Transcript:Mark A. Socinski, MD: We alluded to this when we were talking about the PACIFIC trial and the population with EGFR mutation. One of the things we’ve learned is that, you know I-O is more of a drug for smokers. With all these mutations we have—and we haven’t seen a lot of activity of these immune agents in patients with driver mutations—there have been a lot of emerging data. Your thoughts on that data?

Leora Horn, MD, MSc, FRCPC: It has been a little bit all over the place. The first data that we had were from Justin Gainer, MD, in the EGFR/ALK-positive patient cohort, showing around a 3% overall response rate, which was lower than what we had seen from the subset of the CheckMate, KEYNOTE, and OAK trials. At the American Society of Clinical Oncology [ASCO] 2018 Annual Meeting, we had the French data, again, showing lower overall response rates [ORR] of the immune-target registry, which is patient data looking at multiple driver mutations.

That data showed the ORR to be a bit higher if you had MET mutation; although we know MET mutation occurs in smokers, it’s not like EGFR and ALK mutations, where it’s always nonsmokers; the ORR was a bit higher in patients with ROS1 mutation as well. But the patients with EGFR mutation were around 7%. That data are going to continue to emerge. They are telling us that patients with driving mutations will benefit most from targeted therapy. I-O therefore belongs somewhere in second or third line after we’ve exhausted these options.

Mark A. Socinski, MD: To me, it’s not a first-line option; it has ZOSTER tyrosine kinase inhibitor [TKI] possibilities. One of the issues, getting back to where we started, with the combination of chemotherapy and I-O, is do you administer this to patient populations even if the EGFR and ALK mutations were excluded from the trials? We’ll receive more data regarding the use of I-O in this population. These 2 presentations were very helpful in terms of framing the discussion around I-O in these driver populations. Thomas, your thoughts?

Thomas E. Stinchcombe, MD: Leora is correct, you should exhaust the TKIs. I-O does have a role, but it most likely would be used in combination with chemotherapy in these settings. The enthusiasm for I-O is so great that we sort of ignore selecting patients who are unlikely to benefit; that’s the next step, so that patients don’t get an ineffective therapy.

Mark A. Socinski, MD: Right, but it’s complicated because, as you pointed out, the driver mutations tend to occur, like BRAF, for example. When you look at the registration data, 60% of patients were smokers. That might be a subset in which there might be more utility. There is, of course, KRAS mutation, which we don’t have a good answer for at this time.

Leora Horn, MD, MSc, FRCPC: That one is easier because we don’t have any good drugs at the moment. The hard part is telling patients, because your PD-L1 [programmed death-ligand 1] comes back quickly—you normally get it with your immunohistochemical data. I don’t care that your PD-L1 is 90%.

Mark A. Socinski, MD: This is a great point. I made a slide recently that says, “Genotype trumps everything.” The problem is, you know your PD-L1 first, but you have to resist that temptation to act on it.

Transcript Edited for Clarity.