Geoffrey Y. Ku, MD, discusses data of durvalumab plus tremelimumab in a biomarker-selected patient population and sheds light on the role of immunotherapy in patients with advanced gastric cancer.
Geoffrey Y. Ku, MD
Although prospective screening with an RNA-based interferon (IFN)-γ gene expression signature proved feasible for selecting patients with gastric cancer who would most benefit from the combination of durvalumab (Imfinzi) and tremelimumab, use of this signature failed to lead to a significant improvement in clinical outcomes, said Geoffrey Y. Ku, MD.
In arm E of a 5-cohort study,1 which was presented at the 2019 AACR Annual Meeting, researchers analyzed the ability of a tumor-based IFN-γ signature to identify a subset of patients with advanced gastric cancer who would have a higher likelihood of responding to checkpoint inhibition.
Out of 176 patients who were prescreened in the second- or third-line setting, the signature assay had a 70% rate for providing actionable data; 37.5% of these patients had high expression of tumor IFN-γ gene signature. A total of 19 patients were enrolled in the study. At a median follow-up of 5.8 months, the objective response rate (ORR) was 15.8% (95% CI, 3.4%-39.6%) with the combination. One patient experienced a complete response, 2 experienced partial responses, and 1 achieved stable disease. Median progression-free survival (PFS) was 1.8 months (95% CI, 1.6-1.9) while the median overall survival (OS) was 7.0 months (95% CI, 2.4-7.5).
For the study, patients were prescreened using archival, embedded tissue to determine the IFN-γ gene signature status with the use of a targeted RNA sequencing assay. Patients who tested positive during prescreening were eligible to screen. Those who were enrolled in the study received intravenous durvalumab at 20 mg/kg and tremelimumab at 1 mg/kg every 4 weeks for 4 cycles, followed by durvalumab 10 mg/kg every 2 weeks for ≤12 months.
Data with the unselected patient population, arm D, were presented at the 2018 ASCO Annual Meeting. Results showed a median PFS of 1.8 months (95% CI, 1.6-3.3) and a median OS of 10.6 months (95% CI, 4.8-17.0) in patients treated with third-line durvalumab plus tremelimumab.2 Although there was a higher ORR observed with the combination in the selected patients compared with unselected population, rates of PFS and OS were comparable between the 2 groups. As such, these results do not support further use of this signature for patient selection in this indication and treatment regimen.
In an interview with OncLive during the 2019 AACR Annual Meeting, Ku, a medical oncologist at Memorial Sloan Kettering Cancer Center, discussed these data and shed light on the role of immunotherapy in patients with advanced gastric cancer.
OncLive: Could you speak to the recent headway made in advanced gastric cancer?
Ku: The field is actually moving pretty fast. In the fall of 2017, a little more than 1.5 years ago, pembrolizumab (Keytruda) was approved by the FDA in the third-line setting and beyond for PD-L1—positive tumors. Since then, things have really taken off. We now have data that these drugs are active in the second-line setting, and there are ongoing trials looking at their use in the frontline setting. We're also beginning to combine these agents with radiation in the preoperative setting.
What is known about anti—PD-1 and anti–CTLA-4 in this space?
The evaluation of combination immunotherapy [in gastric cancer] really mirrors what [was done] in melanoma, [a space where it] was found to be a beneficial strategy. The CheckMate-032 study looked at ipilimumab (Yervoy), a CTLA-4 antibody, with nivolumab (Opdivo), a PD-1 antibody. This was a nonrandomized study, and about 50% of the patients who were treated did seem to have activity, but with significant toxicity. That regimen was actually taken forward into a phase III study, the CheckMate-649 study. In this frontline study, investigators randomized patients to receive either chemotherapy alone, chemotherapy with nivolumab, or the combination of nivolumab plus ipilimumab. All we know at this time is that about 1 year ago, the nivolumab/ipilimumab arm was closed, either because of increased toxicity or decreased efficacy. To some degree, we may be at a standstill with regard to CTLA-4 plus PD-1/PD-L1.
What was the rationale for the phase I study of durvalumab with tremelimumab?
Because CTLA-4 plus PD-1 inhibition was promising in melanoma, there was reason to look at the combination in other tumors, including gastric cancer. Durvalumab is a PD-L1 inhibitor; there remains a lack of clarity about whether antibodies like durvalumab, atezolizumab (Tecentriq), and avelumab (Bavencio) are comparable with PD-1 antibodies such as pembrolizumab and nivolumab. They inhibit the same axis, but do they have the same level of activity? At this point, there are a couple of studies that have looked at PD-L1 antibodies in gastric cancer, most notably with avelumab. Those studies have been relatively disappointing.
In this context, durvalumab was evaluated in a 5-cohort study involving second- and third-line patients. With specific reference to the abstract, we were trying to use an interferon-γ signature to try to enrich patients who were more likely to respond.
What were the goals for this research?
The abstract [presented at the 2019 AACR Annual Meeting] focuses on cohort E of the 5 cohorts. This specific cohort was treating patients in the second- and third-line settings with the combination of durvalumab and tremelimumab. Patients had to be preselected on the basis of that IFN-γ signature in their tumors. Retrospectively, patients who had this tumor signature seemed to have a higher likelihood of responding to durvalumab or tremelimumab. The hope was that by preselecting these patients, we could enrich responses in this particular cohort.
What was the design of the trial?
The novelty of the study was that patients could be preselected even while receiving prior lines of therapy. For pragmatic reasons, this required archival tissue to test for the IFN-γ signature. Any patient who had the signature would then be eligible at the time of progression. Such a strategy was selected because it was recognized that it would take several weeks to run the IFN RNA signature. We were trying to run this in real time, so it wouldn't be feasible in patients who needed treatment quickly. It turned out that this was largely feasible.
What were the key takeaways from this research?
The key takeaway was somewhat disappointing. It was feasible to screen patients using the IFN-γ signature, but despite the effort that went into this, it did not seem to result in a higher response rate. Certainly, the expectation was that we would be able to identify a small group of patients who were more likely to benefit from the combination strategy, but that was not likely the case. With the hazard of comparing across small studies, the activity [observed] seemed very similar to [that seen in] an unselected population. Moving forward, this would suggest that a signature like this, at least in a retrospective basis, may not be feasible.
The field of immunotherapy is rapidly evolving in different cancers. Combination strategies are what we are actively looking at in the next phase of evaluations. It does remain unclear whether CTLA-4 in combination with PD-1 or PD-L1 will move forward. It also remains unclear, and it's probably unlikely, that we would preselect patients using an IFN-γ signature.
Beyond this work, what are some other promising trials that are ongoing in gastric cancers?
The 2 studies of most interest are looking at adding a PD-1 inhibitor to chemotherapy in the frontline setting. The first is KEYNOTE-062, which might be reported as early as [June 2019]. The other study is CheckMate-649, which is a study that is nearing completion. If either of these studies are positive, that would really transform the treatment of gastric cancer.