Good morning. My name is Tiffany, and I will be your conference operator today. At this time, I would like to welcome everyone to the Bristol-Myers Squibb 2016 third quarter earnings results conference call. All lines have been placed on mute to prevent any background noise. After the speakers' remarks there will be a question and answer session. John Elicker, you may begin your conference.

Thank you, Tiffany, and good morning, everybody. Thanks for joining the call to discuss our Q3 results and additional announcements that you've seen reflected in our press release. Joining me this morning are Giovanni Caforio, our Chief Executive Officer, and Charlie Bancroft, our Chief Financial Officer. Both Giovanni and Charlie will have prepared remarks. In addition, joining Giovanni and Charlie for the Q&A portion will be Murdo Gordon, who's our Chief Commercial Officer; Francis Cuss, our Chief Scientific Officer; and, given the likely discussion we'll have in the immuno-oncology space, we also have Fouad Namouni, our Head of Oncology Development. Before we get started, I'll take care of the Safe Harbor language. During the 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 our non-GAAP financial measures, which are adjusted to exclude certain specified items. Reconciliations of these financial measures to the most comparable GAAP measures are available on our website. Giovanni?

Thank you, John, and good morning, everyone. We have a lot to talk about this morning. So I'll start by summarizing my views on the performance in the quarter and then turn to our guidance for 2016 and our thinking about 2017. I'll share some thoughts about our long-term strategy and then hand it over to Charlie to provide additional details about our performance and operations. So to begin we just finished another very good quarter. Our business fundamentals are strong as evidenced by our commercial performance. We delivered $4.9 billion in revenues, a full 21% increase over the last year, thanks to good business trends across our portfolio. We made important regulatory advances with Opdivo. We see the positive advisory opinion for the treatment of classical Hodgkin's lymphoma in Europe. Our application for advanced bladder cancer has been accepted for priority review in the U.S. and validated in Europe. And, in the next few weeks, we expect to get a decision on our application for Opdivo for head and neck cancer in the U.S. We had disappointing results from CheckMate -026. The results of these trials do not, however, change our strategy. We've applied learnings from the trial to our entire development program, and we are advancing a broad set of opportunities for Opdivo. We see an even more important role for combinations. Continue to be excited by the longer-term follow-up from study -012 and believe study -227 has the potential to bring real value to patients. More generally, we remain fully confident in our Opdivo development program, which is both broad and promising, with data expected from eight potentially registrational trials over the next 12 months. And, based on our current assumptions for 2017, we expect Opdivo sales to grow globally as well as in the U.S. The…

Thank you, Giovanni. Good morning, everyone. This was indeed an outstanding quarter operationally for the company. And that is particularly evident when you look at our commercial performance. We saw solid growth from our key products, leading to 21% sales growth and an almost doubling of EPS compared to Q3 last year. Overall, FX had a negative impact on EPS of about $0.02. As Giovanni mentioned, we saw strong commercial execution in the quarter. Let me provide some color. Opdivo demand remained strong. In the U.S., we recorded $712 million of sales, up 11% sequentially. Opdivo-based therapy has become established as standard of care in the U.S. across melanoma, renal cell, and second-line non-small-cell lung cancer. While our second-line lung business will be under some pressure, we believe we have opportunities for growth in renal, melanoma, and Hodgkin's, as well as future indications such as bladder and head and neck. In addition, over the next 12 to 18 months, we expect to see important read-outs for several tumor types that we believe will position the brand for further growth. Outside the U.S., we recorded $208 million for Opdivo during the quarter. This does not include approximately $150 million of gross sales for France and Germany in the quarter, which are being deferred. The amount we will actually recognize depends on the final price in each country, and we do expect to reach an agreement for both later this year. From an operational perspective, we continue to see strong adoption in markets where Opdivo has been launched. For example, we continue to see 70% to 80% of PD-1 share in France and Germany. We have also made good process this quarter securing reimbursement among key EU markets such as Denmark and Sweden. Looking beyond Opdivo, we are particularly encouraged to see…

Okay, thanks, Giovanni and Charlie. Tiffany, I think we're ready to go to the Q&A. And, just to remind everybody, in addition to Giovanni and Charlie, Murdo, Francis and Fouad are here to answer any questions you might have. Tiffany?

Your first question comes from the line of Tim Anderson with Bernstein. Your line is open. Timothy Minton Anderson - Sanford C. Bernstein & Co. LLC: Thank you very much. At ESMO, you hinted at a possible early filing strategy with the combination. And I'm wondering if you can share any more details on what this could entail? I think a lot of investors believe this could happen by packaging -012 and -568 together. Would this be the most likely route to early approval? And then on -227, lots of consideration whether that design and that trial will change, and I'm wondering when you'll have more to say on that. And on cut-offs for the different arms, is it possible you could have a different cut-off for monotherapy arm versus the combination arm?

Good morning, Tim. Let me confirm that we did say at ESMO that we had identified a potential approach for an earlier submission. We continue to look at that, but of course it's dependent on data and the agreement of the regulatory authorities, but I'm not going to say more about that for competitive reasons at this point.

Good morning Tim, this is Fouad. For -227, it's a large Phase 3 study, and as we said, you can think of it to be two large Phase 3 studies in one protocol. We are looking at the non-expressers and the expressers, and the study is continuing to recruit, probably for the next two or three weeks in the expressers. And the size of the study will certainly be larger than what was planned initially. We will be obviously looking at these patients in the expresser group before the non-expresser, given timing of the recruitment. This study gives us a lot of optionalities in terms of looking at different cut-offs, looking at different times in the study, and looking at different endpoints in terms of PFS and overall survival. So we do have – we believe we do have a very good optionality in terms of how we think about the steps of the study and how we think about the timing of the endpoints.

Yeah, Tim let me just comment. While we do have an approach for the potential approval combination next year, that's not included in how we think about our 2017 guidance.

Thanks, Tim. Tiffany, can we go to the next question, please?

Your next question comes from the line of Chris Schott with JPMorgan. Your line is open.

Great. Thanks very much for the questions. The first one is just coming back to the 2017 guidance. What gives you confidence providing this EPS range when we've yet to see the rollout of Keytruda in first-line lung and what that could mean for Opdivo? And could you elaborate a little bit more on the assumptions that you're factoring into this preliminary guidance? I guess specifically what gives you confidence that Opdivo can grow in the U.S. next year? My second question is just coming back to -568. I believe several of your Opdivo indications have been either filed or approved based on single-arm Phase 2 studies – I think like Hodgkin's as an example. How do we think about, just maybe more broadly, indications that you would be able to file on Phase 2 data versus indications where you're going to need fully controlled studies? Is this the size of the indication? Is it the lack of a standard of care, existing therapies? Is it the strength of the data? I'm just trying to put in context what we need to think about to look at Phase 2 data versus the broader Phase 3 programs. Thank you.

Chris, this is Giovanni. Let me start, and Murdo will give you some comments on assumptions regarding Opdivo for next year, and then maybe Francis can answer your third question on regulatory strategies. When we think about 2017, obviously we look at the totality of our business. There are very strong trends for our business, as you saw in Q3. We see Eliquis continuing to grow and continue to make inroads in terms of market shares in the U.S. and internationally. We see continued strong performance from our in-line business with Orencia and Sprycel, and we do see strong momentum globally with Opdivo, which includes clearly the three big areas we look at when we look at our business: the international launches that are progressing extremely well, an expansion of our business across multiple indications in the U.S., and then obviously the lung business, which as Charlie said, will be under some pressure. When we look at the totality of those dynamics, we feel fairly confident in the ability of Opdivo to continue to grow, and in the outlook for the total business.

Yeah, Chris, maybe just a little bit more detail. When we think about where we are currently in lung, we feel obviously very good about performance to date, with roughly 60% to 70% share overall in all-comers in second line lung. Now, clearly we're going to have competitive pressure with both the approval of pembro in first line and the reduction of that population of I-O-treated patients from the eligible second-line pool. That will be somewhat rate-limited by testing. We do anticipate that there will be an acceleration in the testing rate, but there's still a slope to that curve, and while Merck will rapidly penetrate that patient population, there will be a certain cadence to how that will occur because of the testing rates. Testing rates right now are about 50%, and we have seen a slight acceleration in that rate more recently. When we think about the second-line opportunity, we continue to feel very good about our profile on the basis of -017 and -057, but clearly second-line shares are going to be under pressure from the advent of Roche in the second-line setting. The one thing I'll just clarify is while Merck were able to secure a label update in second line at the greater than 1% cut-off, that information was included in NCCN Compendia for many, many months now. So that's not necessarily a new competitive event for us in second line. Overall, some of the things that we continue to feel good about in second line are just our competitiveness from a share of voice, as well as the breadth of indications across Opdivo. So while we expect some share pressure in second-line and lung, other sources of growth are other indications, namely the advancement of the regimen in melanoma and that upwards of 40% share first-line melanoma across both types of patients, BRAF positive and mutant. Renal cell carcinoma, our business continues to do well there. We're hopeful we'll have a head and neck indication soon. And then next year, hopefully also bladder in the U.S. As Giovanni mentioned, we're also very pleased with the expansion of our business outside the U.S., where the clear demand trends are very good post-reimbursement events. You can't see all that in our revenue. As Charlie mentioned, we're deferring sales in France and Germany in the quarter, but really our shares are upwards of 70% overall PD-1 dollar market outside of the U.S. So really we're feeling good about a very balanced growth portfolio in I-O. And last but not least – it was mentioned by Charlie – Yervoy has returned to growth in the U.S., and in both France and Germany, where we have the regimen reimbursed, we've seen very good evolution there, too.

Good morning, Chris. So let me reemphasize that we are working with regulatory authorities to find strategies to accelerate our regulatory submissions, whether it's optionality in Phase 3 studies or in the striking data in Phase 2 studies, depending on the circumstances. Just to talk for a moment about CheckMate -568, just remind you that it's a single-arm study looking at the combination of Yervoy and Opdivo in first-line lung. The primary endpoint here is response rate, and it gives us a lot of flexibility to look when and how often we look at the data, and I think we see it as complementary to the CheckMate -012 data. It's particularly useful to provide some information on current and future medical practice, particularly with an I-O/I-O combination. We believe it'll be helpful in characterizing the high response rates we saw in CheckMate -012. And, importantly, it's evaluating the same dose as we've got in CheckMate -227. And we may continue to use -568 to obtain additional data and answer additional questions going forward.

Thank you, Francis.

Thanks, Chris. Tiffany, can we go to the next question, please?

Your next question comes from the line of Gregg Gilbert with Deutsche Bank. Your line is open.

Thank you. First for Giovanni. You talked about learnings from the study that failed. Can you be more specific on what learnings you have applied to the rest of your Opdivo development strategy? Of course without sharing any trade secrets. Would love to understand what that means. Longer term on your OpEx goal of flat in 2020 versus 2016, Charlie, can you talk about the theme there? Is it sort of continuously growing R&D offset by shrinking SG&A, and within that, have you already made decisions about geographies or therapeutic areas that you've decided to de-emphasize? I think it would be a pretty big feat to keep OpEx flat over that timeframe when you're viewed as sort of an optimized, pretty lean model already. Thanks.

Gregg, let me just start by going back for a second about the operating model transformation, and then I'll ask Fouad to comment on learnings from study -026 briefly. As both Charlie and I mentioned, continuing to evolve the way we allocate resources in the company is really driven by the opportunities we see in the portfolio. As we've said in the past, when you look at the next three to five years, we have a great opportunity for growth, which is really driven by the performance of our immuno-oncology franchise with Opdivo and Yervoy and Eliquis as significant drivers of growth. So that opportunity is quite significant. It's really concentrated in terms of the therapeutic areas and also quite concentrated in the top 20 markets around the world. At the same time, we are advancing an early pipeline, which continues to make progress. In that pipeline as time advances, there are programs that are emerging as very high-potential programs. And we are ready to make the investment decisions needed in order to accelerate those programs as the data evolves. As I commented earlier also, we do see a continued opportunity to streamline our G&A, to streamline our infrastructure, to look at our manufacturing network all the time, driven by our portfolio. So I see that we continue to have significant opportunities to allocate resources strategically between now and 2020, and we're quite confident in our ability to do that.

The learnings from CheckMate -026 – and I would add also other PD-1 data we have seen recently – and we believe that it is important that for a program to be successful at improving survival in first line non-small-cell lung cancer in this case, requires the following items. Response rates need, really, to be high. Response rates need to be fast and deep. More importantly, responses must be durable over time. And based on the nature – and of course I would add safety is extremely important for the profile of the program. Based on this and based on the nature of the data we have from Opdivo plus Yervoy and the long-term follow-up that we disclosed recently for the combination of these two agents and from CheckMate -012, we believe that CheckMate -227 has potential to fulfill this criteria. Specifically when you look at Yervoy, which is a Treg depleting agent, optimally spaced at every six weeks with Opdivo, this really becoming probably or potentially a solid backbone moving forward. We view this not only in lung cancer, but we view this combination of immunotherapy agents also working and active in a variety of cancer. I would finish by saying that each time we added a CTLA-4 blockade with Yervoy to PD-1 Opdivo, we have seen the activity increase and almost doubling in a variety of cancer – and importantly, with a long durability of the effect.

Thanks, Gregg. Tiffany, can we go to the next question, please?

Your next question comes from the line of Seamus Fernandez with Leerink. Your line is open.

Oh, thanks very much for the question. So a few here. Maybe first, can you guys discuss a little bit of how the imbalances seen in CheckMate -026 might have occurred? So as you kind of go back and review the trial conduct, have you identified how the imbalances discussed previously occurred, and then how you hope to and plan to avoid this in the CheckMate -227 study? The second question is more for Charlie. As of the end of this quarter, could you just update us on how much has actually been reserved for in France and Germany? I know you guys update that in the Q every quarter now. And just a follow-up to that is, should we assume that's booked this year based on your comments? And does the growth outlook for Opdivo assume that those sales booked are in the base as we think about growth year over year? Thanks.

All right. So let me start with your question on the deferral. So in the quarter, between France and Germany, as I mentioned in my comments, Seamus, we deferred $150 million. Year-to-date it's $265 million. You can think back to Daklinza last year in France, where we also had a similar deferral issue, which we then ultimately booked. As we think about the growth year over year moving from 2016 into 2017, we do assume that the net revenue that we'll book for France and Germany will be in our 2016 results.

Good morning, Seamus. Essentially we believe that CheckMate -026 was unsuccessful as a result of the stratification at the 5% cut-off rather than the narrower patient selection at the higher cut-off of 50%. And in the event of this, this was compounded by unfavorable patient selection and imbalances, as you mentioned, particularly actually that exploratory 50% cut-off. Now, this is a view we – it's now supported by a number of lung and statistical experts we've consulted. And basically we agree with the opinion of Dr. Socinski, who's you'll recall the presenter of the CheckMate -026, that had the Keynote -024 design been used with Opdivo rather than Keytruda, the result would likely have been similar. So as we shift our focus towards CheckMate -227, we're confident that not only the profile of Opdivo-Yervoy regimen, but the lessons we've learned from -026 – and I'll ask Fouad to give you some of those very specifics – when applied to -227 will actually allow for a much better chance of success.

Thank you, Francis. Good morning, Seamus. Two things to see and understand in -227 varying from -026. So first and foremost, going in in -227 with the combination data is different level of activity that we have seen with monotherapy in -026. It's higher response rate, it's much longer PFS with a solid follow-up at this time and an improved safety profile from what we knew in the regimen before. Actually, we allowed the study in the expresser to continue to recruit, and I said earlier, it's nearing probably two to three weeks' time before finishing the recruitment with a larger number of patients than initially planned, and we would expect in the three arms that the size will be nearly doubling what we have seen in CheckMate -026. So this in itself is important to correct for the imbalances, the way we look at it. And I would add obviously one more thing that is different in -227 is we did not build in a crossover like we did in CheckMate -026. And as practical matter, it means the patients that have been randomized to chemotherapy can receive the appropriate standard of care, including PD-1 monotherapy, in second line as and where it is commercially available. However, it's important to note that patients will not receive the regimen of Opdivo-Yervoy when their tumor progress.

Thanks, Seamus. Could we go to the next question, Tiffany?

Your next question comes from the line of Jami Rubin with Goldman Sachs. Your line is open. Jami Rubin - Goldman Sachs & Co.: Thank you. Just a couple follow-up questions on the same topics. What has been the feedback you've gotten from physicians on the fact that Opdivo is infused every two weeks versus Keytruda and now Tecentriq every three weeks? I would think that the longer infusion requirements would be an advantage, particularly in Europe, where they have limited budgets. And I'm just curious to know what your thoughts on that is and how that's going to play out. Also just going back to Seamus' question. I think one of the other issues, too, with CheckMate -026 and Keynote -024 is that both trials were recruiting around the same time, and there was a lot of overlap in clinical trial sites. You guys ended up with the B patients; they got the A patients. Is that a risk with MYSTIC and CheckMate -227? I mean, obviously AstraZeneca's running their own I-O/I-O trial. And what are you doing just to avoid that? I mean, you sort of answered it, but do you have any control over the recruitment of imbalances in patients? Because I think, again, that was something that kind of came as a surprise. Thanks.

Jami, it's Murdo. I'll answer the first question and then pass it over. Feedback from customers has been very positive, actually, on our dosing schedule. There are some customers, to your point, where infusion capacity is constrained, and we've actually seen that improve over time since the launch. So, for example, in the early launch phase in Germany, infusion capacity was somewhat constrained, but it has grown since. So the three week versus two week tends to be an issue that gets considered later in the treatment consideration. So the first consideration's obviously efficacy, where the profile of Opdivo is very strong across all histologies and across all types of PDL-1 expression. And then the other thing that's really influencing treatment decision-making is the breadth of indications that we are now developing for Opdivo. And then I would say the dosing cadence comes into consideration. I'd also stress that in the community setting in the United States in particular, the dosing interval of two week versus three week tends not to be an issue.

And hi, Jami, this is Fouad. Let me answer the second question around competing trial clinical sites, the same clinical sites, and what are the implications for our combination of Opdivo and Yervoy. Actually, sometimes we do have competing trials, and probably you have seen this when we did the monotherapy experiments. What we are seeing in CheckMate -227, is slight difference with -026, is a large number of clinical sites are outside of the United States. We are not concerned about competing trial given the population of patients we are recruiting and given the size. We think the recruitment is going well with -227. The profile of the program in terms of safety and experience of the investigator has been good, so we feel confident that -227 is at the right site of investigation, and the investigator are doing pretty good job there.

Thank, Jami. Can we go to the next question, Tiffany?

Your next question comes from the line of Steve Scala from Cowen. Your line is open. Steve Scala - Cowen & Co. LLC: Thank you. A couple questions. First for Murdo. How often do you expect PDL-1 testing to be repeated in the second line after being done in the first line, as opposed to testing only in the first line and then using that information as a guide to the second line? If another test is thought necessary but not elected, then that would be of course good for Opdivo's second-line share. If it weren't thought necessary, then Merck's position might be strengthened. So would like your thoughts there. And then the second question is, consensus earnings in 2020 are $4.46. To what extent do you believe it already reflects the impact of the operating expense guidance provided today? Not the guidance itself, but the impact of that guidance. So would like your thoughts there. Thank you very much.

Steve, thanks for the question. Maybe just a little bit of background on what we're seeing happening right now. Most patients are being tested in the first-line setting when tissue is available. I think it would be unlikely that they would be rebiopsied and retested in the second line, just primarily because of the trauma to the patient of actually obtaining tissue. And I think, in general, people believe PDL-1 status to be relevant even in old assays that are done. Albeit the newer studies have been done with fresh biopsies, as you saw in the Merck -024 study. We think second-line retesting will be a small percentage of patients. I will say, though, we have been focusing on the second-line-tested patients and making sure that oncologists understand the strength of our data in high-expressing second-line patients from -017 and -057, and we have seen significant increases in our share of that population in second line. We currently get about 50% of those patients and pembro getting the other 50%. So I think we're very competitive, even in a tested second-line patient population.

And Steve, this is Giovanni. With respect to your second question, we're not providing long-term guidance. As I mentioned before, we feel comfortable with our ability to evolve our operating model and manage expenses within the context I described. And in that context of flat OpEx, we are comfortable that we're going to continue to be very competitive in driving R&D execution and commercial execution at the same time.

Thanks, Steve. Tiffany, next question, please.

Your next question comes from the line of Andrew Baum with Citi. Your line is open.

Thank you. Couple of questions, please. Regarding the -027 guidance, should I be thinking it's based on the assumption that the negative impact in the first line from Keytruda isn't going to be seen next year, given the time taken for these patients to progress and therefore it's coming in 2018? And then, second, on CheckMate -026, Francis or Fouad, if you have to hand whether there was any notable differences in the investigator-determined PFS versus the independent-review PFS, particularly for the chemo arm? I'm just curious about the very heavy censoring that you saw in that trial. And then, sorry, one further question. The forest plot shows a difference between squams and non-squams. We never saw the Kaplan-Meier curves. Is that data of potential interest given what you saw in the second-line setting? Thank you.

Andrew, thank you. Let me just start by answer your question about 2017 and 2018. And let me maybe step back and answer your question by saying we definitely see the lung market to be extremely dynamic. We have included the impact we think will happen in 2017 based on the penetration of Keytruda in the first-line setting in the segment you know and how that impacts, as Murdo says, the size of the second-line opportunity because of patients going through. And also, the competitive nature of second-line lung, in which, as Murdo said, we continue to be very well-positioned. I personally think about the lung market particularly when it comes to first line as a very dynamic market, where obviously in 2017 there will be the use of Keytruda monotherapy in high expressers. But ultimately, beginning in 2018 and clearly going forward, I believe it is clear that this will be a combination market, and so you'll see rapidly evolving market dynamics in first-line lung, and over time those impacting second line as well. But those impacts are reflected in our 2017 guidance.

And for the PFS question, Andrew, the PFS was consistent between the investigators and the independent review committee, to answer the question around censoring. For the question around the squamous versus non-squamous, yes, we have reported a better risk reduction on progression and survival in the squamous versus non-squamous as presented at ESMO. This is probably telling us that in the squamous biology of a very inflamed tumor that that's the type of tumor that responds to a PD-1 agent. In the non-squamous, there are tumors that are inflamed, and there are some groups that really see probably some level of inflammation like the squamous. And this could be detected by enriching in the PD-L1 biomarker to have as a marker the inflammation in the non-squamous. In -026 really were broad. We stratified with a 5% cut-off; we may have less inflamed tumor in the adenocarcinoma group. And probably a much better enrichment for inflammation would have showed better results for the adenocarcinoma group of patients. So these are interesting data. The full data will be available at the time we publish it, so you'll have the opportunity to look at it.

Thanks, Andrew. Next question, Tiffany, please?

The next question comes from the line of Marc Goodman with UBS. Your line is open.

Morning. Obviously there's been a major change in the stock price, and I was curious, are investors missing something? What's not appreciated with respect to how you view things versus how other people are viewing things? Second, can you give us a breakdown of Opdivo revenues by indication like you usually do? And, third, what can we expect at the upcoming meetings for the rest of the year – ASH, lung? Thank you.

Marc, this is Giovanni. Let me just go back and reiterate what I said earlier. From our perspective, our fundamentals in the business are very strong. We do have good trends across our business that will continue into next year. Obviously, we believed it was important to really understand the impact that the first-line lung dynamics next year would have on our business. We have reflected that in our 2017 guidance, and I feel very optimistic about long-term growth prospects for the company.

Yeah. And just related to your question, Marc, on the mix of our business by indication, I would think of it in these three buckets: about a third, roughly, I would say between 30% and 40% of our business currently is in lung in the U.S.; then about 20% to 30% of our business is other indications in the U.S.; and then 25% to 35% of our business is ex U.S.

Good morning, Marc. So we have a number of presentations regarding our new I-O agents at the upcoming SITC meeting. These will include safety and efficacy for lirilumab, the anti-LAG-3 for urelumab as well in these early studies, both alone and in combination with Opdivo, and both in solid and hematological tumors. Just to remind you, we do have a PDUFA date coming up for head and neck on November 11. We've had validation of our European submission based on overall survival. We will get – approval is expected before the end of the year in Europe for classical Hodgkin's lymphoma, and obviously in the first quarter next year, we have a PDUFA date for bladder in the United States as well. So there's a lot going to be happening in the next few months.

Thanks, Marc. Next question, please, Tiffany.

Your next question comes from the line of Geoff Meacham with Barclays. Your line is open.

Morning, guys. Thanks for taking the question. Just wanted to get some perspective on the second line. When you look at the potential for sequential I-O use, so Keytruda followed by Opdivo, for example, mechanistically it doesn't make any sense, but I wanted to ask whether it would be worth it to explore that in a larger study. And two, do you think payers could try to limit sequential use? And then just as a follow-up on your OpEx guidance long term. Would you guys characterize this as a change in the R&D organization, having a more disciplined approach? For example, has the hurdle rate going into more expensive Phase 3 changed?

Geoff, let me start from the investment perspective in your second question. We see – first of all, our evolution of the operating model is very consistent in a way with what we've been discussing with all of you over the last few months in terms of allocating resources strategically, accelerating the pipeline, and generating the right leverage as we grow as a company. We've quantified how we can do that more and better today. And my perspective is that from an R&D point of view, it really is about having the resources necessary in order to invest in the highest potential programs. So I would not characterize that as having raised the hurdle. I would characterize that as having the ability to allocate the right resources to R&D in order to accelerate promising programs. And, as Charlie mentioned, as an example next year, we are increasing R&D spend because we see very strong momentum in the early pipeline and clearly with Opdivo and Yervoy. Clearly, on the same side at the same level commercially, we built very competitive commercial organizations in the key markets, and that we continue to be really committed to. It's really about disproportionate resource allocation to the highest-priority areas. And there will be parts of the company in fact that grow or experience very little change, while we will be making decisions to reduce our focus in parts of the company like, for example, some of the G&A functions where we think we can become a leaner organization, actually accelerating the ability to execute at the same time.

And, Geoff, just on your other question about sequencing the treatment of lung cancer patients, what we've seen in the market to date has been really very predictable in terms of what the oncologists and payers are doing, is they're responding to data events that trigger NCCN guidelines and compendia updates and also are respectful of the labeled indications of the products. So I would say in our forecast, we have not assumed any retreatment of pembrolizumab first-line exposed patients in our second-line opportunity. We have taken them out of our second-line pool of eligible patients. And I would say it's very likely, given the clarity of the NCCN Compendium and the clarity of the pembrolizumab indications, that the pre-treated or first-line treated patient population will be very close to the label, greater than 50% expression, and so that limits the pool of previously I-O treated patients that get excluded from second line, at least quite a large pool of second-line patients that are still eligible for Opdivo treatment in the second line.

Thanks, Geoff. Tiffany, can we go to the next question, please?

Your next question comes from the line of Colin Bristow with Bank of America. Your line is open.

Thanks for taking the questions. Just as a follow-up on -227, can you give any color on increased enrollment? Sorry if I missed this, but are you increasing enrollment across all PDL-1 positives or just at the highest strata? And just also, how will this impact the timing of the ultimate readouts and/or early analyses? On your guidance, it looks like from your prepared remarks and the cadence of share buybacks, you're anticipating around an $0.08 to $0.09 tailwind on EPS in 2017. Am I thinking about this math correct? And then just lastly on Eliquis. Were there any significant wholesaler inventory changes in the quarter? Thanks.

Okay, thank you. This is Fouad. Let me start by answering the question on -227 first. So the recruitment was not stopped as initially thought of in scheduling, and the reason is really to increase the size on the expresser part of the Phase 3. And, as you probably know, the non-expresser part of the Phase 3 started a little bit later and is still recruiting. However, I think we believe we reached a level of confidence in terms of the overall number of patients we will have within the next two to three weeks, and probably the recruitment for the expressers will end by then. There is no real impact to the timeline, and I think we still have optionality on the timing of the analysis and the different looks and cut-offs on -227.

Yeah, I'll comment on your share buyback. As I mentioned in my comments, we expect to have a 10b5-1 program, which means we will be buying shares over time, which means that they all don't get retired at once. So I don't expect that kind of tailwind as you described it. I think it will be less than that.

Yeah, and just on the question on Eliquis. We're really pleased with the prescription share evolution of Eliquis. As Charlie mentioned, we've seen some really nice evolution in the Eliquis market share. And in our cardiology market share, we see an even wider gap with new-to-brand patients versus Xarelto. So that continued growth feels really good. In the quarter, we had sales increase about 15% or $68 million, and roughly $59 million of that was due to demand.

Thanks, Colin. Can we go to the next question, Tiffany, please?

Your next question comes from the line of John Boris with SunTrust. Your line is open.

Thanks for taking the questions, and congratulations on the results. Just want to clarify on CheckMate -568. I guess back at ASCO in 2016, you indicated that that was a medical informing or medical affairs study, -568 was, and that -227 was the registrational trial. Since that time, it sounds like you've doubled the size of that clinical trial. Did you have discussions with the FDA where that now is a registrational trial, or is it still a medical informing study? And what's the pathway at least for that study and timing for that study relative to -227? Second question for Giovanni. With you keeping OpEx flat, you have a core set of assets – I'm not sure if they're still core or not – in HIV/HCV. What are your thoughts around that area in terms of pipeline, optionality, divestiture? How are you thinking about those assets at least going forward? And then I'm not sure if you indicated at all on your sales to date, which are about $2 billion in the U.S., what percent of those revenue in lung cancer are between second-line and some first-line use that might have continued within the quarter? Thanks.

Yeah. John, maybe let me start with your question regarding virology, HCV and HIV. We have communicated in the past that we no longer have ongoing discovery efforts in that area. So that's really not a new announcement. And we are continuing to be focused on the core areas that we've described in the past. Clearly, the focus for us beyond immuno-oncology is in heart failure. We have really interesting programs in fibrosis, immuno-science very broadly, some activity in the genetically defined disease areas with a couple of interesting programs. And that's really where we're focused.

For the -568 question, as Francis mentioned earlier, this is a single-arm trial looking at Yervoy plus Opdivo in first-line non-small-cell lung cancer, primarily at the response rate, and which provide us with a lot of flexibility on when we look at the data and how often because it's an open-label, single-arm study. The increase in enrollment will provide us with more data for the spectrum of PDL-1 expression. In particular in the non-expressers and in the highly expressing patients. We view CheckMate -568 as being really complementary to CheckMate -012, and single-arm studies can help inform the current and future medical practice, as you mentioned and Francis mentioned earlier. We believe it will be helpful in reproducing the -012 data, and we will continue to use -568 to ask additional question and have additional answers to the first-line non-small-cell lung cancer subgroups.

Yeah. And just lastly, the question regarding first-line lung versus second line. To date, we have seen some limited off-label use of Opdivo in first-line lung, although it's very small compared to the total second-line business. And going forward, we've assumed very little first-line, off-label use in monotherapy. So any evolution in that would be an opportunity to our 2017 scenario, as was previously described. What we're really thinking about that will drive growth is our growth across the rest of the tumors in the U.S. and the growth in our business outside of the U.S.

Tiffany, I think we have time for one last question.

Your last question comes from the line of Tony Butler with Guggenheim Partners. Your line is open.

Yes. Thank you. Two brief questions, if I may. Fouad, when you discuss durable response or durability, it seems to be a little bit subjective. And I'm curious if you would just spend one minute to define clinically how you and the rest of the community really look at what may be a durable response. And second, Charlie, just one housekeeping. When you think of the -265, ex U.S., Germany, and France, and you have to reverse that contra to revenue, do you have to do it in the same quarter as you get approval in those countries? Or do you have the flexibility to amortize that revenue accordingly? Thank you.

Let me start, Tony, because that's an easy one. The revenue recognition rules are that you have to recognize it once you do have the pricing agreement.

Okay. For the durability of the response, Tony, I think – let me put this into context. We have been seeing a lot of therapies, whether TKIs or chemotherapy shrinking tumors, shrinking cancer, and fortunately for certain period of time and then the tumor grows again because of that short durability. Our experience with immunotherapy, in particular with combination immunotherapy agents like Opdivo and Yervoy, when patients respond, they respond for a very long period of time. And when I say very long period of time, if you look at our data, overall when we do not even reach median of duration of response. And why we look at that and why that's important because that's really a good surrogate for improvement of survival, and improvement of survival is an important endpoint to look at it for lung cancer because that's what basically we'd be looking at the end. So the more we see patients continuing to respond to our combination of immunotherapy, the more we think there's a potential to really show long-term survival in a variety of cancer types.

Fouad, thank you. Thanks, everyone. Again in closing the call, let me say we've communicated the results of important quarter. Our business trends are quite strong across the board. We've made a number of important announcements today. All of this speaks to two things: our focus on execution for the rest of 2016 into 2017, and our confidence in the growth prospects for the company. Thanks everyone, and have a good day.

This concludes today's conference call. You may now disconnect.