Introduction
Lars Leksell introduced his first stereotactic frame in 1949 [1]. Despite numerous modifications over the intervening years, the design and function of the Leksell frame continues to be based on the arc-centered principle [2]. Arc-centered frames position the surgical target in the center of an operating arc. So positioned, the target can be reached from any entry point, as long as the surgical probe is advanced the distance of the radius. This simple yet effective design provides versatility, ease of use, and great accuracy.
The most current Leksell frame, the Model G, is compatible with both magnetic resonance imaging (MRI) and computerized tomography (CT). The G frame may be used for image-guided biopsy or tumor resection, radio-surgery with the Gamma Knife, or so-called "functional” neurosurgical procedures. This topic discusses the use of the Leksell Model G Frame for performing movement disorder surgery. Space limitations restrict the discussion to a description of the author’s technique for performing MRI-guided neuroablative and deep brain stimulation (DBS) procedures. An understanding of the basic use of the Leksell frame is assumed.
Technique For Movement Disorder Surgery
Ideally, movement disorder patients are awake during surgery to facilitate microelectrode recording (MER) and to permit constant monitoring of the patient’s neurological status. If the patient has Parkinson’s disease (PD), antiparkinsonian medications are withheld, beginning 12 hours before surgery, also to facilitate MER and to permit assessment of the patient’s response to stimulation, ablation, or both. Cessation of levodopa/carbidopa preparations may lead to rebound hypertension, and it is essential that the anesthesiologist maintains strict blood pressure control, keeping the systolic pressure less than 140 mm Hg, to minimize the risk of intracerebral hemorrhage.
Frame Application
The importance of careful frame application to the success of functional neurosurgical procedures (especially if one foregoes the assistance of an independent stereotactic targeting workstation) is often underestimated. Targeting adjustments are most easily made when the axial targeting images run parallel to the intercommissural (IC) plane and when there is no sideward tilt (roll) or rotation of the frame relative to the head (yaw). Only when the frame is positioned in this manner do adjustments to one frame coordinate translate into anatomical adjustments exclusively in that direction. The ear bars provided with the fame assist greatly toward this end by holding the frame steady relative to the head until it is fixed in place with the skull pins. The ear bars are uncomfortable and the inner ear is difficult to anesthetize; however, the discomfort lasts only for the brief period required to apply the frame (5 to 10 minutes) and is tolerated by most patients. Patients with low pain tolerances or significant head tremor are sedated with propofol for this stage of the procedure.
When targeting with MRI, it is best to use the lowermost pair of holes provided for the ear bars, as this elevates the frame as much as possible from the shoulders, allowing room for the bulky MRI adapter. In addition, skull pins should be selected whose lengths do not extend beyond the margins of the MRI fiducial box. Pins that extend beyond this border may prevent the frame from fitting within the head coil. The manufacturer sells a set of reusable MRI-compatible pins of varying lengths to address this need.
Finally, the frame should be fixed to the head so that the lateral bars of the base ring (i.e., the Y-axis) are parallel to the zygoma, the anteroposterior (AP) angle of which closely approximates that of the IC line. Many human stereotactic atlases employ the IC plane as the central median to which the deep brain structures are related. Therefore, by affixing the frame parallel to the IC line, the acquired axial images will run parallel to the IC plane, and atlas-derived measurements may be more reliably used for MRI-based targeting. Figure 1 demonstrates proper fixation of the Leksell frame.
The Targeting MRI
The MRI fiducial box and MRI adapter are attached to the frame, and the head is scanned with the frame aligned orthogonal to the scanner axis. A midsagittal Tl-weighted MRI is performed and the IC distance is measured. Contiguous 3-mm thick, axial fast spin echo/inversion recovery (FSE/IR) images (Table 1) are obtained through the region of the IC plane. These images beautifully display the deep brain structures and are reported to resist distortion secondary to the magnetic susceptibility effect [3]. After image acquisition, the "inverse video” function generates the image used for targeting (Fig. 2). If the frame is applied correctly, the anterior and posterior commissures (AC and PC, respectively) will be visible on the same or adjacent slices.
Figure 1 Application of the Head Frame. The head frame is secured to the skull with fixation screws at four points. The head is centered in the frame and the frame is aligned in the anteroposterior direction with the zygoma, approximating the angle of the intercommissural plane.
The distances between the posterior and middle fiducial markers on the lateral localizing plates are measured to confirm that the axial slices are, in fact, orthogonal to the vertical axis of the frame. The difference between these distances should be no more than 2 mm. If the difference is greater than 2 mm, purely transverse images are not being obtained and the frame should be repositioned in the scanner.
Determining the Stereotactic Coordinates of the Surgical Target
Regardless of the site to be targeted, the author initially determines the coordinates of the AC and PC at the midline. This is done for two reasons: (1) to assess the degree of head rotation relative to the frame, and (2) because the stereotactic coordinates of functional neurosurgical targets are determined by their known relationship to the commissures. The degree of head rotation relative to the frame is represented by the inverse tangent of the difference in the X values of the AC and PC divided by the IC distance
When the frame is properly affixed,
Table 1 MRI Scanning Parameters for Stereotactic Targeting
|
Parameter |
Inversion recovery |
Coronal T2-weighted |
|
Relaxation time (TR) |
4000 |
4000 |
|
Echo time (TE) |
17 |
102 |
|
Inversion time (T,) |
140 |
N/A |
|
No. excitations |
2 |
3 |
|
Echo train length |
8 |
12 |
|
Field of view |
24 cm |
24 cm |
|
Matrix |
256 X 192 |
256 X 256 |
|
Slice thickness (mm) |
3 |
2 |
|
Slice interval (mm) |
0 |
0 (interleaved) |
|
Scan type |
Fast scan |
Fast scan |
|
Inversion recovery, 2D |
Flow compensation |
Magnetic resonance image scanning parameters for stereotactic targeting for functional neurosurgical procedures. Scans are performed on a General Electric Echo Speed Scanner (1.5 Tesla).
Figure 2 Fast Spin Echo/Inversion Recovery Magnetic Resonance Imaging. Axial image at the level of the intercommissural plane demonstrating targeting of the ventrolateral (VL) nucleus of the thalamus. Diagonals connecting the opposing anterior and posterior fiducial markers have been constructed to define the center of the targeting area. Anterior/posterior and right/left axes (Y and X axes, respectively) have been constructed with their origins at the center point. The difference in the distance between the posterior and center fiducial markers on either side of the head (segments 1 and 2) is less than 2 mm. The VL target is 5 mm anterior to and 14 mm lateral to the center point of the posterior commissure, which is clearly visualized.
Once the coordinates of the commissures have been determined, the coordinates of the surgical target are calculated based on the anatomical relationship of the target to the IC plane, as documented in various publications [4-9] and atlases of the human brain.
Targeting the Ventral Lateral Nucleus for Medically Refractory Tremor
The ventral lateral (VL) nucleus can be targeted exclusively with axial FSE/ IR images when performing thalamotomy or thalamic DBS lead insertions. The image that best demonstrates the PC is selected. The target for tremor suppression is located 20% to 25% of the IC distance anterior to the PC (typically 5 to 6 mm), and 2 to 3 mm dorsal to the IC plane [4-6]. Target laterality (i.e., the X coordinate) is selected according to the body part that is to be treated. The ventrolateral nucleus of the thalamus is somatotopically organized, medio laterally [4], such that the face is represented medially (10-12 mm lateral of midline), the hand just lateral to the face (13-15 mm), and the leg/foot most lateral, abutting the internal capsule (14-17 mm). The third ventricular and thalamic widths also influence target laterality. When a wide third ventricle is encountered (i.e., >5 mm), laterality should be measured from the ipsilateral wall of the third ventricle and not the midline.
After determining the coordinates for the PC, the target coordinates are calculated as follows:
Even though the VL target is most commonly located 2 to 3 mm superior to the ventral border of the thalamus [4], the author targets to the depth of the PC to ensure that the surgical trajectory will pass completely through VL. The final depth for lesion or DBS lead placement is based on intraoperative neurophysiology.
Targeting Globus Pallidus Pars Internus (GPi)
The pallidum may also be targeted exclusively with axial FSE/IR images, except that the coordinates for GPi are calculated relative to the mid-com-missural point (MCP), the coordinates of which are determined by calculating the means of the coordinates of the commissures. The coordinates for the posteroventral globus pallidus pars internus (GPi) (as originally described by Leksell [7]) are then calculated as follows:
When targeting Gpi, one may also obtain coronal T2-weighted or FSE/ IR images to examine target depth relative to the optic tract, which lies immediately inferior to GPi.
Targeting Subthalamic Nucleus for Parkinson’s Disease
Unlike the thalamus and globus pallidus, the STN is poorly visualized on FSE/IR images; however, the coordinates for STN can be reliably calculated from its relationship with the MCP [9]. Alternatively, the STN may be visualized (although inconsistently), and directly targeted, on thin-cut coronal T2-weighted images (Table 1). Consult the Leksell G Frame manual for instructions on determining target coordinates from coronal MRI.
The coordinates for STN relative to the MCP are calculated as follows:
Operative Technique
After image acquisition and surgical planning, the patient is taken to the operating room and positioned supine on the operating table, which is configured as a reclining chair. The head is immobilized and prepped with betadine, and the field is sterilely draped.
The MRI-derived coordinates and the surgical trajectory are set on the frame. When targeting VL thalamus or STN, the AP angle of approach is 60° to 70° relative to the IC plane, as this approximates the angle of the ventral intermediate nucleus (the subdivision of VL that is targeted for tremor control) of the STN relative to the IC line. An AP angle of 50° to 60° is used for pallidal procedures. Ideally, the lateral angle of approach for all of these procedures would be 90° (i.e., directly vertical), as this would generate a pure parasagittal trajectory, allowing intraoperative neurophy-siological data to be most easily correlated to the parasagittal sections provided in human stereotactic atlases. Moreover, because the somatotopic representation of the contralateral hemibody in the thalamus is oriented in a medial to lateral direction, a pure parasagittal trajectory reduces the risk of crossing representational anatomical planes during a single recording trajectory. Unfortunately, medial trajectories to medial targets such as VL thalamus or STN pass through the ipsilateral lateral ventricle, posing a risk of intra- ventricular hemorrhage. Therefore, lateral angles that are 5° to 10° lateral of 90° are typically used.
The arc is used to mark the desired entry point on the frontal scalp, anterior to the coronal suture and 2 to 3 cm lateral of the midline. Lidocaine (1%) is administered for local anesthesia. After skin incision and hemostasis, a self-retaining retractor is inserted. A skull perforator (Codman, Inc., Rayn-ham, Massachusetts) is used to make a 14-mm burr hole centered on the entry point. The dura is coagulated and incised in a cruciate fashion.
The author uses a combination of MER and macroelectrode stimulation to physiologically refine the anatomically selected target. The details of micro- and semimicroelectrode recording, as well as macroelectrode stimulation for target localization in the thalamus [4-6, 10], globus pallidus [8,11,12], and subthalamic nucleus [13] have been published elsewhere and are beyond the scope of this report. Once proper targeting is confirmed physiologically, the frame is set up for lesioning or DBS lead insertion.
When performing a neuroablation, the lesioning electrode (Radionics, Inc., Burlington, Massachusetts) is inserted to the physiologically defined target and the lesion is made. When ablating GPi, the author performs 4 to 5 lesions along a single trajectory as defined by MER [8]. Thalamotomy lesions are placed at that site where high frequency stimulation (i.e., >130 Hz) arrests tremor without persistent dysesthesia.
Figure 3 DBS Lead Placement. The deep brain stimulation lead is inserted under C-arm fluoroscopic guidance. The circles and cross-hairs are aligned so that a pure lateral image, centered on the target point, is generated.
Deep brain stimulating leads are inserted under C-arm fluoroscopic guidance (Fig. 3). The manufacturer sells "bomb sites” (i.e., circles and cross hairs) that snap into the rings of the frame and allow the surgeon to generate pure lateral fluoroscopic images that are centered on the frame’s target point. When proper placement is confirmed fluoroscopically, the DBS lead is secured at the level of the skull with a burr hole "cap" that is provided with the lead (Model 3387 or 3389, Medtronics, Inc., Minneapolis, Minnesota). The excess lead is encircled around the cap in the subgaleal space. The incision is irrigated with bacitracin saline and closed in a standard fashion.
A postoperative MRI is obtained before proceeding with implantation of the pulse generator. The MRI confirms proper lead placement and demonstrates any hemorrhage related to surgery. If a hemorrhage is found, pulse generator implantation, which is performed under general anesthesia, is delayed. All patients are observed in the neurosurgical observation unit during the evening after lead insertion or neuroablation. Most patients are discharged within 1 or 2 days of surgery.
