Ac o u s t i c neuromas, also known as vestibular. Predictors of hearing preservation after stereotactic radiosurgery for acoustic neuroma

Transcription

1 J Neurosurg 111: , 2009 Predictors of hearing preservation after stereotactic radiosurgery for acoustic neuroma Clinical article Hi d e y u k i Ka n o, M.D., Ph.D., 1,3 Do u g l a s Ko n d z i o l k a, M.D., F.R.C.S.(C), 1,3 Af ta b Kh a n, M.D., 1,3 Jo h n C. Fl i c k i n g e r, M.D., 2,3 a n d L. Da d e Lu n s f o r d, M.D. 1,3 Departments of 1 Neurological Surgery and 2 Radiation Oncology, and the 3 Center for Image-Guided Neurosurgery, University of Pittsburgh School of Medicine, Pittsburgh, Pennsylvania Object. Many patients with acoustic neuromas (ANs) have hearing function at diagnosis and desire to maintain it. To date, radiosurgical techniques have been focused on conformal irradiation of the tumor mass, with less attention to inner ear structures for which there was scant radiobiological information. The authors of this study evaluated tumor control and hearing preservation as they relate to tumor volume, imaging characteristics, and nerve and cochlear radiation dose following stereotactic radiosurgery (SRS) using the Gamma Knife. Methods. Seventy-seven patients with ANs had serviceable hearing (Gardner-Robertson [GR] Class I or II) and underwent SRS between 2004 and This interval reflected more recent measurements of inner ear dosimetry during the authors 21-year experience. The median patient age was 52 years (range years). No patient had undergone any prior treatment for the ANs. The median tumor volume was 0.75 cm 3 (range cm 3 ), and the median radiation dose to the tumor margin was 12.5 Gy (range Gy). At diagnosis, a greater distance from the lateral tumor to the end of the internal auditory canal correlated with better hearing function. Results. At a median of 20 months after SRS, no patient required any other additional treatment. Serviceable hearing was preserved in 71% of all patients and in 89% (46 patients) of those with GR Class I hearing. Significant prognostic factors for maintaining the same GR class included (all pre-srs) GR Class I hearing, a speech discrimination score (SDS) 80%, a pure tone average (PTA) < 20 db, and a patient age < 60 years. Significant prognostic factors for serviceable hearing preservation were (all pre-srs) GR Class I hearing, an SDS 80%, a PTA < 20 db, a patient age < 60 years, an intracanalicular tumor location, and a tumor volume < 0.75 cm 3. Patients who received a radiation dose of < 4.2 Gy to the central cochlea had significantly better hearing preservation of the same GR class. Twelve of 12 patients < 60 years of age who had received a cochlear dose < 4.2 Gy retained serviceable hearing at 2 years post-srs. Conclusions. As currently practiced, SRS with the Gamma Knife preserves serviceable hearing in the majority of patients. Tumor volume and anatomy relate to the hearing level before radiosurgery and influence technique. A low radiosurgical dose to the cochlea enhances hearing preservation. (DOI: / JNS08611) Ke y Wo r d s acoustic neuroma cochlea Gamma Knife hearing preservation stereotactic radiosurgery vestibular schwannoma Abbreviations used in this paper: AN = acoustic neuroma; GR = Gardner-Robertson; IAC = internal auditory canal; PTA = pure tone average; SDS = speech discrimination score; SRS = stereotactic radiosurgery; SRT = stereotactic radiotherapy. Ac o u s t i c neuromas, also known as vestibular schwannomas, are benign tumors arising from the vestibulocochlear nerve sheath. In recent years ANs have been diagnosed more frequently and at earlier stages of presentation because of the widespread availability of MR imaging. The tumor location may be intracanalicular only or extend outside the IAC. There is variable extension of the tumor along the course of the vestibular nerve. 1,19 The most frequent presenting symptoms include tinnitus, hearing loss, gait disturbances, facial numbness, and weakness. 3 Common hypotheses for the development of hearing loss include direct compression of cochlear nerve fibers by an adjacent AN, the development of a conduction block followed by the degeneration of nerve fibers, compression and/or thrombosis of the internal auditory artery, and/or ischemic injury to the cochlea. 3,12 Therapeutic options include observation, microsurgical removal, SRS, SRT, or other forms of fractionated radiation therapy. Presently, radiosurgery is a well-established alternative to microsurgical removal of an AN. 10 In a recent study in which Gamma Knife surgery was utilized, Paek et al. 21 suggested that the maximum 863

2 H. Kano et al. dose delivered to the cochlear nucleus in the brainstem was the only significant prognostic factor for hearing deterioration. Massager et al. 16,17 showed that hearing loss following SRS was related to the volumetric and dosimetric parameters of the intracanalicular tumor and that the dose to the cochlea, rather than the cochlear nucleus, is relevant. We studied tumor anatomy, radiosurgical planning, and the radiation dose delivered to the inner ear structures to define a relationship between radiosurgical technique, tumor extent, hearing function at presentation, and later outcome. Methods Patient Population The University of Pittsburgh Institutional Review Board approved this retrospective study. Between October 2004 and March 2007, 248 consecutive patients with previously untreated unilateral ANs underwent SRS with the Gamma Knife (Elekta Instruments) at the University of Pittsburgh. For this series of patients, we excluded those with neurofibromatosis Type II, unserviceable hearing (GR Class III V) at the time of SRS, and < 6 months of radiological and audiological follow-up data (8 patients with GR Class I, and 2 with GR Class II hearing). Twelve patients with GR Class III hearing, 4 with GR Class IV, and 20 with GR Class V were excluded. The demographic characteristics of this study population of 77 patients are listed in Table 1. There were 40 men and 37 women with a median age of 52 years (range years). Audiography results were evaluated according to the GR classification. 8 Serviceable hearing (useful hearing) was defined as GR Class I II (SDS 50% and PTA 50 db). Before SRS, 46 patients had GR Class I hearing (SDS 70% and PTA 30 db) and 31 patients had GR Class II (SDS 69 50% and PTA db). Radiological results were evaluated according to the Koos system of classification: 11 Grade I, 28 patients (GR Class I in 22 patients, GR Class II in 6); Grade II, 18 patients (GR Class I in 8 patients, GR Class II in 10); Grade III, 25 patients (GR Class I in 12 patients, GR Class II in 13); and Grade IV, 6 patients (GR Class I in 4 patients, GR Class II in 2). Radiosurgery Technique Radiosurgery was performed using a Model C or 4-C Leksell Gamma Knife (Elekta, Inc.). Our radiosurgical technique has been described in detail in previous reports. 5 The procedure began with the application of a Model G Leksell stereotactic frame after inducing conscious sedation and applying a local scalp anesthetic, except in younger children in whom general anesthesia was induced. The tumor was then visualized using high-resolution 3D spoiled gradient recalled acquisition in steady state sequence MR imaging after intravenous contrast enhancement. Fast spin echo T2-weighted MR images were acquired to evaluate tumor extent and inner ear structures. We obtained images of the cochlea, vestibule, and semicircular canals using T2-weighted volume-acquisition MR imaging divided into axial images at 1- to 1.5-mm intervals. Figures 1 to 4 show the cochlear modiolus, indicated by a cruciform mark. Images were exported to a computer workstation for dose planning with Leksell GammaPlan software. In all tumors the radiosurgery volume conformed to the enhancing tumor volume. The median tumor volume was 0.75 cm 3 ( cm 3 ). The median prescription dose delivered to the tumor margin was 12.5 Gy (12 13 Gy). The prescription isodose was 50% in 65 cases. The maximum radiation dose varied from 18.6 to 26 Gy (median 25 Gy). All patients received an intravenous dose of mg methyl- TABLE 1: Summary of characteristics in 77 patients with AN Characteristic GR Class I GR Class II Entire Series no. of patients median age in yrs (range) (22 82) no. of men no. of women median PTA in db (range) 16 (0 30) 40 (18 49) 24 (0 49) median SDS in % (range) 98 (76 100) 84 (56 100) 92 (56 100) median dose to central cochlea in Gy (range) 4.5 ( ) 4.6 ( ) 4.5 ( ) median distance from tumor to IAC in mm (range) 2.7 (0 9.0) 2.1 (0 8.8) 2.5 (0 9.0) Koos grade (no. of patients) I II III IV median target vol in cm 3 (range) 0.47 ( ) 1.1 ( ) 0.75 ( ) median tumor margin dose in Gy (range) 12.5 (12 13) 12.5 ( ) 12.5 (12 13) median max radiation dose in Gy (range) 25 ( ) 25 ( ) 25 ( ) 864

3 Radiosurgery for acoustic neuroma Fig. 1. A and B: Axial T2-weighted MR images obtained in a 67-year-old woman, showing a right AN (Koos Grade III at the time of SRS). C and D: Axial T1-weighted contrast-enhanced MR images showing a right AN. One-sided arrow indicates the well-visualized cochlea; cruciform marks indicate the center of the cochlea, or the modiolus (dose = 3.1 Gy); and the double-sided arrow shows the distance from the lateral end of tumor to the end of the IAC (3.9 mm). The radiosurgery plan is shown to deliver 12.5 Gy to the tumor margin. prednisolone after radiosurgery, and all were discharged from the hospital within 2 hours. Patient Follow-Up Patients were instructed to undergo clinical and imaging assessments after radiosurgery at 6 months, annually for 2 years, and at less frequent intervals thereafter (every 4 years past Year 10). If a new neurological symptom or sign (especially hearing deterioration) developed, the patient was evaluated for audiological deterioration, tumor progression, or any adverse radiation effect, and new audiological testing and MR imaging were performed. All patients had a minimum of 6 months of follow-up (range 6 40 months, median 20 months). Thirty-three patients had follow-ups 24 months. Tumor control was assessed in 2 ways. Radiologically demonstrated tumor progression was strictly defined as any temporary or sustained increase in tumor diameter of at least 1 mm in 2 dimensions or 2 mm in any direction. We assessed hearing preservation with follow-up audiography, using the end points of preserving a specific GR hearing class (Class I II) or serviceable hearing. Based on MR images of each radiosurgical plan, we defined the dose to the central cochlea (modiolus), at the vestibule, and in the middle of the horizontal semicircular canal, as well as the distance from the end of the tumor to the end of the IAC (Figs. 1 4). Radiation doses at these points were derived using Leksell GammaPlan software. Statistical Analysis For statistical analysis we constructed Kaplan-Meier plots for GR class of hearing function, serviceable hearing preservation, and progression-free survival using the date of SRS, follow-up audiological tests or MR images, and last follow-up data. Results were calculated from the day of SRS using the Kaplan-Meier method. Univariate analysis of the Kaplan-Meier curves was performed using the log-rank statistic, with a probability level < 0.05 set as significant. The Fisher exact test was applied to study the relationship between the number of patients who experienced hearing deterioration and a number of other specific variables, with a probability level < 0.05 set as significant. A comparison of continuous variables was performed using the Mann-Whitney 2-sample t-test, again with a probability level < 0.05 set as significant. Multivariate analysis was performed with the Cox proportional hazards model, with a probability level < 0.10 set as significant. Standard statistical processing software (SPSS, version 15.0, SPSS, Inc.) was used. 865

4 H. Kano et al. Fig. 2. A and B: Axial T2-weighted MR images obtained in a 54-year-old man, demonstrating a right AN (Koos Grade II at the time of SRS). C and D: Axial T1-weighted contrast-enhanced MR images showing a right AN. Arrow indicates the cochlea, cruciform mark shows the cochlear modiolus (dose = 2.6 Gy), and double-sided arrow shows the distance from the lateral end of the tumor to the end of the IAC (2.5 mm). The Gamma Knife surgery plan is shown. Results No patient has required any other procedure since the SRS. Tumor control was achieved in 75 (97.4%) of the 77 patients at the last follow-up. Two patients demonstrated a small expansion (2 mm) of the extracanalicular tumor component on early imaging after SRS. Hearing Function Before SRS Twenty-two of 46 patients with GR Class I hearing had an intracanalicular tumor only (Koos Grade I), whereas 6 of 31 patients with GR Class II hearing had an intracanalicular tumor (p = 0.009, Fisher exact test). The mean and median age of patients with GR Class I hearing were 49.4 and 51.0 years, respectively, whereas those for patients with GR Class II were 55.7 and 53.0 years, respectively (p = 0.017, Mann-Whitney t-test). The mean and median tumor volume of patients with GR Class I hearing were 1.20 and 0.47 cm 3, respectively, whereas those of patients with GR Class II hearing were 1.91 and 1.1 cm 3, respectively (p = 0.030, Mann-Whitney t-test). The median radiation dose to the center of the vestibule in patients with GR Class I and II hearing were 6.3 and 7.3 Gy, respectively (respective mean distance 6.57 ± 2.48 vs 7.35 ± 2.55, p = 0.193, Mann-Whitney t-test). The median radiation dose to the semicircular canal in patients with GR Class I and II were both 3.8 Gy (respective mean distance 3.78 ± 0.93 vs 4.13 ± 1.44, p = 0.238, Mann-Whitney t-test). The distance from the lateral end of the tumor to the end of the IAC (lateral extent) was not a predictor of hearing preservation; however, it was significantly associated with the GR class before SRS (mean distance for GR Class II vs 2: 3.44 ± 2.29 vs 2.32 ± 2.10, p = , Mann-Whitney t-test). In addition, the distance from the lateral end of the tumor to the end of the IAC (lateral extent) was associated with the GR class before SRS (mean distance for GR Class I II vs Class III V: 2.89 ± 2.28 vs 2.10 ± 2.40, p = 0.045, Mann-Whitney t-test; GR Class I, II, and III vs V: p = 0.001, 0.026, and 0.007, respectively, Mann-Whitney t-test). Thus, GR Class I was significantly associated with an intracanalicular tumor location, a younger age, a smaller tumor volume, and a longer segment of the IAC free of tumor. Fifty-seven (36 with GR Class I and 21 with GR Class II) of 77 patients had tinnitus before radiosurgery. After radiosurgery, tinnitus improved in 7 patients, worsened in 5, and remained unchanged in 45. Twenty other patients remained tinnitus free after radiosurgery. Hearing Preservation After SRS The last audiological examination revealed that

5 Radiosurgery for acoustic neuroma Fig. 3. A and B: Axial T2-weighted MR images obtained in a 51-year-old man, revealing a right AN (Koos Grade III at the time of SRS). C and D: Axial T1-weighted contrast-enhanced MR images showing a right AN. Arrow indicates the cochlea, cruciform mark shows the modiolus (dose = 4.1 Gy), and double-sided arrow demonstrates the distance from the lateral end of the tumor to the end of the IAC (3.2 mm). patients (58.4%) retained the same (pre-srs) GR hearing class after SRS. Two patients (2.6%) had improved hearing from GR Class II to I. Thirty patients (39.0%) experienced some reduction in hearing after SRS. Hearing was reduced by 1 GR class in 23 patients (8 patients from GR Class I to II and 15 patients from GR Class II to III), by 2 classes in 4 patients (all 4 patients from GR Class I to III), by 3 classes in 2 patients (both from GR Class II to V), and by 4 classes in 1 patient (from GR Class I to V). Audiological tests in 22 patients (28.6%) showed deterioration to unserviceable hearing (GR Class III V) after SRS; thus, 55 (71.4%) of 77 patients maintained serviceable hearing (GR Class I II) following SRS (Table 2). Eighty-four and 56.5% of patients maintained the same GR class at 1 and 2 years post-srs, respectively. The median time for dropping to a lower GR class after SRS was 25.0 ± 5.1 months (mean ± SD). At 1 and 2 years after SRS, 89.3 and 66.8% of respective patients had preserved hearing. Among those whose hearing became unserviceable, the mean time for unserviceable hearing to develop was 31.1 ± 3.5 months. Statistical Analysis We performed univariate analysis using the log-rank and Fisher exact test to assess factors that might influence hearing preservation (remaining within the same GR class or retaining serviceable hearing). The following variables were assessed: sex (male vs female), age ( vs < 60 years), prescription SRS target volume ( vs < 0.75 cm 3 ), radiation dose to tumor margin ( vs < 12.5 Gy), Koos grade (Grade I II vs III IV), intracanalicular tumor only (yes vs no), radiation dose to the central cochlea ( vs < 3.8, 4.0, 4.2, and 4.5 Gy), GR hearing class before SRS (Class I vs II), SDS before SRS ( vs < 70, 80, 90, and 100%), PTA before SRS ( vs < 20 and 30 db), and distance from the lateral end of the tumor to the end of the IAC ( vs < 2 and 4 mm). On univariate testing for the entire series, pre-srs factors associated with an improved rate of hearing preservation (remaining within the same GR class) included a younger age, GR Class I hearing, a higher SDS, and a lower PTA. The pre-srs factors associated with an improved rate of serviceable hearing preservation included a younger age, an intracanalicular tumor only, a smaller tumor volume, GR Class I hearing, higher SDS, and lower PTA (Table 3). In addition to the factors associated with a better chance of preserving the GR class, the Fisher exact test showed that a radiation dose < 4.2 Gy to the central cochlea was significantly associated with better odds of preserving hearing function (p = 0.022). We performed multivariate analysis using the Cox proportional hazards model to assess factors that might 867

6 H. Kano et al. Fig. 4. A and B: Axial T2-weighted MR images obtained in a 47-year-old man, showing a right AN (Koos Grade II at the time of SRS). C and D: Axial T1-weighted contrast-enhanced MR images revealing a right AN. Arrow shows the cochlea; a cruciform mark, the modiolus (dose = 8.6 Gy); and double-sided arrow, the distance from the lateral end of the tumor to the end of the IAC (0 mm), because the tumor filled the entire canal. The patient had poor hearing. influence the rates of hearing preservation. The following variables were assessed: radiation dose to the central cochlea, tumor volume, distance from the lateral end of the tumor to the end of the IAC, and GR hearing class before SRS. On multivariate analysis, the only factor associated with improved odds of preserved hearing (same GR class or serviceable hearing preservation) was GR Class I hearing before SRS (p = and p = , respectively). Predictors of Hearing Outcome Patients with GR Class I hearing before SRS had a significantly better chance of remaining within the same GR class (p = 0.029, log-rank test; p = 0.013, Fisher exact test) and retaining serviceable hearing (p = 0.012, log-rank test; p = 0.070, Fisher exact test). Patients with a PTA < 20 db before SRS remained within the same GR class (p = , log-rank test; p = , Fisher exact test) and had better odds of maintaining serviceable hearing (p = , log-rank test; p = 0.001, Fisher exact test). Fourteen (93.3%) of 15 patients with a pre-srs SDS < 80% exhibited some deterioration in hearing (any drop in GR class) compared with 15 (25.0%) of 60 patients TABLE 2: Hearing outcome after SRS GR Class 868 No. (%) Preserved Hearing Variable Improved Stable Worse Serviceable Unserviceable pre-srs 2 (2.6) 45 (58.4) 30 (39.0) 55 (71.4) 22 (28.6) GR class I 0 (0) 33 (71.7) 13 (28.3) 41 (89.1) 5 (10.9) II 2 (6.5) 12 (38.7) 17 (54.8) 14 (45.2) 17 (54.8) age <60 yrs 0 (0) 41 (69.5) 18 (30.5) 45 (76.3) 14 (23.7) 60 yrs 2 (11.1) 4 (22.2) 12 (66.7) 10 (55.6) 8 (44.4)

7 Radiosurgery for acoustic neuroma TABLE 3: Statistical analysis of hearing preservation for entire series* Maintaining Same GR Class Serviceable Hearing Preservation Parameter Log-Rank Test Fisher Exact Test Log-Rank Test Fisher Exact Test dose to central cochlea vs < 4.0 Gy vs < 4.2 Gy * vs < 4.5 Gy distance from lat end of tumor to end of IAC or < 2 mm or < 4 mm age, vs < 60 yrs 0.001* 0.011* 0.012* Koos grade, Grade I II vs III IV intracanalicular tumor, yes vs no * 0.006* target vol, vs < 0.75 cm * 0.003* GR class before SRS, Class I vs II 0.029* 0.013* < 0.001* <0.0001* SDS before SRS vs < 70% * 0.002* * * vs < 80% <0.0001* <0.0001* <0.0001* <0.0001* vs < 90% <0.0001* <0.0001* <0.0001* <0.001* vs < 100% 0.007* 0.001* 0.017* 0.006* PTA > vs < 30 db * 0.002* 0.001* > vs < 20 db * * * 0.001* * Statistically significant. with an SDS 80% before SRS. Among patients with an SDS 80% before SRS, 96.3 and 67.8% maintained the same GR class at 1 and 2 years, respectively, after SRS. Thus, patients with an SDS 80% had significantly better odds of maintaining the same GR class (p < , logrank test; p < , Fisher exact test) as well as serviceable hearing (p < , log-rank test; p < , Fisher exact test; Table 3). Relationship Between Patient Age and Hearing Preservation Eighteen patients (23.4%) were older than 60 years, and 59 patients (76.6%) were 60 years or younger. Twelve (66.7%) of 18 older patients exhibited some hearing deterioration (any drop in GR class) compared with only 18 (30.5%) of 59 younger patients. Overall, 94.4 and 69.6% of younger patients maintained the same GR class at 1 and 2 years after SRS, respectively; older patients did not fare as well (48.9 and 17.5%, respectively). Thus, a patient age < 60 years was significantly associated with the maintenance of a GR class (p = 0.001, log-rank test; p = 0.011, Fisher exact test; Table 3). Influence of Cochlear Radiation Dose on Hearing Preservation Both the mean and median radiation dose to the central cochlea in this study were 4.5 Gy (range ). Although not significantly different (p = 0.843, Mann- Whitney t-test), the median radiation dose to the central cochlea was slightly higher in patients with hearing deterioration (4.55 vs 4.40 Gy). Overall, patients who received < 4.2 Gy to the central cochlea had significantly better odds of maintaining the same GR class (p = 0.022, Fisher exact test). Among patients < 60 years old, only 3 (13.6%) of 22 who had received < 4.2 Gy to the central cochlea exhibited some hearing deterioration (any drop in GR class) compared with 15 (40.5%) of 37 patients who had received a radiation dose 4.2 Gy. Among patients with a cochlear dose < 4.2 Gy, 94.7 and 86.8% maintained the same GR class at 1 and 2 years after SRS, respectively. Only 66.1% of patients who had received a higher cochlear dose ( 4.2 Gy) maintained the same GR class at 2 years post- SRS (Fig. 5A). The median time for hearing deterioration to a lower GR class was 25.0 months among patients who had received a cochlear dose 4.2 Gy. Similarly, patients who had received the lower cochlear dose (< 4.2 Gy) had better odds of both maintaining a GR class ( p = 0.032, log-rank test; p = 0.024, Fisher exact test;) and preserving serviceable hearing ( p = 0.027, log-rank test; p = 0.031, Fisher exact test; Fig. 5B and Table 4). On multivariate testing of patients < 60 years old, 2 factors were associated with improved hearing preservation: GR Class I before SRS (p = and p = for same GR class or serviceable hearing, respectively) and a central cochlear dose < 4.2 Gy (p = and p = for same GR class or serviceable hearing, respectively). 869

8 H. Kano et al. post-srs. Although not enough patients were treated to reach statistical significance, 80% of the 22 patients younger than 60 years who had received a cochlear dose 4.2 Gy were able to maintain serviceable hearing at 2 years after SRS (Table 5). Discussion Treatment options for patients with serviceable hearing despite an AN include observation, surgical lesion removal, SRS, and fractionated radiotherapy. 9,13,18,22,23,26 For many patients, choosing a specific form of therapy can be difficult. Physicians vary in their advice, published data can be difficult to interpret, and the Internet provides many confusing and unconfirmed opinions. Although hearing preservation was rarely offered as a goal to patients treated before 1990, this objective is now discussed with virtually all patients who have some level of hearing at diagnosis. In the past, radiosurgical techniques have been focused on conformal irradiation of the tumor mass, usually visualized with high-resolution contrastenhanced MR imaging. Less attention was directed at inner ear structures, which were poorly seen utilizing such sequences and for which there was scant radiobiological information. Yamakami et al. 26 collected data from conservative management over a 3-year period and found that one-half of ANs showed growth, one-third of the patients lost useful hearing, and 20% of ANs ultimately required surgical intervention. In a prospective study of radiosurgery versus resection, Pollock et al. 22 have reported that patients who underwent radiosurgery with the Gamma Knife had better results according to a health status questionnaire compared with patients in the resection group, including hearing preservation, normal facial movement, physical functioning, role limitation due to physical health, energy/fatigue, and overall component. Fig. 5. Upper: Kaplan-Meier curves comparing the maintenance of the same GR hearing class with a radiation dose < vs 4.2 Gy in the middle of the cochlea in patients < 60 years old. A radiation dose < 4.2 Gy was significantly associated with better odds of remaining within the same GR hearing class (p = 0.032). Lower: Kaplan-Meier curves comparing the maintenance of serviceable hearing with a radiation dose < vs 4.2 Gy in the middle of the cochlea in patients < 60 years old. Again, patients who had received a radiation dose < 4.2 Gy had significantly better odds of retaining serviceable hearing (p = 0.027). Best Opportunity for Hearing Preservation Among patients < 60 years old with GR Class I hearing before SRS, only 2 (16.7%) of 12 who had received a central cochlear dose < 4.2 Gy exhibited any GR class deterioration, compared with 6 (27.3%) of 22 patients who had received a higher dose. Ninety-three percent of the patients who had received the lower cochlear dose maintained their GR class at 2 years post-srs. In addition, all 12 patients < 60 years and treated with a cochlear dose < 4.2 Gy maintained serviceable hearing at 2 years Stereotactic Radiosurgery for ANs A recent analysis of our current experience with radiosurgery for ANs has indicated lower morbidity with a similar tumor control rate compared with those from our first experience between 1987 and ,6,10 We have reported a 6-year clinical tumor-control rate of 98.1% after Gamma Knife surgery performed with a median tumor margin dose of 13 Gy. In that study, the overall hearing preservation rate remained 70.3%. In previous studies with higher tumor margin doses, we found that cranial neuropathy seemed to occur within 2 years of SRS. 5 Radiosurgery at the current margin doses of Gy provides a high rate of tumor control and low morbidity, and facial neuropathy is rare. Hearing Preservation After SRS With all forms of therapeutic management, hearing preservation has been the greatest challenge. 5 Common hypotheses for hearing deterioration after irradiation include damage to cochlear primary sensory cells, injury to the cochlear nerve by the tumor, injury to the cochlear nerve by radiation, and compression or thrombosis of the 870

9 Radiosurgery for acoustic neuroma TABLE 4: Statistical analysis of hearing preservation for patients < 60 years* Maintaining Same GR Class Serviceable Hearing Preservation Parameter Log-Rank Test Fisher Exact Test Log-Rank Test Fisher Exact Test dose in central cochlea > vs < 3.8 Gy > vs < 4.0 Gy 0.038* 0.035* 0.022* 0.024* > vs < 4.2 Gy 0.032* 0.024* 0.027* 0.031* > vs < 4.5 Gy distance from lat end of tumor to end of IAC > or < 2 mm > or < 4 mm Koos grade, Grade I II vs III IV intracanalicular tumor, yes vs no * 0.016* 0.005* target vol, > vs < 0.75 cm * 0.021* 0.004* GR class before SRS, Class I vs II * 0.008* 0.002* SDS before SRS > vs < 70% * 0.025* > vs < 80% 0.002* <0.0005* 0.014* 0.012* > vs < 90% 0.007* 0.009* 0.007* 0.013* 100% vs < 100% * PTA > vs < 20 db * 0.018* > vs < 30 db * 0.016* 0.005* * Statistically significant. internal auditory artery, leading to ischemic injury of the cochlea. 3,12 Chang et al. 1 have reported on therapeutic outcomes in 61 patients treated using 3-stage radiosurgery with more than 36 months of follow-up. These authors reported useful hearing preservation in 74% of their patients, which they reported as a crude rate as opposed to a 10- year actuarial rate. In our previous series of 110 patients with testable hearing and > 3 years of follow-up, crude hearing preservation rates were 78% for retaining the exact same GR hearing class (Classes I IV) and 77% for serviceable hearing (starting with GR Class I or II hearing and preserving at least Class II). 2 In the present study crude hearing preservation rates were as follows: 61% of patients remained within the exact same GR class and 71.4% retained serviceable hearing. Predictors of Hearing Preservation This study had a mean follow-up of 20 months (range TABLE 5: Hearing preservation in patients < 60 years of age in GR Class I* Central Cochlear Dose (Gy) No. Worse/No Change Maintaining Same GR Class Crude Rate (%) 2-Yr HPR (%) Log- Rank Test Fisher Exact Test No. Worse/No Change Serviceable Hearing Preservation Crude Rate (%) 2-Yr HPR (%) Log-Rank Test Fisher Exact Test < 3.8 1/ / / / < 4.0 1/ / / / < 4.2 2/ / / / < 4.5 1/ / / / * HPR = hearing preservation rate. 871

10 H. Kano et al months), which is relatively short in terms of commenting on tumor control rates and long-term toxicity. Nevertheless, a similar criticism can be made about most of the other published studies on SRS, SRT, and resection. It is generally accepted that almost all cases of postradiosurgery facial, trigeminal, and auditory neuropathy occur 1 24 months after SRS, with a median onset of 6 months post-srs. 14,15,20 Most cranial nerve dysfunction after SRT also occurs in the first few months after treatment, with most researchers suggesting that treatment-related toxicity occurs primarily during the first 36 months. 7,24,25 In our study the significant pre-srs prognostic factors for retaining the same GR hearing class were GR Class I hearing, SDS 80 and 90%, pre-srs PTA < 20 db, and a patient age < 60 years old. The significant pre-srs prognostic factors for serviceable hearing preservation were GR Class I hearing, SDS 80 and 90%, pre-srs PTA < 20 db, and a patient age < 60 years old as well as an intracanalicular tumor only and a target volume < 0.75 cm 3. Patient age was a stronger predictor of retaining the same GR hearing class than was the pre- SRS GR class. Both SDS and PTA before SRS were the strongest predictors for remaining within the same GR class. These results indicate that younger patients with an intracanalicular AN and high-level hearing are the best SRS candidates for hearing preservation. Influence of Cochlear Radiation Dose on Hearing Preservation Paek et al. 21 have reported that 9 (36%) of 25 patients with serviceable hearing despite an AN kept their pre-srs GR hearing class and 13 patients (52%) retained serviceable hearing after SRS, whereas 16 patients experienced hearing deterioration > 20 db from 3 to 24 months after SRS. The only significant prognostic factor for hearing deterioration was the maximum radiation dose delivered to the cochlear nucleus (hearing unchanged vs hearing worsened: 9.1 ± 4.4 Gy vs 7.8 ± 2.6 Gy). The dose to the cochlea was not specified. Massager et al. 17 have reported that hearing preservation after SRS is significantly associated with the intracanalicular tumor volume and the integrated dose delivered across that volume. The average dose delivered to the cochlea was 3.70 Gy in patients with audiological preservation and 5.33 Gy in those with worsening hearing after SRS. In our study, patients who had received a radiation dose < 4.2 Gy to the central cochlea had significantly better odds of maintaining the same GR hearing class (p = 0.022, Fisher exact test). In the group of patients < 60 years old, the 4.2-Gy radiation dose to the central cochlea was associated with significantly better odds of maintaining the same GR class (p = 0.032, log-rank test) as well as serviceable hearing (p = 0.027, log-rank test); younger patients tended to be influenced by the radiation dose to the central cochlea. We can reduce the dose to the central cochlea by using a beam-blocking technique without reducing the tumor margin dose. Conclusions Among patients < 60 years old with GR Class I hearing, 93 and 100% of those who had received a radiation dose < 4.2 Gy to the central cochlea retained the same GR class and serviceable hearing, respectively, at 2 years post-srs. Such data are provocative and may support earlier radiosurgery in younger patients with high-level hearing. Many such patients are currently being observed for imaging-demonstrated tumor growth or hearing deterioration. Thus, we suggest that radiosurgery with the Gamma Knife be considered for younger patients with smaller ANs in an attempt to preserve functional hearing. Disclosure Drs. Lunsford and Kondziolka are consultants with Elekta AB. Dr. Lunsford is a stockholder in Elekta AB. The work described in this report was funded by a grant (H.K.) from the Osaka Medical Research Foundation for Incurable Diseases. References 1. Chang SD, Gibbs IC, Sakamoto GT, Lee E, Oyelese A, Adler JR Jr: Staged stereotactic irradiation for acoustic neuroma. Neurosurgery 56: , Chopra R, Kondziolka D, Niranjan A, Lunsford LD, Flickinger JC: Long-term follow-up of acoustic schwannoma radiosurgery with marginal tumor doses of 12 to 13 Gy. Int J Radiat Oncol Biol Phys 68: , Delbrouck C, Hassid S, Massager N, Choufani G, David P, Devriendt D, et al: Preservation of hearing in vestibular schwannomas treated by radiosurgery using Leksell Gamma Knife: preliminary report of a prospective Belgian clinical study. Acta Otorhinolaryngol Belg 57: , Flickinger JC, Kondziolka D, Niranjan A, Lundford LD: Results of acoustic neuroma radiosurgery: an analysis of 5 years experience using current methods. J Neurosurg 94:1 6, Flickinger JC, Kondziolka D, Niranjan A, Maitz A, Voynov G, Lunsford LD: Acoustic neuroma radiosurgery with marginal tumor doses of 12 to 13 Gy. Int J Radiat Oncol Biol Phys 60: , Flickinger JC, Kondziolka D, Pollock BE, Lunsford LD: Evolution of technique for vestibular schwannoma radiosurgery and effect on outcome. Int J Radiat Oncol Biol Phys 36: , Fuss M, Debus J, Lohr F, Huber P, Rhein B, Engenhart-Cabillic R, et al: Conventionally fractionated stereotactic radiotherapy (FSRT) for acoustic neuromas. Int J Radiat Oncol Biol Phys 48: , Gardner G, Robertson JH: Hearing preservation in unilateral acoustic neuroma surgery. Ann Otol Rhinol Laryngol 97:55 66, Karpinos M, Teh BS, Zeck O, Carpenter LS, Phan C, Mai WY, et al: Treatment of acoustic neuroma: stereotactic radiosurgery vs. microsurgery. Int J Radiat Oncol Biol Phys 54: , Kondziolka D, Lunsford LD, McLaughlin MR, Flickinger JC: Longterm outcomes after radiosurgery for acoustic neuromas. N Engl J Med 339: , Koos WT, Day JD, Matula C, Levy DI: Neurotopographic considerations in the microsurgical treatment of small acoustic neurinomas. J Neurosurg 88: , Lapsiwala SB, Pyle GM, Kaemmerle AW, Sasse FJ, Badie B: Correlation between auditory function and internal auditory canal pressure in patients with vestibular schwannomas. J Neurosurg 96: , Lee F, Linthicum F Jr, Hung G: Proliferation potential in recurrent acoustic schwannoma following gamma knife radiosurgery versus microsurgery. Laryngoscope 112: ,

11 Radiosurgery for acoustic neuroma 14. Levivier M, Massager N, David P: In vivo evaluation of tumor response to radiosurgery: application to vestibular schwannomas. Neurochirurgie 50: , Lunsford LD, Niranjan A, Flickinger JC, Maitz A, Kondziolka D: Radiosurgery of vestibular schwannomas: summary of experience in 829 cases. J Neurosurg 102: , Massager N, Nissim O, Delbrouck C, Derpierre I, Devriendt D, Desmedt F, et al: Irradiation of cochlear structures during vestibular schwannoma radiosurgery and associated hearing outcome. J Neurosurg 107: , Massager N, Nissim O, Delbrouck C, Devriendt D, David P, Desmedt F, et al: Role of intracanalicular volumetric and dosimetric parameters on hearing preservation after vestibular schwannoma radiosurgery. Int J Radiat Oncol Biol Phys 64: , Meijer OWM, Vandertop WP, Baayen JC, Slotman BJ: Singlefraction vs. fractionated linac-based stereotactic radiosurgery for vestibular schwannoma: a single-institution study. Int J Radiat Oncol Biol Phys 56: , Neely JG, Hough JV: Histologic findings in two very small intracanalicular solitary schwannomas of the eighth nerve: II. Onion bulbs. Am J Otol 9: , Niranjan A, Lunsford LD, Flickinger JC, Maitz A, Kondziolka D: Dose reduction improves hearing preservation rates after intracanalicular acoustic tumor radiosurgery. Neurosurgery 45: , Paek SH, Chung HT, Jeong SS, Park CK, Kim CY, Kim JE, et al: Hearing preservation after gamma knife stereotactic radiosurgery of vestibular schwannoma. Cancer 104: , Pollock BE, Driscoll CL, Foote RL, Link MJ, Gorman DA, Bauch CD, et al: Patient outcomes after vestibular schwannoma management: a prospective comparison of microsurgical resection and stereotactic radiosurgery. Neurosurgery 59:77 85, Sawamura Y, Shirato H, Sakamoto T, Aoyama H, Suzuki K, Onimaru R, et al: Management of vestibular schwannoma by fractionated stereotatic radiotherapy and associated cerebrospinal fluid malabsorption. J Neurosurg 99: , Szumacher E, Schwartz ML, Tsao M, Jaywant S, Franssen E, Wong CS, et al: Fractionated stereotactic radiotherapy for the treatment of vestibular schwannomas: combined experience of the Toronto-Sunnybrook Regional Cancer Centre and the Princess Margaret Hospital. Int J Radiat Oncol Biol Phys 53: , Varlotto JM, Shrieve DC, Alexander E III, Kooy HM, Black PM, Loeffler JS: Fractionated stereotactic radiotherapy for the treatment of acoustic neuromas: preliminary results. Int J Radiat Oncol Biol Phys 36: , Yamakami I, Uchino Y, Kobayashi E, Yamaura A: Conservative management, gamma-knife radiosurgery, and microsurgery for acoustic neurinomas: a systematic review of outcome and risk of three therapeutic options. Neurol Res 25: , 2003 Manuscript submitted June 8, Accepted December 5, Please include this information when citing this paper: published online March 13, 2009; DOI: / JNS Address correspondence to: Douglas Kondziolka, M.D., F.R.C.S.(C), University of Pittsburgh, Suite B-400, UPMC Presbyterian, 200 Lothrop Street, Pittsburgh, Pennsylvania kondziolkads@upmc.edu. 873