Dual Checkpoint Inhibition in NSCLC

Paul K. Paik, MD:While combination I-O regimens with chemotherapy have had the earliest data presented at this point, we should also mention the I-O/I-O combination trials that are going on. And there are a whole slew of I-O/I-O combinations that are in existence, too many to really talk about, many of these are still in the phase I setting, for example. And it’s important to identify patients who might be able to enroll in these things and potentially benefit from these trials. There are some later phase trials of I-O/I-O combinations that have either read out in terms of the top-line results or are about to read out. The one from last year that had read out for top-line results was the MYSTIC trial. This was a randomized trial of durvalumab plus tremelimumab versus chemotherapy, and the trial was negative for its primary endpoint of PFS. We don’t know any data beyond that at this point. They are going to take a look at the secondary endpoint of overall survival, and this was based on a PD-L1 expression cutoff of ≥5%. So, again, PD-L1 is the biomarker.

It will be interesting to figure out whether or not sequencing was done in that group of patients to see whether or not tumor mutation burden might emerge as a better biomarker within that subgroup, similar to what the CheckMate-026 investigators did, taking a look retrospectively at TMB as a biomarker. So, what I would say is that while that trial was negative, it still is going to be an important pool of data for us to analyze to try to figure out if there is a better biomarker that’s hidden within that subgroup of patients.

The other trial that read out, actually early just this week, was CheckMate-227. And CheckMate-227 had a few different arms. The most relevant arms were the ipilimumab plus nivolumab arm versus chemotherapy. Here, this trial was different. While it was originally designed to take a look at PD-L1 expression as their requisite biomarker based on that TMB data, the CheckMate-026 trial, which was the nivolumab versus chemotherapy trial, results seem to hew more closely to tumor mutation burden. The investigators created an amendment to take a look at TMB ≥10 mutations per megabase as a biomarker for PFS as a primary endpoint.

And the results were just read out that it’s a positive study. We don’t know any details beyond that. I’m sure the data will be presented in full later on this year at some conference. But here, again, we have the very first I-O/I-O combination data in a randomized phase III trial that’s positive, which is encouraging. It’s encouraging because, again, it provides a whole separate option of just I-O therapy that doesn’t consume, for example, chemotherapy at the beginning; one that may be amenable to, as a result, sequencing later on with chemotherapy. The one caveat to this, of course, is that the I-O/I-O combination we know from earlier trials, like CheckMate-012, has increased autoimmune toxicity in particular. The rate of ≥3 autoimmune adverse events is somewhere about 2- to 3-fold higher when you add ipilimumab to nivolumab than just with single-agent nivolumab. That rate is around 20%.

And so, that is not an insignificant increase in toxicity, especially in the first-line setting. We will need to be careful about seeing what the toxicity data are for CheckMate-227 and trying to think about how to then incorporate the toxicity data with the efficacy data in terms of figuring out in whom to prescribe these medications. Unaddressed, of course, is tumor mutation burden for pembrolizumab-containing regimens. At this point, we don’t know. And so, from that standpoint, it’s going to be difficult figuring out which is the optimum first-line treatment in our patient population now because we’re going to have different biomarkers. We’ll have TMB, we’ll have PD-L1. How will this pan out? How will we select patients with these competing regimens? At this point, we don’t know.

Transcript edited for clarity.