While melanoma brain metastases (BM) are consistently associated with worse survival compared to other histologies, whether they correlate with worse local control (LC) following stereotactic radiosurgery (SRS) is not yet well-defined. In this study of prospectively and retrospectively collected data we investigated the impact of histology and other host, tumour and treatment factors on overall survival (OS) and LC. We analysed 162 patients and 318 BM lesions from various histologies treated with SRS between 2005 and 2011. We included patients who received SRS as first-line treatment, as well as patients who received SRS for residual or recurrent BM following prior surgery, whole brain radiotherapy (WBRT) or both. Median OS for the entire cohort was 8.4 months. Median OS for tumour histologies of melanoma, lung and breast cancer were 5.1, 12.2, and 14.7 months, respectively. On multivariate analysis, melanoma predicted for worse OS (hazard ratio [HR] 1.515, p = 0.003) together with performance status (HR 1.662, p < 0.001) and uncontrolled systemic disease (HR 1.755, p = 0.003). Melanoma histology was also negatively predictive for LC (HR 1.828, p = 0.021) together with increasing tumour size (HR 1.038, p = 0.017). Other factors, including the use of WBRT with SRS, the use of planning treatment volume margins, and prescription dose were not significantly predictive for OS and LC. We conclude melanoma histology also portends poorer LC in the SRS setting. While survival depends significantly on the systemic behaviour of the disease, treatment refinements to reduce local failure still merit exploration, especially in the era of targeted therapies.
Keywords: Brain metastases; Melanoma; Prognosis; Radiotherapy; Stereotactic radiosurgery.
Crown Copyright © 2014. Published by Elsevier Ltd. All rights reserved.