Historical controls for phase II surgically based trials requiring gross total resection of glioblastoma multiforme

Abstract

New treatments for patients with glioblastoma multiforme (GBM) are frequently tested in phase II surgically based clinical trials that require gross total resection (GTR). In order to determine efficacy in such single-arm phase II clinical trials, the results are often compared to those from a historical control group that is not limited to patients with GTR. Recursive partitioning analysis (RPA) can define risk groups within historical control groups; however, RPA analyses to date included patients irrespective of whether a patient had a GTR or not. To provide a more appropriate historical control group for surgically based trials requiring a GTR, we sought to determine survival for a group of patients with newly diagnosed GBM, all of who underwent GTR and were treated on prospective clinical trials. GTR was defined as removal of >90% of the enhancing mass, determined by postoperative magnetic resonance imaging. Of 893 patients with GBM treated during these trials, 153 underwent GTR. The median survival for the GTR group was 71 weeks (95% CI 65-76) which was better than those who did not have a GTR. Within the GTR group, the median age was 54 years (range 25-77 years), and median Karnofsky Performance Score was 90 (range 60-100). Considering only patients with GTR, age at diagnosis continued to be a statistically significant prognostic factor. Patients treated during surgically based phase II studies should be matched with a historical control group restricted to patients with similar pretreatment variables, including GTR.