Gamma Knife radiosurgery for intracranial metastatic melanoma: an analysis of survival and prognostic factors

Abstract
Objective of this study was to evaluate retrospectively the effectiveness of Gamma Knife radiosurgery for intracranial metastatic melanoma and to identify prognostic factors related to survival. Twenty-six patients with intracranial metastases (72 lesions) from melanoma underwent Gamma Knife radiosurgery. In 14 patients (54%) whole-brain radiotherapy (WBRT) was performed as part of the initial treatment, and in 12 patients (38%) immunotherapy and/or chemotherapy was given after Gamma Knife radiosurgery. The median tumor volume for Gamma Knife radiosurgery treated lesions was 1.72 cm3. The median prescribed radiation dose was 18 Gy (range 8–22 Gy) typically prescribed to the isodose at the tumor margin. Univariate and multivariate analyses were used to determine significant prognostic factors affecting survival. Overall median survival was 6 months after Gamma Knife radiosurgery, and 1-year survival was 25%. The median survival from the onset of brain metastases was 9 months and from the original diagnosis of melanoma was 50 months (range 4–160 months). There were no major acute or late GKS complications. In univariate testing, the Karnofsky score equal to or higher than 90% (P < 0.01, log-rank test), supratentorial localization (P < 0.001, log-rank test), intracranial tumor volume less than 1 cm3 (P < 0.02, log-rank test), and absence of neurological signs or symptoms before Gamma Knife radiosurgery (P < 0.003, log-rank test) were significant favorable factors for survival. In multivariate regression analyses, the most important predictors associated with increased survival were a KPS ≥ 90 (P < 0.023), female sex (P < 0.004), supratentorial localization (P < 0.01), and absence of neurological symptoms (P < 0.008). Radiosurgery is a noninvasive, safe, and effective treatment option for patients with single or multiple intracranial metastases from melanoma. Female sex, Karnofsky score ≥90, supratentorial localization and lack of symptoms before the Gamma Knife radiosurgery were good independent predictors of survival.