Gamma Knife surgery for the treatment of melanoma metastases: the effect of intratumoral hemorrhage on survival

Transcription

1 J Neurosurg J Neurosurg (Suppl) 109: , 109:99 105, 2008 Gamma Knife surgery for the treatment of melanoma metastases: the effect of intratumoral hemorrhage on survival And y J. Re d m o n d, M.D., 1 Mi c h a e l L. DiLu n a, M.D., 1 Rya n He b e rt, B.S., 1 Je n n i f e r A. Mo l i t e r n o, M.D., 1 Ra n i De sa i, Ph.D., M.P.H., 4 Jo n at h a n P. S. Kn i s e ly, M.D., 2,3 a n d Ve r o n i c a L. Ch i a n g, M.D., B.Me d.sc i. 1,4 Departments of 1 Neurosurgery and 2 Therapeutic Radiology, and 3 Yale Cancer Center, Yale University School of Medicine, New Haven; and 4 West Haven Veterans Administration Hospital, West Haven, Connecticut Object. Gamma Knife surgery (GKS) improves overall survival in patients with malignant melanoma metastatic to the brain. In this study the authors investigated which patient- or treatment-specific factors influence survival of patients with melanoma brain metastases; they pay particular interest to pre- and post-gks hemorrhage. Methods. Demographic, treatment, and survival data on 59 patients with a total of 208 intracranial metastases who underwent GKS between 1998 and 2007 were abstracted from treatment records and from the Connecticut Tumor Registry. Multivariate analysis was used to identify factors that independently affected survival. Results. Survival was significantly better in patients with solitary metastasis (p = 0.04), lesions without evidence of pre-gks hemorrhage (p = 0.004), and in patients with total tumor volume treated < 4 cm 3 (p = 0.02). Intratumoral bleeding occurred in 23.7% of patients pre-gks. Intratumoral bleeding occurred at a mean of 1.8 months post-gks at a rate of 15.2%. Unlike the marked effect of pretreatment bleeding, posttreatment bleeding did not independently affect survival. Sex, systemic control, race, metastases location, whole-brain radiation therapy, chemotherapy, history of antithrombotic medications, and cranial surgery had no independent association with survival. Conclusions. These data corroborate previous findings that tumor burden (either as increased number or total volume of lesions) at the time of GKS is associated with diminished patient survival in those with intracerebral melanoma metastases. Patients who were noted to have hemorrhagic melanoma metastases prior to GKS appear to have a worse prognosis following GKS compared with patients with nonhemorrhagic metastases, despite similar rates of bleeding pre- and post-gks treatment. Gamma Knife surgery itself does not appear to increase the rate of hemorrhage. (DOI: /JNS/2008/109/12/S16) Ke y Wo r d s Gamma Knife hemorrhage melanoma metastasis outcome stereotactic radiosurgery survival Ma l i g n a n t melanoma is the third most common tumor to metastasize to the brain, following lung and breast cancer. Survival is poor, averaging months, and many patients die of their intracerebral tumor burden. 13 In addition to mass effect, the tendency of melanoma metastatic to the brain to hemorrhage also contributes to the morbidity and mortality of many patients. 20 Treatment options are mainly palliative, Abbreviations used in this paper: GKS = Gamma Knife surgery; SRS = stereotactic radiosurgery; WBRT = whole-brain radiation therapy. and historically have involved surgery plus WBRT or WBRT alone. 25,26 Although once considered to be radioresistant, melanoma metastases can now be effectively controlled with SRS, 1,2,12 extending survival by months. 3 8,10,12 19,21 23,25,27 The propensity of these lesions to hemorrhage, however, appears to limit tumor control after SRS, 13 and some studies show hemorrhage rates of % after SRS treatment. To better understand the relationship between hemorrhagic events and other factors that affect survival in this population of patients, we reviewed the outcomes of patients treated with GKS for melanoma brain metastases at our institution, focusing particularly on the question of whether GKS induces hemorrhage. 99

2 A. J. Redmond et al. Study Population Methods We performed an institutional review board approved retrospective review of medical records at Yale-New Haven Hospital, Yale Cancer Center, and the Yale-New Haven Gamma Knife Center for all patients who underwent their first GKS treatment for intracranial melanoma metastases between July 1, 1998, and December 31, We extracted data on demographics (age, sex, and race), primary disease (primary location, systemic control at the time of GKS), imaging characteristics (number, volume, location of cerebral metastases at the time of GKS, and evidence of intratumoral hemorrhage on imaging before or after GKS), mean GKS radiation dose, WBRT (including the timing of its use in relation to the radiosurgical procedure), chemotherapy, cranial surgery for tumor mass or for hemorrhage (pre- or post-gks), use of anticoagulation or antiplatelet therapy (antithrombotic), and date of last follow-up. Antithrombotic medications were routinely discontinued prior to each procedure except in situations in which medication cessation could result in a life-threatening situation. Following treatment, patients were followed up with imaging studies every 6 weeks for the first 6 months and then every 3 months thereafter. Systemic control was defined as no detectable spread of primary tumor to another organ system other than the central nervous system; involvement of local lymph nodes did not constitute, ipso facto, a lack of systemic control (Table 1). To obtain dates of all patient deaths that had occurred by December 31, 2007, we accessed the Connecticut Tumor Registry and Web-accessible Social Security Administration Master Death Files. Data Analysis Means, standard deviations, and medians were calculated for categorical variables. The 10th, 25th, 75th, and 90th percentiles were calculated for continuous variables. Kaplan Meier plots, the log-rank test, and univariate and multivariate Cox regression analyses were used to examine the correlations between groups, individual variables, and survival. We used survival analysis techniques to examine post-gks survival and adjusted them for patients who were lost to follow-up or still alive at the time of data analysis. In addition, we explored a variety of cutoff points for categorizing continuous variables, and we tested various clusters of nominal and ordinal variable categories. The predictor variables that were examined included age at the time of GKS, sex, race, pathological findings, systemic control, location of metastases, number of metastases, total volume of metastases, WBRT, chemotherapy, cranial surgery, location of primary melanoma, use of anticoagulation or antiplatelet therapy, evidence of prior hemorrhage on imaging at the time of GKS, and hemorrhage after GKS. For all analyses, probability values < 0.05 were considered significant, and probability values ranging from 0.10 to 0.05 were considered trends. Statistical survival analysis was performed and plots were made using a variety of packages within R 2.6 for Windows ( In a specific effort to evaluate the relationship between the radiation dose per unit volume of tumor treated (dose inhomogeneity) and the post-gks rate of hemorrhage, a new continuous variable, dose/volume ratio, was created. Multiple Student t-tests were then used to compare each percentile dose/volume ratio for each lesion between those individuals who experienced hemorrhage after GKS and those who did not. Statistical analysis was performed using Microsoft Excel (2003) and confirmed with statistical software (version 9.1, SAS, Inc.). Study Population Results Fifty-nine patients with a total of 208 lesions were treated between 1996 and The median patient age was 58.5 years (range years) (Fig. 1a). The majority of the patients were Caucasian (58 patients [98%]) and male (44 patients [75%]). All patients undergoing GKS had Karnofsky Performance Scale scores 70 (Table 1). Follow-up data were available for a median of 152 days (range 19 days 5.75 years) and a total of 19,384 patient days from the time of initial GKS. The median number of metastases treated was 2 (range 1 10) (Fig. 1b). Ninety percent of the patients (53 of 59) had supratentorial lesions. Eighty-eight percent of the patients underwent chemotherapy (52 patients), 46% had WBRT (27 patients), 31% had craniotomy (18 patients), and 54% of patients had total volumes treated that were < 4 cm 3 (32 patients) (Table 1). Gamma Knife Surgery All GKS procedures were performed using a model C Leksell Gamma Knife (Elekta Instruments). Doses of Gy to the edge of the tumor were prescribed to all lesions. The tumor edge isodose lines ranged from 40 to 70%. Survival Fifty-six of the 59 patients had died by the conclusion of the study. The median survival after radiosurgery was 4.4 months (range months). The median survival of patients with a single metastasis that was treated using GKS was 6.1 months compared with 4.6 months for patients with multiple metastases (p = 0.04) (Table 2). The total volume of tumor treated was also associated with survival. Those with tumor volumes exceeding 4 cm 3 survived 3.3 months compared with 6.48 months in those who had tumor volume < 4 cm 3 (Table 2). In the group that had no evidence of intratumoral hemorrhage pre-gks the median survival was 6.8 months, whereas those who had previous evidence of hemorrhage had a reduced survival of 2.1 months (Fig. 2, 3a f). In multivariate Cox regression analysis, prior hemorrhage (HR 2.43, p = 0.005), 2 metastases (HR 1.91, p = 0.04), total volume treated exceeding 4 cm 3 (HR 1.89, p = 0.02), and fewer number of GKS sessions (HR 0.61, p = 0.05) were associated with decreased survival (Table 3). Race, sex, systemic control, WBRT, systemic control, chemotherapy, antithrombotic medications, tumor location, primary tumor 100

3 Stereotactic radiosurgery for melanoma metastasis TABLE 1 Study population and treatment characteristics in 59 patients* Treatment No. of Patients (%) p Value Fig. 1. Bar graphs demonstrating the distribution of ages (a), and the number of metastases (b) in our patient cohort (59 patients) with central nervous system melanoma metastases. location, hemorrhage after GKS, and cranial surgery had no independent association with altered survival (Table 1). Intratumoral Hemorrhage Prior to GKS, 14 (23.7%) of 59 patients had radiographic evidence of hemorrhage seen within at least 1 of their intracerebral metastatic tumors. Of the 208 lesions studied, 18 (8.7%) showed radiographic evidence of pre- GKS hemorrhage. Following radiosurgery, 9 (15.3%) of the 59 patients and 9 (4.3%) of 208 tumors developed new bleeding into their GKS-treated metastatic lesions. None of the previously treated lesions showed evidence of hemorrhage during follow-up. The median time to hemorrhage was 1.8 months (range months). The pre-gks and post-gks bleeding rates of 23.7 and 15.2%, respectively, were not found to be statistically significantly different (p = 0.89). Patients with hemorrhage after GKS survived, on average, 86 days compared with 166 days for those without hemorrhage; however, unlike the marked effect of pretreatment bleeding, post-gks bleeding was not found to be statistically significant with respect to survival (HR 1.55, p = 0.2). The presence of pre-gks hemorrhage was not associated with an increased risk of post-gks hemorrhage (p = 0.9) (Table 3). Individual tumor size or total tumor volume, prior hemorrhage, history of antiplatelet or anticoagulation medications, or location did not have any statistically significant effect on intratumoral bleeding rates post-gks. Similarly no significant effect was found between radiosurgery dose/ volume ratio and the rate of bleeding post-gks. Discussion Few studies have addressed the effect of intratumoral hemorrhage on patient outcome following GKS. 12,23 The age (yrs) 0.44 median (IQ range) 58.5 ( ) mean 58.4 range sex 0.03 M 44 (75) F 16 (25) race 0.74 Caucasian 58 (98) Hispanic 1 (2) systemic control yes 8 (14) no 51 (86) no. of GKS treatments (75) 2 11 (19) 3 4 (7) WBRT pre-gks 24 (41) 0.94 post-gks 5 (8) 0.56 ever 27 (46) 0.92 chemotherapy 52 (88) 0.97 craniotomy 18 (31) 0.71 total vol treated (cm 3 ) 0.14 <4 32 (54) 0.04 >4 27 (46) no. of cerebral metastases 0.07 mean 2.7 median 2 mode 1 range 1 10 total 208 solitary metastasis 19 (32) metastases 40 (68) location of metastases supratentorial 53 (90) 0.14 infratentorial 17 (29) 0.99 deep 8 (14) 0.96 * IQ = interquartile. Unless specified otherwise. natural history and biology of melanoma metastases to the brain pre- or post-gks are poorly understood, but may be inferred from retrospective studies. Kondziolka et al. 11 reported that up to 50% of melanoma metastases hemorrhaged in a cohort of patients who were not treated with GKS, and suggested that prior intratumoral hemorrhage may increase the propensity of tumors to subsequently bleed. In patients with melanoma who undergo GKS, the hemorrhage rate prior to treatment has been reported at rates of 9 24% (Table 4). Intratumoral hemorrhage after GKS appears to occur at published rates of %, with variable median times to hemorrhage of months. In this study the median time to hemorrhage was 1.8 months with a post-gks hemorrhage rate of 15.2% consistent with prior reports. This study further found that that the bleeding rates pre- and post-gks are not significantly different, although the published hemorrhage rates as well as the hemorrhage rate in this study post-gks do seem lower numerically than pre-gks. Suzuki et al. 24 postulated that dose inhomogeneity may promote GKS-related intratumoral bleeding through 101

4 A. J. Redmond et al. TABLE 2 Stratified Cox survival estimates* Cox Regression Analysis Variable No. (%) Survival (mos) p Value no. of mets (32) (68) 4.6 vol treated (cm 3 ) 0.02 < 32 (54) 6.48 >4 27 (46) 3.3 systemic control 0.10 yes 8 (14) 6.9 no 51 (86) 4.4 no. of sessions (75) (19) (6) 11.4 hemorrhage pre-gks no 76% 6.8 yes 24% 2.1 * mets = metastases. Stratified survival estimates and probability values were based on final multivariate Cox regression model. Overall p < for the final multivariate Cox regression model. a process whereby acute tumor necrosis precedes vascular obliteration. To examine this theory, we compared lesional dose/volume ratios and found that they had no effect on bleeding rates. Mathieu et al. 13 found that increased radiosurgery volume, not radiosurgery dose or previous irradiation, predicted hemorrhage in melanoma metastases post-srs. Tumor size, total intracranial tumor volume, or history of antithrombotic medication use was not found to affect intratumoral bleeding rates in our study. However, prior intratumoral hemorrhage predicted a worse survival of 2.1 months compared with 6.8 months in those patients who did not experience hemorrhage. Bleeding after GKS, however, did not have the same dramatic effect on survival and was not predicted by pre-gks bleeding. Our findings would lead us to conclude that GKS, regardless of dose and volume, does not contribute to bleeding in intracerebral melanoma metastases. Examination of the numbers alone (pre-gks bleed rate of 23.7 vs 15.2% post- GKS) would suggest that GKS may in fact be protective from hemorrhage, but we were unable to demonstrate this given the low power of this study. Based on these data, it is not evident why patients with evidence of intratumoral hemorrhage on imaging have reduced survival after GKS. Local treatment failure has been proposed as the cause of reduced survival in patients with hemorrhagic metastases. Mathieu et al. 13 demonstrated that hemorrhagic metastases and increased total volume treated were predictors of local treatment failure. This may partly be due to the difficulty associated with creating a treatment plan for poorly delineated lesions containing blood products. In this series no significant evidence of local failure was demonstrated, although increased total volume of tumor treated was associated with poorer survival. Indeed, this may be due to our sample size. Fig. 2. Boxplots (from right to left, clockwise) demonstrating survival when plotted against different variables. Error bars and range of values are shown. 102

5 Stereotactic radiosurgery for melanoma metastasis Fig. 3. Kaplan Meier survival plots of overall survival after SRS for all 59 patients with melanoma metastases showing survival curves stratified for different variables. GKRS = GKS. 103

6 A. J. Redmond et al. TABLE 3 Stratified Cox survival analysis* Cox Regression Analysis Variable HR (95% CI) p Value 2 mets 1.91 ( ) 0.04 >4 cm 3 treated 1.89 ( ) 0.02 no systemic control 0.51 ( ) 0.11 no. of sessions 0.61 ( ) 0.05 hemorrhage pre-grs 2.43 ( ) hemorrhage post-gks 1.55 ( ) 0.20 * Hazard ratios and probability values were derived from the final multivariate Cox regression model. Abbreviation: CI = confidence interval. Hazard ratios < 1 indicate a decreased risk of death during any period of observation; conversely, HRs > 1 indicate an increased risk of death during any period of observation. Overall p < for the final multivariate Cox regression model. There are several limitations to our study. The data we analyzed represent a cohort of patients treated at a single academic medical center. Treatment biases of referral medical and radiation oncologists may have affected our results and outcomes. For example, ~ 40% of our patients received WBRT prior to their GKS session. Indeed, the lack of consensus in the treatment of these patients with either WBRT or GKS first in our catchment may account for our insignificant probability values for WBRT treatment. It is our institution s bias to perform SRS first given its excellent control rates. Much of our demographic, disease, histological, and survival data are similar to those previously reported, and we have no a priori reason to believe that the patients in our study were any different from those of other reported cohorts. Finally, the specific causes of death of each patient are unknown to us. Although we presume it is due to disease progression, we cannot know for certain. Thus, we chose death itself as our statistical outcome, rather than cause of death. Overall, this study confirms the findings of previous studies that have shown that GKS can prolong survival in patients with metastatic melanoma to the brain. The median survival after GKS treatment in this study was 4.4 months. This is comparable to other studies that quote a survival of months. 3,5 8,10,12 19,21 23,25,27 The principal factors that prolonged survival in this study were the presence of solitary metastasis, lower intracranial tumor volume, and the lack of hemorrhagic lesions. The cutoff volume found to be significantly associated with diminished survival in this study was a total lesion volume treated > 4 cm 3. Total tumor volumes exceeding 3 3,27 or 8 cm 3,13 have been shown to be negative predictors of survival. Reduced survival therefore seems to be related to increased tumor burden (size and number of intracerebral metastases) and possibly the aggressiveness of the tumor manifested as pre-gks intratumoral hemorrhage. Although GKS does not seem to affect hemorrhage rates, larger studies are needed to confirm this finding. Conclusions Gamma Knife surgery performed in patients with melanoma metastatic to the brain remains a viable and noninvasive option especially in situations where patients are poor surgical candidates or tumors may be surgically inaccessible. However, survival may be predetermined by the tumor burden and the presence of hemorrhagic lesions at the time of GKS. Our study indicates that GKS does not increase intratumoral bleeding and that doses of Gy result in good local control of tumor. The reason for post-gks hemorrhage is unrelated to dose/volume variability and remains undetermined at this time. Disclosure This study was supported, in part, through funds from the department of neurosurgery at the Yale School of Medicine. There were no financial conflicts of interest. Acknowledgments We are greatly indebted to the patients, physicians, and staff including Linda Ross and Margaret Pinto at the Yale-New Haven Gamma Knife Center who participated in this study. TABLE 4 Selected series of studies showing rates of hemorrhage in patients undergoing SRS for melanoma metastases to the brain* Cases w/ Cases w/ Median No. of Cases/ Median Pre-SRS Post-SRS Time to Authors & Year Modality No. of Mets Survival Bleeding (%) Bleeding (%) Hemorrhage Seung et al., 1998 GKS 55/ wks none 2/55 (3.6) 1 mo Mori et al., 1998 GKS 60/118 7 mos none 3/46 (7) 1 7 mos Grob et al., 1998 GKS 35/70 7 mos none 2/35 (5.7) Mingione et al., 2002 GKS 45/ mos 4/45 (8.9) 3/35 (2.8) Herfarth et al., 2003 LINAC 64/ mos none 3/64 (4.7) Selek et al., 2004 LINAC 103/ mos none 8/103 (7.8) Radbill et al., 2004 GKS 51/ wks 7/51 (14) 5/51 (9.8) 9 wks Gaudy-Marqueste et al., 2006 GKS 106/ mos none 1/106 (0.9) Mathieu et al., 2007 GKS 244/ mos 21/244 (8.6) 38/244 (15.6) 1.8 mos present study GKS 59/ mos 14/59 (24) 10/59 (16.9) 1.8 mos * LINAC = linear accelerator. There was imaging-defined response in 46 patients. Follow-up imaging was available in 35 patients. 104

7 Stereotactic radiosurgery for melanoma metastasis References 1. Brown PD, Brown CA, Pollock BE, Gorman DA, Foote RL: Stereotactic radiosurgery for patients with radioresistant brain metastases. Neurosurgery 51: , Chang EL, Selek U, Hassenbusch SJ III, Maor MH, Allen PK, Mahajan A, et al: Outcome variation among radioresistant brain metastases treated with stereotactic radiosurgery. Neurosurgery 56: , Christopoulou A, Retsas S, Kingsley D, Paddick I, Lindquist C: Integration of gamma knife surgery in the management of cerebral metastases from melanoma. Melanoma Res 16:51 57, DiLuna ML, King JT Jr, Knisely JP, Chiang VL: Prognostic factors for survival after stereotactic radiosurgery vary with the number of cerebral metastases. Cancer 109: , Gaudy-Marqueste C, Regis JM, Muracciole X, Laurans R, Richard MA, Bonerandi JJ, et al: Gamma-knife radiosurgery in the management of melanoma patients with brain metastases: a series of 106 patients without whole-brain radiotherapy. Int J Radiat Oncol Biol Phys 65: , Gieger M, Wu JK, Ling MN, Wazer D, Tsai JS, Engler MJ: Response of intracranial melanoma metastases to stereotactic radiosurgery. Radiat Oncol Investig 5:72 80, Gonzalez-Martinez J, Hernandez L, Zamorano L, Sloan A, Levin K, Lo S, et al: Gamma knife radiosurgery for intracranial metastatic melanoma: a 6-year experience. J Neurosurg 97: , Grob JJ, Regis J, Laurans R, Delaunay M, Wolkenstein P, Paul K, et al: Radiosurgery without whole brain radiotherapy in melanoma brain metastases. Club de Cancerologie Cutanee. Eur J Cancer 34: , Herfarth KK, Izwekowa O, Thilmann C, Pirzkall A, Delorme S, Hofmann U, et al: Linac-based radiosurgery of cerebral melanoma metastases. analysis of 122 metastases treated in 64 patients. Strahlenther Onkol 179: , Koc M, McGregor J, Grecula J, Bauer CJ, Gupta N, Gahbauer RA: Gamma knife radiosurgery for intracranial metastatic melanoma: an analysis of survival and prognostic factors. J Neurooncol 71: , Kondziolka D, Bernstein M, Resch L, Tator CH, Fleming JF, Vanderlinden RG, et al: Significance of hemorrhage into brain tumors: clinicopathological study. J Neurosurg 67: , Lavine SD, Petrovich Z, Cohen-Gadol AA, Masri LS, Morton DL, O Day SJ, et al: Gamma knife radiosurgery for metastatic melanoma: an analysis of survival, outcome, and complications. Neurosurgery 44:59 64, Mathieu D, Kondziolka D, Cooper PB, Flickinger JC, Niranjan A, Agarwala S, et al: Gamma knife radiosurgery in the management of malignant melanoma brain metastases. Neurosurgery 60: , Meier S, Baumert BG, Maier T, Wellis G, Burg G, Seifert B, et al: Survival and prognostic factors in patients with brain metastases from malignant melanoma. Onkologie 27: , Mingione V, Oliveira M, Prasad D, Steiner M, Steiner L: Gamma surgery for melanoma metastases in the brain. J Neurosurg 96: , Mori Y, Kondziolka D, Flickinger JC, Kirkwood JM, Agarwala S, Lunsford LD: Stereotactic radiosurgery for cerebral metastatic melanoma: factors affecting local disease control and survival. Int J Radiat Oncol Biol Phys 42: , Petrovich Z, Yu C, Giannotta SL, O Day S, Apuzzo ML: Survival and pattern of failure in brain metastasis treated with stereotactic gamma knife radiosurgery. J Neurosurg 97: , Radbill AE, Fiveash JF, Falkenberg ET, Guthrie BL, Young PE, Meleth S, et al: Initial treatment of melanoma brain metastases using gamma knife radiosurgery: an evaluation of efficacy and toxicity. Cancer 101: , Samlowski WE, Watson GA, Wang M, Rao G, Klimo P Jr, Boucher K, et al: Multimodality treatment of melanoma brain metastases incorporating stereotactic radiosurgery (SRS). Cancer 109: , Sampson JH, Carter JH Jr, Friedman AH, Seigler HF: Demographics, prognosis, and therapy in 702 patients with brain metastases from malignant melanoma. J Neurosurg 88:11 20, Selek U, Chang EL, Hassenbusch SJ III, Shiu AS, Lang FF, Allen P, et al: Stereotactic radiosurgical treatment in 103 patients for 153 cerebral melanoma metastases. Int J Radiat Oncol Biol Phys 59: , Seung SK, Shu HK, McDermott MW, Sneed PK, Larson DA: Stereotactic radiosurgery for malignant melanoma to the brain. Surg Clin North Am 76: , Seung SK, Sneed PK, McDermott MW, Shu HK, Leong SP, Chang S, et al: Gamma knife radiosurgery for malignant melanoma brain metastases. Cancer J Sci Am 4: , Suzuki H, Toyoda S, Muramatsu M, Shimizu T, Kojima T, Taki W: Spontaneous haemorrhage into metastatic brain tumours after stereotactic radiosurgery using a linear accelerator. J Neurol Neurosurg Psychiatry 74: , Tarhini AA, Agarwala SS: Management of brain metastases in patients with melanoma. Curr Opin Oncol 16: , Wronski M, Arbit E: Surgical treatment of brain metastases from melanoma: a retrospective study of 91 patients. J Neurosurg 93:9 18, Yu C, Chen JC, Apuzzo ML, O Day S, Giannotta SL, Weber JS, et al: Metastatic melanoma to the brain: prognostic factors after gamma knife radiosurgery. Int J Radiat Oncol Biol Phys 52: , 2002 Manuscript submitted June 13, Accepted July 30, Portions of this work were presented in abstract and poster forms at the Leksell Gamma Knife Society Meeting, May 2008, Quebec City, Canada. Address correspondence to: Veronica Chiang, M.D., B.Med.Sci., Department of Neurosurgery, Yale School of Medicine, 333 Cedar Street, P.O. Box , New Haven, Connecticut veronica.chiang@yale.edu. 105