We conducted a randomized, non-comparative, multi center, phase II clinical trial in order to investigate the efficacy of axitinib, an oral small molecule tyrosine kinase inhibitor with high affinity and specificity for the vascular endothelial growth factor receptors, in patients with recurrent glioblastoma following prior treatment with radiation and temozolomide. Forty-four patients were randomly assigned to receive treatment with axitinib (5 mg BID starting dose; N = 22) or "physicians best alternative choice of therapy" that consisted of bevacizumab (N = 20) or lomustine (N = 2). Six-month progression-free survival served as the primary endpoint. The estimated 6-month progression-free survival rate was 34 % (95 % CI 14-54) for patients treated with axitinib and 28 % (95 % CI 8-48) with best alternative treatment; median overall survival was 29 and 17 weeks, respectively. Objective responses according to RANO criteria were documented in 28 % of patients treated with axitinib and 23 % of patients treated with best alternative therapy. A decrease in maximal uptake of 18F-fluoro-ethyL-tyrosine (18F-FET) by the glioblastoma on PET imaging was documented in 85 % of patients at the time of response on axitinib. Corticosteroid treatment could be stopped in four and tapered in seven out of the 15 patients who were treated with steroids at baseline in the axitinib cohort. Most frequent axitinib related grade ≥3 adverse events consisted of fatigue (9 %), diarrhea (9 %), and oral hyperesthesia (4.5 %). We conclude that axitinib has single-agent clinical activity and a manageable toxicity profile in patients with recurrent glioblastoma.