When the frame is placed appropriately, it will be angled 5-10 degrees from the horizontal.

Once the frame is in place, it is simple to apply securing aluminum "quick fix" pins. We usually infiltrate the pin-hole sites with a combination of lidocaine and bupivacaine in a 50:50 ratio. We use 1% lidocalne with 1:100,000 epinephrine solution,whereas bupivacaine is a 0.25% solution. We inject approximately 2.5-5 ml of this local anesthetic at each pin-hole site using a 25-gauge needle directed through the pin hole. This creates an excellent wheal so that when the patient is awakened following MRI, the amount of pain noted at the pin-hole sites is minimal. Lidocaine provides some reduction in the need for anesthetic from the anesthesiologist during the initial portion of the procedure, whereas bupivacaine provides some pain control that can be maintained throughout a considerable portion of the rest of the day until the quick fix pins are removed. When placing the pins, it is important that they only protrude a small amount externally (usually on the order of a few millimeters) from the vertical stabilization bars. This prevents a collision between the pins and the subsequent MRI localizer box or the MRI apparatus. If the pins placed posteriorly protrude significantly, they can interfere with the table of the MRI scanner, making imaging impossible and requiring replacement of the pins. Once the pins are placed, we prefer to place contralateral pins simultaneously, not completely tightening these until all four pins have been placed. This can be followed by subsequent tightening to finger tightness of each of these pin sites. Overtorking of the pins should be avoided since it may distort the frame, thereby altering the accuracy of the system. As soon as the pins are noted to be secured, the ear bars are removed. One can then check the alignment by again noting the tip of the nose appearing just below the inverted U-shaped portion of the stereotactic frame and also noting proper alignment of the frame in the coronal plane with the external auditory canals centered through the lowest of the three ear bar hole sites.

After placement of the frame and while maintained in a significantly sedated state, the patient is transfeued to the MRI table, which is brought out to the operating center next to the MRI scanner.

Three-Dimensional Volumetric
Matrix Acquisition MRI
At this point, the Leksell MIZI localizer and MRI table adapter are mounted onto the frame. These are provided by Elektra Instruments and can be used for other stereotactic procedures such as Gamma Knife radiosurgery and brain biopsy. The fiducial channels on the MRI localizer are filled with copper sulfate and are regularly checked to ensure that there are no air bubbles present that would obscure fiducial markers, making it impossible to utilize those images in the stereotactic plan. Head movement is completely restricted by fitting the MRI table adapter to the corresponding holes on the MRI table head.

It has been noted by our anesthesiologists that other devices may be required to maintain the patient in a sedated state. In particular, nasal air-ways or laryngeal mask devices may be required to maintain adequate airways. In addition, special padding and positioning of the patient may be needed in older patients with kyphotic deformity of the spine to enable them to lay supine with the MRI localizer and table adapter mounted appropriately within the MRI scanner.

At our center, we use a 1.5-tesla scanner (General Electric, Milwaukee, WI) and utilize multiple special imaging sequences for evaluation. To ensure alignment between the MRI scanner and the frame, it is necessary after sliding the head coil back over the patient's head to align the frame assembly parallel to the imager by rotating the screws on the MRI table adapter such that the beam lights are superimposed on and lie parallel to the corresponding axial lines of the MRI localizer.

We routinely use six different MRI sequences. The utilization of special sequences is critical for this stereotactic planning. Much debate has centered in the field about the use of computed tomography (CT), MRI, intraoperative ventriculography, and other measurements in order to localize the AC-PC line and for localization of the target. To resolve anatomic structures with high accuracy and minimal or no distortion, we use an MRI-based system with confirmation of its lack of distortion on a regular basis using CT. This sequencing capability began initially with utilization of such localizing techniques with the Gamma Knife system for stereotactic radio-surgery but has been extended considerably further in order to provide 3D volumetric images with minimal or no distortion.

It is known that if the data through an MRI are obtained as a 3D volumetric sequence and as a matrix, the accuracy is significantly improved and distortion is minimized compared to standard two-dimensional images. Numerous distortion effects occur with two-dimensional imaging and, although the physics behind this are beyond the scope of this chapter, it is important to realize that this distortion is not of small magnitude and sometimes corresponds to several millimeters. This has prompted some to utilize CT. However, although CT lacks the distortion, anatomic image resolution is inherently poor. With the special 3D volumetric matrix acquisition imaging that we currently use, distortion appears minimal (100-200 microns) and certainly less than that produced by the surgical procedure itself.

Initial sequencing involves the obtaining a sagittal T1-weighted MRI. This routine image set is used as a scout to set up the sequences that follow. These images are of moderate resolution, time of echo (TE)14, time of repetition (TR) 500, receiver band width (BW) 16, field of view (FOV) 26 cm, matrix 256 frequency x 192 phase encodes, number of excitations (NEX) 1, slice thickness 5 mm, skip 2.5 mm, 15 slices, time for acquisition 1 min 52 sec.

The second image sequence is also a routine T1-weighted sequence acquired in the axial plane. These are used to check the frame placement and to measure the fiducial markers in order to confirm known measurements. With this sequence, it can be demonstrated that the frame is symmetric and the sides of the box are parallel. We use a fast spin echo (FSE) TI, TE 11, TR 600, number of echoes 3, BW32, FOV 26 cm, 252 x 192, 2 NEX, slice thickness 3mm, skip 0, 9 slices, time for acquisition 1 min 18sec.

The third data set is the one used for most of the targeting. This data set has images that have very high T1 weighting and is obtained in a 3D or volumetric fashion. In the GE system, this is called Spoiled Gradient Recalled (SPGR) images; the GE parameter FAST is also utilized to save time obtaining the data set. This results in images called FSPGR by GE nomenclature. In order to obtain better gray and white matter discrimination, a preparation pulse is used resulting in a T1 inversion recovery (IR) appearance to the images. Gadolinium contrast enhancement is also utilized to improve gray/white matter discrimination (Prohance (Nycomed) 1 cc per pound). The parameters for this axial pulse sequence are: TE minimum fill (5.3), TR is set by the MRI with the FAST option (12.3), flip angle of 10 degrees, BW16, time of inversion (TI) 500, FOV 26 cm, 256 x 256,2 NEX, slice thickness 1.2 mm, 60 slices, time for acquisition 11 minutes.

It is critical that the patient be immobilized during this time. This sequence is prescribed graphically with the lower slices at the top of the sella turcica (Figure 3). This results in images with excellent signal-to-noise ratio as well as excellent gray/white matter discrimination (Figure 4). Because these images have volume elements that are almost cubical (isometric voxels), reconstruction images appear as sharp as the plane in which they were acquired (Figures 5 and 6). This system enables us to localize well the AC-PC line. The line will then be keyed to identification of the subsequent target in relationship to this line (Figure 7).