Evaluating and Optimizing Dentato-Rubro-Thalamic-Tract Deterministic Tractography in Deep Brain Stimulation for Essential Tremor

Abstract BACKGROUND Dentato-rubro-thalamic tract (DRT) deep brain stimulation (DBS) suppresses tremor in essential tremor (ET) patients. However, DRT depiction through tractography can vary depending on the included brain regions. Moreover, it is unclear which section of the DRT is optimal for DBS. OBJECTIVE To evaluate deterministic DRT tractography and tremor control in DBS for ET. METHODS After DBS surgery, DRT tractography was conducted in 37 trajectories (20 ET patients). Per trajectory, 5 different DRT depictions with various regions of interest (ROI) were constructed. Comparison resulted in a DRT depiction with highest correspondence to intraoperative tremor control. This DRT depiction was subsequently used for evaluation of short-term postoperative adverse and beneficial effects. RESULTS Postoperative optimized DRT tractography employing the ROI motor cortex, posterior subthalamic area (PSA), and ipsilateral superior cerebellar peduncle and dentate nucleus best corresponded with intraoperative trajectories (92%) and active DBS contacts (93%) showing optimal tremor control. DRT tractography employing a red nucleus or ventral intermediate nucleus of the thalamus (VIM) ROI often resulted in a more medial course. Optimal stimulation was located in the section between VIM and PSA. CONCLUSION This optimized deterministic DRT tractography strongly correlates with optimal tremor control. This technique is readily implementable for prospective evaluation in DBS target planning for ET.


S
ince recent years the dentato-rubrothalamic tract (DRT) is considered to be an effective target for deep brain stimulation (DBS) in essential tremor (ET) patients. 1,2eterministic tractography is available in stereotactic software which enables depiction of this tract for DBS target planning. 3owever, tractography can be deceptive.[6][7][8] ABBREVIATIONS: DRT, dentato-rubro-thalamic tract; DWI, diffusion-weighted imaging; ET, essential tremor; FA, fractional anisotropy; PSA, posterior subthalamic area; RN, red nucleus; SMA, supplementary motor area; VIM, ventral intermediate nucleus of the thalamus Supplemental digital content is available for this article at www.operativeneurosurgery-online.com.
Optimization of DRT tractography by choosing a standard set of ROIs may prove beneficial for reducing heterogeneity of tractography in DBS targeting.
Furthermore, if optimized DRT tracking indicates a clear target area and shows high correlation with optimal tremor control, it could be used to replace awake test stimulation during DBS for ET, decreasing the burden of surgery and increasing accessibility of this treatment for ET patients. 4ERATIVE NEUROSURGERY Thus far, no standardized DRT depiction has been described, and we evaluated this in a cohort of ET patients after DBS implantation and the correlation of this technique with tremor control.

Patients
We implemented deterministic DRT tractography (Brainlab AG, Munich, Germany) in DBS for ET from October 2017, following results in recent literature describing correlation of tremor suppression and DRT DBS and our own findings in a cohort study. 5,3We did not use a fixed set of ROIs and noticed considerable differences in DRT depiction depending on the ROIs applied.Exact DRT settings used for surgery could not be recovered, as the software does not allow for this.We therefore decided to evaluate different DRT depictions and tremor control after DBS implantation.By using this methodology, we intended to standardize tractography by determining a fixed set of ROIs for DRT depiction and for this DRT to show correspondence with intraoperative tremor control.
The Medical Ethics Committee of our institution reviewed the study protocol and confirmed that the Dutch Medical Research Involving Human Subjects Act (WMO) is not applicable.For this retrospective analysis, patient consent was not considered necessary.

Image Acquisition at 3 Tesla Magnetic Resonance Imaging
Magnetic resonance imaging (MRI) scanning was done using a 3T Elition Ingenia (Philips, Best, The Netherlands) and 32-channel receive coil (Philips).

DRT Tractography-Based Surgical Planning
Our DBS surgical procedure is described in detail previously. 5,6uring target planning, ventral intermediate nucleus of the thalamus (VIM) and posterior subthalamic area (PSA) were aligned in one single surgical trajectory.In order to facilitate this, an estimation of the DRT was depicted using the following ROIs: ipsilateral motor cortex, ipsilateral VIM as depicted by the software (anatomic mapping, Brainlab AG), ipsilateral PSA, ipsilateral red nucleus (RN), ipsilateral superior cerebellar peduncle, ipsilateral dentate nucleus, contralateral superior cerebellar peduncle, and contralateral dentate nucleus.In the period after tractography was introduced, we applied various combinations of these ROIs in order to depict a suitable estimation of DRT position for target planning, resulting in both crossing and noncrossing DRT depictions.The trajectory was planned so that it made contact with this DRT at the level of VIM and/or in its descent towards the PSA.The possibility of trajectory alignment with the DRT depended on other trajectory require-ments such as prefrontal entry point on top of a gyrus and avoiding the ventricle, caudate nucleus, and blood vessels.The corticospinal tract was depicted utilizing the ROI motor cortex and cerebral peduncle (4-mm trajectory distance). 7Test stimulation was evaluated by applying monopolar (ring mode) stimulation (macrostimulation electrode, Elekta) with increasing current strength at 133 Hz and 60 μs.Stimulation was performed using 2 mm intervals (every 2 mm along the trajectory).DBS electrodes used for implantation were Boston Vercise Cartesia 8 contact directional.

Patient Assessment and DBS Programming
DBS programming was performed by specialized nurses, without knowledge of the position of the DRT.Tremor control was considered successful when complete tremor suppression (grade 0) or almost complete suppression (grade 1) was observed. 5

Optimization of DRT Tractography Through Comparing Various ROI Combinations
][10][11][12] Figure 1 illustrates how each ROI in the separate anatomic areas was drawn on axially orientated slices.The position of VIM was indicated by the anatomic segmentation function of the Brainlab software (Brainlab AG).
Depiction of DRT-PSA was conducted, then the software based VIM ROI was added (DRT-VIM), facilitating evaluating anatomic position and influence of the VIM ROI.Crossing fibers were also separately depicted for DRT-PSA (crossing fibers were not depicted for DRT-VIM).Lastly a DRT was depicted with the RN included as ROI, the DRT-RN, facilitating evaluating influence of the RN ROI.Crossing fibers were also evaluated for this method.This resulted in depicting 5 DRTs per trajectory; 10 per patient.We studied whether the course of the specific DRT made sense from an anatomic point of view and if the electrode (as localized by intraoperative cone beam computed tomography) was situated within the tract.
The fractional anisotropy (FA) was set at 0.2 (standard software setting) when starting tractography.This was adjusted for obtaining a (visual) smoothly depicted track including the known anatomic landmarks.For every track, the maximum angulation was set at 20 • (standard software setting).This was increased to 80 • after a DRT was depicted for evaluating possible changes.For every track, a minimum length of 80 mm was set and kept (standard software setting).

FIGURE 1. Regions of interests (ROIs) applied for deterministic tractography of the dentato-rubro-thalamic-tract. Panels showing the different ROIs, in blue, visualized on an axial 3 Tesla T2 MRI. All ROIs were created on axial MRI slices, aligned to the anteriorposterior commissure line. A shows the motor cortex region ROI. The motor cortex was recognizable as the first horizontally orientated gyrus directly after the vertically orientated supplementary motor area (1). The ROI always included the region of the hand knob (2). B
shows the ventral intermediate nucleus ROI as depicted by the software after automatic thalamic segmentation, 2 mm above the posterior commissure.C shows the posterior subthalamic area ROI.For identification, the parallelogram method was used. 9D shows the red nucleus ROI.The ROI was created on the maximum diameter of the nucleus and enlarged by 2 to 3 mm.E shows the superior cerebellar peduncle ROI, recognizable on a slice with clear "tram track" connection visible from the cerebellum to the brainstem.F shows the dentate nucleus ROI.The ROI was created on the maximal diameter of the nucleus.

Analysis of Active Contacts Relative to Standardized DRT
The distance from DRT to center of active contact was measured, using axial orientated imaging, in millimeters by depicting a circle around the DBS electrode.The location (along the electrode trajectory) and width of this circle were adjusted so distances could be measured.The DRT was reached when its outer rim was touched.

Data Analysis
Due to the exploratory nature of the study, data were analyzed descriptively.Numerical data are presented as mean ± standard deviation (range).

DBS Placement and Intraoperative Tremor Control
From October 2017 until March 2020, a total of 37 DBS leads were placed in 20 ET patients.There were 10 male patients, and age at surgery was 69 ± 9 (45-84) yr.
In 35 placements, a single trajectory was needed for complete intraoperative tremor control during test stimulation, and in 2 placements, 1 additional trajectory was needed (case number 2 and 8 in the tables).

Evaluating and Optimizing the DRT After DBS Placement
In all hemispheres in which DBS electrode placement was performed, a construction of DRT-PSA depiction was possible with the subset of ROIs chosen in the ipsilateral motor area, ipsilateral PSA, ipsilateral cerebellar peduncle and dentate nucleus.In 34 (92%) cases the DBS electrode was situated within this DRT-PSA.Crossing DRT fibers (contralateral cerebellar peduncle and dentate nucleus included as ROIs) were depictable in 8 (22%) of the 37 DBS-implanted hemispheres.
The location of VIM, as segmented by the software, was located medial to the DRT-PSA in 27 of 37 (73%) hemispheres and inside DRT-PSA for the remaining hemispheres.Adding VIM as a ROI for the tractography changed the course of the DRT depiction medially (Figure 2).In 31 (84%) hemispheres, the DBS electrode was situated within this DRT-VIM.After adding VIM as ROI, crossing fibers were not separately reviewed.

. Optimizing deterministic tractography of dentato-rubro-thalamic-tract. A-G show the different possible depictions of DRTs which are used for comparison. The above panel shows an axial, the bottom a coronal view on a 3-Tesla T2 MRI of the thalamic and subthalamic area (commissural aligned imaging). The green DRT is depicted with ROIs in the ipsilateral motor area, ipsilateral PSA, ipsilateral cerebellar peduncle and dentate nucleus (DRT-PSA). The DRT-VIM (purple) is depicted after adding VIM (depicted by the software) as an ROI. The DRT-RN (orange) is depicted with ROIs in ipsilateral motor area, ipsilateral red nucleus, ipsilateral cerebellar peduncle and dentate nucleus (DRT-RN). The crossing DRT-RN (red) crosses the midline at the level of the red nucleus, and is depicted with
When we constructed the DRT-RN depiction, with ROIs in ipsilateral motor area, ipsilateral RN, ipsilateral cerebellar peduncle and dentate nucleus, the electrode was situated within this DRT-RN in 20 (54%) of the hemispheres.The RN ROI was located more medial compared to the DRT-PSA in 17 (46%) hemispheres, creating a more medial orientated depiction of the DRT-RN (Figure 2).Crossing fibers could be depicted in 12 (32%) hemispheres.
Maximum angulation settings did not alter the depiction of the different DRT projections.Range of FA settings for all was between 0.14 and 0.37.
A complete overview is given in the Table.

Neuroanatomical Location of Active Contacts Relative to the DRT-PSA
The evaluation of contact localization was only performed for the DRT showing highest correspondence with intraoperative tremor control.The main purpose of current analysis is to determine whether DRT depiction can serve as readily visible DBS target for intraoperative tremor control.Moreover, it is expected that intraoperative tremor control corresponds with postoperative tremor control.
The number of active contacts located within the DRT-PSA was 30 (84%), of which 28 (93%) showed successful tremor control.The 2 active contact points located in DRT-PSA with suboptimal tremor response represents one patient.In 3 active contact points distance to DRT was 2 mm, of which all showed tremor control.In 4 active contacts the distance to DRT-PSA was between 3.0 and 4.8 mm, all showed insufficient tremor response.Three contact points were located at more than 4.5 mm distance of which all had severe stimulation-induced gait ataxia.No electrode revisions were performed.
The conventional standard VIM stereotactic coordinates as applied by our group, ie, 15 mm lateral, 8 mm anterior, 2 mm dorsal relative to the posterior commissure on the anteriorposterior commissure aligned MRI, were located within DRT-PSA for 13 trajectories (35%).

Tremor Alleviation and Side Effect Profiles
Tremor scores at baseline and during follow-up at 13 ± 8 (2-32) mo are listed in the Supplementary Table .Successful tremor control was achieved in 32 out of 37 contralateral body  20) Y ( 20) Y ( 21)

Surgical Complications
Two patients developed severe dysarthria and gait ataxia shortly after surgery, one due to bilateral hemorrhage surrounding the electrodes and one due to delayed-onset edema.A temporary nasogastric feeding tube was needed.Recovery took months and mild balance difficulties persisted.One patient developed infection of the left burr hole cover.Optimizing DRT tractography was feasible using a fixed set of ROIs in the motor cortex, the PSA, the superior cerebellar peduncle and the dentate nucleus (all ipsilateral).The DRT-PSA (constructed after DBS surgery) showed a high correspondence with tremor response, both for intraoperative test stimulation and for postoperative short-term follow-up. 3,8This comparison between different deterministic approaches for DRT depiction and concurrent evaluation with tremor control for DBS in ET has not been described previously, and provides a short cut for stereotactic planning that can be implemented directly.
Applying tractography for DBS in a standardized manner is a promising technique to optimize the outcome of surgery.][11] However, careful evaluation of the outcome of tractography is necessary as adding certain ROIs can incorrectly influence its course. 9Including a RN or VIM ROI regularly resulted in a more medial depiction of the DRT.This variability may have hampered in finding the correspondence between DRT and tremor control in previous studies. 9,12Continuous evaluation of tractography methods and correspondence with DBS effects and side effects will make this visualization technique more robust and reliable for DBS.
The VIM itself is part of the DRT and the DRT-PSA must therefore run through the actual VIM.The tractography here provides a better estimation of the optimal location for treatment than either standard stereotactic atlas coordinates or the softwarebased VIM ROI.
Another important finding in DRT depiction is whether or not to display crossing fibers.Findings in the current study and others suggest there is both a crossing and noncrossing branch of the DRT on DWI that can be used for DBS target planning. 3,10When using deterministic tractography supported by commercially available software, it is sensible to pursuit a solely noncrossing DRT.Ultimately, the value of the applied DRT for DBS is not determined by its course, but by influence on effectiveness of DBS therapy.
It is of importance to acknowledge there is only one anatomic VIM or DRT.It is unclear if the different tracts or VIM areas constructed using DWI and the Brainlab software coincide with the actual neuroanatomy.Current study only evaluates correlation with tremor control (correlation with anatomy would require combining imaging with postmortem histological analysis).

Implementation of DRT in DBS for ET
Depiction of the DRT-PSA seems to provide a target area for optimal intraoperative tremor control.Implementation during target planning and long-term follow-up are needed for evalu-ation of optimal (stimulation) target selection within, ie, section, within the DRT.4][15] In the current study, we again found this area to show optimal tremor response.For complete alignment of trajectory angle with the DRT, an entry point in the supplementary motor area (SMA) can be needed.However, more anterior angles avoid penetrating the SMA and possibly prevent the occurrence of clinically symptomatic severe edema or hemorrhage.In 2 patients, this resulted in months of severe disturbances in balance, swallowing, and speech.
Considering its reproducibility and correspondence with tremor control, targeting the DRT-PSA in DBS for ET may eliminate the need for intraoperative test stimulation. 16Moreover, occurrence of intraoperative stunning regularly made test stimulation futile.Occurrence of stimulation-induced dysarthria was not a reliable predictor for occurrence postoperative and balance cannot be tested intraoperatively. 13,17his deterministic DRT tractography method and intraoperative 3D imaging may facilitate electrode placement under general anesthesia.However, experience in DBS for tremor, DBS under general anesthesia, and tractography are other essentials. 18

Study Limitations
First, the evaluation of the DRT-PSA was not set up in a prospective manner.Second, only one type of software was used.Third, DRT-based DBS and occurrence of side effects could not be thoroughly analyzed due to the short-term follow-up. 17,19,20ourth, DRT depictions were constructed using ROIs based on literature, and our own experiences, additional combinations, and/or ROIs were not explored.

CONCLUSION
This optimized deterministic DRT tractography strongly correlates with optimal tremor control.This technique is readily implementable for prospective evaluation in DBS target planning for ET.

VOLUME 21 |
NUMBER 6 | DECEMBER 2021 | 533 FIGURE 2. Optimizing deterministic tractography of dentato-rubro-thalamic-tract.A-G show the different possible depictions of DRTs which are used for comparison.The above panel shows an axial, the bottom a coronal view on a 3-Tesla T2 MRI of the thalamic and subthalamic area (commissural aligned imaging).The green DRT is depicted with ROIs in the ipsilateral motor area, ipsilateral PSA, ipsilateral cerebellar peduncle and dentate nucleus (DRT-PSA).The DRT-VIM (purple) is depicted after adding VIM (depicted by the software) as an ROI.The DRT-RN (orange) is depicted with ROIs in ipsilateral motor area, ipsilateral red nucleus, ipsilateral cerebellar peduncle and dentate nucleus (DRT-RN).The crossing DRT-RN (red) crosses the midline at the level of the red nucleus, and is depicted with ROIs in the ipsilateral motor area, and RN, contralateral cerebellar peduncle and dentate nucleus.The definite DBS electrode (yellow) is displayed.A is located 2 mm above the commissural line, B at the commissural line, C is located 2 mm below the commissural line, D is located 4 mm below, E is located 6 mm below, F is located 8 mm below and G is located 10 mm below.Supplemental Videos 1 and 2 show the same DRT depictions in axial and coronal orientation, respectively.
vertical column indicates case numbers.Second column indicates hemisphere of electrode placement, "left" or "right." Third column indicates if the dentato-rubro-thalamictract (DRT) was depicted for DBS surgery (yes or no).Fourth column indicates if crossing fibers of DRT (X indicates crossing fibers) were depicted for DBS surgery.Fifth column indicates if standardized DRT (DRT-PSA) depiction was possible.The fractional anisotropy (FA) applied is stated between brackets.All FA numbers stated are multiplied with 100, facilitating clarity.Sixth until ninth column are similar to the fifth, indicating DRT-VIM, XDRT-PSA, DRT-RN, XDRT-RN, respectively.Tenth column indicates if standard stereotactic VIM coordinates (VIMstx) and VIM as depicted by the stereotactic commercially software were situated in DRT-PSA (VIMsoft).Eleventh column indicates the number of macroelectrode passes (enabling test-stimulation) made before placement of definite DBS electrode.Twelfth column indicates if the DBS electrode was located in DRT-PSA.When not, the distance in millimeters (mm) is stated between brackets.The 13th and 14th columns are similar to the eleventh, for DRT based on the VIM ROI (DRT-VIM) and DRT based on the RN ROI (DRT-RN).Distances to electrode position are not stated (as correspondence with intraoperative tremor control was inferior to DRT-PSA).sides (86%), or 16 of 20 patients (80%).Five patients (25%) experienced gait disturbances.Three patients (15%) experienced stimulation-induced dysarthria.Lowering pulse width significantly reduced gait or speech disturbances in 2 patients.During test stimulation, 4 patients (20%) experienced dysarthria of which one experienced dysarthria during postoperative stimulation.No involuntary muscular contractions were noted during surgery or follow-up.