Everett Vokes, MD: You mentioned a number of trials that are very information to this. Naiyer, I wonder if you could just go over those a little bit more in detail? When should pembrolizumab, alone, be used? There’s an alternative of combining pembrolizumab with chemotherapy. Should nivolumab be used? Is it similarly active in this setting? What do we know?

Naiyer Rizvi, MD: In terms of first-line therapy, and I totally agree with Fred that it’s really important to have PD-L1 resolved, upfront. I’m sure Ram Suresh would agree. And I think the data is that if you give pembrolizumab first, and then you get chemotherapy later, for example, versus if you get chemotherapy first and crossover to pembrolizumab, your survival is better if you get the pembrolizumab first. So, I think that data is really compelling in that we should get a biopsy. For our newly diagnosed patients, as Fred said, the EGFR and the ALK are patients who were not included on the pembrolizumab fist-line trial. And not uncommonly, we see patients with results that come back that are PD-L1-positive and EGFR-mutation positive, for example. I think everyone here would agree that the EGFR mutation results trumps the PD-L1 result, and you should get a targeted therapy.

Pembrolizumab, first-line in more than the 50%-positive patients (which is about 25% to 30% of patients), is clearly ensconced in first-line therapy. The chemotherapy combinations are a little more controversial. The data around that is the KEYNOTE-021g trial, where patients were randomized to pemetrexed, carboplatin, and pembrolizumab versus chemotherapy alone. It was a 60-patient per arm phase II randomized study, and I think the data are very compelling that there’s a progression-free survival advantage. There’s also an overall survival advantage, updated at the European Society for Medical Oncology (ESMO) 2017 Congress, with a hazard ratio that continues to improve.

So, I think there’s definitely something to the story. But, my stance is that we think there are phase III trials, which are imminently apt to read out—one being KEYNOTE-189, which is the same study design as KEYNOTE-021g; also, the IMpower150 study, which is the atezolizumab chemotherapy, bevacizumab trial. So, I think that first-line therapy, pembrolizumab, for sure (of the chemotherapy combinations). I tend to do it a bit on an ad hoc basis. Certainly, for a patient who has significant disease burden and fairly explosive disease, I’ve done that a couple of times because the time to response with the combination is faster. So, I think that if you’re really nervous about the patient, it’s reasonable to give them the combination.

Fred Hirsch, MD, PhD: Naiyer, can I ask you a question? KEYNOTE-024 was based on the PD-L1 assay. The chemotherapy combination is for all comers, independent of PD-L1 expression. Currently, with the data we have, would it be a reasonable approach, for patients with high PD-L1 expression, to use pembrolizumab as monotherapy? And for patients below 50%, would it be reasonable to offer pembrolizumab and chemotherapy still under the condition that they don’t have another driver? Is that the paradigm which is reasonable?

Naiyer Rizvi, MD: I think it’s an approach. I’m not ready to change my practice for chemotherapy combinations, upfront, based on the small trial that was done. So, I kind of do it ad hoc at this point. I don’t know what others have experienced.

Suresh Ramalingam, MD: In that study, admittedly with the small sample size, you don’t have the ability to slice different PD-L1 subgroups. But, when they look at the response rate based on PD-L1 expression, in the less than 50% group, the response rate was actually similar, if not lower, for the chemotherapy plus pembrolizumab group compared to chemotherapy alone arm. The biggest benefit was seen for the combination in the highest PD-L1 expressing group of patients. So, it’s hard to, at this point, make the case that if your PD-L1 is low, you get chemotherapy plus pembrolizumab. If the PD-L1 is high, you get pembrolizumab alone. Fred, as you and Naiyer mentioned, you can only learn so much from a 120-patient trial, regardless of how many updates and subsets you do. We really need to look at the phase III trials that hopefully will lead out very soon.

Everett Vokes, MD: I agree with that. But, looking at the trial, weren’t there 2 segments where 1 was exactly like you say—I think maybe between 25% and 49%, or 10% and 49%? And then, there was a lower group where there was, again, a difference in favor. I know we’re dissecting a small trial into too many subgroups, so it is certainly fair to say the question is unanswered. But, you get into a question of, what does this mean in that setting? Taking your hypothesis, that that is the group of patients that should get chemotherapy and pembrolizumab, then you could actually argue (as some institutions are now doing), to not do PD-L1 testing and simply give everybody chemotherapy and the PD-1 inhibitor. I think Naiyer, certainly, wouldn’t go along with that? And I take it that Ram [Suresh], you wouldn’t? I wouldn’t either. So, it’s really an area of confusion, I think.

Suresh Ramalingam, MD: I also want to point out that this is all for non-squamous patients. For squamous patients, we don’t have data on chemotherapy plus PD-1 or PD-L1 at this point.

Transcript Edited for Clarity