Thanks everybody for being here. My name's John Elicker. We're here to discuss ESMO 2019. Really fortunate to have Dr. Samit Hirawat here talking to folks in the investment community. Samit will provide an overview of the clinical data that's been presented with, obviously, a focus on the 227 data. One of the things that I just ask that you respect is Samit is somewhat restricted based on his prior experience at Novartis. We have other folks here who can also comment on anything within the oncology development portfolio, Fouad Namouni and Rosanna Rickford. Then Chris Boerner is going to provide some commercial perspective on how we think about the future. The next slide, if we could go to that. Oh, we have a. Well, it took care of the legal. Just everybody validate that.
Samit, it's all yours.
John always runs faster than I expect him to, so that his famous slide came and went just like in the presentations. Good evening. Thank you for your time. My name is Samit Hirawat, and before I start, because I'll introduce myself a little bit. I've been an oncologist and I'm an oncologist by training. I've been in the pharmaceutical drug development learning for the past 20 years or so. I started my career at Pharmacia, which soon became Pfizer. I moved over to a startup biotech, PTC Therapeutics, where I learned drug development for pediatric indications, genetic disorders such as Duchenne muscular dystrophy, cystic fibrosis, et cetera. For the past 12 years, I was at Novartis doing drug development and oncology. For the last years, I was the head of oncology, hematology, and cell therapy development.
In June of this year, I joined BMS as the Chief Medical Officer and Head of Global Drug Development here. I'm pleased to be here, and I look forward to interacting with a lot of you later today and in the future as time allows. Now let's shift over to see what is going on from a BMS immuno-oncology portfolio at ESMO. We can probably divide that into three large buckets. Number one, more emerging and with a dual IO inhibition utilizing YERVOY and OPDIVO. This combination has shown prolonged activity in terms of continued overall survival benefit when we look at the CheckMate 067 trial in melanoma that was presented earlier today by Dr. Larkin. This is the longest follow-up in any melanoma trial in the first-line setting.
Second, we saw the data just now for CheckMate 227, and I will talk more about it in the subsequent slides, so I'll not belabor that right now. Second bucket, we now see long-term follow-up data in the early disease setting. That is the presentation for the three-year follow-up in the adjuvant setting for nivolumab, the only trial of its kind to compare nivolumab or immunotherapy versus active immunotherapy, that is ipilimumab. Third bucket is emerging data in new tumors. Data were not present before. These data will be presented tomorrow and day after. Once again, utilizing the combination of YERVOY and OPDIVO in cervical cancer in the first and second line to be seeing that data. Second, looking at the combination of nivolumab plus docetaxel in prostate cancer. Third, the phase III trial of nivolumab versus paclitaxel or docetaxel in second-line treatment of esophageal cancer.
These data will be presented on Monday. Now, what about lung cancer? For any given patient in lung cancer, what are the important things? Every patient looks forward to achieving a response or stabilization of disease for a prolonged duration. Toxicities that they experience to be manageable and not life-threatening or take them away from receiving more therapies. Number three, prolongation and the overall survival so that they can live longer. I believe that OPDIVO plus YERVOY meets that muster. It has shown improvement in overall survival compared to chemotherapy in patients with PD-L1, more than 1% in non-small cell lung cancer, also shown the superior data in PD-L1 less than 1%, although these data descriptive in terms of their analysis in the negative population.
We also have seen responses with this combination that are deep and durable, and I'll again share the data with you in a little bit. Last but not the least, safety profile of this combination is quite manageable and is similar to the data that we now know is present in renal cell carcinoma when low dose ipilimumab has been utilized in combination with nivolumab. You saw this slide earlier, so I will not go into the combination, the mechanisms of action. If you look at nivolumab, it basically restores the T cell function that has been suppressed. On the other hand, it's important to add ipilimumab here because it causes de novo T cell activation and occurrence of novel clones that lead to emergence of memory, which probably contributes to the deeper and more especially, the longer-lasting responses and durability of those responses.
This is a biological hypothesis that has and shown to be true and demonstrated an improvement in overall survival in two indications where the drugs are approved. Renal Cell Carcinoma. Now we have a 30-month follow-up, and you can see the landmark analysis over there that were presented earlier, and today you saw the 5-year follow-up. There you also see the addition of each of the components, ipilimumab to nivolumab, and you see the separation of the curves and stabilization of that curve, which is such a hallmark for immunotherapy. Designed for CheckMate 227. Let's talk about lung cancer now. Certainly, the primary endpoint was evaluated in part 1A of the study, where patients with more than 1% PD-L1 expression were enrolled.
Descriptive analysis, due to statistical analysis plan that was impacted due to the tumor mutational burden based progression-free survival analysis was conducted in part 1b , and you also saw that data, and I'll share that more in the coming slides. Coming to the primary endpoint. In patients with PD-L1 more than 1%, we certainly saw a very long follow-up of 29.3 months, longer than what has been shown in the prior trials of immunotherapies for a single agent in combinations with chemotherapy. This is much mature. We do see the primary endpoint was met. We do see the stabilization of the curve that is starting to occur beyond that 2-year landmark, and we do see that there is benefit to these patients in the long run.
More importantly, the blue curve versus the green curve, we do see the superiority of the green curve or the combination of ipilimumab plus YERVOY or nivolumab plus ipilimumab over single agent, and we do see that stabilization continuing to have that separation of the curves. We have looked at the data for the KEYNOTE-042 trial that the discussant had brought up today. If you look at the curves, number one, with a longer follow-up, this curve is much more mature. KEYNOTE-042 had shown the data for 18 months. Second, if you look at the curves, they look quite similar for nivolumab versus that of pembrolizumab. Third, ASCO 2018, Leena Gandhi presented the discussion on the 042 trial, and her words were, "Nivolumab and pembrolizumab look similar. They are the same in terms of their activity." You heard from Dr. Solange Peters today. Response rates are very important.
More importantly, three times higher responses compared to chemotherapy in terms of complete response. Twice the number of complete responses compared to single agent nivolumab. Even more important, the right side of the slide. While responses can be seen with chemotherapy, they are very short-lasting, and unfortunately, these patients progress. Second, if you look at the medians, combination of nivolumab plus ipilimumab has a duration of response that is four times longer than chemotherapy alone and about nine months longer compared to single-agent nivolumab. That combination is absolutely producing a 50% non-progression at the two-year landmark analysis. Coming to the descriptive analysis in the PD-L1 less than 1% negative patients. Once again, we see that the addition of ipilimumab to nivolumab shows that improvement in overall survival. Again, the curves have separated out over here in the long follow-up of this trial.
It is, of course, no surprise that when you combine the data for the negatives and the positive, the overall survival benefit is seen in patients with negative as well as positive tumors for PD-L1 expression. You can see even after 2 years, the curves are beginning to stabilize and the separation of the curve continues to be there and as we follow these patients for the longer duration. This is very important. I think there is a discrepant opinion. We tend to disagree with what the discussant said. I think there is a spectrum of opinions that will come up. Dr. Solange Peters presented the safety data in a certain way. The discussant presented it another way. It's important to pay attention to this slide and look at the numbers. I think cherry-picking numbers can have its own effects and its own conclusions.
What we do see over here, that the duration of treatment with the combination therapy, nivolumab plus ipilimumab, is longer than chemotherapy. Thus, that 12% discontinuation due to adverse event is not something that I would really embark on and look at separately. Number 2, if you look at the all grade toxicities, compare those numbers of nivolumab plus ipilimumab compared to the chemotherapy combination. They stand out, especially the hematological toxicities for the grade 3, 4 events. Now we have, obviously, continuing and emerging experience in the community as they use this combination for two other indications, RCC and melanoma. Even more important, because there is a discussion around chemotherapy plus IO combination, how does that safety compare to the chemo-free regimen of nivolumab plus ipilimumab? Once again, we see over here, this picture speaks 1,000 words.
On the right side, you see the combination of nivolumab plus chemotherapy showing an impact on the overall safety profile, as opposed to the green bars on the left side, where we don't see as high a chemotherapy-induced or rather nivolumab plus ipilimumab toxicity. Let me just conclude. I don't think we've cured lung cancer. I think we have still more work to do. There is still an unmet medical need, and what physicians or patients are looking for are agents and treatment options that can give them the durability as well as the increase in the overall survival. We do believe that ipilimumab plus nivolumab offers that unique profile in terms of improving the overall survival, leading to durable responses and deep responses of patients being alive at two years.
Thus, as Dr. Peters said, OPDIVO plus YERVOY potentially is a treatment option for patients in the first-line treatment of non-small cell lung cancer. With that, thank you so much, and I'll ask Chris to come up and give you a commercial perspective. Thank you.
Thank you. Good evening, everyone. It's a pleasure to be here and see all of you again. Let me just pick up where Samit concluded. For those of you on the phone, I'm on slide 17. Let me start by walking through some of the dynamics in first-line lung cancer. As I think all of you know, first-line lung cancer is a disease in which IO therapy has become well-established, both as a monotherapy as well as in combination with chemotherapy. You see that in the PD-L1 positives as well as in the negatives, although to a lesser extent in the negatives. From this meeting, you also see that this is continuing to be a very hot area of interest.
You see a continuous flow of data readouts as well as a number of important studies. All of that said, though, I do think it's important that we also remember that this is still an area of significant unmet need. Let me characterize that in a few ways. First, this is an area where there's still a need for agents that deliver long-term survival. Less than 20% of the patients diagnosed today with lung cancer are going to be alive in five years. Second, there's a need for agents that deliver durable responses. The vast majority of patients who are treated with standard of care IO plus chemotherapy will relapse within a year. Finally, I would point out there's still unmet need with respect to the depth of responses.
Complete responses are exceedingly rare in lung cancer. In fact, less than 1% of patients who are, again, treated with standard of care IO plus chemotherapy are going to get a complete remission. While we need to be very cognizant of the competitive dynamics in first-line lung cancer, we must also not come to the conclusion that we've closed the book on the opportunity to help patients in first-line lung cancer. It is in that regard that we do believe that OPDIVO plus YERVOY has a role in first-line lung cancer based on a differentiated profile. I'll just highlight a few things. First, this will be the only dual IO agent in first-line lung cancer to demonstrate long-term survival.
As you just saw in the data that was presented by Samit, you see very compelling CRs and durability of responses, and obviously, this could be an option for patients who don't want or cannot tolerate chemotherapy. We do believe that dual IO therapy will have a role to play in first-line lung cancer. It's still but we have had the opportunity over the last month to share these data with about 100 physicians, and what they've told us is broadly characterized on this slide. First, and not surprisingly, these physicians point out that IO monotherapy and combination therapy is well-established in first-line lung cancer, and that there are patients who need a chemo combination regimen. Patients, for example, who have very aggressive disease, where you need to get a response very quickly, those are patients who should be getting a chemo combination regimen.
They also point out the need for new treatment options in first-line lung cancer, and there are a number of aspects of the OPDIVO, YERVOY regimen that they find attractive, and let me highlight just a few. First, they're very interested in the flattening of the overall survival curve that you see with this regimen. That is particularly true, by the way, of physicians who have experience with the regimen in either renal cell or metastatic melanoma, because they draw parallels between what they've seen in those two diseases to what in lung cancer. In addition, they find the landmarks as compelling. They also find very compelling the CR rates and the durability of responses that you see here.
Importantly, for those physicians, particularly those who again have experience with the regimen in melanoma and renal cell, they find the combination of OPDIVO and low dose YERVOY as having a very manageable safety profile. With respect to the specific patients that they would consider using this regimen in, the reality that they've gotten with the data over the last month, they highlight a number of potential patient populations. First, those patients who have less aggressive disease, where they don't need to get that immediate response required from chemotherapy. In addition, they highlight subgroups of patients where there either isn't an established standard of care or where existing agents perform less well. Those would be patients, for example, who are squamous cell, patients potentially who are PD-L1 negative.
Obviously, there are those patients who will ask for a regimen or where, again, that patient can't tolerate chemotherapy. What we've seen from physicians is very much a willingness to find potential patients who could benefit from dual IO therapy. One thing we can say definitively today is that everywhere we had the opportunity to put OPDIVO plus YERVOY into the market, we've established it in that market. What you see here are data from melanoma and renal cell, and for those of you on the phone, I'm on page 19. This is U.S. data, and what you see in these tumors is that OPDIVO plus the YERVOY regimen is either established or in the case of melanoma, has a dominant share in these tumors. I would point out that both of these are tumors where we have considerable competition.
For example, in both melanoma and renal, you see competition from targeted therapy, some chemotherapy, and importantly, you also see competition from other IO agents. I'll call out renal cell in particular, because we've had a number of conversations over the last number of months with the introduction of IO and TKI therapy, what that would do to our market share of OPDIVO plus YERVOY, and you can see that in our indicated population, we retain just shy of 40% market share. All of that is, of course, driven by a very clinically compelling profile as well as very strong commercial execution. You could reasonably ask, what does that all mean for lung cancer? I'd highlight a few things. First, in lung cancer, a very high percentage of physicians who treat lung cancer also treat melanoma and renal.
In fact, in the U.S., roughly around 80% of lung cancer treaters are also treating in renal cell or melanoma. The second thing I would highlight is that there's strong familiarity among lung cancer treaters with OPDIVO, and that's based on the very strong position that we have in our second-line indication. In fact, in second line, we still maintain a dominant position from a commercial standpoint, and that's in spite of significant competition in that space. The third thing I would highlight is that, and this is probably most importantly, is that about 50% of treaters in lung cancer have experience with OPDIVO and YERVOY in other tumors, and those physicians account for about two-thirds of the potential patient population in first-line lung cancer.
The other thing I would point out is that this is a market we know well as a company, we know well from a commercial and medical standpoint, and we certainly have shown an ability to compete and compete effectively. If I would summarize this, I would say this is a market where there's a lot of crossover across tumors between lung, melanoma, and renal in terms of treating. This is a market where physicians well, and it's importantly a market where their experiences in melanoma and renal are particularly relevant to the ability to treat those patients in lung cancer. From a commercial standpoint, I'd leave you with the following on slide 21. First, we are keenly aware of the competitive dynamics in first-line lung cancer. We will be entering after IO has been established in that setting.
That said, there is still areas of important unmet need in this space. This is an area where physicians are continuing to look for novel therapies. Many of the characteristics that they see with OPDIVO plus YERVOY in lung cancer are characteristics that they find particularly relevant and necessary for patients in lung cancer. It's an area that we know well. It's an area that we've demonstrated an ability to compete. I can say definitively that as we go through the discussions that will take place in the coming months, particularly regulatory discussions, the opportunity that ultimately we have commercially in this space, I feel very confident in our ability from a commercial standpoint to capture that opportunity. Slide 22. Let me pull up a little bit from lung cancer for the next couple of minutes.
Since we were first approved with OPDIVO in second-line melanoma in 2014, we have managed to build a franchise for OPDIVO alone of just over $6.7 billion globally. We've been approved across 19 indications in 67 countries across 11 tumors. I would say that the vast majority of those 19 indications, we have established and maintained a leading commercial position. We have a very to start with, as we sit here today. As we think forward about the number of growth opportunities we have, this is where that growth will come from. We have over 20 potential growth opportunities with OPDIVO and YERVOY. I'll just highlight a few. In lung cancer, beyond the data that was presented today with 227, we obviously have CheckMate 9LA.
For those of you who don't recall, 9LA is testing the hypothesis of whether the addition of 2 cycles of chemotherapy to dual IO therapy provides benefit in this space. As I mentioned before, patients who we know need a chemotherapy combination, so this, depending upon the data, could be a very important treatment option for them. That data should read out in the first half of 2020. Beyond lung cancer, there are a number of other metastatic studies that will read out over the next few years in renal cell with 9ER, esophageal, and gastric cancer, among others. As we've talked about in a number of different settings, we have a very large early-stage program which will begin to read out in 2020, and that program will potentially add to the leadership position that we have at melanoma today with OPDIVO.
Obviously, all of these IO life cycle opportunities set on top of the combined opportunity associated with bringing the BMS pipeline and the Celgene pipeline together. We have, in previous settings, talked about the big six launch opportunities. You all may be familiar with the fact that fedratinib, branded Inrebic, was approved a few weeks back for myelofibrosis. We have important PDUFA dates coming up over the next two quarters for luspatercept and ozanimod. liso-cel will have important data that will be presented at ASH, and as we get into 2020, there will be a number of other opportunities. What I would leave you with is we've got a very established franchise in IO with OPDIVO and OPDIVO plus YERVOY. We're obviously excited about the data that we saw with 227, and that sits on top of a number of other growth opportunities that we have with the brand.
With that, I will stop and invite Samit up to the stage, and we can begin Q&A. As John mentioned, we have a number of other folks here today who can also chime in for the Q&A. Thank you. I don't think it's on. There we go. Just a reminder for those of you in the room, in addition to Samit and Chris, I should have mentioned earlier, Giovanni Caforio, our Chairman and CEO, is here. You all know him. Adam Lenkowsky runs our U.S. business, is here as well. Fouad Namouni, many of you know. Suding Meyer, who's been running the lung development program specifically. Dheena, who's our resident hematology expert, is here as well. With that, we'll take questions. Start with Chris Schott. Need a microphone over here.
Great. Thanks. Chris Schott at JP Morgan. Two questions. The first one, like the YERVOY that was very solid. I know there was storage for it, but a bit of that segment.
Yeah. I think it's a bit too early to talk about whether or not we would envision that being on label. Obviously, we're happy with the benefit that we saw in that setting. I think that what I can say right now is if you level it back up to what we're hearing from physicians, what physicians are telling us is, first of all, they like the profile of OPDIVO plus YERVOY. They see aspects of that profile that they can apply to different patients within their practice. I think that we typically talk about the patient population in lung cancer as exclusively as based on PD-L1 segments. I think in reality, and Samit can speak to this, as can Fouad, when you're sitting in front of a patient, you're actually having a different conversation.
I think the way that physicians have seen this data, as we've shown it to them, as they've got comfortable with it, is that they think about this data as potentially being useful in particular subsets. That is certainly one that they have mentioned. They've also highlighted squamous populations. They've talked about patients who have less aggressive disease. I think as physicians get more comfortable with the data over time, they'll be interested in using it in particular subsets, that being one. From a commercial standpoint, see what the label says, and we'll obviously adjust our commercial efforts accordingly. Anything to add?
No, I think you've covered it all.
Maybe just a second one for that was, when you think about the commercial approach you're going to take, talk a little bit about, think about just the academic side of the market versus community. I just think about the profile you hear, it's the business nuance kind of benefits in terms of depth of response, launch et cetera. You're kind of fighting, I guess, the headline and also final payer access.
Right.
Okay, talk a little bit about how you're going to approach those.
Yeah.
We see more potential in one setting.
Right. Let me start, and then maybe I'll ask Adam Lenkowsky, who leads our U.S. business, to chime in as well. First of all, I think with respect to academic versus community, I think we've obviously gotten some good support for the data that we've seen with 227 in the academic setting. You have to remember roughly 80% of the opportunity sets in the community setting here. What I would say is, absolutely true, there is a focus, particularly in meetings like this, on hazard ratios. I think that we also have to remember that very rarely does a physician sit in front of a patient and talk about hazard ratios. What they talk about are they talk about landmarks, they talk about the durability of the responses that they're going to get.
In that respect, I think the data that we've seen here can potentially play well, again, for patients who are not either currently well-served by existing therapies or where the data are particularly compelling versus the current standard of care. I think in that respect, the way we will be approaching this from a commercial standpoint is very much in line with the way physicians talk to patients about therapies in lung cancer and elsewhere. Again, I think in our case, what we're going to continue to do is leading up to approval. We're going to be talking to physicians about this data. As physicians have gotten more comfortable with it, they begin thinking about what particular subsets of their patients this could be most applicable to. Adam, do you want to?
Just echoing some comments. We do know that the majority of physicians who treat lung cancer are in the community, about 80%, Chris said. We think about our current user base in renal cell carcinoma, about 70% of our business comes from the community setting. There's a lot of familiarity with OPDIVO and OPDIVO Yervoy. When we think about our renal indication, the low dose of Yervoy, physicians are getting very comfortable utilizing and managing the adverse events, which is leading to the performance that Chris has shared, with a share of roughly 35%-40%.
Yeah, I think the only thing I would add to that, Adam alluded to it, is that there's been a lot of discussion about the toxicity of this regimen. I think first of all, it is important to remember that we're talking about OPDIVO and low dose Yervoy. I think in some ways, you have an empirical answer to the ability of physicians to manage that, and that is that market share in first-line renal cell. In spite of a number of different competitive options in that space, most recently IO plus TKI, you see market share hanging around 40%. That's five months, by the way, post the most recent IO plus TKI approval. Thanks. Can we go to Andrew? It's nice to see you. Andrew Baum. Chris and Samit, a note from the webcast, they're having trouble hearing the question.
If you could maybe repeat it for the people on the line.
Thank you. It's Andrew Baum from Citi. You spoke about subgroups where it would be at not the first option to use OPDIVO, namely rapidly progressing disease. I could also think of some other subgroups, so patients with a history of prior autoimmune disorder, for example, maybe the elderly patient. If we were to put some cuts of that, I guess I'm trying to come up with what really is accessible to you. You've obviously got a PD-L1 negative out where your data is most interesting.
Right.
You've got your rapidly progressing 15%, let's say, prior autoimmune disease, 10%.
Yeah
ballpark, just give me some guidance on the numbers to realistically get an assessment of where you compete with the value proposition, assuming that the durability argument isn't yet fully mature.
Right. I think, Chris, can you repeat the question?
Okay. Sorry, it was a question, maybe I'll just synthesize it, if you will. If we could try to characterize the percentage of patients who might not be eligible for this therapy, and thus, by extension, those patients who might be eligible. Maybe I'll start and then turn it over to Samit. What I would say is that, first of all, sort of difficult to give an exact percentage for a couple of reasons. Number one, many of the things that you talked about are overlapping, it's difficult to segment them out. What I'd say is if you look at the fall-off on Kaplan-Meier curves, you see roughly 30%-40% of patients who fall off very quickly in the first six to nine months. That's a rough approximation, which is we have very significant disease.
That said, what we've heard from physicians is that how they think about the aggressive nature of the disease or a patient who would be eligible for chemotherapy is very much dependent upon the patient who's sitting in front of them. You can't simply look at ECOG performance status for patients who have bulky disease. In many cases, they're clinical characteristics that present themselves. I can't give an exact percentage, but what I can say is as physicians have thought about the profile of OPDIVO and compared it to PD-L1 monotherapy as well as combination with chemotherapy, they are starting to identify specific subtypes of patients that might be eligible for this regimen. Anything to add?
No, I think you've said it all, but the only thing I would say is that when we look at the various analyses, the subgroups, age does not really play a role in terms of the activity. Neither does gender in a major way, neither does the histology. In fact, squamous and non-squamous do the same. I think it is a patient-by-patient analysis afterwards that when the patient comes in, data aside, and you say, "Okay, performance status 0, squamous cell, not a very aggressive disease. What options do I have?" You have two or three options, then you say, "Okay, do I try the chemo or do I not try the chemo? Do I give a single agent? Do I do a combination?" I think that is the way.
I don't know the numbers in terms of what the share is of each part of the puzzle, but that's how the decision will be made, I think.
Just one follow-up question. You gave the frequency of most common adverse events. With immune-related adverse events, they're less frequent but often more severe. Could you just, if it was in the presentation, I apologize if I missed it?
Right.
Just the instance of immune-related adverse events and their severity at the start.
The question, if you believe-
Yeah. I got it. I'll repeat the question and then maybe I'll ask Fouad or Sabine to take that because you know the depth of the data. The question is around immune-related adverse events in the trial and compare that to the other arms. Do you have this data?
Yeah.
Come on, Sabine. There you go.
I don't have the concrete number here, but what I can tell you is, when we look at the immune-related adverse events, we do see them somewhat more frequently than the monotherapy and obviously more frequently than the chemotherapy, toxicity profile being different. What we also see is that the manageability is the same as with the monotherapy. Those algorithms that we have established for the monotherapy, and also in these other tumor types, they work equally well for lung cancer.
I would add, what we see with low-dose YERVOY in lung cancer is very consistent safety profile immune-related adverse events as we have seen in cancer as a whole.
The only thing that I would add to what was just said is, again, I come back to a point that was made in Solange's presentation today, which is that you're not seeing any new toxicity events or any new profile of toxicity with this regimen from what has been previously presented, very similar to what you see in for example. There, I think, thanks to the efforts of medical and commercial, we've gotten physicians comfortable with how they manage this toxicity, and that's reflected in the market share that we see.
Yeah. I would remind all of us, this was released in the New England Journal of Medicine today, so the totality of the data is included in the paper.
Thanks. Terence?
Terence Flynn, Goldman Sachs. First, just on the discussion, I think you mentioned there was nothing
On the mic. On the mic.
Second, on the CheckMate 9LA, just remind me the stats plan for that study and how the elderly population will be handled. Thanks.
Thank you.
Can you repeat the question, Samit?
Sorry. There are two parts to the question. I think one is about the stats plan for the 9LA study, and I'll ask Fouad or Sabine to comment on that in terms of the totality. The second part is the additional biomarker analysis that is to be done. Certainly, any large trial, of course, requires multiple looks in the biomarker, and that work is being done at this time. The data that you saw today were related to the PD-L1 and then, of course, the TMB part of the trial, so more data will be shared as they become available. I think those.
Just to follow up on the biomarker question, and then I go to 9LA. We do see today a improvement in overall survival regardless of PD-L1 status. It seems to us from the totality of the data that PD-L1.
As a biomarker when it comes to the combination of OPDIVO and YERVOY in first-line or small cell lung cancer. We have also shown that when it comes to overall survival endpoint, tumor mutational burden with the cutoff that we used of 10 mutation per megabase does not seem to be also a predictive biomarker. In addition to that, obviously, we continue, as Samit said, to do more work to understand what are, if there are other biomarkers. We believe today this combination actually is having and showing it in all the population of patients, regardless of histology and regardless of biomarkers. From a 9LA perspective, obviously, as you guys know, we do not comment on our statistical analysis plans, or we don't disclose them, and it's really a part of the study.
The idea there is obviously to try to look at two cycles of chemotherapy, as Chris mentioned earlier, at the same time as having a dual immunotherapy inhibition with OPDIVO and YERVOY. The study is in all comers for biomarker. It is in all histologies for 9LA, but obviously there are analyses that are pre-planned in this study to look at PD-L1 expression and also to look at tumor mutational burden. I think we look forward to see, hopefully, the data first half of next year and be able to share it when we have it.
Thanks, Terence. John.
Jon Miller from Evercore ISI. I guess a lot of folks at the conference are talking about hazard ratios, where it seems like you're flagged a little bit. You've mentioned the landmark results, which look pretty compelling and pretty similar to your competitors. A lot of this hazard ratio shortfall is being driven by chemo arms, by the comparator arm, which presumably in large, randomized, well-run clinical trials ought to be pretty tight. What's driving the difference between all these comparator arms that's causing such a large swing in your observed hazard ratios?
Thank you for your question. The question is around ratios in the various trials and what drives that. As was also mentioned today, it is very important when we look at the crossover from chemotherapy to immunotherapy, this is a trial with the highest amount of crossover. 43% of the patients in the PD-L1 positive population crossed over from chemotherapy to receive immunotherapy. Only 19.3% of the patients had crossed over in KEYNOTE-042 and KEYNOTE. You saw yesterday in the atezolizumab trial, it was only 30%. One of the artifacts or one of the reasons for driving that comparator arm to perform better in CheckMate 227 is probably that crossover.
Hazard ratios are, of course, a function of the overall curve, then when you start to compare these, looking just at the hazard ratio and that one point estimate is certainly marred with a lot of confusion. That's one of the things that I think from a discussing point of view, was a little confusing, because if you look only at the hazard ratio, they may look similar. Unless you start to look at what was the comparator arm and how would that perform, what the crossover was, a little bit of an in-depth analysis was required to really glean the benefit. That's why the landmark analysis is becoming important to start to see what the benefit was at two years and of course, with the longer term follow-up.
I think the other thing that stood out as we began to go through this data with a fairly large number of physicians over the last month is that, of course, they're looking at the totality of the data, but they're also looking at the tail of that curve. Especially as I mentioned before, those physicians who have experience with OPDIVO and YERVOY in melanoma and renal, they're beginning to draw parallels between the data that they see there and what they start to see in the flattening of that curve in the out years. I think for those of you who attended the melanoma presentation with the five-year data, I think that that's also starting to play in a little bit of can you imagine that type of scenario may play out. Now, obviously, that's an empirical question. We have to wait for the data.
I think that is an important component of what physicians who see this data and find ways in which they can insert it into their practice are also looking at.
Just to add, I think a very important point is this is a trial with 29 months follow-up, and the others that you've seen are much shorter. That tail becomes even more important when you see the steepness of the curve in some of the other trials compared to the flattening of this curve in this trial.
That's also something that came out of the discussions we've had with these physicians over the last month or so. Seamus.
Thanks. Seamus Fernandez from Guggenheim Partners or Securities. Just a couple of quick questions. The discussant focused in on the 50% plus group, and based on the majority of your effect was coming from that group. I guess, sort of to Andrew's point, we've got a bit of a gap between one to 49 versus the negatives. Can you just help us understand how you guys are going to approach that part of the argument and that discussion? The second question is, can you just give us your level of confidence or at least your vision of what the label looks like at the final analysis? Is it in PD-L1 positives, or is there an opportunity that you see to add the negatives?
The final question, and forgive me for kind of going in this direction, the comparisons of hazard ratios versus how the chemo performs, that also starts to walk us into a very dangerous structure of starting to think about the upper end of the median overall survival.
versus the comparators or the competitors and do cross-trial comparisons. You guys have either the lowest with this data set of the group with the 17-month median overall survival versus looking at a landmark analysis. I'm just trying to get a better sense of when does this discussion, from your perspective, really start to shake out. Is it at three years when we're looking at the three-year analysis, and then we're kind of saying, "Okay, we have seen a meaningful differentiation at three years. That's where we expect that data point, and a lot of what we hope is going to be very nice, we'll see it then.
Yeah. Maybe I'll take the question around how we think about the high expressers versus one to 49, et cetera. I can probably answer the question that you asked about what the label's going to look like. I think we're just in discussions now with regulators. Not surprisingly, I'm not going to be able to provide much color around that. All I can say on that point is I feel pretty comfortable with the fact that whatever that label looks like ultimately, we have a team in place that knows how to operate in this space. They know these physicians well. They know this tumor well, and I'm pretty certain that we're going to have a team that can capitalize on that opportunity.
With respect to the greater than 50 versus 1 to 49 in non-expressers, what I would say is obviously in the greater than 50 population, you've got an established monotherapy IO treatment today. What I would also say, though, Samit can speak to this, is that what we have seen very clearly in this data set is that PD-L1 expression tends not to correlate with the responses that we've seen with OPDIVO plus low-dose YERVOY. As physicians have gotten comfortable with that, they've begun to sort of get out a bit of that rubric of 1 to 49 and greater than 50 or less than 1. They've started to think about, "Okay, what is the patient that's sitting in front of me?
What's that patient's characteristics?" That's how they've begun to say, "This is how this particular dual IO regimen could fit into their practice." I think the short answer is that, in essence, physicians sort of come out of that one to 49 greater than 50 rubric and begin thinking about the patient who I'm looking at across the desk. The reality is we sit within a competitive environment in which IO is established, and certainly in the greater than 50%, we're going to have to deal with monotherapy. I think physicians generally will start to look at IO/IO therapy in a slightly different way. Do you want to pick up on the other questions?
Just one thing to add to the more than 50%, less than 50%. I think we've seen activity and efficacy and overall survival superiority in the negatives. We've seen it in the high expressers. If you look at a continuum of it, to start saying that it will not work between the one to 49, how did that number come by? Is there any scientific rationale for that? I don't think there is one. It's an arbitrary cutoff. If it's a continuous variable, ideally there should be a dose response or treatment response or exposure response relationship. None of that exists. We've done the analysis, and PD-L1, while it could be like high TMB, could be a prognostic factor where patients with high PD-L1. If anything, it doesn't mean that drugs don't work, such as nivo plus YERVOY doesn't work in one to 49.
I think these subcuts or subpopulation analysis should be viewed with caution. You've already addressed the part about regulatory status. I think once again, the data speaks for itself. We obviously cannot comment what the labels will look like in the third one with the commercial version.
Yeah, the question about hazard ratio landmarks and medians and time.
Yeah. Look, I think the data already, this is a very long follow-up of the lung cancer trials at 29.3 months. If you were to compare it with 18 months, that's not appropriate. As you already said, cross-trial comparisons are always marred with, again, confusion and caveats have to be put on. From our perspective, these data are already speaking for themselves. We've shown the addition of the ipilimumab to the nivolumab, that it does bring value to the overall survival. We've shown already the comparison versus nivo, versus chemo, comparison versus nivo-chemo. I think all of those are additive in terms of information available today. These data will continue to evolve, such as the data from others as well, and people will continue to compare that, but I don't think we need to wait for that.
I think the data are here, and I think all we're saying is this could be one of the treatment options amongst others for any given patient.
Thanks, Samit. Richard?
Richard Wagner.
Richard.
I was just going to ask one. Sorry, Richard. Just for the question on long-term sort. Two years ago, when we started with the YERVOY and OPDIVO in melanoma, we did not have the five-year survival that we have seen this morning. What we have is deep responses, complete responses, and durability of response using this dual immunotherapy indication. I think we are seeing this feature that is, we think, unique for immunotherapy combination. Depth of response all the way to complete responses and durability of response. One would hope that following more patients longer as we have been doing in breast cancer, we can hope to see similar effects.
Richard Wagner, Wolfe Research, asking on behalf of Tim Anderson. Now that we know from today's results CTLA-4 has activity in first-line lung in combination with the PD-1. The next trial that would really be of interest would be OPDIVO plus YERVOY versus the KEYTRUDA chemo combo, pitting 227 against KEYNOTE-189 or KEYNOTE-407, if not in all PD-L1 negative patients. Under the assumption that today's results are only going to gain your product modest boost in usage in patient types to be defined through clinical practice, why not run a trial like this? It seems you don't have much to lose. That's the first question. Second question, returning to the patients who would be attracted to this IO/IO combo, chemo-free combo.
More of a market research question, what percentage of patients today in practice are so averse to chemo that they either refuse chemo or take a singlet instead of the dominant doublet? Put another way, how many patients from market research who consented to chemotherapy, but only because there was no alternative? A last question going to, if you will, for the 9LA. Based on what you saw today from 227, what's your sense that the two doses of chemotherapy as opposed to the full four doses that are given in POSEIDON will be a sufficient boost to the results? Thank you.
The first question was regarding comparing nivo/ipi versus pembro plus chemo. The second question is around market research.
Okay.
The third question is around two cycles in. The first one is provocative, certainly. Thank you for that provocation. At the current time, as far as I know, in the three months that I've been here, we don't have those plans. We don't intend to do that. We've shown over here, nivo chemo versus nivo/ipi. We talked about it earlier, nivo and pembro, they have similar curves if you look at them, with the caveat, two different trials, with the caveat, longer follow-up for nivo, shorter follow-up for pembro. I don't think, this is my personal opinion, and certainly chime in others, that we need to do that trial. That's number one.
Do you want to take the third question?
From the 9LA perspective, the data shown today does not change our opinion. We remain as encouraged as Fouad was already mentioning. We remain as encouraged in terms of the overall outcome, and we're really looking forward to that outcome in the first half of 2020. We haven't changed our thinking around addition of chemotherapy to the doublet of.
I would say, maybe I'll start the answer to the question of what % of patients would prefer a non-chemo option, and then I'll turn it over to Adam and see if he wants to add anything. I would say a large % of patients would prefer a non-chemo option. I think the challenge that we've seen in lung cancer is that outside of very high expressing PD-L1, they've really not had that option available to them. I think the decision as to whether or not a patient is ultimately given a non-chemo option is a function of a few things. It's a function of whether or not the patient asks for it, because we do know from other settings that when patients ask for therapies, that physicians in many cases actually adhere to what the patient has asked for.
A second component is essentially what the physician is speaking to with respect to other options for that particular patient setting in front of them. I think that's where a good, robust discussion will likely take place now that you have a dual IO or will potentially have a dual IO option available to these patients. I think it's very much going to be individual. Adam, anything to add?
I think you said it really well, Chris. We've done market research not only with physicians but also with patients as well, across a number of tumors, including that the majority of patients prefer a non-chemo regimen. There's fear, as we may know, the stigma of patients who are on chemotherapy. Obviously, typically with patients who are on chemo, quality of life of patients on therapy is very poor, and we see there's quality of life versus OPDIVO YERVOY in other tumors and where OPDIVO YERVOY is superior. This really is an option for patients that either don't want or don't need a chemotherapy treatment. The 9LA study, as we said, if patients are rapid progressers, does provide an option for a chemo-sparing regimen, which is just two cycles and then ipi/nivo.
Right. Thanks, Richard. We're coming up on the end of the hour. Are there any questions before we close it down? We've got one right over here.
[Dr. Chin for these talents]. Chris, can you remind us if CheckMate 9LA allows crossover, and if so, why should we not be worried about that from a similar chemo performance perspective?
The question is around crossover in 9LA.
Crossover in 9LA. 9LA was not designed with the crossover that's built into it. Patients, wherever they are recruited in the world, they will have option for a second-line therapy, or they may not have an option for second therapy, depending where they are. There is no crossover built in 9LA.
I think that is going to wrap it up. Thank you, Samit and Chris, and thanks for all of you in the room, and thanks for everybody on the phone for joining us on Saturday. Have a great rest of the night. Great rest of the day. Appreciate it.
Thank you.
Thank you.