• brain metastases;
  • stereotactic radiosurgery;
  • whole-brain radiotherapy;
  • recursive partitioning analysis class



The authors investigated whether stereotactic radiosurgery (SRS) alone improved outcomes for patients in recursive partitioning analysis (RPA) Classes 1 and 2 who had 1 to 3 brain metastases compared with whole-brain radiotherapy (WBRT).


Data regarding 186 patients in RPA Classes 1 and 2 who had 1 to 3 brain metastases and who received either 30 to 40 grays (Gy) of WBRT (n = 91 patients) or 18 to 25 Gy SRS (n = 95 patients) were analyzed retrospectively. Eight other potential prognostic factors were evaluated regarding overall survival (OS), entire brain control (BC), local control (LC) of treated metastases, and brain control distant from treated metastases (distant control [DC]): Those 8 factors were age, sex, performance status, tumor type, number of brain metastases, extracranial metastases, RPA class, and interval from tumor diagnosis to radiotherapy.


On multivariate analysis of OS, age ( risk ratio [RR], 1.51; P = .024), Karnofsky performance status (KPS) (RR, 1.98; P = .002), and extracranial metastases (RR, 2.26; P < .001) were significant, whereas the radiation regimen was not significant (P = .89). On multivariate analysis of BC, only the radiation regimen (RR, 1.33; P = .003) was found to be significant. On multivariate analysis of LC, radiation regimen (RR, 1.63; P < .001) and sex (RR, 1.62; P = .022) were significant. On multivariate analysis of DC, KPS (RR, 1.85; P = .049) and extracranial metastases (RR, 1.69; P = .047) were significant. The radiation regimen was not found to be significant even on univariate analysis (P = .80). In RPA class subgroup analyses, BC and LC were better after SRS than WBRT for patients in RPA Classes 1 and 2, whereas OS and DC did not differ significantly.


For patients in RPA Classes 1 and 2 who had 1 to 3 brain metastases, SRS alone was associated with improved BC and LC compared with 30 to 40 Gy WBRT, whereas OS and DC were not significantly different. Similar results were observed in separate subgroup analyses of patients in RPA Class 1 and RPA Class 2. Cancer 2007. © 2007 American Cancer Society.