High-grade gliomas (HGGs), including glioblastoma, are a heterogeneous group of primary brain tumors, associated with devastating neurological sequelae and limited survival. In 2005, a randomized phase III study established postoperative radiotherapy and temozolomide as the standard of care for patients with resected, newly diagnosed glioblastoma. Despite this progress, almost all patients relapse and therapeutic options in the recurrent setting are limited. The optimum approach for recurrent HGG is challenging because of tumor resistance and the worsening performance status of the patients. As glioblastoma is a highly vascular tumor and has high levels of vascular endothelial growth factor, there has been interest in the use of the vascular endothelial growth factor targeting, monoclonal antibody bevacizumab. In a series of phase II studies, bevacizumab alone or with irinotecan showed improvements in tumor response, disease control, and survival compared with historical controls. These results led to the licensing of bevacizumab for glioblastoma in the USA, but a contrasting view was adopted by the European Medicines Agency, because of (deemed) modest response rates and lack of direct comparisons with other agents. Against this background, Gil and colleagues conducted a retrospective review of 130 patients with recurrent HGG treated with bevacizumab and irinotecan and showed an encouraging median progression-free survival of 5.1 months (95% confidence interval, 4.4-5.9) and a median overall survival of 9.0 months (95% confidence interval, 6.7-11.2), in agreement with other series. In this editorial, the context and implications of these results are discussed, with a particular focus on the possible need and design of randomized phase III trials.