Article Text
PDF
Abstract
Background Visualisation of the dorsolateral subthalamic nucleus (STN) remains challenging on 1.5 and 3Tesla T2-weighted MRI. Our previously defined hotspot, relative to the well-visualised medial STN border, serves as an MRI landmark for dorsolateral STN identification in deep brain stimulation (DBS). We aimed to validate this hotspot in a separate trial cohort of Parkinson’s disease (PD) patients and refine its location.
Methods In this post hoc analysis of a randomised controlled trial, in which the hotspot was taken into account during target planning, responses to DBS were evaluated using hemibody improvement on the Movement Disorder Society–Unified Parkinson’s Disease Rating Scale motor examination and compared with our historical cohort, as well as dopaminergic medication reduction. Then, a refined hotspot was calculated and the Euclidean distance from individual active contacts to the refined hotspot was correlated with motor improvement.
Results The first quartile of the hemibodies (poor responders) showed an average improvement of 13%, which was higher than the —8% in the historical control group (p=0.044). Dopaminergic medication reduction was greater in the current cohort compared with the historical cohort (p=0.020). Overall variability of hemibody motor improvement was reduced in the current cohort compared with the historical control group (p=0.003). Motor improvement correlated to the Euclidean distance from active contact to the refined hotspot (2.8 mm lateral, 1.1 mm anterior and 2.2 mm superior to the medial STN border) (p=0.001).
Conclusion We validated the hotspot for dorsolateral STN targeting in DBS for patients with PD and showed an improved motor response in poor responders, a reduced variability in motor improvement and a greater dopaminergic medication reduction. We then refined the hotspot at 2.8 mm lateral, 1.1 mm anterior and 2.2 mm superior relative to the medial STN border, which visualises a readily implementable target within the dorsolateral STN on lower field strength MRI.
- PARKINSON'S DISEASE
- NEUROSURGERY
- CLINICAL NEUROLOGY
- ELECTRICAL STIMULATION
- MRI
Data availability statement
Data are available on reasonable request.
This is an open access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited, appropriate credit is given, any changes made indicated, and the use is non-commercial. See: http://creativecommons.org/licenses/by-nc/4.0/.
Statistics from Altmetric.com
WHAT IS ALREADY KNOWN ON THIS TOPIC
Ultrahigh field MRI is not widely available and clear visualisation of the dorsolateral border of the subthalamic nucleus (STN) remains challenging on 1.5T and 3T MRI. We defined an optimal theoretic deep brain stimulation (DBS) location in the dorsolateral STN, relative to a patient-specific reference point that is well visualised on lower field strength MRI.
WHAT THIS STUDY ADDS
We validated this hotspot for dorsolateral STN targeting in DBS for patients with Parkinson’s disease and showed an improved motor response in poor responders, a reduced variability in motor improvement, and a greater dopaminergic medication reduction after implementation of the hotspot.
HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY
This hotspot enables dorsolateral STN targeting on lower field strength MRI and is readily implementable worldwide.
Introduction
Deep brain stimulation (DBS) of the dorsolateral sensorimotor part of the subthalamic nucleus (STN) is effective for reducing motor symptoms (eg, tremor, rigidity, bradykinesia) in patients with advanced Parkinson’s disease (PD).1–10 The dorsolateral STN became visible in recent years by using 7 Tesla (T) T2-weighted MRI.11 12 However, ultrahigh field MRI is not widely available and clear visualisation of the dorsolateral border of the STN remains challenging on 1.5T and 3T MRI due to (a) the small size and oblique orientation of this nucleus13 and (b) the poor delineation of the ventrolateral border of the STN with the substantia nigra (figure 1).14 15 Although specialised MRI sequences (eg, susceptibility-weighted imaging and quantitative susceptibility mapping) have improved direct STN visualisation, various practical limitations, limited validation and geometric distortion issues have been noted.15 Routine, non-optimised, T2-weighted imaging remains the most commonly used sequence, with a well-validated 3T MRI protocol for visualising the STN.14 16 A readily implementable MRI landmark that helps to identify the dorsolateral STN on routine T2-weighted images is, therefore, essential for DBS targeting and may contribute in improving motor symptoms, in addition to reducing individual variability in motor improvement in PD patients undergoing STN DBS.4 8 10 17 18
Axial and coronal 3T and 7T T2-weighted MRI. (m)STN, (medial) subthalamic nucleus; SNr, substantia nigra.
In previous work, we defined an optimal theoretic DBS location in the dorsolateral STN relative to a patient-specific reference point that is well visualised on 1.5T and 3T T2-weighted MRI: the medial STN border. This point is readily identifiable at the intersection of the medial STN border with the line drawn at the anterior border of the red nucleus (ie, the Bejjani line13) at its biggest diameter on the axial MRI plane, perpendicular to the anterior commissure-posterior commissure (AC-PC) line.19
Compared with the traditional midcommissural reference point (MCP)-based STN targeting, the medial STN border enabled visualising a target (or hotspot) within the dorsolateral STN which showed significant correlation with motor outcome.19 Implementation of the medial STN border as a patient-specific reference point has in recent years been done across different centres, confirming its generalised applicability in STN DBS.20 21
The previously defined hotspot was localised 2.8 mm lateral, 1.7 mm anterior and 2.5 mm superior to the medial STN border. In our centre, this hotspot has since then been applied for target planning in the STN DBS procedure.
In the current study, we prospectively evaluated the implementation of this hotspot relative to the medial STN border as a patient-specific reference point for DBS targeting and analysed its consistency for correlation with lateralised motor improvement, in a recent randomised controlled trial (RCT) cohort of patients with PD who underwent STN DBS surgery at our institution. We aimed that implementation of the hotspot in dorsolateral STN targeting would be associated with an increase in motor improvement and a decrease in individual variability of motor symptoms. Also, we used the current cohort to further refine the theoretic hotspot within the dorsolateral STN.
Methods
Patients
Patients were selected from a single-centre RCT comparing bilateral STN DBS surgery under general versus local anaesthesia, that included patients with PD between May 2015 and March 2019.22 23 Motor symptoms were scored in standardised conditions according to the Movement Disorder Society–Unified Parkinson’s Disease Rating Scale Motor Examination (Part III) (MDS-UPDRS ME)24 by a specialised research nurse. For the current analysis, the OFF medication scores at baseline were compared with the OFF medication/ON stimulation scores 6 months after surgery. Dopaminergic medication was adjusted according to clinical needs, and levodopa equivalent daily dose (LEDD) was recorded at every follow-up appointment.
Surgical procedure
Patients underwent DBS electrode implantation under local anaesthesia (awake) or general anaesthesia (asleep). The implantation was frame-based (either the Leksell Coordinate Frame G or the Leksell Vantage system; Elekta AB, Stockholm, Sweden) and intraoperative cone-beam CT (iCBCT; O-arm O2 Imaging System, Medtronic, Dublin, Ireland) was used for stereotactic registration and intraoperative electrode localisation. One to three parallel tracks for microelectrode recordings (MERs) were used to guide electrode placement by identifying the electrophysiological dorsal STN border in all patients, both in surgery under general anaesthesia (asleep) and under local anaesthesia (awake). In awake cases, macrostimulation was performed along the trajectory with the longest track or most powerful signal of STN activity, favouring the central trajectory in case of comparable results, testing for symptom relief and adverse effects of stimulation. When exploration showed a good therapeutic effect and a high threshold for adverse effects, the definitive electrode was placed.25 In asleep cases, propofol was paused for 20 min for STN activity to return, and cessation lasted maximally 45 min while high-dose remifentanil was continued under careful observation of vital parameters and possible patient movements to ensure that wake-up did not occur. The definitive electrode was placed in the track with the longest section or most powerful signal of STN activity without macrostimulation, also favouring the central trajectory in case of comparable results. Final electrode placement was done using electrodes with four stimulation levels (Vercise Cartesia Directional lead, Boston Scientific, Marlborough, Massachusetts, USA) and a subcutaneous or subpectoral, infraclavicular pulse generator (Vercise PC, Boston Scientific) was implanted in the same surgical session.
Image acquisition
Preoperative sequences and parameters used with the 3T MRI (3T Elition Ingenia, Philips, Best, The Netherlands; equipped with a 32-channel receive coil (Philips)) were 3D sagittal T1-weighted magnetisation-prepared rapid acquisition gradient echo sequences with contrast enhancement; slice thickness 0.9 mm, FOV 256×256×170 mm, flip angle 8°, voxel size 0.5×0.5×0.9 mm, repetition time (TR) 8.8 ms, echo time (TE) 4.0 ms, echo train length (ETL) 242 and acquisition time 4.11 min. For T2-weighted 3D Turbo spin echo (TSE) sequences, the parameters were: slice thickness 0.6 mm, FOV 250×250×180 mm, flip angle 90°, voxel size 0.56×0.56×0.56 mm, TR 2500 ms, TE 231 ms, ETL 133 and acquisition time 5.27 mins.
On availability, some patients voluntarily underwent 7T MRI (7T Achieva system, Philips, Best, The Netherlands; equipped with a 32-channel receive coil (Nova Medical, Wilmington, Massachusetts, USA)): 3D T2-weighted TSE (slice thickness 0.7 mm, TR 3000 ms, TE 324 ms, ETL 182, FOV 250×250×190 mm, flip angle 100°, voxel size 0.7×0.7×0.7 mm and acquisition time 7 mins). Eligible patients were those without any metal implants regardless of compatibility with lower field-strength MRI (ie, 1.5T, 3T).
For stereotactic registration, the iCBC was used in the ‘stereotactic mode’ of the O-arm (40 cm FOV, 192 slices, 120 kV and 150 mA). After placement of both electrodes, a second iCBCT in ‘high-definition mode’ (20 cm FOV, 192 slices, 120 kV and 150 mA) was performed for electrode localisation and accuracy evaluation.
Hotspot-based target planning
Target planning of the STN was performed using AC-PC aligned axial and coronal orientated T2-weighted MRI. The maximum diameter of the red nucleus and the Bejjani line was used during target planning to determine the target central within the hypointense STN at a level of 820 and a width range of 1080–1180 for optimal resolution window. Angle adjustments were made to maximise the length of the trajectory within the dorsal part of the STN. The previously described hotspot (2.8 mm lateral, 1.7 mm anterior and 2.5 mm superior to the medial STN border19) was taken into account during planning (eg, figure 2). Other trajectory requirements were a gyral prefrontal entry point with avoidance of blood vessels, caudate nuclei and ventricles. The DBS planning software used was SurgiPlan (Elekta AB, Stockholm, Sweden) and since October 2017 Brainlab Origin (Brainlab AG, Munich, Germany).
Workflow hotspot-based STN targeting on 3T T2-weighted MRI. Step 1: Right STN targeting (orange line) using Bejjani line, at lateral, posterior, inferior relative to MCP with lateral angle degrees and anterior angle 30°. Step 2: Hotspot (green dot) at lateral, 1.1 anterior and 2.2 superior to the medial STN border. Step 3: Trajectory alignment with hotspot using lateral angle adjustment to 26°. MCP, midcommissural point; STN, subthalamic nucleus.
The hotspot validation and further refinement
The MDS-UPDRS ME scores at baseline and after 6 months follow-up were used to calculate percentual change in lateralised (hemibody) OFF medication and OFF medication/ON stimulation scores. Hemibody sub scores were calculated from the MDS-UPDRS ME items on rigidity (3.3), bradykinesia (3.4–3.8) and tremor (3.15–3.18). Subsequently, hemibodies were categorised into three groups, based on percentual MDS-UPDRS ME improvement: (1) non-responding (<30%), (2) responding (between 30% and 70%) and (3) optimally responding (>70%). To assess the effect of hotspot implementation, these groups were compared with historical control groups from our previous study.19
For the evaluation of consistency for correlation with lateralised MDS-UPDRS ME scores, we first determined the medial STN border, then a corresponding hotspot based on this cohort was calculated and finally the Euclidean distance between individual active contacts and the refined hotspot was correlated with the lateralised MDS-UPDRS ME scores.
The medial STN border coordinates were determined in left and right mesencephalic area as previously described.19 In brief, the maximum diameter of the red nucleus in the axial plane was determined using both axial and coronal-orientated AC-PC aligned T2-weighted images. Subsequently, a line perpendicular to the AC-PC line was drawn, coinciding with the anterior border of the red nucleus along its trajectory. The point of intersection with the medial STN border (medial border of hypointense crescent area lateral to the red nucleus) was then determined.
Active DBS contacts were located on CT by using contact distance measurements and visual inspection of electrode contact artefacts (figure 3). In case of bipolar or double monopolar stimulation, stereotactic coordinates of the midpoint between the active contacts were chosen as centre of stimulation.
Electrode contact visualisation (Vercise Cartesia Directional lead, Boston Scientific) using inline planes. The level of the metal artefacts coincide with the electrode contact points.
Similar to the previously described methods,19 a theoretic ‘hotspot’ was calculated by averaging X, Y and Z active contact coordinates of the ‘optimally responding’ group, relative to the medial STN border as well as to the MCP. After that, the Euclidean distance, that is, 
, from individual active contacts to both defined hotspots were calculated separately and were finally correlated with motor improvement of corresponding hemibodies.
Statistical analyses
Numerical data are presented as mean with SD or median with IQR after testing for normality (Shapiro-Wilk >0.9). The UPDRS26 ME scores of the historical control cohort were converted to MDS-UPDRS ME equivalents using a validated formula.27 Bootstrap for independent samples t-test was used to compare quartile means with SEM of outcome variables between the current cohort and the historical control group. The χ2 was used to analyse differences in (sub)group proportions between the current cohort and the historical control group. Differences in means between (sub)groups were analysed using the independent samples t-test. Variability of the outcome measure in the current cohort was compared with the historical control group using Levene’s test for equality of variance. Correlations were analysed using Pearson’s two-tailed correlation coefficient. SPSS Statistics V.28 (IBM) was used for all statistical analyses. A p<0.05 was considered statistically significant.
Results
Patient characteristics
In total, 218 DBS electrodes were implanted in 109 patients with PD. The average age at surgery was 61±8 years and 31 patients were female (28%). Average disease duration of 11±5 years and all patients underwent bilateral DBS electrode placement. Two patients were lost to follow-up (one retracted participation and one dieddue to unrelated cause prior to follow-up assessment). An overview of patient characteristics is published elsewhere.23
One patient was excluded from the current analysis due to missing OFF medication evaluation 6 months after surgery (this patient refused this assessment). As a result, 106 patients were included in the current analysis, representing 212 DBS electrodes and corresponding hemibodies. The average time from baseline OFF medication assessment to surgery was 2.7±1.5 months. Follow-up OFF medication/ON stimulation scoring was performed 9.3±1.9 months on average after baseline OFF medication assessment. The mean LEDD at baseline was 1542±503 mg and was lowered to 631±316 mg on average at follow-up after DBS. An overview of MDS-UPDRS ME scores at baseline and follow-up is presented in table 1.
Motor examination scores
Reduction of LEDD at follow-up was significantly higher in the current study (821±408 mg) compared with the historical control cohort (610±547; independent t(136)=2.36, p=0.020). The average improvement of motor symptoms, reflected by the change from baseline to follow-up MDS-UPDRS ME score during the OFF medication state, was not different between the two cohorts: −26.2±16 (50.5% improvement) in the current study vs −22.2±14.3 (40.3% improvement) in the historical control group (independent t(136)=1.26, p=0.209). No difference in MDS-UPDRS ME improvement in the ON medication state was found between the current study (−3.2±9.3 score change or 9.9% improvement) and the historical control group (−2.4±7.5 score change or 6.2% improvement; independent t(136) = 0.46, p=0.649).
Per group MDS-UPDRS improvement
37 body sides were categorised as non-responding hemibodies (17%), 108 as responding hemibodies (51%) and 67 as optimally responding hemibodies (32%). Overall, 63% of the DBS leads were placed in the central MER-trajectory. Others were placed using the anterior (26%), lateral (7.7%), medial (1.9%) or posterior (1.4%) trajectory. Monopolar stimulation was performed in all except seven electrodes (97%) at 6 months follow-up. Stimulation rates were set at 130 Hz (in 17 electrodes the rates were adapted with a range from 71 to 185 Hz), and a pulse width of 60 µs was used in all except 3 electrodes (80 or 90 µs). Current controlled stimulation amplitudes ranged between 0.9 and 3.1 mA, with a median of 2.2±0.6 mA. Reasons for adapting stimulation rates, pulse width or amplitude were either related to side effects or optimisation of therapeutic effect. Between non-responding body sides and optimally responding body sides, no statistically significant differences in stimulation rates and amplitude settings were found (non-parametric independent t(2)=0.219, p=0.639 and 0.104, p=0.747, respectively).
55 (52%) of the included patients underwent the DBS procedure awake vs 51 (48%) patients asleep, with no statistical difference in proportion between the hemibody subgroups of motor outcome (χ2 (2, N=212)=0.831, p=0.660). Baseline levodopa responsiveness for optimally responding body sides was significantly higher compared with non-responding hemibodies (75% vs 54%; independent t(102)=−6.47, p<0.001) and responding body sides (75% vs 60%; independent t(173)=−5.48, p<0.001). No statistical difference in proportions of preoperative target planning using 3T MRI (150/212 or 71%) vs 7T MRI (62/212 or 29%) was found between the hemibody subgroups of motor outcome (χ2 (2, N=212)=4.443, p=0.108).
Comparing the outcome categories in the current study to our historical control group, no statistically significant difference in proportions was found between non-responding hemibodies: 17% vs 26%; responding hemibodies: 51% vs 46%; and optimally responding hemibodies: 32% vs 28% (χ2 (2, N=277)=2.409, p=0.300). When comparing the quartile means of the percentage hemibody motor improvement between the current study and the historical control, a statistical difference was found for the first quartile only, favouring the current cohort (table 2).
Hemibody subgroup proportions and average motor improvement
Overall variability of hemibody motor improvement as measured by statistical variance was reduced significantly in the current cohort compared with the historical control group (Levene’s test=8.926, p=0.003).
The medial STN border and hotspot refinement
Stereotactic coordinates of medial STN borders relative to the MCP varied considerably, with a range of 6.3–12.8 mm lateral (mean 8.9±1.1 mm), 0.3–5.2 mm posterior (mean −2.3±0.7 mm) and 3.5–6.5 mm inferior (mean −5.0±0.6 mm). The theoretic stimulation hotspot (ie, mean stereotactic coordinates of active contacts of the optimally responding hemibody group) was found at 2.8 mm lateral, 1.1 mm anterior and 2.2 mm superior relative to the medial STN border, and 11.8 mm lateral, 1.2 mm posterior and 2.7 mm inferior relative to the MCP. An overview of mean stereotactic coordinates of the active contacts is presented in table 3, categorised by non-responding, responding and optimally responding hemibodies.
Mean stereotactic coordinates of active contacts
The mean Euclidean distance of the non-responders to the refined hotspot was significantly different from the mean distance to that hotspot of both the responders and the optimal responders. No statistically significant difference between mean Euclidean distances of responders and optimal responders was found. This was found using both the medial STN border and the traditional MCP as a reference point. Corresponding boxplots are shown in figure 4.
(A) Euclidean distance from active contacts to the refined ‘hotspot’ defined relative to the medial STN border (ie, 2.8 mm lateral, 1.1 mm anterior, 2.2 mm superior), categorised by non-responding (<30% MDS-UPDRS ME change), responding (30%–70% MDS-UPDRS ME change) and optimally responding (>70% MDS-UPDRS ME change) corresponding hemibodies. Average Euclidean distances between non-responders and responders were statistically significant (independent t(143)=2.25, p=0.026*), as well as between non-responders and optimal responders (independent t(102)=3.22, p=0.002**). (B) Euclidean distance from active contacts to the refined ‘hotspot’ defined relative to the MCP (ie, 11.8 mm lateral, 1.2 mm posterior, 2.7 mm inferior), categorised by non-responding, responding and optimally responding corresponding hemibodies. Average Euclidean distances between non-responders and responders were statistically significant (independent t(143)=2.13, p=0.035 as well as between non-responders and optimal responders (independent t(102)=2.74, p=0.007**). Outliers were not depicted in both graphs. MCP, midcommissural point; MDS-UPDRS ME, Movement Disorder Society–Sponsored revision of the Unified Parkinson’s Disease Rating Scale Motor Examination; STN, subthalamic nucleus.
The correlation between refined hotspot and motor improvement
Both for medial STN border and MCP, the Euclidean distance from the theoretic stimulation hotspot to individual active contacts was negatively correlated with motor improvement of corresponding hemibodies (p=0.001, figure 5). Similar results were found when adjusting for baseline levodopa responsiveness (Pearson’s correlation=−0.17; p=0.016 when using the patient-specific medial STN border as reference point and Pearson’s correlation=0.17; p=0.016 when using the traditional MCP). The significant results sustained after removal of univariate outliers.
(A) Significant negative correlation (Pearson’s correlation −0.23; p=0.001) between percentual MDS-UPDRS ME hemibody score change and Euclidean distance from active contact to the refined ‘hotspot’ defined relative to the medial STN border (ie, 2.8 mm lateral, 1.1 mm anterior, 2.2 mm superior). (B) Significant negative correlation (Pearson’s correlation −0.19; p=0.007) between percentual MDS-UPDRS ME hemibody score change and Euclidean distance from active contact to the refined ‘hotspot’ defined relative to the MCP (ie 11.8 mm lateral, 1.2 mm posterior, 2.7 mm inferior). MCP, midcommissural point; MDS-UPDRS ME, Movement Disorder Society–Sponsored revision of the Unified Parkinson’s Disease Rating Scale Motor Examination; STN, subthalamic nucleus.
Discussion
In current study, we evaluated the implementation of the hotspot for dorsolateral STN targeting in DBS for PD. The average motor response of the poorly responding hemibodies to DBS increased compared with our historical cohort,19 and overall variability of hemibody motor improvement was reduced significantly. Lowering the proportion of poor responders in DBS for PD and reducing variability in motor improvement after DBS is key in discussing the expectations of the therapy in the outpatient clinic.
The overall motor improvement of 50.5% in the current study is similar to the estimated mean STN DBS motor improvement at study level of 49.6% (95% CI 45.6% to 53.6%) published in a recent meta-analysis.28
The refined hotspot is located slightly more posteriorly (0.6 mm compared with the previous hotspot published in 2018) located in the dorsolateral STN. The central channel often showed a long trajectory of typical STN activity (in both asleep and awake surgery), and a more optimal therapeutic window during awake test stimulation (data are not shown) than the anterior channel. Other factors contributing to more central channel selection are optimisation of the DBS target by implementation of the 2018 hotspot, asleep surgery without test stimulation for trajectory selection, and 7T MRI with optimised dorsolateral STN depiction. From an anatomical perspective, a more posterior location situates the hotspot closer to the subdivision with most projections to cortical motor areas. In the current cohort, the use of 7T MRI vs 3T MRI for preoperative target planning was not identified as a confounding factor for motor improvement after implementation of the hotspot. Investigating the clinical outcome of STN DBS due to 7T MRI implementation, however, did not lay within the scope of the current study. Future research remains warranted to study whether clinical outcome differences are found between surgical planning using 7T and 3T MRI. Previously, Patil et al 14 found evidence that 3T MRI sufficiently and accurately visualise STN measures in three dimensions that corresponded very well with STN dimensions defined by large cadaveric studies. Therefore, the clinical relevance of improved demarcation between the STN and the substantia nigra remains a subject of debate. Other imaging advancements, such as connectivity-derived segmentation of the STN using 7T MRI, may contribute to improving clinical outcome.
We correlated the distance between the point of stimulation and the refined hotspot with clinical outcome. In line with the 2018 study, a negative correlation between the three-dimensional distance from individual active contacts to the refined theoretic hotspot and motor improvement of corresponding hemibodies was found using the medial STN border as a reference point. In contrast to the 2018 study, this correlation was also found using the MCP as a reference point. Nonetheless, the refined hotspot based on the medial STN border showed a stronger correlation with motor improvement with less variability, compared with the MCP. Overall, these findings support our hypothesis that clinical effectiveness is optimised by employment of the medial STN border as a patient-specific reference point in stereotactic STN DBS surgery.
Compared with previous studies investigating alternative targeting methods from AC-PC-based targeting, our hotspot locates in proximity of their optimal contact positions. Andrade-Souza et al 29 used the red nucleus as an internal fiducial on MRI for targeting the STN to approach their optimal stimulation location. This was located 12.12 mm lateral, 2.41 posterior and 2.39 mm inferior relative to the MCP of 28 DBS electrodes, compared with 11.8 mm lateral, 1.2 mm posterior and 2.7 mm inferior relative to MCP in our current study. More recently, Conrad et al 30 localised their optimal point of stimulation at 11.75 lateral, 1.84 mm posterior and 1.08 mm ventral to the MCP, using a computational electrical field model that integrated anatomical, clinical and electrophysiological information of 20 patients (38 DBS electrodes). Although this is slightly different from our results in the y-direction and z-direction, the optimal location for stimulation seems consistently dorsolateral within the STN, presumably the sensorimotor part. These small differences might be explained by the relatively small sample sizes. Houshmand et al 16 compared RN-based targeting with traditional MCP-based targeting and found that RN-based targeting was more accurate and precise in STN midpoint targeting. As can be expected, average STN midpoint coordinates of 58 patients were found medioventrally relative to our hotspot.
Limitations
Our study has several limitations. Although the employment of the hotspot was performed in a prospective manner, it should be noted that the final determination of electrode placement was based on MER. Also, for each implanted electrode, the centre of stimulation was based on the position of the active electrode as visualised on iCBCT without correction for possible intraoperative brain shift. In previous work, we confirmed that stereotactic MRI and O-arm iCBCT produce comparable coordinates in the stereotactic space without clinically relevant differences in electrode localisation.31 In order to minimise the error due to brain shift, a fibrin glue was applied in the burr hole and a gyral entry point was used. Although the electrodes were configured in ring mode (as opposed to field steering) at the time of follow-up assessment, it is possible that in some cases the centre of stimulation slightly shifted from the centre of the active electrode as this is conditional on the stimulation rate, pulse width and amplitudes. Future prospective implementation of the hotspot and its evaluation regarding STN DBS outcomes should, therefore, potentially calculate the three-dimensional distance between the individual volume of tissue activated (VTA) and the hotspot. Recently, a VTA-based biomarker was proposed for selecting DBS implantation trajectories by Rao et al that correlates well with motor outcomes.32 Although prospective validation of this biomarker is pending, it shows potential to reduce the length of the surgical STN DBS procedure and postoperative programming.
Our subgroups based on motor improvement (<30%, 30%–70% and >70% improvement on de MDS-UPDRS ME) are clinically relevant, yet statistically arbitrary. In addition, the historical control cohort evaluated motor outcome using the original UPDRS. No reliable formulas to convert UPDRS III subscores, that is, hemibody scores or tremor items, into MDS-UPDRS equivalents are available to date.27 33 We, therefore, compared the percentage hemibody improvement between the groups. The use of quartile means with bootstrapping allowed statistical comparison between the current cohort and our historical control group. Mean OFF medication/OFF stimulation scores were not available for the historical cohort, so a possible change in baseline motor score after electrode implantation could not be compared with the current study population.
The current study did not comprise neuropsychological outcomes or side effect, and all motor outcomes were only obtained relatively short term. Further studies are needed to explore the sustainability of our findings at long-term follow-up, including neuropsychological outcomes and side effects.
Conclusion
We validated the hotspot for dorsolateral STN targeting in DBS for patients with PD and showed an improved motor response in poor responders, a reduced variability in motor improvement and a greater dopaminergic medication reduction after implementation based on this hotspot. We refined the hotspot to 2.8 mm lateral, 1.1 mm anterior and 2.2 mm superior to the medial STN border. This hotspot enables dorsolateral STN targeting on lower field strength MRI and is readily implementable.
Data availability statement
Data are available on reasonable request.
Ethics statements
Patient consent for publication
References
Footnotes
Contributors EB: conceptualisation, methodology, acquisition of data, formal analysis, visualisation, writing–original draft, editing and final approving. RH: acquisition of data, writing–review, editing and final approving. RMADB: acquisition of data, writing–review and final approving. MB: acquisition of data, writing–review and final approving. PvdM: conceptualisation, methodology, acquisition of data, visualisation, writing–review, editing and final approving. RS: conceptualisation, methodology, acquisition of data, writing–review, editing and final approving. MB: guarantor, conceptualisation, methodology, acquisition of data, visualisation, writing–review, editing and final approving.
Funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.
Competing interests RH reported grants from Dutch Brain Foundation. RMADB reported institutional grants from Hersenstichting Charitable Organisation, Netherlands Organisation for Health Research and Development, Stichting Parkinson Nederland, GE Healthcare, Medtronic, Lysosomal Therapeutics and Neuroderm. MB received funding from the Amsterdam UMC TKI-PPP grant (2021 call), the EU Joint Programme–Neurodegenerative Disease Research (JPND) project (2021 call), stichting ParkinsonFonds, Medtronic, all paid to institution and outside the submitted work. RS reported personal fees from Medtronic and Boston Scientific. No other disclosures were reported.
Provenance and peer review Not commissioned; externally peer reviewed.