Effectiveness and Safety of Combined Cranial Stereotactic Radiation and Anti-PD-1 Therapy for Melanoma Brain Metastases

Abstract

Even though it has been proven that anti-programmed death-1 (anti-PD-1) therapy increases overall survival (OS) among metastatic melanoma patients, radiotherapy is still needed either for management of unresectable brain metastases (BM) as in postoperative setting. Thus, it is necessary to define effectiveness and safety of anti-PD-1, nivolumab in this cohort, and cranial stereotactic radiation (CSR) combined. We selected individuals with detected BM who were treated with CSR within 6 months of receiving nivolumab. Analyses were retrospectively done in 53 patients with metastatic melanoma treated with nivolumab on an expanded access program at a single institution. The median number of previous systemic treatments before nivolumab was 3. The primary end points were BM control and OS; secondary end points were toxicity, brain progression free-survival (BPFS), and whole-brain radiotherapy free-survival (WBRTFS). Nine patients with 38 BM who underwent CSR from May 2014 to October 2016, were analyzed. Eight (89%) of these patients received CSR before and one (11%) during nivolumab treatment, with a median time of 49 (38-186) days before the first nivolumab cycle. Five (13%) lesions were treated with stereotactic fractionated radiotherapy (SFRT), from which 3 were resected and had the tumor bed irradiated. The remaining 33 (87%) lesions underwent stereotactic radiosurgery (SRS), single fraction, with a median dose of 18,5 Gy and a median volume of 0,6 cc. None acute toxicity was observed. Six lesions were considered as local BM failure through a median follow-up of 16 months, and WBRT was performed in 3 patients after BM disseminated failure. Kaplan-Meier estimates for BM control were 79,7% and 72,5% at 12 and 24 months, respectively, while the median OS from the date of CSR was 22,4 months. The median BPFS and WBRTFS were 15,37 and 21,9 months, respectively. In this study, the OS and BM control were superior to those ones established for standard treatments. Neurotoxicity did not appear to be an issue in our cohort and neither on the scarce available studies about this matter. Further prospective investigational studies are necessary to confirm our conclusions.