Amgen Inc. (AMGN) Earnings Call Transcript & Summary

My name is Ian, and I will be your conference facilitator today for Amgen's conference call in conjunction with the ASCO 2020 Virtual Scientific Program. [Operator Instructions] I would now like to introduce Arvind Sood, Vice President of Investor Relations. Mr. Sood, you may now begin.

Thanks, Ian. Good afternoon, everybody. Thanks for joining us. Would have been great to see you all in person as we normally do at ASCO, but given this year's virtual format, I'm glad that at least we have an opportunity to connect virtually. So we have presented some important data at this year's ASCO, particularly on AMG 510, or sotorasib, as we now call it, which is our novel small molecule inhibitor of KRAS G12C. Now one portion of the study known as CodeBreaK 100 relates to activity in advanced colorectal cancer patients, and the other is Phase I data in patients with advanced solid tumors other than non-small cell lung cancer and colorectal cancer. We also presented some updated results from a Phase I dose escalation study of our BiTE molecule, AMG 330, in patients with relapsed and refractory acute myeloid leukemia. So to discuss this data in broader detail, I'm joined today by Dr. David Reese, our Executive Vice President of Research and Development; and Dr. Greg Friberg, who is our Vice President of Global Development and oncology therapeutic area head. I'm also delighted that we are joined today by 2 of our clinical investigators involved in the CodeBreaK 100 study: Dr. Marwan Fakih, who is a professor in the Department of Medical Oncology and Therapeutics Research at the City of Hope in Duarte, California. And Dr. Fakih is the first author on the colorectal cancer poster. Also joining us is Dr. David Hong, who is Deputy Chairman of Investigational Cancer Therapeutics at the MD Anderson Cancer Center in Houston, Texas. Dr. Hong is the first author on the other tumors poster. We have corresponding slides that go along their presentation, and they should have been posted to the Events page within the Investors section of our website, amgen.com. So with that, I would like to turn the call over to Dr. David Reese. Dave?

Thank you, Arvind, and good afternoon, everyone. It's great to talk to you here today about some of the data that we're presenting at ASCO. If you turn to Page 3, those of you who are following on with the slide deck on the website, as Arvind mentioned, you'll see our agenda. I'll make some introductory remarks. Greg will then provide an overview of the -- what we believe are the key data that we're presenting here at ASCO. We won't go through this in great detail because the presentations and abstracts are available online now, as you know, and then we will open it up for a Q&A period with both Greg and myself as well as Drs. Fakih and Hong. If you turn to Slide 5 in the associated webcast deck. What I'd like to do is start by setting a little bit of the context here and identifying a couple of our strategic imperatives as we move forward. Our oncology portfolio, as we have discussed before, is built on first-in-class molecules. We are always focused on areas of high unmet medical need and where we believe we can deliver a unique capability that serves patients' needs. We are keenly interested in combination therapies across a range of both hematologic and solid tumors, and that remains a focus of our development programs. Now as you're aware, we've got upcoming data updates in 2020. And as we note here, the Phase I lung cancer dose expansion study, we have submitted and are hoping to present in the second half of the year at a scientific congress. And then, of course, I know all of you are keenly interested in the availability of the potentially pivotal data from our fully enrolled Phase II trial in non-small cell lung cancer. As I have said before, we are aiming to have sufficient follow-up in that trial so that we have at least 6 months of follow-up from time of first response in all patients because the last patients were enrolled at the end of last year. And as you'll see from data that Greg will share in just a moment, most responses occur within the first 3 months or so of therapy. You can see that, that pushes that data availability into the second half of the year. We are on track with that trial. I suspect that because of the follow-up criteria that I just outlined, we will probably just miss availability for ESMO, and our goal is to present the data as quickly as possible either at another scientific forum depending on the evolution of the schedules of scientific congresses in the second half of the year or potentially through top line data given the interest of the entire field in these data. So we'll provide ongoing guidance on that data availability, but I think the punch line here is that we are committed to very timely release of the data. In addition, over the course of the year, we will have program updates for some of our half-life extended BiTE programs, including those targeting DLL3 in small cell lung cancer, PSMA in prostate cancer and, of course, BCMA in multiple myeloma. I am encouraged by some of the data that we're seeing through emerging from these programs, and we look forward to sharing those with you a little bit later in the year. Now of course, if you move to Slide 6 in the associated presentation, many of you, of course, are interested in the sotorasib program and where we stand. I'd like to highlight a few salient facts for you today. As you're aware, there's a continued significant unmet medical need for many patients with the G12C KRAS mutation, primarily in solid tumors. And to date, we have now enrolled over 400 patients in the sotorasib development program, representing 13 different tumor types. And of course, we'll talk about a number of those today. As you know, the Phase II study in lung cancer is fully enrolled. I am pleased to announce today that we have also fully enrolled a Phase II monotherapy study in colorectal cancer. That will provide further insight into the potential utility of the drug in colon cancer. Also, importantly, we would like to announce that we have recommenced enrollment in the Phase III head-to-head lung cancer trial against Taxotere. Some of you may recall that we were just on the verge of initiating enrollment in that trial when widespread lockdowns from the epidemic occurred. That trial was placed on pause but is now actively screening and will begin dosing patients momentarily, and I'm optimistic about what we're going to be able to do as that study moves forward. In addition, we are now actively enrolling again in the various combination studies. In fact, patients are being dosed as we speak, and Greg will talk a bit more about this a little later on. If you move to Slide 7, we've simply provided references to the abstracts, some of which we will be discussing in a moment. Those of you who have registered for ASCO, of course, then have full-out access to these presentations and associated slide decks. So with that, I'll turn things over to Greg, who'll take you through a little bit of the key data.

Thanks, Dave. It's my real pleasure today to be able to present some of the data that we're showing at ASCO this year. This, of course, includes data for AMG 510. Their name now, we call it sotorasib. And sotorasib is our first-in-class KRAS G12C inhibitor. If we move forward to Slide #9, you can just see a quick overview of the mechanism of action of the drug. I'm not going to spend much time here. We're all familiar with this. But just in brief, this is a small molecule inhibitor. It's a covalent binder of the G12C mutation for KRAS. It's dosed once a day, and when it binds to the inactive GDP-bound confirmation of the protein, it leads to signal transduction -- inhibition of this mutant KRAS phenotype. If we move on to Slide 10, you can just see the introductory description of the first tranche of data we're going to go over. And this is a good time just to take a step back and remind you of some of the nomenclature here. We call CodeBreaK 100 the combined Phase I/II protocol that includes several different arms, several different tranches of data. The 3 tranches of Phase I data include, of course, non-small cell lung cancer data that we're not talking about today and 2 other tranches. One is the colorectal cancer subset as well as a third tranche, which, as Arvind described, includes the other tumors. These represent, again, 3 distinct tranches. And what we're presenting, as a quick reminder, is of the Phase I portions of these studies. We're very fortunate to have Marwan Fakih with us today to answer questions once I've gone through presenting this. But just in brief, if you move on to Slide 11, you can see the patient characteristics listed for the colorectal cancer study. Of note, we've now treated in the Phase I portion, 42 patients with G12C-positive KRAS mutations. This represents about a doubling of the number of patients at the highest dose, 25 at the highest dose since our last presentation at ESMO. Of note, we have longer follow-up time now, about 8 months of follow-up. And as a quick reminder, these patients are unfortunately folks who have had prior therapies fail them before. They've had 3 -- a median of 3 prior lines, meaning they're on their fourth or fifth line plus of therapy for the most part. Moving to the next slide, on Slide 12. We have a quick summary of the adverse events. And I think the punchline here is that there really have been no significant changes that we've seen in this group of 42 patients since the last time we presented. You can see, of course, of note that with regard to treatment-related serious adverse events, treatment-related fatalities and treatment-related AEs leading to treatment discontinuation, we've seen none in any of those 3 categories. Moving onward. If we move to Slide 13, we have a summary of the tumor responses in colorectal cancer. And what you can see here again in the 42 patients that were treated, there were 3 partial responses. All of them were in the 960-milligram once-a-day cohort, and you can see the durability of those patients. They are all 3 confirmed at this point. One's out pretty briefly at 1.4 months, and the other 2 are in the 4-month range. Interestingly, of course, beyond just the partial responses, if you look in the 960-milligram cohort, you see another -- approximately 2/3 of patients on top of this, who had stable disease, and we'll dive into that in a little bit. The median durability of that stable disease was measured at 4.2 months. So if we move to Slide 14, we are presenting for the first time some of the Kaplan-Meier estimates for progression-free survival, and we'll go over overall survival in a second. Now these are single-arm studies. These are descriptive cohorts. But what you can see here in these 42 patients is that the median progression-free survival appears to be about 4 months, slightly higher, probably quite similar in the 960-milligram cohort at this point at 4.2 months. And again, just to put this in context, when patients are typically in this fourth-, fifth-line therapy for colorectal cancer, approved therapies, such as Stivarga and Lonsurf, have documented progression-free survivals on the order of 2 months with really very rare responses on the order of 1% to 2%. On the next slide, Slide 15. We have the Kaplan-Meier estimates for the patients on the study for overall survival. And what you see here is that for all doses, we have a 10.1 month estimate for median. And if you break that down into just the 960-milligram cohort, that median has not yet been reached. Again, context here is probably warranted. If we look again at the approved therapies that I noted, typically 6 to 7 months overall survival in large studies has been noted. And probably, the KRAS patients even in those studies benefited less from those therapies. Moving on to Slide 16. You can see our waterfall plot. And just as a quick review, you see the 3 confirmed partial responses on the right side of the curve. Of note, there are other varying degrees of shrinkage noted here, including 11 out of 23 patients who were measurable at the 960-milligram cohort had some shrinkage and 18 out of 39 overall. If we move to the next slide, we're presenting on Slide 17 for the first time the so-called spider plots or spaghetti plots that actually give you a bit more detail with regard to tumor growth kinetics, tumor shrinkage kinetics. And a couple of take-home points here that are worth noting, one would be that, if you see, as Dave alluded to, patients who are destined to respond tend to do so fairly early in their treatment within the first or second scan. And secondly, you see that the kinetics of those who have a RECIST-stable disease includes patients who have a varying degree of tumor shrinkage and/or stabilization as well. Moving on to Slide 18. You can see our swim lanes for the patients who were treated at the 960-milligram cohort. There's 25 of these patients listed here, and 8 of them remain on treatment. So in summary, really, as Dave had noted, we were encouraged by the data that we've seen, and we have enrolled a separate single-arm Phase II cohort that has recently completed enrollment, and we look forward to letting that data mature. Shifting gears and moving to a second presentation that was done -- that went online today. We'll be presenting a tranche of data from the same CodeBreaK 100 study for the other tumors. And as a reminder, this is the G12C mutation-positive patients with every other histology other than non-small cell lung cancer and colorectal cancer. So on Slide 20, you get a flavor of who these patients are. Interestingly, we only had 6 of these when we presented our data at ESMO previously. We now have 25 of them across 11 different tumor types. You can see, again, the median follow-up is a little less mature than with the other cohorts, representing the fact that many of these patients have been enrolled since our last update of the data. Of the 11 diseases that are here, the most common are pancreatic with 10 and appendiceal with 4, and then we have 2 each of unknown primary and endometrial with a smattering of single digits in some other diseases as well. Of note, all but 2 of these 25 patients were treated with the 960 milligrams a day dosing. If we could advance to Slide 21. Slide 21 is a quick overview of the adverse events in this tranche of 25 patients. And again, we don't see any significant differences to what we've been presenting already: no fatal AEs, no AEs listing -- leading to treatment discontinuation. We did have one SAE that was deemed to be treatment-related by the investigator of pneumonia. Moving on to Slide 22. We have a summary of a waterfall plot for these patients for those that have had enough time to be evaluated and for whom were measurable. And what you see here is 3 confirmed partial responses. The 3 diseases that were included appendiceal cancer, endometrial cancer and melanoma. You see as well that there were a variety of patients who had what looks like the measurable tumor shrinkage that was classified as stable disease at the time of the data cutoff and included 5 patients with pancreatic tumors who had some degrees of shrinkage. Moving on to Slide 23. We have our typical swim lane plots with the detailed data for these patients that were on the study. There's 22 of these patients who have responses evaluable who are on this plot. And the first takeaway here is, of course, that these lines aren't nearly as long as for some of the other diseases. This is early, and we need to, therefore, again, take the data with a grain of salt. We do know that with regard to the partial responses, the melanoma patient and endometrial patient remains on treatment. And we have an additional 7 patients, including 4 pancreatic cancer patients out of the 8 who are evaluable, who remain on study. So I think just to wrap up the overview of the other tumor types, we -- as Dave had highlighted, we have Phase II portions of this CodeBreaK 100 study, and we do continue to enroll exactly these patients who fall into this other tranche into a large Phase II cohort to really better inform this data and understand more broadly across these diseases where G12C is more rare how the drug might perform. Moving on to Slide 24. Just in summary, as Dave has noted, we've treated over 400 patients across the CodeBreaK 100 studies. We are also exploring 6 different combination cohorts in a master protocol that we call CodeBreaK 101. It was a study that was highlighted in the Trials in Progress section of ASCO. And if we move on to Slide 25, I just want to give you a little bit of insight into how that study is designed and what it looks like. So Slide 25, again, shows the schematic for this program. It was designed, as I mentioned, as a master protocol, trying to be more efficient than running individual Phase Ib combinations on their own. As noted in the trial and progress poster, there -- these are, of course, protocols that have multiple parts. There are safety parts that really have a minimum of 6 to 10 patients, and then they can be followed by larger expansions of up to 60 patients each. We've announced 6 of the sub-protocols in the study so far. There are more on the way. But of those 6, they're combinations with a MEK inhibitor, a PD-1 inhibitor, a SHP-2 inhibitor, a pan-erbB tyrosine kinase inhibitor, a PDL-1 inhibitor and an EGFR inhibitor that includes with and without chemotherapy. If we move on to the next slide, I just want to take the last few moments and reflect on the data that's been presented for our AML BiTE, AMG 330. This is being presented by Dr. Ravandi, and I would just say that this now -- since the update that we last provided at ASH '18 about 18 months ago, we've been able to increase the exposure, the dose levels more than threefold in these patients. We have about 20 more patients, 6 more cohorts. Now these are fourth-, fifth- and later-line AML patients. They're very sick Phase I patients. We have been able to increase the dose, and we've been very pleased to see our first MRD-negative complete response in those studies as well. I'll just wrap up the quick summary in saying that we're continuing to escalate. We're fine-tuning dosing schedule. We're learning with this CD33 target in AML how to manage side effects, including cytokine release. And all of those learnings, all of the target coverage and the AE mitigation profile learnings really will apply across our platforms, including our half-life extended CD33 platform, AMG 673. So with that, if we move to the next slide, I think we've -- I'm going to hand it back to Dave, and we're entering our question-and-answer portion, and I thank you all for your time.

Well, thanks, Greg. We'll move in directly into the Q&A period right now. We will be joined, as we mentioned at the outset, by Drs. Fakih and Hong. And so you may feel free to direct questions to them as well. I should point out that both of them are deeply experienced investigators and also have quite a bit of clinical experience now with sotorasib or AMG 510. So Ian, why don't we go ahead and open up the line for questions?

[Operator Instructions] Our first question is from the line of Yaron Werber with Cowen.

Maybe I have 2 questions. Maybe the first one, just a housekeeping. If I remember correctly, I think I heard you say that the 3 PRs for colon cancer are now confirmed. Is that true? And what about for the KRAS, the non-colon, non-small cell, the other, is that confirmed? And then I have a question for the one about KOLs.

Yes. This is -- Dave, go ahead.

No. Go ahead, Greg.

Thank you. So this is Greg Friberg. Yes, they are confirmed, both in the colorectal as well as the other, meaning that they've had a follow-up scan that was at least 4 months after the original scan that showed at least 30% shrinkage.

Okay. And the question, I mean, I guess, to both Amgen and the KOLs, I mean, is -- what's the thought about doing -- is there a thought to do a monotherapy head-to-head study against Stivarga? I think you alluded what the standard of the care is. It's about -- I think it got approved and literally no responses and about a 1.2 months' PFS or so. Is there a thought to do a head-to-head against Stivarga in fourth-line CRC monotherapy to get a quick registration?

Well -- and thanks, Yaron. Let me start with that, and then I'll ask Dr. Fakih to comment on the general clinical context for these patients and his impression of the data in colorectal cancer to date. So we're obviously taking a look at potential monotherapy route to approval in colorectal cancer. As I mentioned, we have fully enrolled a Phase II trial in colorectal cancer. And I think we'll look to see initial results from that study at the end of this year and very first part of next year before making a final determination in all likelihood on a Phase III head-to-head trial in that particular setting. I would point out that we are also keenly interested in combination therapy in colorectal cancer, and we hope to advance those investigations in the arms of the trial that Greg showed you a few minutes ago as quickly as possible. So Dr. Fakih, Marwan, perhaps we can get you to comment on the general clinical context here in terms of monotherapy and standards of care. And you may be on mute, Marwan.

You got it. I was on mute. Sorry about that. So I agree the median progression-free survival so far is quite interesting. It does look favorable compared to both Stivarga and Lonsurf, trifluridine, and so does the rate of progression-free survival, whether at the 3 months' point or the 6 months' point because with both Stivarga and Lonsurf at 6 months, it's a 10% progression-free survival rate. And at the 4 months' point with both Stivarga and Lonsurf, your progression-free survival rate is roughly between 20% and 25%. So this almost doubles the rate of progression-free survival at 4 months and 6 months, which is quite interesting. I think as alluded to earlier, there's probably 2 challenges with the monotherapy. Number one, such a study will require a large number of centers because it's -- the frequency of the mutation is low. And then the question becomes do we jump on, on monotherapy? Or if we're going to have some signal of efficacy from any of the combination arms, would it make more sense at that point to look at the combination as a possible registrational venue? Of course, those are -- that's my own opinion, that if you do have a much stronger signal from the combination, that it does make a lot more sense that, at that point, especially with a potential -- rare patient population or relatively low number of -- or low percentage of patients, to go on with the combination. But as far as clinical significance, I do regard these numbers clinically significant for a third-line and fourth-line setting for metastatic colorectal cancer, and I do agree that they look more favorable -- considerably more favorable than what we would expect with the available drugs, Lonsurf and Stivarga.

And our next question is from the line of Terence Flynn with Goldman Sachs.

Maybe just one clarification. First is just with respect to the -- Dave, with respect to the timing of the 510 Phase II lung data, did you mean that you're going to miss the abstract submission cutoff in August for ESMO? Or did you mean you're going to miss ESMO in September? And then on the lung cancer trial, the Phase III that you just recommenced enrollment in, just wondering if there's the flexibility to add a combination arm to that trial in the event that you start to see some interesting data out of the CodeBreaK 101 study.

Thanks, Terence, and very good questions. Yes. In terms of the first question regarding the Phase II data availability, I think it's very unlikely that we will be -- have data in time for a presentation at ESMO. So my anticipation is that the data will come a little bit later than ESMO. And again, as I said, we're committed to ensuring that we have the appropriate dissemination of those results in a timely fashion. In terms of the Phase III trial, we are -- we have discussed the potential of adding additional arms, of course, that introduces some methodologic challenges, statistical challenges, but that is something that we have considered upfront, including something like a combination arm, should those emerging data look interesting.

Our next question is with the line of Michael Yee with Jefferies.

Maybe a question for David. In terms of the combination studies that are ongoing, could you comment -- I guess, that's laid on Slide 25. Maybe just pontificate or comment about some of those arms and what you're more interested in, for example, in lung and colorectal. And I know you've said it's kind of been on pause, but is that back up and running? And maybe just talk about the timing of all that in the combination study. And then a follow-up related to lung. If you go back to the lung program, you previously said you would be doing mutation analyses and sequencing. Is there anything you've learned in any of those responders or nonresponders that might be helpful in thinking about future designs or whatever? Appreciate that it's a bit competitive.

Yes. Thanks, Mike, for those questions, also 2 very important questions. In terms of timing on the combination, on -- so those are, by and large, up and running. We've enrolled and actually started dosing patients in several of those arms, and we'll provide guidance over the course of the year in terms of when we can anticipate data presentations. All of these combinations were chosen based on biology and the potential for additive or synergistic effects in various tumor types. Of course, the relevant combinations may differ across -- and will differ across malignancies given the underlying mutational background of various tumor types. So more to come on all of that, but I'm pleased with our progress on that, and thank you for your interest in the combination therapy trial. In terms of biomarker assessments, we actually have a very extensive biomarker program in play right now. And we hope later in the year to prevent -- present the first tranche of data from that. This includes mutational analysis and various other assays that we have done as part of the program. So I think quite a bit of biomarker data are being amassed, and we will be presenting those over the course of the year. I'd also like to invite Dr. Hong to join the combination -- or excuse me, the conversation and comment specifically on the various combinations and how he sees those and their potential utility perhaps in different indications. David, would you like to comment?

Yes. Can you guys hear me?

Yes, we can. Please go ahead.

Can you all hear me?

Yes. Go ahead.

Yes. So -- hello, can you hear me? Hello?

Yes...

Yes, we can hear you.

Sorry. So thanks, Dave. I'm very excited about, to be honest with you, all of these combinations. When we were talking about this before the CodeBreaK study with Amgen, and we laid out kind of our kind of best one -- combinations, and they took that to heart and they've already -- we've already implemented them. And what's cool is that we've now kind of -- this CodeBreaK master protocol, which kind of allows us to kind of put these combinations as kind of -- almost kind of cassettes into the protocols that have allowed us to really kind of expedite the activation of these trials. And we don't have to take a gazillion steps in order to activate many of these combinations. And so as David alluded, we've already gone ahead and started to enroll in several of these combinations. And all of them are exciting, I mean, the MEK inhibitor, the SHP-2. I do think that -- particularly, I'm excited about the EGFR plus 510 and possibly in combination with chemo in that context because I think, to some extent, the story of 510 will be similar to the BRAF story in colorectal, where, as you all know, that story, the response rates were not that high. But when you combine it with other agents like an EGFR or a MEK inhibitor or chemo combination, you saw the higher response rates and so forth and, obviously, other benefited ratios, such as progression-free survival, overall survival. So I'm -- I think that, that is -- in some of the tumor types other than lung, I think that may be where we're going to see activity, augmented activity, I guess, with the 510 molecule. Just going back to my abstract. Particularly, when you look at that waterfall graph, specifically that pancreatic water graph, I know that none of those were PRs, but I put a couple of those patients on. Those were -- those are significant responses in patients who were highly refractory. And what it tells me is that I wouldn't be surprised if we combine this with a chemo combination such as FOLFIRINOX or gem-Abraxane, where we could get additional benefit with the combination. So there's kind of -- I don't want to say there's an endless combinations to come about, but there's lots of combinations that we can look at, different reiterations that hopefully will begin really allowing us to take advantage of 510. And Marwan and will, I think, attest to this: one of the things that I -- is optimistic is that 510 in itself is an incredibly well-tolerated drug. The toxicity profile, as you've seen, is mostly Grade 1, some Grade 2 and very few Grade 3 toxicities. And what it tells us is that, likely, we're going to be able to combine this with chemo, with immunotherapy, with other small molecules without any significant [ TOC collapse ]. So that's also what, as a Phase I doc, I'm excited about.

And our next question is from the line of Geoffrey Porges with SVB Leerink.

First question I'd like to ask is just about the nature of identifying the G12C mutation. These tumor types, such as colorectal or pancreatic, are regularly genotyped in this way. And is there potential to move towards liquid biopsy? And then I just want to follow up on the comments on pancreatic cancer. I'd love to hear how frequently this mutation has identified pancreatic cancer and whether you have created an expansion cohort already based on the signal in pancreatic cancer.

Thanks, Geoff, and I'll take the question on the diagnostics and then ask Greg and, potentially, our investigators to comment on the remainder of your question. So we are moving forward with both tissue-based diagnostic partnership as well as a liquid biopsy partnership. Both of those are under full steam, and our anticipation is that they will be ready for the sort of joint filing that would permit approval of the diagnostic at the time of initial approval of the drug. So my full expectation at this point is that, that is not -- will not be rate limiting and that we will have both tissue and liquid diagnostic approaches available to treating physicians. Greg, perhaps I can ask you to start and comment on the remainder of the question.

Sure. Thanks, Dave. Yes, the literature would suggest that in pancreatic cancer, the rate of G12C positivity is about 1% to 2%. Of course, finding the patients is the issue. We have been able to find a fairly large cohort of them. And as I mentioned, we have a Phase II cohort for this other category, of which we expect that pancreas given that they're the #1 from a prevalence standpoint that we've been able to find in our protocol will continue to fill up that cohort. If I could ask Dr. Fakih just to comment as a GI oncologist about the sequencing in his patients.

Sure. So pretty much every single colon cancer patient with metastatic disease should have KRAS testing. Now the -- I think what's missing right now is the awareness factor in the community because this -- the drug has not been approved yet. AMG 510 is still experimental, and not everybody has followed the data. So there are some patients with KRAS G12C who are walking around that, unfortunately, are not being recognized. But I think as the awareness is there, then these patients should be captured. I think the other thing is that some labs do report present -- KRAS mutation present versus not present rather than putting the raw data there. And in that setting, it's not that they don't know, but that's how they're reporting their data. The incidents in our patients, when we look at 500, 600 patients that we have genotyped their tumors on over the last few years, it's about 3% to 4% of our patients with colorectal cancer have a KRAS G12C. But I can tell you that we've seen second, third opinions coming in for a Phase I study, and nobody had picked up that they had a KRAS G12C. I think the challenge in pancreas a little bit, it's not a standard of care to do a KRAS testing in that -- in those patients. While some academic practices and some physicians are doing genomic profiling, one of the reasons maybe those aren't being identified as much is also the fact that it's not a standard of care that every patient with pancreas cancer should have a KRAS mutation assay.

And our next question is from the line of Mohit Bansal with Citi.

Great. And congrats on all the progress. One question from -- both from Amgen's point of view as well as investigators there. So first, before that, how big is the Phase II part of the colon cancer trial? Now the question is, could that size be registrational enabling because -- asking because on, one hand, you have low responses, but the norm is low responses in this particular cancer. And also, you are actually going after a subset of colon cancer patients. So the conditions which apply for likes of Stivarga, they may not apply here. So I just wanted to get your thoughts on if this can be registrational enabling.

Yes. So I'll ask Dr. Friberg to comment on some of the specifics on the Phase II colorectal cancer trial.

So with regard to the Phase II tranche within CodeBreaK 100, we haven't disclosed what the size is, but it's fair to say that it's larger than our experience so far. With regard to whether or not this would be of registrational [ or additional ] opportunity or not, it's really too early to say. I think what we want to make sure of is that we get better estimates around not only this response rate [ I think ] but also these progression-free survival and overall survival curves. Of course, it is a single arm, and so we'll have to see how this data is moving forward and whether or not a monotherapy versus -- as I think Dr. Fakih and Dr. Hong nicely pointed out, or whether or not a combination would represent a better opportunity for patients. That book is still being written.

And our next question is from the line of Ronny Gal with Bernstein.

If I can bring in 2. First, regarding CodeBreaK 101. You've initiated both a PD arm and a PDL-1 combination. And I'm kind of wondering if there's a hypothesis here for a difference about the mechanism of action of the 2 agents that would lead you to need to do separate arms for both. And second, kind of looking at the data you presented, it seems that the big differentiation is -- potentially would be on OS versus PF or ORR for this molecule as -- for AMG 510 as monotherapy. And I was wondering if the thinking is for the registrational trial to use OS as the primary endpoint. Or are we still looking at kind of like PFS and the combination of PFS and ORR?

Great. Thanks, Ronny. Let me handle that. And then perhaps I'll ask Dr. Fakih to comment on the specifics of colorectal cancer and study designs. And as you noted, in CodeBreaK 101, we have both PD-1 and PDL-1 arms. You may recall from the paper that we published in Nature last year, there was some evidence of upregulation of the immune response. And one of the intents here is to see if there is any difference in response to PD-1 versus PDL-1 inhibitors. We also, of course, want to get safety data in both of these combinations. So I think that was really the biologic and clinical rationale between including both in this multi-arm combination trial. In terms of overall survival. Overall survival could well be, in fact, an important endpoint in advanced colorectal cancer monotherapy trial. And let me ask Dr. Fakih to comment on his views of what, as a clinician, he would like to see in terms of endpoints in such a study.

Yes. I mean, I think if we are investigating this as a third-line, fourth-line refractory patient population or at least failed 2 lines of therapy, including the standard chemo, probably, overall survival would be the cleanest endpoint and the most clinically relevant endpoint. [ That ] this is a targeted therapy, and I think there's precedent previously on -- regarding approval based on PFS. And I do think that the progression-free survival is a very valid endpoint, especially if you have a PFS exceeding 4 months. But in general, I have no doubt that in the setting where it would be investigated, these 2 endpoints are likely going to correlate well because this would be in a pretreated patient population. If the FDA does allow, I do think PFS would be the way to go and would be a little bit more reflective of the treatment options -- of the treatment effectiveness in that setting, especially if there are venues at that point when such a study is conducted for some of these patients to cross over, for example. Or if the data is so strong that the study allows a PFS primary endpoint and allows a crossover of the control group to the treatment group and focusing on a PFS as the main endpoint, I think that probably would be the most favorable way to proceed from a patient perspective because we would be able to allow all patients to receive that treatment.

And our next question is from the line of Evan Seigerman with Crédit Suisse.

So one on housekeeping, you had mentioned the Phase I updated non-small cell lung cancer in the dose expansion cohort. Any reason why we're not going to get data from the dose escalation cohort? And then looking at the colorectal kind of program and overall [ back -- I just wanted to ask ] about the need for combination therapy. It's clear that that's kind of where you're going. I think for Dr. Fakih, any thoughts as to what rational combinations you would like to see and how you would potentially see this program move forward?

Great. Thanks, Evan. So I'll ask Greg to address the first question on the expansion cohort from the Phase I non-small cell lung cancer trial. And then Dr. Fakih can comment on combination therapies of interest to him in colorectal cancer. Greg?

Yes. Thanks, Dave. When it comes to the Phase I portion of CodeBreaK 100, we think of the escalation patients and the expansion patients in a tranche, and we wanted to present an update of the lung cancer patients all at once. Now unfortunately, COVID has shaken up some of the conferences that we're aware of, including World Lung Congress (sic) [ World Conference on Lung Cancer ] that got moved into next year, and that's forced us to really reevaluate our options to present that data. We really hope to be able to publish the data or present it at a congress before the end of the year. Of course, that requires that the data be accepted, but we're very much looking forward to presenting that data together with appropriate follow-up.

Okay. Great. Thank you. And then...

Great. Thank you, Greg. I think -- go ahead, Marwan.

Yes, I'll take the second question. Yes. So I think to me, the combination with a monoclonal antibody targeting EGFR has been always on my mind. Theoretically, given the -- given what we know about the same MAP kinase pathway and the fact that EGFR is overexpressed and overactive in colorectal cancer compared to other malignancies, such as lung cancer, and given the synergy of combining BRAF inhibitor plus anti-EGFR inhibitors, and this is kind of a similar pathway or at least it affects the same MAP kinase pathway. I would like to turn your attention to recent publications by a group from MD Anderson -- from Memorial Sloan Kettering that collaborated with an Italian group looking at KRAS G12C mutated cell lines as well as PDXs and shows a clear difference in EGFR phosphorylation in colorectal cancer compared to lung cancer. The senior author Misale is. It was published about a week ago in Cancer Discovery. And it shows a very, very, very strong synergy between AMG 510 and an anti-EGFR monoclonal antibody. And that synergy does not happen as much in lung cancers with KRAS G12C. So I think there's now preclinical data both in cell lines as well as patient PDXs with colon cancer with G12C on that paper that has -- is public for review that strongly puts a rationale -- strongly supports the combination of an anti-EGFR plus AMG 510 in patients with colorectal cancer. So I'm very excited about such a combination in patients with G12C mutation.

And our next question is from the line of Alethia Young with Cantor Fitzgerald.

Congrats on all the progress with this program. One question, one follow-up. I just want to know, when you look at kind of some of the other tumors, beyond pancreatic of sorts, which you sort of discussed it, what are -- are there strong biologic rationale to potentially kind of study the other KRAS mutations there? And then secondarily, maybe this is more for Amgen. I mean, what are the prospects of maybe running a basket trial of sorts once you get past some of the work in non-small cell to maybe flesh out that signal?

Yes. Thank you, Alethia. I'll take the second part of your question and handle that and then ask Dr. Hong to talk about some of the other tumors in the rationale. In terms of basket trial, as Greg mentioned, we actually have opened a Phase II trial that is, in essence, a basket trial, monotherapy trial, for nonlung cancer, noncolorectal cancer indications. As Greg also noted, we anticipate that there will be a fair number of patients with pancreatic cancer in that trial, but we're also hoping to get additional patients with tumors, such as endometrial cancer and appendiceal cancer, where there's also a measurable rate or frequency of G12C mutation. So we've -- that has a sample size that permits enrollment of as many of these patients as we can really get on. And once we increase the numbers, I think we'll have a sense of the utility of monotherapy there. But Dr. Hong, perhaps we can ask you about the sort of nonlung, noncolorectal, nonpancreatic tumors and what your thinking is right now.

Yes. This is an interesting area. And I think now that kind of this barrier of tumor-agnostic therapy has been breached with other therapies, it's always interesting to see if we could -- can possibly do that with any other agent. I think it's just too early for us now to know given the relatively small numbers in the current cohort here to what extent 510 will have efficacy across all tumors that have KRAS G12C. Obviously, there's hints here that looks like there's clear activity. We just -- I want to -- we have 2 abstracts beyond the 510 that I published with some fellows: one by Dr. Nusrat, who's now a member a faculty member at Sloan Kettering and [ Dr. Kee ], who's a Phase I fellow with us. In those 2 days, as we looked at, one, our tissue database of about 42,000 patients and looked at just prevalence of -- presence of the KRAS G12C mutation. Obviously, we -- in that data set of 42,000 patients, about 2.4% or about 8,000 patients were identified -- sorry, about -- yes, about 2.4% of patients were identified with G12C. Most of those were non-small cell lung and colorectal. But the third most common were actually gyne tumors and particularly endometrial in that data set. And if you saw our -- the abstracts, there were 2 endometrials. And one of those, patients has had a really profound and, actually, a fairly long response. As the data matures, could we see a higher response rate or equivalent response rate with monotherapy in nonsmall cell lung cancer and gyne tumors like endometrial? Possibly. Just -- the data is just too early to say. So we'll see what the Phase II data says. But I do think that we are going to see activity in other tumor types. The question whether or not we'll see broad activity across all different histologies, like a larotrectinib and entrec fusion, that's still, I think, up in the air.

And our next question is from the line of Geoff Meacham with Bank of America.

I just have a few. So the first one on the colorectal data, just want to get the KOLs' perspective of how good they think the durability of response rates are as well as OS, PFS just compared to line of therapy. In other words, what would you normally expect in third-line plus? And were there themes that you can identify biomarker-wise in either nonresponders or even super responders? And then for [ lung cancer ], Greg, on CodeBreaK, I think this is obviously a smaller cohort, but what would you say are the triggers for expanding the program outside of colon or lung? Do you need to see, for example, a bigger n or mature PFS to get into another registration Phase II outside of colon or lung?

Great. Thank you, Geoff. So let me just say a word again about the biomarker-related question. As I mentioned, we're amassing a large amount of biomarker data as we speak, and we anticipate presenting that later in the year. So at that point, we can have a more fulsome discussion about potential patterns of response or resistance. So we can have Dr. Fakih comment on the durability question as he see it for these patients who have often had multiple prior lines of therapy in colorectal cancer. So Dr. Fakih?

Yes. I mean, in general, the patients who have a PR or CR will have a more sustained durability than somebody with stable disease only. I think it's encouraging at this point that the progression hasn't occurred yet in the 3 cases that have been reported. Obviously, when you have 3 patients with an objective response, it's hard to project what is a median. I think what would be more important in such a setting in patients with third line and fourth line and especially that you're really dealing more with a very large cohort of stable disease is the progression-free survival. And I think that's where I would focus on as far as the main primary endpoint in such a delayed line of treatment in patients with colorectal cancer. And as I alluded to earlier, the fact is we're seeing almost double the historic control with trifluridine and with regorafenib. And I think the fact that this is coupled with a very favorable safety profile, I think this is very meaningful for our patients. And I would have no hesitation prescribing such a drug in a third-line, fourth-line setting given how safe it is for a patient with KRAS G12C. I think, however, if we can get a better outcome with combination, that would definitely be the way to go as we learn more from the combination treatments.

Great. Thank you, Dr. Fakih. Greg, can you address the portion of the question that was directed to you?

Sure. Yes, thanks for the question, Geoff. With regard to the tumors outside of lung and colon that we're studying right now, it's a tough question. These are single-arm studies. These are Phase I populations. When we see an objective response, it's something that we can point to and, I think, credibly use that data. Of course, we would like to see multiples in any one given tumor type. When we're dealing with more disease stabilization, looking at time-based endpoints, I think we need to have a sufficient number of patients out at sufficient time points, whether that's 6 months or 12 months. These are the kind of data that we use to try to interpret those results. But in a world where these diseases may be on the more rare side, we can also have the opportunity to potentially bundle them. But there is no one set of rules to trigger expansion in these larger tumor types. We right now are trying to collect as many patients as we can in this tranche and -- in the Phase II. And when we've looked at that data, we're hopeful, again, that it's going to give us some insights when we look at it even on a patient-by-patient basis.

Our next question is from the line of Carter Gould with Barclays.

Great. Congrats on the progress, guys. I guess, first, given sort of the coherent messaging we're hearing around safety, has there been any interest -- maybe ask, Amgen, your willingness or interest to explore higher doses here or potentially other dosing paradigms like b.i.d. dosing. That's the first question. And then specifically, with the combination cohorts, any more specific pancreatic-type regimens potentially being added to that study?

Yes. Thanks, Carter. In terms of dosing, I would point -- I'll start, and then I'll ask Greg to provide some color, and then he can talk about the combination therapy cohorts as well. We're quite happy with the monotherapy dose that we currently got. As I've noted before, we really wanted to achieve a Cmax that was above a threshold, substantially above a threshold for a period of time that would allow complete abrogation of KRAS signaling through a dosing interval, i.e., a day. We think we have achieved that. Now that being said, of course, any program that is early -- and it's worth pointing out that we still actually are relatively early despite the progress and numbers of patients enrolled. We, of course, will explore different dosing paradigms, including potentially twice-daily dosing. And when we have amassed such data, we'll be happy to talk about those. Let me turn it over now to Greg to talk about the combination therapy cohort.

Yes. Thanks, Dave. Just as a point of reference, I think it was on Wednesday we reached 22 months since our first patient dose. So certainly, this program continues to move quickly. With regard to the combinations, we are envisioning a variety of, particularly chemotherapy, combinations that we can and plan to add to the master protocol moving forward, some of which include active agents in pancreatic cancer. So it's absolutely something that is on our mind. What we want to do is explore as many hypotheses based on some of the preclinical data that I think Dr. Hong referred to as well, and that does include combinations with some cytotoxics used in pancreatic cancer.

And our final question is from the line of Jay Olson with Oppenheimer.

And I appreciate the comments on PFS. And when I look at Slide 15, it looks as if there might be a slight hint of an OS benefit for the 960-milligram dose. So could you talk about the potential for these high disease control rates that we're seeing for sotorasib in colorectal cancer to translate into an OS benefit? And then as a follow-up, if your Phase II study were to become registrational for colorectal cancer, what is the OS threshold you would need to reach? I think you mentioned a median PFS of 4 months, but what sort of median OS would you think might be required for registration?

Yes. Perhaps I'll ask Dr. Fakih to address those questions about the sorts of OS criteria that he would like to see in colorectal cancer. Dr. Fakih?

Yes. So I mean I think I am excited about an OS of 10 months? Because historically speaking, it's expected to be about 7 months with regorafenib or Lonsurf. However, saying that, it's -- one has to really kind of recognize this is really a selected patient population on a single-arm trial, and one cannot exclude sometimes the potential of patient selection biases in these settings. And given the lack of randomization, I am not really sure that you can use from a single-arm trial an overall survival endpoint, even if it is 10, 11 months or even if it's 12 months in that setting, as a support for a registrational path. Although I cannot speak for the FDA, I think with targeted therapy, historically speaking, I think the FDA has looked more at response rates in such settings and coupled probably with protracted progression-free survivals. But I think from a single-arm perspective, I'm not really sure that the OS would be something -- even if it's 11 months in a single-arm trial, that would be something that would be registrational. I would have my doubts.

All right. Well, thank you, Dr. Fakih, for that important perspective. And seeing as we're well past the hour, I think we'll go ahead and conclude at this point. I really want to thank all of you for joining us. We remain quite excited about not only the sotorasib program but our BiTE programs and other oncology programs. And we look forward to providing additional updates on our earnings call and in the course of scientific congresses and publications over the remainder of this year. And as always, our Investor Relations team is available if you have follow-up questions. I would also like to extend again our warm thanks to Drs. Fakih and Hong, who are deeply experienced and, I think, have given us valuable perspective today. Thank you, everyone.

Thank you.

Great. Take care, everybody.

Ladies and gentlemen, we thank you greatly for joining us today. You may now disconnect.