Background: The safety and efficacy of concurrent pembrolizumab (anti-PD1) and stereotactic radiosurgery (SRS) for brain metastases (BM) is unknown. Methods: Patients with melanoma or NSCLC, 1-10 brain metastases, ≥ 1 extra-cranial lesion, age ≥ 18, and ECOG 0-1 were treated with anti-PD1 every 3 weeks. SRS was administered 1-2 days after starting anti-PD1. SRS used three different radiation arms: Arm A used 6 Gray (Gy) in 5 fractions (fx), Arm B used 9 Gy in 3 fx, and Arm C used 18-21 Gy in single fx. Primary endpoint was grade 3 CNS toxicity at 3 months (CTCAE v 4.0). Secondary endpoints were overall survival (OS), local control (LC) within the SRS field, intra-cranial progression free survival (IC-PFS), extra-cranial progression free survival (EC-PFS), rate of extra-cranial clinical benefit, and immunological changes. OS, LC, IC-PFS, and EC-PFS were estimated using the Kaplan-Meier method, and covariates were compared using log-rank tests. 95% confidence intervals for 6-month and 12-month were estimated using Greenwood’s formula. Results: 25 patients were treated from 2016 until 2020. The mean age was 61. The mean number of CNS lesions was 2.7. The mean number of extra-cranial lesions was 2.5. Six were enrolled on Arm A, 12 on Arm B, and 7 on Arm C. 21 had melanoma. 4 had NSCLC. Of the melanoma, 8 were BRAF-, 10 were BRAF+, and 3 had unknown mutation status. 12 patients (48%) had progressed on prior immunotherapy and/or other oncological therapies. The trial met its primary endpoint, with no grade 3 CNS toxicity at 3 months. Two patients (8%) experienced ≥ Grade 3 anti-PD1 related toxicity, and no grade 5 toxicity was noted. The median OS was 32.8 months. The 6 and 12 month OS were 79.1% (56.5-90.8%) and 67.8% (43.3-83.5%), respectively. The 1 year OS was similar between previously treated and treatment naïve patients (71.8% vs. 65.6%), suggesting some role for SRS in overcoming therapy resistance. However, with longer follow-up, the OS trended worse (p=0.07) for previously treated patients. LC was 95.7% (72.9-99.4%) at 6 and 12 months. IC-PFS at 6 months was 69.1% (45.8-83.9%), and at 12 months was 57.5% (33.7-75.5%). The EC-PFS at 6 and 12 month was 54.5% (32.1-72.4%) and 43.6% (22.3-63.2%), respectively. Clinical benefit, which was defined as a best overall response of stable disease or better according to RECIST 1.1, occurred in 12 patients (48%). No outcome differences were noted amongst the three different SRS arms. 70% of the patients demonstrating early activation (within 3 weeks of starting SRS/anti-PD1) of CD8+PD1+Ki67+ T cells demonstrated a clinical benefit. 100% of patients that failed to show early activation of CD8+PD1+Ki67+ T cells progressed. Conclusions: Concurrent pembrolizumab (Anti-PD-1) and SRS is safe and effective. Early activation of CD8+PD1+Ki67+ T cells correlates with improved outcome. Further trials testing pembrolizumab and SRS are justified. Clinical trial information: NCT02858869