Stereotactic radiosurgery has become a common therapeutic choice for patients with acoustic tumors (vestibular schwannomas). Experience with radiosurgery now extends over thirty years. During the late 1980's and early 1990's, patients and their doctors chose radiosurgery or resection based mainly on early outcomes data from limited patient series (3-10). In 1987 we began a prospective assessment of the response of patients with acoustic tumors to gamma knife radiosurgery. Both early and later (10-15 year) outcomes were determined through the use of serial imaging studies, hearing and facial function examinations, and physician-based evaluations (11). Because expected outcomes may be different for patients with solitary tumors, or those with neurofibromatosis type-2, we have analyzed these patient populations separately.

Patient Characteristics: University of Pittsburgh
Over twelve-hundred patients underwent stereotactic radiosurgery for an acoustic tumor (vestibular schwanoma) at the University of Pittsburgh over a 19-year interval. These included 1123 patients with solitary tumors and 84 with NF-2. Fifty percent of tumors were left-sided and fifty percent were right-sided. Fifty percent of patients were female. The patient age range was 12 to 95 years (mean, 56 years).

A resection had been performed in 11% of patients. Most patients had normal facial function (House-Brackmann grade 1)(12). The Gardner/Robertson scale was used to code hearing function (13). "Useful" hearing before radiosurgery was noted by 33% of patients.

In our last review of 45 patients with NF-2, prior resection was performed in 13 (16). Multiple resections were performed in four patients. Normal facial function before radiosurgery was present in 74%, normal trigeminal function in 75%, and useful hearing (Gardner/Robertson grades 1+2) in 31%.

Technique of Gamma Knife Radiosurgery
All patients underwent stereotactic radiosurgery using the Gamma Knife (Elekta Instruments, Atlanta GA) supplemented with local anesthesia and intravenous sedation as necessary. Children under the age of 12 years with NF-2 had radiosurgery under general anesthesia. Radiosurgery was performed with computed tomographic (CT) imaging between 1987 and 1991. Subsequent patients underwent radiosurgery using magnetic resonance imaging (MRI) after a prospective comparison study confirmed the accuracy of MR-based stereotactic targeting (14). Multiple irradiation isocenters were used to conform the radiation margin to the intracanalicular and extracanalicular tumor components (7). The 50% isodose line was used to cover the tumor margin in 696 patients with solitary tumors (88%). An initial tumor margin dose of 18-20 Gy was selected based on the initial experience from the Karolinska group in Stockholm (8). By 1992 was decreased further to a margin dose of approximately 14 Gy. Repeated re-evaluations of the cranial nerve response prompted additional small decreases in dose in order to preserve cranial nerve function (7,15). At the present time the usual dose to the tumor margin is 12 to 13 Gy depending on the tumor size, hearing status and prior history. This dose range has been in use for over ten years (16,17). Dose selection in individual patients is based on the factors of tumor volume, prior surgical history, hearing status, facial motor function, and patient desires. After radiosurgery, patients are discharged from hospital the same day. They may return to their activities immediately.

Follow-Up Evaluations
Serial imaging studies (MRI, or CT when MRI was contraindicated) were requested every six months for the first two years, annually for the next two years, and then bi-annually. Serial audiograms were obtained at 6 - 12 month intervals in patients with hearing. Contrast-enhanced imaging studies were used to define the tumor response and to identify any peri-tumoral imaging changes. Before and after radiosurgery, each tumor was measured in five separate dimensions (three extracanalicular and two intracanalicular) using a method previously reported (1). A significant imaging change using this caliper technique was defined as a difference of + 2 mm.

The Long-term Experience
We continue to evaluate a cohort of patients managed before 1992 who are a minimum of 10 years out from their procedure (n=162). This study represented results of our initial techniques (11). The majority of irradiated acoustic tumors (approximately 70%) decreased in size over time. Nine patients had tumors that increased in size and all were identified within the first three years after radiosurgery (11). Enlargement represented either true neoplastic tumor growth (n=4) or tumor death with an expansion of the tumor margins as the central portion of the tumor became necrotic. In the latter patients (n=5) subsequent imaging studies confirmed tumor volume regression. Four patients underwent resection. No further increase in tumor volume was identified in any patient with further follow-up (15). Patients returned to their routine activities immediately. In our 5 to 10 year review, three patients developed hydrocephalus and required a ventriculoperitonal shunt (11). All new or worsened post-radiosurgery deficits occurred within 28 months of radiosurgery and no patient described a treatment related problem after the third year.

Current Experience with Solitary Tumors
Refinements in technique followed a continued review of results. In 1991 we began to use MRI-based stereotactic planning since CT-based planning did not show well the intracanalicular portion of the tumor. With MRI, we could image the tumor and regional neural structures in greater detail. This facilitated the use of multiple small irradiation isocenters for more conformal radiosurgery. With this type of radiosurgery, cranial nerve morbidity dropped precipitously. Similarly, our analysis of hearing preservation in NF-2 patients (below) showed significant gains.

One hundred and ninety-two patients had radiosurgery between 1992 and 1997 and were eligible for extended follow-up (15). The maximum follow-up in this cohort was 65 months. The median tumor margin dose was 13 Gy. The actuarial five-year clinical tumor control rate (no need for any additional treatment) was 97%. One patient underwent a resection six months after radiosurgery. Five year actuarial rates of developing any facial weakness, facial numbness, hearing level preservation, and preservation of testable speech discrimination were 1.1+ 0.8%, 2.6+1.2%, and 71+4.7%, and 91+ 2.6% respectively. At a tumor margin dose of < 13 Gy, the rate of facial neuropathy was 0%, and above 13 Gy, 2.5% (usually mild and transient). Tumor diameter did not significantly affect results. Reports from other centers have shown similar results (37).

Neurofibromatosis type-2
Serial imaging studies of 45 tumors over a median 36 month follow up (range, 6-120 months) found that 16 tumors (36%) had regressed, 28 tumors (62%) remained unchanged in size, and 1 tumor (2%) demonstrated progression (16). Loss of central contrast within the tumor was observed in some patients and thought to reflect tumor necrosis.

At our last detailed review, the mean period of clinical followup was 41 months (range, 6-120). No patient demonstrated improvement in their clinical examination after radiosurgery. Thirty patients (67%) maintained a stable exam and 15 patients (33%) demonstrated some degree of clinical deterioration. Two patients (4%) died during the follow up period secondary to unrelated illnesses. The median Karnofsky score after radiosurgery was 80. Thirty-five patients (78%) were able to carry out normal daily activities at the time of the last examination (Karnofsky > 80).

Of the 14 tumors associated with useful hearing (Gardner-Robertson grades 1 or 2) at the time of radiosurgery, 6 (43%) demonstrated no change in hearing class during the follow-up period. Eight other patients lost all functional hearing (defined as absent speech discrimination) at a mean 6 months from radiosurgery (range, 3-15 months). The overall rate of hearing preservation in the series was 43%. In 1992, we began to use MRI-guided stereotactic planning with increasing numbers of smaller isocenters. By specifically dividing the population into those patients treated prior to 1992 and those treated after 1992, the difference in hearing preservation again becomes apparent. Prior to 1992, 5 patients with useful hearing (grade 1 or 2) were treated. All patients subsequently lost speech discrimination. After 1992, 9 patients with useful hearing underwent radiosurgery. Six of the patients (67%) had hearing preservation at the time of last examination (16). Thirty-one tumors (69%) were associated with intact facial nerve function (House-Brackman grade I) at the time of radiosurgery. The overall rate of facial nerve preservation (grade 1) was 81%. Thirty-six of the treated tumors were associated with intact trigeminal nerve function. Three patients (8%) experienced trigeminal distribution sensory loss at a mean 5 months (range, 4-5 months) from radiosurgery. One patient subsequently recovered all trigeminal function, while 2 patients manifest residual deficits. The overall rate of trigeminal nerve preservation was 94%.

© 2001-06 The Department of Neurosurgery, at the University of Pittsburgh. This web site is intended for educational purposes only. The information provided through this web site should not be used for diagnosing or treating a health problem or disease. Please see your physician before taking any medical action. Do not duplicate in any form. All rights reserved. Please read full legal disclaimer and privacy policy.

Please direct all web-specific inquiries to our web manager.