Deep brain stimulation is a well-established procedure

Transcription

1 J Neurosurg 110: , 2009 Avoiding the ventricle: a simple step to improve accuracy of anatomical targeting during deep brain stimulation Clinical article Lu d v i c Zr i n z o, M.D., M.Sc., F.R.C.S.Ed. (Ne u r o.su r g), 1,2 Ar j e n L. J. va n Hu l z e n, M.Sc., 3 Al e s s a n d r a A. Go r g u l h o, M.D., 4 Pa t r i c i a Li m o u s i n, M.D., Ph.D., 1 Mi c h i e l J. Sta a l, M.D., Ph.D., 3 An t o n i o A. F. De Sa l l e s, M.D., Ph.D., 4 a n d Marwa n I. Ha r i z, M.D., Ph.D. 1,5 1 Unit of Functional Neurosurgery, Sobell Department of Motor Neuroscience and Movement Disorders, Institute of Neurology, Queen Square, London; 2 Victor Horsley Department of Neurosurgery, National Hospital for Neurology and Neurosurgery, Queen Square, London, United Kingdom; 3 Department of Neurosurgery, University Medical Centre Groningen, The Netherlands; 4 Department of Neurosurgery, David Geffen School of Medicine, University of California, Los Angeles, California; and 5 Department of Neurosurgery, University Hospital of Northern Sweden, Umeå, Sweden Object. The authors examined the accuracy of anatomical targeting during electrode implantation for deep brain stimulation in functional neurosurgical procedures. Special attention was focused on the impact that ventricular involvement of the electrode trajectory had on targeting accuracy. Methods. The targeting error during electrode placement was assessed in 162 electrodes implanted in 109 patients at 2 centers. The targeting error was calculated as the shortest distance from the intended stereotactic coordinates to the final electrode trajectory as defined on postoperative stereotactic imaging. The trajectory of these electrodes in relation to the lateral ventricles was also analyzed on postoperative images. Results. The trajectory of 68 electrodes involved the ventricle. The targeting error for all electrodes was calculated: the mean ± SD and the 95% CI of the mean was 1.5 ± 1.0 and 0.1 mm, respectively. The same calculations for targeting error for electrode trajectories that did not involve the ventricle were 1.2 ± 0.7 and 0.1 mm. A significantly larger targeting error was seen in trajectories that involved the ventricle (1.9 ± 1.1 and 0.3 mm; p < 0.001). Thirty electrodes (19%) required multiple passes before final electrode implantation on the basis of physiological and/or clinical observations. There was a significant association between an increased requirement for multiple brain passes and ventricular involvement in the trajectory (p < 0.01). Conclusions. Planning an electrode trajectory that avoids the ventricles is a simple precaution that significantly improves the accuracy of anatomical targeting during electrode placement for deep brain stimulation. Avoidance of the ventricles appears to reduce the need for multiple passes through the brain to reach the desired target as defined by clinical and physiological observations. (DOI: / JNS08885) Ke y Wo r d s deep brain stimulation stereotactic accuracy ventricle Deep brain stimulation is a well-established procedure used in the treatment of a variety of chronic neurological conditions. 8,11,22,27 Functional neurosurgeons take great pains to improve the accuracy of electrode placement in the brain. 1,2 Stereotactic anatomical targeting is an essential first step in every functional procedure. 4,16,17 Improved accuracy at this stage is highly desirable since this is likely to result in superior clinical outcomes and should minimize the need for multiple passes within the brain with the increased risk that this entails. Abbreviation used in this paper: DBS = deep brain stimulation. J. Neurosurg. / Volume 110 / June 2009 To date, frame-based stereotactic methods remain the gold standard in terms of targeting accuracy. 3,18 The drive behind this study arose from an ongoing audit performed by 2 of the authors (L.Z. and M.I.H.) of patients undergoing DBS electrode placement at a London center, in which the targeting error for each implanted electrode was assessed. One electrode in our London series had a much greater than expected targeting error (3.7 mm compared with the mean targeting error of 1.3 ± 0.7 mm, unpublished data). During electrode implantation, transgression into the lateral ventricle was suspected based on a decrease in impedance measure- 1283

2 L. Zrinzo et al. ment; this suspicion was later confirmed on postoperative stereotactic imaging, and the electrode was subsequently repositioned. On the basis of this anecdotal incident, it became policy at this London center to ensure that the planned electrode trajectory would avoid the ventricle whenever possible. There are no published guidelines or data as to whether the ventricle should be avoided when planning DBS electrode trajectories. We therefore approached DBS centers that did not routinely avoid the ventricle during the procedure to study whether transgression of the ventricular system influenced targeting accuracy. We also considered whether ventricular involvement had an impact on the number of brain passes required to achieve satisfactory physiological and/or clinical observations during surgery. Methods The targeting error during placement of 162 DBS electrodes was determined at 2 centers (57 in Groningen, 105 in Los Angeles); electrodes were implanted in 109 consecutive patients (33 in Groningen, 76 in Los Angeles). Patient data are summarized in Table 1. The Leksell Stereotactic System Model G (Elekta Instrument AB) was used at both centers. The planned target was defined on preoperative stereotactic MR imaging, and the initial stereotactic coordinates were calculated and recorded, as were the angles on the arc and ring of the frame when the final electrode was implanted. Postoperative stereotactic imaging was used to determine the final electrode location. 7 Operative Technique: Groningen All procedures were performed by the same neurosurgeon (M.S.) with the patient in the supine position. After the Leksell G frame and localizer box were mounted on the head, 1.5-T MR images were obtained (Sonata Vision; Siemens). A 3D multiplanar T1-weighted scan (104 slices; voxel size mm; TR 30.0 msec; TE 5 msec), a T2-weighted coronal scan (40 slices; voxel size mm; TR 5000 msec; TE 112 msec), and a turbo inversion recovery series (turbo factor 11; voxel size mm; TR 6510 msec; TE 14 msec) were obtained. These MR imaging data sets were exported for use with stereotactic planning software and spatially fused (@Target; BrainLAB). The anterior and posterior commissure coordinates were determined and the planned target was localized by direct visualization of anatomical structures and their relationship to neighboring anatomical landmarks. The resulting stereotactic coordinates were used to plan the initial trajectory. Semimicroelectrode recordings were performed intraoperatively, together with macrostimulation and assessment of clinical effects. In a number of cases, the original image-directed trajectory and coordinates were altered in response to the physiological and/or clinical observations. After the final semimicroelectrode recording was completed, the semimicroelectrode was replaced with a quadripolar lead (Type 3389 or 3387; Medtronic, Inc.) introduced into the brain through a rigid metal cannula. The cannula TABLE 1: Summary of patient and electrode data Parameter Groningen Los Angeles Total patient demographics male female PD dystonia ET other tremor pain electrode placement lt hemisphere rt hemisphere STN GPi motor thalamus other* * Other placements included the periaqueductal gray, sensory thalamus, or posterior hypothalamus. Abbreviations: ET = essential tremor; GPi = globus pallidus pars interna; PD = Parkinson disease; STN = subthalamic nucleus. used was an in-house development specifically designed to allow careful withdrawal without displacement of the implanted lead (unpublished method). All error calculations were based on the final coordinates used for quadripolar lead implantation. The electrode was secured to the skull with Histoacryl cement after removal of the stylet. After every procedure, the final position of the DBS lead was verified on frontal and lateral digitized skull radiography using a localizer box attached to the stereotactic frame; T1- and T2-weighted MR images were obtained the day after surgery. All postimplantation images were imported into the planning system and fused with the preoperative stereotactic MR images. Fiducial markers engraved on the sides of the radiographic localizer box allowed precise stereotactic localization of the leads without introduction of errors from image fusion. The stereotactic coordinates of the most distal and most proximal electrode contacts were defined and recorded. Postoperative MR images allowed visualization of the electrode trajectory through the brain as well as investigation of lateral ventricle involvement (Figs. 1 and 2). Operative Technique: Los Angeles The operative technique used in Los Angeles has been reported previously and is summarized here briefly. 10,21 The Leksell stereotactic frame and MR Indicator (Elekta) were applied to the patient s head in the anteroom of the interventional MR imaging suite and adjusted parallel to the infraorbital meatal line (Reid s baseline) such that the stereotactic frame was parallel to the anterior commissure posterior commissure plane. 32 Preoperative 1.5-T MR images (Sonata; Siemens) were obtained; axial T1-weighted volume images (voxel size J. Neurosurg. / Volume 110 / June 2009

3 Avoiding the ventricle improves DBS targeting accuracy Fig. 2. A: Axial stereotactic T1-weighted MR images obtained in a patient at the Los Angeles center. Hypointense DBS lead artifacts (arrowheads) are shown. The left DBS lead can be seen within the left frontal horn (upper) indicative of ventricular involvement before reentering parenchyma further along the trajectory (lower). The DBS lead artifact in the right hemisphere is clearly seen embedded within brain parenchyma. B: Coronal T1-weighted MR image (nonstereotactic) obtained in a patient at the Groningen center. The DBS lead artifact can be seen to traverse the lateral ventricle in the left hemisphere and is clearly embedded within brain parenchyma throughout its trajectory in the right hemisphere. Fig. 1. Upper: Postimplantation stereotactic MR image showing the trajectory view along the implanted DBS lead. The distal portion of the electrode generates a cylindrical hyposignal with a more bulbous terminal portion. The trajectory of the implanted DBS lead can be established in stereotactic space as can the stereotactic coordinates of the electrode contacts. These coordinates can be transposed to the preoperative stereotactic MR images to provide precise localization of the contacts in relation to the visualized anatomy. Lower: The targeting error (d) is defined as the shortest distance between the final target and the electrode trajectory and is represented by a perpendicular line dropped from the intended target (gray cross) to the electrode trajectory (T). Given the coordinates of the final target and of the distal and proximal electrode contacts, the targeting error (d) can be calculated geometrically using Heron s formula. Inset: The vector components (x, y, and z) of the targeting error are shown. J. Neurosurg. / Volume 110 / June 2009 mm; TR 11.4 msec; TE 4.4 msec) and T2-weighted axial and coronal images (voxel size mm; TR 4000 msec; TE 96 msec) were imported into planning software (iplan; BrainLAB) for computer-assisted target determination. Single-channel microrecordings were obtained using a NeuroTrek system (Alpha-Omega). Recordings were taken 2.5 cm above the target depth, and progressed either to 2 mm beyond the target, or a point beyond the target that showed patterning suggestive of the substantia nigra when targeting the subthalamic nucleus or CSF silence when targeting the posteroventral globus pallidus pars interna. Microelectrode recordings were not used when targeting the ventral intermediate thalamic nucleus, ventral posterior thalamic nucleus, or periaqueductal gray. On occasion, physiological and/or clinical observations along the initial trajectory led to further tracks being made before the final lead was implanted. The initial coordinates were adjusted to reflect the final intended stereotactic target. Electrodes were fixed in position using the Stimloc device (Medtronic, Inc.). Intraoperative stereotactic MR images were obtained with the frame in situ immediately after lead implantation so that electrode position could be documented and hemorrhage ruled out. Intraoperative images were imported into planning software and reconstructed along the axis of the electrodes. On MR images, the DBS lead generated a cylindrical hyposignal with a more bulbous terminal portion, generated by the exposed contacts (Fig. 1 upper). 40 A template was placed along the center of the electrode artifact on greatly magnified postoperative images to represent the center of the 4 electrode contacts (2 and 3 mm apart on the 3389 and 3387 electrodes, respectively). The coordinates of each contact were then calculated using the planning software. Data Processing The operative procedure in both centers provided coordinates for the final target and also defined the electrode trajectory in stereotactic space. When modifications to the initial trajectory were undertaken in response to physiological and/or clinical observations, the final target coordinates were adjusted accordingly. These adjustments involved geometric calculation of the final target coordinates based on detailed intraoperative records on the spatial relationship of the final trajectory to the initially planned trajectory. The targeting error was defined as the shortest distance 1285

4 L. Zrinzo et al. between the final target coordinates and the electrode trajectory. This targeting error was calculated geometrically with Heron s formula for every implanted electrode (Fig. 1 lower). 9 Postoperative MR images were also used to evaluate the relationship of the electrode trajectory to the lateral ventricles. Contiguous axial images covering the length of the electrode trajectory were studied, and electrodes were classified into 2 groups. If the electrode artifact was embedded in brain parenchyma from cortex to tip, the electrode was recorded as having avoided the ventricle (Group 1). If the signal artifact was seen to traverse the ventricle or touch the ventricle wall on any of the axial images, the electrode was considered to have involved the ventricular system (Group 2; Fig. 2). Statistical Analysis Descriptive statistics used for the observed targeting error include the mean, SD, and 95% CIs of the mean. An independent 2-sample t-test was used to analyze differences in targeting error between electrodes that involved the ventricle and those that did not. The chi-square test was used to analyze whether there was a significant relationship between ventricular involvement and the requirement for multiple brain passes to obtain the desired target, as defined by intraoperative clinical and physiological observations. The level of significance for all statistical analyses was chosen at probability values < Results The mean targeting error for all electrodes was 1.5 mm with an SD of 1.0 mm; the 95% CI of the mean targeting error was mm. Twenty-two (39%) of 57 DBS electrodes implanted at the Groningen center, and 46 (44%) of 105 DBS electrodes implanted at the Los Angeles center were found to have involved the ventricular system. The mean targeting error for electrodes that did not involve the ventricle was 1.2 mm (SD 0.7 mm; 95% CI of the mean mm). Electrodes with trajectories involving the ventricle showed a significantly greater targeting error: 1.9 mm (SD 1.1 mm; 95% CI of mean mm) (p < 0.001). Deep brain stimulation electrodes whose trajectories avoided the ventricle therefore showed a 36% reduction in mean targeting error over those electrodes with trajectories involving the ventricle. A targeting error > 2.0 mm was present in 37% of electrodes with a trajectory that penetrated the ventricle but in only 14% of electrodes with a trajectory that avoided it. Results were consistent within individual centers, with the mean targeting error for all electrodes at Groningen and Los Angeles being 1.5 mm (SD 0.8 mm; 95% CI mm) and 1.5 mm (SD 1.0 mm; 95% CI mm), respectively. The error for electrode trajectories that did not involve the ventricle was 1.2 mm (SD 0.7 mm; 95% CI mm) at Groningen and 1.2 mm (SD 0.8 mm; 95% CI mm) at Los Angeles. Electrodes with a trajectory involving the ventricle exhibited significantly more targeting error at both centers: 1.9 mm (SD 0.8 mm; 95% CI mm) at Groningen and 1.9 mm (SD 1.2 mm; 95% CI mm) at Los Angeles (p < 0.001; Table 2 and Fig. 3). Reduced anatomical targeting accuracy may be expected to require additional brain passes to achieve the desired target. We therefore looked at the number of times that clinical and physiological observations resulted in a change in the initial trajectory. Thirty electrodes (19%) required multiple passes prior to final electrode implantation. In the Groningen center, multiple tracks were made for 4 (11%) of the 35 electrodes that avoided the ventricle and 6 (27%) of the 22 electrodes that involved the ventricle. The findings were similar in the Los Angeles center, where 7 (12%) of 59 electrode trajectories that avoided the ventricle required multiple tracks, and 13 (28%) of 46 that involved the ventricle had required multiple tracks. The direction of movement from initial to final trajectory was analyzed. Movements were found to be scattered in all directions without bias toward a medial direction. The chi-square test was performed using combined data from both centers and confirmed a significant association between ventricle involvement and an increased requirement for multiple brain passes (p < 0.01). For the 132 electrodes that did not require more than 1 pass, the targeting error remained significantly greater for those involving the ventricle than those avoiding it (Group 2: 1.8 mm, SD 1.2 mm; Group 1: 1.1 mm, SD 0.7 mm; p < 0.001). Discussion In the present study, we have demonstrated that the accuracy of anatomical targeting in functional neurosurgery can be significantly improved by ensuring that electrode trajectories do not transgress the walls of the lateral ventricle, whenever possible. Accuracy and precision in electrode placement is of great importance in functional neurosurgery. Indeed, centers where such procedures are performed place great emphasis on preoperative imaging and targeting. 4,29 It follows that an equal effort should be made to verify that the target area has been reached. The gold standard in achieving this verification is postoperative stereotactic imaging that localizes the implanted electrodes within the preoperatively defined stereotactic space without introducing image fusion errors. 38 Routine pre- and postoperative stereotactic images are obtained in all patients undergoing DBS electrode implantation in the centers involved in this study. An ongoing audit of targeting accuracy at the London center was instrumental in highlighting a deviation from the expected level of accuracy on a single occasion when the electrode trajectory inadvertently transgressed the ventricle. This incident resulted in a change of practice such that increased emphasis was placed on ensuring that ventricular transgression did not occur whenever possible. A literature search provided another anecdotal report on the adverse effects of ventricular transgression on lead placement accuracy, 19 but we were unable to find a study examining the relationship between targeting accuracy and ventricular involvement in the electrode s trajectory. We defined targeting error in this study as the shortest distance between the final target and the electrode trajectory. We submit that it is this error that is most clinically relevant when implanting a quadripolar electrode that spans a 1286 J. Neurosurg. / Volume 110 / June 2009

5 Avoiding the ventricle improves DBS targeting accuracy TABLE 2: Summary of results All Electrodes (mm) Avoided Ventricles (mm) Involved Ventricles (mm)* Center Mean ± SD 95% CI of Mean Mean ± SD 95% CI of Mean Mean ± SD 95% CI of Mean Groningen 1.5 ± 0.8 ± ± 0.7 ± ± 0.8 ± 0.3 Los Angeles 1.5 ± 1.0 ± ± 0.8 ± ± 1.2 ± 0.3 both 1.5 ± 1.0 ± ± 0.7 ± ± 1.1 ± 0.3 * All trajectories that involved the ventricles had significantly increased targeting error compared with those that avoided them (p < 0.001). length of 6 or 9 mm from the center of the most proximal to that of the most distal contact (3389 or 3387 DBS leads, respectively). This contact arrangement allows deviation along the axis of the electrode to be readily resolved by selection of adjacent contacts for stimulation. However, perpendicular deviation of the electrode trajectory from the intended target point is most likely to result in a lead placement error that requires additional tracks through the brain. 13 Fig. 3. Box-and-whisker graph showing combined data from both centers in the study. The 95% CI of the mean (box) and the SD (whiskers) of the targeting error is shown for all electrode trajectories and for Groups 1 and 2. There is a significantly larger targeting error for trajectories involving the ventricle (Group 2) compared with those solely traversing brain parenchyma (Group 1; p < 0.001). J. Neurosurg. / Volume 110 / June 2009 The authors of a number of published phantom studies have assessed the targeting accuracy of various stereotactic frames. 23,28,41 Additional factors may further affect targeting accuracy in vivo, including deformation of the stereotactic frame when mechanical loading is applied to the skull, 35 MR imaging distortion, 39 and brain shift. 12,26 Numerous other variables that could influence targeting accuracy including unilateral versus bilateral surgery, order of electrode insertion during bilateral surgery, and the laterality of surgery were considered beyond the scope of the present study. The in vivo perpendicular targeting error of the Riechert-Mundinger frame has been examined with preand postoperative stereotactic CT scans, and the mean ± SD of the targeting error has been reported as 1.3 ± 0.8 mm. 13 Postoperative MR images have been used increasingly to determine the accuracy of electrode placement. 15,24,29 31,33,36 However, studies that depend on fusion of pre- and postimplantation images to localize implanted electrodes may also introduce errors of image fusion. We used postimplantation stereotactic MR imaging to assess targeting accuracy at the Los Angeles center and based calculations on stereotactic radiograms at the Groningen center. Both methods avoid the introduction of image fusion errors. Despite the different imaging techniques used in Los Angeles and Groningen, the observed targeting accuracy was remarkably similar. This result suggests that, with adequate quality control, MR imaging distortion may not be a clinically significant issue in anatomical targeting for functional neurosurgery. 5,6,20,24,25,34 In the present study, combined and individual data collected at 2 centers demonstrated conclusively that trajectories involving the ventricle are accompanied by a significant reduction in targeting accuracy. When trajectories avoided the ventricle, there was a 36% reduction in mean targeting error compared with trajectories involving the ventricle (p < 0.001). The absolute difference in mean targeting error between the 2 groups may at first glance seem clinically irrelevant (1.2 vs 1.9 mm). However, the impact of ventricular involvement on targeting error in individual electrodes is best appreciated by referring to Fig. 4. A much greater proportion of electrodes in Group 1 showed targeting errors of < 1.0 mm. The targeting error was < 2.0 mm in 86% of electrodes from Group 1 compared with 63% of electrodes from Group 2. We have not attempted to correlate targeting accuracy with clinical outcome in the present study; nonetheless, we consider a targeting error of > 2.0 mm to be clinically relevant. 1287

6 L. Zrinzo et al. Fig. 4. Bar graph showing the percentage distribution of targeting error in the 2 groups of electrode trajectories. Note that 86% of electrode trajectories in Group 1 had targeting errors < 2.0 mm compared with 63% of electrodes in Group 2. Also, in contrast to Group 2, no electrodes from Group 1 had a targeting error of > 4.0 mm. This study does not elucidate why electrode trajectories that involve the ventricle result in such a reduction in targeting accuracy. We hypothesize that the ventricle wall presents a mechanical obstacle to the advancing electrode and that this may result in deviation from the intended target through various mechanisms. The cannula and stylet do not have infinite rigidity and may be slightly deformed by the rigidity of the ventricle wall. Even a very small deviation several centimeters above the target area would result in a measurable increase in the observed error at target. Likewise, the advancing electrode may transfer a vector force to the brain as it meets the resistance of the ventricle wall, deforming the brain such that the target area could be minimally displaced by the time the electrode reaches the end of its trajectory. After implantation of the DBS lead, withdrawal of the rigid cannula would allow the brain to spring back to its usual anatomical position, carrying with it the more flexible DBS electrode away from the intended stereotactic coordinates. Additionally, increased CSF loss and subsequent brain shift may also occur with ventricle penetration. These potential sources of error can be avoided by electrode trajectories that do not involve the ventricle wall. Other factors may dictate the planned trajectory, such as the location, shape, and orientation of the target structure. In our experience, the planned trajectory can usually be adjusted to ensure ventricular avoidance while positioning as many electrode contacts as possible in the commonly used anatomical structures in functional neurosurgery (subthalamic nucleus, posteroventral pallidum, motor thalamus, and zona incerta). In some instances it may be impossible to access certain targets satisfactorily without involving the ventricle. A transventricular route may have to be considered for other DBS targets, which may have an impact on the initial accuracy when targeting these structures (such as the pedunculopontine nucleus in patients with parkinsonism; the anterior nucleus of the thalamus in those with epilepsy). 14,37,42 Clearly, avoidance of the ventricles should not be the only factor that dictates the electrode trajectory in individual cases. However, neurosurgeons may wish to consider the ventricle as an additional factor when planning the surgical trajectory. We found a significant association between ventricle involvement of the electrode trajectory and an increase in the requirement for multiple brain passes (p < 0.01). One explanation for this association may simply be that multiple brain passes lead to ventricle penetration. However, the direction of movement from initial to final trajectory was scattered, without a medial bias as would be expected if this were the case. An alternative notion is that the reduced anatomical targeting accuracy seen in electrodes with trajectories involving the ventricle increases the probability that additional brain passes will be required to reach the target as defined by intraoperative clinical and physiological observations. Indeed, the proportion of electrodes requiring multiple passes through the brain more than doubled when the trajectory involved the ventricle compared to those that avoided it. Improved targeting accuracy when avoiding the ventricle may therefore not only be statistically significant but also clinically relevant. This simple precaution of avoiding the ventricle in the planned trajectory has theoretical benefits that may include a lower risk of hemorrhage (both by avoiding ependymal vessels and decreasing the number of brain passes), reduced operative time, and potentially improved clinical outcome. Conclusions Improved anatomical targeting accuracy is highly desirable in functional neurosurgery, regardless of the individual surgeon s particular surgical protocol. Ensuring a surgical trajectory that avoids the ventricles is a simple step when using modern planning software that improves the accuracy of anatomical targeting and appears to reduce the need for multiple passes through the brain to reach the desired target. Disclosure Medtronic provided financial support for the study in the form of travel expenses for Dr. Zrinzo. Drs. Hariz and Limousin occasionally receive honoraria for lectures by Medtronic. The authors have reported no other conflicts of interest J. Neurosurg. / Volume 110 / June 2009

7 Avoiding the ventricle improves DBS targeting accuracy Acknowledgments This study was undertaken at University College London/ Uni versity College London Hospitals (UCL/UCLH) and was partly funded by the Department of Health National Institute for Health Re search (NIHR) Biomedical Research Centres funding scheme. The Unit of Functional Neurosurgery, Queen Square, London, is sup ported by the Parkinson s Appeal. Dr. Zrinzo thanks Medtronic for its financial support of this study. The authors acknowledge Dr. Thomas Foltynie for his contribution toward the final manuscript. References J. Neurosurg. / Volume 110 / June Bejjani BP, Dormont D, Pidoux B, Yelnik J, Damier P, Arnulf I, et al: Bilateral subthalamic stimulation for Parkinson s disease by using three-dimensional stereotactic magnetic resonance imaging and electrophysiological guidance. J Neurosurg 92: , Benabid AL, Krack PP, Benazzouz A, Limousin P, Koudsie A, Pollak P: Deep brain stimulation of the subthalamic nucleus for Parkinson s disease: methodologic aspects and clinical criteria. Neurology 55 (12 Suppl):S40 S44, Bjartmarz H, Rehncrona S: Comparison of accuracy and precision between frame-based and frameless stereotactic navigation for deep brain stimulation electrode implantation. Stereotact Funct Neurosurg 85: , Breit S, LeBas JF, Koudsie A, Schulz J, Benazzouz A, Pollak P, et al: Pretargeting for the implantation of stimulation electrodes into the subthalamic nucleus: a comparative study of magnetic resonance imaging and ventriculography. Neurosurgery 58 (1 Suppl):ONS83 ONS95, Burchiel KJ, Nguyen TT, Coombs BD, Szumoski J: MRI distortion and stereotactic neurosurgery using the Cosman-Roberts-Wells and Leksell frames. Stereotact Funct Neurosurg 66: , Carter DA, Parsai EI, Ayyangar KM: Accuracy of magnetic resonance imaging stereotactic coordinates with the cosmanroberts-wells frame. Stereotact Funct Neurosurg 72:35 46, Chen CC, Pogosyan A, Zrinzo LU, Tisch S, Limousin P, Ashkan K, et al: Intra-operative recordings of local field potentials can help localize the subthalamic nucleus in Parkinson s disease surgery. Exp Neurol 198: , Coubes P, Cif L, El Fertit H, Hemm S, Vayssiere N, Serrat S, et al: Electrical stimulation of the globus pallidus internus in patients with primary generalized dystonia: long-term results. J Neurosurg 101: , Coxeter HSM: Introduction to Geometry, ed 2. New York: Wiley, De Salles AA, Frighetto L, Behnke E, Sinha S, Tseng L, Torres R, et al: Functional neurosurgery in the MRI environment. Minim Invasive Neurosurg 47: , Deuschl G, Schade-Brittinger C, Krack P, Volkmann J, Schafer H, Botzel K, et al: A randomized trial of deep-brain stimulation for Parkinson s disease. N Engl J Med 355: , Elias WJ, Fu KM, Frysinger RC: Cortical and subcortical brain shift during stereotactic procedures. J Neurosurg 107: , Fiegele T, Feuchtner G, Sohm F, Bauer R, Anton JV, Gotwald T, et al: Accuracy of stereotactic electrode placement in deep brain stimulation by intraoperative computed tomography. Parkinsonism Relat Disord 14: , Halpern CH, Samadani U, Litt B, Jaggi JL, Baltuch GH: Deep brain stimulation for epilepsy. Neurotherapeutics 5:59 67, Hamid NA, Mitchell RD, Mocroft P, Westby GW, Milner J, Pall H: Targeting the subthalamic nucleus for deep brain stimulation: technical approach and fusion of pre- and postoperative MR images to define accuracy of lead placement. J Neurol Neurosurg Psychiatry 76: , Hariz MI, Krack P, Melvill R, Jorgensen JV, Hamel W, Hirabayashi H, et al: A quick and universal method for stereotactic visualization of the subthalamic nucleus before and after implantation of deep brain stimulation electrodes. Stereotact Funct Neurosurg 80:96 101, Hirabayashi H, Tengvar M, Hariz MI: Stereotactic imaging of the pallidal target. Mov Disord 17 (3 Suppl):S130 S134, Holloway KL, Gaede SE, Starr PA, Rosenow JM, Ramakrishnan V, Henderson JM: Frameless stereotaxy using bone fiducial markers for deep brain stimulation. J Neurosurg 103: , Khan MF, Mewes K, Gross RE, Skrinjar O: Assessment of brain shift related to deep brain stimulation surgery. Stereotact Funct Neurosurg 86:44 53, Landi A, Marina R, DeGrandi C, Crespi A, Montanari G, Sganzerla EP, et al: Accuracy of stereotactic localisation with magnetic resonance compared to CT scan: experimental findings. Acta Neurochir (Wien) 143: , Lee MW, De Salles AA, Frighetto L, Torres R, Behnke E, Bronstein JM: Deep brain stimulation in intraoperative MRI environment comparison of imaging techniques and electrode fixation methods. Minim Invasive Neurosurg 48:1 6, Limousin P, Pollak P, Benazzouz A, Hoffmann D, Le Bas JF, Broussolle E, et al: Effect of parkinsonian signs and symptoms of bilateral subthalamic nucleus stimulation. Lancet 345:91 95, Maciunas RJ, Galloway RL Jr, Latimer JW: The application accuracy of stereotactic frames. Neurosurgery 35: , Martin AJ, Larson PS, Ostrem JL, Keith Sootsman W, Talke P, Weber OM, et al: Placement of deep brain stimulator electrodes using real-time high-field interventional magnetic resonance imaging. Magn Reson Med 54: , Menuel C, Garnero L, Bardinet E, Poupon F, Phalippou D, Dormont D: Characterization and correction of distortions in stereotactic magnetic resonance imaging for bilateral subthalamic stimulation in Parkinson disease. J Neurosurg 103: , Miyagi Y, Shima F, Sasaki T: Brain shift: an error factor during implantation of deep brain stimulation electrodes. J Neurosurg 107: , Mueller J, Skogseid IM, Benecke R, Kupsch A, Trottenberg T, Poewe W, et al: Pallidal deep brain stimulation improves quality of life in segmental and generalized dystonia: results from a prospective, randomized sham-controlled trial. Mov Disord 23: , Orth RC, Sinha P, Madsen EL, Frank G, Korosec FR, Mackie TR, et al: Development of a unique phantom to assess the geometric accuracy of magnetic resonance imaging for stereotactic localization. Neurosurgery 45: , Patel NK, Plaha P, Gill SS: Magnetic resonance imaging-directed method for functional neurosurgery using implantable guide tubes. Neurosurgery 61: , Pollo C, Villemure JG, Vingerhoets F, Ghika J, Maeder P, Meuli R: Magnetic resonance artifact induced by the electrode Activa 3389: an in vitro and in vivo study. Acta Neurochir (Wien) 146: , Pollo C, Vingerhoets F, Pralong E, Ghika J, Maeder P, Meuli R, et al: Localization of electrodes in the subthalamic nucleus on magnetic resonance imaging. J Neurosurg 106:36 44, Reid RW: Observations on the relation of the principal fissures and convolutions of the cerebrum to the outer surface of the scalp. Lancet 124: , Schrader B, Hamel W, Weinert D, Mehdorn HM: Documentation of electrode localization. Mov Disord 17 (3 Suppl):S167 S174,

8 L. Zrinzo et al. 34. Simon SL, Douglas P, Baltuch GH, Jaggi JL: Error analysis of MRI and Leksell stereotactic frame target localization in deep brain stimulation surgery. Stereotact Funct Neurosurg 83:1 5, Treuer H, Hunsche S, Hoevels M, Luyken K, Maarouf M, Voges J, et al: The influence of head frame distortions on stereotactic localization and targeting. Phys Med Biol 49: , Vayssiere N, Hemm S, Zanca M, Picot MC, Bonafe A, Cif L, et al: Magnetic resonance imaging stereotactic target localization for deep brain stimulation in dystonic children. J Neurosurg 93: , Velasco AL, Velasco F, Jimenez F, Velasco M, Castro G, Carrillo-Ruiz JD, et al: Neuromodulation of the centromedian thalamic nuclei in the treatment of generalized seizures and the improvement of the quality of life in patients with Lennox- Gastaut syndrome. Epilepsia 47: , West J, Fitzpatrick JM, Wang MY, Dawant BM, Maurer CR Jr, Kessler RM, et al: Comparison and evaluation of retrospective intermodality brain image registration techniques. J Comput Assist Tomogr 21: , Wu TH, Lee JS, Wu HM, Chu WF, Guo WY: Evaluating geometric accuracy of multi-platform stereotactic neuroimaging in radiosurgery. Stereotact Funct Neurosurg 78:39 48, Yelnik J, Damier P, Demeret S, Gervais D, Bardinet E, Bejjani BP, et al: Localization of stimulating electrodes in patients with Parkinson disease by using a three-dimensional atlasmagnetic resonance imaging coregistration method. J Neurosurg 99:89 99, Yu C, Apuzzo ML, Zee CS, Petrovich Z: A phantom study of the geometric accuracy of computed tomographic and magnetic resonance imaging stereotactic localization with the Leksell stereotactic system. Neurosurgery 48: , Zrinzo L, Zrinzo LV, Tisch S, Limousin PD, Yousry TA, Afshar F, et al: Stereotactic localization of the human pedunculopontine nucleus: atlas-based coordinates and validation of a magnetic resonance imaging protocol for direct localization. Brain 131: , 2008 Manuscript submitted July 19, Accepted December 8, Please include this information when citing this paper: published online March 20, 2009; DOI: / JNS Address correspondence to: Ludvic Zrinzo, M.D., Unit of Functional Neurosurgery, Sobell Department of Motor Neuroscience and Movement Disorders, Institute of Neurology & National Hospital for Neurology and Neurosurgery, Queen Square, London WC1N 3BG, United Kingdom. l.zrinzo@ion.ucl.ac.uk J. Neurosurg. / Volume 110 / June 2009