Stereotactic Radiosurgery of Skull Base Tumors

Published on 13/06/2015 by admin

Filed under Surgery

Last modified 22/04/2025

Print this page

This article have been viewed 2200 times

Chapter 65 Stereotactic Radiosurgery of Skull Base Tumors

P. Ashley Wackym, Christina L. Runge-Samuelson, David R. Friedland

Management of many skull base tumors has shifted in recent years away from surgical resection and towards control of growth. This is particularly true for vestibular schwannomas, i.e., acoustic neuromas, and is increasingly applicable to glomus jugulare tumors. The principal modality for such treatment is gamma knife surgery although other conformal radiation treatment systems are available. Gamma knife surgery is advantageous in requiring a single session for treatment of most skull base lesions, which increases its appeal to both surgeon and patient. This chapter will focus primarily on the methods used in treating skull base tumors with gamma knife surgery.

Gamma knife surgery, similar to microsurgery, has advantages and disadvantages which must be thoroughly discussed with the patient.^1,² For the patient it is alluring to undergo an outpatient procedure rather than microsurgical management that requires a much longer period of care. Further, gamma knife outcomes show excellent tumor control and, with current methods, low cranial nerve morbidity. Gamma knife surgery is a viable treatment modality for the appropriate patient as defined by age, medical history, tumor characteristics and physical findings. As such, many neurotologists now offer gamma knife surgery as part of their armamentarium for managing vestibular schwannomas and glomus tumors.³

Several institutions world-wide offer training courses for physicians and radiation physicists at centers having the Leksell Stereotactic System or Leksell Gamma Knife. To date more than 1500 neurotologists, neurosurgeons, physicists, and radiation oncologists have received such training. In addition, the parent company, Elekta Instrument AB (Stockholm, Sweden), offers basic and advanced training courses and workshops. Courses typically consist of didactic lectures, observation of patient treatment, and practical hands-on training. Further, all new installations of Leksell Gamma Knife are accompanied by a one-week on-site start-up training for the neurotologists, neurosurgeons, radiation oncologists, and physicists comprising the gamma knife treatment team.

PATIENT SELECTION

Opting for gamma knife surgery over observation or microsurgical resection is a complex decision. There are the preferences of the informed patient, the comfort and experience of the surgeon, the patient’s medical history and condition, and the characteristics of the tumor. While there are no definitive measures defining or restricting the use of gamma knife surgery, particular guidelines can inform the decision making process.

Although a tissue diagnosis is not typically acquired prior to gamma knife treatment, radiographic and clinical diagnoses of vestibular schwannoma and glomus jugulare are sufficient to initiate a discussion of gamma knife surgery. Other potential neoplasms amenable to gamma knife treatment by the neurotologist are cerebellopontine angle meningiomas, posterior fossa and jugular foramen non-vestibular schwannomas, temporal bone metastatic lesions and primary vascular neoplasms. An absolute contraindication to gamma knife treatment would be tumors extending too far inferiorly to enable placement into the centrum of the collimator helmet. Gamma knife surgery is also contraindicated in large tumors causing life-threatening brainstem and central aqueduct compression. Such large tumors, in the absence of clinically significant problems, provide a relative contraindication to gamma knife surgery as post-treatment swelling may cause obstructive hydrocephalus requiring emergent intervention. Typically, vestibular schwannomas greater than 2.5 cm in the cerebellopontine angle should be cautiously approached if gamma knife proves the best option given other medical concerns. Most surgeons will not treat vestibular schwannomas greater than 3.0 cm in maximum axial dimension within the cerebellopontine angle because of the risk of post-treatment obstructive hydrocephalus.

Other guidelines for gamma knife surgery require clinical judgment as to the medical condition of the patient, the expected growth and potential morbidity of the tumor, the functional status of the patient, audiometric and vestibular performance, age and expected life-span of the patient. Individualized treatment plans depend on a frank and thorough dialogue between physician and patient as to the options available, risks and benefits of each approach, and expected outcomes based upon evidence-based reviews or an analysis of each institution’s outcomes.

PREOPERATIVE COUNSELING

Informed consent for gamma knife surgery requires the surgeon to discuss alternative options such as observation and microsurgical resection.² The risks and benefits of these alternatives should be frankly described and compared to gamma knife treatment. Many patients have received information from the Internet or from physicians with limited experience with gamma knife and may have erroneous information. Common misconceptions include the expectation that gamma knife surgery completely removes the tumor and that hearing will improve, or conversely that cranial nerve morbidities are significant. These need to be addressed with evidence-based reports and information.

One statistic, which is particularly alarming to patients considering gamma knife surgery is that there have been eight cases of malignancy within vestibular schwannomas (as of 2002).⁴ Four of these patients had been previously treated with radiosurgery. While it remains possible that these four malignancies developed after the radiation treatment, it is more likely that these malignant tumors were misdiagnosed as benign at the outset of evaluation and treatment.

Delayed development of radiation-induced neoplasms was addressed by Pollock and colleagues in 1998.⁵ They reviewed more than 20,000 patients treated with radiosurgery worldwide and found no increased incidence of new neoplasm development (i.e., benign or malignant). A retrospective cohort study comparing the Sheffield, England radiosurgery patient database with the national mortality and cancer registries identified a single new astrocytoma among those treated.⁶ Based on their national incidence figures, 2.47 cases would have been predicted. The risk of radiosurgery induced malignancy in patients with neurofibromatosis type 2 (NF2) and von Hippel-Lindau disease was similarly studied.⁷ Of 118 NF2 and 19 von Hippel-Lindau disease patients, totaling 906 and 62 patient-years of follow-up data, respectively, only two cases of intracranial malignancy were found. Both of these were in NF2 patients. One was thought to have arisen before the radiosurgery; the other was a glioblastoma diagnosed three years after radiosurgery. Gliomas may occur in as many as 4% of NF2 patients and the single case may not represent an increased risk. It was suggested that the late risk of malignancy arising after irradiation must be put in the context of the condition being treated, the treatment options available to these individuals, and their life expectancy.

Despite the findings of the studies just reviewed, it is important to counsel patients about the possibility of malignant transformation or induction. A handful of tumors suggestive of radiation induced malignancy have been reported among the tens of thousands who have undergone gamma knife treatment. Lustig and colleagues reported the development of a squamous cell carcinoma following radiation treatment of vestibular schwannoma.⁸ Hanabusa and colleagues reported the malignant transformation of a vestibular schwannoma following gamma knife surgery.⁹ There was histologic evidence of vestibular schwannoma following a retrosigmoid resection. Four years after this resection, recidivistic tumor was identified, and the patient was subsequently treated with gamma knife surgery. Six months post-treatment, the tumor had grown, and the patient underwent surgical resection via a combined retrosigmoid-translabyrinthine approach. Abnormal mitotic figures were observed on histologic sections, and the diagnosis of malignancy was assigned.

SURGICAL TECHNIQUE

The Gamma Knife Unit

The first gamma knife unit (Elekta Instrument AB, Stockholm, Sweden) was installed in Stockholm, Sweden in 1968, and it was not until 1987 that the first gamma knife (model U) was installed in the United States at the University of Pittsburgh. The gamma knife model B (1996) is the unit currently most used throughout the United States. The gamma knife model C was introduced three years later and the major upgrade consisted of an automatic positioning system (APS). The unit is otherwise quite similar to the model B and both contain 201 radioactive isotope cobalt 60 (⁶⁰Co) sources and beam channels. Due to physical restraints these units can only treat lesions intracranially or along the skull base. During 2008, a completely redesigned gamma knife unit, named Perfexion, is being introduced. It uses 192 ⁶⁰Co sources, has a single collimator helmet with variable diameters, and can treat lesions within the entire head and neck, down to the level of the clavicles.

The basic principle of gamma knife surgery is to provide focused radiation to the tumor while minimizing radiation delivery to surrounding tissues. As such, a semicircular shield called the collimator helmet is used to generate approximately 200 individual gamma radiation “beams.” In the center of the helmet, where the beams meet, radiation delivery is maximal, but along each individual radiation tract tissue exposure is relatively low. When the collimator helmet is locked into position, the 201 openings of the collimator helmet coincide with the cobalt sources. There is a shielded chamber within which the ⁶⁰Co sources are contained, and stainless steel shielding doors protect the treatment room from the ⁶⁰Co sources. There is a treatment couch with an adjustable mattress that slides into the gamma knife unit together with the collimator helmet and the patient. Figure 65-1 schematically shows the orientation of the components of the gamma knife model, Leksell Gamma Knife® 4C and Figure 65-2 shows the overall appearance of the gamma knife model, Leksell Gamma Knife® 4C.

Figure 65-1 Gamma knife surgery. Schematic illustration of the Leksell Gamma Knife 4C which utilizes the automatic positioning system.

Figure 65-2 Leksell Gamma Knife 4C.

When treatment is initiated, the treatment couch is automatically moved from its idle position into the treatment unit together with patient and helmet. Once the couch is docked in its treatment position, the helmet collimator and corresponding collimators in the unit form a beam channel, allowing the radiation that is continuously emitted by the sources to reach the patient. At the end of each irradiation “shot,” the couch is automatically withdrawn, either to its idle position or to a position outside the radiation focus to reposition the patient for the next irradiation shot. There are four interchangeable helmets by means of which the size of the collimator (that part of the treatment unit that shapes the beam) can be changed between 4 mm, 8 mm, 14 mm and 18 mm. The combination of four different sized collimators and repositioning the patient in the three-dimensional space defined by the stereotactic headframe are effective to deliver the radiation dose selectively and conformally to radiosurgical targets of any shape.

Frame Attachment

The stereotactic head frame is used to coordinate the location of the tumor within the collimator helmet. As such, proper placement is of utmost importance to providing adequate treatment. There are two general principles guiding head frame placement for gamma knife surgery. First, the target should be as close to the center of the frame as possible. This prevents possible collisions of the frame with the sides of the collimator helmet especially when trying to align laterally extended tumors in the center of the unit. Second, the frame attachment should be stable. This prevents movement and ensures accuracy and correlation among the pre-treatment imaging study, workstation treatment plan, and delivery of focused radiation. These principles should be addressed at the time of frame attachment. In lateral targets, such as vestibular schwannomas or glomus tumors, the frame should be shifted toward the tumor side. In skull base tumors the frame should also be positioned lower than for treatment of more superior intracranial lesions. Anterior-posterior alignment should also be accounted for and can be adjusted by varying the lengths of the pins used to secure the frame. To ensure stability, avoid screw fixation in bone flaps, cranioplasty materials, burr holes, or skull defects.

The method of anesthesia used during frame placement is surgeon and patient dependent. In our program, either sedation with versed and fentanyl, or monitored anesthesia with propofol, followed by injection of local anesthetic at the pin sites is used. Figure 65-3A shows the typical array of tools used for the frame attachment. A variety of screw lengths allow the surgeon to choose those ideally suited for the individual location of the posts and tumor. The placement of the frame should begin with an accurate orientation of the location of the target within the patient’s head. Ideally, the target should be located within the fiducial range and placed centrally within the frame thereby avoiding later collisions with the collimator helmet and granting sufficient accuracy for the stereotactic target definition.

Figure 65-3 A, Stereotactic headframe at the time of assembly. Pins used for fixation prior to imaging, treatment planning, and gamma knife radiosurgery are seen in the foreground. B, With a towel placed on the vertex of the head, the magnetic resonance imaging (MRI) fiducial box is balanced on the head while topical anesthetic is infiltrated at the pin sites. C, While the headframe and attached plastic MRI fiducial box is secured to the skull using the pins, an assistant stabilizes the assembly in place. Note the tightening of pins in opposing vectors. D, After the frame has been secured to the skull, measurements are taken through defined entrance points in a plastic dome representing the size and position of the collimator helmet relative to the frame, head and face. These values are incorporated into the treatment planning software to create a wire grid representing the patient’s head. The post height and pin length are also measured and are inputted into the treatment planning software. These values are used in the calculations used to predict collisions between the collimator helmet and the stereotactic headframe.

The stereotactic frame is assembled and preliminarily supported by using external auditory canal support pins, a Velcro band, or a stereotactic fiducial box. When using a fiducial box to facilitate frame placement, it is important to use the MRI fiducial box, rather than the CT or angiography fiducial box, since this is the smallest of the three plexiglass fiducial boxes (Figs. 65-3B and 65-3C). Asymmetric frame placements are possible and do not impair the accuracy of imaging. The frame can be shifted from side to side or can be moved as far as possible to the front or back to facilitate centering of the tumor. The frame is stabilized against the patient by an assistant and the surgeon should adjust the lengths of the posts to maintain relative tumor position. A low position of the anterior posts can help avoid anterior collisions with the collimator helmet for skull base posterior fossa tumors. In critical positions, collisions can sometimes be avoided by using the curved posts in the anterior position.

Once post position is determined the screws can be inserted. The surgeon and assistant should work on diagonally opposing screws to provide the best stability without changing the desired frame position. For asymmetric frame placement apply the longest screws first, thereby defining the desired distance of the target to the frame. Protrusion of the screws from the posts should be kept to a minimum to avoid collisions. Approximately 8 to 10 mm is considered to be sufficient but at our institution we prefer to limit this projection to 4 to 6 mm. If a screw extends further it should be exchanged for a shorter screw.

Measurements of the frame and placement are then performed to allow the computer to identify any potential collisions after the plan is formulated. These measurements are required for the frame and skull section in Leksell GammaPlan treatment planning software. Measurements include the length of the four posts and the length of the screws that protrude from the posts. Additionally, the volume of the head is measured using the plastic collimator bubble, simulating the relationship of the frame to the treatment collimator helmet (Fig. 65-3D). This concludes frame placement and the patient may proceed to imaging.

Imaging

Treatment planning requires imaging of the tumor with respect to the frame as determined by specific fiducial boxes. The MRI fiducial box clips to the frame and care should be taken to ensure that it is flush and square during imaging. The MRI fiducial box has a Z-shaped channel on each side filled with copper sulfate to generate position markers for each axial slice. The box should be checked prior to each use to ensure the channels are filled with solution and no air bubbles are present. The patient, with head frame and fiducial box, is secured into the head holder on the MRI sliding table. For imaging acoustic neuromas and glomus tumors we typically order axial 3D SPGR (spoiled gradient recalled) acquisition with T1 weighting and double dose IV contrast. Before the patient leaves the scanner images are reviewed and the distance between fiducial registration markers is validated for accuracy.

Many centers acquire only MRI scans for treatment planning. We prefer to also acquire a non-contrast CT scan through the temporal bone to aid in planning. There is evidence of distortion of MR images and correlation with CT scans at the time of planning can aid in reducing radiation delivery to critical structures such as the cochlea and facial nerve.¹⁰ A CT fiducial box is affixed to the frame, the patient secured in the holder attached to the table, and an axial scan through the temporal bone and skull base acquired. Both CT and MR images are imported into the Gamma Knife workstation. Axial scans are defined, and coronal and sagittal reconstructions generated for each.

Treatment Planning

Leksell GammaPlan is the dedicated software treatment planning system for Leksell Gamma Knife. Dose planning for gamma knife surgery means precisely conforming the isodose distribution to the target. The isodose distribution is built up by a number of individual shots or isocenters. The Leksell GammaPlan software is designed to help the operator as much as possible to perform this procedure and is quite straightforward to use.

Currently, for vestibular schwannomas, the routine prescription is 12 to 14 Gy delivered to the 50% isodose line. The 50% isodose line shows where 50% of the prescribed dose lies. In the case of gamma knife treatments the dose is frequently prescribed to the 50% isodose line. This ensures that the periphery of the tumor will receive at least the prescribed dose, that the dose will be higher than the prescribed dose inside the tumor, and that the dose will fall off rapidly outside the tumor thus sparing critical structures.

Dose planning using Leksell GammaPlan involves composing shots to develop a conformal isodose. By definition, this includes the whole target but spares the surrounding healthy tissue. Figure 65-4 shows an example of a vestibular schwannoma. The target is well positioned on the screen and magnified for good visibility. When the shot menu is opened, one can select the size of the collimators. The size of the collimator is selected based on the tumor shape and the gaps in coverage of the 50% isodose line displayed over the tumor. Shots are placed sequentially to cover the target as effectively as possible. Changing the position of the shots, adding additional shots, and adjusting the relative weight of shots quickly leads to a conformal dose plan.

Figure 65-4 Initial treatment planning at the gamma knife workstation involves building a three-dimensional model of the tumor. Determination of the conformation of the treatment plan follows placement of the shots and assignment of the radiation dose delivered to the specified isodose line.

The dose plan can be checked using Leksell GammaPlan with the three-dimensional (3D) image or the measurement tools, such as dose volume histograms. While the subject of conformity index is beyond the scope of this chapter, an excellent review of available methods has been published.¹¹ Leksell GammaPlan indicates the point in the stereotactic space where a global maximal dose can be found. Leksell GammaPlan also calculates the individual shot times. Once the treatment plan has been determined to be appropriate by the gamma knife team (surgeon, radiation oncologist, and radiation physicist), the stereotactic coordinates and irradiation times are printed and used during the gamma knife treatment.

An automated approach to initial treatment planning has been developed by Elekta Instruments AB. This software assists the treatment planner in generating a good initial dose plan quickly and is termed “the Wizard.” This is an interactive tool that helps the operator develop the dose plan. The operator first selects the shot size and the degree of density with which the Wizard should fill the target. A mouse click instructs the Wizard to fill the target with shots. If the initial dose distribution is not sufficient, a mouse click on the run button instructs the Wizard to optimize the plan by moving and weighting the shots. This interaction results in a better dose plan, and after a few more changes a satisfactory dose plan can be created. However, in this our experience, manual placement of the shots, particularly for vestibular schwannomas, has always resulted in a better treatment plan.

Fine-tuning is made with small adjustments in shot position and weight, allowing optimization of the dose plan. Leksell GammaPlan allows the creation of different plans for the same target. This allows the surgeon and oncologist to follow different strategies and later compare plans and select the best plan for the actual treatment. Treatment plans can utilize as few as one or two shots, such as when treating trigeminal neuralgia, or over 10 shots when treating a large vestibular schwannoma within the cerebellopontine angle and filling the internal auditory canal. With the enhanced capabilities of Leksell Gamma Knife C, plans with 20 shots or more can easily be implemented in a timely manner, since the model C does not require manual adjustments of coordinates in between each shot by the gamma knife treatment team. This allows improved conformity and selectivity of gamma knife surgery, potentially reducing the risk of complications.

To shape the dose distribution to avoid critical structures, one or more of the 201 collimators can be replaced with a closed shield called a plug. One can select spherical areas called shields with different diameters and place them over risk centers in the brain, cranial nerves, or cochlea. Once the shields are put in place, the Leksell GammaPlan software closes off all beams that would irradiate through the shielded area. The result is a modified dose plan in the low isodose lines with only little effect on the target peripheral isodose. The beam channels that need to be plugged can be seen in the plug pattern. The plug patterns can be merged for all shots of the same size so that the operator only has to plug the helmets for the treatment once.

In the final plan, the peripheral dose is set to a value, which is assessed as optimal for a particular patient. Indication, size, and location of the target are taken into account, as well as clinical experience. The peripheral isodose is usually set to the 50% isodose line. This is exactly half the maximum dose in the target, referred to as the hot spot (Fig. 65-5). Along the 50% isodose line the dose gradient is usually the steepest ensuring sufficient dose within the target, while the dose level outside falls steeply, sparing the surrounding healthy tissue. Leksell Gamma Plan can also display the absolute dose values if desired. It will show the point in the stereotactic space where the global maximum dose can be found. With vestibular schwannoma it is valuable to complete this exercise, as the maximal dose at the “hot spot” should be positioned well away from the facial nerve and cochlea. In addition, plotting the absolute dose lines will help in determining the actual level of radiation delivered to surrounding structures.

Figure 65-5 Gamma knife surgery. Selecting the Absolute Dose Level and Display Isodose options allows verification that the maximal radiation dose is not delivered near critical structures, such as the facial nerve. In this example, 14 Gy delivered to the 50% isodose line was prescribed. As shown, the maximum dose (28 Gy) is delivered to the center of the tumor (smallest circle). The largest circle represents the 20% isodose line where 6 Gy of radiation is delivered.

When the dose planning is completed, Leksell GammaPlan checks the shots for collisions with the collimator helmet and sorts the plan according to collimator size. The team performs quality assurance steps to check the accuracy of the X, Y, and Z coordinates and to ensure the plan treats the correct side. All relevant data are documented including details of the treatment plan, targets, dose volume histograms, snap shots and images. When the treatment set-up has been finalized the treatment protocol is exported to Leksell Gamma Knife. This is via a special secured direct serial connection. Leksell GammaPlan only accepts valid and verified treatment plans for export. In addition, a protective design limits the transfer of a treatment plan to the Leksell Gamma Knife to one patient at a time. Once the data have been transferred to the operator’s console, it is verified, and the patient can be treated.

For the model C unit, the operator does not have to enter the treatment room during a run. However, with the model B the treatment team enters the treatment room after each shot is delivered and manually adjusts the X, Y, and Z coordinates, as well as the gamma angle, i.e., the pitch of the head, if necessary. With both the model B and model C, the team has to change the collimator helmet manually when necessary, as dictated by the treatment plan. Detailed treatment and physics protocols are viewed and printed out.

Treatment

Treatment can be performed automatically using the automatic positioning system or manually using trunnions. For the model B, manual setting of the X, Y, and Z coordinates as well as the gamma angle if necessary is accomplished by the treatment team. The Y and Z coordinates are set with the Y, Z slides on the y-bar attached to the coordinate frame, whereas the X coordinate and the gamma angle are set with the trunnions. It is imperative to have a check and balance in place that consists of visual verification of each coordinate by a different team member. Y coordinates need to be verified prior to setting the Z coordinate as the latter will obscure the scale on the Y axis. It is preferable to set the X coordinate of the trunnion on the shorter side first as this will provide more room to manipulate the patient and head frame within the collimator helmet. These coordinates need to be manually changed between each shot on the model B unit.

With the automatic positioning system, the treatment is controlled from the operator console. Once the treatment starts, the selected run is carried out automatically. Before repositioning, the couch will move out a short distance to bring the patient out of treatment focus. At this point, the APS will move the patient’s head to the next target position. A run consists of all shots for a specific collimator helmet size. Additional runs are performed after manually changing the collimator helmet.

After all runs have been completed the head frame is removed. The anterior fixation sites are dressed with antibiotic ointment and adhesive bandages. The posterior sites are dressed with antibiotic ointment. Often pressure needs to be held to control bleeding and occasionally a staple may need to be used on the posterior sites. Typically patients will experience a transient headache after removal of the frame and some develop nausea and emesis. We typically pre-medicate with decadron and ondansetron prior to frame removal. Patients are observed for several hours post-treatment and discharged home with pain medication and follow-up appointments.

GAMMA KNIFE SURGERY OUTCOMES

Just as is the case with other forms of medical and surgical therapy, the techniques and outcomes of gamma knife surgery for vestibular schwannomas have evolved and improved over time. Tumor control and facial nerve motor preservation occurs with virtually all vestibular schwannoma patients treated with current gamma knife protocols. Areas of continued focused investigation include the effects of radiosurgery on hearing and balance, and methods of improving outcomes.

The University of Pittsburgh group has the largest clinical experience in treating vestibular schwannomas with gamma knife surgery. Lunsford and colleagues summarized their experience with 829 vestibular schwannomas treated between 1987 and 2002.¹² This extensive clinical experience included an average tumor volume of 2.5 cm³ and a median margin dose to tumor of 13 Gy. They reported tumor control in 97% of patients at 10 years, and facial nerve (motor) dysfunction in < 1% of patients. Trigeminal nerve symptoms occurred in < 3% of patients and typically occurred with large tumors reaching the level of the trigeminal nerve. No reporting of balance function was included in their analyses.

The reporting of hearing preservation has limited representation in the entire 829 patients. Hearing outcomes data were presented in only 267 patients and “5-year actuarial rates of hearing level preservation and speech preservation” were reported in 103 patients. They reported “unchanged hearing preservation” in 50% to 77% of these patients, and this method of reporting auditory performance points to the difficulty in interpreting the outcome of most of the studies reporting hearing outcome in patients with vestibular schwannoma who have been treated with gamma knife surgery. They also stated that “for patients with intracanalicular tumors, hearing preservation rates in those treated with 12.5 to 14 Gy at the margin showed 90% preservation of serviceable hearing.”¹³ Unfortunately, pretreatment and longitudinal data are not available in these reports.

In an earlier series of 190 patients the average dose to the tumor margin was reduced to 13 Gy, and excellent tumor control was achieved at 97.1%.¹⁴ In this study, issues highlighted earlier with reporting of hearing outcome are equally apparent. They reported “hearing-level preservation” in 71 ± 4.7% of patients. They also reported a “preservation of testable speech discrimination ability” in 91 ± 2.6% of subjects. Obviously, testable speech discrimination ability is far different than useful hearing, and it is unfortunate that these authors did not report the actual auditory thresholds or speech discrimination ability. Most importantly, these were not reported as a function of time post-gamma knife surgery. In addition, they reported that “hearing levels improved” in 10 (7%) of 141 patients who exhibited decreased hearing defined as Gardner-Robertson grades II to V before undergoing gamma knife surgery. Based on our clinical observations and those of other centers, this picture is far more complex over time than is represented in these publications.

Prasad and colleagues from the University of Virginia reported their series of 200 vestibular schwannomas treated with gamma knife surgery over a 10-year interval in 2000.¹⁵ Of these patients, 153 patients had follow-up data including 96 with primary treatment and 57 with secondary treatment. They reported no hearing pre-gamma knife in 105 patients, including 53 of 96 primary treatment and 52 of 57 secondary treatment patients. The Gardner-Robertson grading system and subjective assessment of hearing was used; however, no pure-tone average or speech discrimination data were reported. Unfortunately, their data set included audiometric data from only 48 patients, and the intervals of audiometric testing were not reported. Despite these limitations, they found that, except for one patient, no change in hearing was observed in the first two years after gamma knife surgery. Their data also showed that the greatest change in Gardner-Robertson grade occurred between years two and four post-gamma knife; however, without understanding the assessment intervals, the precise onset of the hearing loss is unknown. No outcomes regarding balance function were reported.

Kim’s group at the Seoul National University reported the hearing outcomes in 25 patients with vestibular schwannomas with serviceable hearing.¹⁶ The median tumor volume was 3.0 cm³ (0.16 to 9.1 cm³), and the dose used was 12 ± 0.7 Gy at the 49.8 ± 1.1% isodose line. They reported the hearing outcomes using the Gardner-Robertson grading system, pure-tone averages, and speech discrimination scores. Pre-gamma knife, interim post-gamma knife, and last post-gamma knife data were reported. Similar to our experience, they found that in 16 patients the hearing deteriorated > 20 dB three to six months post-gamma knife and that this hearing loss continued for 24 months. The only prognostic factor for hearing deterioration that they identified was the maximum dose to the cochlear nucleus.

In the Medical College of Wisconsin Acoustic Neuroma and Skull Base Surgery Program, we have established a clinical pathway for all of our patients undergoing gamma knife surgery for primary or secondary treatment of their tumors. Pretreatment they undergo a complete videonystagmography test battery, a complete audiologic assessment, and facial nerve electromyography. At six-months intervals post-treatment, each patient undergoes a gadolinium enhanced MRI as well as an audiologic test battery and caloric testing to assess peripheral vestibular function. In addition to other standard reporting methods, we have also presented the data in a longitudinal manner for their objective auditory thresholds (Fig. 65-6), speech discrimination ability (Fig. 65-7), and degree of vestibular paresis (Fig. 65-8). We have recently published an expanded cohort of 54 patients with a median follow-up interval of 54.7 months.¹⁷ This report focused on the longitudinal outcomes in vestibular function and changes in the Dizziness Handicap Inventory before and after gamma knife surgery.

Figure 65-6 Auditory function over time after gamma knife surgery treatment of unilateral vestibular schwannomas. Three-frequency averages of pure-tone thresholds (PTA-3) in dB HL at 0.5, 1, and 2 kHz were determined for all patients with measures at the preoperative time and at least one postoperative interval. The PTA-3 difference was calculated for each time interval relative to the preoperative PTA-3. The differences are plotted as a function of postoperative time interval, with zero representing the preoperative time. A positive difference value indicates a higher or poorer, postoperative PTA-3. In general, over time, the vast majority of patients were found to have PTA-3s that were poorer or similar to preoperative PTA-3s, although a few individuals showed some initial improvement (e.g., subject 5). The greatest changes in PTA-3 were measured at six months post-treatment although continued changes were observed up to five years post-treatment.

Figure 65-7 Speech recognition testing was performed using the Northwestern University Auditory Test No. 6 (NU-6) monosyllabic words. The stimuli were presented at 40 dB sensation level, i.e., above speech recognition threshold, or if this was too loud, at the patient’s most comfortable listening level. Speech recognition was scored in percent correct. As with PTA, the differences between pre- and postoperative speech recognition were calculated and plotted as a function of postoperative time interval. Positive values are consistent with an improvement in speech recognition. Approximately half of the patients showed improvement in speech recognition at six months post-treatment, while the other half showed a decrease in performance. Of those patients who experienced a reduction in speech discrimination ability, there was a greater range of change than that observed in the patients who enjoyed an improvement in speech discrimination ability. It should be noted that the greatest changes in speech discrimination ability occurred at six months post-gamma knife treatment. However, over time, the patients generally demonstrate speech recognition performance similar to or poorer than pre-treatment performance.

Figure 65-8 Vestibular paresis was determined with bithermal caloric testing. A positive difference value indicates greater vestibular paresis post-gamma knife surgery. Both degradation and improvement in vestibular paresis are observed across patients. Within a patient, the postoperative degree of vestibular paresis generally tends to remain stable over time after the relatively large initial change observed at the six month post-treatment assessment. In those patients who had continued reduction in their vestibular function, there was continued difficulty with dysequilibrium until vestibular compensation was complete and the vestibular paresis stabilized.

It is clear that most of the change in hearing and balance function occurs during the first six months after gamma knife surgery; however, continued but less rapid worsening of function can occur up to 12 months. These objective measurements correspond well to the transient facial nerve dysfunction, trigeminal nerve dysfunction, tinnitus, and dysequilibrium occurring in our patients with vestibular schwannomas undergoing gamma knife surgery.^3,^10,¹⁷ A possible mechanism underlying these changes is that there is an initial increased size of the tumor after radiosurgery. Typically this post-treatment edema persists for six months; however, this may remain for up to one year.^3,^10,¹⁷ The labyrinthine artery, a branch of the anterior inferior cerebellar artery, provides essentially all of the blood supply to the cochlea and vestibule and it is likely that the post-radiation edema compromises this blood supply to the inner ear. The resulting inner ear devascularization could certainly explain the rapid change in hearing and balance function seen at the six-month post-treatment assessment in our patients (see Figs. 65-6 and 65-8).

Several of our patients have had tumor control or regression and improvement of hearing and vestibular function. This is clearly divergent from the natural history of vestibular schwannomas. In contrast, worsening of auditory and vestibular function and the development of disequilibrium has occurred in a number of our patients. Continued systematic studies of these patients and expansion of the cohort of patients studied are important to determine the efficacy of gamma knife surgery and to compare to other forms of radiotherapy, as well as microsurgery and expectant management. Recognition of symptoms such as disequilibrium and knowledge regarding the expected time course of vestibular paresis progression are important not only for patient counseling but provide the opportunity to intervene with vestibular rehabilitation or nonspecific vestibular suppression until compensation has been completed, should this be needed clinically.¹⁷

One final issue to consider is tumor growth after radiosurgery (Fig. 65-9). It is important to appreciate that there is an increased size of the tumor after radiosurgery. In fact, we observed a statistically significant increase in tumor size for patients whose tumors extended outside of the internal auditory canal six months after gamma knife surgery and a statistically significant decrease at one year post-treatment.^3,¹⁰ Typically, post-treatment edema persists for six months; however, this may remain for up to one year. Consequently pretreatment counseling should include this information. There have been anecdotal cases discussed and occasionally reported that describe increased tumor size early after radiosurgery. The challenge is in making a decision about whether to resect these tumors and when.^2,^5,¹⁸^–²² Pollock and colleagues emphasized the need to demonstrate sustained tumor growth by serial MRI before making the decision to operate and also to review the case with the surgeon who performed the radiosurgery before a surgical decision is made.⁵

Figure 65-9 Gamma knife surgery. Example of serial magnetic resonance imaging studies of a small left vestibular schwannoma. Note at six and 12 months post-gamma knife radiosurgery the tumor is larger than pretreatment. By 18 months the tumor is smaller.

Another related controversy is whether facial nerve dissection and preservation are more difficult during microsurgical resection after radiosurgery. On one end of the spectrum, descriptions of no increased difficulty have been reported;⁵ and, on the other end of the spectrum,¹⁹^–²² markedly increased difficulty in separating the tumor from the facial nerve and poorer facial nerve function outcome have also been reported. The report of Watanabe and colleagues included a histopathologic analysis of the resected facial nerve.²⁰ They found microvasculitis of the facial nerve, axonal degeneration, loss of axons, and proliferation of Schwann cells. In light of the mechanism of delayed effects following radiosurgery, these findings are not surprising. Moreover, these findings emphasize the need for the neurotologist to be certain that the treatment plan avoids high radiation doses to the facial nerve. Recall as described earlier that a dose of 12 Gy delivered to the 50% isodose line means that the maximum tumor dose is 24 Gy. If the treatment plan delivers this maximal dose to the area of the facial nerve, it should be expected that greater radiation effects will be observed. For this reason, if the neurotologist and the patient have made a decision to resect a tumor previously treated with radiosurgery, it is important to review the treatment plan to determine the amount of radiation delivered to the facial nerve to counsel the patient appropriately preoperatively.

ALTERNATIVE TECHNIQUES

As noted earlier, tumor size and location may dictate that a method other than gamma knife be considered. Indeed, alternative methods of radiosurgery are available for treating a wide variety of skull base neoplasms. These include the Peacock (NOMOS Inc., Cranberry Township, PA), the SmartBeam IMRT (Varian Medical Systems Inc., Palo Alto, CA), the Precise (Elekta, Inc., Stockholm, Sweden), and the CyberKnife (Accuray, Sunnyvale, CA). Among the more common of these modalities is CyberKnife, which will be briefly reviewed here.

CyberKnife Stereotactic Radiology

Overview of Treatment Planning

The CyberKnife stereotactic radiosurgery system utilizes a compact 6-MeV linear accelerator, a computer-controlled robotic arm with six degrees of freedom, and an image-guidance technology that does not depend on a rigid stereotactic frame and thereby enables treatment of extracranial sites (Fig. 65-10). Potential benefits of this approach include: 1) increased access to and coverage of any target volume including the ability to treat lesions in and around the cranium that are unreachable with other systems, for example, in the lower posterior fossa and foramen magnum; 2) enhanced ability to avoid critical structures; 3) capability to treat lesions in the neck and spine; 4) ability to treat lesions throughout the body; 5) delivery of highly conformal dose distributions; 6) option of fractionating treatment; and 7) potential to target multiple tumors at different locations during a single treatment, e.g., skull base and neck.

Figure 65-10 Ceiling-mounted diagnostic-energy x-ray sources emit low-dose x-rays through the patient’s tumor treatment area. Amorphous silicon image detectors capture x-ray images from ceiling-mounted diagnostic-energy x-ray sources to produce live radiographs. The operating system (typically located adjacent to the treatment room) correlates patient location detected by image guidance system with reconstructed CT scan and directs the robot to adjust position accordingly. The compact linear accelerator mounted on a computer-controlled robotic arm which adjusts position to maintain alignment with the target, compensating for any patient movement and uses X-band technology for mobility.

Rights were not granted to include this figure in electronic media. Please refer to the printed book.

The CyberKnife treatment planning system is designed to support the radiosurgery team in determining the optimal plan, including beam weight, targeting positions, dose distributions, and other factors for each patient’s treatment. The CyberKnife stereotactic radiosurgery system permits the following planning and delivery options: 1) inverse planning; 2) nonisocentric delivery; and 3) hypofractionation. In contrast to most gamma knife procedures, CyberKnife is CT based. MR images can be fused with the CT to provide optimal information on soft tissue as well as skeletal anatomy. CT angiography can be used when vascular skull base lesions such as arteriovenous malformations or extensive glomus jugulare tumors are to be treated with this technique.

The flexibility of the robotic arm supporting the linear accelerator allows the CyberKnife to implement a wider range of treatment plans than other systems. Furthermore, because the system does not require the use of a stereotactic head frame temporarily attached to the patient’s head, it allows scanning, treatment planning, and quality assurance to take place at any time prior to treatment itself. The CyberKnife system provides a range of treatment options, including the ability to use either forward or inverse treatment planning. With forward treatment planning, the radiation oncologist determines what dose to deliver from a particular targeting position. The total dose within the lesion is then calculated by the system software. With inverse treatment planning, the radiation oncologist specifies total dose to be delivered to the tumor. The surgeon and radiation oncologist are then able to set boundaries to protect adjacent critical structures. The software subsequently determines targeting positions and the dose to be delivered from each targeting position. While other stereotactic radiosurgery systems offer the inverse planning option, the number of possible plans is limited by the constraints of the delivery system.

ALTERNATIVE APPLICATIONS

Although the majority of stereotactic radiosurgery performed by the neurotologist will be for vestibular schwannoma, other neoplasms and pathologies may be amenable to radiotherapy.²⁴ Several of these were noted earlier in the section on patient selection and this section will focus on a few common pathologies for which the neurotologist may be the primary surgeon.

Paragangliomas, more specifically glomus jugulare tumors, are becoming more commonly addressed with primary radiotherapy than with surgical resection. The other chemodectomas such as glomus vagale and carotid body tumors are located too low in the neck for most gamma knife units in use today. Further, these tumors do not typically carry the same morbidity as glomus jugulare tumors and are still commonly addressed surgically. In an effort to reduce the cranial nerve palsies that often accompany glomus jugulare resection, gamma knife surgery has been employed.

An evaluation of 42 patients with primary or recurrent/persistent glomus jugulare tumor undergoing gamma knife surgery showed excellent tumor response.²⁵ Approximately 1/3 of tumors shrank and 2/3 showed no size change. A single 3.9 cm tumor was found to have increased 99 months after treatment with 12 Gy at the margin and was re-treated. Progression-free survival was 100% at seven years and 75% at 10 years. Six patients had complications related to treatment. Five of 26 patients with intact hearing at the time of treatment had subjective decline within the first year. Objective measures of hearing were not performed. One patient had facial parasthesias, one had vocal fold paralysis (the re-treated subject), one had vertigo and imbalance, and one had post-treatment migraine requiring admission.

A meta-analysis of glomus jugulare treatment and outcomes compared stereotactic radiosurgery to surgical resection.²⁶ Neurological deficits in those treated with gamma knife, CyberKnife or LINAC showed no change in 58.2%, improved in 39% and permanently worsened in 2.8%. Such deficits included complaints of hearing loss, dizziness, dysphagia, voice change, shoulder dysfunction and headache. Overall, there was an 8.5% incidence of cranial nerve complication with 75% of these being transient. Permanent deficits occurred in three of 141 patients all of which involved facial motor dysfunction; none of which reached House-Brackmann grade VI. Tumor control was achieved in approximately 98% of individuals at 39 months.

Conventional surgery for glomus jugulare had a complete resection rate of approximately 92%; some of which represented more than one surgery for resection. The recurrence rate at a mean of 82 months was 3.3%. The mortality rate was 1.3% for conventional surgery compared with 0% for radiotherapy. Cranial nerve deficits varied widely among surgical reports but on average the facial nerve was affected in 4.4% to 11%; the glossopharyngeal in 26% to 42%; the vagus in 13% to 28%; the spinal accessory in 25% to 26% and the hypoglossal in 5% to 21%. Other morbidities included a CSF leakage rate of 8.3%, aspiration in 5.5% and wound infection in 5.5%. Although cranial nerve deficits occur more frequently with conventional surgery, most reports note that the long-term impact of such dysfunction is relatively small. It is important for the surgeon to take into account patient function, age and general health, and tumor size when discussing and weighing treatment options for glomus jugulare tumors.

Meningiomas are the second most common benign neoplasm of the cerebellopontine angle and can often present with deficits similar to vestibular schwannomas. Total resection results in excellent tumor control rates and, for all cranial locations, shows a 15-year progression-free survival rate of approximately 68% to 75%.^27,²⁸ Experience with partially resected or inoperable meningiomas, however, has shown that radiation therapy can produce excellent tumor control in the majority of cases. The use of stereotactic radiosurgery as a primary treatment to avoid or reduce the incidence of surgical and neurological deficits is increasingly common.

Elia and colleagues reviewed stereotactic radiosurgery outcomes for meningioma published since 2001.²⁸ In over 1500 patients the 5-year progression-free control rate was 93.4%. The complication rate ranged from 2.5% to 13% and included neurological and vascular toxicities. Many of these were for tumors around the optic chiasm and carotid arteries and included dosages up to 20 Gy. Kreil and colleagues recently published their series on the treatment of 200 skull base meningiomas with gamma knife surgery.²⁹ There were 21 patients with cerebellopontine angle lesions. Of 20 patients with preoperative hearing loss (not quantified), one improved and 19 remained stable; none showed deterioration. Tinnitus remained stable in seven of seven patients. Vertigo was present in 25 skull base meningiomas and improved in eight and worsened in none. Given the low incidence of complication and the high rate of tumor control, stereotactic radiosurgery should be strongly considered in tumors around sensitive neural structures and in patients medically unsuitable for conventional surgery.

In addition to tumors, the neurotologist is often consulted for facial pain syndromes, most notably trigeminal neuralgia. Functional stereotactic radiosurgery using gamma knife has been employed in the treatment of trigeminal neuralgia. In the series of meningiomas reported by Kriel there were 25 patients with preoperative trigeminal neuralgia due to tumor of which 16 improved.²⁹ There were two induced cases of trigeminal neuralgia but these were transient. Such findings indicate that radiation to the trigeminal nerve can induce functional changes.

Gorgulho and De Salles reviewed surgical and stereotactic treatments for trigeminal neuralgia.³⁰ Among current treatments, long-term improvement was noted in 70% to 75% of microvascular decompressions, 58% to 77% of radiofrequency rhizotomies, 32% of balloon compressions, 17% to 50% of glycerol rhizotomies, and 45% to 57% of stereotactic radiosurgeries. Immediate improvement was noted in over 90% of patients with stereotactic radiosurgery. Recurrence rates were highest with glycerol rhizotomy and much lower and very similar among the other modalities. Stereotactic radiosurgery was noted to be particularly attractive because it is the least invasive of these methods.

Many different treatment protocols for trigeminal neuralgia have been attempted.³⁰ In their review, Gorgulho and De Salles identified several patterns with regards to gamma knife treatment for trigeminal neuralgia that affect outcomes. The root entry zone of the trigeminal nerve, not the nerve proper, should be the preferred target as dosage delivery to this area seems to correlate with pain relief. A minimal dosage of 70 Gy and maximal dosage of 90 Gy should be prescribed. The incidence of post-treatment numbness with this prescription dose ranges from 3% to 55% but bothersome numbness persists in only about 4% to 12%. Treating a longer section of the trigeminal nerve proper does not improve pain control and increases the incidence of post-treatment numbness. Likewise, higher dosage to the nerve does not improve pain control and increases numbness. The overall incidence of complications with stereotactic radiosurgery for trigeminal neuralgia is significantly lower than all other techniques. As with other benign diseases, potential long-term effects of radiation treatment need to be considered in younger individuals.