Mark A. Socinski, MD: OK, let’s transition. We’ve had a nice discussion on stage III disease and have talked about a number of different issues. Let’s transition to stage IV disease. And again, in my mind, with stage IV disease, the first thing is to define any potential targets, and to identify those patients who are going to get a greater benefit from targeted agents. We have a number of targets and a number of approved agents. That’s not going to be our focus today. One of the issues is what to do if you don’t have a target. Then I do think the issue of chemoimmunotherapy is the topic of discussion. I think all of us would agree that staining for PD-L1 [programmed death-ligand 1] by IHC [immunohistochemistry] is a standard of care in stage IV disease.

Corey J. Langer, MD: Absolutely.

Mark A. Socinski, MD: I want to get Corey’s perspective on those patients who are high strainers, greater than 50% is the benchmark we use. What are your thoughts for that population?

Corey J. Langer, MD: So that was a game changer. Marty Reck, MD, PhD’s, original presentation at ESMO [the European Society for Medical Oncology] was just 3 years ago. It was not that long ago.

Mark A. Socinski, MD: Actually, October will be 3 years. So less than 3 years.

Corey J. Langer, MD: You and I were there.

Mark A. Socinski, MD: Yes, we were.

Corey J. Langer, MD: Roughly a third of patients with wild-type advanced non—small cell lung cancer have at least 50% expression. And KEYNOTE-024 showed definitively, in that population, that single-agent, non-chemotherapy, pembrolizumab, far and away exceeded standard chemotherapy doublets with platinum. The progression-free survival improved on the order of about 4 months, so there was a major improvement in response rate. And most importantly, there was a survival advantage that was detectable within a year—70% versus the high 50% range. And over time, as the data matured, that survival advantage, if anything, has improved further. The median survival in that trial was 30 months—unprecedented for single-agent pembrolizumab—compared to about 14, 14-1/2 months for chemotherapy alone. So the control arm was doing at least as well or better than it had done historically.

KEYNOTE-042, which was the plenary session last year at ASCO [the American Society of Clinical Oncology], I view as practice-affirming but not practice-changing. That same cohort, granted that was a trial that expanded to 1% or higher, but if we just looked at 50% or higher again, the positivity of that trial was driven by the 50% or higher group. The median survival and the data are a little more mature, about 20 months versus 12 months. You could argue it’s a better slice of the real world. There’s a more global trial. There was a higher percentage of folks with comorbidities in smokers. But nevertheless, including that population, there was a major survival advantage.

We’ve got to remember that with KEYNOTE-024, well over 1600 patients were screened for PD-L1. The ultimate accrual was just over 300 patients. So there’s a lot of attrition. A lot of folks do not qualify for these trials. There’s a presentation here, at this meeting, that will actually look at trial qualifications. Probably 50% of patients that we see would never get on these trials.

Mark A. Socinski, MD: I must say, ever since you first presented KEYNOTE-021 Cohort G, one of your many claims to fame…. This trial looked at the response rate of chemoimmunotherapy in the greater than 50% range. So there are many patients for which even though they stain high, who may be more symptomatic, and may have a higher tumor burden and that sort of thing, I actually will use the combination of chemoimmunotherapy, even in the 50% range.

Corey J. Langer, MD: Who have more fulminant disease.

Mark A. Socinski, MD: Correct. So how does this extend to the less than 50% group? Let me ask Heather. Heather, in the 1% to 49% population? We saw this, KEYNOTE-042.

Heather A. Wakelee, MD: Right.

Mark A. Socinski, MD: And your thoughts?

Heather A. Wakelee, MD: I always emphasize looking for driver mutations, because I think in the community there’s still a push towards, “Oh, my PD-L1’s back the next day. I’m just going to put him on immune therapy alone. Or chemoimmunotherapy in that 1% to 49% group.” So assuming no driver, the patient is 1% to 49%.

Mark A. Socinski, MD: Can I make just 1 point before you do that? I say, whenever I talk about this, that you should completely ignore PD-L1 status until you have the genomic testing done.

Heather A. Wakelee, MD: That’s exactly what I say to people also. I think that’s an important point. So it is important to emphasize that because if you don’t know, you don’t know.

Mark A. Socinski, MD: Right.

Heather A. Wakelee, MD: I’ve had multiple patients who came to get a second opinion from me. They were keyed up to get their immune therapy the next day because they had a PD-L1 of 70%. I say, “Wait, you’re a never-smoking Asian person who’s 50 years old.” So that was obvious. But there are others for which it’s not so obvious. And again, a couple people had ALK. That would have been a very bad choice for them to go on single-agent immune therapy.

But getting back to your key question, for patients without driver mutations, if they have a PD-L1 of 1% to 49%, I give them chemoimmunotherapy unless there’s an absolute contraindication for them getting chemotherapy; and a very frail patient; or where we are in California, lots of people are absolutely refusing chemotherapy. And I can be pretty persuasive.

Mark A. Socinski, MD: I would hope so.

Heather A. Wakelee, MD: But sometimes we end up in that setting. We will talk about the data. I will tell them that with immune therapy alone, to me, if they’re not higher than 50%, the probability of that working is quite low.

Mark A. Socinski, MD: Yes. One of the problems I have with that is, to Corey’s point, those patients wouldn’t have been allowed on the trial.

Corey J. Langer, MD: If they’re really frail, they would have been PS2 [performance status 2], most likely, and they theoretically could not have been enrolled.

Mark A. Socinski, MD: And so, I worry about the activity of immunotherapy as monotherapy in this population.

Heather A. Wakelee, MD: I’m thinking of a patient from last week. She’s one of those patients where when you meet her, she’s not looking good. But the next week, she’s looking much worse. You can almost see the cancer growing, right? And so, she was someone for whom we were waiting for the driver mutation. She was a never-smoking woman in her 60s. She was tested outside, so I only had my ALK and ROS and EGFR, and those came back the day before I had her set up for chemoimmunotherapy. Her performance status was definitely a 2 by the time we started, but she was someone for whom I knew that her performance status had gone down solely from her disease. So just like in the small cell setting where we’ve known for a long time that they’re going down because of the disease, you’re going to push your chemotherapy window beyond what you might do in a trial. But for other folks who are older, more frail, that’s where I think about immunotherapy in that 1% to 49% group. But I tend to try to encourage chemoimmunotherapy in that setting.

Mark A. Socinski, MD: Yes. And just to be aware of the exploratory analysis in the 1% to 49% group, it did not achieve statistical significance. And I still struggle with this cut point of 20%. I don’t know if our pathologists are that good at figuring out whether it’s 20%. That seemed to be if you were higher…that hasn’t influenced my practice. Any others?

Paul K. Paik, MD: No.

Corey J. Langer, MD: No.

Heather A. Wakelee, MD: And there are some patients, we’ll get to this a little bit later, for which PD-L1 for those particular trials was the marker. But we know it’s not the whole truth.

Mark A. Socinski, MD: Right.

Heather A. Wakelee, MD: It’s not like when you find an EGFR mutation and that’s the truth. There are shades of gray here that we don’t fully understand. So there are some patients who are still going to have a response who don’t have PD-L1, and that’s where, in that 1% to 49% expression, I don’t feel like it’s a terrible choice in the right setting, but with the full disclosure conversation.

Mark A. Socinski, MD: Right. I’ve always been concerned about what I refer to as the early dip phenomenon. In each of those Kaplan-Meier curves on KEYNOTE-042, all of those different subsets, there’s an early disadvantage to immunotherapy. I just don’t know how we identify those patients, from day 1, to do that.

Transcript Edited for Clarity