Amgen Inc. (AMGN) Earnings Call Transcript & Summary

Well, good morning once again, and thank you for joining us for our second virtual oncology day. I'm Yaron Werber, biotech analyst at Cowen. And it's a great pleasure to moderate the session today with Amgen. The 2 gentlemen here really need no introduction. We have David Reese, who's EVP of R&D; and Arvind Sood, VP of IR. Gentlemen, thanks so much for joining us. We really appreciate it.

Good morning.

There's lots to talk about, especially the Lakers and the Celtics, but maybe we'll leave that for the last 2 minutes because that's going to be -- that's an ongoing battle.

So let's start first by -- there's a lot of data coming up, obviously, on LUMAKRAS at ASCO. The latest abstract from CodeBreaK 100 in non-small cell, OS has not been reached yet. Obviously, the study is continuing to accrue events over time. The subgroup analysis actually was very informative. So in the STK11 mutant, the response rate was about 40%. It was about 39% on the wild type. So you're retaining nice activity. But we also now saw that KEAP1 mutants do have an impact, right? So the patients who have concomitant KEAP1 mutations, the response rate was 20%. If they're a KEAP1 mutant -- if they're KEAP1 wild type, it was 44%. And if you looked at the STK11 mutants that didn't have KEAP1 mutations, the response rate was 50%, right? So with all of that in mind, as you think now about bringing LUMAKRAS to first line and you're going to go into an STK11 mutant population, presumably is monotherapy, how do you think about trial design and stratification?

Yes. No, it's a great question. And maybe we can start first at a higher level, which is as you see in the abstract, the mutational profile in terms of predicting response or resistance is complex. There's probably no simple answer here. I think our sense now, and you'll see more data on a few days at ASCO, but our sense is that this is not an analog to the receptor tyrosine kinases and kinase inhibitors, where you can often predict the gatekeeper mutation that will confer resistance, for example. This is -- there are going to be multiple different hits on parts of the pathway. It's also important that simply assaying a single gene outside of KRAS is probably not enough in the STK11 in KEAP1 story is very informative there, as you can see in the abstract. I won't repeat the details, but knowing both of those is important as a guide to likelihood of response. Nevertheless, patients with all of the patterns do -- some patients do respond with all of the patterns, indicating that the story is even more complex than that. And to me, in the field, that is really the fundamental next task, is going a layer deeper, figuring out the circuitry and really trying to come up with signatures of response and resistance. And that's the focus of a lot of our effort right now.

So as you think about -- you're currently, I believe, in discussions or you're planning on being with discussion with the regulatory authorities. Is there a path forward with focusing on a certain mutation like an STK11 mutation and concomitant mutations into first line? Is the agency open to that?

I think potentially, although I think one thing that is very likely is given the ubiquity and the belief in checkpoint inhibitors in first line, ultimately, it's very likely some sort of randomized trial would be required. That will be the focus of ongoing discussions and development. I think it's probably premature to draw any firm conclusions.

Okay. So maybe in specific -- I mean when you look at the data on STK11 mutants, they don't respond as well to PD-1, right?

Right.

So maybe there's still a randomized trial with a PD-1 against the PD-1. Is there room to do a LUMAKRAS or a single-agent activity, almost like a...

Yes, there might be. One could even envision potentially a 3-arm trial or a trial with sequential therapy where you've got, for instance, induction therapy for a period of time with LUMAKRAS. So these are all the sorts of things that we're thinking about as we go forward in the development program.

Yes. And as you think about induction, what does induction mean to you? Is it -- I think it's -- I think to me, induction means you start with LUMAKRAS whether it's 2 cycles or 3 cycles. I guess the question is what happens after that. Do you add KEYTRUDA on and you keep LUMAKRAS on? Or is it -- are you inducing and then switching...

Yes. I think there's potentially interest in both approaches. And we're -- as you know, we're looking at how to optimize use either with or sequentially with checkpoint inhibitors, and that's one of the things that we need to sort out as well as the durations or what we would call schedule. How long do you use a one agent followed by another, for example.

And based on the preclinical models, how long do you think induction should happen with a KRAS inhibitor? Is 1 cycle enough, 2 cycles, 3 cycles?

Yes. Our sense is that pretty quickly the biologic rationale here, if we step back for a moment, is that we've got preclinical data and some of this was published in our Nature paper from a couple of years ago, suggesting that you get upregulation of PD-L1 upon treatment of KRAS G12C mutant tumors with LUMAKRAS. So -- and the sense is that, that happens relatively quickly, meaning within days to weeks as opposed to many months. And so that would indicate that you don't need a 6-month induction period how to optimize that is something that you could only really figure out empirically in the clinic.

Okay. Another question that I have relating to the subset sort of analysis. When we now go back to the World Lung data back in the fall, specifically looking at PD-L1 expression, the ORR was 48% in the TPS low, where it defined is less than 1%, was 36% in the TPS 1, equal to or greater than 1%. How does that also kind of fit into your thoughts about first line with checkpoint inhibition. Typically, if you have PD-L1 high, right, you can do with -- you can get away with KEYTRUDA or PD-L1 or PD-1 alone. If you're in a low, you need chemo. So I guess there's many questions we can go from there.

Yes. So obviously, one interest we have is a combination with chemotherapy alone in patients who would not be eligible for checkpoint inhibitors because of a contraindication or because they're predicted to have a very low likelihood of response. So again, I think this is going to be personalized medicine writ large as we go forward.

And what about -- is there a chance to, in a PD-L1 low to do a combination with a checkpoint and maybe replace chemo. I'm just thinking in -- if we look at the activity also whether you have -- to the point, if we have a PD-L1 low, the response is actually fairly high or higher than a PD-L1. Can you replace chemo and do a combination with a checkpoint but you still need to get to some kind of combinability with the checkpoint?

Yes. You still need to -- and again, the question is, could you -- do you even alter that PD-L1 expression upon treatment with the KRAS inhibitor. So again, a lot of questions. I'd remind everyone, we've been in the clinic 30, 32 months, something like that. It took 40 years to get into the clinic. And there are some outstanding questions yet in terms of optimization.

In terms of -- as you think about combo with a checkpoint, a G12C inhibitor, your data has been fairly clean. There's been a little bit of GI here and there, really Grade 1, 2. As you think about KEYTRUDA, what are some of the areas where you need to kind of watch for overlapping in terms of some side effects?

Yes. I mean, I think, if you look at the side effect profiles of the single agents, where there is overlap is obviously where we pay attention. But development of combination regimens in oncology is obviously something that we've been doing for many decades now. And it is -- our sense is that given that the single agent or monotherapy profile is so clean that, that can be a guide and where we've got potential overlap is where we'll pay real attention. But again, these are questions that will only be answered in the clinic.

And so the next step, it sounds like you're still working on dose and schedule on the combo, if I remember correctly from the last conference call?

Yes, I think that's a fair conclusion.

And some of it, as you mentioned, is increasingly, it sounds like you're moving toward a sequential regimen, too, where you really see the benefit over time. The -- when you look at that in context of all the other combinations that you're looking at, via potentially a SHIP2 to or an EGFR, what's going to be the trickiest area for you in terms of overlapping tox? I mean each one has its own...

Yes. I think they all have their own profiles. And it's really -- there's no generalized statement that we can make there. It's going to be based on the adverse event profile of the combination partner.

And when you're looking at combo, do you have a sense, do you need -- how much do you need -- how much coverage do you need on the SHIP2 access?

Yes. I think that's a great question, and that's one we're working hard on now because some of -- in some of these combinations, you may only need to sort of tickle one or both of the targets to get a real synergistic effect. And so that's one thing that we're working out. And that obviously affects thoughts about how you dose in combination. And of course, there's -- in oncology, routinely, we vary doses in combinations depending on the efficacy and adverse event profiles.

And so maybe based on the preclinical data and maybe there's obviously some data now in the public domain from Novartis' SHIP2, there was an appreciable single-agent activity, and I think they're still looking at the dosing schedule even in monotherapy. They did say that you need to get to at least 20 mg per kg to adequately cover the pathway based on the biomarker, but that's assuming adequate coverage. Any sense based on preclinical data, how much SHIP2 coverage do you really need to have?

Yes. I don't know that we know enough to actually answer that question right now. It's a great question, but I couldn't honestly tell you how much SHIP2 coverage, we think, we need in combination. And it's a hard -- it's a very hard question to answer biochemically.

And why is that, in what sense?

Just the nature of the assays, especially when you start looking at combination regimens in cell line data, for example.

And as you think about combination regimens is the fact that it looks like at least one SHIP2 inhibitor doesn't have appreciable single-agent activity, right? It was a 20% stable disease. And I think the doses that they're using or even the doses potentially are going to be lower in combination. Does that sway your enthusiasm for this pathway? Or what do you learn from this data so far?

I think what we learned is what we've at least seen reported in the abstracts, which is not really any monotherapy activity. And so the outstanding question is, do you get something in combination? Will it add to KRAS inhibition alone? And we'll answer that in the clinic. I think biologically, there's still possibility there, and it's absolutely worth testing.

And SHIP2 is fairly high up peripherally, right? fits into the MAP kinase pathway. Based on what you know, and I recognize you might not know a lot so far, when you think about escape mechanisms to a G12C, how does a SHIP2 potentially shut down emergent resistance? Or how does the synergy really happen? Because that will fit into a question of how much coverage you really need to have.

It may help if you're getting secondary mutations on KRAS itself, for example. So where you've now generated, for instance, partial resistance or a clone. And so a downstream hit, you may help in that instance. And then, of course, we're looking further downstream in the pathway as well, for example, with MEK inhibition, which -- there are alternative or bypass signaling pathways that can be used to ultimately activate the MAP kinase pathway. And with a downstream hit, you may help shut off some of those as well.

And so maybe let's talk about MEK because I think on the MEK combo, you were able to get through the high dose of LUMAKRAS.

Yes. Exactly. And so we're moving forward there. Now we really just need to ask a question, do we get efficacy that we think is enhanced beyond single agent. And the MEK combination is obviously of interest in tumors outside of lung cancer as well, such as colorectal cancer.

Yes. And for lung cancer specifically, MEK has not really been used in lung cancer, right? Certainly not as a single agent and hasn't really been successful in combination so far. So is the activity again going to be probably in shutting down resistance? Or is there potential synergy with the KRAS otherwise?

My guess is it's probably more the resistance pathways. And if you've got bypass tracks that are coming through the non-receptor tyrosine kinase pathway.

Okay. Got it. A question also, and I know we're kind of very systematically covering the whole pathway here. When it comes down to pan-erbB inhibition, the challenge has been, just with this approach in general, is the toxicity. The fact that it's got downstream modulation could be fairly attractive. I guess the same question comes with the CDK4/6, there, the therapeutic window potential is better. How do you think about those 2 relative to each other?

Yes. Well, it's very hard to speculate relative to each other. I think with the pan-erbB inhibitors, the question will be safety and combinability. And obviously, we're going to answer that in the clinic as well. Because there you get both HER2 as well as EGFR-related adverse events.

I'm going to shift over. Let's shift over to colon cancer. There, the combo you're moving into, we're still going to see the updated mono data and now you're moving to combo as well, including EGFR antibodies. Are you using the high dose of LUMAKRAS in that regimen?

We fully enrolled a Phase II trial, which was done using the 960-milligram dose that we filed for in non-small cell lung cancer. So we'll have that readout in hand, and we'll be able to share the data over the course of the year, I think.

Okay. Let me move to bema and the FIGHT Phase II data in first-line FGFR2 positive patients. So there's a lot to talk about here as well. About 30% of patients who were screened ultimately had FGFR2b positivity. And 96% of those were by IHC, 17% were by circulating tumor DNA, about 13% by both. And I'll dive into the data in the second. The data for both was fairly robust, about a 2/3 reduction in a hazard ratio of 0.63 in IHC positive and even 0.1 in those who are both IHC positive and ctDNA positive. So the first question is how often are people doing ctDNA screening?

It's not routine right now, and that's obviously something that as part of the development program that we would want to introduce now. This is not some exotic new technology at this point. So it really just requires education. The immunohistochemical assay is very standard, the sort that pathologists are quite familiar with. And so I don't think that's going to be a barrier at all. And most of the patients, as you noted, will be included or deemed eligible based on overexpression as determined by immunohistochemistry.

And where do you think FGFR2 antigen expression fit in with respect to also intracellular FGFR2 amplifications or fusions. Do they coexist? Or are they exclusive of each other?

Well, amplification, of course, can coexist with overexpression. And then this is reminiscent of the HER2 story, which is where I did my initial training. So this kind of brings me full circle in my career in certain ways. The mutational tumors such as gall bladder cancer, for example, where mutations occur, that's a separate really entity molecularly, I think. It's one where that is clearly a driver mutation. It's a subset, small subset of these patients. I don't think you've got much overlap between those that have got overexpression in accompanying mutational status based on the data we've seen today. Obviously, we're going to generate a lot more data in terms of this sort of molecular profiling.

And so as you move into your pivotal that's going to be in combination with chemo, presumably FOLFOX combo versus FOLFOX alone in first-line FGFR2 gastric cancer. As you think about the landscape at the end, how is this going to fit in with PD-1? as PD-1s are beginning to move into first line?

Yes. So the PD-1s are coming in, there's clearly an effect. That's great for patients. And we had anticipated that, of course, as we were acquiring this program. I think there's plenty of room for both. Will there be some overlap between the populations? Yes. I would anticipate what some are calling triplet therapy. Is that something we'll investigate as well? So both bema plus, PD-1 plus the backbone chemotherapy. I think in many parts of the world for some time to come, chemotherapy alone would remain a standard. And so there will be the opportunity to add a bemarituzumab to those backbone regimens, for example, in many parts of East Asia, where the disease is highly prevalent. So -- but I think our guess is there will be a very comprehensive development program here. We've got to sit down with regulatory authorities now that we've completed the acquisition. That's just 2 or 3 weeks ago so -- but we are working furiously on those plans and getting in front of regulators to discuss the pivotal program.

And in terms of next indication, whether it's lung cancer or other FGFR expressing tumors, what comes to mind next? And how do you differentiate again where PD-1 is a standard?

Yes. I think squamous cell carcinoma of the lung is probably top of the list, given that that's where right now, there's the best evidence of -- appreciable frequency of FGFR2b overexpression. So I think that's top of our list right now. I would anticipate a broad Phase II signal-seeking program. There are other GI malignancies, triple-negative breast cancer, for example, where there's evidence that, again, a subset of tumors have FGFR2b overexpression. And so I think we will embark on a Phase II signal-seeking program to determine where there's promise for additional indications beyond gastric cancer.

Come back -- I'm getting questions from the audience. I'll come back to them in a minute. Just a quick question on keratitis that was seen in the FIGHT program. It is -- as you would expect, it increases, obviously, with duration of therapy. FGFR2 is expressed on the corneal epithelium. Are there any plans to introduce some prophylactic measures? What could those do and how long...

Yes, absolutely. And it's worth pointing out that in the FIGHT trial, there was no prophylaxis. And so our belief is that it's clearly, I think, an on-target toxicity. Because as you point out, this receptor is expressed in corneal epithelium, that's well defined. We think it disrupts the sort of normal physiology of that eye. And there are ways to restore that, for example, with lipid -- high lipid content eye drops and things like that. These are the sorts of things, prophylaxis measures that we will incorporate into the development program going forward. And these are -- we've talked to a large number of oncologists, of course, who were experts in the area. They view this as something that's manageable, that you have to be aware of and look for. But it's par for the course for oncologists in terms of dealing with adverse events.

But when you look at the data by week 25 and higher, it was at about 15% for Grade 3. Did they discontinue? Or did they take a treatment holiday? Or how does that work in the study?

Yes. Ultimately, many of those discontinued if they were developing Grade 3. Of course, we'll build in specific management programs into the development program going forward. And clearly, here being aware and I think it became clear as Five Prime learn more that if you recognize in the very early stages, developing keratitis and intervene, you could also have a real effect. So all of that learning, we're taking on board. But the plan would be to engage in aggressive prophylaxis and management.

Okay. Quick question. I want to move on to AMG 757, the DLL3 where the data, it will be updated again at ASCO for small cell lung cancer. So here, you're now looking at the 10th dose level. Recall, up to now, we showed about a 14% response rate. But now that you have the 10th dose level, you have unconfirmed PRs in about 5 out of 8 patients or 63%. Can you give us a sense the data at ASCO, is it going to be updated further in...

That will be updated. It will have the confirmed response rate, duration of response. I think the -- this is a program I'm keeping a very careful eye on. Our hope is over the coming months, we settle on a dose regimen or perhaps even comparing to and start thinking about and talking to regulators about what makes sense for this program going forward. I think as most of you are aware, third-line small cell lung cancer, there's no real standard of care. The prognosis is not good for these patients. There's a huge unmet medical need. And so my hope is that we can finish off dose optimization in the coming months and think about what the next steps are in this program. I think between this program and the prostate cancer program targeting PSMA, AMG 160, we have unequivocal evidence or proof of concept that we can generate durable responses in solid tumors, and that is exciting. So the question now is how do you just optimize outcomes with these agents.

Is the thought, David, in third-line small cell to go potentially head-to-head or really due to single-agent strategy or has accelerated and maybe a...

Yes. I think those are things we need to sort out. We need to have the right conversations with regulatory authorities. Again, there's certain -- globally, there's certainly beyond second line, there's no standard of care at all. And many patients actually don't traverse past second line because of lack of availability in the toxicity of existing treatments.

Okay. And then a question from the audience, it goes back to LUMAKRAS. In addition to, obviously, you filed -- and you obviously have Phase III is ongoing. Are there any other monotherapy Phase II studies that are currently potentially pivotal?

Yes. That's very hard to speculate on. I mean the next big Phase II readout is the colorectal cancer trial, which is a decent-sized Phase II study. I think that will give us a pretty definitive insight into whether there's a monotherapy path forward or whether combination therapy makes more sense in colorectal cancer.

And remind us that study is purely mono? Or does it have an arm with ERBITUX?

Yes, that's purely monotherapy.

And so the arm with ERBITUX is really now going to be a Phase III study?

The ERBITUX trial, that's a part of the -- EGFR inhibitor trial is...

Phase I.

Part of the master protocol. And just to remind everyone, that protocol has been designed so that each arm goes through a safety phase and dose-finding phase, and then can be expanded into what's, in essence, an expansion cohort that's basically a Phase II study. But it's -- they're single arm.

And that's -- okay, that's a part of the code, the initial CodeBreaK study. Okay. So that has not been expanded yet at this point, the ERBITUX LUMAKRAS is still part of the Phase I?

Yes. I don't think we've talked about what our plans are there yet.

Okay. I'm getting more questions. I'll tell you what, maybe just in the last minute or so, to wrap up with teze, we just saw some updated data at ATS that again, continues to look very good with NAVIGATOR. There was also data from the SOAR study. There -- it really does appear that trial design might have been the culprit relating to how long the patients try maintenance for and the overall treatment duration when the endpoints were measured. Are you thinking about running that study again with a slightly different design? And what's the timing for that?

Yes. So we're working with our colleagues in AstraZeneca on what the follow-up plans are. As you point out, just very quickly, we think there were methodologic factors that influenced the outcome. The period in which patients could be weaned was 36 weeks, basically double what previous trials had. So they -- in addition, patients -- or physicians could make multiple attempts to wean these patients off corticosteroids, whereas previous trials would limit that to 1 or 2 attempts. And then finally, the placebo response rate was 46%. Those first 2 factors probably contributed to that. I think another contributor is that the nature of this patient population has changed over time with the availability of biologics. Many of these were low eosinophil phenotype patients. And one question is whether oral corticosteroids have any effect at all in that patient population. And if they don't have much of an effect, you're just as likely to be weaned off if you're on placebo as if you're on tezepelumab. So all of those factors are things that we're taking into account as we think about next steps here.

All right. Terrific. Well, David and Arvind, thank you so much for joining us. We really appreciate it. This was extremely insightful, as always.

Great. Thank you. Our pleasure.

[indiscernible] over the next few weeks. And for everybody, there is -- the next session is going to start at 12 with Gilead, and that requires for you to just sign in to the other login. Thank you so much.

Thank you, Yaron. Thanks for having us over.