Brain shift during bur hole–based procedures using interventional MRI

Clinical article

Full access

Object

Brain shift during minimally invasive, bur hole–based procedures such as deep brain stimulation (DBS) electrode implantation and stereotactic brain biopsy is not well characterized or understood. We examine shift in various regions of the brain during a novel paradigm of DBS electrode implantation using interventional imaging throughout the procedure with high-field interventional MRI.

Methods

Serial MR images were obtained and analyzed using a 1.5-T magnet prior to, during, and after the placement of DBS electrodes via frontal bur holes in 44 procedures. Three-dimensional coordinates in MR space of unique superficial and deep brain structures were recorded, and the magnitude, direction, and rate of shift were calculated. Measurements were recorded to the nearest 0.1 mm.

Results

Shift ranged from 0.0 to 10.1 mm throughout all structures in the brain. The greatest shift was seen in the frontal lobe, followed by the temporal and occipital lobes. Shift was also observed in deep structures such as the anterior and posterior commissures and basal ganglia; shift in the pallidum and subthalamic region ipsilateral to the bur hole averaged 0.6 mm, with 9% of patients having over 2 mm of shift in deep brain structures. Small amounts of shift were observed during all procedures; however, the initial degree of shift and its direction were unpredictable.

Conclusions

Brain shift is continual and unpredictable and can render traditional stereotactic targeting based on preoperative imaging inaccurate even in deep brain structures such as those used for DBS.

Abbreviations used in this paper:AC = anterior commissure; DBS = deep brain stimulation; GPi = globus pallidus internus; PC = posterior commissure; PD = Parkinson's disease; ROI = region of interest; STN = subthalamic nucleus; VR = Virchow-Robin.

Object

Brain shift during minimally invasive, bur hole–based procedures such as deep brain stimulation (DBS) electrode implantation and stereotactic brain biopsy is not well characterized or understood. We examine shift in various regions of the brain during a novel paradigm of DBS electrode implantation using interventional imaging throughout the procedure with high-field interventional MRI.

Methods

Serial MR images were obtained and analyzed using a 1.5-T magnet prior to, during, and after the placement of DBS electrodes via frontal bur holes in 44 procedures. Three-dimensional coordinates in MR space of unique superficial and deep brain structures were recorded, and the magnitude, direction, and rate of shift were calculated. Measurements were recorded to the nearest 0.1 mm.

Results

Shift ranged from 0.0 to 10.1 mm throughout all structures in the brain. The greatest shift was seen in the frontal lobe, followed by the temporal and occipital lobes. Shift was also observed in deep structures such as the anterior and posterior commissures and basal ganglia; shift in the pallidum and subthalamic region ipsilateral to the bur hole averaged 0.6 mm, with 9% of patients having over 2 mm of shift in deep brain structures. Small amounts of shift were observed during all procedures; however, the initial degree of shift and its direction were unpredictable.

Conclusions

Brain shift is continual and unpredictable and can render traditional stereotactic targeting based on preoperative imaging inaccurate even in deep brain structures such as those used for DBS.

Abbreviations used in this paper:AC = anterior commissure; DBS = deep brain stimulation; GPi = globus pallidus internus; PC = posterior commissure; PD = Parkinson's disease; ROI = region of interest; STN = subthalamic nucleus; VR = Virchow-Robin.

Frame-based and frameless stereotaxis are methods that can be used to access small targets in the brain through minimally invasive entry sites. These targets are then biopsied, ablated, injected, or electrically stimulated. Although the target size and location in each case is slightly different, accuracy is of the utmost importance. Error at any point during navigation, setup, registration, and surgery can produce less than accurate results.

Advances in image quality and registration techniques continue to improve accuracy and speed. Procedures that use preoperative images for targeting, however, are based on the assumption that the brain does not move with respect to external coordinate systems. This assumption is violated if shift or deformation of brain structures occurs after images are obtained.19

To date, brain shift during minimally invasive, bur hole–based procedures such as implantation of deep brain stimulation (DBS) electrodes and stereotactic brain biopsy is not well characterized or understood. Many studies of brain shift are difficult to generalize due to inconsistencies in craniotomy size, amount of lesion removed, patient position, and dosage of osmotic therapy. These variations make the already complex phenomena of brain shift even more difficult to analyze. In this study we use a novel method of comparing serial high-field interventional MR images in a group of patients undergoing DBS electrode placement to evaluate and understand brain shift in small bur hole−based procedures. Serial MR images were obtained before and at multiple time points after creation of a bur hole during MRI-guided electrode implantation with the patient's head rigidly fixed and immobile at the isocenter of a 1.5-T MRI scanner. This technique for analysis of brain shift minimizes many of the variables that exist in other studies. There is no brain retraction, osmotic therapy, lesion resection, or variation of patient positioning. Causes of brain shift in this study are limited to those associated with intracranial air entry (due to CSF loss or to equalization of intracranial and atmospheric pressure) or to mechanical force on the brain resulting from the intracranial passage of guidance catheters.

Methods

Patient Selection and Demographics

A total of 39 patients underwent 47 consecutive operations for implantation of a deep brain stimulator for either Parkinson's disease (PD) or dystonia at the University of California, San Francisco (UCSF) Medical Center. These patients underwent unilateral or bilateral placement of their DBS electrodes via a bur hole procedure. Table 1 shows the patient population, procedure, location, target, and disease process. Three surgeries were excluded from the analysis. One patient was intentionally moved between MRI sequences, invalidating the methodology used to assess brain shift. In 2 additional cases, data were analyzed but the control values for fixed points in the cranium were above our acceptable threshold (> 0.5 mm, see below). Therefore, 44 separate operations were included in this study. Six patients were operated on twice either due to staged unilateral placement of stimulators (5 patients) or replacement of stimulators after infection (1 patient). The average age at the time of surgery was 58 ± 13 years old, and 48% of the patients were female. Fifteen patients had unilateral stimulator placement and 29 had bilateral placement. The target was the subthalamic nucleus (STN) in most patients, but in 5 patients the globus pallidus internus (GPi) was implanted.

TABLE 1:

Summary of demographic and clinical characteristics of 39 patients who underwent 47 DBS lead placement procedures

Procedure No.Patient No.Age (yrs), SexSide of DBS PlacementPrevious DBS PlacementTargetDisease
1147, FltnoSTNPD
22*45, MrtnoSTNPD
32*45, MltyesSTNPD
4370, MltnoSTNPD
5445, MbilatnoSTNPD
65*74, MltnoSTNPD
75*75, MrtyesSTNPD
8676, MltnoSTNPD
97*56, MltnoSTNPD
107*56, MrtyesSTNPD
11864, MbilatnoSTNPD
12967, MltnoSTNPD
131065, FltnoSTNPD
141159, MbilatnoSTNPD
151256, FbilatnoSTNPD
161348, MbilatnoSTNPD
171456, MbilatnoSTNPD
181562, FbilatnoSTNPD
191665, FbilatnoSTNPD
201759, MbilatnoSTNPD
211828, FbilatnoGPidystonia
221964, FbilatnoSTNPD
232058, MbilatnoSTNPD
242167, FbilatnoSTNPD
252259, MbilatnoSTNPD
262350, MbilatnoSTNPD
272458, MbilatnoSTNPD
282168, FbilatyesSTNPD
292562, FbilatnoSTNPD
302666, MbilatnoSTNPD
312765, FbilatnoSTNPD
3228*46, FltyesSTNPD
332968, FbilatnoSTNPD
343059, FbilatnoSTNPD
353149, FbilatnoSTNPD
363272, MbilatnoSTNPD
373362, MbilatnoSTNPD
383471, FbilatnoSTNPD
393557, MbilatnoSTNPD
4036*59, FrtnoGPiPD
413713, FrtnoGPidystonia
423865, FbilatnoSTNPD
433914, FbilatnoGPidystonia
4436*59, FltyesGPiPD

Patients 2, 5, 7, 28, and 36 had staged unilateral DBS placement. Patient 28's first surgery was performed using a microelectrode.

Patient 21 suffered an infection after the first surgery and underwent explantation and subsequent reimplantion.

Interventional MRI and Procedure

Serial MR images were obtained and analyzed using a 1.5-T magnet (Philips Integra) prior to, during, and after the placement of the DBS electrodes via frontal bur holes. In this series of patients, placement was performed using the Medtronic NexFrame deep brain access trajectory guide, with the entire procedure performed in the bore of the MRI under sterile conditions. Our surgical technique with this system has previously been described in detail.31 In brief, each patient's head is secured into a carbon fiber head holder designed to mount directly onto the MRI unit gantry. The patient is in the supine position, lying flat, with his or her head extended no more than 10°. The patient remains in this position throughout the remainder of the procedure, including bur hole placement and imaging. Coronally oriented incisions are made, followed by 14-mm frontal bur holes, cruciate dural opening, and mounting of the NexFrames. At this point, the gantry (with the patient already in position) is moved to the magnet isocenter. The table and patient are fixed in this position until the completion of surgery. All images used for this analysis were therefore obtained with the patient stationary at the isocenter after bur hole placement and dural opening. The first high-resolution T2-weighted axial MR images (2D-turbo spin echo squences) were obtained with a 2-mm slice thickness and 1-mm voxel size and aligned such that the axial images passed through the plane of the anterior commissure (AC) and posterior commissure (PC). Targets for the electrodes were then selected and, using several additional lower-resolution MR images, the NexFrame was used to place a stylet/sheath assembly and subsequently the DBS electrode into the brain. During the trajectory planning, sulci, vascular structures, and the ventricles were avoided. This process was then repeated on the contralateral side if the procedure was bilateral. After placement of both electrodes, the second set of high-resolution T2-weighted axial MR images was obtained, again with a 2-mm slice thickness. If any misplacement of the lead was noted and found to be inappropriate, an additional placement of the electrode was made. After this second placement, additional high-resolution T2-weighted axial MR images were obtained. This occurred in 9 patients. Following confirmation of appropriate location, the DBS leads were secured into place and the scalp was closed.

Measuring Shift

The MRI console tool assigns a Cartesian coordinate to each point in space within the bore of the scanner. This x, y, and z coordinate system is maintained throughout all imaging performed during the procedure, with the origin of the coordinate system (0 x, 0 y, 0 z) at the magnet isocenter. As long as the patient's head or MRI table is not moved during the procedure, this coordinate system can be used to measure movement of a specific structure from image to image. The center of a specific unique structure in the brain was identified by placing a region-of-interest (ROI) shape over the desired location. This ROI was then assigned a specific x, y, and z coordinate by the MR console. The ROI can then be overlayed onto a subsequent MR image. If movement of the center of this location was noted, the ROI shape was adjusted. The new x, y, and z coordinates were then recorded. The distance between these 2 points is the Euclidean distance. This was calculated by using the formula = √[(Δx)2 + (Δy)2 + (Δz)2]. In each operation, 15 unique superficial and deep brain structures were recorded, and the magnitude, direction, and rate of shift were calculated. These structures included at least one data point in each patient from both frontal lobe surfaces, both temporal lobes, and both occipital lobes. Also one point was obtained at the midline of the AC and the PC and at least one midline structure. Finally at least one data point was obtained from a flow void in the pallidum bilaterally and a point at each lateral border of the third ventricle. These unique structures included flow voids, Virchow-Robin (VR) spaces, unique surface structures, and unique deep structures. All points were measured in the MR image slice that was in the AC-PC plane. Each slice used in this study was 2 mm in thickness, and therefore small shifts in the Δz direction could not be measured. Axial MRI is most sensitive for evaluation of shift movement in the x and y directions; since the majority of shift recorded is in the y direction, it is likely that the magnitude of shift in the z direction has a minimal overall effect on the total shift and likely underestimates total shift. Measurements were recorded to the nearest 0.1 mm. In the anteroposterior axis, the posterior direction, which in this study is in the direction of gravity, was defined as positive. Shift anteriorly was therefore defined as negative movement. Lateral distance was identified as toward or away from the midline. Figure 1 is an example from one patient where there was a significant amount of ipsilateral frontal surface shift.

Fig. 1.
Fig. 1.

Comparison of unique structures using ROIs. A T2-weighted axial MR image is obtained at 2 separate time points during surgery. On the left is the pre–electrode placement image used for targeting. On the right is the post–bilateral electrode placement image. Each red outline is a specific ROI identifying a structure used to calculated shift. Obvious shift can be noted in the left frontal lobe surface (circled).

Assignment of Ipsilateral Shift and Contralateral Shift

Fifteen of the 44 procedures were unilateral insertion of DBS electrodes. In each of these cases, the shift on the ipsilateral and contralateral side of the brain, in reference to the electrode placement, was calculated separately. If bilateral bur holes were made and DBS electrodes were placed, both sides of the brain were considered ipsilateral and therefore any shift was categorized as ispilateral shift.

Control Points

In each operation, a unique set of points within the bony structure of the cranium were identified. Typically, 3 points were collected, one each from 3 of the 4 quadrants. Each quadrant was defined by the midsagittal line and the coronal plane at the AC-PC midpoint. These unique points included diploic spaces and unique bony structures. These same points were obtained on the subsequent scans. All points were obtained on the same image sequence and slice as the initial data point from the first image. If in any case, a control point showed more than 0.5 mm of movement, data from that case were subsequently not used. This occurred in 3 cases, and all data points showed an increase in magnitude of shift, signaling that the patient had moved from the original setting (in one case this was expected, as the patient was moved intentionally during the procedure).

Statisical Analysis

The effect of patient sex on brain shift was examined using chi-square analysis. To determine if age affected the brain shift, a linear regression analysis was performed. Statistical significance was set at p < 0.05. A paired t-test was used to compare averages of shift magnitude. All statistical analysis was performed with JMP (SAS).

Results

Control Points

The average shift of control points in each of the 4 quadrants was identified using the diploic spaces and unique bony structures (Table 2). There was no significant difference in the shift in each quadrant. The total shift recorded was 0.1 ± 0.1 mm from 153 data points. The range of shift was 0.0–0.4 mm. The largest standard deviation was 0.1 mm.

TABLE 2:

Control point data

LocationsMean Shift in mm (± SD)No. of Data Points
ant ipsilat skull0.1 ± 0.170
ant contralat skull0.0 ± 0.115
pst ipsilat skull0.1 ± 0.159
pst contralat skull0.0 ± 0.09
total average of control shift0.1 ± 0.1153

Four control points from each patient were recorded and averaged. The values reported for mean shift represent the means (and SDs) for the average of these 4 values for each patient. Ant = anterior; pst = posterior.

Magnitude of Brain Shift

Brain Shift Magnitude by Location

The average shift and range of shift were identified in specific locations throughout the brain. All points were assigned a location as well as a side of approach. The overall shift of all points recorded was 0.7 mm with a range of 0–10.1 mm. Table 3 summarizes the shift by location.

TABLE 3:

Brain shift based on location*

LocationShift Ipsilat to Bur Hole (mm)No. of Data PointsShift Contralat to Bur Hole (mm)No. of Data PointsShift Midline to Bur Hole (mm)No. of Data Points
MeanRangeMeanRangeMeanRange
frontal lobe surface1.40.0–10.1811.00.1–8.612
frontal lobe flow void/VR0.90.0–4.1280.30.2–1.010
VR pallidum0.60.0–2.1830.30.3–0.817
border of 3rd ventricle0.50.0–2.1740.30.0–0.815
ant temporal lobe flow void/VR0.60.1–3.0560.20.0–0.410
ant temporal lobe surface0.20.0–0.880.40.2–0.63
pst temporal & occipital lobe flow void/VR0.40.0–1.9380.30.0–1.017
occipital lobe surface0.40.0–2.1370.20.0–0.98
midline flow void0.920.0–10.057
AC0.430.0–2.044
PC0.300.0–0.944

Shift ipsilateral to the bur hole and shift contralateral to bur hole are compared. Amounts of shift measured in various locations, anterior to posterior, are compared. Most shift was identified in the frontal lobe surface on either side.

Frontal Lobe and Surface

In the anterior brain, the frontal lobe surface, midline flow voids, and distinct areas within the frontal lobe were identified. All points were anterior to the AC. The ipsilateral frontal lobe surface in unilateral procedures and the frontal lobe in bilateral procedures were noted to have the most amount of average shift with an average magnitude of 1.4 mm (range 0.00–10.1 mm) from 81 unique points. Surprisingly, in unilateral procedures the contralateral frontal surface had the second greatest amount of shift of all locations. This side is not exposed directly to the bur hole, yet the average shift was 1.0 mm (range 0.1–8.6 mm) from 12 unique points. Overall, in unilateral procedures, the ipsilateral side of the anterior brain showed more movement than the contralateral side when each location was compared with the corresponding opposite side.

Basal Ganglia—Deep Structures

The locations closest to the DBS target sites in the basal ganglia, the STN and the GPi, included recorded locations of VR spaces in the pallidum (VR pallidum), the border of third ventricle, and the AC-PC plane. On the ipsilateral side and in bilateral procedures, the largest average of shift was in the VR pallidum, which was 0.6 mm (range 0.0–2.1 mm). On the contralateral side in unilateral procedures, the largest average of shift was 0.3 mm in the VR pallidum and the border of the third ventricle (range of shift in the contralateral basal ganglia was 0.0–0.8 mm). The AC-PC points were also evaluated, and the greatest average of shift was 0.4 mm in the AC point (range 0.0–2.0 mm). Overall, the ipsilateral basal ganglia showed more movement than the contralateral side. In the deep structures, 4 patients (9%) had a shift of more than 2.0 mm, 9 patients (20%) had a shift of 1–1.9 mm, and 31 patients (70%) had a shift of less than 1 mm.

Occipital and Temporal Lobes

The points located in the occipital and temporal lobes had the least amount of shift. Anterior temporal lobe points included the anterior temporal lobe and sylvian fissure. Occipital and posterior temporal points were any point posterior to the posterior margin of the midbrain. The smallest amount of shift was noted in unilateral procedures on the posterior temporal/occipital lobe surface on the contralateral side (average 0.2 mm, range 0.0–0.9 mm). The ipsilateral sides of the occipital and temporal lobes showed more shift than the contralateral side. Overall, the temporal and occipital lobe structures had less shift than more anterior areas.

Correlation of Location and Shift

One hypothesis was that patients with the most shift in surface structures would also have more brain deformation in deeper structures. To investigate whether there was a correlation, we first identified one unique structure on the surface that had the most shift in each patient. In the same patient we also identified the deep structure with the maximum shift. Then the greatest surface shift values were plotted against the maximum shift of a deep structure in the same patient. A linear relationship was expected to show that when the frontal lobe and surface have a greater shift, those same patients would also have a greater shift in the deeper structures. These points were plotted (Fig. 2), but no correlation (R2 = 0.36) was noted between shift magnitude on the surface and in deep structures. Therefore, whether a patient has a large or small amount of frontal surface shift, the shift of the deeper structures remains unpredictable. There was also no correlation found between patient age or sex and shift magnitude.

Fig. 2.
Fig. 2.

Cortical surface versus deep pallidum shift. A plot of the maximum cortical surface shift versus the maximum deep pallidum shift in each patient shows no linear relationship.

Evaluation of Shift and Time

To evaluate for a correlation between procedure length and brain shift, we next examined the maximum deep and cortical shift during the time between the MRI studies. The amount of time between the first and last scans that were used for data gathering ranged from 40 to 123 minutes, with an average of 80 minutes. As in Fig. 2, the maximum deep and cortical shift was then plotted in Fig. 3, except now it was plotted versus time. In each case, deep and cortical, there was a significant correlation between time and shift (R2 = 0.04 and R2 = 0.04). The slope of the line (representing the rate of shift) shows that the superficial cortical structures moved at a greater rate than the deeper structures over the course of time.

Fig. 3.
Fig. 3.

Maximum deep and cortical brain shift versus time in between scans. The maximum cortical surface and deep structure shift are plotted against the time in between MRI scans where data points were gathered. The slope of the lines represents the rate of shift in mm/min.

Magnitude of Shift in Brain Quadrant

In each case, the magnitude of shift was calculated in each brain quadrant. Each quadrant was identified in the same method as was used for the control samples, with the midsagittal line and the anterior-posterior coronal centered at the AC-PC midpoint. To identify where the most shift was occurring, the magnitude of shift was plotted against the distance of that shift from the center of the brain, which was defined as the AC-PC midpoint (Fig. 4). The hypothesis was that points farther from the center of the brain were more likely to shift in the anterior quadrants as the surface sagged, however posteriorly, the shift would be more uniform. Figure 4 identifies the anterior ipsilateral quadrant or anterior quadrant in bilateral procedures as having the most brain movement and also the most variability of shift. The structures farther from the AC-PC midpoint, which include the frontal surface of the brain just beneath the bur hole, show the most scatter, indicating the most variation in brain shift. The contralateral anterior quadrant shows much less variation and more consistently less shift overall. In fact, with the exception of the anterior ipsilateral quadrant, all quadrants showed a relatively uniform shift throughout the brain. However, one point near the brain surface in Procedure 10 (Patient 7, unilateral procedure) had a large amount of movement. This patient had a shift on the ipsilateral frontal surface of 5.6 mm and a shift on the contralateral frontal surface of 7.8 mm. Interestingly, 1.5 months prior to this procedure, this patient underwent a unilateral DBS lead placement (his first DBS surgery). During that procedure, the maximum shift of the frontal surface was 0.3 mm. This further emphasizes the challenges associated with predicting brain shift.

Fig. 4.
Fig. 4.

Magnitude of shift in each brain quadrant related to distance from the AC-PC midpoint. Brain shift in millimeters is plotted against overall depth of a structure in each of the 4 quadrants, identified by the AC-PC midpoint. Note that the frontal ipsilateral quadrant shows the most shift and also the most variability of shift.

Evaluation of Pass/Target Ratio on Shift

In our series of 44 operations, a total of 73 targets were planned, and 82 passes of DBS electrodes were made to obtain acceptable electrode placement. The overall pass/target ratio was 1.12. Multiple passes are often made in various bur hole–based procedures including DBS surgery and stereotactic biopsies. This could cause a direct effect on shift, as this is the only time the surgeon is manipulating the brain. In Table 4, we evaluated the relationship between this ratio and overall shift. The effect of multiple passes on the ipsilateral and contralateral brain in unilateral approaches was evaluated, as well as a comparison of bilateral approaches with a pass/target ratio of 1 to an approach where it is greater than 1. In each comparison, more than one pass to the target resulted in a significant increase in shift on both the contralateral and ipsilateral side. It should be noted that the shift calculations were repeated by normalizing for the length of procedure, and the correlations remained the same. This data were normalized to the overall time of shift evaluation and are therefore reported in shift rate (mm/hr).

TABLE 4:

Relationship between number of passes/target and mean shift*

DescriptionNo. of Passes/Targetp Value
1>1
unilat bur hole
 contralat side shift rate in mm/hr0.3 (95% CI 0.2–0.4)0.8 (95% CI 0.0–1.7)0.02
 ipsilat side shift rate in mm/hr0.4 (95% CI 0.3–0.5)0.9 (95% CI 0.5–1.3)<0.001
bilat bur holes
 overall shift rate in mm/hr0.5 (95% CI 0.4–0.6)0.9 (95% CI 0.6–1.2)<0.001

Values are normalized for time. Patients who required more than 1 pass to obtain an acceptable target were found to have significantly more shifts on both ipsilateral and contralateral bur hole base procedures.

Shift in Unilateral Versus Bilateral Bur Hole Procedures

A comparison of the brain shift seen in patients who had unilateral bur holes versus bilateral bur holes for implantation of DBS leads is shown in Table 5. For patients who had bilateral bur holes, the average shift was 0.7 mm (95% CI 0.6–0.8 mm). For patients who underwent unilateral bur holes the shift was divided into the ipsilateral side and the contralateral side. The ipsilateral side had a shift of 0.4 mm, and the contralateral side had a shift of 0.3 mm. The data are reported in Table 5. In addition, consideration of the pass/target ratio was evaluated and controlled for to isolate the effect of multiple bur holes. On both the ipsilateral and contralateral side of the unilateral bur hole on the first pass there was less shift seen in comparison with total shift rate seen in the brain of a patient undergoing bilateral bur holes placement. However, there was no significant difference in shift on the ipsilateral side of the brain compared with bilateral bur hole and electrode placement when more than 1 pass was performed.

TABLE 5:

Mean shift in unilateral versus bilateral bur hole procedures*

DescriptionUnilat Bur HoleBilat Bur Holesp Value
1 pass
 contralat side shift in mm0.3 (95% CI 0.2–0.4)0.7 (95% CI 0.6–0.8)0.001
 ipsilat side shift in mm0.4 (95% CI 0.3–0.5)0.7 (95% CI 0.6–0.8)0.01
2 passes
 contralat side shift in mm0.9 (95% CI 0.0–1.9)1.1 (95% CI 0.7–1.5)0.70
 ipsilat side shift in mm1.1 (95% CI 0.7–1.5)1.2 (95% CI 0.8–1.6)0.68

Shift observed on the first pass was slightly less in unilateral bur hole procedures than in bilateral bur hole procedures. If more than 1 pass was performed, no significant difference was noted between shift in unilateral and bilateral bur hole procedures.

Direction of Brain Shift

The direction of shift was then investigated to determine if the deformation was primarily in the direction of gravity or in other directions.

Lateral and Anteroposterior Brain Shift

The direction and magnitude of shift was assessed by looking at the movement on the ipsilateral and contralateral side compared with the bur hole. Figure 5 identifies movement in the y position which is aligned in the antero/posterior direction, or the direction of gravity. Negative shift is movement of the structure in the direction of gravity, while positive shift is against gravity. This is plotted against the position of the structure in the anterioposterior direction on the x-axis. Locations located anterior to the AC-PC midpoint are located on the left and those in the occipital lobe are on the right of the chart. In the frontal lobes, on both the ipsilateral and contralateral sides, there is an overall direction of shift in the direction of gravity. However, the shift of structures located in the temporal and occipital lobes have a more unpredictable movement. Deformation in both the anterior and posterior directions is noted.

Fig. 5.
Fig. 5.

Magnitude of anteroposterior brain shift. The anterior and posterior shift is plotted against the anterior and posterior location in the brain. More shift is noted in the anterior ipsilateral brain.

Figure 6 identifies the magnitude of lateral shift of each unique location evaluated. As in Fig. 5, these structures are listed in their position from anterior to posterior. Medial movement is denoted by a bar line toward the middle of the figure and lateral movement is toward the lateral edge of the figure. Although the magnitude of shift appears greater in the anterior brain closer to the bur hole, the direction of movement is erratic and movement occurs in both medial and lateral directions.

Fig. 6.
Fig. 6.

Magnitude of lateral brain shift. The lateral and medial shift is plotted against the anterior and posterior location in the brain. More shift is noted in the anterior ipsilateral brain; however, the direction of shift moves in both directions throughout the brain.

Magnitude of Shift in Each Direction

The overall shift observed was 0.7 mm among all patients and all structures (Table 6). Additionally we evaluated the directional magnitude of shift in each scalar dimension. The movement of each structure was then deconstructed to the x and y directions. The majority of deformation is in the y direction, or the direction of gravity (0.5 mm) as seen in Table 6. Then, in each patient, structures with the most movement on either the surface or deep areas were identified. Points with the maximum movement were chosen to emphasize the overall movement in a specific area of the brain. Like the overall shift, the shift in the superficial and deep structures is also greatest in the y direction. Also noted is that the maximum surface shift is almost 3 times the maximum shift seen in the deep structures.

TABLE 6:

Magnitude of shift in directions x and y*

Unilat or BilatAverage Δx Shift in mmAverage Δy Shift in mmTotal Shift in mm
total brain0.3 (0.0–4.2)0.5 (0.0–9.9)0.7 (0.0–10.1)
max surface shift0.9 (0.0–4.2)1.7 (0.0–9.9)2.2 (0.2–10.1)
max pallidum shift0.4 (0.0–2.0)0.6 (0.0–2.0)0.8 (0.1–2.1)

Maximum shift was noted in the direction of gravity in surface and deep structures. Shift ranging from 0.1 to 2.1 mm was noted in deep structures.

The direction of shift of each point in the pallidum and putamen was graphed in Fig. 7 for 8 patients. The magnitude of each location is identified next to the arrow. Although there was a significant amount of shift in the direction of gravity, this was not the case at each location. Shift was recorded in all directions.

Fig. 7.
Fig. 7.

Pallidum and putamen shift. The magnitude and direction of specific deep structures were identified in 8 patients. Each vector was then placed on the specific location with the magnitude labeled. Shift is recorded in all directions.

Discussion

Here we describe a new technique for accurate measurement of brain shift in bur hole–based procedures. Kelly et al. were the first to report on brain shift in 1986 when they noted the movement of steel balls on the surface of the brain during the course of a craniotomy.12 Hill et al. and Roberts et al. continued this investigation further when they began to monitor intraoperative brain surface shifting.9,28 Initially, emphasis was placed on the cortical surface with high-resolution monitoring due to ease of access. Others have used optical flow methods and intraoperative imaging to measure shift both on the surface and in the subsurface structures.8,15,18,19,28 Since then, surgeons have attempted to predict, quantify, visualize, and minimize brain shift throughout their operative cases.

Visualization of brain shift during surgery has been attempted with several techniques. First, intraoperative ultrasonography11 and intraoperative MRI were used to evaluate brain deformation in 3D space.4,15,19,28 To date, intraoperative MRI, CT, and ultrasonography are the only methods available to obtain live information on brain shift. Of these three, MRI is the only modality with the sensitivity to outline the deep structures that are often the target during bur hole based procedures such as DBS or brain biopsy. The magnitude of the shift throughout the brain has been reported to have a significant range. Several studies have shown that shift in deep structures ranges from 0 to 4 mm8,13 and on the surface can be more than 10 mm.8,9,28 One study demonstrated brain shift toward the direction of gravity of approximately 1 mm without a cranial opening even being performed, merely due to position.9 Some reports suggest that shift is primarily in the direction of gravity;20 others report that shift is not correlated to gravity and is unpredictable.7 Brain shift studies also report that craniotomy size affects the magnitude of brain distortion.9 In contrast, other studies report that shift is independent of cranial opening and orientation of the patient.28 The collective conclusion that can be drawn is that shift is difficult to predict.

Magnitude and Quantification of Shift

Brain shift was observed in all locations of the brain in our series, despite the use of a small bur hole with minimal dural opening and brain manipulation. The largest magnitude of shift (10 mm) was observed at the frontal surface closest to the bur hole. This amount of shift can become critical if it is adjacent to eloquent surface cortical areas such as the motor or speech centers.19 Usually, superficial shift is not surprising, as it is clearly visible to the surgeon and therefore monitored carefully.19 Subcortical or deep shifting is more elusive and concerning as eloquent structures lie close together and this shift is not apparent to the surgeon.

In our study we looked at brain shift in 44 bur hole–based procedures and found the overall shift to be 0.7 mm on average (range 0–10 mm). In the deep structures we noted an average shift of 0.6 mm (range 0–2.1 mm). This is similar to other reported magnitudes of shift using different methodologies.8,9,13,29 These include studies that show shift in the deep structures ranges from 0 to 4 mm.8,13 On the surface we report a range of shift from 0 to 10 mm. Others groups have also focused on the shift of surface structures. Hill et al., as well as other investigators, also report maximum magnitude of shift at the surface of 10 mm.8,9,28

In functional neurosurgery, one well-described brain location of interest is the STN.5,6,27 Reports have shown that misplacement of an electrode by as little as 2 mm can cause ineffective stimulation and poor clinical outcome.10,21,27 In the deep structures surrounding the STN we found that 4 patients (9%) had recorded amounts of shift of more than 2 mm. Therefore, if we used purely anatomical electrode placement without the use of intraoperative MRI, at least 9% of our patients would be in jeopardy of having electrodes outside the optimal location.

Single Pass Versus Multiple Passes

During bur hole–based procedures, passage of an instrument is typically the sole moment when the surgeon will manipulate the brain. A single pass would be expected to cause the least amount of shift, but the question remains of how much a second pass effects brain shift. Our data show almost a doubling of shift rate when more than one pass is used in both unilateral and bilateral bur hole–based procedures. Preoperative planning and intraoperative imaging should both be used to target as efficiently and accurately as possible. If, however, an electrode is not adequately placed on first pass, a surgeon must realize that brain shift is continuing and may even worsen.

Unilateral Versus Bilateral Bur Holes

We analyzed the effects of bilateral and unilateral bur hole procedures on overall brain shift. A procedure with bilateral bur holes is likely to result in increased CSF loss and pneumocephalus and therefore may be associated with increased shift. Our data showed that in unilateral bur hole procedures, ipsilateral structures are likely to undergo an average of 0.4 mm of shift and contralateral structures 0.3 mm of shift during the first pass of a DBS electrode. Bilateral procedures averaged a slightly higher rate of 0.7 mm. Although this difference is small, it was significant. If more than one pass is performed, however, there is no difference between unilateral and bilateral bur hole procedures with respect to the amount of brain shift. Understanding that shift may occur in between electrode placement in bilateral procedures and that this shift is greater than in unilateral procedures again emphasizes the importance of intraoperative images to ensure target accuracy.

Direction of Shift

The most prominent direction of shift is expected to be greatest along the downward axis in line with gravity.28 Our data support gravity playing a dominant role in brain deformation. Each of our patients was positioned in the supine position. Considering this, one might expect the lateral shift to be minimal. However a lateral shift of 0.9 mm was found on the surface, and a lateral shift of 0.4 mm was found in the deep structures. Although these shifts are small, they are not inconsequential. In some cases, intraoperative shift is actually against gravity, further making shift extremely difficult to predict. Nabavi et al. found similar instances of shift in unexpected directions.18

Predictability of Shift

Our data provide evidence that there is no predictability in overall shift direction, rate, or magnitude when performing minimal bur hole procedures. Despite small trends in localized areas of the brain, any structure can move in any direction during surgery. If we also consider the many physiological changes that are not accounted for in this study, such as intracranial pressure, systolic blood pressure, CO2 levels, and brain edema, prediction of shift seems almost impossible.19 In addition to all these parameters, the length of surgery also affects the total amount of shift that will occur. We show that the amount of shift and the rate of shift can be different for the superficial and deep cortical structures, further complicating any prediction. Despite these dilemmas, multiple brain shift prediction models have been derived and studied.16,25 Many models demonstrate an almost monotonous unidirectional motion primarily accounting for CSF drainage and gravity, which is not supported by our data. One promising method modifies preoperative imaging with several strategic intraoperative points to increase accuracy.28 Modeling shows promising data, but currently these algorithms are still unable to fully predict brain shift accurately enough for clinical use.30

Eventually as more intraoperative MRI data are accumulated, better predictive brain simulation algorithms might be able to predict all aspects of brain shift in different pathological states such as brain tumors. These will need to include variables including tumor type and density, intracranial structures, edema, intracranial pressure, osmotic agents, positioning, etc. As of now, this is not yet available, and therefore we must rely on other modalities such as intraoperative updating of images for consistently accurate and reliable neuronavigation.

Accuracy and Error

Accuracy of any stereotactic procedure is of the utmost importance. Image distortion of the MRI scan used for targeting is one source that has been proposed to cause error. Multiple studies have shown that this distortion is minimal at 1.5-T field strength (less than 1–2 mm) and that, despite this distortion, accuracy of device placement can be achieved.23,32 Other factors that can affect error include slice thickness and voxel acquisition size. In our study, we report data in the x and y coordinates. The shift in the z direction is underestimated, as this vector is primarily out of plane with the axially acquired MRI slices used for analysis. Axial MRI is most sensitive for evaluation of shift movement in the x and y directions; since the majority of shift recorded is in the y direction, it is likely that the magnitude of shift in the z direction has a minimal overall effect on the total shift. The voxel size in the MR images used in our study was 1 mm. Many of the shift measurements we report are less than 1 mm, especially in the deep structures, which may lead some to question the accuracy of such small measurements.26 However, measurement with subvoxel accuracy when the target structure is larger than the voxel size is a well-established concept in image-guided stereotactic surgery.14 The target structure only occupies a portion of the voxels around its boundary, and the signal intensity that each of these voxels reports is proportional to the fraction of the pixel that does not contain the object (this is the concept of “signal averaging” at the voxel level). The center of the object can therefore be determined based on the partial filling characteristics of these pixels on the periphery, and subvoxel accuracy can be achieved.

Finally, in contrast to many other brain shift investigations, the movement of unique structures in this study is not evaluated and analyzed by computer programs but by hand measurements on the MR console. Although our method eliminates the errors associated with computer modeling, there is a small amount of human error in the measurement of each structure shift in our experimental method. Despite these limitations, confidence of accuracy in our study was provided by the use of control objects in the bony calvaria that would not be expected to shift. One hundred fifty-three controls were measured, with an overall shift of 0.1 mm. Therefore, the error introduced by human measuring was minimal and substantially less than the reported shift values.

Limitations of Study

For this investigation it should be mentioned that there is an initial shift that remains uncaptured. The high-resolution serial images used in this study were obtained after bur hole placement; the initial scans on average were obtained 20 minutes or more after the dura was opened. This initial shift is likely to occur immediately after dural opening. Therefore, in these cases the actual shift was likely greater than the measured shift. Because our targeting MRI occurred after this initial shift had occurred, this did not affect the overall ability to implant the DBS electrodes at the desired target. The fact that targeting is performed on a scan obtained after initial brain shift has occurred is one of the potential advantages of the interventional MRI-guided technique.

Prevention of Shift

Multiple factors causing brain shift can be controlled throughout bur hole–based procedures performed without intraoperative imaging, during which brain shift cannot be detected or monitored in real time. Shift is typically initiated by intracranial air entry, which may be due to loss of CSF and/or due to equalization of intracranial and atmospheric pressure.2,16,17 Other causes of shift specific to stereotactic surgery include the deformation of tissue caused by the tip of a rigid electrode during implantation. Although not tested in our study, several reports suggest ways to reduce brain shift. Cerebrospinal fluid loss and pneumocephalus can be reduced by minimizing the time from dural opening to lead implantation, flooding the bur hole with saline irrigation after opening, avoiding suctioning of CSF, sealing the dural opening with fibrin glue as soon as practical, and reducing factors contributing to negative intracranial pressures, including hypovolemia and hypocapnia.22,24,33 Also, by avoiding the sulci and ventricular walls and ensuring adequate pial opening, a surgeon can decrease any resistance when advancing a catheter, electrode, or other device.3,34 One report even identifies a new bur hole technique that minimizes pneumocephalus and therefore reduces brain shift.1 Finally, our data show that reducing the number of passes to reach a target can reduce shift. Minimizing passes can be achieved with careful preoperative planning and with the help of intraoperative imaging. With real-time, intraoperative image-guided approaches such as MRI and good surgical technique, brain shift can be monitored and accounted for, which may reduce the effect of shift on clinical outcome.26

Conclusions

We describe a novel way of measuring brain shift throughout superficial and deep brain structures in bur hole–based procedures where interventional MRI was used for implanting DBS electrodes. Brain shift is continual, unpredictable, and can render traditional stereotactic targeting based on preoperative imaging inaccurate even in deep brain structures such as those used for DBS. Intraoperative MRI has the advantage of providing the most updated images for targeting and monitoring of device placement and may be the best method to compensate for effects of brain shift during surgery.

Disclosure

Dr. Martin reports receiving support from MRI Interventions for non–study-related clinical or research efforts. Mr. Sootman reports an employee relationship with Medtronic. Dr. Larson reports receiving support from MRI Interventions for the study described and honoraria from Medtronic. Dr. Starr reports receiving research funding from MRI Interventions.

Author contributions to the study and manuscript preparation include the following. Conception and design: Ivan, Martin, Larson. Acquisition of data: Ivan, Yarlagadda, Saxena, Martin, Starr, Larson. Analysis and interpretation of data: Ivan, Yarlagadda, Saxena, Larson. Drafting the article: Ivan, Yarlagadda. Critically revising the article: all authors. Reviewed submitted version of manuscript: all authors. Approved the final version of the manuscript on behalf of all authors: Ivan. Statistical analysis: Ivan. Administrative/technical/material support: Sootsman. Study supervision: Martin, Larson.

This article contains some figures that are displayed in color online but in black-and-white in the print edition.

References

  • 1

    Coenen VAAbdel-Rahman AMcMaster JBogod NHoney CR: Minimizing brain shift during functional neurosurgical procedures – a simple burr hole technique that can decrease CSF loss and intracranial air. Cent Eur Neurosurg 72:1811852011

    • Search Google Scholar
    • Export Citation
  • 2

    Elias WJFu KMFrysinger RC: Cortical and subcortical brain shift during stereotactic procedures. J Neurosurg 107:9839882007

  • 3

    Elias WJSansur CAFrysinger RC: Sulcal and ventricular trajectories in stereotactic surgery. Clinical article. J Neurosurg 110:2012072009

    • Search Google Scholar
    • Export Citation
  • 4

    Ganser KADickhaus HStaubert ABonsanto MMWirtz CRTronnier VM: [Quantification of brain shift effects in MRI images.]. Biomed Tech (Berl) 42:Suppl2472481997. (Ger)

    • Search Google Scholar
    • Export Citation
  • 5

    Guehl DEdwards RCuny EBurbaud PRougier AModolo J: Statistical determination of the optimal subthalamic nucleus stimulation site in patients with Parkinson disease. J Neurosurg 106:1011102007

    • Search Google Scholar
    • Export Citation
  • 6

    Hamid NAMitchell RDMocroft PWestby GWMilner JPall 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 76:4094142005

    • Search Google Scholar
    • Export Citation
  • 7

    Hartkens THill DLCastellano-Smith ADHawkes DJMaurer CR JrMartin AJ: Measurement and analysis of brain deformation during neurosurgery. IEEE Trans Med Imaging 22:82922003

    • Search Google Scholar
    • Export Citation
  • 8

    Hata NNabavi AWells WM IIIWarfield SKKikinis RBlack PM: Three-dimensional optical flow method for measurement of volumetric brain deformation from intraoperative MR images. J Comput Assist Tomogr 24:5315382000

    • Search Google Scholar
    • Export Citation
  • 9

    Hill DLMaurer CR JrMaciunas RJBarwise JAFitzpatrick JMWang MY: Measurement of intraoperative brain surface deformation under a craniotomy. Neurosurgery 43:5145281998

    • Search Google Scholar
    • Export Citation
  • 10

    Holl EMPetersen EAFoltynie TMartinez-Torres ILimousin PHariz MI: Improving targeting in image-guided frame-based deep brain stimulation. Neurosurgery 67:2 Suppl Operative4374472010

    • Search Google Scholar
    • Export Citation
  • 11

    Jödicke ADeinsberger WErbe HKriete ABöker DK: Intraoperative three-dimensional ultrasonography: an approach to register brain shift using multidimensional image processing. Minim Invasive Neurosurg 41:13191998

    • Search Google Scholar
    • Export Citation
  • 12

    Kelly PJKall BAGoerss SEarnest F IV: Computer-assisted stereotaxic laser resection of intra-axial brain neoplasms. J Neurosurg 64:4274391986

    • Search Google Scholar
    • Export Citation
  • 13

    Khan MFMewes KGross RESkrinjar O: Assessment of brain shift related to deep brain stimulation surgery. Stereotact Funct Neurosurg 86:44532008

    • Search Google Scholar
    • Export Citation
  • 14

    Maciunas RJGalloway RL JrLatimer JW: The application accuracy of stereotactic frames. Neurosurgery 35:6826951994

  • 15

    Maurer CR JrHill DLMartin AJLiu HMcCue MRueckert D: Investigation of intraoperative brain deformation using a 1.5-T interventional MR system: preliminary results. IEEE Trans Med Imaging 17:8178251998

    • Search Google Scholar
    • Export Citation
  • 16

    Miga MIRoberts DWHartov AEisner SLemery JKennedy FE: Updated neuroimaging using intraoperative brain modeling and sparse data. Stereotact Funct Neurosurg 72:1031061999

    • Search Google Scholar
    • Export Citation
  • 17

    Miyagi YShima FSasaki T: Brain shift: an error factor during implantation of deep brain stimulation electrodes. J Neurosurg 107:9899972007

    • Search Google Scholar
    • Export Citation
  • 18

    Nabavi ABlack PMGering DTWestin CFMehta VPergolizzi RS Jr: Serial intraoperative magnetic resonance imaging of brain shift. Neurosurgery 48:7877982001

    • Search Google Scholar
    • Export Citation
  • 19

    Nimsky CGanslandt OCerny SHastreiter PGreiner GFahlbusch R: Quantification of, visualization of, and compensation for brain shift using intraoperative magnetic resonance imaging. Neurosurgery 47:107010802000

    • Search Google Scholar
    • Export Citation
  • 20

    Obuchi TKatayama YKobayashi KOshima HFukaya CYamamoto T: Direction and predictive factors for the shift of brain structure during deep brain stimulation electrode implantation for advanced Parkinson's disease. Neuromodulation 11:3023102008

    • Search Google Scholar
    • Export Citation
  • 21

    Papavassiliou ERau GHeath SAbosch ABarbaro NMLarson PS: Thalamic deep brain stimulation for essential tremor: relation of lead location to outcome. Neurosurgery 54:112011302004

    • Search Google Scholar
    • Export Citation
  • 22

    Patel NKHeywood PO'Sullivan KLove SGill SS: MRI-directed subthalamic nucleus surgery for Parkinson's disease. Stereotact Funct Neurosurg 78:1321452002

    • Search Google Scholar
    • Export Citation
  • 23

    Patel NKPlaha PGill SS: Magnetic resonance imaging-directed method for functional neurosurgery using implantable guide tubes. Neurosurgery 61:5 Suppl 23583662007

    • Search Google Scholar
    • Export Citation
  • 24

    Patel NKPlaha PO'Sullivan KMcCarter RHeywood PGill SS: MRI directed bilateral stimulation of the subthalamic nucleus in patients with Parkinson's disease. J Neurol Neurosurg Psychiatry 74:163116372003

    • Search Google Scholar
    • Export Citation
  • 25

    Paulsen KDMiga MIKennedy FEHoopes PJHartov ARoberts DW: A computational model for tracking subsurface tissue deformation during stereotactic neurosurgery. IEEE Trans Biomed Eng 46:2132251999

    • Search Google Scholar
    • Export Citation
  • 26

    Petersen EAHoll EMMartinez-Torres IFoltynie TLimousin PHariz MI: Minimizing brain shift in stereotactic functional neurosurgery. Neurosurgery 67:3 Suppl Operativeons213ons2212010

    • Search Google Scholar
    • Export Citation
  • 27

    Richardson RMOstrem JLStarr PA: Surgical repositioning of misplaced subthalamic electrodes in Parkinson's disease: location of effective and ineffective leads. Stereotact Funct Neurosurg 87:2973032009

    • Search Google Scholar
    • Export Citation
  • 28

    Roberts DWHartov AKennedy FEMiga MIPaulsen KD: Intraoperative brain shift and deformation: a quantitative analysis of cortical displacement in 28 cases. Neurosurgery 43:7497601998

    • Search Google Scholar
    • Export Citation
  • 29

    Roberts DWMiga MIHartov AEisner SLemery JMKennedy FE: Intraoperatively updated neuroimaging using brain modeling and sparse data. Neurosurgery 45:119912071999

    • Search Google Scholar
    • Export Citation
  • 30

    Skrinjar ONabavi ADuncan J: Model-driven brain shift compensation. Med Image Anal 6:3613732002

  • 31

    Starr PAMartin AJLarson PS: Implantation of deep brain stimulator electrodes using interventional MRI. Neurosurg Clin N Am 20:1932032009

    • Search Google Scholar
    • Export Citation
  • 32

    Sumanaweera TSAdler JRGlover GHHemler PFvan den Elsen PAMartin D: Method for correcting magnetic resonance image distortion for frame-based stereotactic surgery, with preliminary results. J Image Guid Surg 1:1511571995

    • Search Google Scholar
    • Export Citation
  • 33

    Zonenshayn MSterio DKelly PJRezai ARBeric A: Location of the active contact within the subthalamic nucleus (STN) in the treatment of idiopathic Parkinson's disease. Surg Neurol 62:2162262004

    • Search Google Scholar
    • Export Citation
  • 34

    Zrinzo Lvan Hulzen ALGorgulho AALimousin PStaal MJDe Salles AA: Avoiding the ventricle: a simple step to improve accuracy of anatomical targeting during deep brain stimulation. Clinical article. J Neurosurg 110:128312902009

    • Search Google Scholar
    • Export Citation

If the inline PDF is not rendering correctly, you can download the PDF file here.

Article Information

Contributor Notes

Address correspondence to: Michael E. Ivan, M.D., 505 Parnassus Ave., Box 0112, San Francisco, CA 94117. email: ivanm@neurosurg.ucsf.edu.Please include this information when citing this paper: published online May 2, 2014; DOI: 10.3171/2014.3.JNS121312.
Headings
Figures
  • View in gallery

    Comparison of unique structures using ROIs. A T2-weighted axial MR image is obtained at 2 separate time points during surgery. On the left is the pre–electrode placement image used for targeting. On the right is the post–bilateral electrode placement image. Each red outline is a specific ROI identifying a structure used to calculated shift. Obvious shift can be noted in the left frontal lobe surface (circled).

  • View in gallery

    Cortical surface versus deep pallidum shift. A plot of the maximum cortical surface shift versus the maximum deep pallidum shift in each patient shows no linear relationship.

  • View in gallery

    Maximum deep and cortical brain shift versus time in between scans. The maximum cortical surface and deep structure shift are plotted against the time in between MRI scans where data points were gathered. The slope of the lines represents the rate of shift in mm/min.

  • View in gallery

    Magnitude of shift in each brain quadrant related to distance from the AC-PC midpoint. Brain shift in millimeters is plotted against overall depth of a structure in each of the 4 quadrants, identified by the AC-PC midpoint. Note that the frontal ipsilateral quadrant shows the most shift and also the most variability of shift.

  • View in gallery

    Magnitude of anteroposterior brain shift. The anterior and posterior shift is plotted against the anterior and posterior location in the brain. More shift is noted in the anterior ipsilateral brain.

  • View in gallery

    Magnitude of lateral brain shift. The lateral and medial shift is plotted against the anterior and posterior location in the brain. More shift is noted in the anterior ipsilateral brain; however, the direction of shift moves in both directions throughout the brain.

  • View in gallery

    Pallidum and putamen shift. The magnitude and direction of specific deep structures were identified in 8 patients. Each vector was then placed on the specific location with the magnitude labeled. Shift is recorded in all directions.

References
  • 1

    Coenen VAAbdel-Rahman AMcMaster JBogod NHoney CR: Minimizing brain shift during functional neurosurgical procedures – a simple burr hole technique that can decrease CSF loss and intracranial air. Cent Eur Neurosurg 72:1811852011

    • Search Google Scholar
    • Export Citation
  • 2

    Elias WJFu KMFrysinger RC: Cortical and subcortical brain shift during stereotactic procedures. J Neurosurg 107:9839882007

  • 3

    Elias WJSansur CAFrysinger RC: Sulcal and ventricular trajectories in stereotactic surgery. Clinical article. J Neurosurg 110:2012072009

    • Search Google Scholar
    • Export Citation
  • 4

    Ganser KADickhaus HStaubert ABonsanto MMWirtz CRTronnier VM: [Quantification of brain shift effects in MRI images.]. Biomed Tech (Berl) 42:Suppl2472481997. (Ger)

    • Search Google Scholar
    • Export Citation
  • 5

    Guehl DEdwards RCuny EBurbaud PRougier AModolo J: Statistical determination of the optimal subthalamic nucleus stimulation site in patients with Parkinson disease. J Neurosurg 106:1011102007

    • Search Google Scholar
    • Export Citation
  • 6

    Hamid NAMitchell RDMocroft PWestby GWMilner JPall 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 76:4094142005

    • Search Google Scholar
    • Export Citation
  • 7

    Hartkens THill DLCastellano-Smith ADHawkes DJMaurer CR JrMartin AJ: Measurement and analysis of brain deformation during neurosurgery. IEEE Trans Med Imaging 22:82922003

    • Search Google Scholar
    • Export Citation
  • 8

    Hata NNabavi AWells WM IIIWarfield SKKikinis RBlack PM: Three-dimensional optical flow method for measurement of volumetric brain deformation from intraoperative MR images. J Comput Assist Tomogr 24:5315382000

    • Search Google Scholar
    • Export Citation
  • 9

    Hill DLMaurer CR JrMaciunas RJBarwise JAFitzpatrick JMWang MY: Measurement of intraoperative brain surface deformation under a craniotomy. Neurosurgery 43:5145281998

    • Search Google Scholar
    • Export Citation
  • 10

    Holl EMPetersen EAFoltynie TMartinez-Torres ILimousin PHariz MI: Improving targeting in image-guided frame-based deep brain stimulation. Neurosurgery 67:2 Suppl Operative4374472010

    • Search Google Scholar
    • Export Citation
  • 11

    Jödicke ADeinsberger WErbe HKriete ABöker DK: Intraoperative three-dimensional ultrasonography: an approach to register brain shift using multidimensional image processing. Minim Invasive Neurosurg 41:13191998

    • Search Google Scholar
    • Export Citation
  • 12

    Kelly PJKall BAGoerss SEarnest F IV: Computer-assisted stereotaxic laser resection of intra-axial brain neoplasms. J Neurosurg 64:4274391986

    • Search Google Scholar
    • Export Citation
  • 13

    Khan MFMewes KGross RESkrinjar O: Assessment of brain shift related to deep brain stimulation surgery. Stereotact Funct Neurosurg 86:44532008

    • Search Google Scholar
    • Export Citation
  • 14

    Maciunas RJGalloway RL JrLatimer JW: The application accuracy of stereotactic frames. Neurosurgery 35:6826951994

  • 15

    Maurer CR JrHill DLMartin AJLiu HMcCue MRueckert D: Investigation of intraoperative brain deformation using a 1.5-T interventional MR system: preliminary results. IEEE Trans Med Imaging 17:8178251998

    • Search Google Scholar
    • Export Citation
  • 16

    Miga MIRoberts DWHartov AEisner SLemery JKennedy FE: Updated neuroimaging using intraoperative brain modeling and sparse data. Stereotact Funct Neurosurg 72:1031061999

    • Search Google Scholar
    • Export Citation
  • 17

    Miyagi YShima FSasaki T: Brain shift: an error factor during implantation of deep brain stimulation electrodes. J Neurosurg 107:9899972007

    • Search Google Scholar
    • Export Citation
  • 18

    Nabavi ABlack PMGering DTWestin CFMehta VPergolizzi RS Jr: Serial intraoperative magnetic resonance imaging of brain shift. Neurosurgery 48:7877982001

    • Search Google Scholar
    • Export Citation
  • 19

    Nimsky CGanslandt OCerny SHastreiter PGreiner GFahlbusch R: Quantification of, visualization of, and compensation for brain shift using intraoperative magnetic resonance imaging. Neurosurgery 47:107010802000

    • Search Google Scholar
    • Export Citation
  • 20

    Obuchi TKatayama YKobayashi KOshima HFukaya CYamamoto T: Direction and predictive factors for the shift of brain structure during deep brain stimulation electrode implantation for advanced Parkinson's disease. Neuromodulation 11:3023102008

    • Search Google Scholar
    • Export Citation
  • 21

    Papavassiliou ERau GHeath SAbosch ABarbaro NMLarson PS: Thalamic deep brain stimulation for essential tremor: relation of lead location to outcome. Neurosurgery 54:112011302004

    • Search Google Scholar
    • Export Citation
  • 22

    Patel NKHeywood PO'Sullivan KLove SGill SS: MRI-directed subthalamic nucleus surgery for Parkinson's disease. Stereotact Funct Neurosurg 78:1321452002

    • Search Google Scholar
    • Export Citation
  • 23

    Patel NKPlaha PGill SS: Magnetic resonance imaging-directed method for functional neurosurgery using implantable guide tubes. Neurosurgery 61:5 Suppl 23583662007

    • Search Google Scholar
    • Export Citation
  • 24

    Patel NKPlaha PO'Sullivan KMcCarter RHeywood PGill SS: MRI directed bilateral stimulation of the subthalamic nucleus in patients with Parkinson's disease. J Neurol Neurosurg Psychiatry 74:163116372003

    • Search Google Scholar
    • Export Citation
  • 25

    Paulsen KDMiga MIKennedy FEHoopes PJHartov ARoberts DW: A computational model for tracking subsurface tissue deformation during stereotactic neurosurgery. IEEE Trans Biomed Eng 46:2132251999

    • Search Google Scholar
    • Export Citation
  • 26

    Petersen EAHoll EMMartinez-Torres IFoltynie TLimousin PHariz MI: Minimizing brain shift in stereotactic functional neurosurgery. Neurosurgery 67:3 Suppl Operativeons213ons2212010

    • Search Google Scholar
    • Export Citation
  • 27

    Richardson RMOstrem JLStarr PA: Surgical repositioning of misplaced subthalamic electrodes in Parkinson's disease: location of effective and ineffective leads. Stereotact Funct Neurosurg 87:2973032009

    • Search Google Scholar
    • Export Citation
  • 28

    Roberts DWHartov AKennedy FEMiga MIPaulsen KD: Intraoperative brain shift and deformation: a quantitative analysis of cortical displacement in 28 cases. Neurosurgery 43:7497601998

    • Search Google Scholar
    • Export Citation
  • 29

    Roberts DWMiga MIHartov AEisner SLemery JMKennedy FE: Intraoperatively updated neuroimaging using brain modeling and sparse data. Neurosurgery 45:119912071999

    • Search Google Scholar
    • Export Citation
  • 30

    Skrinjar ONabavi ADuncan J: Model-driven brain shift compensation. Med Image Anal 6:3613732002

  • 31

    Starr PAMartin AJLarson PS: Implantation of deep brain stimulator electrodes using interventional MRI. Neurosurg Clin N Am 20:1932032009

    • Search Google Scholar
    • Export Citation
  • 32

    Sumanaweera TSAdler JRGlover GHHemler PFvan den Elsen PAMartin D: Method for correcting magnetic resonance image distortion for frame-based stereotactic surgery, with preliminary results. J Image Guid Surg 1:1511571995

    • Search Google Scholar
    • Export Citation
  • 33

    Zonenshayn MSterio DKelly PJRezai ARBeric A: Location of the active contact within the subthalamic nucleus (STN) in the treatment of idiopathic Parkinson's disease. Surg Neurol 62:2162262004

    • Search Google Scholar
    • Export Citation
  • 34

    Zrinzo Lvan Hulzen ALGorgulho AALimousin PStaal MJDe Salles AA: Avoiding the ventricle: a simple step to improve accuracy of anatomical targeting during deep brain stimulation. Clinical article. J Neurosurg 110:128312902009

    • Search Google Scholar
    • Export Citation
TrendMD
Cited By
Metrics

Metrics

All Time Past Year Past 30 Days
Abstract Views 0 0 0
Full Text Views 529 379 18
PDF Downloads 351 212 5
EPUB Downloads 0 0 0
PubMed
Google Scholar