Gamma knife surgery for brain metastases: indications for and limitations of a local treatment protocol.

OBJECTIVE: The purpose of this retrospective study was to evaluate results of a local treatment protocol using gamma knife surgery (GKS) for brain metastases without upfront whole brain radiation therapy (WBRT). METHODS: Results for 521 consecutive patients satisfying the following 3 criteria were analysed: 1) a maximum of 3 tumours with a diameter of 25 mm or more; 2) no prior WBRT; 3) no surgically in accessible large (>30 mm) tumours. Large tumours were surgically removed and all smaller lesions were treated by GKS without up front WBRT. New lesions, detected with follow-up MRI, were appropriately treated with repeat GKS. Overall survival (OS), neurological survival (NS), qualitative survival (QS) and new lesion-free survival (NLFS) curves were calculated and the prognostic values of covariates were obtained. OS and NS were compared according to tumour number. RESULTS: In total, 1023 separate sessions were required to treat 4562 lesions. The primary organs were lung in 369 patients, gastro-intestinal tract in 70, breast in 33, urinary tract in 24, and others/unknown in 25. The median OS period was 9.0 months. On multivariate analysis, the significant prognostic factors for OS were found to be extracranial disease (risk factor: active), Karnofsky performance status (KPS) score (<70) and gender (male). NS and QS at one year were 85.6% and 73.0%, respectively. The only significantly poor prognostic factor for NS was carcinomatous meningitis. NLFS at 6 months was 68.9%. For both OS and NS, the differences between a few (10) tumours had a significantly poorer prognosis than those with