Mark A. Socinski, MD: In my mind, in KEYNOTE-189, we have the control arm with a median survival of 11 months. And then, in IMpower150, we have the control arm with a median survival of 14 months.

Benjamin P. Levy, MD: Remember, those are nonsquamous patients.

Mark A. Socinski, MD: Right. My point is that not everyone should get bevacizumab because of the contraindications. And so, it’s hard for me. I think this is a situation where it’s very difficult to make cross-trial comparisons.

Benjamin P. Levy, MD: It is.

Karen Kelly, MD: It’s very difficult.

Mark A. Socinski, MD: If not impossible.

Corey J. Langer, MD, FACP: There are different populations.

Karen Kelly, MD: True. There are different populations. I think this is the one place where cross-trial comparisons are really not going to be helpful.

Mark A. Socinski, MD: Right.

Karen Kelly, MD: We’ll be doing a lot of cross-trial comparisons because we think that the patient populations are similar, like in KEYNOTE-042 and KEYNOTE-189. But in this setting, they’re not.

Corey J. Langer, MD, FACP: This is really the only trial that I’m aware of, certainly that’s been reported, that’s included EGFR and ALK. They were explicitly excluded from KEYNOTE-024. They’re excluded from KEYNOTE-189. They were excluded from our KEYNOTE-021 cohort G pilot randomized phase II study. I believe they were excluded from KEYNOTE-042, and certainly from the pembrolizumab trials. So, the final common pathway for patients with oncogenic drivers remains to be chemotherapy. Eventually, the TKIs will stop working. This regimen then becomes, I would say, the standard of comparison in that population.

Karen Kelly, MD: I think the fear, though, is that the sequencing is not going to be done upfront like it should be, and that these patients will just be placed on the triplet combination. I really want to emphasize that, again, the standard of care is to do the testing and then make the treatment decision. As you pointed out, Ben, we may not be doing testing in anybody. People may not continue to do testing. So, I think it’s really important that people remember the workflow.

Benjamin P. Levy, MD: And that TKIs come first.

Karen Kelly, MD: TKIs come first. It’s a different set of patients. You’re talking about upfront versus pretreated patients with TKIs. So, the workflow process is a little different.

Corey J. Langer, MD, FACP: Sometimes, though, the dilemma comes up that PD-L1 results come back first. You’re waiting for EGFR and ALK. I’m sure that all of us have dealt with this situation where the PD-L1 result is astoundingly positive,70%, but they also happen to be EGFR-mutant positive, or ALK-positive. The oncogenic driver, I agree, trumps chemotherapy or chemotherapy/immuno-oncology combinations in that population.

Karen Kelly, MD: Right, but you don’t want to put them on this if they’re 100%. I’ve had a patient who’s had 100% PD-L1 expression and also had an EGFR mutation. She did not do well on immunotherapy. We’ve seen this, time and time again. Even if you get that result back, you have to wait for those other results.

Corey J. Langer, MD, FACP: I agree.

Benjamin P. Levy, MD: We’re going to learn about the role of single-agent pembrolizumab for an EGFR-positive patient with a PD-L1 greater than 50%. It’s small numbers—8 patients. It’s a prospective study. The response rate is 0.

Mark A. Socinski, MD: I tell my referring doctors that PD-L1 should only be considered after you know the results of EGFR, ALK, ROS1, and BRAF. That’s the NCCN recommendation. Those drivers trump PD-L1, and I think that, as Ben just pointed out, the real effectiveness of immunotherapy in those oncogenic driver subsets is less than what you would expect.

Transcript Edited for Clarity