Durvalumab Demonstrates Durable Responses in Recurrent GBM


Findings from a recent phase II study showed the PD-L1 inhibitor durvalumab generated durable responses in bevacizumab-naïve patients with recurrent glioblastoma multiforme (GBM).

David Reardon, MD

David Reardon, MD

Findings from a recent phase II study showed the PD-L1 inhibitor durvalumab generated durable responses in bevacizumab-naïve patients with recurrent glioblastoma multiforme (GBM). Findings of the study were presented at the 2016 Society for Neuro-Oncology Annual Meeting.

In a 30-patient cohort, the 6-month progression-free survival (PFS) rate was 20.0% (6 patients; 90% CI, 9.7-33.0). The median PFS was 13.9 weeks (95% CI, 8.1-24.0). Of these 6 patients, 3 had wild-typeIDH1status and 3 had mutatedIDH1.

“Among the 6 patients who were progression-free at 6 months, 5 of them remained progression-free for more than 1 year, while 1 patient was ongoing at 48 weeks,” said lead author David A. Reardon, MD, Dana-Farber Cancer Institute.

Overall best responses were partial response in 4 patients (13.3%) and stable disease in 14 patients (46.7%) for a disease control rate of 60.0% (18 patients). The median overall survival (OS) was 28.9 weeks (95% CI, 22.9 to not estimated). The 6-month OS rate was 59.0% (90% CI, 42.6-72.2) and the 12-month OS rate was 44.4% (90% CI, 28.9-58.9).

This ongoing phase II, multicenter, open-label study has 3 patient populations: newly diagnosed patients with GBM with unmethylatedMGMTpromoter (n = 37), recurrent GBM that is bevacizumab-naïve, and recurrent GBM that is bevacizumab-refractory (n = 17). The bevacizumab-naïve population is divided into 3 groups based on concomitant therapy: no bevacizumab (n = 30), bevacizumab at 10 mg/kg once every 2 weeks (n = 32), and bevacizumab at 3 mg/kg every 2 weeks (n = 32).

This presentation focused on the 30 patients receiving no concomitant bevacizumab. These patients were treated with single-agent durvalumab at 10 mg/kg IV every 2 weeks for up to 12 months. The primary efficacy endpoint for this cohort was PFS at 6 months (PFS-6) and the historical benchmark of PFS-6 of 10% was used, based on historical data from the pre-bevacizumab era. The 30-patient intent-to-treat population included patients who received any dose of durvalumab and had at least 1 baseline and 1 post-baseline tumor assessment.

The first patient was dosed on March 5, 2015, and the data cutoff for this presentation was June 30, 2016. Durvalumab was discontinued by 25 patients (80.6%): 1 for adverse event (3.2%), 2 for physician decision (6.5%), and 22 for progressive disease (71.0%).

The safety analysis set of patients (n = 31) had a median age of 54.0 years (range, 24-77) and was 83.9% male. Measurable lesions at baseline occurred in 24 patients (77.4%). At treatment initiation, 9 patients (29.0%) were on steroids for GBM. The methylation status ofMGMTwas methylated for 9 patients (29.0%), unmethylated for 15 (48.4%), and unknown for 7 (22.6%). TheIDH1gene was mutated in 4 patients (12.9%), wild-type in 22 (71.0%), and unknown in 5 (16.1%). The median time from GBM diagnosis to enrollment was 40.3 weeks (range, 1.9-123.4). Prior to enrollment, 25 patients (80.6%) had 1 relapse and 6 patients (19.4%) had 2 relapses.

The median duration of exposure to durvalumab was 11.1 weeks (range, 0-60), and the median number of doses was 6.0 (range, 1-29).

No grade 4 or 5 treatment-related adverse events (TRAEs) occurred among the 31 patients in this safety analysis set. Additionally, no significant immune-related toxicities or clinically significant laboratory trends were noted. Among non-neurological TRAEs, the only grade 3 event reported was 1 patient with decreased lymphocyte count. Among neurological TRAEs, 1 patient experienced grade 3 headache and 1 patient experienced grade 3 dizziness.

Durvalumab is a human IgG monoclonal antibody that binds to PD-1 and CD80, allowing T cells to recognize and kills tumor cells without the need for cytotoxic activity that is antibody-dependent and cell-mediated. PD-1 is expressed by many lymphocytes that infiltrate GBM, and its ligand, PD-L1, is expressed by 61% to 100% of GBM tumors. The transcription and expression of PD-L1 in GBM is increased when the PTEN suppressor gene is lost, which occurs in 40% to 50% of GBM tumors.

In his concluding remarks, Reardon said that pending analyses are evaluating immunocorrelative biomarkers, and that the data he discussed warrant additional studies of durvalumab in GBM.


Reardon DA, Kaley T, Dietrich J, et al. Phase 2 study to evaluate the clinical efficacy and safety of MEDI4736 (durvalumab [DUR]) in patients with glioblastoma (GBM): results for cohort B (CUR monotherapy), bevacizumab (BEV) naïve patients with recurrent GBM. Presented at 2016 SNO Annual Meeting; November 17-20, 2016; Scottsdale, Arizona. Abstract ATIM-04.

Related Videos
Related Content