New Data for Checkpoint Inhibitors in NSCLC

Transcript:Mark Socinski, MD: We talked about adding immunotherapy to chemotherapy as a strategy. One of the other strategies is adding immunotherapy agents together. And, obviously, we’ve been interested in the PD-1, PD-L1 axis with CTLA-4. So we have a couple of trials. I believe the MYSTIC trial has been done for quite some time in terms of its accrual, and I think the CheckMate-227 trial is close to being done if it’s not done already. But these trials are exploring these combinations. What are your thoughts?

Tracey Evans, MD: We have the early stage data on the combination of durvalumab, which is a PD-L1 inhibitor with tremelimumab, which is a CTLA-4 monoclonal antibody. And the hope is that these combinations of immunotherapy may capture more of those patients who are not high PD-L1 expressors. There are data that seem to show that there is an improved response rate when these agents are used in combination. The phase Ib data from the durvalumab and tremelimumab study showed a response rate of 23% in these patients, and it didn’t seem to matter what their level of PD-L1 expression was. And the MYSTIC study is a randomized study with that combination—durvalumab by itself or with standard chemotherapy.

Then you have the nivolumab and ipilimumab combinations, where they actually did seem to depend upon the PD-L1 level. The response rates were better in those patients that did not express PD-L1 than we tend to see with the checkpoint inhibitors by themselves. But, still, the people who had a higher level were an enhanced population who had a better response rate. The problem with the combinations is that they have increased toxicity. The CTLA-4 antibodies are much more difficult relative to the PD-1 inhibitors. The side effects we see with these agents are autoimmune events. The one I most commonly see is hypothyroidism developing as people stay on these agents. But when I treated a patient on a study of a combination, I ended up with a case of aplastic anemia. We’ve had perforations from colitis with the combinations. In the durvalumab/tremelimumab study, 30% had to discontinue treatment due to treatment-related toxicities. So, that’s a pretty high rate of toxicities that we’re seeing with these combinations.

Mark Socinski, MD: So, the question is, is the bang worth the buck at the end of the day? We don’t know yet.

Alexander Drilon, MD: Yes, and I personally don’t think we’re seeing enough of an augmentation of response. As you mentioned, it was on the order of 20% or more with a combination, which is what you see with single-agent, second-line therapy. And, in fact, it may be telling that the application for the nivolumab/ipilimumab approval was recently withdrawn. So, that might give us an early hint of what the final results will look like.

Mark Socinski, MD: Yes, that’s a good point.

Vali Papadimitrakopoulou, MD: It was a limited set of patients that was reported, but the response rate was as high as 45% in that small group of patients. So, I think we need to give the data some time to mature.

Mark Socinski, MD: Actually, to Alex’s point, I was impressed that MYSTIC did have an independent DSMB (data and safety monitoring board), and it went to full accrual. Even though the toxicity may be worse, it didn’t stop the trial like we’ve seen with a number of new agents in lung cancer, where trials were stopped because of toxicity. So, I’m reassured by that. But I do share your concern that in the average lung cancer patient, that the toxicity scale may be tipped with immunotherapy combinations, whereas I don’t have that same feeling with chemotherapy in this particular setting. Getting back to October 9th, Alex, in Copenhagen, we also saw the presentation of the OAK trial, which brought us yet another option in the second-line setting. Could you walk us through the atezolizumab OAK data?

Alexander Drilon, MD: Atezolizumab, as opposed to nivolumab and pembrolizumab, is a PD-L1 inhibitor—a monoclonal antibody against PD-L1. So, going back to that toxicity element, the thought was, can we maybe see a better toxicity profile if we use one of these trials? And durvalumab, that you mentioned, is also a PD-L1 drug. So, the efficacy that we saw, talking about that first, the response rate was 15%, hazard ratio was 0.73, and the overall survival was 14 months. It was similar in design to the previous CheckMate and KEYNOTE trials, where the control arm was docetaxel. But I don’t think that I saw a dramatic difference in toxicity with atezolizumab and the other agents that were previously tested. One thing that struck me was that, if you look at the range of patients who had various shades of staining, for this particular drug, you look not just at the tumor cells but also at the infiltrating component. And if you look from 0 up to 3+, there was some benefit in those subsets—granted that it was cued toward more benefit in the higher staining patients.

I think my take home message from this is, if you have someone who needs an immune checkpoint inhibitor in the second-line setting, giving atezolizumab is reasonable, especially considering the schedule that’s given every 3 weeks—and not 2 weeks for nivolumab—in the absence of having enough tissue for PD-L1 staining.

Mark Socinski, MD: We saw data—you mentioned durvalumab, Tracey mentioned the ATLANTIC study- your thoughts on the activity of that agent?

Jared Weiss, MD: I think the activity there was sufficient to justify the ongoing studies that we’ve discussed. We had different cohorts in ATLANTIC. It started with all-comers, and then there was a cutoff at 25%, and a later cutoff at 90%. As far as efficacy, what I took away from that, mostly, was that there were real durable responders. That phenomenon was there, and it was tolerable—justifying the ongoing studies.

Mark Socinski, MD: Let’s assume that the patient is not strongly PD-L1-positive, gets first-line, and arrives at second-line. You have 3 choices. What’s your go-to second-line choice of immunotherapy?

Vali Papadimitrakopoulou, MD: It used to be nivolumab up until the recent approval of atezolizumab. But because of the scheduling issues, I think atezolizumab is going to become increasingly adopted, in my practice at least.

Mark Socinski, MD: Jared?

Jared Weiss, MD: I have the greatest comfort with pembrolizumab, just having experience back to early trials. So, if they’re PD-L1-positive, I use that. If they’re not, due to the schedule, I use atezolizumab.

Tracey Evans, MD: We tried to put together a pathway on this very idea, and I was hoping we could choose the agent that cost least. They all come out similarly extremely expensive, so that one was a watch. But we’ll use the Q3 week approaches and our pathway, because we’re trying to keep things as similar as we can—pembrolizumab if there’s any PD-L1 positivity, and if it’s 0, then atezolizumab.

Mark Socinski, MD: And Alex, at Memorial Sloan Kettering Cancer Center?

Alexander Drilon, MD: Same as Jared. If positive, pembrolizumab. If negative, or you can’t do PD-L1, atezolizumab.

Mark Socinski, MD: Atezolizumab for scheduling?

Alexander Drilon, MD: Correct.

Jared Weiss, MD: And now that we’re testing for first-line, there shouldn’t be unknown in second-line. And one final point on that unknown is that I think it’s even worse, in a way, to have a PD-L1 unknown than molecular, not because of extent of benefit but because this is just an IHC. This takes a slide. There’s no reason it should take more than 2 days to come back.

Tracey Evans, MD: Are you going back and testing the people who are already on first-line or through first-line?

Mark Socinski, MD: Is there any role for testing it beyond first-line?

Tracey Evans, MD: No, I don’t think there is.

Vali Papadimitrakopoulou, MD: Okay, so I need to clarify that. If we’re talking about standard of care therapy…

Mark Socinski, MD: Yes, that’s what I’m talking about.

Vali Papadimitrakopoulou, MD: If we’re talking about clinical trials, I think there is a great deal of questioning. What do you do after progression on immunotherapy?

Mark Socinski, MD: Right. That’s my next question.

Vali Papadimitrakopoulou, MD: Should you be testing for PD-L1 in that setting? What other biomarkers should we be looking for? I think that’s going to be keeping us busy for the next few years.

Tracey Evans, MD: It’s interesting. We have a patient with brain metastases who was progressing all over and who we put on second-line nivolumab without testing her, and she’s having a fabulous response. So, now that we’re testing everybody, my nurse went back and tested her and she was 0. She’s having a fabulous response.

Transcript Edited for Clarity