David Reardon, MD
Findings from separate trials exploring the PD-1 inhibitor pembrolizumab (Keytruda) and the PD-L1 inhibitor durvalumab (MEDI4736) point to a role for checkpoint inhibitors in the treatment of glioblastoma multiforme (GBM), based on encouraging efficacy signals and safety data with the two agents reported by David Reardon, MD, at the 21st Society for Neuro-Oncology (SNO) Annual Scientific Meeting.
Reardon said that these results mark important firsts in the field: “There has been a lot of anticipation regarding the role of checkpoint inhibitors for glioblastoma and whether we’ll see results in any way similar to the exciting results that have been observed in other cancer indications with this new class of cancer therapeutics.”
spoke with Reardon, clinical director, Center for Neuro-Oncology, Dana-Farber Cancer Institute, while at the SNO 2016 Annual Meeting to discuss these two trials, next steps, and what these advancements mean for the field of neuro-oncology.
OncLive: Let’s start with the KEYNOTE-028 trial.
: The KEYNOTE-028 trial was a “basket” clinical trial looking at the PD-1 inhibitor pembrolizumab across a number of solid tumor types, including recurrent GBM. All of the patients treated on this study received single-agent pembrolizumab. The GBM basket included patients with any recurrence: about one-third of the patients were enrolled at second or higher recurrence, and about two-thirds were treated following their first recurrence. These were bevacizumab-naive patients who had progressed after standard of care therapy. All of the patients received pembrolizumab on a standard dosing schedule of 10 mg/kg every 2 weeks.
What were the results of this trial?
This was a single-arm study, so there's no control population, but there are well-established historical control data for recurrent GBM to judge our outcomes. The study revealed that these agents were quite well-tolerated for the glioblastoma population—there were really no unusual or unexpected side effects observed which is always reassuring in the brain cancer setting.
What we observed from an efficacy point of view was also encouraging, but the results do have to be understood in the context of a single-arm, fairly small study—nonetheless, they are encouraging. We saw a rate of progression-free survival (PFS) at 6 months of 45%. The historical benchmarks for recurrent glioblastoma with salvage therapies are in the order of about 10%. Bevacizumab (Avastin), which is FDA-approved on an accelerated basis in the United States for recurrent GBM, has a PFS–6-month rate of about 40%. So what we saw with this checkpoint-inhibitor data is comparable with data on bevacizumab, in terms of PFS at 6 months.
What was really exciting, though, was not only were patients maintaining PFS rate at 6 months, but a subset of these patients maintained their durability well beyond that and have remained progression-free for up to 80 weeks. This suggests there may be a very exciting tail of the curve for glioblastoma patients in a way that's analogous to what's been observed in other cancer indications, such as melanoma treated with a PD-1 or CTLA-4 checkpoint inhibitor.
We also saw an encouraging overall survival rate on this study, although the follow-up and maturation of the data are ongoing.
What are the next steps for this trial?
These are very encouraging results for pembrolizumab, and we certainly are moving forward with other clinical trials, utilizing this particular PD-1 checkpoint inhibitor. We have completed a larger, randomized, phase II study of recurrent glioblastoma patients who were randomized to receive pembrolizumab monotherapy versus pembrolizumab plus bevacizumab in the recurrent setting. That trial randomized about 80 patients between the 2 arms. We expect a read out probably early next year. Hopefully, that will build on this smaller, single-arm study experience.
There are also a number of other investigator-initiated trials going forward with pembrolizumab, particularly in promising combinatorial regimens for glioblastoma patients.
We don't have results from this yet, but also ongoing is an evaluation of some of the immuno-correlative biomarkers. The patients on this study were required to have tumors that were positive for PD-L1 expression, which we know has been an important predictor in other cancers of patients who are more likely to benefit from checkpoint blockade. We are looking at that, and the data are forthcoming, not only PD-L1 as a potential biomarker, but other immuno-correlative biomarkers that were collected and are currently being analyzed and will be reported subsequently as well.
Finally, we're moving forward with a large registration study to evaluate pembrolizumab in the recurrent GBM setting that will be a large, randomized, phase II, multicenter, multinational study.
These initial results have really triggered a cascade of additional significant studies going forward, all the way to a planned registration study for this drug.
What has been learned thus far from the MEDI4736 trial?
This is a phase II clinical trial testing the PD-L1 inhibitor durvalumab in patients with GBM. It is a 5-arm study that includes 1 arm of newly diagnosed patients, and then 4 arms of recurrent glioblastoma patients. We're reporting on 1 of the arms for the recurrent GBM patients, involving 32 bevacizumab-naive patients who received single-agent durvalumab therapy. Again, similar to the KEYNOTE-028 pembrolizumab study, we saw a very well-tolerated side effect profile in patients treated with durvalumab.
In addition to the drug being well-tolerated in recurrent GBM patients, we also saw very encouraging efficacy results. The durvalumab study is the first PD-L1 inhibitor to be evaluated in patients with GBM, and there's been a lot of discussion of how a PD-L1 will compare relative to a PD-1 inhibitor. This is an important study to report out on that outcome.
What we have seen with this phase II study is that the results appear to be comparable to what's been reported with the PD-1 inhibitors. Specifically, we saw very exciting evidence of activity. Although there was a low rate of overall radiographic response, about half of the patients achieved stable disease with PD-L1 inhibitor therapy. What we saw in this study regarding PFS at 6 months was about 20% of the patients, which again is superior to historical benchmarks for recurrent GBM patients.
Even more exciting was the durability of this PFS in a subset of patients. All 6 patients who were progression-free at 6 months have remained progression-free for over a year. Some are at the 15-month mark and going strong. That again suggests there's this tail of the curve of a subset of patients who are having a remarkably durable benefit with this PD-L1–targeting immunotherapeutic for recurrent GBM.
This study is also looking at immuno-correlative biomarkers in the hope that we'll be able to better identify who are the patients more likely to benefit and who are maybe less likely. And, what can we potentially do to develop combinatorial regimens or other therapies to help improve their response rate and bring in a higher percentage of patients responding more durably?
What do these trials and new advances mean for oncologists and their therapeutic approach?
The KEYNOTE-028 and MEDI4736 studies are the first phase II trials reported on immune checkpoint inhibitors for glioblastoma patients. Both are in the recurrent setting which is always a challenging group of patients, but these findings indicate that the 2 agents appear to have a benefit and a role for this indication.
The medical oncology community should be paying attention to the development of immunotherapeutics for brain cancer and using the exciting results with these studies to be thinking about the development of immunotherapy for brain cancer. My hope is we'll see more and more studies validating the role of immunotherapy for brain cancer, as we have seen for melanoma, lung cancer, kidney cancer, and other cancer indications.
What do you think is the big takeaway from these trials?
The take-home message from the immunotherapy trials that have been presented here at the 2016 SNO meeting is that, very similar to other types of cancers, there appear to be a subset of patients with brain cancer who are also deriving significant benefit from these agents. Unfortunately, just as with other cancers, it seems to be a minority, but at least this research provides some exciting preliminary results and proof of principle that these agents can also have an impact on this cancer indication as well. It gives us data upon which we can build and move forward.