The fourth sequence is used for localizing the entry point of the surgical device. This is also a 3D or volumetric sequence. However, since it is only used for delineation of sulci, gyri, and draining veins, a quicker lower resolution image will suffice: FSPGR: TE minimum fill (5.3), TR is set by the MRI with The FAST option (12.1), flip angle of 20 degrees, BW 16,FOV 26 cm, 256 x 256, 2 NEX, slice thickness 3 mm,28 slices, time for acquisition 3 min 19 sec (Figures 8-10).

The last two sequences are essentially the same except for the plane of acquisition. They attempt to simulate a conventional proton density image of a dual echo spin echo T2-weighted image, but with a higher signal-to-noise ratio and with shorter acquisition times. The final weighting of the image is a compromise between tissue contrast and being able to visualize the stereotactic frame fiducial markers (Figures 11 and 12) This axial FSE-IR sequence is as follows: TE 44 effective (EF), TR 4500, echo train (ET)16, BW 32, TI 150, FOV 26 cm, 256 x 256, 4 NEX, slice thickness 3 mm, skip 0, 12 slices, time for acquisition 4 min 48 sec. Additional parameters include flow compensation in slice, extended dynamic range, and saturation bands superior, inferior, right, and left. The right and left hands are placed graphically within the image between the fiducial markers and the area of interest so as to decrease the artifact obscuring the fiducial markers.

The coronal FSE-IR images are obtained; TE 50 EF, TR 2500, ET 16, BW 16, TI 150, FOV 26, 256 x256, 4 NEX, slice thickness 3 mm, skip 0, 14 slices, time for acquisition 5 min 30 sec, flow compensation in slice, extended dynamic range, and saturation bands superior, inferior, right, and left. Again, the right and left saturation bands are placed medial to the lateral fiducial markers (Figures 13-15). For imaging of the globus pallidus and excellent anatomic depiction of the globus pallidus, internal capsule and optic tract are required and these fiber tracts and nuclei are best visualized utilizing the ESE-IR sequences.

Each of the multiple sequences has its own advantage and each advantage is integral to the total planning process. Once the sagittal localizing images are obtained, axial images are required to determine the X (lateral) and the Y (anteroposterior) target coordinates, while the coronal images will be utilized to determine the X and Y (vertical) coordinates. Targets are then obtained using measurements based on the relative distance from the mid commissure; further evaluation will direct the final target visualization.

Some familiarity with the use of the fidudials is required in performing this surgery. The standard globus pallidus target as defined in the literature is 2 mm anterior to the midpoint of the AC-PC line, 20-22 mm lateral to that line, 4-6 mm inferior to the AC-PC plane, and level with the floor of the third ventricle. Prior to this measurement, however, one must determine the MRI coordinates of the center of the frame by joining the diagonally opposing fiducials using built-in MRI software. The image (Figure 16) that shows the AC-PC line is then utilized to determine the AC-PC line, and subsequent early target recognition.

Much of the further work can be performed as it is transferred through the network to the planning software system, which is currently placed within the surgeon's office to facilitate planning for subsequent surgical treatment. The software provided with the GE MRI system and with the SurgiPlan planning software (Elekta) determines the X and Y coordinates automatically (Figures 17 and 18).

When planning software is not available and only the MRI software can be used, the logic of the determination of coordinates is important. The Elekta-Leksell system has simplified this by making the midpoint of the frame the 100/100/100 coordinate of the stereotactic system. Coordinate changes are in millimeters and, therefore, the coordinate numbers themselves are a physical measurement with a distinct relationship to the center of the frame. To determine the vertical coordinate of a target (Z) noted within a plane of section, it is necessary to add a constant (40 mm) to the distance measured between the posterior fiducial and the middle fiducial on both sides. Minimal discrepancy should be noted between these sides and, in fact, if the frame is placed in the method described, the values should be identical. If the distance between the right and left is greater than 2 mm, the MPI frame should be replaced so that it is more truly orthogonal and/or adjustments be made to realign the frame to MRI gantry.

After first obtaining images using the SPGK and ESE-IR sequences, the initial localization of the target based on "standard" coordinates can be checked. We have found that the standard target is not adequate for the final target and adjustments can then be made in the software planning system under the third part of the operative procedure.

It is important to understand that the target will lie in the tail of the pallidum. It is therefore critical that the globus pallidus not be visualized in the adjacent coronal MRI slice that is 3 mm posterior to the target. The use of the FSE-IR T2-weighted sequence is very helpful because of its visualization of the fiber tracts. Its definition of the anatomic nuclei is not as clear, but this has previously been obtained using the SPGR highly T1-weighted images.

Final target localization will be adjusted further as discussed in the computer software planning stage; however, it is important to note that the lateral coordinate of the target must show the target lying within the globus pallidus interna just lateral to the optic tract with its coordinate corresponding to the superior border of the optic tract itself above the choroidal point. The final target should be at least 2 mm away from the internal capsule and also 2 mm away from the optic tract so that neither structure will be injured by the lesioning or DBS procedure.

Sterotactic Planning with Computer Software
We utilize the SurgiPlan software planning system for further surgical planning and analysis of the volumetric data. Other excellent systems are available, including the StealthStation, which is manufactured by Sofamor-Danek-Medtronic Corp. Initially, we used an internally developed software system, but we have found the Surgiplan system to be quite capable in its localization capability (Figures 19-24).

Critical to the surgical planning is the cooperative involvement of the neurosurgeon and the neuroradiologist so that precise localization of the target can be obtained; in addition, safety and effectiveness must be achieved by carefully planning the trajectory and entry point. At our center, the same neuroradiologist performs all planning with the same neurosurgeon so that constant feedback can be maintained based on past experiences.

Once the target has been effectively localized and noted to be in a safe location, effort must be placed on a safe entry and trajectory to the target. Although this portion of the process is done by few surgeons, it provides increased safety for the patient. As part of the MRI sequencing, we obtain surface images of the cerebral cortex so that an entry site can be chosen with a high degree of safety. Typically, we choose a trajectory that passes through the frontal lobe within the center of the superior or middle frontal gyrus and lateral to the ventricular system. The ring angle is typically approximately 65° ± 5°, which will also be the trajectory relationship to the AC-PC line since the frame was initially placed parallel to the AC-PC line. The arc angle is then adjusted to 80° ± 5° on the right side and l00° ± 5° on the left. Using this procedure, it is frequently found that such an entry site followed by the trajectory to the target provides a path that avoids any arteries or veins visualized on the surface image views and also provides the trajectory that avoids injury to eloquent neural structures. The surface MRI view is useful, for without such planning, an initial twist drill or burr hole placement could incidentally be located over a cerebral artery or vein and electrode passage could result in a catastrophic hemorrhage. This complication has not been seen in any of our patients, possibly because of this added step in the preoperative planning (Figures 25 and 26).

With the SurgiPlan software system, trajectory slices are possible so that every stage of the trajectory can be visualized in terms of its potential harm as an electrode is passed toward the target. Fine adjustments to the entry point can be made to avoid these critical structures or avoid passage through the ventricular system in the patient with large ventricles.

The planning software subsequently prints out X, Y, and Z coordinates, a ring angle and an arc angle, and the coordinates for the entry point. The entry point coordinates are not directly utilized during operative planning but are used by the computer system in creating the trajectory itself An estimate of accuracy can then be obtained and is usually accurate within several hundred microns and always less than 0.5 cm accuracy so that the results from imaging and planning can be used effectively during the surgical procedure.

Lesioning (thalamotomy) and placement of DBS electrodes within the Vim nucleus of the thalamus both utilize the same planning as mentioned above, varying only in terms of target definition. The desired "standard" target for planning in this regard in relation to the AC-PC line is 4 mm posterior to the midpoint of the AC-PC line with lateral extension of 12 to 14 mm and 1 mm below the AC-PC plane. Visual confirmation of the target is likewise as impoftant in the Vim nucleus of the thalamus as it was in the globus pallidus. In many instances, surgeons have found the Vim nucleus of the thalamus target to lie at least 10 mm lateral to the lateral wall of the third ventricle. However, some adjustments still need to be made because of the varying sizes of the thalamic nucleus, particularly in patients with significant atrophy or a large ventricle. Target localization, therefore, must be adjusted frequently in the mediallateral direction; placing the lesion midway between the lateral ventricle wall and the internal capsule provides a safe corridor of placement.

As in the description for pallidotomy, the same sequences were utilized for identification in this case of the internal capsule and its proximity and, in some instances, some visualization of nuclear detail within the thalamus itself. Lesioning or placement of a DBS electrode may then proceed as described below. In the future, similar stimulation procedures may be used in the subthalamic nucleus for treatment of PD and other disorders; however, this awaits approval and is mainly being done at a few centers under special protocols.

Surgical Procedure
After completion of the preoperative planning (the majority of the time for the surgical endeavor itself), the patient can be brought to the operating room for lesioning or DBS placement. As in the discussion above, we describe the pallidotomy and then include the thalamotomy and DBS electrode placement procedures. The procedure for DBS electrode placement is described in full by the product manufacturer and is not described here.

The patient is initially brought to the operating room and placed supine; the frame is fixated to the Mayfield head clamp and the Leksell frame using a special Mayfield adapter. Intravenous antibiotics are infused prior to skin incision and usually include Nafcillin (25 mg/ kg) and gentamicin (1 mg/ kg). Steroids are almost never required during this procedure. Blood pressure and oxygen saturation are continuously monitored, usually on the arm ipsilateral to the site of lesioning or DBS so that the contralateral extremity can be utilized for testing during surgery. To evaluate the tremor and its eradication at the time of surgery, we have found it useful to utilize either writing tests with paper or a cup with liquid that the patient can readily see. Grounding pads are placed beneath the patient's buttock while lesioning or electric stimulation is performed.