Zev A. Wainberg, MD: The role of immunotherapy in gastric cancer is based on the premise that, relative to a lot of other malignancies, gastric and GEJ cancers do have a good number of neoantigen formations. That is one of the principle mechanisms why perhaps immunotherapy might work in this cancer. When compared to colorectal cancer, for example, or pancreatic cancer, we have a fair number of neoantigen formations in these malignancies that could make them sensitive to immunotherapy. Beyond that there are other subtypes, such as MSI-high, which represents only a small percentage of patients with gastric and GE junction cancer, but is particularly sensitive to checkpoint inhibition. The roles of the EBV subtype and TCGA subsets are still being evaluated.

It makes scientifically credible sense that the group of patients who are virally mediated, such as the EBV group, may have particular sensitivity to the checkpoint inhibitors as well. But some of the retrospective analyses are only being done now, so we’re not entirely sure whether those data will be supported by the clinical data. It’s going to take a little more time to figure out which are the optimal subtypes for this disease, where immunotherapy may have a role. It’s an exciting time in many ways because there are a lot of immunotherapy drugs out there, and there are a lot of potential combinations out there.

Yelena Y. Janjigian, MD: The KEYNOTE-059 study was presented at ASCO this year, and it targeted patients beyond second-line therapy, third-line, and beyond; a heavily pretreated patient population. Patients were permitted to go on cohort 1 regardless of PD-L1 status. So, PD-L1 tumor expression was tested and known, but both PD-L1—positive and PD-L1–negative tumors were permitted to enter.

This was a monotherapy pembrolizumab study in treatment-refractory metastatic disease. What the study has demonstrated, and what we know, is that tumors have a dramatic response to pembrolizumab. There were a majority of patients who were PD-L1—positive—up to 50% to 60% of tumors positive by the PD-L1 immunohistochemistry test—and the PD-L1 positivity was hopefully, in this population, used to select patients who were the most likely to benefit from pembrolizumab.

The overall response rate in PD-L1 positive patients in the third-line setting was 21%. Again, these were unprecedented data, and therefore, they granted approval to pembrolizumab based on these data, based on the response rate and promising overall survival, and pending some of the follow-up phase III studies demonstrating actual long-term benefit from these agents.

A similar agent, nivolumab, was also approved in Japan for use in the third-line setting, also in metastatic gastric cancer, in a randomized phase III study against placebo. For the first time, we’re seeing that a similar proportion of patients with metastatic gastric cancer in Asia and the West are benefiting from these drugs, from the same biology and similarities that we see, which is quite reassuring.

Interestingly for the nivolumab trial, the ATTRACTION-2 study, PD-L1 was not predictive of response or survival with anti—PD-1 therapy. So, there’s a lot to be learned about PD-L1 as a biomarker, but right now as it stands in the KEYNOTE-059 study, responses in PD-L1–positive patients were higher, although there were PD-L1–negative patients who responded as well. The FDA approval right now is for PD-L1–positive patients only for pembrolizumab in third-line therapy or beyond.

Alan P. Venook, MD: The value of checkpoint inhibitors in gastric cancer seems to be comparable to what we’re seeing in many other solid tumors. In the rare patient who’s MSI high, that is clearly an indication for the checkpoint inhibitors. Generically, there is activity, perhaps 15% to 20% response rates, in the KEYNOTE studies with Keytruda, which is pembrolizumab. They’re pretty much across the board. There’s a 15% to 20% response rate whether first-line or second-line or third-line treatment, and it’s not clear to me who benefits and who doesn’t. There has been interest in looking at PD-L1 expression. It’s not clear to me that that makes a difference. In general, though, I think for patients with gastric cancer it’s hard to not try a checkpoint inhibitor at some point, because there are some patients who have dramatic responses.

With pembrolizumab in particular, though, they’ve done a series of studies and to my knowledge—their results are not all in—on average, there’s about a 20% response rate. One of the challenges, of course, is that these studies accrue remarkably quickly, and so we might have 2 or 3 patients on a study. You don’t get much experience with it. I will say, though, that in the patients we’ve treated on-study with pembrolizumab, we’ve seen a couple of dramatic responses. And all you have to do is see 1 response to believe that the treatment may have effect. Why these checkpoint inhibitors would work in gastric cancer or GE junction cancer compared to others? We don’t understand, although those cancers may be more related to toxins and environmental stimuli that may make the chances a little more mutated. The data remain to be shown definitively for what the role for these checkpoint inhibitors are in gastric cancer.

Transcript Edited for Clarity