Effect of surgical modification of deep brain stimulation lead trajectories on radiofrequency heating during MRI at 3T: from phantom experiments to clinical implementation

Abstract

OBJECTIVE

Radiofrequency (RF) tissue heating around deep brain stimulation (DBS) leads is a well-known safety risk during MRI, resulting in strict imaging guidelines and limited allowable protocols. The implanted lead’s trajectory and orientation with respect to the MRI electric fields contribute to variations in the magnitude of RF heating across patients. Currently, there are no surgical requirements for implanting the extracranial portion of the DBS lead, resulting in substantial variations in clinical lead trajectories and consequently RF heating. Recent studies have shown that incorporating concentric loops in the extracranial lead trajectory can reduce RF heating. However, optimal positioning of the loops and the quantitative benefit of trajectory modification in terms of added safety margins during MRI remain unknown. In this study, the authors systematically evaluated the characteristics of DBS lead trajectories that minimize RF heating during 3T MRI to develop the best surgical practices for safe access to postoperative MRI, and they present the first surgical implementation of these modified trajectories.

METHODS

The authors performed experiments to assess the maximum temperature increase of 244 distinct lead trajectories. They investigated the effect of the position, number, and size of the concentric loops on the skull. Experiments were performed in an anthropomorphic phantom implanted with a commercial DBS system, and RF exposure was generated by applying a high specific absorption rate sequence (B_1+rms = 2.7 μT). The authors conducted test-retest experiments to assess the reliability of measurements. Additionally, they evaluated the effect of imaging landmarks and perturbations to the DBS device configuration on the efficacy of low-heating trajectories. Finally, two neurosurgeons implanted the recommended modified trajectories in patients, and the authors characterized their RF heating in comparison with unmodified trajectories.

RESULTS

The maximum temperature increase ranged from 0.09°C to 7.34°C. The authors found that increasing the number of loops and positioning them closer to the surgical burr hole, particularly for the contralateral lead, substantially reduced RF heating. These trajectory modifications were easily incorporated during the surgical procedure and resulted in a threefold reduction in RF heating.

CONCLUSIONS

Surgically modifying the extracranial portion of the DBS lead trajectory can substantially reduce RF heating during 3T MRI. The authors’ results indicate that simple adjustments to the lead’s configuration, such as small, concentric loops near the burr hole, can be readily adopted during DBS lead implantation to improve patient safety during MRI.

Since the introduction of deep brain stimulation (DBS), it has provided remarkable therapeutic benefits to patients with movement disorders and other neurological diseases.^1–4 During DBS, electrical stimulation is delivered to specific subcortical targets via implanted electrodes connected to a programmable implantable pulse generator (IPG) placed in the clavicle region via subcutaneous extensions.

MRI is a versatile neuroimaging modality that offers potential benefits for patients with DBS devices. It is well suited for postoperative monitoring, target verification, and localization of the electrodes,^5,6 as well as monitoring the functional effects of stimulation on affected brain networks.^7–9 The need for MRI in patients with DBS implants is increasing, and an estimated 66%–75% of these patients will require an MRI examination within 10 years of device implantation.¹⁰

Radiofrequency (RF)–induced heating is the primary limiting factor for performing MRI examinations in patients with DBS devices. The implanted DBS lead behaves like an antenna when coupled with the electric field of the MRI transmit coil, resulting in localized tissue heating.^11,12 This phenomenon, known as the antenna effect, increases the specific absorption rate (SAR) of RF energy deposited in the tissue surrounding the lead’s tip. Because of safety concerns associated with RF heating, DBS manufacturers have established stringent device-specific guidelines for MRI protocols. Currently, most neuroimaging procedures are limited to a field strength of 1.5T in a horizontal closed-bore scanner. The heating-related thresholds remain conservative, with B_1+rms < 1.1 μT or whole-head SAR < 0.1 W/kg, 30 times below the FDA’s limit for scanning in the absence of implants.¹³ These guidelines limit routine clinical MRI for patients with DBS devices.

Efforts aimed at mitigating RF heating for patients with DBS systems have increased in recent years. These contributions include modifying the material and design of DBS leads,^14,15 introducing novel MRI head coil technology to induce a region of low electric field that coincides with the implanted lead’s trajectory on a patient-specific basis,^16–19 and potentially utilizing ultra-high-field²⁰ and vertical open-bore scanners with different orientations of the magnetic and electric fields.^21,22 While these approaches are promising, their clinical utility is still limited since they require changes to existing DBS or MRI technology or methodologies.

Surgical modification of the extracranial portion of the DBS lead trajectory is an alternative method for reducing RF heating by minimizing the coupling of the leads with the MRI electric fields without altering the DBS system or MRI methodologies.^23,24 Currently, precise surgical guidelines are established for implanting the intracranial trajectory of the lead where the entry point on the skull and the angle of insertion are predetermined to target an intended brain structure. There is a lack of similar guidelines for implanting the extracranial portion of the lead, as this does not contribute to the therapeutic effects of DBS. This results in substantial variations in the implanted extracranial lead trajectories across patients, leading to highly variable and unpredictable RF tissue heating.²⁵ Baker et al. first introduced the concept of manipulating the DBS lead trajectory into a looped configuration using a trajectory formation device.²⁶ More recently, one proposed modified trajectory configuration involved surgically shaping the extracranial portion of the DBS lead into concentric loops near the respective burr hole.^23,25 The effect of the concentric loops was predominant for the contralateral DBS leads, demonstrating an 18-fold reduction in the maximum 1g-averaged SAR in simulations with a loaded transmit body coil.²³ However, the optimal positioning of the loops remains unknown, and other studies that implemented loops near the surgical burr hole fixated the lead trajectory throughout their studies,^27–29 without fully evaluating pertinent RF heating-related parameters, including loop dimensions and the trajectory’s position in the phantom.

The goal of this study was to perform the first large-scale, systematic evaluation to identify characteristics of extracranial DBS lead trajectories that minimize RF heating during MRI at 3T. We assessed the RF heating of 244 unique lead trajectories with a commercial DBS system, focusing on parameters such as the location of the trajectory on the skull, trajectory topology (i.e., the number of concentric loops), and the size of the loops. Additionally, we evaluated the RF heating of the lead trajectories under varying IPG configurations and imaging conditions to determine the robustness of our experimental methods and the reliability of the low-heating trajectories. Finally, we demonstrated the feasibility of translating low-heating lead trajectories from phantom experiments to DBS patients.

Methods

Lead Trajectory Parameters

We systematically examined the effect of three parameters of the extracranial lead trajectory on RF heating: the position of the loops on the skull that are surgically and anatomically feasible, the number of concentric loops (1–3 loops), and the diameter of the concentric loops (2.5–4.5 cm with 0.5-cm increments) (Fig. 1B). These parameters were selected based on prior studies that demonstrated their nontrivial effects on the variations in the magnitude of RF heating across trajectories.^{11,23,26,27,30} To ensure precise replication of the intended trajectories during the experiments and to guarantee reproducibility, we created 3D models of the lead trajectories and the anthropomorphic phantom in a computer-aided design (CAD) tool (Rhino 7.0, Robert McNeel & Associates). We then translated the intended trajectories to a commercial DBS system using 3D-printed trajectory guides. The evaluated trajectories are categorized by trajectory parameters and are shown in Supplementary Figs. 1 and 2.

FIG. 1.

Extracranial DBS lead trajectory development process. A: Lead trajectories were prepared on the 3D-printed skull phantom to determine feasibility the of the trajectories. B: Three-dimensional models of all trajectories were created in a CAD tool. Trajectory parameters included the position of the loops on the skull, topology, and diameter of the loops. C: Trajectories were replicated with a commercial DBS system. Experimental trajectories matched their digital counterparts. Figure is available in color online only.

Phantom Fabrication

While most RF heating studies of active implants are performed with box-shaped phantoms, the electric field distributions, and consequently SAR, are substantially different in a box-shaped phantom than in human body models.³¹ To address this limitation, we fabricated an anthropomorphic phantom based on CT images of a patient with a DBS device (Supplementary Fig. 3); details about the phantom fabrication are provided elsewhere.³² To ensure anatomical representation, we compared the size of the skull against the skull dimensions of 53 other patients with DBS devices, including the maximum head length, maximum head breadth, and distance from the most superior point on the head to an axial plane intersecting the eyebrows (Supplementary Fig. 3). The maximum breadth, length, and distance of the skull used in the experiments were 141 mm, 185 mm, and 91 mm, respectively (additional details are included in Supplementary Figs. 4 and 5). The skull also included a grid with 20-mm spacing to assist with specific positioning of the lead trajectory (Supplementary Fig. 3).

Experimental Setup

A total of 244 measurements were performed to identify DBS lead trajectories that minimize RF-induced heating. Experiments were conducted with an Abbott full DBS system consisting of a 40-cm lead (model 6172), a 50-cm extension (model 6371), and the Infinity-5 IPG. The lead was implanted in the right hemisphere, with the entry point on the right side of the skull and the angle of insertion similar to what is done for targeting the subthalamic nucleus (Fig. 2A). The IPG was placed in either the right or left pectoral region to represent ipsilateral or contralateral lead trajectories (n = 94 and n = 150, respectively). The skull was filled with an agar-based gel (σ = 0.47 S/m, σ_r = 78) prepared by mixing 32 g/L of edible agar (Landor Trading Company, gel strength 900 g/cm²), 5 g/L of sodium benzoate (Sigma Aldrich), and saline solution (1.55 g NaCl/L). Using an agar-based solution to fill the skull was advantageous compared to a polyacrylamide gel, as it formed a semisolid gel that kept the leads in place. The rest of the phantom was filled with 18 L of saline solution (σ = 0.50 S/m, ε_r = 80) to mimic the conductivity of average human tissue. The maximum temperature increase (ΔT_max), was measured in the gel surrounding the lead tip using fiber-optic temperature probes (Osensa, resolution 0.01°C). An additional temperature probe was placed in the phantom away from the DBS device to ensure that the temperature did not increase elsewhere in the phantom. Experiments were performed in a 3T Prisma MRI scanner using the body transmit coil (Siemens Healthineers) (Fig. 2B). The phantom was placed in the head-first, supine position and oriented such that the eyebrows/tip of the DBS lead were at the scanner’s isocenter. RF exposure was generated with a T1-weighted turbo spin echo dark fluid pulse sequence (TR 2750 msec, TE 8.2 msec, FA 170°, acquisition time 381 seconds, B_1+rms = 2.7 μT).

FIG. 2.

A: Three-dimensional rendering of an example lead trajectory in the 3D-printed phantom. Temperature probes were attached to the DBS lead. B: Experimental setup on a 3T Siemens Prisma scanner. C: Phantom setup for configurations where the lead was contralateral or ipsilateral with respect to the IPG. D: The effect of imaging and implant-related parameters including the landmark (brain, chest, and abdomen), imaging plane and its phase encoding direction, and IPG location and configuration on ΔT_max was evaluated. Axial slices were originally acquired. The placement of the trajectory on the anterior surface of the IPG was the original configuration. Figure is available in color online only.

Test-Retest Analysis

To evaluate the reliability of our measurements, we repeated the experiments with 18 randomly selected trajectories, of which 9 were contralateral and 9 were ipsilateral with respect to the IPG. These trajectories represented those that generated RF heating in the top and bottom 20%, as well as those in between these thresholds. We calculated the intraclass correlation coefficient (ICC) and the 95% confidence interval using R software (version 4.2.1, https://www.r-project.org/) in RStudio (posit.co) based on a two-way mixed-effects model, single rating, and absolute agreement.³³

Effect of Patient Positioning, Imaging Parameters, and IPG Configuration

Experiments were repeated with the 18 randomly selected trajectories incorporated in the test-retest analysis to determine the effect of slight changes in patient positioning on RF heating. Originally, the eyebrows/tip of the DBS lead was located at the scanner’s isocenter. The second and third patient positions were 20 mm superior and inferior to the eyebrow level (Fig. 2D) as different indications for brain MRI may require different imaging locations.

Additionally, we performed experiments at the chest and abdomen imaging landmarks with another group of 12 randomly selected trajectories (6 contralateral and 6 ipsilateral with respect to the IPG) that represented the top and bottom 20% of RF heating during head imaging. These experiments were performed to demonstrate that the trajectories that exhibited minimum heating during brain imaging do not induce excessive RF heating during imaging of other body regions (i.e., chest and abdomen).

Finally, experiments were repeated with 6 randomly selected trajectories (3 contralateral and 3 ipsilateral with respect to the IPG) to assess the effect on RF heating when perturbing the imaging plane, location of the IPG, and the trajectory of the extension around the IPG.³⁴ Experiments were performed for acquisition of coronal (right-left and feet-head phase encoding) and sagittal (anterior-posterior and head-feet phase encoding) slices (Fig. 2D). The location of the IPG was translated 10 mm superior to its original position and 20 mm in the medial and lateral directions (Fig. 2D). The trajectory of the excess extension around the IPG was originally placed anterior to the surface of the IPG; however, the trajectory was placed posterior to the surface of the IPG in the additional experiments (Fig. 2D). These experiments were performed to assess the robustness of trajectory modification in reducing RF heating when subjected to various perturbations that may occur in practice.

We calculated the absolute difference in ΔT_max between the original configuration and each configuration where a change was made to the landmark, imaging plane, or IPG-related configuration. This difference in ΔT_max is indicated as |ΔT_max2 − ΔT_max1|.

Surgical Implementation of Modified Lead Trajectories in Patients

To assess if low-heating trajectories could be readily adopted during DBS surgery, two neurosurgeons (J.M.R. at Northwestern Memorial Hospital and J.P. at Albany Medical Center) were instructed to implement low-heating trajectories in their patients undergoing DBS surgery. For some patients (J.P.), we used curved mayo scissors passed posterior and to the left of the incision that were opened to their widest position to create a pathway for a coiled lead to be inserted. We then coiled the lead on itself in 2 or 3 concentric circles at the burr hole before passing the rest of the lead toward the temporal lobe where it would be later connected to the extension (Fig. 3A). For other patients (J.M.R.), the scalp was routinely elevated circumferentially from the skull to facilitate a tensionless closure using a periosteal elevator. The lead was then anchored with the burr hole ring, clip, and cap. If this lead were to be connected to an ipsilateral IPG, the electrode tip was secured in its protective cap and then passed under the scalp to the location of a previously determined extension incision that would be used during the IPG implant stage. If this lead were to be connected to a contralateral IPG, the lead was passed to an extension with a blocking pin that had been placed at the time of the IPG implantation and the two were connected (Fig. 3B). In either case, the electrode was then coiled into 2 or 3 loops that were placed under the scalp posterior to the main linear incision (Fig. 3C). Additional details regarding the surgical approach are provided in Supplementary Fig. 6 and Video 1.

FIG. 3.

A and B: A curved Mayo scissor was used to create an opening for the extracranial loops at Albany Medical Center (A) as opposed to a tensionless closure at Northwestern Memorial Hospital (B). C: Concentric loops implemented near the surgical burr hole, similar to the phantom experiments. D and E: Three-dimensional surface-rendered views of CT images of patients with unmodified (highlighted in magenta, D) and modified (highlighted in blue, E) DBS lead trajectories that were replicated during phantom experiments. Figure is available in color online only.

Following the procedures, lead trajectories were segmented from postoperative CT images using 3D slicer (http://slicer.org) and processed in the Rhino CAD tool. Models of the trajectories were 3D printed to help with trajectory replication during phantom experiments. Furthermore, experiments were performed with lead trajectories from the same neurosurgeons prior to receiving instructions on implementing low-heating trajectories to compare the effectiveness of surgical lead modification. Retrospective use of patients’ imaging data for the purpose of modeling was approved by Northwestern Memorial Hospital and Albany Medical Center’s institutional review boards.

Results

RF Heating Measurements

Leads that were contralateral with respect to the IPG had higher heating compared with leads that were ipsilateral to the IPG based on a two-sample one-tailed t-test (p = 2.44 × 10⁻²³). The mean ± SD ΔT_max was 3.44°C ± 1.93°C (0.24°C–7.34°C) for contralateral leads (Fig. 4C). The mean ΔT_max was 1.26°C ± 1.17°C (range 0.09°C–4.75°C) for ipsilateral leads (Fig. 4D). All loops were coiled in the clockwise direction, as illustrated in Supplementary Fig. 7.

FIG. 4.

Superposition of all the trajectories evaluated in this study. There were 150 trajectories for the lead contralateral to the IPG (A) and 94 trajectories for the lead ipsilateral to the IPG (B). Distribution of ΔT_max for all lead trajectories contralateral (C) and ipsilateral (D) with respect to the IPG. Superposition of low heating trajectories where ΔT_max < 1°C to illustrate the anatomical position of these trajectories (E and F). Figure is available in color online only.

Effect of Spatial Orientation

Low-heating trajectories had concentric loops positioned closer to the surgical burr hole (Fig. 4E and F). This trend was especially apparent for leads that were contralateral to the IPG. For lead trajectories where ΔT_max < 1°C, 16 of 20 contralateral and 27 of 48 ipsilateral lead trajectories were within 40 mm (radially) of the surgical burr hole. The impact of the anatomical location of the concentric loops was consistent regardless of the number of concentric loops and the size of the loops.

Effect of Number of Loops

For contralateral leads, the mean ΔT_max values were 4.99°C ± 1.03°C, 2.07°C ± 1.14°C, and 0.73°C ± 0.40°C for single, double, and triple loop configurations, respectively (Fig. 5A–C). For ipsilateral leads, the mean ΔT_max values were 1.79°C ± 1.39°C, 0.79°C ± 0.82°C, and 1.19°C ± 0.56°C for single, double, and triple loop configurations, respectively (Fig. 5D–F). As the number of concentric loops increased from 1 to 3 in contralateral trajectories, ΔT_max decreased (correlation coefficient, r = −0.83). For the ipsilateral trajectories, there was a weaker negative correlation between ΔT_max and the number of concentric loops (r = −0.29).

FIG. 5.

Distribution of ΔT_max categorized by the number of concentric loops with the lead contralateral (A–C) and ipsilateral (D–F) to the IPG. Distribution of ΔT_max categorized by the diameter of the concentric loops with the lead contralateral (G–K) and ipsilateral (L–P) to the IPG. Figure is available in color online only.

Effect of Loop Size

The diameter of the loops in the trajectories did not correlate with ΔT_max (r = 0.17 and r = 0.11 for contralateral and ipsilateral trajectories, respectively) across the different loop topologies. Table 1 provides the mean ΔT_max for each loop size for leads contralateral (Fig. 5G–K) and ipsilateral (Fig. 5L–P) to the IPG.

TABLE 1.

Effect of loop size on ΔT_max

Loop Diameter, cm	Mean ± SD of ΔTmax (°C)
	Leads Contralat to IPG	Leads Ipsilat to IPG
2.5	2.11 ± 1.22	0.71 ± 0.59
3.0	4.04 ± 1.98	1.52 ± 1.62
3.5	4.11 ± 2.01	1.43 ± 1.22
4.0	3.35 ± 1.95	1.46 ± 1.09
4.5	3.38 ± 1.36	1.11 ± 0.65

Test-Retest Measurements

The repeated experiments demonstrated excellent reliability (ICC 0.96, 95% CI 0.91–0.99). Figure 6A shows the ΔT_max values for the 18 test-retest experiments.

FIG. 6.

A: ΔT_max from the test-retest experiments (ICC 0.96, 95% CI 0.91–0.99). B: The effect of IPG configuration and imaging plane on ΔT_max. C: The effect of imaging position within the brain on ΔT_max. The original imaging landmark corresponded to 0 mm. D: The effect of the imaging landmark at the chest and abdomen on ΔT_max. A-P = anterior-posterior; F-H = feet-head; ID = trajectory number; H-F = head-feet; R-L = right-left. Figure is available in color online only.

Sensitivity Analysis

Changing the patient position (by ± 20 mm) to mimic different brain imaging scenarios produced minimal differences in ΔT_max. The mean |ΔT_max2 − ΔT_max1| was 0.52°C ± 0.98°C, where ΔT_max1 corresponded to the ΔT_max value when the phantom was at the central position and ΔT_max2 occurred when the phantom was superior or inferior to the central position. Figure 6C shows the ΔT_max at 3 brain imaging locations.

Altering the imaging landmark from the brain to the chest and the abdomen resulted in a reduction of RF heating around the DBS lead tip. The mean ΔT_max values were 0.49°C ± 0.43°C and 0.16°C ± 0.15°C when imaging was performed at the chest and abdominal landmarks, respectively (Fig. 6D). On the other hand, the mean ΔT_max was 2.58°C ± 2.58°C for the same lead trajectories when the experiments were conducted with the landmark located at the level of the DBS lead tip. Thus, low-heating lead trajectories could mitigate RF heating regardless of the imaging landmark.

Similarly, altering the location of the IPG or the trajectory of the extension around the IPG produced minimal differences in ΔT_max (Fig. 6B). The mean |ΔT_max2 − ΔT_max1| was 0.66°C ± 0.53°C across all IPG-related configuration changes, where ΔT_max1 corresponded to the original configuration of the IPG. Furthermore, changing the slice acquisition and the phase encoding direction did not affect ΔT_max, where the mean |ΔT_max2 − ΔT_max1| was 0.66°C ± 0.53°C and ΔT_max1 corresponded to axial slice acquisition with phase encoding in the anterior to posterior direction (Fig. 6B).

Effectiveness of Surgically Implemented Modified Trajectories

Based on the findings from the phantom experiments, DBS lead trajectories with 2 or 3 concentric loops, approximately 3 cm in diameter, were placed near the burr hole in 6 new patients (Fig. 3D and E). After the surgery, modified trajectories were segmented from CT images and replicated in phantom experiments with the Abbott DBS system, resulting in a mean ΔT_max of 1.24°C ± 0.31°C (range 0.87°C –1.57°C). We compared the RF heating of the surgically modified trajectories with 6 unmodified lead trajectories previously implemented by the same neurosurgeons. The mean ΔT_max was 3.42°C ± 0.57°C (range 2.46°C–4.21°C) for the unmodified trajectories (Fig. 7), demonstrating a threefold reduction in RF heating.

FIG. 7.

ΔT_max for the surgically modified and unmodified lead trajectories. There was almost a threefold reduction in the mean ΔT_max between unmodified and modified lead trajectories. Figure is available in color online only.

Discussion

There has been a steady increase in the application of modern neuroimaging techniques for guiding and interpreting the outcomes of DBS therapy.^35–38 Most DBS devices are approved for MRI at 1.5T under certain conditions, and one device is approved for 3T MRI, but the majority of the patients at our institutions still receive DBS devices that are 1.5T MRI conditional. Thus, there are strong incentives to work toward making 3T MRI compatible with DBS, especially with the presence of off-label imaging at 3T MRI.³⁹ Three-Tesla MRI confers better contrast-to-noise ratio, making it easier to delineate small abutting structures that can be crucial to localizing electrodes in areas like the subthalamic nucleus.⁴⁰ Furthermore, highly accelerated sequences with improved sensitivity that make use of functional MRI desirable for DBS are only accessible at 3T.⁴¹ Finally, while 3T MRI is generally more sensitive to susceptibility effects, even this can act as an advantage when localizing electrodes in iron-rich areas of the brain and for revealing pallidofugal and striatonigral fiber tracts.⁴²

The main barrier to MRI at 3T for patients with DBS systems is potential RF-induced heating around the lead tip. It is well recognized that the trajectory of elongated conductive implants, such as leads encountered in neuromodulation and cardiovascular implantable electronic devices, has a significant effect on MRI-induced RF heating.^23,43 Safety studies on the RF heating of DBS systems have traditionally evaluated trajectories that elicit the worst-case scenario heating or looped trajectories without specificity.^27–29 Here, we present the first large-scale study on how characteristics of the extracranial DBS lead trajectory affect RF heating and quantify the extent of RF heating reduction by surgical modification of the lead trajectory. Effective trajectory characteristics were also readily implemented in new patients undergoing DBS surgery.

We found that placing 2 or 3 concentric loops within 40 mm of the surgical burr hole was most effective for reducing RF induced heating. Our results were consistent with the earliest introduction of a DBS lead management prototype that formed 2.25 successive loops ranging from 1.8 to 2.3 cm in diameter near the burr hole.²⁶ Other in vitro studies have incorporated loops in the DBS lead trajectories, including a single 3-cm loop,³⁰ 1–3 counterclockwise loops fixed on to the external surface of the phantom,²⁷ 2 loops on the external surface of the phantom,²⁸ and a single loop around the burr hole.²⁹ While all these trajectories can be classified as having looped topologies, their geometries and positioning were still vastly different. This present study provided a direct comparison to determine the characteristics of the trajectory that most effectively reduced RF heating; it is important to increase the number of concentric, overlapping loops near the surgical burr hole, but changing the size of the loops did not yield the same effect. These results also made it possible to refine recommendations on how to create low-heating trajectories that can be easily implemented.

Additional experiments were performed after perturbing the imaging landmark, IPG-related configuration, and the imaging plane to determine their effect on RF heating and if low-heating trajectories were resilient to such changes. Changing the position in the head resulted in minimal changes in RF heating and changing the imaging landmark resulted in consistently lower ΔT_max across low, average, and high heating trajectories. Translating the position of the IPG in the superior, medial, and lateral directions also induced minimal changes in ΔT_max from the original configuration. Lastly, changing the slice acquisition and phase-encoding direction resulted in minimal changes in ΔT_max. This is expected, as the specific slice and the encoding directions correspond with the MRI gradient coils that do not produce gradient-induced heating for medical devices with elongated leads such as a DBS system.

Our preliminary clinical results demonstrated that modifying the extracranial DBS lead trajectory is feasible within the current surgical procedure without increasing the complexity or duration. The modified DBS lead trajectories reduced RF heating during 3T MRI by almost 3-fold compared with the unmodified lead trajectories previously implanted by the same neurosurgeons. This demonstrates great potential for widespread adoption of surgical modification of the lead trajectory, as both neurosurgeons in this study presented different surgical practices to produce the concentric looped trajectories.

However, there are several limitations, primarily regarding the generalizability of our results. Key factors that affect MRI-induced RF heating include the electric length of the internal wires connecting the electrodes to the pulse generator and the dielectric properties of the lead material encasing these wires. For example, DBS leads may have tightly woven helical wires that are much longer than the lead’s visible length or nearly straight wires. Such factors can vary substantially across different DBS manufacturers and across different lead models produced by the same manufacturer. To further explore this, we recently conducted a series of experiments with a full DBS system from Boston Scientific. These results were consistent with the results obtained with the Abbott full DBS system, primarily the effect of extracranial lead trajectory characteristics on RF heating and the performance of the patient-derived modified and unmodified trajectories.⁴⁴ Furthermore, given the limited cohort of 12 patient-derived modified and unmodified trajectories, the relative temperature rise differences observed between the modified and unmodified trajectories do not indicate the absolute heating outcome during clinical MRI.

Conclusions

Surgically modifying the extracranial DBS lead trajectory while focusing on increasing the number of concentric loops and the loops’ placement can effectively mitigate RF heating during 3T MRI. The reduction in RF heating was most apparent for leads that were contralateral with respect to the IPG as the parameter space is larger. Clinical adoption of the trajectory specifications was feasible, and subsequent experiments confirmed low RF heating. Overall, this method can enable safer imaging during MRI at 3T for patients with DBS systems.

Supplementary Material

Video

VIDEO 1. Example implementation of a modified DBS lead trajectory during surgery at Northwestern Memorial Hospital. © Jasmine Vu, published with permission. Click here to view.

ABBREVIATIONS

DBS

deep brain stimulation

ICC

intraclass correlation coefficient

IPG

implantable pulse generator

RF

radiofrequency

SAR

specific absorption rate

ΔT_max

maximum temperature increase