VEGF signaling through VEGFR-2 is the major factor in glioblastoma angiogenesis. CT-322, a pegylated protein engineered from the 10th type III human fibronectin domain, binds the VEGFR-2 extracellular domain with high specificity and affinity to block VEGF-induced VEGFR-2 signaling. This study evaluated CT-322 in an open-label run-in/phase 2 setting to assess its efficacy and safety in recurrent glioblastoma. Eligible patients had 1st, 2nd or 3rd recurrence of glioblastoma with measurable tumor on MRI and no prior anti-angiogenic therapy. The initial CT-322 dose was 1 mg/kg IV weekly, with plans to escalate subsequent patients to 2 mg/kg weekly if tolerated; within each CT-322 dose cohort, patients were randomized to ±irinotecan IV semiweekly. The primary endpoint was 6-month progression-free survival (PFS-6). Sixty-three patients with a median age of 56 were treated, the majority at first recurrence. One-third experienced serious adverse events, of which four were at least possibly related to study treatment (two intracranial hemorrhages and two infusion reactions). Twenty-nine percent of subjects developed treatment-emergent hypertension. The PFS-6 rate in the CT-322 monotherapy groups was 18.6 and 0.0 % in the 1 and 2 mg/kg treatment groups, respectively; results from the 2 mg/kg group indicated that the null hypothesis that PFS-6 ≤12 % could not be rejected. The study was terminated prior to reaching the planned enrollment for all treatment groups because data from the completed CT-322 2 mg/kg monotherapy treatment arm revealed insufficient efficacy. Despite biological activity and a tolerable side effect profile, CT-322 failed to meet the prespecified threshold for efficacy in recurrent glioblastoma.