Poor prognostic indicators for patients treated with radiosurgery for brain metastases.

Abstract

32 Background: Radiosurgery (SRS) is a standard palliative treatment for patients with limited number of brain metastases. Growing evidence supports the use of SRS for more extensive disease. As SRS is increasingly used in advanced cancer, we sought to identify predictors of survival after SRS to help better inform patients about prognosis. Methods: We reviewed patients treated with SRS for brain metastases at the University of California, San Francisco (UCSF) from Jan 2010-Dec 2013. Before SRS, all patients were screened for appropriateness of SRS at a multidisciplinary conference. Post-SRS overall survival (OS) was estimated by Kaplan-Meier method and compared by log-rank tests. Frequencies of death within 30 days of SRS were compared by chi-squared tests. Results: Demographic and imaging data was available for 326 SRS patients. At SRS consult, 90% patients were Karnofsky Performance Status (KPS) ≥ 70. Median number of brain metastases at consult was 2 (range 1-37). Median follow up was 13 months. Median OS after SRS was 11 months. Median OS was shorter for patients with KPS < 70 (4 months) compared to those with KPS ≥ 70 (12 months) (p < 0.001). Overall, frequency of death within 30 days of SRS was 5.4%. Within 30 days of SRS, 13% of patients with KPS < 70 died compared to 3.8% with KPS ≥ 70 (p = 0.03). Death within 30 days of SRS was more common among patients with uncontrolled (10%) versus controlled primary tumors (3%) (p = 0.02). Both uncontrolled primary tumor and KPS < 70% were associated with death within 30 days of SRS on multivariate analysis (p = 0.02 and p = 0.03, respectively). Patients with both KPS < 70 and uncontrolled primary tumors had frequency of death within 30 days of 29%. Factors not associated with death within 30 days were age, extracranial metastatic disease burden, histology, number of brain metastases, prior whole brain radiotherapy, SRS, or neurosurgery. Conclusions: Death within 30 days of SRS within the UCSF cohort is uncommon, likely due to multidisciplinary screening for patients expected to live long enough to benefit from SRS. Most patients who die within 30 days of SRS have both KPS < 70 and uncontrolled primary tumors. The potential for poor prognosis should be discussed with these patients during shared-decision making.