John E. Elicker - Bristol-Myers Squibb Co. Giovanni Caforio - Bristol-Myers Squibb Co. Charles A. Bancroft - Bristol-Myers Squibb Co. Thomas J. Lynch, Jr. - Bristol-Myers Squibb Co. Murdo Gordon - Bristol-Myers Squibb Co..

Umer Raffat - Evercore ISI Jami Rubin - Goldman Sachs & Co. LLC Andrew S. Baum - Citigroup Global Markets Ltd. Seamus Fernandez - Leerink Partners LLC Tim Anderson - Sanford C. Bernstein & Co. LLC Chris Schott - JPMorgan Securities LLC David R. Risinger - Morgan Stanley & Co. LLC Vamil K.

Divan - Credit Suisse Securities (USA) LLC Gregg Gilbert - Deutsche Bank Securities, Inc. Geoffrey Meacham - Barclays Capital, Inc..

Good day and welcome to the Bristol-Myers Squibb 2017 Third Quarter Results Conference Call. Today's conference is being recorded. At this time, I would like to turn the conference over to Mr. John Elicker, Senior Vice President, Corporate Affairs and Investor Relations. Please go ahead, sir..

Thank you, Tracy, and good morning, everybody. Thanks for joining the call today to review our third quarter earnings. With me are Giovanni Caforio, our CE; Charlie Bancroft, our CFO. Both Giovanni and Charlie will have prepared remarks.

And then, as well, Murdo Gordon, our Chief Commercial Officer, and Tom Lynch, our Chief Scientific Officer are here for Q&A. Take care of the Safe Harbor language first. During this call, we'll make statements about the company's future plans and prospects that constitute forward-looking statements.

Actual results may differ materially from those indicated by these forward-looking statements as a result of various important factors, including those discussed in the company's SEC filings. These forward-looking statements represent our estimates as of today and should not be relied upon as representing our estimates as of any future date.

We specifically disclaim any obligation to update forward-looking statements, even if our estimates change. We'll also focus our comments on non-GAAP financial measures, which are adjusted to exclude certain specified items. Reconciliations of these non-GAAP measures to the most comparable GAAP measures are available on our website.

Giovanni?.

Thank you, John, and good morning, everyone. Before we get started, I want to say a few words about the situation in Puerto Rico. As you know, this is a significant humanitarian issue, and we are very focused on ensuring our people are safe and that we support ongoing efforts to provide disaster relief.

We have two manufacturing plants as well as commercial operations on the island. Our first priority following the hurricanes was to make sure our people were safe. And I'm pleased to report we've accounted for 100% of our colleagues on the island. With respect to our operations, we do not expect to have any supply disruption.

There's been a tremendous effort by our teams to make this happen, and that effort is ongoing. They've done a great job executing against our contingency plans, despite very difficult operating conditions. Going forward, it remains essential the infrastructure is restored as rapidly as possible. Now, getting back to the quarter.

I'm pleased that we had another good quarter, with continued strong performance across the company. We grew our revenue by 7%, with double-digit growth for Opdivo, Eliquis and Orencia. In oncology, Opdivo had another very good quarter. Despite increasing competition, Opdivo's performance across all indications continued to be very strong.

Globally, we delivered a 38% increase. In the U.S., we maintained a strong share of the second line lung market in the high-30s. We are also seeing really good performance in other tumors, with strong shares in renal and melanoma. Outside of the U.S., we are seeing strong performance, as we move along the launch curve in those markets.

Our regulatory and clinical achievements positioned us well for continued growth of our I-O portfolio. We secured FDA approvals for Opdivo in both liver and colorectal cancers, as well as an approval in Japan for gastric cancer. And the FDA also expanded the Yervoy indication to include pediatric melanoma patients.

Our sBLA for Opdivo in adjuvant melanoma was granted priority review and given breakthrough designation. We announced the results from CheckMate-214 in renal cell cancer, which was stopped early for overall survival in intermediate to poor risk patients.

We believe this is an important result that validates the rationale for combining Opdivo and Yervoy. We presented important data at World Lung (04:40) on TMB in small cell lung cancer, and we're looking forward to presenting new data on our IDO at SITC.

Looking ahead, there are a number of milestones with the potential to significantly expand our efforts, to bring the benefits of immuno-oncology to a broader set of patients. We expect to see the final results for the combination in first-line non-small cell lung cancer from CheckMate-227 in the first half of 2018.

We are also looking forward to seeing data for Opdivo in first-line HCC and for the combination in head and neck and small cell lung cancer next year. And our next wave I-O agents have tremendous opportunity to build further approaches, to expanding the number of patients who benefit from combining I-O agents within our portfolio.

I'm really looking forward to the initiation of registrational studies for IDO and LAG-3. Outside of oncology, our portfolio performance was strong, with Eliquis delivering 39% growth. We see continued growth opportunity for Eliquis where real-world data is reinforcing its important role in reducing stroke in patients with atrial fibrillation.

We're looking forward to additional real-world data for Eliquis later this year at AHA. I'm very encouraged to see our diversified pipeline advance, with good progress on our FGF21 agent in NASH. We are in discussions with the FDA and plan on beginning the next phase of our program in 2018.

We're also progressing our pipeline in immunoscience, where we expect two early-stage assets, our BTK inhibitor and our TYK-2 inhibitor to deliver proof-of-concept data in the coming months. These are promising medicines in areas where there is an important need for new treatment options.

During the quarter, we pursued several strategic business development deals, to further build our portfolio and pipeline, specifically in immuno-oncology. We acquired IFM, which expands our assets in innate immunity, an area of promising early science. We also signed an agreement with Clovis to start a joint development program for its PARP inhibitor.

And our agreement with Halozyme positions us to enhance the delivery of our cancer treatments. I believe we have the most promising early pipeline we have had in a long time, and our capabilities to deliver that pipeline are evolving significantly.

We have strengthened our investment in translational research, tumor biology and real-world data, as a part of our integrated R&D approach. As Tom can describe in more detail, we believe that these capabilities position us well, to continue to demonstrate the value of our medicines and regimens for patients.

And so, with that, I'll turn the floor over to Charlie. Thank you..

Thanks, Giovanni. Good morning, everyone. This was another important quarter for the company, with strong sales growth across our key brands. Let me begin with a brief discussion of product performance starting with I-O. As Giovanni mentioned, we delivered strong commercial performance, as we maintained leading share in the U.S.

for Opdivo in second-line lung and renal. We also drove continued adoption in melanoma and other indications. International growth was robust, and we achieved additional reimbursement approvals, including non-small cell lung cancer in the UK and lung and renal in Australia.

Looking forward in building on our strong foundation, we believe we have meaningful opportunities ahead with potential approvals in adjuvant melanoma and first-line renal cell. For Yervoy, we see near-term market dynamics as somewhat mixed.

We expect some pressure due to adoption of Opdivo in adjuvant melanoma, but potential for extended use of the regimen in intermediate poor risk first-line renal cell.

Our strong commercial execution across the portfolio was also driven by exceptional performance for Eliquis, which saw a 39% worldwide growth in the quarter, bringing the product to almost $5 billion in annualized revenue. As we described in the past, the impact of the U.S.

coverage gap increases considerably in Q3 and Q4 compared with the first half of the year. Though this put a headwind on U.S. sales during the quarter, the 8% sequential increase in TRx reflects strong underlying demand growth.

Globally, Eliquis is the number one NOAC in the U.S., Japan and Canada and is the leading brand in terms of new-to-brand scripts in important markets such as Germany, the UK and Spain.

With Eliquis continuing to extend its lead within the NOAC class and the category poised for continued expansion, we believe that Eliquis is well-positioned for growth going forward. Orencia and Sprycel delivered strong performance during the quarter, delivering year-on-year growth of 10% and 8%, respectively.

This is meaningful growth for products at this stage of their life cycle and illustrates the very good commercial execution we've seen across the portfolio. Turning to our non-GAAP P&L. I'd like to start by discussing our gross margin.

While we are pleased by the demand trends and our performance during the quarter, gross margin was pressured due to product mix, particularly the strong growth of Eliquis, coupled with the declines in virology. We also took a one-time, $70 million inventory charge related to HCV, given the competitive dynamics we are seeing globally.

Going forward, we expect mix will continue to be the main driver of gross margin. With respect to OpEx. We are delivering on the approach we outlined about a year ago. We've been investing in R&D while finding efficiencies across the rest of our business.

Because of the importance of our pipeline for realizing our long-term potential, we continue to view investment in R&D as a key priority.

Regarding our tax rate, during Q3, the tax rate was unfavorably impacted by both a reduction in expected credit for the Puerto Rico excise tax from hurricane-related operational disruptions on the island, and from adjustments to our prior year tax liability.

Briefly with respect to capital allocation, business development remains a top priority, and Giovanni outlined several key transactions that we executed during the quarter. Additionally, we bought back $225 million of stock during the quarter, and we plan to buy an additional $250 million for the end of the year.

Turning to guidance for the remainder of 2017. We are revising our tax rate to 22%, mainly due to Puerto Rico, and I expect our gross margin to be towards the lower end of the range. Having said that, we are encouraged by the business outlook going into the fourth quarter, and we are increasing our non-GAAP EPS range to $2.95 to $3.05.

To close, our year-to-date performance has been very strong. Across the company, we've continued to execute and deliver outstanding commercial performance and made important progress in advancing our pipeline.

We believe there are tremendous opportunities across our portfolio of marketed products and pipeline assets, and are well-positioned for future success. I'll now turn it back to John for Q&A..

Thanks, Charlie. Tracy, I think we're ready to go to questions. And just as a reminder, in addition to Giovanni and Charlie, both Tom and Murdo are here. Go ahead, Tracy..

Thank you, sir. We will now take our first question from Umer Raffat from Evercore. Please go ahead..

Hi, guys. Thanks so much for taking my question. I had a two-part question on tumor mutation burden. First, on the CheckMate 9LA trial, is the type of TMB test that you're using similar to what was done in CheckMate-026? I'm just trying to understand, what's the base test that you intend to use in that trial and perhaps other trials.

And then, second, not on statistical plan, but can you confirm that there's been no change to the primary analysis of CheckMate-227? Thank you very much..

Umer, thank you. This is Tom. Thank you very much for your question. And I think the part about TMB is a very interesting one and I think that what this indicates is how we're learning so much more about lung cancer. When you think about lung cancer, it used to be a disease where it was just squamous cell and adenocarcinoma.

Now, you have a disease where it's EGFR and ALK and ROS1 and RET fusion proteins and PD-L1 and increasingly, we're starting to understand that TMB is also an important potential biomarker in lung cancer. And you saw three important data sets that we're very happy to present.

The first was our CheckMate-026 data that you alluded to at AACR this year, which showed that you could select a group of patients that seem to benefit from Opdivo monotherapy.

We showed similar data in bladder cancer, and then, of course, at the World Lung Cancer Meetings last week in Yokohama, we showed data that in small cell lung cancer, we could select patients for both Opdivo monotherapy and Opdivo and Yervoy combination therapy, using TMB as an important biomarker.

So we think this is something that's important for us to look at in many of our studies, in a wide variety of tumors. Now, your specific question about how do we measure it, I think, is a very good one. And I think this is a field that is evolving.

When you think about how we did it with CheckMate-026, we did whole exome sequencing as a mechanism for doing it, and our plans for CheckMate 9LA will be to use the FoundationOne panel that's a sequencing set of a defined set of genes.

And both – the good news is that, as we showed at ACR, whether you do whole exome sequencing or whether you do the Foundation panel, there appears to be excellent concordance between the two types of assay. You don't get the exact same result, but the results are very concordant with each other.

I do think this is an area that's going to evolve over the next several years, as TMB becomes a more important marker.

I also think it's important to think about, how does one measure TMB? Will it be all tumor based, will it be serum-based, I think that's another area of great interest, and where our translational medicine teams will be putting a lot of their attention. Your second question is regarding the stat plan.

And I think I just want to say one thing about CheckMate-227. We really aren't going to be talking about the statistical plan of CheckMate-227. As you know, we have co-primary endpoints of PFS and OS, and we have the optionality of looking at that, but we're not today going to talk about the stat plan of CheckMate-227.

But I think your question on TMB is quite important, and one that we continue to look at with great enthusiasm, across a broad variety of tumors..

Thanks for the question, Umer.

Tracy, can we go to the next one please?.

We will now take our next question from Jami Rubin from Goldman Sachs. Please go ahead..

Thank you. Just if I can follow-up on Opdivo. I think Giovanni, you said that the CheckMate-227, we would see the final analysis during the first half of 2018. Can you confirm whether or not there will be an interim – has an interim, in fact, happened and already passed? Is that still a possibility? And then, secondly....

Jami, we lost you..

Can you hear me?.

There you go. You're back..

Okay. Sorry about that. I don't know what happened there. Anyway. As we think about trends next year, you will have hopefully frontline renal, adjuvant melanoma, which are two very important, big opportunities. At the same time, the dynamics will start to change in second-line lung as KEYTRUDA has been used in the frontline for over a year.

So can you just sort of high-level help us to think about the pushes and pulls to the Opdivo franchise next year? Bottom line, will sales be up, flat, down, just sort help us to think about that. Thanks very much..

Jami, this is Giovanni. Thanks for your questions. So with respect to my comment, let me just say, we've consistently said the final analysis for CheckMate-227 and the timing in the first half of next year, that's really not changed.

I'll ask Tom to give you a perspective on interim analysis, and then, maybe Murdo to give you his perspective on current trends for Opdivo. We are, obviously, not going to give any perspective at this point on 2018 more specifically. It's premature to do that, but it's important to discuss existing trends..

And Giovanni, thanks, and Jami, thanks for your question about the interim. So what we said about the interim is that, and as you know, it's event-driven. And you really can't – you don't know when events are going to happen.

You have to wait for the events to actually happen and that could be extremely frustrating, because you'd love to be able to say on date X, we're going to have all the events, but they happen when they happen. We predict based on looking at things.

We think that the interim will happen either at the end of 2017 or early in 2018, and that's when we expect the interim analysis to happen at that point. But as we've said all along, we expect to have data on CheckMate-227 in the first half of 2018 and we're looking forward to being able to share that data with you when we have that..

And Jami, thanks for the question related to trends on Opdivo. We're, obviously, really pleased about the performance year-to-date on Opdivo and the source of that growth that we're experiencing having stable trends in second-line lung cancer in the U.S. and good growth and penetration of the same opportunity ex-U.S.

as we've secured reimbursement, and most recently, seeing reimbursement for Australia and the UK come online ex-U.S. Now, in the U.S., beyond lung, we also continue to hold really high shares in RCC, despite having some strong competition in cabo in second-line. And we also are doing well across our I-O portfolio in first-line metastatic melanoma.

And if we look at catalyst for growth going forward, clearly, we're excited about the great data from Opdivo in the adjuvant setting and the NCCN guidelines were updated to reflect those data, so that could lead to some NPS in the near term.

We're also hopeful that the strong data from the CheckMate-214 first-line renal cell carcinoma trial, the combination of Opdivo and Yervoy, will be submitted quickly and reviewed quickly, hopefully, for the benefit of patients looking to experience an OS benefit in first-line renal. That's another strong catalyst for growth.

And we're just recently launching in hepatocellular carcinoma and the anecdotal early feedback on that is also very strong. So I know there were questions on whether or not we would grow this year in the U.S.

I think we've been very good in terms of the commercial and medical and customer facing execution and I feel that we would do everything we can to have a strong performance next year. But obviously, there are a lot of variables and a lot of data sets still to read out and we'll be talking to you over the quarters as we go..

Tracy, can we go to next question, please?.

We will now take our next question from Andrew Baum from Citi. Please go ahead..

Thank you. Three questions, please. From memory, I think about 65% of patients in CheckMate-026, you had access to TMB baseline levels, I'm obviously thinking about missing data.

What percentage should we be thinking about for CheckMate-227? Second, just to confirm that if, as we believe, the TMB baseline part of your final (22:15) analysis, I'm assuming I would say (22:18) by definition has to be part of the interim and you can mark that in a theoretical way of ex U.S.

(22:22)? And then, finally, would you like to share with us any of the differential activity (22:30) combination strong (22:33) CheckMate-012 and CheckMate-568 also with the combination compared to NIVO plus chemo? Thank you..

Andrew, Tom Lynch. Thank you. I did not hear your second question. I'm sorry, I got the first and the third.

Just repeat the second?.

Sorry, I'll repeat that. Sure, of course. So I was saying that assuming that TMB – patient selection base from TMB baseline today is (23:02) part of your final analysis.

One, should we assume that, by definition, it also has the full part (23:08) of the interim OS analysis?.

Thank you. Okay. So let's talk – Andrew, thank you for your question and let's talk a little bit about where we are with CheckMate-227 and the study. I think a couple of things to say. And as we pointed out for quite some time, CheckMate-227 is not just one study, it's a family of studies.

It's a study in expressers, it's a study in PD-L1 non-expressers, and it's a third study in all-comers. And this family of studies, we think, is going to be more than – we think we know is more than 2,400 patients with advanced lung cancer. It's going to give us a lot of information about how to treat patients in first-line non-small cell.

When you think about the amount of data that will be coming out in 2018, both from us and from other companies and corporate groups around the world, I think a year from now, we're going to have a much better idea about how to treat patients, how to identify subsets of patients who are going to benefit, and have to really learn how to give patients the best options for treating patients in that setting.

Now, your first question comes down to the point about TMB. And you mentioned that we're able to look back in study CheckMate-026 and get about 60% of patients that were able to have a viable (24:30) TMB.

And I think one of the things we've said for quite some time is we will continue to do as much work as we can with our studies, retrospectively, looking at TMB when possible.

And as we've said, we believe TMB is an (24:44) important enough that we're going to be looking at it in study CheckMate 9LA, and we'll be looking at it in CheckMate 9LA at that point. So we think TMB has potential as a potential biomarker. Your second question is about the concept of looking at TMB in an interim analysis.

What we've said about the interim analysis at our last earnings call is that overall survival would be looked at in the interim analysis and that's the extent that we can comment on the stat plan and the interim analysis at this point was just to disclose what the endpoint is for the interim analysis that will be done when the events occur.

And then, finally, your question on IPI and NIVO in CheckMate-012 and CheckMate-568, I think a couple of things about the IPI and NIVO. One is, this has been a really important year for IPI and NIVO. It's been a year where just in the past five months, we've seen great confirmation of the value of this combination.

It works in melanoma, it works in patients with renal cell, we could not have been happier with the results of study CheckMate-214, and those are studies where we've shown survival advantages in that setting. And we've got five additional tumors where we see clear response rate benefits to the use of IPI and NIVO compared to monotherapy.

We await, obviously, PFS and survival benefits in those trials where we've only seen response rate data thus far. So we believe this combination has great potential benefit for patients with cancer. So one of the reasons we did study CheckMate-568 is to get more experience with IPI and NIVO in that setting.

So we believe we have a family of studies in lung cancer, which are going to help define the best way to treat patients with non-small cell lung cancer, and we really look forward to sharing that data in 2018 as it becomes available..

Tracy, can we go to the next question, please?.

We will now take our next question from Seamus Fernandez from Leerink. Please go ahead..

Great. Thanks for the question. So first off, just wanted to ask on next year, as we think about 2018, Charlie, can you just give us a general sense of the pushes and pulls as it relates to the gross margin? The commentary on $70 million of inventory write-down was certainly helpful.

But just trying to get a general sense of what you see as the key pushes and pulls around the gross margin and overall mix for 2018. And then – and maybe Murdo can comment on just sort of the dynamics that he thinks would be the big contributors there as well.

And then, on TMB, just wanted to get a better sense of whether or not the blood-based FoundationOne test already has a validation set. So I know that you've got the ability – typically, what I think the FDA requires is there is an exploratory data set followed by a validation data set.

And I'm just trying to understand if you can comment on whether or not Bristol considers the blood-based FoundationOne test to be a validated test to apply, whether it be as part of CheckMate-227 or how are you using it in the CheckMate 9LA study? Thanks so much..

Okay.

Charlie?.

Okay. Thanks, Seamus. For 2018, by and large, our gross margin has been impacted by mix, notwithstanding this particular quarter where we had the HCV inventory write-off of $70 million. And going forward, it will be guided by mix.

And as you know, Eliquis, which has been a nice growth driver for us, has a margin below 50% given the structure of our P&L and our relationship with Pfizer. So you can think of it in the sense of Eliquis growth, although really good on the top line does have a headwind on the margin side.

And we're also, although, towards the end of HCV, by and large, we do lose Reyataz and Sustiva in the U.S. at the end of this year and also very high-margin products. So I would think about it in the context of Eliquis being a drag. We are losing some higher-margin products.

So you can think about the mix being the most important factor in our gross margin going forward..

And Seamus, just on trends, obviously, we're really pleased with Eliquis demand trends that we're seeing. We're up 8% sequentially quarter-over-quarter in TRx demand and up 37% over same quarter prior year in demand. So the top line demand curve for Eliquis looks very good.

As Charlie said, obviously, it has had a lower gross margin than some of our other products.

Just another couple of dynamics within Eliquis to keep in mind and I can make a comment on this year, and that is for the third quarter, we did see an increase in the number of patients entering the coverage gap, so that showed some decline in net sales for the quarter in terms of percentage quarter-over-quarter.

But overall, we are stable to some slight increase in our gross-to-net, partly due to DoD channels and some strengthening of our access position in the marketplace with some strategic contracting that was done. So overall, Eliquis looks really good.

I-O is just really difficult to pin down, because of the number of new data sets and events, both for us and our competition, so we are holding a wide range of potential scenarios for 2018.

We, obviously, are, as I said to Jami, we're very pleased with our trends and the quality of execution around the world, and we have some very new Q4 catalyst for growth that should help us pin down what 2018 looks like. But clearly, we feel good about what we've been able to do despite competition in the marketplace with I-O.

And as Giovanni mentioned in his opening comments, rest of the portfolio with both Sprycel and Orencia have good trends for (31:01)..

And Seamus, just to answer your question on the blood-based Foundation test, as you know, the three data sets that we reported on tumor mutational burden have been tumor-based evaluations, and that's been in study CheckMate-026 in our bladder cancer experience, and then, also in study with small cell. So those are blood- and (31:24) tumor-based.

My recommendation would be to ask the folks at Roche and Foundation. As you know, at ESMO, Roche presented data looking at serum-based and blood-based mutational assays. And I think you'd have to ask them any specific details about where they stand in terms of the validation of the serum-based assays.

What I had mentioned earlier was that I personally think that there is a real interest in looking at these.

I think these might be very promising and the early looks at them have shown that we seem to be getting very interesting signals from the blood-based assays, and whether these complement serum-based assays – tumor-based assays, or whether they are used instead of, I think, will be an interesting question to find in the next coming years..

Tracy, can we go to the next question, please?.

We will now take our next question from Tim Anderson from Bernstein..

Thank you. A few questions, please. Can you just [Technical Difficulty] (32:26) maybe by mid 2018 [Technical Difficulty] (32:31) on I-O drugs..

Tim, you're really breaking up..

Can you hear me?.

A little better..

Is this any better?.

Yeah, that's a little better..

Can you just [Technical Difficulty] (32:49) perhaps which non I-O drugs [Technical Difficulty] (32:54) is, as more time has elapsed, you still kind of 100% confident [Technical Difficulty] (33:05) biomarker assay, [Technical Difficulty] (33:09)? And last question, again, tumor mutation burden.

Do you have any prospective data at this point with TMB, whether it's been released publicly or not? Looks like all the analyses are really done on retrospective data sets.

[Technical Difficulty] (33:29), out of Roche analyses and out of even your analyses of CheckMate-026 [Technical Difficulty] (33:37) consistent OS benefit, there's a clear PFS benefit [Technical Difficulty] (33:41) in either your data from CheckMate-026 or Roche's data, if you can kind of explain that?.

Tim, this is Giovanni. So unfortunately, you were breaking a little. First question, I believe, was about non-oncology assets and opportunities for potential registrational study starts going into 2018. As we've mentioned before, we are working actively on advancing three important programs, LAG-3, IDO, and our FGF21 program in NASH.

Over the next few months, we will also see proof-of-concept data on the BTK program and the TYK-2 program. And we look forward to seeing that data as well. So the pipeline in oncology and outside of oncology is progressing. You had two questions.

One was whether there is any updated perspective on our PD-1 test, I believe, and the answer there is, I don't think we have really any news to report there, but I'll ask Tom to answer that question, and also give you a perspective on the potential availability data with respect to TMB..

Yes. Again, Tim, our apologies, it was very, very hard to hear that question. I'd say, regarding the PD-L1 test, we presented data at recent meetings. It shows that we do believe the PD-L1 test we have is reproducible and is similar to that which is being used by other companies evaluating their drugs.

I think your second question about TMB and its ability to predict response rate PFS and overall survival, that's what sounded the question was about. I would say that, as with any biomarker, all those endpoints become incredibly important.

And we will look at response rate PFS and overall survival in any group of patients we would look at, in any given data set that we would look at, to be able to determine which is important. I think the relative balance between PFS and OS immuno-oncology is one that is incredibly important. And we love studies that show OS benefits.

I think, as many of our agents are beginning to show excellent second- and third-line activity in other areas, PFS may become something that we will look at more often. And I don't think there's any reason to believe we wouldn't look at that in a group of TMB patients in the future, as we would a group of PD-L1 selected patients in the future.

And I'm sorry if that's not exactly what you asked, but that's the best of our ability to hear..

Tracy, can we go to the next question, please?.

We will now take our next question from Tony Butler from Guggenheim Partners. Please go ahead..

Hey, this is Daniel (36:41) for Tony Butler.

Could you tell us a little bit about what data we can expect at SITC from your IDO1 inhibitor program? So, would we see some initial efficacy data and, maybe in general, do you see TDO1 or IDO2 as a problem when it comes to IDO1 inhibition? And then, second question concerning TMB, given the fact that mutation frequency in non-small cell lung cancer is highly variable across individual patients, what do you think is the proportion of patients in non-small cell lung cancer that are TMB high? Thank you..

So Tony (sic) [Daniel] (37:30), this is Tom. Thank you. I'll answer the first two questions on IDO. I think I'd say a couple of things about IDO in general. We think IDO is an important target, and we felt this way at Bristol-Myers for the past two to three years.

As you know, we purchased the compound from Flexus that is now the BMS IDO inhibitor, and we have a strong partnership with our colleagues at Incyte, which we're working with as well. And, as you know, we presented data with our IDO inhibitor at AACR this year.

This suggested that there may be some pharmacokinetic and pharmacodynamic advantages to our IDO inhibitor. We see excellent inhibition of kynurenine levels and we're able to see good receptor occupancy at doses that appear well tolerated in patients. Now, whether those PK/PD dynamics will be differentiating to the agent, I think we don't know.

I think only clinical trials will tell us if it matters for patients. But we think IDO is an important endpoint.

We think so much about it that we are proceeding with two separate lines of registrational studies, and we've talked about our plans with Incyte, and we also have plans to begin our studies in melanoma and in non-small cell lung cancer, with the Bristol IDO.

Now, as you mentioned, we will be presenting our first clinical efficacy data on IDO at SITC this year, and we look forward to seeing that data at the SITC meeting. I want to say a couple of things about what to expect.

You have to remember, this has been a remarkably quick development of this agent from Phase I, all the way into the point where we're now beginning registrational studies. And so, the data does take time to emerge.

You will be seeing some efficacy data on one or two tumors at SITC, but most importantly, you'll be seeing some safety data, which we think is very important to get out there at the SITC meeting as well. So some efficacy data and some safety data will be presented at SITC coming up this year.

So again, that will be our first important IDO data on our drug. And as we mentioned, we expect to have our studies with our drug registrational studies beginning at the end of this year. Second set of questions comes down to TMB, again, and thinking about what the mutational frequency will be in non-small cell lung cancer.

I think it's safe to say that we believe that this is something which is still not what 100% known at this point.

I think until you do prospective studies like we're doing in CheckMate 9LA, where you really have a sense of what the overall prevalence of different stratifications of TMB will be, where I think it's safe to say at this point is we think it's a substantial portion of patients with non-small cell lung cancer and that this is an important group of patients to treat.

I really do think that you need to see what the TMB percentage is in a group of prospectively treated patients. To really know what the final number is, whether it's 40% to 50%, I think those bands are fairly generous at this point and I think one really wants to see what that looks like.

I think what you can say is it's definitely an important percentage of patients with non-small cell lung cancer that may qualify as having elevated or higher TMB..

Thanks.

Tracy, can we go to the next question, please?.

We will now take our next question from Chris Schott from JPMorgan. Please go ahead..

Great. Thanks very much. Just two questions here.

First on the Opdivo-Yervoy combo in renal, can you just, A, talk about the commercial opportunity here as we think about both products? And B, I know it's a different indication, different dose, but any learnings from that study that influence either your thinking or confidence, whether it's around CheckMate-227 or some of the additional combo studies that you're running in different tumor types.

My second question was coming back to TMB. I guess how do you see this balancing relative to PD-1 status, which also seems to predict response in some tumor types, as we think about using these as predictive markers going forward? Thanks very much..

Chris, thank you. So I'll ask Murdo to give you some perspective on renal, and then, there are a couple of questions for Tom..

Yes, thanks for the question, Chris. When we think about the total renal opportunity for our I-O franchise, we're really excited about the Opdivo plus Yervoy regimen in frontline. Just scale-wise in terms of patient population, the frontline opportunity is roughly twice the size of the second-line opportunity in terms of treated patients.

So if you think about the second-line size of patient population being about 8,000 per year, you're about to double that in the U.S. in the first-line RCC treated setting.

So I think what we've been able to demonstrate in that trial is compelling overall survival data and obviously, there'll be an update for that presentation at SITC in a little while. And I think depending on labeling, et cetera, that could be a very, very interesting opportunity for us.

We have done a nice job in second-line share versus TKI's over 50% right now and holding. And because this is a slower-moving malignancy in terms of progression, the effect on second-line will take longer to materialize than you might see in other malignancies.

So at least for next year, you're going to have a nice overlap of an uptake in first-line, if all goes well, and a continuing robust business in second-line..

And Chris, I'll follow-up now with some comments on lung cancer. So the first question is what can you read-through from renal to lung from CheckMate-214 to CheckMate-227. And I think, as I've said in the past, it's very difficult to read-through on efficacy, because you're really looking at different tumors and slightly different schedules.

But what you can read-through a little bit on, I think, is tolerability. And I think what we're so happy about CheckMate-214 is that it was a well-tolerated regimen in CheckMate-214 and patients were able to tolerate the regimen. As you know, the CheckMate-227 dose is very similar of Opdivo-Yervoy, with Opdivo being three (44:14), Yervoy being one.

Although in the lung cancer regimen, Yervoy is given continuously as opposed to the renal. So there's slight differences in the length of duration, but the toxicity of the early part of the regimen looks very similar and tells us but (44:30) it gives us confidence that the tolerability will be good. And so, I think that that's important.

And tolerability is critical, because obviously, if you can't get dose into patients, you're not likely to see the benefits. So it's good to see tolerability in that setting. Your second question is a really good one, which is how does TMB relate to PD-L1 as a marker.

And this gets me thinking about how important it is to do good translational medicine studies and the translational medicine and science is really at the heart of what we're trying to do at Bristol-Myers Squibb from an R&D perspective.

Six months ago when I started, I told you it's going to be important for us to invest in our translational medicine capacity. We've done that. We're beginning – you can see our work on TMB as a good example of that.

I ask you to look at the SITC meeting, and you can see a lot of very interesting preclinical data looking at drugs like OX40 and IDO, and LAG-3 in terms of how they worked mechanistically. And also, this year, we're very happy to announce the appointment of Dr.

Saurabh Saha, who is now our new leader of translational medicine at Bristol-Myers Squibb, and he brings a wealth of experience in terms of thinking about how to design drugs, both in the cancer space and in the non-cancer space which brings me back to the question of how do TMB and PD-L1 relate to each other and will these be separate sets of patients? And again, great question.

And again, as a lung cancer person, I think about the diversity of tumor markers that we have in lung cancer. We have ALK, we have EGFR, we have RET fusion, we have RAS, we have ROS, we have TMB, and we have PD-L1. We know already that while EGFR and ALK don't usually occur together, we certainly can have situations of co-mutation.

We think that RAS tends to be exclusive, but we have seen episodes or examples where you can see both of them. These are the kinds of questions that we're going to need to define prospectively what that relationship between TMB and PD-L1 are, to help us define the groups of patients that are going to benefit most from our drugs.

So really interesting questions to look at in the next couple of years..

Tracy, can we go to the next question, please?.

We will now take our next question from David Risinger from Morgan Stanley. Please go ahead..

Yes, thanks very much and thanks for providing all the details. I wanted to just ask a couple of follow-ups on CheckMate-227.

So could you just clarify your message on CheckMate-227, please? Specifically, is TMB going to be used as part of the primary endpoint analysis and if you set, what TMB high versus medium versus low means? And then, second, given the materiality, should we expect a press release on the CheckMate-227 interim once the look occurs no matter what the results are? Thank you..

David, this is Giovanni. Thanks for your question. Let me just say, the discussion we've had today on CheckMate-227 is really very consistent with what we've said before with respect to the timing for the final analysis, the potential and timing for the interim analysis, the flexibility we retain with respect to the statistical plan.

And at this point, we are really not in a position to provide any additional insight on the study..

And I think, David, with regard to our press release, we'll evaluate that at the time. I think it's premature to speculate on that.

Tracy, can we go to the next question, please?.

We will now take our next question from Vamil Divan from Credit Suisse. Please go ahead..

Great. Thanks so much for taking my question. So one question, just following up I think on the prior comments, regarding CheckMate-214 and the read-through to lung. And you mentioned you likely saw (48:41) on the safety side. I'm just curious, we did still see I think it's about 22% of patients that discontinued due to side effects in CheckMate-214.

Is that level of discontinuation due to the adverse events okay in lung cancer or do you think you need to have a better profile when we see that data? And then, second, with all the focus on the combination, I was just curious maybe you can give us a little more color on what's going on in the market right now with melanoma and just in terms of the percentage of use that's with PD-1 monotherapy, how much if it's a (49:13) combination with Yervoy? Is there a difference in terms of community-based doctors versus academic centers or U.S.

ex-U.S.? Just how to think about where the use of the combination maybe more and where is monotherapy doing there? Thanks..

So Vamil, thank you for your first question. Again, it's a very interesting point you raised in terms of what level of drug do you need to get into patients to see a benefit. It's really hypothetical, but it's impossible to know at this point what the level is.

What you can say it in CheckMate-214 is, is that that level of discontinuation did not force the survival benefit that was seen in renal cell. Whether that is seen in lung or not, we really can't answer until we see the data, but it's a very valid question..

Yeah. And Vamil, just on metastatic melanoma, we continue to see really strong first-line marker (50:14) shares for the regimen. So roughly 30% of all new first-line metastatic melanoma patients receive Yervoy plus Opdivo in combination. And then, if you add up all the remaining monotherapy, it's about another 40%.

So you've got a large percentage in combination, a large percentage in Opdivo, and then, the rest made up of KEYTRUDA.

We're also seeing some changing dynamics in the adjuvant setting, so that will also change some of what's happening in metastatic, with the update to the NCCN guidelines and hopefully, in the future an approval from the FDA, we would expect rapid uptake of Opdivo in the adjuvant setting, which will change dynamics on Yervoy.

So if you're just looking at Yervoy sales, the dynamic underneath Yervoy sales will be declined to some extent on the off-label 3-milligram use that's currently in adjuvant, being replaced by Opdivo 3-milligram and then, continuing to be used in combination with Opdivo in metastatic.

So just a dynamic to keep in mind as you think about total melanoma going forward, but we have some strengthening trends on our Opdivo plus Yervoy regimen, and we're very focused on trying to improve our performance in BRAF patients..

Thanks, Vamil.

Can we go to the next question, please, Tracy?.

We will now take our next question from Gregg Gilbert with Deutsche Bank. Please go ahead..

Thank you, folks. I have a few.

Tom, I know you'll say it depends on the data, but how important do you think Incyte's IDO readout in melanoma is in the first half of next year, both in terms of melanoma, but also read across to other settings? Secondly, on the NASH compound, are you confident that you have FDA's buy-in yet, as it relates to this compounding Phase III ready, based on how you did Phase II? And just a cleanup one on the Halozyme collaboration you mentioned, is this just a chair time benefit, sort of shorter infusion time angle, or is there more to that story? Thanks..

So Gregg, thanks for your question.

So first, regarding the Incyte, I will give you a little more color than saying it depends upon the data, because one of the things that I think that we have, one of the things that's influenced our decision-making about IDO as a class, are some of the interesting data that we've seen with our own drug, with epacadostat.

What we've seen with epacadostat and Opdivo, we've seen very promising data with epacadostat and Opdivo. And so, I think that looking at the readout that comes from Pembro and epacadostat, I think will certainly be of interest to us as we go forward.

Now how this relates to other tumor types, melanoma and – I mean, how it relates to lung cancer and kidney cancer and neck cancer and whatever other cancers we look at IDO in, I think that's very difficult to know. But we will, obviously, be paying attention to that.

I think there's a couple things you have to remember between – with antibodies, when you think about the difference between antibodies and small molecules, antibodies are remarkably specific and they have high specificity for their targets.

Small molecules do differ, and they can differ by a methyl group here or there that could explain both differences in activity and differences in toxicity. So I think we have to be a little careful about looking too much across readouts of different agents and, ultimately, at the end, it does depend upon the data, so you're correct about that.

Second question comes down to where we are with FGF21 and NASH. And again, this is a really important area, something that – an area that we're very excited to be in.

It's an area of tremendous excitement and, as I'm sure you know, last week were the big liver meetings, and these were terrifically important meetings where data from a number of companies, including ours, were presented, and we presented an update of our data at an oral session at The Liver Meeting showing that our FGF21 drug is able to reduce hepatic fat.

We're in the middle of discussions with the FDA, in order to be able to design the appropriate studies to be able to register our drug, and we look forward to coming up with what the final study design will be in consultation with them.

I think one thing that is important to note is that we do believe that biopsy will be necessary in a Phase III trial, to be able to demonstrate efficacies, that indirect or surrogate markers like hepatic fat or Pro-C3 probably won't be enough, and the FDA said as much last week at The Liver Forum that they had, where they basically said they'd like to see more biopsy data from companies, and that was not specific to Bristol-Myers, that was a comment to all the companies that were developing drugs in that setting.

So we're looking at our options in terms of how to operationalize this, and we look forward to being able to announce, on clinicaltrials.gov, our plans for moving forward with the timing of FGF21. But you're correct, it will require biopsy as an endpoint in those studies. And I'll turn it over to Murdo on Halozyme..

Yeah. Thanks, Gregg. On Halozyme, we're really excited about this opportunity to partner with them on their technology. And yes, you highlight, there's definitely a large patient convenience and infusion chair time benefit to having an injectable version, versus an infusion version of our products.

But we also have to think beyond this, in terms of where site of care may take place in the future. It is possible that site of care may move from academic hospitals, academic cancer centers, to more office-based treatment, particularly if patients are on these treatments beyond a one-year timeframe and they have large travel distances.

So that's one area. The other thing to think about is, at least in the U.S., an injectable product will be reimbursed through a different payment channel, so that's something to think about as a potential advantage in the future.

And last but not least, we have over 11 different assets in this agreement that we could use potentially in combination with one another in different ways, that this technology might allow us to improve again how patients receive their administrations.

Now, this is also something that is related to new payment models going forward in terms of outcomes for patients. And I think patient reported outcomes and how they experience treatment could be important. So there's a number of different strategic advantages to that agreement..

Gregg, let me just add on that, because I think it's really important as part of our strategy as a company, given the number of assets we have in development, the potential for combinations, to continue to lead in terms of innovating delivery models. Murdo's comments are really important in the U.S.

A significant part of our business is outside of the U.S., where the constraints with resources at hospital level and the number of distribution on oncology sites is, in some cases, a limiting factor.

The other thing I would say is that one of our key priorities across multiple tumor types has really been, from the beginning, to accelerate immuno-oncology into the adjuvant setting.

And clearly, when you look at patients that are healthy and (58:09) in the adjuvant setting, clearly, thinking about different delivery models is going to be very important for us as well..

Well, Tracy, we're coming up to the bottom of the hour, I think we have time for one more question..

We will now take our next question from Geoff Meacham from Barclays..

Good morning, guys and thanks for the question. I just have a couple of quick ones. You guys are committed to CTLA-4 as an important mechanism in I-O. Would it make sense to make next gen CTLA-4 as the foundational combo going forward? I wasn't sure what further de-risking is needed for that program.

And then, on the quarter, I wanted to follow up on an earlier question on second-line lung trends. Opdivo sure looks to be largely stable in the U.S., but what can you tell us on overall patient flows and new start trends? Is KEYTRUDA used in first-line having an impact downstream? Thank you..

So Geoff, thanks for the question on CTLA-4. We believe CTLA-4 is incredibly important to us. It's a mechanism that we think, again, as I said earlier that the Opdivo-Yervoy combination has delivered and continues to deliver and continues to improve impact for patients broadly.

Now, we do believe that there is some opportunities for looking at our next gen CTLA-4 compounds. And there are two of them. One is the non-fucosylated formulation, which seems to have greater potency and that has entered clinical trials this year.

And the second is the Probody technology that we're looking at with our partner, CytomX, again, that looks very exciting and we hope to begin clinical trials in that very soon.

So I think with either of those two approaches, once we see evidence of both the safety and efficacy, we would certainly be inclined at that point to move those agents forward as a partner for Opdivo.

I do believe that they have the potential – just like CTLA-4 is foundational to us now, they have the potential to offer that kind of benefit to patients down the road..

Yes, and Geoff, just on second-line, the shares, as you mentioned, are holding up nicely. That's a very strong performance commercially and with our medical organizations. We are seeing slightly lower impact on the second-line eligible pool of immuno-oncology patients.

So that percentage reduction is roughly in the neighborhood of 5% in terms of volume of patients who have not seen an I-O agent in first-line before. So that's a little bit less of a reduction than we'd anticipated, so that's helping Opdivo. We're also holding onto some first-line non-promoted share (01:00:47) and that is also stable.

With respect to the penetration of Pembro in first-line, it's as we expected..

So thanks, everyone. Just in summary, at the end of the call, let me just say this was a good quarter for us. I am pleased with continuing strong trends with respect to commercial execution. We made good progress on a number of fronts in the R&D pipeline and execution overall.

We have a number of important events ahead of us, which have the opportunity to continue to broaden the set of opportunities we have ahead of us. And I look forward to continuing, obviously, to talk to all of you. Thanks, again, and have a good day..

This concludes today's call..

Tracy, thanks..

Thank you for your participation. You may now disconnect..