Hello, everyone. I am Pascal Soriot, CEO of AstraZeneca. Welcome to the First Half 2016 Results Conference Call for Investors and Analysts. The presentation is posted online for you to download. We plan to spend a good half hour on the presentation and then we will leave plenty of time for Q&A. In total, we have one hour and fifteen minutes together. So please turn to Slide 2. I'm joined today by Luke Miels, EVP for Global Product and Portfolio Strategy, Global Medical Affairs and Corporate Affairs; Marc Dunoyer, our CFO; and Sean Bohen, our EVP for Global Medicines Development and Chief Medical Officer. I will provide a shorter review, and then I'll hand over to Luke for an update on the growth platforms and the ongoing launches of new medicine. As usual Marc will cover the financials and our guidance, Sean will provide a pipeline and news update, and you'll see there's a lot of exciting news in that pipeline. At the end I will provide concluding remarks before we take your questions. Please turn to Slide 4. The second quarter of 2016 continued with in line with our expectations, and we are on track for the year. Total revenue reflects the loss of exclusivity, including Crestor, but also the phasing of externalisation revenue. We saw our end investment stabilizing with Q2 growth more in line with our full-year commitments. We're still funding a very attractive pipeline, which has delivered important news lately. As far as SG&A costs, they continue down, which support the full-year commitment of a material decline. EPS overall was impacted by the FluMist inventory write-down, and all in all the business performed as we expected, and we therefore maintain the full-year 2016 guidance. Importantly, we are all very excited about this, the pipeline continued to…

Thanks, Pascal. And if I can ask you all to move to Slide 8 please. So I'm pleased to report the growth platforms have demonstrated significant progress in the half. These platforms now represent over 60% of total revenue, and encouragingly new oncology now represents over $0.25 billion in sales. And these sales reinforce our confidence that our strategy is very much working and very much on track. Please now turn to Slide 9. With Brilinta and Japan performing solidly and in line with expectations and even better in the second quarter, I'll prioritize my time today to focus on respiratory, diabetes, emerging markets, and of course new oncology. Please turn to Slide 10. Our respiratory franchise delivered 1% sales growth during the half, with the performance in emerging markets and the resilience of Symbicort in volume terms offsetting the competitive pressures in the U.S. and also the EU. In the half year, Symbicort established itself as the number one medicine globally in the growing ICS/LABA class in volume terms. Sales were down 6% in the half, reflecting the competitive nature of the overall market. In the U.S., solid volume growth was partly offset by pricing as a result of new contracts. And for Europe, while sales continued to be affected by slower in-class market growth and competitive pressures from both branded competition and analog, Symbicort maintained its market leadership in the highly competitive ICS/LABA market environment. And we aim to strengthen our position with the ongoing launches of the pressurized device across the EU. In emerging markets we maintained leadership where sales grew by 25% with China growing by 33%. Pulmicort sales grew by 10% in total, largely driven by a 23% increase in emerging markets, and the business in China continues to expand with more nebulizer centers established…

Thanks, Luke, and hello everyone. I'm going to spend the next few minutes taking you through our performance in the first half. Please turn to Slide 17. Before we go into the numbers, I wanted to highlight something you may have noticed in our announcement this morning. In view of the recent SEC guidelines and requirements around reporting of GAAP and non-GAAP measures, we have decided to proactively implement changes in our material straightaway. You'll notice that we are now giving equal weight to the reported and core performance, with reported numbers now coming first. By doing this, we are announcing what is becoming clear and more consistent disclosure. We will strive to improve our disclosure over time. In the spirit of the changes we are making, you can see on this slide the reported P&L of the half and for the second quarter. Please turn to Slide 18. Turning to the core P&L, the total revenue decline of 3% in the half reflected a fall in product sales by 2%, with the effect of losing exclusivity in Crestor in the United States particularly impacting the performance in the second quarter. Looking at externalisation revenue, we are planning good level of activity in the second half of the year. There were a few things; however that completed in first half. The 5% rise in the cost of sales in the half reflected the mix of sales as well as a $47 million write-down of inventory levels for FluMist in the United States. The core gross profit margin declined by 1 percentage point in the half to 82%. Core R&D cost growth slowed to 3% in the second quarter and full year total core R&D costs are expected to be at the similar level to last year at constant exchange rates.…

Thank you, Marc. I will now move on to our exciting pipeline. As we've alluded to earlier, we are seeing very good progress with the pipeline and in particular the medicines in late stage clinical trials. First I plan to review the second quarter late stage pipeline headlines, but also go a bit more into detail on a few items of particular interest. At the end, I will review our upcoming news flow. Please turn to Slide 24. Starting with respiratory and autoimmunity, Benralizumab saw a leadout of two positive Phase III trials, CALIMA and SIROCCO. These trials will form the backbone of a regulatory submission before the end of the year for an intended indication in severe uncontrolled asthma. Further last week, the dermatologic and ophthalmic drugs advisory committee, appointed by the U.S. FDA, voted unanimously recommending approval for Brodalumab for adults with moderate to severe plaque psoriasis. We will continue supporting our partners Valeant in the United States and LEO in the EU, in bringing this important medicine to patients as soon as possible. In cardiovascular and metabolic diseases, Brilinta saw the last patient of a total of about 19,000 patients recruited in the Phase III THEMIS trial, the last Company sponsored major outcome study for Brilinta. To read out, in 2018, THEMIS is testing Brilinta in patients with Type 2 diabetes and coronary heart disease with no histories of myocardial infarction or stroke. Qtern, previously known as Saxa/Dapa, was approved in the EU for Type 2 diabetes. Also the regulatory resubmission for the fixed dose combination of the two diabetes medicine was accepted by the FDA in the U.S. We expect to hear back from the agency during Q1 of 2017. We received a complete response letter from the FDA for ZS-9 as a potential medicine to…

Thank you, Sean. Please turn to Slide 32. Before we end, let me just quickly summarize; first of all our financials are on track, and as you heard we are reconfirming our guidance for the year. Second, the pipeline is accelerating. We now have 14 new potential medicines in Phase III and under registration. The third message for you today is the oncology pipeline is progressing very rapidly, in particular the IO-IO combination programs that are really moving along quite quickly. And finally we are looking forward to sharing upcoming newsflow over the next 18 months that we think has the potential to transform AstraZeneca. So we will now go to the Q&A.

[Operator Instructions] We can now take the first caller who is Jo Walton at Credit Suisse. Go ahead Jo.

Thank you. In the spirit of talking about clinical trial results that are coming through next year, I wonder if we could also look at some of the costs that might be coming through next year. So if you've got all of these new products coming through, how realistic is it to assume that you can continue to decline your SG&A as you go through 2017 as you have done through 2016? A second one on foreign exchange please, because we are going to be seeing a foreign exchange benefit at the bottom line which we're not seeing at the top line, I wonder if you could give us some help as to which line item we should see it in? Presumably it will be an expansion of the cost of the gross margin and an apparent reduction in cost of goods, but just to check that please. And finally, you've highlighted the credit rating. I wondered why you'd chosen to do that. Does this tell us something about you feel that you've got capacity to buy some more products in?

Thank you, Jo. So I'll ask Marc to answer the ForEx question and also the credit rating question. Maybe quick point on the SG&A cost, and Luke, if you have anything to add please jump in. The reason we feel confident we can continue merging our SG&A costs, Jo, is that first of all we have productivity programs in place that will continue to deliver benefits over the next year or two. And secondly, it's important to keep in mind our pipeline is now shifting to more specialty care products. If you look at the newsflow it's going to be very much driven by oncology over the next 12, 15 months. Products also like Benralizumab, all of those medicines that require R&D costs but less SG&A cost, as you know, because they're more targeted customer groups. So with the pipeline shift and redeployed to specialty care products, our expectation is that we can manage our SG&A cost as we have guided we would. I'll ask Marc to answer the other two questions.

So, Jo thank you very much for the question on the ForEx. So the slight benefit that we are anticipating obviously if the existing ForEx are continuing to the end of the year. In the spirit of great ForEx volatility this is a very difficult thing to predict. The lines of the P&L that are particular influence these rates we're maintaining to the end of the year would be the cost lines. Basically this would be SG&A as well as R&D cost, but also as you suggested the cost of good lines. All this line would be receiving some slight improvement for the end of the year. This is due to the dramatic decline of the sterling pound versus dollars and to a small extent, but much lower extent, on the Swedish kroner versus dollars.

Credit rating, Marc?

On the credit rating, so I think there was no hidden agenda, no any hidden message underneath. It was just to first of all to say thank you to our creditors and also to present to the outside world that our portfolio of debt maturities is now more balanced than it was in the past, and there was nothing more than explaining how our portfolio of debt is structured.

Thank you, Marc. Sachin Jain, Bank of America. Sachin go ahead.

First one for you, Pascal, in the introductory comments you thanked shareholders for their patience, which Marc reiterated when discussing the dividend. I wonder if there's any particular backdrop for that. Shares broadly flat last two years with investors left apparently not giving credit to pipeline progress and the inflection you believe in. So just wondered whether you saw any risk of another corporate crystallizing the value you see quicker? Then just a couple of pipeline questions, first on the Tesaro PARP data that came out recently. How do you view that from a competitive standpoint, given the positive data in gBRCA wild-type and the absolute PFS improvement versus what you have on the label? Secondly on respiratory, Glaxo was very vocal in triple yesterday in potential market share gains versus Symbicort. Wonder if you could remind us of the timelines of your triple and its relative competitive profile? And then a final one on IO, does the Merck headline data in first-line PD-L1 positive lung with the PFS benefit change your view of PFS versus OS read for MYSTIC? Thank you.

Thank you. Sachin. Lots of great questions there. Sean, maybe you could let me take the Tesaro question and also the triple and the PFS, and Luke could cover the Symbicort market share. Is that okay?

Triple timing. Yes.

And then I'll start with the backdrop of the - you can comment again, here again no special message here. I mean the value proposition to investors has always been that's how we've always presented. It has always been that we are going or we are going through a transition period where we are losing patent protection on a number of products, we are rebuilding, retooling, rebuilding the Company and investing in our pipeline and essentially creating long-term value that will end up being reflected in our share price. In the meantime essentially we were rewarding shareholders for their patience with a progressive dividend policy and sustaining the dividend. Having said that, we suddenly recognized that we asked our shareholders to be patient as we retool. So as we are now coming through our infection point, we're not yet at the end of our second phase of this journey we are going through. That journey will end at the end of 2017 and by then we will be totally finished with our patent expires. But as we are going through an inflection point losing the last patent-protected product and starting on the journey of rich newsflow, we thought it was the appropriate time to say again thank you for your continued support of a number of years. Now as to whether we would be exposed, we currently have been very aware over the last few years that as we create value, then at some point our pipeline becomes attractive, and so hopefully it becomes attractive to our shareholders. But of course, it may be attractive to anybody. I can't speculate more than - I don't want to speculate, but suddenly it is clear that there is value in our pipeline that we hope our shareholders will recognize. Sean, do you want to cover the triple question on…

Do you want me to do triple first?

Yes. I mean, in the order you want.

Okay, well let me start, thank you for the question, Sachin. Let me start with your Tesaro PARP question. We have great confidence in PARP inhibition and the DDR portfolio. When we see a competitor molecule targeting the same target have positive data that certainly increases that confidence. The question you asked has to do with a comparison between the two molecules, and that's just very difficult to do when you don't have the full data from the trial in particular. We don't know the pre-treatment characteristics of the patients, and I'm sure as Tesaro gets a chance to present their data, we will get more information be able to make some sort of assessment. We have great faith in SOLO-2 in this population, and as I pointed out the Study 19 data looks quite interesting to us. That update with survival is encouraging, and as well we're wondering if we'll be able to duplicate this long payout, which would be very interesting. I'll move on to the IO one then I'll go to the Vespi because there's a part probably there for Luke. So IO, so again with Merck, they definitely reported top line that their PFS was positive in first-line PD-L1 positive. Non-small cell lung cancer, and in their press release said that they hit OS as well. Again, we haven't seen the exact data, so it's a little hard for me to speak about what it means for MYSTIC. We are very confident in the MYSTIC trial design, in that PFS and OS being called primary endpoints increases the strength of that trial to demonstrate the benefit of IO and combo IO across a broad patient population. Specifically the part of the - the piece of information we're missing, Sachin, is that we don't know how much crossover there was in the Merck trial from the control arm to other IOs versus how much we might have in MYSTIC and whether that might confound the OSN point versus the PFSN end point. So that is the one big piece that I really don't have the information to comment on. Last with Bevespi, approved in the U.S. earlier this year. Our plan is to file next year in the EU. And the reason that those filing dates are different, we include the symptom data, the symptom improvement data in the first filing for the EU. And that data becomes available a bit later. Its availability was delayed slightly because that trial is also the one that's enabling China for us, and China implemented a new review, their Human Genetic Resources Administration, and that lead to a bit of a delay in implementation of the trial in China.

So, Luke do you want to call the market share point that [indiscernible].

And I'll just cover the triple from the commercial perspective first. I mean it's a very attractive option, that's why we have one. Depending on the geography you've got between 20% or 50% of patients are on a free triple. I mean from our view, the focus is on cannibalizing those compounds and converting them to a fixed triple in the face of generics. If you look at ICS/LABA, this is a little bit of a quarterly debate. Maybe you guys can find this a bit entertaining at times. I think it depends on how you cut the data. If you look at Q2, Symbicort was up 1.24%, Advair was down 2.18%, and Brio was up 1.29%. So if you look at the Company level, it's 1.24% versus 0.89%, if you look at within the brands there is a mix. I think you'd expect that. Our focus is very much on making the case that fast control with Symbicort is an attractive option versus once a day. And if we look into 2017, we've got a very good sense of where we're heading in terms of access, and we expect to finalize this over the next four weeks.

The only thing I would add, Sachin, I think Luke, you are talking about U.S. market shares, right? And then if you go beyond the quarter, if you look at the last few weeks actually, Sachin, you will see that in terms of new Rx, MBRx in the US, Symbicort is making some very nice progress and we expect that to be reflected in our total share in Q3. So we continue to make progress, and with the triple maybe the additional point is that as we said before it's an MDI. So there is a place for DPI, there is a place for an MDI. The two technologies are complementary, somehow competitive of course, but also somehow complementary, addressing different patient needs. We think we have a good technology with the pill delivery. What we see now that we are starting to discuss the respi in U.S. with some of our customers and physicians, we see a lot of excitement about the technology and the ability of the technology to deliver later medicines in a very effective manner. So we think that the triple pill will have a nice potential. Moving to James Gordon, JPMorgan.

Hello. Thanks for taking my question. Two pipeline and one launched question. On the pipeline, so continuing the theme of DDR, regarding parts of your competitors have suggested that the level of PARP trapping is key to how efficacious the PARPs are. I think Medivation's PARP is much stronger in terms of trapping, Tesaro's also seems to have more trapping ability than Lynparza. So do you think that could be an explanation for different efficacy; does trapping matter? One question on respiratory, which we saw the Benralizumab data coming up at ERS. Where is it you're seeing the differentiation for the product versus Nucala? Is it a more efficacious product, or is it around dose frequency? Then just one product, one launch product question. Nexium was surprisingly strong, so sales in the U.S. were up sequentially about 25%, whereas the IMS sequentially data down about 20%. So what happened on Nexium and can that be sustained?

Should we start with maybe DDR? Sean, you want to cover that?

Sure yes, there's a story out there about trapping, and what I'm going to have to do is I'm going to have to refer you in the near future to a paper that will come out of our IMED group and be published in a scientific journal that we believe refutes the contention that trapping makes any difference in terms of the effective PARP inhibition, that is just reversible inhibition versus trapping. What was the second question?

Yes, Benralizumab.

Differentiation of Benralizumab. You'll see the detailed data in a bit, and it is true that we tested - we did test the two dose schedules every four weeks and every eight weeks. I can't talk about the data until it's made public, but I think you'll be able to see where our enthusiasm comes from when you see the data. As far as differentiation goes, what we previously published is that Benralizumab depletes eosinophils more rapidly and more completely than other drugs that we've seen in this class. And it is our hope that then translates into a superior profile.

You want to call the next [indiscernible]?

Yes, sure so, James, I think the key thing is we're actually able to retain a little bit more business than we expected. It was around one-third of Rx. And this was driven by some government and Part D programs, but I think the key thing is if you look into Q3, you should not assume that we retain them and that you see a more traditional loss of exclusivity trajectory at that point.

Thanks, Luke. Shall we move to Tim Anderson of Bernstein. Tim, go ahead.

Thank you. Just a few questions on externalisation just to clarify, after the May SEC guidance came out is that going to change at all how you elect to put both revenues and expenses in your non-GAAP results? I know it's changing how you present GAAP and non-GAAP on things like earnings calls, but does it change how you're going to actually treat the non-GAAP P&L? Second question on Farxiga going into 2017, can you talk about likely formulary positioning increasing, decreasing, staying the same? And would you agree that EMPA-REG, even though it's value enhancing for the whole SGLT2 class, showing a benefit of that class of drugs, it might actually lead to more price competition in this category as companies other than Lilly compete for formulary share by lowering their price? And last question if you could just make some high level comments on Brexit and the potential impact to the drug industry in the intermediate term across the various potential areas like regulatory outlook?

So let me just maybe start with Brexit. Farxiga, Luke, will you cover this one and externalisation, Marc? Brexit, it's really hard to comment at this point. We are at the very beginning of the discussions between the UK and the EU, and we'll have to see how the process evolves over time. One thing that we'll engage in discussions across a variety of aspects, but one of the most important ones for us is indeed the regulatory process. Today we have a single process for our true up and we would hope that it is retained or at least, at the very least a very effective process of mutual recognition, so we don't have to duplicate efforts to get approval in the UK. That simply adds costs and delays and doesn't help anybody, the patients and other payers and the rest of the industry. That's one aspect that is really critical for us, and there's a few others of course, but that is the most important. It's really too early to judge and we hope that everybody is going to be practical about those things, but we'll have to see. Farxiga, Luke?

So Tim, I think it's best to summarize the diabetes contracting environment in the U.S. as dynamic and relatively unpredictable, which I think to be fair is how it’s been for some time. If we look at commercial, we've got a strong position there. I mean we've got around 90% of lives covered. Where we need to do some more work is Part D, and we're very focused on that. In terms of EMPA-REG being a disruptive element, I think in the end it is a very positive force. What we watch closely of course is the class growth of the SGLT2, and if you look at quarter one it was around 2% whereas in quarter two it went up to 7%, which I think is very encouraging. And our focus of course is capturing as much of that as possible. And I think linked to that, our focus is very much on converting patients who may be placed on a DPP4 onto an SGLT2, and when that decision is made that we make sure that patient is a Farxiga patient.

So regarding exchange revenues, I do not believe that SEC guidelines will have any impact, so the short answer is let me remind you that the external revenues are only containing sustainable source of revenues in opposition to what we classify as other income, which are usually one off. We are retaining - in the case of external revenues we are retaining funds of significant influence on the asset, and there's no disposal of the intellectual property. This is why we classify those as external revenues, and we intend to provide the best possible transparency on this source of revenues.

Thank you, Marc. Keyur Parekh, Keyur you want to go ahead Goldman Sachs.

Thanks, Pascal. A couple of questions, please. First one on kind of the PARP landscape, would you expect, assuming that the mid-operative data holds out as the press release suggests, would you expect them to get approved a label beyond just the BRCA mutated and germline BRCA mutation? And if so, how do you think from a commercial perspective Lynparza standard Tesaro product? And secondly, Sean, you mentioned kind of on the MYSTIC study there was a potential for a development of single arm to be seen as potentially you being able to file on that. Can you just remind us how MYSTIC is powered from a statistical perspective? Is it [Durva Creme versus Derma] versus standard of care, or is it kind of the pool population versus standard of care? What is the statistical design of the study? Thank you.

Both questions are for you, Sean.

Yes, so with regard to the Tesaro label and what's expected there, I expect that their label will actually have the multiple mutations that they tested and we focused on BRCA, but we also have shared some data from the Royal Marsden on prostate cancer with a more complex definition of BDR. I think that this will be an evolving story there. We have confidence in SOLO-2 converting us to full approval, and have Fast Track Designation in fact for that indication, and we talked about when that data is available. You had a competitive landscape question about that, which I don't is that Luke?

I can believe spend on that I mean I think right now if we look at scripts in the U.S., around two-thirds of them have a variance, and in the non-promoted areas around its low teens in prostate and around 6% is breast. So I mean I think this is a very rapidly evolving area. I mean Sean has referred to a paper that is coming out; we think that this is going to further crystallize things. We also very, very focused on the HRR test, and we have our views there. But ultimately I think the key is that we need to remain ahead of the curve in terms of new indications, but also critically if you look at DNA damage response, the capability that we'll have to combine agents such as AZD1775, Lynparza, and other agents will position us in a strong position in the mid-term.

And then you had a question about MYSTIC. So we don't get into the detailed statistical analysis plan and design. I can refresh you with the things that we have communicated about MYSTIC. So MYSTIC, it's all commerce, so we have the ability to analyze PD-L1 positive versus PD-L1 negative and all of those patients there are three arms; the standard of care, chemotherapy, Durvalumab, or Durva plus Treme. And then as you all had noticed from earlier this year, we did change the plan to bring to elevate overall survival to a co-primary end point with progression free survival. Earlier this year and in order to enable the power of that, we significantly increased the size of the trial, it's about 1,100 patients.

Thanks, Sean. So next question is from Nicolas Guyon at Morgan Stanley. Nicolas go ahead.

Hi, thanks for taking my questions. I have two actually. The first is a follow-up on MYSTIC again, I'm afraid. So I guess an important question here is how do we separate the PFS data that you'll have next year? In the context of chemo, I think the one is rated roughly six months of PFS, Merck & Co chemo IO combo roughly 10 months, and no available data for IO mono-agent yet. What’s could be the appropriate comparator, and what would you consider as clinically relevant for the PFS for MYSTIC; that's number one? And second question is on liquid tumors, could you update us on the Celgene collaboration and the next important coming clinical results please? Thank you.

Sean.

So with regard to the meaningful PFS end point, I think it's difficult again to do these cross-trial comparisons. I think that the data we've seen to date certainly increases our confidence particularly in the PD-L1 positive non-small cell lung cancer patient. That we will have efficacy you know just to be blunt, it doesn't look like the PD-L1 versus PD-1 so far have really differentiated from each other in terms of how they work within that population. So we anticipate Durvalumab will work well and show a meaningful benefit. The second aspect of it is of course the ability to see a difference with Treme. And so obviously our primary strategy is IO combination Durva/Treme. And the place where we had previously said we were most likely to see the benefit of that is in the PD-L1 negative population, which is now not really benefiting from IO therapy. And the last thing I will say though is that there was some interesting data presented from BMS work at ASCO that suggested at least in some of their experience that their CTLA-4 PD-1 combination was showing benefit even in the PD-L1 positive population not our primary hypothesis but mistake is testing that hypothesis nonetheless. I'll move on to Celgene. The next we haven't reported on data readouts we’ve reported on what trials we have ongoing and initiated and it started primarily in multiple myeloma where the combination trials are ongoing in Phase I obviously there you have to get your safety data before you can move on and expand the trial. We also have non-hodgkin's lymphoma and chronic lymphocytic leukemia with Durvalumab, either with linoleamide anti-B cell antibodies, also Durvalumab and Rituximab chop in diffuse large B cell lymphoma that's high risk. Those trials are to start by end of year. This year and the also some Myelodysplastic syndrome trials that start second half of this year.

Thanks Sean. Richard Parkes, Deutsche Bank. Richard go ahead.

Hi, thanks for taking my questions. Firstly, on the guidance, you've managed to - you've reiterated the earnings guidance despite the FluMist write-down, which given the lost revenues in that write-down it would have taken you towards the lower end of your expectation. So I'm wondering what's changed since the beginning of the year that's maybe positively offset that. And maybe you could talk about your expectations for externalisation revenue for the year, and how that's evolved in that context. That's the first question. Second question is on plans for FluMist. I'm assuming that without a return to the U.S. market are slightly to the an overall loss-making operation. So just wondering when you might make a decision on the path forward for that franchise,? And then finally on Farxiga. Your competitors are highlighting slowing NRX in the SGLT2 class in the U.S., I'm just wondering if you could discuss what you're seeing now, your expectation, and what you might be able to do to help to return the cost of growth? Thanks.

Thanks Richard. So the guidance is a range and so we can be at the top end or the bottom end of the range, and it is true that the FluMist event actually tends take us towards the lower end of the range. So there's not really a big change that we can see from what we guided to at the beginning of the year. We're still in the range we indicated, and that's why we reconfirm the guidance. As far as FluMist, we are certainly working hard to - at this stage our teams are working hard looking at trying to understand better this discrepancy that we saw between the CDC data [indiscernible] but also the UK health authority data and planning for next year. We still believe FluMist is a strong brand, it's a good product, there's a good place for it, and we stand behind it. We have to kind of there will be more work to decide what we will do with it moving forward. That's going to take a little longer. And on Farxiga, Luke do you want to comment?

So Richard, it's interesting if you look say from early 2014 until now, the class was on a very steep exponential growth curve until around July of 2015, when you had the DKA letter, the diabetic ketoacidosis letter. And the class essentially moves sideways for some time, and I think that was while physicians, particularly primary care physicians, where I think potentially waiting for regulators and opinion leaders to form a judgment. The issue around DKA at the time of course was it had the classic symptoms of DKA, but there were some other elements which made it difficult to understand and characterize that patient. That's now happened, and now what we're seeing in the last quarter is the beginning of the class growth again. I think it's early days, it's not where it was before. However, if you look at the fundamental benefits that this class delivers, it's a very effective product life. Farxiga is very effective at addressing A1C, and you have the added benefit of weight loss of course. And on the back of that we have one outcome study in place, and hopefully we have a few more that should ensure that this class returns to its historical growth.

Thanks, Luke. We've experienced this in the past. Crestor was suddenly a product that was established over a period of time, which was difficult at the beginning I think the SGLT2 class has enormous potential, as Luke said, and we just need to work through this initial issues that the class is experiencing. Andrew Baum, Citi. Andrew go ahead.

Good afternoon. Three questions for Sean, please. Firstly, could you tell us whether you'll be updating the follow up for the Phase Ib for Treme non-small cell lung cancer trial? Since the historic data, asthma is obviously coming up. Number two, looking at that same trial, there were at least a couple of cases of pseudo progression, which I know is more common with CTLA4 than necessarily PD-L1 monotherapy. When we think about the PFS endpoint for MYSTIC, on the one hand we have the recent Merck trial seeming to validate its utility in this particular setting, but on the other hand you now have CTLA4 included in the regimen with the risk of pseudo progression. How do you think about that and mitigate the potential impact in terms of diluting any PFS signal you may see? And the final question is on your refractory CLL trial with Acalabrutinib versus Ibrutinib. Should we expect that data at the end of 2017? I note it's not on your chart, but looking at the timelines one wonders whether it's possible. Thank you.

Okay, so starting with the 006 update, the non-small cell lung cancer IO trial you're talking about. We will update. We haven't submitted to do that yet, so it won't be as ASMO and then we won’t be able to communicate when we update until we submit and get the update accepted at a meeting to present it. But we will plan to update that data in due time. Pseudo progression, yes, pseudo progression it's interesting that you bring it up. It was a very, very hot topic as you appreciate for a while with IO, and it has sort of gone away, and we've seen in long that you do seem to get PFS benefit regardless of whether the pseudo progression is a real entity or not. It doesn't seem to be so prominent that you lose it. And then of course obviously MYSTIC is now designed to have overall survival as a co-primary endpoint. So we have another way to capture the benefit of IO. With regard to the Durva/Treme combo being potentially different than Durvalumab, I really don't think we have enough data to conclude that at this point in time, but we do feel like MYSTIC is well designed and well powered and now stronger with both the PFS and the OS endpoints as co-primaries. And there was a third one?

Acalabrutinib.

Acalabrutinib. It's too early for us to report on that because we have the elements of accrual rate and forecasting when the accrual will end. Then of course when we report out we'll be dependent on the event rate, because these are event-driven trials. And so it's too early for us to forecast that timeline for when the head-to-head data will be available.

Thanks Sean. Simon Baker, Exane. Simon go ahead.

Thanks for taking my questions. Three, if I may. Firstly going back to Tim's question on externalisation revenue. You said that it's unimpacted by the SEC guidelines, but I was wondering what the impact of IFRS 15, the new revenue recognition standard, would be if at all on the reporting externalisation revenue? Then secondly onto the gross margin. I wonder if you could give us your updated thoughts on how the gross margin evolves through this year and beyond. Obviously we know about the impact of FluMist, but also I was wondering about the potential impact of the genericization of Crestor on the gross margin? And then finally on R&D, you've over the last couple of years focused your commercial therapeutic focus down to the key target areas that you now have. I see recently there seems to be evidence that the same is happening in R&D. There have been reports of closure of U.S. neuroscience R&D, and also reductions in the non-oncology R&D efforts in the U.S. at MedImmune. I just wondered if there was more to go there in terms of fitting the R&D footprint to the commercial footprint, notwithstanding what you've said in the past about not wanting to try and box in and potentially lose good scientists by overly focusing the R&D effort. Thanks so much.

Thank you, Simon. Maybe I'll start with your last question and Marc, you could cover the externalisation revenue question and the gross margin one? So the focus was not a commercial focus, it was a development, and commercial focus was really building capabilities in development actually to start with and of course also commercially around the three core areas. And what we said was always that we would actually partner the science or the products that come out of our science in the other therapy areas. So we continue doing this; we suddenly have further fine tune or our investments in the early phase if you look at it, immuno-oncology portfolio of early projects is suddenly increasing. We now have ADCs that are antibody drug conjugates that are also progressing through frankly call it development. So in oncology, we have an extremely rich research and development pipeline, a set of early programs. So we have to suddenly focus our sales here quite a lot. It's not a commercial focus, it's actually at the end of the day, we have strong sense across the whole portfolio where we suddenly have to place our bets and focus our resources on a few things. Having said that, we still have some activities in autoimmune and in CNS, and also we have antibodies that we'll continue developing for infections. But suddenly they need us to focus even our research and development effort. Marc, do you want to comment about the IFRS?

Yes, regarding the question on IFRS 15, the date of implementation of this new guideline is for 2018. So we are still presently are evaluating whether it has any influence on the way we report. We'll be confirming that and letting you know whether it has any implication for the way we describe or exchange revenues. For the time being we do not think this is necessary, but as we continue our investigation we will provide more transparency on this.

Gross margin, second half? Or this year and beyond, sorry, so I’m only going to comment on the gross margin for the rest of the year. I believe the level of gross margin for the first half is a good indication for the second half. Obviously, as the mix of products evolves over time, it's a bit more difficult to project, but I believe it's a very good indication. For the later years, I think it should remain still a very high gross margin, but it will be influenced as I was just explaining by the mix of products in our portfolio. Well you have a certain sort of if you can imagine a downward pressure coming from losing Crestor, which has a higher gross margin. On the other hand you're going to have upward pressure coming from growth in oncology products that have a high margin. So we just have to monitor what this mix looks like. So should we move to Steve Scala at Cowen. Steve go ahead.

Thank you, I have a few questions. First, how does AstraZeneca view the CheckMate 568 trial? That's Bristol's single arm lung trial of the Nivo/Ipi combo, which is to report in January. Bristol states that it is a medical informing trial, but does Astra view it as a competitive threat to MYSTIC because Bristol could have combo data first and has two agents that are approved? The second question is just to clarify, does AstraZeneca believe it needs to hit both PFS and OS to file MYSTIC, or is only one or the other sufficient? And then lastly, it appears the THEMIS trial has slipped from 2017 to 2018 and I’m just wondering why? Thank you.

Okay, so the first thing is on the BMS data. It's not a randomized control trial, so I think its ability can position it particularly in the United States to use these drugs off label. They can if they find the data compelling. But we think it's unlikely - BMF has said it's unlikely to be a label enabling trial. MYSTIC PFS, , yes we can file on PFS and MYSTIC’s original design focused primarily on PFS, because it is being enrolled in places where crossover from the chemo arm to IO is a possibility. And the question was might that confound the OS endpoint, but we'll find that out later. And then separately we have Neptune with combination in all comers, and that's an OS primary endpoint trial as the sole primary endpoint, which is confirmatory as well. THEMIS, what happens with these trials is you get the enrollment rate, you get the event rate, that's another event-driven trial. That affects the maturity, and so if the events come slower than you'd like, you end up moving the timeline. We try to be reasonably conservative, but we don't always get it quite right. And so we update as we get more data on when we expect the outcome.

Yes, good. Thanks very much. Seamus Fernandez at Leerink. Go ahead Seamus.

Thanks very much. Can you guys hear me?

Yes. Go ahead.

Okay, great. So just wanted to ask a of couple questions. So from - just to clarify on Andrew's question with regard to the timing of the head-to-head trial of Acalabrutinib versus Ibrutinib, is it - I didn't hear if you stated whether or not it was possible that we could have a readout sooner than what was stated on clinicaltrials.gov which is a 2019 primary completion. And can you just remind us what the goals of that evaluation are, evaluation is? I would presume that it's primarily safety, so it would just be helpful to know how that trial is designed and when you would feel you have confidence in certainly non-inferiority with the prospect of perhaps a better safety profile. The second question is again on MYSTIC. The agency seems to have offered and demonstrated quite a bit of flexibility as it relates to tests and cutoffs in oncology clinical trials. Can you just help us understand if there may be that type of flexibility with regard to the PD-L1 test, if the PD-L1 test could be utilized as a primary analysis in the MYSTIC study, and if you could move the cutoff? Because I believe your cutoff has historically been 25%, is it possible that given the data that you've seen from other companies and the analysis from or the alignment of some of the other clinical data sets, looking at your PD-L1 test versus others it seems like altering that perhaps down to even 1% might be possible. So just wondering what the flexibility is on some of those endpoints. And then just the last question, the FLAURA study, are we still on track for a second half read out on the FLAURA study? Thanks.

Thank you, Seamus. Sean, it's all for you.

Okay. Acalabrutinib, so we expect the relapse refractory data our current base and again these are event-driven so just think the THEMIS qualification I gave applies here. 2018, the relapse refractory head-to-head you mentioned too; so the question there is there are two aspects to this obviously. There's superiority and efficacy, and then there's the safety profile. And we’ll get both of those out of there. We believe there is an opportunity for superiority and with two attributes. One is being able to stay on Acalabrutinib longer due to ability to tolerate, and the second is we've shared some data on our inhibition level of BTK, kinase activity at trough with the twice daily dosing with Acalabrutinib. And it's greater in the mid to high 90s than mid 80s that Pharmacyclics has shared with Ibrutinib. Now the question is will that translate into a difference in efficacy, and that's a clinical question. And then the front line is 2019, going to MYSTIC and cutoffs, we agree with you completely that the data sets are a little confusing by virtue of the different assays. We do think that this will be illuminated over time, and at this point we have the ability to compare our assays. And some of that data has already been shared publicly. The other part is a trial design question. Could we change the cutoff from where we are, and the answer is we can rewrite the analysis plan as long as the trial is blinded and we haven't done an analysis. So there is an opportunity for us to look at that and change that in the analysis plan. Now that's the kind of thing you can only do once because then you have to put your nickel down and look at the data and then you're positive or negative. FLAURA is on track for second half of 2017. I think that was the question.

Thanks very much. Emmanuel Papadakis at MainFirst. Emmanuel, go ahead.

Hey. Just a couple of question around cash generation actually for Marc. The first was I saw you booked a $347 million charge as other in the quarter IFRS reconciliation. I couldn't find any detailing of the exact components of that. Any color would be very helpful, particularly which element of that might be cash versus non. And the second was just around cash generation for the full-year. I think Marc, you had previously commented it would be broadly in line with last year $3 billion or so was relatively weak if my math was correct in Q2. Just wondering if you'd reaffirm that expectation, or are we looking above or below? Many thanks.

Okay. So the first question on the $347 million, which is one of the non-core adjustments, it contains two groups of expenses. The first one - they are mainly provision, legal provisions, and the other is the value of contingent consideration. These are the - this is the two components of the $347 million. We do not provide a sort of product-by-product detail of the legal fees or legal compensation.

Considering that those aren't provisions and…

Okay, no question are they cash or non-cash, they are for the time being non-cash.

And the second is cash generation for the full-year?

So cash generation for the full year should be in line with what you have what you seen in the first half, but obviously this is a dependent to some extent the type and structure of the externalisation and other income agreements that we would sign in the second half. We have also indicated that we would have a very active second half of the year, but this needs to be considered in this. But I think it's - where you are now roughly in line with the first half would be very good projection.

Thanks, Marc. Jeffrey Holford at Jefferies. Jeff, go ahead.

Thanks very much. Given that a lot of your IO focus in the future is going to be on combinations, I wonder if you can give us your thoughts on how you think pricing of combination, of IO agency is going to be done, whether you think we're going to get into caps type models where they're limited revenue per patient. Just to help thinking about longer term modeling as some of these. Also just wonder if you can give us a bit of an update on your OX40 program? I know that you're committing an asset into mid- and late stage programs now. Just wondering when we’ll see those mid- and late stage programs and what the focus of those will be? Thank you.

Thank you. Sean, do you want to start with OX40 and the other mid-stage programs?

Yes, sure I'll go to OX40. So as you may know, we selected a particular OX40 agonist antibody MedImmune has as the lead molecule in which to invest. And we're in the process of studying that in combination with Durvalumab. And depending upon when we are able to define the dose, we will then initiate the late stage trials. Where to go with it, I think is something that you follow whatever signal you generate in earlier stage trials if you get one, and then I think also there is of course a patient selection hypothesis around OX40 that we will also test in the clinical program.

Thank you, comments on pricing, Luke?

Yes. So I think, Jeff, we're already seeing this across geographies. There is essentially a PDX or PD1 price being established. I think the bigger question is if what we think, and it's planning to happen with IO, IO happens then what's likely to be the evolution of pricing with the CTLA4 class. I think it's fair to say that the current pricing in melanoma is unlikely to remain there. We have a lot of strategic flexibility and a lot of good ideas in terms of what we can do with the combination of Durva and Treme. I won't go into them today, but again, we think we can bring something that's very competitive and very attractive to countries and patients. We're optimistic.

Thanks, Luke. Thanks, it's clear that, Jeff, we believe that we'll have to manage cost of those combinations, and we have many ideas, as Luke said. Those will be different country by country, so there's a whole list of options to manage the cost and it depends on the payers and the logistics of what we can do country by country.

So let me just close here. This was really a very busy session, lots of great questions, thank you for that. The exciting part is that it actually is a wealth of questions reflect the strength and the works of our pipeline, and the fact that over the next 12 months we're going to have a news flow that suddenly will actually give a lot more proof points of our progress and clarify the landscape for us for the next few years. Thank you again for all your interest, and we certainly look forward to meeting many investors during the award shows over the next 10 days. Thank you again. Thank you, bye-bye.