H. Jack West, MD: We have talked about KEYNOTE-189 and we’ve talked about KEYNOTE-407 as the real chemo/I-O [immuno-oncology] standards for patients with nonsquamous or squamous non—small cell, respectively, except for the high–PD-L1 patients where pembro [pembrolizumab] alone, based on KEYNOTE-024, is a very strong choice. Those are not the only chemo/I-O trials, and one was actually just very recently approved based on the IMpower150 trial. This was the arm that had carbo [carboplatin]/paclitaxel/bevacizumab with atezolizumab. So it beat carbo/paclitaxel, and bev [bevacizumab] for the nonsquamous population across the spectrum of PD-L1 based on the Roche assay. But does this have a place? And if so, where?

Charu Aggarwal, MD, MPH: So I’ve been struggling since the approval came out on where I would use it, not just since the approval came out but even when the data were presented, just because we do have other randomized trials showing us an advantage in terms of overall survival. We have all these KEYNOTE trials that you just mentioned. I just don’t see myself using a quadruplet just yet. I think it increases cost, it increases toxicity, and I’m getting very little benefit at the end after using all 4 drugs.

Hossein Borghaei, DO: Yes, I agree. OK, so if I have somebody with such a degree of renal failure, I cannot use pemetrexed if they’re nonsquamous, which is where the approval of this is. All right, it gives me an option. But when you look at the bottom-line data, up to a third of those patients in that arm had to come off the study because of toxicity. So it is a more toxic regimen, and I agree that I’m not sure of what we’re getting in terms of clinical efficacy and all of that.

H. Jack West, MD: I think sadly what you’re most reliably getting is more alopecia and neuropathy.

Hossein Borghaei, DO: Neuropathy, yes.

H. Jack West, MD: I think the clearest path to utility, I would say, is the EGFR mutation—positive patients. Again, these are not a group who have been well studied with these trials, but this was the first that showed that these patients could benefit at least as much, perhaps even more.

Hossein Borghaei, DO: But the approval doesn’t include that, right?

H. Jack West, MD: Right. The approval does not include that, which is unfortunate because that is, to me, the setting where I would be most inclined to use it.

Hossein Borghaei, DO: I agree.

H. Jack West, MD: But we don’t have that. Now, you both raised the issue of, “Well, it’s positive, but how does it really compare?” The reality today, not just for IMpower150 but for trials like IMpower130, which was with a carbo and nab-paclitaxel backbone versus carbo/nab-pac [nab-paclitaxel], and atezo [atezolizumab] in nonsquamous using the nab-pac because of its nonsteroid requirement, or the very-similar-to-KEYNOTE-189 IMpower132 trial, that looked at cis [cisplatin] or carbo and pembro with or without atezo. These are both positive trials. The IMpower132 trial didn’t make it on overall survival [OS] but was positive for PFS [progression-free survival]. IMpower130 with the nab-pac was positive for PFS and OS, but is there any place that these could bubble up, or are they better than chemotherapy but just not better than the standards as they are in late 2018, based on KEYNOTE-189 or KEYNOTE-407?

Hossein Borghaei, DO: I’m not sure what you’re necessarily asking.

H. Jack West, MD: These are both nonsquamous trials, so let me back up. The question is, we have a couple of other trials in advanced nonsquamous disease. One is cis or carbo with pemetrexed and atezo, and that beat chemotherapy alone. That’s IMpower132. And we have IMpower130, which was carbo and nab-paclitaxel with or without atezo. These are both positive trials, but do they change the field at all, or are they just 1 more positive result but one we don’t need because it didn’t beat the real comparator today, which is KEYNOTE-189 in the nonsquamous population?

Hossein Borghaei, DO: Yeah, so that’s an interesting question. To me that goes to the heart of whether there are differences between these PD-1s and PD-L1s that we’re seeing. I think it is possible that we’re beginning to see some differences within PD-1 antibodies versus PD-L1. And biologically, you can come up with a scenario where perhaps certain combinations would work better with a PD-1 and certain combinations would work better with a PD-L1 antibody. Maybe chemotherapy is one of them, I’m not 100% sure. I think they just add, as you said, to me, to the fact that, OK, we have another positive study. I’m not sure they’re as much of a game changer as we’ve had with the KEYNOTE-189 on that.

Obviously, in an area where, again, we don’t have head-to-head comparisons, it’s hard because remember how we had the study using carbo and pemetrexed/bev versus carbo/Taxol/bev and we all thought pem [pemetrexed] was going to be better because we thought the data were better? But at the end of it, it ended up being a very equal regimen in terms of survival. So it is possible that in the absence of a phase III head-to-head comparison, it’s hard to make that kind of conclusion.

But it does look, from the trial perspective, from the fact that KEYNOTE-189 had positive results within 10 to 12 months of follow-up, and the other ones are taking a little bit longer to have that kind of a survival advantage, that perhaps 1 antibody is a little bit better or 1 combination is a little bit better than the other. I’m not sure whether that answers your question or not, but I’m beginning to wonder whether we’re beginning to pick up these differences between the PD-1s and the PD-Ls that in the second-line setting we had no evidence for superiority of one versus the other.

Charu Aggarwal, MD, MPH: So I see your point. There may be differences in the antibodies. But the hazard ratio on the IMpower studies that Jack just mentioned, the 130 and the 132, just don’t seem as compelling as the KEYNOTE-189 and the 407 studies. And coming back to your argument about carbo/Taxol/bev versus carbo/pem/bev, they did demonstrate, in a randomized trial, that yes, maybe they were not that different. I don’t think we have any randomized data to say that carbo/pem/atezo is similar or dissimilar to carbo/pem/pembro. I think looking at the data as they are, I would say that IMpower130 and 132 don’t really add much to my practice. Good to know, but I don’t really need to use that to integrate anything. It doesn’t change what I do.

Hossein Borghaei, DO: Exactly.

Charu Aggarwal, MD, MPH: And I’m just intrigued by why those differences exist. So I agree that there may be antibody differences, but we just don’t know whether it’s patient population, whether it’s trial design, whether it’s follow-up. I don’t know.

Hossein Borghaei, DO: So let me add this to it. At this past ESMO [European Society for Medical Oncology Congress], there was a very interesting presentation, and this had to do with just reviewing some of these studies. But there was one slide as part of that presentation, and I’m sorry I don’t remember the presenter’s name right now, and we can look this up, but there’s one slide that shows that the rate of antidrug antibodies that can actually be neutralized is higher with some of the antibodies that we’re using in the immunotherapy world than the others.

H. Jack West, MD: A French study.

Hossein Borghaei, DO: It is a French analysis, but I think the data actually came from the FDA, if I remember the slide correctly. So it is possible that some of the differences we’re beginning to see have to do with the fact that some of these antibodies are causing more neutralizing antidrug antibodies versus the others. And percentages, the way it was presented, were such that you could look at this study and say, “Now, I can see if 30% of patients had a neutralizing antidrug antibody, that would make this study less positive than the comparator that has a 4% chance of causing it.” So we might be getting to an area where we’re going to have to look at this a little bit more carefully. So it’s possible.

H. Jack West, MD: I think that we are, as a group, moving away from the concept based on the remarkably similar results in the second-line setting. These agents are essentially fungible, interchangeable, and equivalent. There may be real differences. It’s hard to know how much of it is methodologic. They do have different assays, and obviously SP142 is an outlier there. But we are seeing that consistently the atezo trials are either barely positive or just a little bit behind the pembro trials or just barely negative. We’ve seen some of the nivo [nivolumab] trials fail in this setting and maybe durva [durvalumab], particularly in metastatic disease, or avelumab may fall short. I am certainly less cavalier about saying that these are essentially the same agents. I don’t feel nearly that confident that that’s the case anymore. And it may be. I think we need to get more information about the antidrug antibodies. That’s a really important observation, but our first foray into this that we may learn more about.

Transcript Edited for Clarity