Bevacizumab for Newly Diagnosed Glioblastoma
Abstract and Commentary
By Susan C. Pannullo, MD, FAANS
Associate Professor of Clinical Neurological Surgery, Director of Neuro-Oncology, Weill Cornell Medical College
Dr. Pannullo reports no financial relationships relevant to this field of study.
Synopsis: Two recent randomized, placebo-controlled clinical trials of treatment of glioblastoma multiforme with bevacizumab have not shown superiority over standard treatment.
Sources: Gilbert MR, et al. A randomized trial of bevacizumab for newly diagnosed glioblastoma. N Engl J Med 2014;270:699-708.
Chinot OL, et al. Bevacizumab plus radiotherapytemozolomide for newly diagnosed glioblastoma. N Engl J Med 2014;270:370:709-722.
Glioblastoma multiforme is a frustrating and com-
plex disease with poor survival and very limited treatment options. The standard therapeutic approach is maximal safe resection, followed by radiation therapy and systemic chemotherapy. The survival benefits of the current therapies remain poor, prompting constant attempts to add to these modest benefits through the addition of novel
A pair of articles in the February 20, 2014, issue of the New England Journal of Medicine present data from the Radiation Therapy Oncology Group (RTOG) 0825 trial and the Avastin in Glioblastoma (AVAglio) trial. These two Phase 3, randomized trials examined the role of bevacizumab, a humanized monoclonal antibody that targets vascular endothelial growth factor, in treatment of patients with newly diagnosed glioblastoma multiforme. Subjects in both studies received the standard 6 weeks/5 days per week of radiation therapy to a maximum of 60 Gy with concurrent oral temozolomide followed by maintenance temozolomide. Both studies randomly assigned subjects 1:1 to intravenous bevacizumab vs placebo. These trials were prompted by previous Phase 2 studies suggesting safety and efficacy of bevacizumab in patients with glioblastoma multiforme at recurrence, as well as at time of initial diagnosis.
The RTOG and AVAglio trials shared a relatively common study design. Both trials enrolled adults with newly diagnosed, histologically centrally confirmed supratentorial glioblastoma, with good performance status, stable or decreasing steroid doses, good wound healing, no postoperative hemorrhage on imaging, and adequate hematologic, hepatic, renal, and coagulation parameters. Biopsy without resection was permitted in the AVAglio study. In the AVAglio trial, bevacizumab (or placebo) was started at the same time as chemoradiation; in the RTOG study, bevacizumab infusions (vs placebo) began at week 4 of chemoradiation. There were additional minor differences in the study designs: In RTOG 0825, subjects received maintenance temozolomide for a maximum of 12 4-week cycles with q2 week bevacizumab vs placebo, while the AVAglio maintenance phase was planned for six 4-week cycles of temozolomide with concurrent q2 week bevacizumab vs placebo followed by a q3 week bevacizumab vs placebo monotherapy phase. Both trials continued treatment until disease progression or unacceptable toxicity occurred. In the RTOG trial, the study was unblinded at time of progression and the subject could continue or start (if on the placebo arm) bevacizumab. Similarly, in AVAglio, subjects could receive salvage therapy at progression. Coprimary endpoints in both trials were overall survival and progression-free survival.
A total of 637 subjects were randomized in the RTOG study (312 to bevacizumab, 309 to placebo, with 8 excluded from each group) and 921 in the AVAglio study (458 bevacizumab, 463 placebo). Neither trial demonstrated survival advantage for "upfront" use of bevacizumab. In RTOG 0825, median survival in the bevacizumab group was 15.7 months vs 16.1 months for placebo, not statistically significant. In AVAglio, median survival was 16.8 months for the bevacizumab arm and 16.7 month for placebo.
Both trials demonstrated increased duration of progression-free survival with bevacizumab (RTOG = 10.7 months bevacizumab vs 7.3 months placebo; AVAglio = 10.6 months bevacizumab vs 6.2 months placebo). However, this parameter has been of particular debate as an outcomes measure, due to the belief that bevacizumab may produce radiographic changes on MR imaging as a result of alterations in blood-brain barrier permeability that may cause regression of tumoral gadolinium enhancement used to define tumor burden. This effect has been termed "pseudoregression" and mirrors the "pseudoprogression" that has been described as a confounder of imaging analysis in clinical neuro-oncology and a disruptor of image-based outcomes in brain tumor clinical trials.
Patients in RTOG 0825 had greater deterioration of neurocognitive measures, perceived symptom severity, and health-related quality of life in the bevacizumab-treated subjects compared with the placebo group. Interestingly, time to deterioration of several similar measures was delayed in the AVAglio trial. The AVAglio study results were notable for increased rate of adverse events, including arterial thromboembolic events, bleeding, wound healing complications, gastrointestinal perforation, and heart failure, in the bevacizumab group vs placebo.
Negative trials are always discouraging and disappointing, but are important nevertheless. In this orphan disease with few potential subjects, large trials are difficult to complete. Further complicating the issue of trials for glioblastoma are patient and physician biases that may encourage potential subjects to seek therapies without participation in a randomized trial. The RTOG 0825 and AVAglio studies emphasize the value of rigorously performed randomized clinical trials. Although potential subjects (and investigators) often "fear" the placebo arm as being "worse" than the study drug arm, the study drug, in the end, may be no better than the placebo. In fact, in some aspects (e.g., adverse events in the bevacizumab arm, neurocognitive function in the RTOG study), outcomes on the study drug subjects may be worse. These two trials are particularly disappointing as they cast an unfavorable light on what was thought to be a promising therapeutic approach in a devastating disease that has few treatment options and a median survival of just over 1 year. Discussions are occurring regarding plans to pool the raw data from these studies and evaluate the combined data in an additional attempt to confirm the findings and understand the discrepancies between them. In addition, examination of molecular marker subgroups may reveal a subpopulation of glioblastoma patients who would derive benefit in the future. Finally, these studies call attention to the need for refinement of imaging techniques used to evaluate response to brain tumor therapies.