Ticiana A. Leal, MD: I think the PD-L1 [programmed death-ligand 1]—negative population is certainly the most challenging population. If you look at KEYNOTE-189, though, I think as a standard treatment option, the PD-L1–negative population did seem to benefit from chemotherapy in combination with pembrolizumab. And certainly, if you look at KEYNOTE-407, similarly, they also benefitted. So I think for the PD-L1–negative population, in general, as of now, those are patients who do seem to benefit from chemoimmunotherapy, although we realize and recognize that they don’t do as well overall. There are a lot of questions about: What is the optimal therapy for the PD-L1 population? What is the optimal biomarker? And do they have another marker that does suggest there’s a benefit with immunotherapy, talking about TMB [tumor mutation burden] and others that are coming along?

Mark A. Socinski, MD: In my opinion, are there substantial preclinical data looking at the importance of VEGF as involved in the tumor microenvironment and immunosuppression? And of course we had what I think is probably the most impressive trial lately, IMpower150. I’m just kidding.

Mark A. Socinski, MD: But that kind of tested the hypothesis. Just so everyone knows the design, the control arm was essentially ECOG 4599—carboplatin/paclitaxel/bevacizumab. And then the 2 investigational arms were either an add-on strategy, which you take advantage of the anti—PD-L1, anti-VEGF effect, or the substitution strategy. To make a long story short, it was a positive trial. It met PFS [progression-free survival] and OS [overall survival]. I guess since it did get incorporated in the NCCN [National Comprehensive Cancer Network] guidelines, it was guideline changing. I don’t know if it’s practice changing, for various reasons we can discuss.

But there were a couple of subset analyses that were particularly interesting, that were preplanned. One was in the EGFR/ALK space, and the other was in the presence of liver metastases, which was based on the observation from ECOG 4599, that there was a substantial benefit due to bevacizumab in the liver metastases population. So that was a special subgroup here. I might be interpreted as being biased by discussing more of the results. I just want to get other people’s thoughts. Let’s start with the EGFR stuff, right? Because this is the only trial to date that’s included EGFR.

Heather A. Wakelee, MD: It is. Well, IMpower130….

Mark A. Socinski, MD: I’m sorry, yes.

Heather A. Wakelee, MD: Right? And so, that’s an interesting compare and contrast. For IMpower150 in the EGFR patient population, there definitely was an overall survival benefit that was shown, again, in a subset, albeit planned, but it was still a subset analysis. And so I think we have to be a bit cautious in how to interpret that. I have given that regimen to some patients, and I’ve had at least a couple of them who do seem to be benefitting from it. And so, I think there is something there that needs further exploration. We also need to talk about the IMpower130 trial, which was the same carboplatin, and the same nab-paclitaxel, and the same atezolizumab, plus or minus atezolizumab, but without the bevacizumab. And that also allowed for the EGFR patient population. In both cases, it was a small percentage, 10%, 20%. But that was completely negative for the EGFR patient population. So the 3-drug regimen didn’t cut it. You needed the bevacizumab. Is that just because of random chance from 1 trial, or is it really a biological impact? I truly believe there is a VEGF interaction in the setting of EGFR, and we’re going to see the RELAY trial presented at ASCO [the American Society of Clinical Oncology meeting]. We have erlotinib, bevacizumab, and lots of preclinical data. So maybe, but I’m still trying to understand it fully.

Mark A. Socinski, MD: My favorite slide at Santa Monica this past year was Dr John Heymach’s slide that said the best biomarker for VEGF therapy is EGFR.

Heather A. Wakelee, MD: The EGFR mutation, right. And John’s work has been, he’s really been leading a lot of that, too.

Mark A. Socinski, MD: That’s right. Corey?

Corey J. Langer, MD: If you go back and look at your trial, IMpower150, the separation in the PFS curves are actually a bit more pronounced beyond the median. Whether that’s just serendipitous or that is a reflection of potential synergy between checkpoint blockade and angiogenesis inhibition… The chemotherapy is done at that point.

Mark A. Socinski, MD: Right.

Corey J. Langer, MD: It was 8.3 versus 6.9 months. The chemotherapy would be done by 4 to 5 months. I think that adds further speculation about the potential interaction. And frankly, the hazard ratio is 0.55. It was even lower for those who were in the TKI [tyrosine kinase inhibitor]-refractory group, where, frankly, I have actually applied this regimen. In the real world, in clinical practice, I have used the 4-drug combination. Ironically, it’s not approved in that population, or at least not formally approved.

Mark A. Socinski, MD: In the United States.

Corey J. Langer, MD: In the United States.

Mark A. Socinski, MD: Outside the United States…

Corey J. Langer, MD: The rest of the world is a little bit more reasonable, in this regard. But at least in the United States, it’s not. But based on those data, based on the hazard ratios, and based on the publication, I’ve managed to get insurers to cover it.

Transcript Edited for Clarity