Background: Preclinical studies indicate that local recurrence after radiotherapy (RT) of Glioblastoma is dependent on recovery of tumor vasculature following RT via hypoxia-driven SDF-1 (CXCL12) secretion. We hypothesize that blocking the CXCL12/CXCR4 axis would improve local control. Methods: Newly diagnosed Glioblastoma patients were recruited to this Phase I study of a 4 week intravenous infusion of Plerixafor (Mozobil), a CXCR4 antagonist, with concurrent Temozolomide (TMZ) and radiation therapy. Three patients were treated with Plerixafor at 8.3 µg/kg/hr and six patients at 16.6 µg/kg/hr, while being monitored for dose limiting toxicities (DLTs) including grade ≥ 3 hematologic or non-hematologic adverse events. Patients underwent dynamic susceptibility contrast perfusion MRI (DSC-MRI) for quantification of relative cerebral blood volume (rCBV) values by region-of-interest analysis. Pharmacokinetic (PK) analysis of Plerixafor plasma levels were collected. Results: Since August 2014, 9 patients completed therapy with upfront RT/TMZ and Plerixafor with no DLTs observed, leading to the recommended phase II dose of 16.6 µg/kg/hr. Currently, 7 of 9 patients are alive, with the longest survival since diagnosis being 18 months. Plasma Plerixafor levels reached our target of 100 ng/ml at the first time point, 7 days into therapy. DSC-MRI at 1 and 6 months post-RT revealed lower rCBV in the tumor bed compared to the postoperative pre-RT scan (P < 0.02, one way ANOVA). Conclusions: Plerixafor was safely escalated to the target dose of 16.6 µg/kg/hr with no DLTs observed. A marked decrease in rCBV in the radiation treatment field suggests enhanced local treatment effect in support of our hypothesis that inhibition of the SDF1/CXCL4-mediated vasculogenesis pathway in the post-RT period enhances radiation. Recruitment into the phase II study is ongoing to evaluate these preliminary observations. Clinical trial information: NCT01977677