Targeting Accuracy of the Leksell Vantage Stereotactic System for Deep Brain Stimulation Surgery: A Retrospective Review

Abstract

BACKGROUND: 

Compared with the Leksell G frame, the new Leksell Vantage stereotactic headframe offers many benefits during deep brain stimulation (DBS) procedures.

OBJECTIVE: 

To define the “real-world” targeting accuracy of the Vantage frame for performing DBS lead implants.

METHODS: 

Retrospective review of all patients who underwent DBS using the Vantage frame with our surgeon between July 2021 and May 2022. Target and actual coordinates of implanted DBS leads were determined in BrainLab.

RESULTS: 

Thirty-three electrode trajectories (17 left-sided, 16 right-sided) were included. On the right, actual electrode placement was, on average, 0.95 mm more medial, 1.06 mm more posterior, and 0.55 mm more superior compared with planned trajectories. The vector error was 1.93 ± 0.91 mm. On the left, actual electrode placement was 0.49 mm more lateral, 0.66 mm more posterior, and 0.67 mm more superior than planned trajectories. The vector error was 1.68 ± 0.66 mm. After excluding all electrodes that were repositioned after microelectrode recording and/or test stimulation, right-sided electrodes (n = 10) were 0.88 mm more medial, 0.94 mm more posterior, and 0.76 mm more superior than planned trajectories. The vector error was 1.74 ± 1.01 mm. Left-sided electrodes (n = 14) were 0.49 mm more lateral, 0.63 mm more posterior, and 0.49 mm more superior than planned trajectories. The vector error was 1.62 ± 0.68 mm.

CONCLUSION: 

This study demonstrated the accuracy of the Leksell Vantage frame in comparison with the traditional Leksell G frame.

ABBREVIATIONS:

ANT

anterior thalamus

GPi

globus pallidus internus

PD

Parkinson's disease

STN

subthalamic nucleus

VIM

ventral intermediate nucleus.

The success of deep brain stimulation (DBS) surgery is dependent on accurate implantation of the stimulating lead(s) within the therapeutic target. Potential sources of suboptimal targeting include magnetic resonance image distortion, imprecise registration of targeting fiducials, imprecise image fusion, operator error, and the mechanical accuracy of the targeting system used.

For a generation, the Leksell Coordinate G Frame (Elekta Instruments) was a mainstay for performing frame-based stereotaxis; however, in 2019, Elekta introduced the Vantage frame, which features a number of notable improvements. Perhaps most significant is that the frame itself is made of a nonmetallic, glass fiber–reinforced epoxy composite material that limits both MRI and computed tomography (CT) artifact, purportedly enhancing the fidelity of preoperative imaging.¹ The Vantage frame also lacks the G frame's front cross-bar, which blocks the patient's face, providing easier airway access in the event of an emergency² and allowing for better visualization of the patient's face, which can be beneficial during awake surgeries. The elimination of the front bar also increases patient comfort, as does the elimination of the G frame's notoriously uncomfortable ear bars, which are used during frame application.¹

To date, little has been written about the “real-world” performance of the Vantage frame, and so we report our initial experience in using the Vantage frame for DBS surgery, focusing on its targeting accuracy in our hands.

METHODS

Study Population

This retrospective study was performed with the approval of our Institutional Review Board (Study Number: 2022P000366). The need for patient consent was waived. Billing records were used to identify all consecutive patients who underwent DBS lead implant surgery for any diagnosis and at any target between July 2021 and May 2022. Operative reports were referenced to confirm the use of the Leksell Vantage frame. Any patient who underwent surgery using the Leksell G frame was excluded. Two patients whose intended targeting coordinates were unavailable were also excluded (Figure 1).

FIGURE 1.: Flowchart of patient selection. Twenty-four patients who collectively received 43 electrode implants were identified as having undergone DBS surgery with our surgeon between July 2021 and May 2022. Three patients were excluded for use of the Leksell G frame, 2 patients were excluded for lack of preoperative trajectory coordinates, and 1 patient underwent DBS initially with the STarFix system but ultimately required revision surgery with the Vantage frame and was therefore included in our cohort. Overall, 19 patients who received 33 electrode implants were included in the study. DBS, deep brain stimulation.

Surgical Technique

The senior author's technique for performing DBS surgery has been detailed elsewhere.^3,4 In brief, he uses a frame-based, MRI-guided technique for initial anatomic targeting. Microelectrode recording (one pass at a time) is used to confirm proper targeting only of the subthalamic nucleus (STN). Macroelectrode stimulation is performed on all patients and at all targets as a quality control measure. Most patients are operated on while awake; sedation is used at the patient's request. To minimize cerebrospinal fluid losses, the dural opening is only large enough to accommodate the electrode guide tube. A sharp corticectomy is always performed to ease insertion of the cannula within the brain. After closure, and before leaving the operating room, a stereotactic CT is performed (Airo, BrainLab) to check for both intracerebral hemorrhage and proper lead placement. The latter is accomplished by fusing the CT to the preoperative MRI containing the operative plan and using automated lead detection software (BrainLab, Inc) to check the position of the lead(s) relative to the planned trajectories.

Study Design

Relevant clinical parameters and operative information were gleaned from the electronic medical record. Targeting data were extracted from the targeting software used to perform the surgeries (BrainLab, Inc). This included the coordinates relative to the intercommissural plane of the intended target for each electrode implant and the actual coordinates of the ventral most contact of the implanted lead as determined using the lead detection software. Intended and actual coordinates of all electrodes were then compared, and the Euclidean distance (vector error) between the 2 was calculated using the equation: $\sqrt{∆x^2}+∆y^2+∆z^2$. To assess the accuracy of the frame in isolation, we repeated the analysis after eliminating those electrodes that were adjusted from the original targeting plan based on physiological recordings or responses to stimulation. Finally, the vector errors associated with electrodes implanted first or second during bilateral procedures were compared to examine the possible effect of brain shift on targeting accuracy. Unilateral electrodes were considered first-side implants.

Statistical Analyses

Stata version 14.0 (StataCorp) was used to perform all analyses. Welch's t-tests were used to compare vector error between first and second implants. P values of ≤ .05 were considered significant.

RESULTS

Patient Characteristics

Overall, 19 patients who underwent a total of 33 electrode implants were included in this analysis (Figure 1). Of these, 14 had Parkinson's disease (PD), 3 had essential tremor, 1 had refractory epilepsy, and 1 had laryngeal dystonia. The therapeutic target was the STN for 10 patients (all with PD), globus pallidus internus for 4 patients (PD), ventral intermediate nucleus for 4 patients (essential tremor and laryngeal dystonia), and the anterior thalamus for 1 patient (epilepsy).

Fourteen patients received bilateral, and 5 received unilateral implants. Two of the patients who received bilateral globus pallidus internus implants underwent staged surgeries, meaning that their second electrode was implanted roughly 1 month after the first electrode, and new preoperative imaging was obtained. Of the 5 patients who received unilateral implants, 2 underwent surgery to reposition previously implanted leads and 1 underwent surgery to replace an infected electrode that had previously been removed.

Electrode Coordinates

Of the 33 lead implants included in this study, 17 were left-sided and 16 were right-sided. On the right side, actual electrode placement was, on average, 0.95 ± 0.78 mm more medial, 1.06 ± 0.81 mm more posterior, and 0.55 ± 1.03 mm more superior than the planned trajectories. The mean three-dimensional vector error was 1.93 ± 0.91 mm (Figure 2A). On the left side, actual electrode placement was 0.49 ± 0.92 mm more lateral, 0.66 ± 0.72 mm more posterior, and 0.67 ± 0.93 mm more superior on average compared with planned trajectories, and the mean vector error was 1.68 ± 0.66 mm (Figure 2B and Table 1).

FIGURE 2.: Targeting error of A, right-sided and B, left-sided electrodes. Targeting error (ie, the difference between intended and actual placement of the ventral most electrode contact) in the xy plane of A, right-sided and B, left-sided electrodes. Electrodes that were repositioned after microelectrode recording and/or test stimulation are highlighted in blue, whereas electrodes that were not repositioned are highlighted in red.

TABLE 1. - Difference Between Intended and Actual Coordinates for All Electrodes

Patient	Target	Side	Implant order	Difference between intended and actual coordinates
				Lateral right (mm)	Anterior (mm)	Inferior (mm)	Vector distance (mm)	Radial error (mm)
1	VIM	Left	1st	−0.03	−1.81	−0.30	1.83	1.81
2	GPi	Left	1st	−1.16	−1.24	−2.81	3.28	1.70
3	VIM	Left	1st	0.04	−1.04	−0.57	1.19	1.04
4	STN	Left	2nd	−1.41	−0.15	0.51	1.51	1.42
5	STN	Left	1st	−0.57	−1.13	−1.05	1.64	1.27
6	STN	Left	1st	−0.43	−1.29	−0.94	1.65	1.36
7	GPi	Left	1st	−2.55	0.09	−0.29	2.57	2.55
8	STN	Left	1st	−0.51	0.08	−2.53	2.58	0.52
9	STN	Left	2nd	−0.99	−0.25	−0.69	1.23	1.02
10	VIM	Left	1st	−0.30	−1.57	−0.49	1.67	1.60
11	GPi	Left	1st	−0.51	0.69	−0.69	1.10	0.86
12	VIM	Left	1st	−0.72	0.12	−0.29	0.79	0.73
13	ANT	Left	2nd	1.46	−0.39	−1.16	1.91	1.51
14	STN	Left	1st	−0.81	−1.30	0.73	1.70	1.53
15	GPi	Left	1st	−1.21	−1.29	−0.85	1.96	1.77
16	STN	Left	2nd	0.75	−0.25	−0.58	0.98	0.79
17	STN	Left	1st	0.64	−0.48	0.59	0.99	0.80
			Average	−0.49 ± 0.92	−0.66 ± 0.72	−0.67 ± 0.93	1.68 ± 0.66	1.31 ± 0.51
18	VIM	Right	2nd	−2.08	−2.63	0.86	3.46	3.35
19	GPi	Right	2nd	−2.52	−2.20	−2.39	4.11	3.35
20	VIM	Right	2nd	−0.88	−0.48	−1.03	1.44	1.00
21	STN	Right	1st	−0.27	−0.63	0.44	0.81	0.69
22	STN	Right	2nd	−1.06	−1.01	−1.35	1.99	1.46
23	STN	Right	2nd	0.23	−0.04	−2.36	2.37	0.23
24	STN	Right	1st	−0.16	−1.40	−0.43	1.47	1.41
25	STN	Right	2nd	−0.78	−1.28	−0.81	1.70	1.50
26	STN	Right	1st	−1.64	−0.34	1.29	2.11	1.67
27 a	GPi	Right	2nd	−0.8	0.06	−0.40	0.90	0.80
28	STN	Right	1st	−0.82	−1.79	−0.89	2.16	1.97
29	ANT	Right	1st	−1.96	−0.02	−1.40	2.41	1.96
30	STN	Right	2nd	−1.26	−1.97	0.09	2.34	2.34
31 a	GPi	Right	2nd	−0.15	−0.79	0.24	0.84	0.80
32	STN	Right	1st	−0.84	−1.21	−0.34	1.51	1.47
33	STN	Right	2nd	−0.22	−1.17	−0.34	1.24	1.19
			Average	−0.95 ± 0.78	−1.06 ± 0.81	−0.55 ± 1.03	1.93 ± 0.91	1.57 ± 0.88

ANT, anterior thalamus; GPi, globus pallidus internus; STN, subthalamic nucleus; VIM, ventral intermediate nucleus.

^aIn the second implant of a staged procedure, new preoperative imaging was obtained.

When the electrode position was compared based on whether the electrodes were implanted first or second, the average vector error was 1.68 ± 0.68 mm for first implants (n = 21) and 2.02 ± 0.94 mm for second implants (n = 12). The difference in vector error between first and second implants was not significant (P = .27). For the 2 patients who underwent staged procedures, both electrodes were considered first implants. See Table 2 for more details.

TABLE 2. - Difference Between Intended and Actual Electrode Coordinates Based on Implant Order

First implant	Second implant	df	t	P
All patients
Vector error (mm)
1.68 ± 0.68 (n = 21)	2.02 ± 0.94 (n = 12)	18.82	−1.13	.27
Radial error (mm)
1.35 ± 0.53 (n = 21)	1.60 ± 0.96 (n = 12)	15.66	−0.83	.42
No repositioning
Vector error (mm)
1.60 ± 0.73 (n = 16)	1.80 ± 0.99 (n = 8)	12.03	−0.49	.63
Radial error (mm)
1.38 ± 0.57 (n = 16)	1.47 ± 0.80 (n = 8)	11.64	−0.29	.73

After excluding the electrodes that were repositioned after microelectrode recording and/or test stimulation (because purposeful repositioning meant that the target trajectory was no longer the initially planned location), we found that right-sided electrodes (n = 10) were, on average, 0.88 ± 0.80 mm more medial, 0.94 ± 0.73 mm more posterior, and 0.76 ± 0.84 mm more superior compared with the planned trajectories, with a mean vector error of 1.74 ± 1.01 mm (Figure 2A). Left-sided electrodes (n = 14) were 0.49 ± 1.02 mm more lateral, 0.63 ± 0.74 mm more posterior, and 0.49 ± 0.87 mm more superior on average than the planned trajectories, and the mean vector error was 1.62 ± 0.68 mm (Figure 2B). See Table 3.

TABLE 3. - Difference Between Intended and Actual Coordinates for Electrodes Not Repositioned After Microelectrode Recording or Test Stimulation

Patient	Target	Side	Implant order	Difference between intended and actual coordinates
				Lateral right (mm)	Anterior (mm)	Inferior (mm)	Vector distance (mm)	Radial error (mm)
1	VIM	Left	1st	−0.03	−1.81	−0.30	1.83	1.81
2	GPi	Left	1st	−1.16	−1.24	−2.81	3.28	1.70
3	VIM	Left	1st	0.04	−1.04	−0.57	1.19	1.04
4	STN	Left	2nd	−1.41	−0.15	0.51	1.51	1.42
5	GPi	Left	1st	−2.55	0.09	−0.29	2.57	2.55
6	STN	Left	2nd	−0.99	−0.25	−0.69	1.23	1.02
7	VIM	Left	1st	−0.30	−1.57	−0.49	1.67	1.60
8	GPi	Left	1st	−0.51	0.69	−0.69	1.10	0.86
9	VIM	Left	1st	−0.72	0.12	−0.29	0.79	0.73
10	ANT	Left	2nd	1.46	−0.39	−1.16	1.91	1.51
11	STN	Left	1st	−0.81	−1.30	0.73	1.70	1.53
12	GPi	Left	1st	−1.21	−1.29	−0.85	1.96	1.77
13	STN	Left	2nd	0.75	−0.25	−0.58	0.98	0.79
14	STN	Left	1st	0.64	−0.48	0.59	0.99	0.80
			Average	−0.49 ± 1.02	−0.63 ± 0.74	−0.49 ± 0.87	1.62 ± 0.68	1.37 ± 0.52
15	GPi	Right	2nd	−2.52	−2.20	−2.39	4.11	3.35
16	VIM	Right	2nd	−0.88	−0.48	−1.03	1.44	1.00
17	STN	Right	1st	−0.27	−0.63	0.44	0.81	0.69
18	STN	Right	2nd	−1.06	−1.01	−1.35	1.99	1.46
19	STN	Right	1st	−0.16	−1.40	−0.43	1.47	1.41
20 a	GPi	Right	2nd	−0.80	0.06	−0.40	0.90	0.80
21	STN	Right	1st	−0.82	−1.79	−0.89	2.16	1.97
22	ANT	Right	1st	−1.96	−0.02	−1.40	2.41	1.96
23 a	GPi	Right	2nd	−0.15	−0.79	0.24	0.84	0.80
24	STN	Right	2nd	−0.22	−1.17	−0.34	1.24	1.19
			Average	−0.88 ± 0.80	−0.94 ± 0.73	−0.76 ± 0.84	1.74 ± 1.01	1.46 ± 0.81

ANT, anterior thalamus; GPi, globus pallidus internus; STN, subthalamic nucleus; VIM, ventral intermediate nucleus.

^aIn the second implant of a staged procedure, new preoperative imaging was obtained.

When coordinates for all electrodes that were not repositioned were compared based on whether the electrode was implanted first or second, the average vector error was 1.60 ± 0.73 mm for first implants (n = 16) and 1.80 ± 0.99 mm for second implants (n = 8). The difference in vector error between first and second implants was not significant (P = .63). For the 2 patients who underwent staged procedures, both electrodes were again considered first implants. See Table 2.

Microelectrode Recording and Electrode Repositioning

Microelectrode recordings were performed during surgery for 7 patients (14 electrodes), all of whom received STN implants to treat PD. Eight of these electrodes were repositioned based on those recordings. Five electrodes were positioned more inferiorly and 2 electrodes more superiorly along the planned trajectory (ie, a second pass was not required). One electrode was moved 2 mm anterior to the planned trajectory.

One patient's right-sided ventral intermediate nucleus electrode was repositioned 2 mm anterolateral to the planned trajectory because the patient experienced facial paresthesias at low-amplitude settings during stimulation at the original target.

DISCUSSION

This study sought to determine the mechanical accuracy of the new Leksell Vantage frame by comparing the intended and actual electrode coordinates for 33 lead implants. Overall, the discrepancy between intended and actual electrode coordinates was small, and there were no obvious biases noted in the directionality of the imprecision. Moreover, the accuracy noted is consistent with similar reports of targeting accuracy for DBS and more than acceptable for performing the DBS procedure. To our knowledge, this is the first such analysis regarding the accuracy of the Vantage frame.

Targeting Accuracy and Discrepancies

In one 2015 study of 116 DBS leads implanted using the G frame, the vector error (defined as the vector distance between the intended and actual electrode coordinates) was determined to be 2.5 ± 1.2 mm. The absolute lateral/medial error was 1.4 ± 1.0 mm, the absolute anterior/posterior error was 1.2 ± 1.0 mm, and the absolute superior/inferior error was 1.3 ± 0.9 mm.⁵ In comparison, our study found the vector error to be 1.74 ± 1.01 mm for right-sided electrodes and 1.62 ± 0.68 mm for left-sided electrodes, and the error in each three-dimensional plane was <1 mm on both the right and left sides. This suggests that the Vantage frame provides a mechanical accuracy that is similar to and perhaps slightly better than that provided by the older G frame. Explanations for this improvement include the nonmetallic composition of the frame, the frame's resistance to torque because of its solid single-component design, and the superior articulation of the operating arc to the frame.

Impact of Brain Shift

Brain shift, or the deformation/displacement of the brain during surgery, is of critical concern to DBS surgeons, particularly when performing bilateral procedures. In one 2009 study of 25 patients undergoing bilateral DBS lead implantation, X-ray and MRI performed after the first lead was implanted revealed up to 1.3 mm of brain shift at the anterior commissure and up to 2.5 mm of brain shift along the planned implantation trajectory of the second electrode.⁶ In another study of 28 electrode implants, brain shift in the contralateral direction was found after the first electrode implant.⁷ Brain shift such as that described in these studies can theoretically undermine the accuracy of the second lead implant because the position of the therapeutic brain target is no longer accurately represented by the preoperative images used for trajectory planning.

We compared the vector error between electrodes implanted first or second during a bilateral procedure, and although the error was greater for second side implants, the difference was not significant. This may be due to the senior author's technique, which minimizes cerebrospinal fluid losses.

Generalizability and Limitations

As the first study examining the real-world accuracy of the new Leksell Vantage frame, this study should prove to be useful to DBS surgeons who are considering a switch to the Vantage frame. Our study demonstrates that the Vantage frame is accurate when used properly and has several features that make it preferable to its predecessor. Our study cohort was limited because our surgeon only recently began using the Vantage frame, so future reviews will be needed to confirm our results.

One limitation of this study is its reliance on an accurate fusion of the preoperative MRI to the intraoperative CT, a process that no doubt accounts for some unquantifiable portion of the error that we measured. It is also possible that the presence of subdural air in the intraoperative CT scan both degraded the quality of the image fusion and altered the apparent location of the implanted lead. There are conflicting conclusions in the literature regarding the latter point. Kim et al⁸ found that subdural air in the immediate postoperative scan accounts for a significant shift in the apparent location of the DBS lead tip, most notably along the Y axis. By contrast, Bentley et al⁹ found that although subdural air elicits a bend in the DBS lead wire on immediate postoperative CT scans, the anatomic position of the lead tip actually shifts very little in this process. Our results support the latter conclusion although we still advocate using any and all techniques to minimize the amount of air that is permitted to enter the subdural space during DBS surgery. Finally, it is also possible that artifact from the electrodes on the intraoperative CT may affect the perceived position of electrodes although the lead detection software is presumably designed to accommodate that artifact.

CONCLUSION

This study demonstrated the targeting accuracy of the Leksell Vantage frame and described discrepancies that our surgeon experienced between intended electrode trajectory and actual electrode placement. The accuracy of the Vantage frame was found to be potentially better than that of the traditional G frame, suggesting that the Vantage frame is a very strong alternative to the G frame when performing DBS. Brain shift during DBS procedures also had no effect on Vantage accuracy in our study. Because the Vantage frame is more comfortable for patients, it may therefore prove to be the preferred option for DBS surgery in the future.