Radiosurgery for brain metastases: the Tuebingen experience

Purpose: To retrospectively investigate the effectiveness of linear accelerator based radiosurgery (RS) in the treatment of brain metastases (BM). Material and methods: Of 55 patients with a total of 72 BM, 41 patients had a single brain metastasis and 14 patients had two or three metastases. Median tumour dose of 15 Gy (range 8-20 Gy) was prescribed to a median isodose surface of 90% (range 70-100%) encompassing the target volume. Results: The median survival time (MST) for all 55 patients was 7 months [95% confidence interval (CI), 5-10 months] and 2-year survival is 18%. There was no significant difference between patients who had one brain metastasis and those with either two or three metastases (log rank P=0.7565). Multivariate analysis in patients with a single BM showed that interval between primary diagnosis (PD) to BM, maximum size of metastasis, and histology (renal cell carcinoma and melanoma versus others) were independent prognostic factors influencing survival. Local control was obtained in 66/72 (92%) metastases. Actuarial local control at 24 months was 52%. Only age (less than or equal to50 years vs >50 years) and histology (renal cell versus others) influenced local control in the univariate analysis in patients with a single BM. In multivariate analysis, size, histology (renal cell and melanoma versus others), activity of extracranial metastatic disease, age, interval from PD to BM and location (midline versus other) independently influenced local control, while the dose was not significant for our patient group. Only one patient developed radiographically suspected RS-induced necrosis after previous whole brain RT. Conclusion: RS was effective and little toxic in BM. Identification of prognostic factors must be performed to gain knowledge on patients most likely to benefit from this procedure. (C) 2002 Elsevier Science Ireland Ltd. All rights reserved.