Transcript:Benjamin Levy, MD: Let’s move on to the last segment, and that is the role of tyrosine kinase inhibitors in squamous cell carcinoma—and that may dovetail into a conversation about an unenriched group of patients that do not have EGFR mutations and whether there’s a role for TKI therapy in this group of patients. I’ll set the historical perspective and then maybe Ed can talk about the LUX-Lung 8 trial.
You look at the data for TKIs in an unenriched patient population. My take is, it’s been mixed. We have the BR-21 study, which looked at erlotinib compared to best supportive care in a second- or third-line setting in patients with both adenocarcinoma and squamous patients, and showed a survival advantage. We have to be careful with subset analyses, but there was a subset analysis in the squamous cell population to show that benefit and survival did carry over.
We had that study showing that TKIs may be beneficial, at least compared to best supportive care, in a squamous cell patient population. We have the SATURN study, which was a maintenance erlotinib study after 4 cycles of chemotherapy, patients randomized to erlotinib versus best supportive care. And in the squamous cell patient population, we did see that the PFS advantage shown in the intention-to-treat population did carry over into the squamous cell patient population. We have to be realistic about what we can glean from subset analyses that aren’t powered to show such effects.
I guess counter to that has been the TKI-vs-the docetaxel story in the second-line, and we saw data from the TAILOR trial showing superiority of docetaxel over a TKI for a specific group of patients who are EGFR wild-type. And in the subset of patients who had squamous, there was a benefit with docetaxel that approached statistical significance. I would argue that the benefits seen have to be weighed in the context of toxicities or the discrepant toxicities between these two drugs. Nevertheless, it seemed in the TAILOR, and even the DELTA study, subset analyses from the squamous cell patient population seemed to favor cytotoxic chemotherapy over TKI. So a lot of conflicting data.
This is a contentious area: to use or not to use TKIs in patients that don’t harbor EGFR mutations. We’ve come a long way in molecular testing in delivering TKIs up front for these patients. The question is, do these drugs have any clinically meaningful benefit in a patient population that is not historically harboring EGFR mutations, particularly squamous cell. Lots to cover. Ed, maybe you can talk about the LUX-Lung 8 trial and then maybe your perspectives on the use of this drug.
Edward S. Kim, MD, FACP: I think you get a nice overview there. I’ll add that being the PI of INTEREST—which was a gefitinib-vs-docetaxel 1400-patient study that showed 7.9-month survival versus 8.1 in a noninferiority fashion—wasn’t good enough to get that drug approved. We have to go back and thank ISEL for everything we have today. And why do I say that? ISEL was a negative study that was gefitinib versus best supportive care, and Nick Thatcher reported that and it missed by two weeks in the subset analysis. And what that chain of reaction caused was that IPASS—which was the first-line gefitinib-vs-carboplatin/paclitaxel study that was going to be a global western trial—because of the subset indications found in ISEL, the negative, the never-smokers, and the Asians, got shipped to Asia. That’s why the biomarker component of IPASS was exploratory; it was just in Asian female never-smokers, and we got lucky. Awesome. That leads us to all these head-to-heads that are ongoing. We have three drugs in this space right now—gefitinib, erlotinib, and afatinib—and now even Tagrisso (osimertinib) has some coverage in this area as well, although we wouldn’t use that up front right now because it targets T790M.
But LUX-Lung 8 was a large study: 795 patients in squamous, afatinib versus erlotinib. It was worldwide. Jean-Charles Soria presented his data, and everything was a little bit more favorable in the afatinib side. And this is a pattern that we’re seeing, maybe because it’s a newer-generation drug. We’ve seen this, as well, when afatinib goes against gefitinib. And it goes to show you that a drug that was thought to be very toxic that no one could tolerate at 40 mg is pretty tolerable, and dose reducing to 30 or to 20 is simple, just like what we do with erlotinib going from 150 to 100 to 75. When you see these incremental benefits head-to-head in a population that’s supposed to be equal, then I think it makes us open our eyes a little bit because that might be the little sway that you choose one drug or another. We frankly are impressed with the deletion 19 data with afatinib. We don’t think that’s not the case with some of the others. We’re just asking for the data and want to see, and we haven’t seen it yet. And so the more data like this, like LUX-Lung 8 and LUX-Lung 7 that begins to accumulate, you start thinking, well, maybe afatinib is a preferred drug up front, especially in deletion 19 patients.
Benjamin Levy, MD: Paul, your thoughts. Clearly, this is controversial. I have been on the side of using TKIs as a second- or third-line in an unenriched group of patients who can’t tolerate chemotherapy. And for my practice, that’s been a fair amount. Nevertheless, a lot of conflicting data. What is your experience with afatinib in the squamous cell population, or even TKI, and should it be something that we consider at some point?
Paul K. Paik, MD: It’s probably easiest, because of the data, to address LUX-Lung 8 specifically. And I’m glad you brought up ISEL, and you brought up BR-21 before, because that exercise that people did through ISEL and BR-21 looking at subsets, clinical characteristics that trended toward improvement or had significant improvements in certain subsets, that make sense now in retrospect; women, Asians, never-smokers. It’s important to know that the LUX-Lung 8 population was not homogeneous. There was about a 5% to 6% rate of never-smokers, of nonpure squamous lung cancers, adenocarcinomas. And if you look at the subset analysis in terms of clinical characteristics, the improvements trended toward favoring these subgroups, much like they did for ISEL and BR-21.
And in terms of the table,-one thing, that’s your first foray into asking, “Well, what really is going on here?” And then you get to the discussion where they report a 6% EGFR mutation rate in each arm, the afatinib and erlotinib arms. You have to wonder, well, okay, there was a 6% EGFR mutation rate in each arm. What exactly is going on? You had a 6% overall response rate for the afatinib arm and 3% for the erlotinib arm. Is the benefit being driven by this subset of patients? The discussion mentioned that because EGFR mutations were balanced, this did not explain the difference in efficacy between afatinib and erlotinib. I would argue that’s not the right question to ask.
In this day and age where we’re used to—and we’ve talked about this at length during this session, picking a biomarker and then a drug—it stands to reason that we would like to see data that show what the response rates are in that population that did not have that alteration. In other words, is the benefit overall just being driven by that small cohort that had EGFR mutations? Until I see those data, the LUX-Lung 8 study is in a gray area.
I think Ed is right. In aggregate, where we like to be dogmatic, there were no classical EGFR mutations detected in TCGA’s analysis, and in the sequencing we’ve done, we really have not seen any great frequency of EGFR mutations. The pathway is still there; amplification still occurs. There’s efficacy with necitumumab. There was with cetuximab. There is probably something that’s there. We just don’t know what it is. That leg work still needs to be done for a drug that’s going to be expensive [and] whose benefit is still not that great overall.
And I’m glad you mentioned TAILOR and DELTA in terms of chemotherapy versus EGFR TKIs. At best, this shifts a TKI down the road. I have not been using EGFR TKIs as a result. We’ve been sequencing squamous patients now routinely for about a few years, so if I detect an EGFR mutation, which we have on occasion, then we’ll offer it pretty soon in terms of the treatment course. But absent that, I have not been routinely giving a TKI. Part of this is philosophical, part of it is not, but that’s sort of the way that I’ve come down on it.
Edward S. Kim, MD, FACP: Really where the TKIs have gone in squamous tumors has been kind of like how before we had mutations in non-squamous ones.
Paul K. Paik, MD: That’s exactly right, yes.
Edward S. Kim, MD, FACP: It’s been that [patient with] poor performance status, can’t give them chemotherapy, getting close to hospice, and let’s give it to them there. And that’s where they’re subsetted right now.
Benjamin Levy, MD: So both of you would fall in line with using a TKI after all other options are exhausted or a patient who has not tolerated chemotherapy?
Edward S. Kim, MD, FACP: I think you have checkpoint inhibitors and you have docetaxel/ramucirumab in patients who can tolerate that. Would I toss up single-agent gemcitabine versus a TKI? Probably at that point because we’re in third- or fourth-line.
Benjamin Levy, MD: I would agree. I think that in light of all the newer options we have and drugs that have shown survival advantages, TKIs have been bumped for me. I have had some clinical scenarios where I’ve used them as third-line. I have, just for instance, a patient who got beat up on chemotherapy frontline and swore to me and himself he would never go back to chemotherapy. We gave him immunotherapy. He’s doing very well. He keeps telling me once this drug, nivolumab, stops working, there’s no way I’m going back to chemotherapy. And in that circumstance, we can consider giving him a TKI and offering something meaningful if he still wants to take it.
Transcript Edited for Clarity