Local control of brain metastases by stereotactic radiosurgery in relation to dose to the tumor margin

J Neurosurg 104:907 912, 2006 Local control of brain metastases by stereotactic radiosurgery in relation to dose to the tumor margin MICHAEL A. VOGELBAUM, M.D., PH.D., LILYANA ANGELOV, M.D., SHIH-YUAN LEE, M.S.P.H., LIANG LI, PH.D., GENE H. BARNETT, M.D., AND JOHN H. SUH, M.D. Brain Tumor Institute and Departments of Biostatistics, Neurosurgery, and Radiation Oncology, Cleveland Clinic, Cleveland, Ohio Object. The maximal tolerated dose (MTD) for stereotactic radiosurgery (SRS) for brain tumors was established by the Radiation Therapy Oncology Group (RTOG) in protocol 90-05, which defined three dose groups based on the maximal tumor diameter. The goal in this retrospective study was to determine whether differences in doses to the margins of brain metastases affect the ability of SRS to achieve local control. Methods. Between 1997 and 2003, 202 patients harboring 375 tumors that met study entry criteria underwent SRS for treatment of one or multiple brain metastases. The median overall follow-up duration was 10.7 months (range 3 83 months). A dose of 24 Gy to the tumor margin had a significantly lower risk of local failure than 15 or 18 Gy (p = 0.0005; hazard ratio 0.277, confidence interval [CI] 0.134 0.573), whereas the 15- and 18-Gy groups were not significantly different from each other (p = 0.82) in this regard. The 1-year local control rate was 85% (95% CI 78 92%) in tumors treated with 24 Gy, compared with 49% (CI 30 68%) in tumors treated with 18 Gy and 45% (CI 23 67%) in tumors treated with 15 Gy. Overall patient survival was independent of dose to the tumor margin. Conclusions. Use of the RTOG 90-05 dosing scheme for brain metastases is associated with a variable local control rate. Tumors larger than 2 cm are less effectively controlled than smaller lesions, which can be safely treated with 24 Gy. Prospective evaluations of the relationship between dose to the tumor margin and local control should be performed to confirm these observations. KEY WORDS brain metastasis radiosurgery local control S TEREOTACTIC radiosurgery has become an important and widely used tool for the treatment of brain metastases. Multiple retrospective studies and one National Cancer Institute sponsored cooperative group prospective study have demonstrated either survival benefit, local control benefit, or both when SRS was added to WBRT in appropriately selected patients. 2,3,5,7,8,10,12,14 It has remained unclear, however, whether SRS is equivalent, superior, or inferior to resection, and there has been some debate about the utility of SRS for surgically treatable lesions. 3,4,13 This debate may be influenced, in part, by the nonuniform dosing strategies that have been the practice for SRS. We would expect that the efficacy of SRS would be linked to the prescribed dose, which is commonly set at an isodose line that completely covers the tumor to be treated. To date, however, there has been only one published report in which the efficacy of SRS has been examined as a function of the prescribed dose, 16 and this study included only Abbreviations used in this paper: ANOVA = analysis of variance; CI = confidence interval; HR = hazard ratio; KPS = Karnofsky Performance Scale; MTD = maximal tolerated dose; RTOG = Radiation Therapy Oncology Group; SRS = stereotactic radiosurgery; WBRT = whole-brain radiation therapy. J. Neurosurg. / Volume 104 / June, 2006 patients whose tumors were 2 cm or less in maximal diameter. The MTD for SRS in patients with brain tumors who have been previously treated with fractionated radiotherapy was established by the RTOG in protocol 90-05. 15 In the RTOG study, it was assumed that the MTD would depend on the maximal tumor diameter, and three groups of tumors were defined: Group 1, maximal diameter 20 mm or less; Group 2, 21 to 30 mm; and Group 3, 31 to 40 mm. The MTD for each group was as follows: Group 1, 24 Gy (highest dose tested, MTD not reached); Group 2, 18 Gy; and Group 3, 15 Gy. It would naturally follow that any evaluation of SRS treatment efficacy should be performed with the use of the MTDs that have been established for each of these groups. Nevertheless, with rare exceptions, this has not been the case. Since 1997, our group has routinely attempted to follow the RTOG dosing guidelines for treatment of brain metastases. We now report our finding, established in a series of 202 patients, that local control following SRS is dependent on the dose to the tumor margin. Clinical Material and Methods Between January 1997 and June 2003, 436 patients harboring 626 tumors underwent SRS at the Cleveland Clinic 907

M. A. Vogelbaum, et al. for treatment of one or multiple brain metastases. Of these, 202 were identified who met the study entry criteria. These criteria included use of the RTOG 90-05 radiosurgical dosing schedule (81 of the 436 patients did not meet this criterion) and a required minimum of 3 months of clinical follow up with at least one imaging study available for our review (131 of the remaining 355 patients did not meet this criterion). Reasons for deviation from the dosing scheme included proximity of the tumor to optic apparatus, proximity to other critical functional structures, or location in the brainstem. Twenty-two patients had brain metastases from small-cell carcinoma and were not included in this analysis. Hence, 375 tumors were treated in this subset of 202 patients. Patients could have either newly diagnosed brain metastases treated with SRS prior to or in lieu of WBRT, or residual or recurrent tumors treated after initial WBRT. Twenty-seven patients underwent a second or third radiosurgical procedure for newly discovered brain metastases; in no case was a brain metastasis that had been previously treated with SRS treated again with this therapy. Data for this study were extracted from a patient registry approved by the Cleveland Clinic Institutional Review Board. The following data were obtained from the medical records: patients age and sex, the histological type of the primary tumor, status of metastatic disease, status of primary tumor, KPS status, maximal lesion diameter and volume, number of metastases, use and timing of WBRT, and SRS treatment prescriptions. Two hundred two patients with 375 treated brain metastases were included in this study. The median overall follow-up duration (broken down by tumor) was 10.7 months (range 3 83 months). The patients clinical characteristics are listed in Table 1. All patients were treated with the Leksell Gamma Knife, model B or C (Elekta AB, Stockholm, Sweden). Treatment doses for each lesion were based on the dose criteria established in RTOG 90-05 and followed in RTOG 95-08. 2,15 In summary, the prescribed dose to the tumor margin was set according to each lesion s maximal diameter: for lesions 20 mm or less, the prescribed dose was 24 Gy; for 21- to 30- mm lesions, the dose was 18 Gy; and for 31- to 40-mm lesions, the dose was 15 Gy. In most cases, the prescribed dose was at or near the 50% isodose line (at the tumor margin). These doses were used regardless of whether patients had undergone previous WBRT. Treatment planning was performed by a neurosurgeon, who used MR imaging volumetric acquisitions or T 1 - weighted images with Gd contrast to define the lesion volume. A computerized tomography scan was also performed to reduce the chance of spatial errors created by warping of the MR images. All plans were reviewed by the radiation oncologist and medical physicist. All patients were treated with 6 mg of dexamethasone (or the equivalent dose of methylprednisolone) given intravenously at the time of SRS. Subsequent doses of steroid drugs were given as clinically indicated. The treating physicians included a total of six neurosurgeons and three radiation oncologists. The primary endpoint was local lesion control. We evaluated the relationship between dose to the tumor margin and growth control rate. Local progression was determined by review of clinical notes and reports on MR images. In a number of cases, additional studies (surgical biopsy procedures, [ 18 F]fluorodeoxyglucose positron emission tomography scans) had been performed to confirm the clinicoradiographic diagnosis, and these reports were reviewed. Twenty-seven patients underwent multiple sessions of radiosurgery due to distant brain progression, and these new tumors were included in the analysis (for a total of 240 treatment sessions). We also evaluated the relationship between dose to the tumor margin and survival. Survival duration was measured from the radiographically confirmed diagnosis of brain metastases until time of death or last patient contact. The date of death was confirmed with use of the Social Security Death Index. Statistical Analysis Patient and tumor characteristics were compared across three SRS dose groups by using the ANOVA F-test, Kruskal Wallis test, chi-square test, and the Fisher exact test, wherever appropriate. For univariate analysis, the Kaplan Meier method was used to summarize the time to local treatment failure/death for the three groups being compared. The log-rank test was used to test for difference in survival. Multivariate analysis with the Cox model was performed to account for the effect of covariates. Robust standard errors were used to adjust for possible correlations between multiple tumors in the same patient in the analysis of local failure. All tests were two-sided. The Type I error of tests was 0.05. The analysis was performed using SAS version 9.0 software (SAS Institute, Cary, NC) and S-PLUS version 6.2 (Insightful Corp., Seattle, WA). Results Patient and Tumor Characteristics The majority of patients were younger than 65 years of age and had a KPS score greater than 70. In more than half, the primary disease was controlled but the patients had active systemic metastases. Most had more than one brain metastasis, and 18% had more than three. The majority of metastases were from non small cell lung cancer. Forty-eight patients (24%) were treated with SRS alone, 117 (58%) were treated with WBRT prior to SRS, and 37 (18%) were treated with WBRT following SRS. Treatment Variables The treatment variables, sorted by prescribed dose to the tumor margin, are shown in Table 2. There were no statistically significant relationships between treatment group and sex, KPS score, status of primary disease, or timing of WBRT. Patients in the group receiving the 24-Gy dose were slightly younger than those in the 15- or 18-Gy treatment groups. A significantly greater number of patients in the 24- Gy group had extracranial metastasis or multiple brain metastases. These patients also had more tumors located in the supratentorial compartment. A larger number of tumors in the 24-Gy treatment group were melanomas; more metastases in the 18-Gy group were from a lung cancer primary lesion. Kaplan Meier Analysis of Time to Local Failure Figure 1 shows the Kaplan Meier curves of time to local treatment failure, stratified by SRS dose. Comparisons among the three groups show that a dose of 24 Gy to the tumor margin had a significantly lower hazard (longer time 908 J. Neurosurg. / Volume 104 / June, 2006

Dose-related efficacy of stereotactic radiosurgery TABLE 1 Demographic and tumor characteristics in 202 patients with 375 metastases* Characteristic No. of Patients (%) sex male 87 (43) female 115 (57) age (yrs) median 57 range 25 85 KPS Score 50 1 (0.5) 60 9 (4.5) 70 34 (16.8) 80 45 (22.3) 90 93 (46.0) 100 20 (9.9) primary tumor controlled 131 (65) systemic metastases 109 (54) RPA Class I 44 (22) II 148 (73) III 10 (5) primary tumor by no. of patients breast 37 (18) GI 8 (4) lung 87 (43) renal 27 (13) melanoma 22 (11) other 10 (5) unknown 11 (5) by no. of tumors breast 70 (19) GI 8 (2) lung 166 (44) renal 42 (11) melanoma 54 (14) other 16 (4) unknown 19 (5) no. of metastases 1 83 (41) 2 50 (25) 3 30 (15) 4 15 (7) 5 13 (6) 6 9 11 (5) WBRT timing by no. of patients ND 48 (24) pre-srs 117 (58) post-srs 37 (18) by no. of tumors ND 72 (19) pre-srs 240 (64) post-srs 63 (17) * There were 375 metastases in 202 patients; percentages were calculated based on 202 patients for every characteristic except primary tumor and WBRT timing, as indicated. Abbreviations: GI = gastrointestinal; ND = not done; RPA = recursive partitioning analysis. to local failure) than the 15- and 18-Gy groups (p = 0.0005; HR 0.277, CI 0.134 0.573), whereas the 15- and 18-Gy groups were not significantly different from each other (p = 0.82; HR for 18 Gy compared with 15 Gy, 0.925; CI 0.476 1.8). Other covariates, such as age, KPS score, site of primary cancer, lesion location (supratentorial compared with infratentorial), the total number of brain metastases, whether the primary disease was under control, the presence of extracranial disease, and the timing of WBRT, were entered into a Cox model together with SRS dose, and stepwise and backward model selection procedures were applied, but J. Neurosurg. / Volume 104 / June, 2006 FIG. 1. Kaplan Meier curves of time to local treatment failure, stratified by SRS dose. A tumor margin dose of 24 Gy had a significantly longer time to local failure than in the 15- and 18-Gy groups (p = 0.0005; HR 0.277, CI 0.134 0.573). none of these covariates were significantly associated with time to local failure. Table 3 shows the estimated proportions of patients with time to local failure exceeding 3, 6, 9, and 12 months. The median duration of local control for lesions treated with 24 Gy was not reached; for 18 Gy it was 11.57 months (95% CI 8.6 19.37) and for 15 Gy it was 11.83 months (95% CI 6.03 41.9) (p 0.0001 by log-rank test for 24 Gy compared with 18 or 15 Gy). The 1-year local control rate was 85% (95% CI 78 92%) in tumors treated with 24 Gy, compared with 49% (30 68%) in tumors treated with 18 Gy, and 45% (23 67%) in tumors treated with 15 Gy. Kaplan Meier Analysis of Survival Time, Based on Minimal Dose for Each Patient In this analysis, if a patient had multiple lesions, the dose was set to correspond to the smallest one used for SRS. The analysis was similar to that for time to local control. No statistically significant difference was found among the three treatment groups (p = 0.9364; Fig. 2). Two covariates were found to be significantly associated with survival. They were 1) total number of lesions (p = 0.0005; HR for each incremental number, 1.165; CI 1.069 1.269) and 2) systemic disease (p = 0.0062; HR for presence compared with absence, 1.551; CI 1.133 2.123). Kaplan Meier Analysis of Survival Time, Based on MTD for Each Patient In this analysis, if a patient had multiple lesions, the dose was set to correspond to the largest one used for SRS. The analysis was similar to that for time to local control. No statistically significant difference was found among the treatment groups (p = 0.5362; Fig. 3). Two covariates were found to be significantly associated with survival. They were 1) total number of lesions (p 0.0001; HR for each incremental number, 1.229; CI 1.130 1.337) and 2) primary tumor controlled (p = 0.0235; HR for controlled compared with not controlled, 0.680; CI 0.488 0.95). Systemic disease was borderline significant at the 0.05 level (p = 0.0548; HR of presence compared with absence, 1.368; CI 0.994 1.884). Discussion Stereotactic radiosurgery is an important tool for the treat- 909

M. A. Vogelbaum, et al. TABLE 2 Treatment variables for SRS sorted by prescribed dose to tumor margin in 375 metastases* Dose & No. of Tumors (%) Total No. Variable of Lesions 15 Gy 18 Gy 24 Gy p Value no. of tumors 375 41 85 249 sex of patients female 223 26 (63) 41 (48) 156 (63) male 152 15 (37) 44 (52) 93 (37) 0.0561; chi-square test age in yrs mean SD 58 13 59 11 55 12 median 57 59 54 range 32 84 39 85 25 85 0.0562; ANOVA F-test no. of patients w/ KPS score of 50 0 (0) 1 (1) 0 (0) 60 2 (5) 7 (8) 7 (3) 70 12 (29) 12 (14) 34 (14) 80 7 (17) 19 (22) 66 (27) 90 17 (41) 42 (49) 115 (46) 100 3 (7) 4 (5) 27 (11) 0.1322; Kruskal Wallis test primary lesions controlled no 126 15 (37) 36 (42) 75 (30) yes 249 26 (63) 49 (58) 174 (70) 0.1085; Fisher exact test ECD no 165 21 (51) 54 (64) 90 (36) yes 210 20 (49) 31 (36) 159 (64) 0.0001; chi-square test total no. of lesions 1 22 (54) 34 (40) 33 (13) 2 9 (22) 17 (20) 58 (23) 3 5 (12) 12 (14) 42 (17) 4 2 (5) 7 (8) 38 (15) 5 3 (7) 6 (7) 40 (16) 6 9 0 (0) 9 (11) 38 (15) 0.0001; Kruskal Wallis test tumor location infratentorial 57 9 (22) 16 (19) 32 (13) supratentorial 318 32 (78) 69 (81) 217 (87) 0.1844; chi-square test primary lesion breast 70 9 (22) 13 (15) 48 (19) GI 8 1 (2) 2 (2) 5 (2) lung 166 17 (41) 44 (52) 105 (42) melanoma 54 4 (10) 4 (5) 46 (18) other 16 3 (7) 4 (5) 9 (4) renal 42 5 (12) 9 (11) 28 (11) unknown 19 2 (5) 9 (11) 8 (3) 0.0806; chi-square test WBRT timing ND 72 13 (32) 18 (21) 41 (16) pre-gks 240 25 (61) 50 (59) 165 (66) post-gks 63 3 (7) 17 (20) 43 (17) 0.0975; chi-square test median max tumor diam (cm) 3.3 2.4 1 range 2.9 4.5 2 3 0.1 2 0.0001; ANOVA F-test * Diam = diameter; ECD = extracranial disease; GKS = gamma knife surgery; SD = standard deviation. ment of patients with brain metastases. It is of particular utility for the treatment of patients with unresectable brain metastases, in whom the use of SRS after WBRT is associated with improved local control, and, in the case of single brain metastases, improved survival when compared with WBRT alone. 2 In patients who have brain metastases that can be surgically removed, however, few studies have been conducted to address the question of whether SRS or surgery is the better treatment modality. In one published study, 4 62 patients who underwent surgery for removal of brain metastases were retrospectively matched with 31 patients who underwent SRS. The authors reported an improvement in the median survival duration for patients treated with surgery or radiosurgery (16.4 compared with 7.5 months). They believed that this difference was due to a higher local recurrence rate in the patients treated with SRS. It is important to note that in this report the range of tumor volumes treated with SRS was 0.41 to 8.25 ml (median 1.96 ml) and that the range of prescribed doses to the tumor margin was 17 to 22 Gy (median 18.7 Gy). By extrapolation it would follow that the majority of tumors were less than 2 cm in maximal diameter. Hence, the fact that the highest prescribed dose was 22 Gy and that the median was less than 20 Gy strongly suggests that the SRS dosing scheme used in this study was well below the MTD reported in RTOG 90-05. 15 Bindal, et al., 4 noted that their dosing regimen was similar to that used in many other studies. 1,6,7,9,11,12,18 20 We report in this retrospective analysis of our single-center SRS experience that the efficacy of SRS, specifically as it relates to local control, is significantly dependent on the prescribed dose to the tumor margin. We found that tumors treated with a dose of 15 or 18 Gy to the tumor margin had an approximately threefold increase in risk of local failure compared with tumors treated with 24 Gy to the margin. Local control at 1 year was 85% for tumors treated with a 910 J. Neurosurg. / Volume 104 / June, 2006

Dose-related efficacy of stereotactic radiosurgery TABLE 3 Estimated percentage of patients in whom time to local failure exceeded 3, 6, 9, and 12 months* Dose & Local Control Rate (95% CI) Factor 15 Gy 18 Gy 24 Gy total no. of lesions 41 85 249 follow-up interval 3 mos 100% 99% (96 100%) 100% no. of lesions at risk 31 56 166 6 mos 71% (54 88%) 87% (77 96%) 92% (87 97%) no. of lesions at risk 18 37 92 9 mos 63% (44 81%) 64% (49 80%) 85% (78 92%) no. of lesions at risk 13 18 60 12 mos 45% (23 67%) 49% (30 68%) 85% (78 92%) no. of lesions at risk 6 8 37 * Metastases are categorized according to prescribed SRS dose. The probability value for all follow-up intervals was less than 0.0001 and was calculated using the log-rank test. These values represent the number of lesions that were controlled and were still being followed with MR imaging or computerized tomography studies at each interval (lesions at risk of further spread). prescribed dose of 24 Gy, compared with 49 and 45% for tumors treated with 18 or 15 Gy, respectively. These findings are remarkable, especially in light of a slightly higher preponderance of melanomas, which tend to be radioresistant, in the 24-Gy dose group. It is difficult to compare our results directly with most of the earlier series because, with rare exceptions, the control rates were not reported as a function of tumor diameter or prescribed dose. In one study of radiosurgery in which the maximal tumor diameter was limited to 2 cm or less, 1- year local control rates of 96% were reported for SRS plus WBRT and 77% for SRS alone. 16 Our local control rate of 85%, which includes both patients who received SRS alone and those who underwent SRS plus WBRT for similar-sized lesions appears to show a similar control rate. We are not aware of other studies in which the local control rates in tumors larger than 2 cm were specifically examined. The results reported here support the suggestion that analyses of local control rates post-srs that lump together tumors that are larger and smaller than 2 cm may yield unrealistically high (for large tumors) and low (for small tumors) estimates of the efficacy of this therapy. In the RTOG 90-05 protocol, the acute ( 3 months) and chronic ( 3 months) risks of SRS were determined and stratified by lesion size. Notably, SRS was performed following the completion of fractionated radiotherapy regimens in these patients. For tumors larger than 30 mm, there were no cases of acute toxicity and a 14% incidence of chronic toxicity in the 15-Gy treatment group. This dose was selected as the MTD after a 50% risk of acute and chronic toxicity was identified for the 18-Gy treatment group. For tumors between 21 and 30 mm, there were no cases of acute toxicity and a 20% incidence of chronic toxicity in the 18-Gy treatment group. This dose was selected as the MTD after a 38% risk of acute and chronic toxicity was identified for the 21-Gy treatment group. These findings indicate that the prescribed doses for metastatic tumors larger than 2 cm cannot be increased safely and that resection remains an important tool for the management of these lesions. It must be noted, however, that the J. Neurosurg. / Volume 104 / June, 2006 FIG. 2. Kaplan Meier analysis of survival time, using minimal dose for each patient. In this analysis, if a patient had multiple lesions, the dose was set to correspond to the smallest one used for SRS. The analysis was similar to that for time to local control. No statistically significant difference was found among the treatment groups (p = 0.9364). RTOG 90-05 protocol included tumors of various histological types (both primary and metastatic brain tumors) that had been previously treated with a variety of fractionated radiotherapy regimens and that the true MTD for radiationnaïve metastases may be higher. The MTD for tumors that had a maximal diameter of 2 cm was not actually reached in the RTOG 90-05 protocol. The investigators chose to end the study after reaching a prescribed dose of 24 Gy, despite finding no cases of acute and only a 10% incidence of chronic toxicity, which was below their definition of dose-limiting toxicity. Notably, Shehata, et al., 16 observed even lower rates of Grade 3 or 4 neurotoxicity for tumors treated with doses of less than 20 Gy in combination with WBRT. We did not observe an effect of the timing of fractionated radiotherapy on the local control rate observed after treatment with SRS. This finding is consistent with other published reports. 17 A more definitive evaluation of the role of WBRT in patients treated with SRS was initiated in a prospective randomized study (American College of Surgeons Oncology Group Z0300). The results and conclusions of this study are limited by the fact that it is a retrospective analysis of the experience at a single institution. Further prospective evaluation of the FIG. 3. Kaplan Meier analysis of survival time, using MTD for each patient. In this analysis, if a patient had multiple lesions, the dose was set to correspond to the largest one used for SRS. The analysis was similar to that for time to local control. No statistically significant difference was found among the treatment groups (p = 0.5362). 911

M. A. Vogelbaum, et al. dose-dependent efficacy of SRS is necessary to confirm our findings, and additional studies should be combined with a dose-escalation strategy for certain situations (for example, radioresistant histological types or unresectable tumors). One difficulty with assessment of local control after radiosurgery is the similarity in the appearance of tumor progression (treatment failure) to that of radiation necrosis (treatment toxicity, but not failure) on standard MR imaging. We found that six cases of clinically suspected tumor progression in our series were subsequently revealed to be radiation necrosis on [ 18 F]fluorodeoxyglucose positron emission tomography scanning or biopsy sampling. Improved imaging techniques will be necessary to resolve this important distinction, and we believe that this technical limitation serves as the most significant barrier to a prospective evaluation of the relationship between radiation dose to the tumor margin and local control. Conclusions The findings in this study have relevance to the management of metastatic brain tumors in the clinic. When considering the treatment options for a patient with newly diagnosed brain metastases, clinicians are faced with the question of whether surgery or radiosurgery would be the best management strategy. Although radiosurgical treatment has the advantages of less morbidity and patient discomfort as well as no need for hospitalization, based on the results of this study we would suggest that its efficacy may not be durable for lesions larger than 2 cm due to a restriction in the tumor-margin dose that can be used safely. Although many other factors, including tumor location, extent and status of systemic disease, and medical comorbidities must be taken into account when selecting the most appropriate management strategy for these patients, based on our results we would suggest that although radiosurgery may be a preferable option for metastases less than 2 cm in maximal diameter, it may not be the best strategy for many tumors larger than 2 cm. Acknowledgment We thank Martha Tobin for her editorial assistance. References 1. Adler JR, Cox RS, Kaplan I, Martin DP: Stereotactic radiosurgical treatment of brain metastases. J Neurosurg 76:444 449, 1992 2. Andrews DW, Scott CB, Sperduto PW, Flanders AE, Gaspar LE, Schell MC, et al: Whole brain radiation therapy with or without stereotactic radiosurgery boost for patients with one to three brain metastases: phase III results of the RTOG 9508 randomised trial. Lancet 363:1665 1672, 2004 3. Auchter RM, Lamond JP, Alexander E, Buatti JM, Chappell R, Friedman WA, et al: A multiinstitutional outcome and prognostic factor analysis of radiosurgery for resectable single brain metastasis. Int J Radiat Oncol Biol Phys 35:27 35, 1996 4. Bindal AK, Bindal RK, Hess KR, Shiu A, Hassenbusch SJ, Shi, WM, et al: Surgery versus radiosurgery in the treatment of brain metastasis. J Neurosurg 84:748 754, 1996 5. Chidel MA, Suh JH, Greskovich JF, Kupelian PA, Barnett G H: Treatment outcome for patients with primary nonsmall-cell lung cancer and synchronous brain metastasis. Radiat Oncol Investig 7:313 319, 1999 6. Engenhart R, Kimmig BN, Hover KH, Wowra B, Romahn J, Lorenz WJ, et al: Long-term follow-up for brain metastases treated by percutaneous stereotactic single high-dose irradiation. Cancer 71: 1353 1361, 1993 7. Flickinger JC, Kondziolka D, Lunsford LD, Coffey RJ, Goodman ML, Shaw EG, et al: A multi-institutional experience with stereotactic radiosurgery for solitary brain metastasis. Int J Radiat Oncol Biol Phys 28:797 802, 1994 8. Fuller BG, Kaplan ID, Adler J, Cox RS, Bagshaw MA: Stereotaxic radiosurgery for brain metastases: the importance of adjuvant whole brain irradiation. Int J Radiat Oncol Biol Phys 23: 413 418, 1992 9. Kihlstrom L, Karlsson B, Lindquist C, Noren G, Rahn T: Gamma knife surgery for cerebral metastasis. Acta Neurochir Suppl 52:87 89, 1991 10. Kondziolka D, Patel A, Lunsford LD, Kassam A, Flickinger JC: Stereotactic radiosurgery plus whole brain radiotherapy versus radiotherapy alone for patients with multiple brain metastases. Int J Radiat Oncol Biol Phys 45:427 434, 1999 11. Loeffler JS, Alexander E III: Radiosurgery in the definitive management of CNS malignancies. Front Radiat Ther Oncol 27: 227 244, 1993 12. Mehta MP, Rozental JM, Levin AB, Mackie TR, Kubsad SS, Gehring MA, et al: Defining the role of radiosurgery in the management of brain metastases. Int J Radiat Oncol Biol Phys 24: 619 625, 1992 13. O Neill BP, Iturria NJ, Link MJ, Pollock BE, Ballman KV, O Fallon JR: A comparison of surgical resection and stereotactic radiosurgery in the treatment of solitary brain metastases. Int J Radiat Oncol Biol Phys 55:1169 1176, 2003 14. Sanghavi SN, Miranpuri SS, Chappell R, Buatti JM, Sneed PK, Suh JH, et al: Radiosurgery for patients with brain metastases: a multi-institutional analysis, stratified by the RTOG recursive partitioning analysis method. Int J Radiat Oncol Biol Phys 51: 426 434, 2001 15. Shaw E, Scott C, Souhami L, Dinapoli R, Kline R, Loeffler J, et al: Single dose radiosurgical treatment of recurrent previously irradiated primary brain tumors and brain metastases: final report of RTOG protocol 90 05. Int J Radiat Oncol Biol Phys 47: 291 298, 2000 16. Shehata MK, Young B, Reid B, Patchell RA, St Clair W, Sims J, et al: Stereotatic radiosurgery of 468 brain metastases 2 cm: implications for SRS dose and whole brain radiation therapy. Int J Radiat Oncol Biol Phys 59:87 93, 2004 17. Sneed PK, Suh JH, Goetsch SJ, Sanghavi SN, Chappell R, Buatti JM, et al: A multi-institutional review of radiosurgery alone vs. radiosurgery with whole brain radiotherapy as the initial management of brain metastases. Int J Radiat Oncol Biol Phys 53: 519 526, 2002 18. Somaza S, Kondziolka D, Lunsford LD, Kirkwood JM, Flickinger JC: Stereotactic radiosurgery for cerebral metastatic melanoma. J Neurosurg 79:661 666, 1993 19. Sturm V, Kober B, Hover KH, Schlegel W, Boesecke R, Pastyr O, et al: Stereotactic percutaneous single dose irradiation of brain metastases with a linear accelerator. Int J Radiat Oncol Biol Phys 13:279 282, 1987 20. Valentino V, Mirri MA, Schinaia G, Dalle Ore G: Linear accelerator and Greitz-Bergstrom s head fixation system in radiosurgery of single cerebral metastases. A report of 86 cases. Acta Neurochir (Wien) 121:140 145, 1993 Manuscript received May 25, 2005. Accepted in final form January 18, 2006. Address reprint requests to: Michael A. Vogelbaum, M.D., Ph.D., Brain Tumor Institute and Department of Neurosurgery, ND40, Cleveland Clinic Foundation, 9500 Euclid Avenue, Cleveland, Ohio 44195. email: vogelbm@neus.ccf.org. 912 J. Neurosurg. / Volume 104 / June, 2006