An identical proportion of patients in both treatment groups were alive at 12 months in the CheckMate 143 trial

Overall survival (OS) was not different in patients with recurrent glioblastoma who received single-agent immune checkpoint inhibitor therapy compared with those who received an anti-angiogenesis agent, the first, randomized, phase 3 study of its kind has found.

At a median follow-up of 9.5 months, median overall survival (mOS), the primary end point of the study, was 9.8 months (95% CI, 8.2-11.8 months) for patients treated with the programmed cell death-1 (PD-1) inhibitor, nivolumab, versus 10.0 months (95% CI, 9.0-11.8 months) for patients treated with the anti-angiogenesis agent, bevacizumab, David Reardon, MD, Dana-Farber/Harvard Cancer Center, Boston, and colleagues reported in JAMA Oncology.

At 12 months, OS was identical in both groups at 42%, investigators added.

Overall survival was also “generally similar” in different patient subgroups, although tumor type and baseline use of corticosteroids did moderate outcomes, investigators added.

“To our knowledge, the CheckMate 143 randomized clinical trial is the first phase 3 study investigating the use of a PD-1 inhibitor in patients with recurrent glioblastoma,” Reardon and colleagues observed.

“[While t]he study did not meet the primary end point of OS, [t]he safety profile of nivolumab in patients with glioblastoma was consistent with that in other tumor types,” researchers noted.

Some 184 patients were randomized to nivolumab and another 185 patients were assigned to bevacizumab.

Patients had grade IV recurrent glioblastoma or gliosarcoma after receiving standard-of-care surgical resection, radiotherapy and temozolomide.

Patients had to have undergone surgery a minimum of 28 days and radiation a minimum of 12 weeks before study enrollment.

Those assigned to nivolumab received an intravenous (IV) dose of 3 mg/kg of nivolumab every other week while those assigned to bevacizumab received an IV dose of 10 mg/kg, also every other week.

Treatment was continued until disease progression or unacceptable toxicity.

“No statistical difference was observed in the risk of death between groups,” investigators reported.

At study end point, 83.7% of patients treated with nivolumab had died as had 79.5% in the anti-angiogenesis group.

Median progression-free survival favored bevacizumab at 3.5 months (95% CI, 2.9-4.6 months) compared with median PFS of 1.5 months (95% CI, 1.5-1.6 months) in the PD-1 inhibitor arm (P<0.001), the authors note.

Similarly, the overall response rate (ORR) was higher in bevacizumab-treated patients at 23.1% (95% CI, 16.7-30.5%) compared with 7.8% (95% CI, 4.1-13.3%) for nivolumab-treated patients.

On the other hand, responses were more durable for patients treated with nivolumab at 11.1 months compared with 5.3 months for those treated with bevacizumab.

In an exploratory analysis, investigators identified two factors that were associated with a longer median survival: methylated MGMT promoter glioblastoma lack of baseline corticosteroid use.

In the small subgroup of patients with methylated glioblastoma and no steroid use at baseline, there was a trend towards improved survival with nivolumab at 17 months compared with bevacizumab at 10.1 months (Hazard Ratio (HR), 0.58 (95% CI, 0.30-1.11), the authors noted.

“Longer mOS was observed in patients with methylated tumors than in patients with unmethylated tumors in both treatment groups,” Reardon and colleagues also pointed out.

Interestingly as well, OS in patients treated with nivolumab and who received steroids at baseline was shorter at only 7 months compared to 12.6 months for nivolumab-treated patients not on steroids at baseline.

Thus, investigators suggested that patients with methylated MGMT promoter glioblastoma and no baseline corticosteroid use may derive some benefit from immune checkpoint inhibition.

Similar rates of grade 3 and 4 treatment-related adverse events (TRAEs) were observed in both groups at 18.1% and 15.2% for nivolumab and bevacizumab, respectively.

Rates of grades 3 to 4 neurological TRAEs were also similar between the 2 groups, at 4.4% for patients treated with nivolumab and 1.2% for those treated with bevacizumab.

Commenting on the findings , editorialist Linda Liau, MD, PhD, University of California, Los Angeles and colleagues cautioned that these findings need to be carefully considered in the context of certain limitations in the study design itself.

First, almost two-thirds of patients on the nivolumab arm and almost 56% of patients in the bevacizumab arm received additional treatment when they discontinued the trial drug.

Some underwent additional surgery or radiation or were treated with cytotoxic, investigational or immunologic drugs with different therapeutic targets.

“Inconsistencies in the treatments patients received after coming off the study could have perturbed the OS results,” the editorialists suggested.

Second, the significant difference seen in PFS rates between the 2 treatment arms are also difficult to interpret, given that responses to immunotherapy as seen on imaging studies tend to evolve over time and clinicians are still not skilled at differentiating true disease progression from pseudoprogression, Liau pointed out.

Despite these limitations, “the hypothesis that MGMT methylation may be a predictive biomarker for patients with glioblastoma who could benefit from immunotherapy is an important finding,” the editorialists felt.

Liau and colleagues themselves reported a survival advantage for patients with MGMT-methylated glioblastoma where 3-year survival rates exceeded 46% in a group of patients with MGMT-methylated tumors following treatment with an autologous dendritic cell vaccine compared with only 11% in the MGMT-unmethylated group.

The observation that patients taking a corticosteroid at baseline did worse than those who were not may be explained by the fact that steroid-induced immunosuppression hampers the action of therapies that work by activating the immune system such as the PD-1 inhibitors.

“[G]lioblastoma still remains one of the most lethal of all human cancers and there has been no U.S. Food and Drug Administration-approved immunotherapeutic treatment for brain cancers to date,” Liau and colleagues observed.

“The role of immunotherapy for glioblastoma certainly deserves continued investigation,” they concluded.

  1. Treatment of recurrent glioblastoma with either an immune checkpoint inhibitor or an anti-angiogenesis agent resulted in the same poor rates of median overall survival in a first-of-its kind randomized trial.

  2. Be aware that there was a small signal that patients with methylated glioblastoma not on steroids might still derive some benefit from immune checkpoint inhibition.

Pam Harrison, Contributing Writer, BreakingMED™

The study was funded by Bristol-Myers Squibb.

Reardon has received grant funding from Acerta Pharmaceuticals, Incyte, Midatech, Omniox, and Tragara; grant funding and personal fees from Agenus, Celldex, EMD Serono, and Inovio; and personal fees from Advantagene, Genentech/ Roche, Merck, Merck KGaA, Monteris, Novocure, Oncorus, Oxigene, Regeneron, Stemline Therapeutics, and Taiho Oncology.

Liau reported participation (PI or Co-PI) in sponsored clinical trials related to immunotherapy for glioblastoma and funded by Northwest Biotherapeutics and by Merck.

Cat ID: 115

Topic ID: 78,115,730,115,935,129,192