus II retinal prosthesis. Magn Reson Imaging 2012;30:382– 389.

Cabot E, Lloyd T, Christ A, et al. Evaluation of the RF heating of a generic deep brain stimulator exposed in 1.5 T magnetic resonance scanners. Bioelectromagnetics 2013;34:104–113.

36.

29.

Shellock FG, Crivelli R, Venugopalan R. Programmable infusion pump and catheter: evaluation using 3-Tesla magnetic resonance imaging. Neuromodulation 2008;11:163–170.

Dietrich O, Reiser MF, Schoenberg SO. Artifacts in 3-T MRI: physical background and reduction strategies. Eur J Radiol 2008;65: 29–35.

37.

30.

Langman DA, Goldberg IB, Judy J, et al. The dependence of radiofrequency induced pacemaker lead tip heating on the electrical conductivity of the medium at the lead tip. Magn Reson Med 2012;68: 606–613.

Viano AM, Gronemeyer SA, Haliloglu M, Hoffer FA. Improved MR imaging for patients with metallic implants. Magn Reson Imaging 2000;18:287–295.

38.

Zhou D, Greenbaum E. Implantable neural prostheses 2: techniques and engineering approaches. Berlin: Springer Science & Business Media, 2010.

39.

Graf H, Lauer UA, Schick F. Eddy-current induction in extended metallic parts as a source of considerable torsional moment. J Magn Reson Imaging 2006;23:585–590.

40.

Christ A, Kainz W, Hahn EG, et al. The virtual family—development of surface-based anatomical models of two adults and two children for dosimetric simulations. Phys Med Biol 2010;55:N23–38.

31.

32.

Rezai AR, Finelli D, Nyenhuis JA, et al. Neurostimulation systems for deep brain stimulation: in vitro evaluation of magnetic resonance imaging-related heating at 1.5 Tesla. J Magn Reson Imaging 2002;15: 241–250. Smith CD, Kildishev AV, Nyenhuis JA, et al. Interactions of magnetic resonance imaging radio frequency magnetic fields with elongated medical implants. J Appl Phys 2000;87:6188.

Month 2015

9