Good afternoon. Big pleasure to be back here in London and to present a very good set of results for the last year. It's also a pleasure to have two new colleagues today with me, Bill Anderson and Michael Heuer, who will present for the first time, our divisional year-end results. Before we get started, I'd just like to make two logistics comments. One is, unlike last year, we will not have breakout sessions after the presentations, so we'll have a five-minute break to rearrange the set up here, but all of you can stay in this room for a combined longer Q&A session. So don't run away after the presentations. And secondly for those who join us via webcast, you can submit questions via webcast, or also call in by telephone. Good. With that, let me go right into the results. You have seen sales up by 7%. Core EPS actually up by 19%. And even if we exclude the U.S. tax reform impact, still 8%, slightly ahead of the sales growth. And with this we will see a dividend increase to CHF 8.70. Both pharma and Diagnostics, up by 7%. Pharma, as you will see in a moment, very much driven by the new product portfolio. Diagnostics, again a strong growth, above the market. Very much driven by the main business areas, Centralized Diagnostics. And again, here, the immunodiagnostics business, which has grown double digits. You see a good momentum in terms of sales, and very much driven, again, by our new medicines, such as Ocrevus, HEMLIBRA and the new cancer indications or products we have launched more recently. More specifically here, what you can see is CHF1.2 billion impact due to the entry of biosimilars, and this is primarily Europe, of course, for MabThera and for Herceptin. And on the other hand, CHF3.2 billion of new sales for the newly launched medicines, so clearly overcompensating the decline, which we have due to the more mature portfolio. On an operating profit level, you can see 9% growth, so we could slightly improve our margins. Core EPS, again, 19%, as I've said, and also a continued strong free cash flow. Now let me shift to the portfolio, and I'd like to start with the number of breakthrough therapy designations because it is an indicator for the quality of the portfolio. And last year, again, we had 6 new breakthrough therapy designations, and we started off this year with a very important breakthrough therapy designation for Kadcyla, which opens up a new segment for us in the adjuvant setting. Bill will certainly cover that. But I mentioned that because even more so now than ever, the differentiation of the medicines, the benefits, which you bring to patients, that's really what matters. That is what will separate the winners and losers in this industry. It does not only provide you with the chance to replace existing medicines and get to market share positions, importantly, it gives you the pricing power with payers if you bring true medical benefit. And what we have also seen is that authorities around the world start to differentiate on the regulatory process. So if you come with something which is innovative where they give you a breakthrough therapy designation or priority review, timelines are actually very fast. On the other hand, if it's a less differentiated medicine, you quickly run into timing issues. So the quality of the portfolio is really what makes the difference in the future. Now still we are proud to also have a high number of new medicines. We are at a record now with 16 new molecular entities in late-stage development. You have seen a rich clinical news flow over 2018. I won't go into details here, but I'd like to summarize again in particular also with the good uptake of the new medicines. We are very confident to get through this portfolio transition to further grow, in spite of the entry of biosimilars, which are -- has, of course, fully started now in Europe and which we expect in the years to come in the United States. Dividend, you have seen, CHF 8 70. Let me now quickly shift to 2019. And again, we expect a rich clinical news flow. We have just got Xofluza approved in the United States and we expect two new molecular entities to be launched in 2019; polatuzumab in hematology and Entrectinib in a number of different tumors. Perhaps even more important from a commercial point of view, in 2019, if you look at the immediate sales impact in the short term, is actually the line extensions. So we expect HEMLIBRA for the non-inhibitor segment in Europe. Kadcyla, again, I mean, stunning data in the adjuvant segments, so that should make a difference. VENCLEXTA is off to a good start. The sales are booked with AbbVie, we have a profit sharing agreement for that product. And importantly, Tecentriq. You have seen the strong clinical results for two new indications, both for very severe form of cancers, where there are not many medical -- really no medical options. Triple negative breast cancer on the one hand, and small cell lung cancer on the other hand, so we hope to get this approved in 2019, and that should certainly drive the growth of Tecentriq. We expect to file for another 2 new molecular entities, satralizumab, this is a molecule for an autoimmune disease, which affects the eye. It's a rare disease, a molecule which comes from Chugai. And then we look forward to risdiplam where we see growing evidence that this could really make a difference for patients. Accordingly, we got breakthrough therapy designation status and I think prime designation as well in Europe to bring this to patients as soon as possible. And I should also not forget Diagnostics. Even though we typically focus on the pharma portfolio, we have -- as you know, we have launched a new line of instruments, in particular the cobas e 801, which had an extremely successful uptake. And now we are coming right away again with the cobas Pro, which should give us, again, momentum in the so-called mid to high throughput segment in the Centralized Lab. So also, good prospects on the portfolio side as far as Diagnostics is concerned. Right. With this I'd like to close. You have seen the outlook for sales in the low to mid-single digits. EPS, again, to grow in line with sales and on that basis, we should be able to further increase the dividend. Thank you very much. And with this, I hand over to you, Bill. Thank you.

Thanks, Severin. So it's nice to see a number of familiar faces that I have met, either in my job as the head of global product strategy at Roche a few years back or as head of Genentech but I look forward to getting to know the rest of you in the years ahead. Yes, we obviously feel really good about these results, so I'm pleased to share them, starting with the sales. And as Severin mentioned, 7% overall growth. If you look at the geographic break out, I think it would be important to emphasize the U.S. result of 14%. This is actually a record-breaking result for us. In fact, the historical sales growth record for Genentech was CHF2 billion, and that was in 2006, believe it or not. And so here we are in 2018, with about CHF 2.8 billion growth, and so a new high watermark, driven largely by the new product launches and I'll say a little bit more about the product split on the next slide. On Europe, with deep decreases that are mostly driven by biosimilars, but partially offset by the new product portfolio. Japan, you've got a little of the same story, new products getting going, and biosimilars just starting to impact. And then in international, a very strong showing, particularly from China, where we were able to get, for the first time, MabThera, Herceptin and Avastin on the national drug formulary and so now literally those drugs have been made available to hundreds of millions more in terms of population. And we're seeing the benefit of that and so good momentum there. In terms of the margin picture and profits, also a nice answer. And royalties and other operating income, we had a good gains here based in part on third-party patents from drugs like Humira, Cosentyx that rely on our historical IP portfolio as well as royalties from partnered products, this is products like Xolair, Lucentis, VENCLEXTA, ex-U. S. On the cost-of-sales side, this includes some royalty payments that we make out, but also about an 8% increase in COGS plus period costs, so this is sort of the manufacturing part. And so it's 8% increase, but that's on volume growth of 12%. So I think pretty good cost containment there even though the number is growing a little bigger than the sales number. M&D, I think this largely reflects various efforts to hold down M&D cost, so that we can maximize our investment in the R&D portfolio. And I think this is a trend that will continue. We've put a lot of effort into really getting more and more productive in marketing and sales, just from a standpoint of the name of the game is innovation, and the place where the main innovation happens is in R&D, so to the extent we can hold our marketing and sales cost down and pump that back into the pipeline, we have a gain, and so that's why you see M&D growing at 4% and R&D growing at 6%. And I think that's something we look to continue as we move forward. And then G&A is an 8% increase, but on a quite small base, and this largely reflects the legal costs associated with the various acquisitions we made as well as just ongoing work to defend our IP portfolio. So this is the product view, and I think what's most encouraging here is if you look at the top 4 growth drivers, these are all new products or newer products. And then if you look at #5, Lucentis, Lucentis will be exposed to biosimilar competition in the coming years. However, in 2018, we showed a major gain in that we demonstrated with the Lucentis device that can be implanted in the eye and can extend the dosing interval from 6 months to even a year versus the monthly regimen that Lucentis has today. And that will not be subject to biosimilar competition because we have the patents around how to do that and how to put it in the device. And so we believe that the Lucentis franchise will be a very important franchise for some time to come for the company. And obviously, a big advance for patients. I think if you look down at the bottom of the slide, you can see the biosimilar impact, primarily happening in Europe and you can see again, Japan in yellow, offset in part by the very strong growth in international. Those orange bars, and that's primarily in China. So it's a bit of a balanced picture there, but obviously that's been hard on Europe. A bigger impact on MabThera, lesser on Herceptin. In 2019, that will probably kind of reversed itself as most of the damage has already happened on MabThera, and Herceptin will be more affected in itself as most of the damage has already happened on MabThera and Herceptin will be more affected in 2019. Then looking at the therapy areas, in oncology, I think we're pretty pleased that we were able to grow oncology 2% in a year where we had major losses in two, very strong growth on Perjeta as the APHINITY regimen has proved to be quite popular. It's a curative setting, and doctors and patients are choosing to go for the maximum potential on cure. I think the hematology franchise, you see Gazyva growing 40% and I'll say a little bit more about the future of Gazyva because we're continuing to accrue indications there. And then towards the bottom, you see some increase in momentum with drugs like TECENTRIQ and ALECENSA. This is 1 piece of clinical data I wanted to highlight. We presented this in San Antonio at the San Antonio Breast Cancer Conference. This is Kadcyla used in patients who had undergown neoadjuvant therapy with either Herceptin or Herceptin plus PERJETA and failed neoadjuvant. And so these basically by failing neoadjuvant, they become high risk for recurrence of metastatic disease. And so these patients were randomized to receive either Herceptin or Kadcyla, and what we showed was about a 50% improvement in disease-free survival and a 50% reduction in the risk of metastatic disease. This has now been awarded by the FDA, just in the last days, breakthrough status and we'll be on the market. We've file already. Should be on the market in the U.S. very soon in Q1, and shortly thereafter in Europe. I think this is a big advance for patients, and will be a nice growth driver for Kadcyla. The other thing I would highlight on this slide is the fixed dose combination which is proceeding with Herceptin plus PERJETA. This is subcutaneous fixed dose combination, and again, we think it'll offer a nice convenience advantage for practices and for patients. And also, a nice opportunity to put Herceptin and PERJETA together. This is another I think, many years and this is the combination of TECENTRIQ plus nab-paclitaxel, the IMpassion 130 study. And what you can see, this is the PFS result in the PD-L 1 positive population, but also, a very strong results on OS. And we filed this data and also anticipate approval from FDA in the first half of the year and shortly thereafter by EMA. So again, a really positive study for TECENTRIQ and one that, we think, is going to put continued momentum into TECENTRIQ. Just a slight sort of summarizing where we stand in hematology. I wouldn't attempt to try to go into all of these, but I would note, we've now -- we had eight breakthrough therapy designations in hematology, quite a broad array of medicines and a broad number of diseases. Maybe to go a little deeper on this, so looking at hematological diseases and the green line that you see is sort of the area that we have covered historically with either MabThera or Gazyva, and more recently, VENCLEXTA. And what you see in the middle there, we have a couple of new things coming, polatuzumab vedotin which Severin mentioned, we filed, will be launching around the middle of the year and targeting refractory DLBCL or patients that have relapsed DLBCL. Really nice data here, and we've actually got two angles on this, because we've got pola and then we also have the next two molecules that are listed. Our 2 different T-cell bispecifics that are targeting NiCd 20 and they recruit T-cells -- CD 3 positive T-cells to attack the tumor. And we've seen some really quite remarkable results there. So we'll be expanding our depth in those therapy areas. And then if you look on the chart on the right side, you can see how we're going into things like myelodysplastic syndrome with VENCLEXTA, relapsed multiple myeloma, also VENCLEXTA, then we have a couple of molecules aiming at AML, and we have already gotten an approval for VENCLEXTA in AML. This actually goes a little deeper on that, so this is VENCLEXTA plus either HMA or LDAC and we really see this as a new standard of care in its first line unfit AML. So about 50% of patients that are not seen as fit for intense chemotherapy. And we believe that it'll be used probably somewhat also in patients that are fit, because of the compelling clinical data. And you can see the data over there in the table on the right, so things like complete responses in 44% or 55% of patients versus the historical figures of 20%. Minimal residual disease negative rates of up to 50% with VENCLEXTA and quite nice overall survival data. So this has been approved, and again, I think it's one reason why we have very high confidence that VENCLEXTA is going to become a major drug in hematology in the years ahead. I mentioned that T-cell bispecifics, and this slide includes some data from 2 of them. So we have 1 of them in gRED and 1 in pRED, and it's an interesting example of how having 2 early R&D centers can really complement, because we're in a position now of having a lot of data and really understanding both the mechanism and the disease. And it'll put us in a good position to decide whether we are taking 2 drugs forward or whether we pick one. But you can see, again, some very nice results in patients who typically do not respond well to therapies. So look at the CR rates in -- again in various populations where you can see ranging from 20% up to 38%. And these are results that we might expect to see from something like CAR-T, only these are therapies we can put in a vial and ship on a truck. And so it's a much simpler way, I think, to treat than cell-based therapy in this case. So stay tuned for more on these. The immunology franchise, which continues to grow nicely. Products like Esbriet, ACTEMRA and Xolair. And again you can see now, we've achieved CHF 8 billion of sales worldwide, and again, we see additional sales potential for these. I won't say more on these right now in the interest of time, but we can talk more about them in the Q&A. The neuroscience franchise, which continues to deliver, and Ocrevus, in particular, in terms of the sales. It's been a very strong uptake. This has been the fastest uptake of any drug in the history of Roche, and one of the most powerful drug launches, I think, in the world. You can see the breakout between U.S. and international. I think what we're seeing, the pattern is, when we first launch, we get a lot of primary progressive patients. And now we're seeing more and more it's relapsing MS and patients in earlier lines of therapy switching -- originally, it was mostly switching from later lines of therapy. Now we're seeing switching from early lines and new patients. In fact, we're getting now -- our most recent data is we have about 39% -- 39% share of switches and new starts in relapsing MS. And the other 13 therapies are sharing the other 61%. So we feel really good about this. This is clearly the #1 therapy in MS now in terms of patients choosing. And by the way, our leading indicators have been very strong in terms of new patient requests for access assistance from our access assistance group in the U.S. And basically, we think we're going to continue to have strong penetration in 2019 and beyond. Let me say a few things about risdiplam. So risdiplam is our small molecule gene splicing modifier for SMA and this has been really quite a rapid journey because we've only had this in the clinic for a few years, and now we're starting to have more and more clinical data accrue. You can see in the chart, this study that we ran, the so-called FIREFISH study part 1. And what this shows is, the increasing levels of functionality as patients stay on therapy, and the blue line on the bottom is the five to seven-month old patients. And so these were patients who had been exposed to the disease without treatment for 5 to 7 months, and were quite ill at this point and gone on risdiplam. And you can see very large gains. But even larger gains and a better outcome when the patients started younger. We've actually initiated a new study called RAINBOWFISH, which is accruing patients that are 0 to 6 weeks old, so basically, newborns. And we hope to see yet better data with that, but we're very pleased with what we're seeing in terms of how it compares to competitive therapies. And the regulators are very impressed as well. They've urged us to basically collect the data that we've got from the type 1 study as well as the type 2, 3 studies, which are called SUNFISH, and to submit them and we plan to begin those submissions in the middle of this year. We'll have sort of rolling submissions and we look forward to approval of this molecule, hopefully, in 2020. And then the last product I want to talk about is HEMLIBRA because a lot is happening with HEMLIBRA. We just got approval in the non-inhibitor patients in Q3 of last year -- at the end of Q3 in the U.S. And we've had very strong uptake. You can see essentially, we've been virtually doubling sales every quarter. I'd say the early read is that there's not 1 kind of patient that is getting HEMLIBRA. In some cases, it's babies. In other cases, it's adults. It's a broad mix of patients that are looking for better alternatives to frequent Factor VIII dosing. We've also launched now in the inhibitor market in Germany, France, the U.K., had very positive reception. And I think most exciting for us is that, we expect to hear -- it's really imminent decision from CHMP on a noninhibitor approval in Europe. And we anticipate that will be in severe patients. And severe patients are patients with 1% or less level of Factor VIII -- of normal Factor VIII levels. And that is the major market, because these are the patients that tend to be on prophylaxis. These are patients that tend to use the most Factor VIII. And it's about -- it's 50% to 60% of total patients with hemophilia, but it's a much larger share than that of the hemophilia marketplace. And so that's the anticipated target for Europe. But in every case, we think we're going to have strong penetration in the U.S. and Europe in inhibitors, and then increasingly strong penetration in the non-inhibitor population in the years to come. All right. So this is a chart, I think we've been watching and waiting for it to develop for virtually a decade, which is this wave of the pipeline and pipeline progress and what is it going to translate to in terms of sales. And this is the sales by quarter. So Q4 of 2018, I think the full year figure was 22% of sales. But in Q4, we were up to 24% of total sales were from these new products. And I think, again, as you can see that rate is growing. You see the list of products and the timeline on the right. And then if you look at the yellow box on the far right, these are the next products up, so as Severin mentioned, we intend to see pola and entrectinib both launching around the middle of the year. And then satralizumab which is a drug coming from Chugai for neuromyelitis optica, which is a severe neurological condition. It somewhat resembles MS, but it's different, it's targeting the optic nerve and the spinal cord. And patients have blindness, but also inability to walk and motor function, lower motor function especially. So we're looking forward to bringing that out, and then risdiplam, that I mentioned, in SMA, are probably the next ones up. The late-stage news flow. I won't go through the list. I think you all have been tracking it. I think what's encouraging is, we've had both from a Phase III readouts as well as regulatory progress, it really was a banner year. And we -- if we listed all the Phase III readouts, I think we were at about 75% positive rate and that was with a high standard where we excluded studies that were positive in terms of meeting primary endpoints if we didn't believe the clinical meaningfulness was high. So we're above our target rate and we want to keep it up. But good progress there. This is the 2019 key news flow. And I think maybe what stands out on this is, obviously, those 2 approvals you see with TECENTRIQ plus chemo, a few lines down in triple negative breast cancer and in small cell lung cancer, both areas where TECENTRIQ has great data. It's sort of all alone for now in cancer immunotherapy in those in terms of the data. And with U.S. and EU approval ahead, we think we'll have nice momentum for TECENTRIQ. So I think that's all I'll cover for now, and I'd like to invite Michael up. Thank you.

Thanks, Bill. Good morning. Good afternoon, everybody. It's the first time to be with you in this set up. I'm happy to have the opportunity to show the Diagnostics results to you. You can imagine being the first time with you and being in the position to show some, I think, kind of positive results makes it much easier. With this, I would like to share with you where is Diagnostics ending 2018. And in this context, I mean, we had a great sales year, by all means. And with 7% growth and about CHF 13 billion sales in Diagnostics, it was probably one of the best years ever we had. And I would like to also mention that the biggest contributors for this success were, basically, all businesses we have with Centralized and Point of Care Solutions leading the way, which are our most important part of the portfolio. As Severin already mentioned, immunodiagnostics is part of it, with 20th consecutive year of double-digit growth. And meanwhile, already about 40% of our divisional sales. In this context, also, Molecular Diagnostics did well. And Diabetes Care, in the past, we had not always the best in shares -- news to share with you. With Diabetes Care, we reached a certain level of stabilization. I mentioned that already before. And I'm quite positive and happy that we've reached that level last year. Tissue Diagnostics continues double-digit growth for the 10th year in a row. Positive is also that all countries, all regions contributed to this great success. And great is also the development in the Americas. North America, 7% growth, 25% of our sales. Latin America, 9% growth, and Asia Pacific -- Asia Pacific with 13% growth. China thereof 16% growth. And the emerging markets in total already achieving about 40% of our divisional sales. In Japan, 6%; EMEA, 3%, also great development; and above market growth. The Diagnostics Division has also been supported by the different business areas. And as I mentioned, immunodiagnostics, 11% growth. And this in a year where, as you know, our key competitors have launched their new platforms, their new serum work area diagnostic systems. And I must say, we had a great year. And we expect this to continue. Molecular Diagnostics point of care, 1 important area, point of care did very well, especially with a new lab-in-a-tube device, a point-of-care PCR system that we launched first in the United States, which addresses a need, for example, in flu testing in the doctor's office. And the flu season last year, as you know, was very heavy in the United States and in the world in general. And the sales of this point-of-care device went very well with 162% growth and others. I don't want to go too much in all the details. What does it mean in terms of profit? The division performed well also in profit and with 9% profit -- core operating profit growth our achievement was, I think, a great result. two percentage points higher than sales. And to summarize this now, sales grew above market, we've won market share and the profit above sales. This is the way we want to go. This was also based on certain specific factors. Last year, for the first year, we did not get PCR license royalties anymore. This was a negative impact. We could compensate for that. And we also invested still highly in the R&D side. We are investing very much into new systems, in disruptive technologies, digital technologies as well as parameters to complement our complete portfolio. And the complete portfolio is very much driven by the integrated core lab strategy that we have, expanding entry into new disciplines, expanding our current disciplines, adding new test parameters to make it very convenient for customers to decide to bet basically their core lab's requirements on 1 integrated solution, which is the one from Roche. We had analyzers and assays, the broadest portfolio of assays on these integrated solutions, automation software and as well as the area of digital solutions in the area of lab decision support systems as well as clinical decision support systems. Cobas e 801 was a great success story already. We achieved now close to 2,000 placements end of last year. About 1,000 in last year alone and we also launched the cobas e 801 in China, a major market for us. Cobas pro was launched on time and on budget in December, and this is a major product for us, which I want to allude a little bit more on this slide. We are targeting the medium to large laboratories where we are already market-leading with the cobas 6000/8000 devices where we have about 15,000 cobas 6000s already in the market, 5,000 cobas 8000s and right in this area, the cobas pro will be helping us protect our installed base but at the same time, win new customers. We have run several already launch events with customers. They really love it and we have great expectations for this system going forward. I also mentioned, it is important to add tests. On all these platforms that we have in the marketplace we run last year 20 billion different tests. And the tests are the ones that generate the profit and are important for our further growth. And to continue expanding the menu, we, for example, invested into the growth hormone portfolio to complete it, and we achieved it last year. And this is also one of the examples that helps us further increase our penetration in this market and helping our customers to have the best choice of a broad portfolio to be used in their integrated core lab solutions. And that also means we need to connect different disciplines. We do that by adding, not only Roche disciplines that we have, like now the HbA1c testing high-volume system, which is unique in the marketplace because HbA1c in diabetes is one important test that requires more and more high-volume devices. But also, adding our Molecular Diagnostics devices, which are unique. Nobody else can integrate that into core lab right now. And also competitor's systems, if we don't have it in our portfolio like, for example, some infectious diseases tests that we offer them with other connected systems. The global access program is something that this -- and you see Severin on that slide, is a one of those areas of that are very close to our heart, not only from a business perspective, but also from a social responsibility perspective. We have, last year, expanded the access to HIV viral load testing in low resources countries, especially in Africa, to 14.2 million viral load tests in Africa. 14.2 million test means 14.2 million infected people that are monitored with our test on a regular basis, and this helps them to maintain the viral load levels low and be able to live a normal life. Otherwise, in the past, they would have just died. And here, we have expanded our portfolio. Our installed base in the countries. Basically, in all major African countries we have our infrastructure placed, so this test can be done. And with the launch of the cobas Plasma Separation Card, we're now also testing of newborns, babies, as well as people in absolutely low infrastructure situations to just with a drop of blood on this separation card, sent it to the labs and get the results back in a reasonable time. One part, I said is laboratory, efficiency, tests and the other part is digitalization. And in digitalization we have invested especially with great collaboration with the University of Leeds in the NHS system in the U.K., the uPath enterprise solution software that allows the laboratories, pathology labs, using our digital pathology solutions to just exchange their slide views, their interpretations also with their peers and provide better results, better interpretation of results to the oncologists, for example. And coming to an end, we have the key launches in 2018. We made 11 out of 13. In purpose of saving time, I will not allude too much to this in total, but you can ask me afterwards. I mean, it was a great year also from this side. And I want to focus on 2019. In 2019, we have a broad portfolio of solutions to be launched to the marketplace that covers instruments, devices. It covers tests, assays, coagulation and disruptive technologies, like in microbiology, with the vivoDx MRSA test, important health care situation in the marketplace and also a lot of software solutions. NAVIFY Tumor Board sequencing, Mutation Profiler and even Diabetes Care, we are working on a lot of new software solutions because, as we speak, our division is in full focus on transforming our division -- divisional processes into the new area of digitalization across the whole value chain. And that makes us believe that we're on the right track going forward. Thank you very much. Now I hand over to Alan.

Thanks, Michael. Great performance. Good. Great to see you. Thanks for coming. I think it was really an outstanding year, I would say. And let me emphasize right at the beginning, I think we always told you that we can grow in a period of biosimilar impact. And I think we had a great start. I think there's quite a momentum in 2018. And that will certainly help us out. Yes, let me go through the highlights, and basically, I can do my presentation on that slide already. With a sales growth of 7%. I think my colleagues did a great deal in explaining that, so I don't want to go back to that. But the biosimilars impact, at the same time, was a minus CHF 1.3 billion. Don't get irritated, Severin had a minus CHF 1.2 billion on his slide, that's just the currency impact. So it's perfectly the same number with the minus CHF 1.3 billion. The core operating profit is up plus 9%. I will come back to that and go through the P&L. Core EPS growth, up 19%. Excluding the tax reform, 8% so evidently margins went up. And then the dividends further increase to Swiss CHF 8.7 billion proposed. Cash flow, I will tell a story about that. I will touch upon net debt for all of you who are concerned that we get a net cash positive rather soon. And you will say that we managed that pretty well. And then the net financial result, well, good news and perhaps a little bit of a challenge ahead of us, so I will talk about that. And then you see the IFRS net income, up 24%. I will go through the noncore items later on. Yes, when you look at the performance, the 7%, I think, core operating profit -- oh let me say first, look at the currency. Currency impact was rather low in 2019 -- so at 2018, it helped us. 2019, I'll come back to that, but we assume as we always do, that might also be relatively small. And so I think you see the core operating profit going up. I'll go through that, core net income goes to the 20%. Then you see really core EPS up 19%? Quick remark here. You might ask yourself, why is the momentum coming down a little bit and that's a noncontrolling interest, that's basically Chugai. Chugai had a fantastic year and evidently they had even a higher dynamic than we had and that brings the dynamic a little bit down. IFRS net income, we'll talk about that and I will also touch upon the operating free cash flow. On free cash flow, I won't talk about so when you say okay, operating free cash flow is growing by CHF 1 billion roughly. Why is free cash flow growing by CHF 1.4 billion? Tax. That's really another residual, if you like, of the U.S. tax impact. Now let's go through the P&L. And let me start with the royalties and other operating income. And I think that Bill and Michael told the story already. I think the CHF 200 million is really just basically a little bit of higher royalty income on the pharma side evidently and the other piece is really outlicensing income that we have had. That's basically it. For me, it's important to say it's not really pushed by any product disposal gains that we have had. Not at all. Cost of sales, up 8%. And let me say here that in total for the group, volume went up 11%. We had 12% on the pharma side, 9% on the Diagnostics side so you'll see a really a huge volume momentum moving forward. And I think that's really quite good. When you look at the gross margin, the gross margin really when you compare half year to half year, 2017 to 2018, it's pretty comparable. When you look at the gross margin overall for the full year, it's exactly the same than in 2017. So I think really that number makes total sense. I know that you've seen the first half and there were relatively high hopes for the full year. Nevertheless, I have to say the volume growth was really outstanding. You might have also noticed the sales increased in the last quarter, which speaks for the volume increase in total. M&D are CHF 400 million up. I think here that's a very reasonable spend with all the launches that we have had, which we showed was important and we invested well. R&D up CHF 640 million in total, CHF 550 million on the pharma side and I hope you agree that made sense with all the things that we have moved forward. And CHF 100 million on top on the Diagnostics side. And then G&A, it's CHF 100 million on top and that's really basically half of that is the legal costs from our defense on the biosimilar side and the other half is coming from the acquisitions so FMI, Flatiron and so on. Good core operating profit up by, as you can see on the slide, CHF 1.7 billion, 9% and even the margin went up for the group as a whole. Now when I talk about margins I think you see them here. Margin went up for the group in total. It went up, respectively, on the Pharma side. And it also went up on the Dia side and I might even say, it even went up in Diabetes Care, which I think is important to mention after a couple of years where you have not seen a lot of growth in that business. So I think very, very happy with that. The core net financial results. As you see, the fantastic development that we have had in 2018, but you also see at first sight, there's sort of a little bit of challenge ahead and that's the interest expense. And you see first time for a long time, the interest expense went up. As you know, we have CHF 18.8 billion of gross debt on the balance sheet, so everybody who gets dragged away by the net debt, that's 1 element. Gross debt is the other element, so CHF 18.8 billion. We pay roughly CHF 605 million for that CHF 18.8 billion. I think we optimized that rather well over the years. So we've really come -- came to a floor now and, okay, we will see how that goes on. I will talk about that on the next slide. Let me mention here the equity securities, quite a positive impact and that's really our -- if you like, minority investments in the Roche venture fund. One is pretty known, that's Abaxis, we had a stake in AveXis, the company was sold to Novartis. We had a gain from that, and then we have another point where we had another investment in the venture fund, in a company which is listed and certainly there's some fluctuations for that. The other company gave us roughly CHF 200 million on top. So I think quite a nice development here. Currency is interesting. Currency is hedging costs. And when you look really at what I've said before on currency, you would argue well, it wasn't a lot of impact. But during the year, the volatility, especially on the other currencies was quite stronger compared to last year, so we had a little bit of higher hedging cost here. So as I said, 20% improvement roughly compared to 2017. But when you look, really, at the longer period, I think that was quite a ride from CHF 1.4 billion in 2015 to the roughly CHF 600 million in 2018. But I have to say now, for '19, it's not really looking so great. I think on one hand certainly, when you look at -- at the investment of the venture fund, that's something we don't really have control in because these are minorities. And that's not really in our hands. We'll see how share prices will develop. We have now to report now the unrealized gains here. So we'll see how that moves. But I can imagine that there might be an additional burden around CHF 200 million here when it comes to the net financial result in 2019. Yes, another success story is the tax rate. 26.6% and the U.S. helped us quite a bit. So now finally, I think the U.S. comes down to the average of the OECD with that tax rate. Gave us a 6.9 percentage point improvement. Now to 19.7%. We expect the group tax rate to be around 20% in 2019. So it looks like a sustainable effect and certainly we appreciate that. Yes, the noncore items. First of all, let me start with the core operating profit that went up CHF 1.5 billion. Then you see the global restructuring plans, the number came a little bit down from CHF 1.2 billion in 2017 to roughly CHF 900 million in 2018. And these programs are part of our transformation of the company structure, and these are very targeted programs but they all contribute to the financial flexibility moving forward. When you look at the amortization of intangible assets, that's basically Esbriet, because Esbriet with the significant amortization that we have every year so intangible assets go down and, therefore, amortization goes down. And then you see the impairments of intangible assets here. Quite comparable to what we have had last year. Certainly, what's significant is that we wrote down the goodwill for InterMune. You have seen what we have done last year. Interesting is, that has does pretty well. I think it has grown with 19% in 2018. In the last quarter, even was 26%. And that's even led to the point that the intangible asset for that we rolled up a little bit. I think really the perspective is better compared to 2017. Nevertheless, are we really matching with Esbriet our original expectations that we have had at the time when we bought InterMune? Well, okay there is a certain deviation and that lead to that point. For everybody who's interested, I think what enabled us to do that goodwill impairment was really the fact that we changed the structure of our cash-generating units. In the past, all these pharma goodwills were under the Pharma Division, if you like, which is highly cash-generative, as you can imagine. Now, we have stripped down really to -- we call it strategy, we call it product, and we call it technology. And in these 3 buckets and that enables us to be a little bit more flexible when it comes really to the goodwill impairments itself. Alliance and business combination, you see the positive impact last year, this year is slightly negative impact. I think last year, we benefited from released considerations last year -- contingent considerations I should say, which we didn't have in 2018. And then legal and environmental, we reduced a provision for one product quite significantly by CHF 200 million. We didn't have that in 2018. And then the total financial results and taxes, and taxes once again play a role here and certainly the U.S. tax reform by 1.8 over 274, which increased the IFRS net income by about CHF 2 billion to roughly CHF 11 billion an increased in constant rates of 24%. So let me talk about cash. Sometimes overlooked in our industry. We have attributed by a good net working capital management very much driven by the inventories. I think we had a reasonable development on the receivables side. We had the slight increase on the accounts payable, major point was really that we brought the inventories in the right direction, which makes sense when you look at the significant sales growth that we have had in 2018. Other networking capital is more about provisions that we had for royalty payments and whatever so nothing really here of major impact, if you like. Then inventories in -- or the investments in PP&E, Chugai, Genentech, I can mentioned here a little bit all over the place, and then investments in intangible assets, I think that's what we're doing for the pipeline and which just makes sense. Good, when you look at the continuous improvement of the operating free cash flow, I think that really moved in the right direction over the years. We had slow years but I think now I think we're really at a very significant level and let's see what we have in store for the next years. Yes, I said for all those who are really interested in net debt and when may we get net cash positive, you see really we managed that pretty well in the year 2018. We started with a net debt level of CHF 7 billion and we ended with a net debt level of CHF 5.7 billion. And why that? Well, we had quite a significant operating free cash flow generation, as I explained, then we had, well, CHF 3.9 billion outflow for taxes and treasury. That was last year roughly CHF 4.4 billion, so the tax reform shows, if you like. And then I think, well, we have the dividend and we have the transactions that we have done. The M&A transactions. And the CHF 5.7 billion outflow, we were able to overcompensate with our cash flow generation. On the other hand, I would say, we kept net debt pretty much stable. Good. Quick remark on the balance sheet. Well, cash and marketable securities, always good when you have cash on hand, we increased that a little bit. The other current assets, well, inventories went down, as I explained before. The noncurrent assets, well on one hand. Goodwill, well, evidently went down, intangible assets went up and that was basically a match, and then really the increase itself came from PP&E. Current liabilities, I think here accounts payable went a little bit up, provisions went a little bit up and when you look at the non-current liabilities, then it was really about the pension liabilities and this is just an interest rate topic, which came up, interest rates went slightly up, so liabilities went down. And then you see the equity, which is now at a rate of 39%. Net debt-to-total assets now at 7% year-end 2017, it has been at 9%. Good, outlook. Well, I talked about currency already. And about the impact. And what you see is that, the U.S. dollar gave us really a slightly negative impact on a full year basis. When you look really at the euro, gave us a slightly positive impact on a full year basis. And in total, it was pretty much balanced because when you look at the right-hand side and the impacts that we have seen in Q1, half year, Q3 and also full year, rather low. And then you know what we're doing with the currency rates. We keep everything stable end of 2018, and then project everything until year-end 2019. And based on that, which is very hypothetical and certainly not the truth, we expect a very small impact on sales core operating profit and core EPS for the full year 2019. But this will change for sure. That's the current view based on the assumption that I've explained. Good. I think we have to give you the right base for the 2019 guidance when it comes to core EPS, and I'm happy to do so because, as you know, we have in our core EPS incorporated a currency impact, which basically comes from the net financial expenses. And you see, really, what we have reported as core EPS is CHF 18.14. You need to increase that by CHF 0.16. Where is that coming from, the currency loss is CHF 160 million, so it's a loss of CHF 160 million. We will find that on Page 54 in the finance report that is outlined. When you then take yes, this impact you apply a certain tax rate. Why not 19.7%? Evidently, these are hedging costs, if you like, so well, it happens in Switzerland, perhaps 19.7% is a little bit on the higher side here. But when you then divide that through 860 million shares, you get to the CHF 0.16. And that gives you a starting base for 2019 of CHF 18.3. Good. I think Severin said everything about the outlook already. Once again, thanks for coming, and we're happy to take your questions after we've rearrange the setup. Thank you.

Two minutes or whatever. Two minutes. A - Severin Schwan: Okay. I think we're all set again. Sachin, Can you take us off? Can we have a mic in the second row, please? Thank you.

Sachin Jain from Bank of America. 3 questions, please. Firstly, on U.S. biosimilars, I wonder if you can give us some color on what sort of erosion you expect in the U.S. in the second half, kind of like a big -- a key delta. So any color on how many competitors for each product do you expect to have full labels. And your time line is based on settlement dates or assumption of non-settled companies. Any color you can give there, and I know there is some sensitive information. Secondly, on midterm margins, Severin. Roughly a year ago, you cautioned against dramatic expansion. Consensus has reset to a limited expansion of your 100 basis points over the next couple of years with portfolio rejuvenation. I wonder where you are now more positive on the midterm, also noting that the '19 guidance also implies margin expansion to offset the net financial pressure that I alluded to. And then finally, just a quick product question on Huntington's. And again, you referenced it on the wires this morning. Any color on discussions with the regulators as to what data could be behind the best-case filing time line?

Thank you very much. Yes, perhaps I'll take the margin question to get us started, and then, Bill, if you can comment on the U.S. biosimilars. And I'm thinking, yes, it's true. That was, I think, a year ago or so. And at that time, there was -- I think I was interviewed actually. I remember at the time. And the -- we all remember. And the context, or at least the context I perceived during this interview, was kind of margin will expand and go into the sky. And what I said is, well, confident about our growth through the biosimilars. The new medicines have comparable margins. But they won't go through the sky. Now here we are a year later. It's true, we could expand the margins a bit. And we'll see for the coming years. I mean, it is very good to see that we have this good momentum from the new medicines because, eventually, that makes the difference. If you have new differentiated medicines, that's actually what counts in the long term. If you don't rejuvenate your pipeline, eventually, your margins will go down. And if you launch new innovative medicines, you can keep margins or potentially even expand margins. Now on a more granular level, I would say the following. We have comparable cost profit margins. There is a bit of a negative effect due to the royalty load we have associated with OCREVUS, and you have seen OCREVUS, of course, is taking a bigger share as we have this tremendous success in the market. But on the other hand, we expect that to be compensated with VENCLEXTA. As you know, we don't book sales for VENCLEXTA, and as we go forward and share the profits for that medicine, that, of course, should also compensate. So broadly, I'd say I'd expect a stable development as we go forward, and you see this also reflected in the guidance. If I look at the 2019 guidance, actually, the much bigger challenge from an operational point of view is to digest the decline in the royalties because we'll lose the Cabilly patent. That's where the pressure is, and that's where we have to work hard to continue with our productivity initiatives. There is a lot ongoing throughout the value chain, but we are confident that we can digest it. And that is reflected in the guidance. I hope this gives you a bit more color. And with this, you had the crystal ball for biosimilars in the U.S.

I'm afraid I may disappoint you with my answer on U.S. biosimilars, not with the numbers, but with the lack of specificity because, frankly, there's 2 obstacles. One is that nobody knows exactly what's going to happen and how it's going to play out. There's ongoing sort of litigation and ongoing sort of disputes over IP. And it is rather difficult to see how that's going to play out and then what choices various biosimilar companies might take with respect to whether they launch with partial labels versus full based on settlements, et cetera. So I think there's that level of uncertainty, and then the other level of uncertainty is what will the uptake be when they arrive. And frankly, there's not lots of precedents to look at on that. So I think we've baked in some impact in the second half, and it's not a huge impact because it happens in the second half and there's only so much that can happen. So I'm not sure I can really say much more than that because I don't want to tell you more than I know, how about that.

Can I have the number in guidance?

No, we have a plan, and -- but I think the guidance reflects the middle of what we think could be. And with respect to the question on Huntington's disease, yes, I mean, I think there's a lot of excitement in the community and with the regulators around the potential to lower the Huntington's protein that we've demonstrated already. And there is a will to move and expedite the approval of the product. And we'll -- I don't think it's -- it's not set, so we don't have a hard and fast time line internally, but I think it's something that we would hope could happen in terms of filing potentially in the next 12 months.

In case if we go this direction. Yes. Can we have the mic here in the middle? We'll come back to you.

Richard Vosser from JPMorgan. Perhaps we can start with OCREVUS and just perhaps you could talk about the U.S. market and the retention of patients, proportions that are returning. It seemed a little bit softer this quarter in the U.S. for OCREVUS. And maybe talk about -- you alluded to more sales in relapsing-remitting. Is PPMS now declining -- or not declining, but growing at a lesser rate? Second question, just on HEMLIBRA, perhaps you could give us a framework on how to think about the uptake in inhibitors in 2019 and maybe the early reaction in 2018 in the fourth quarter to the launch.

Did you mean inhibitors or non-inhibitors.

Sorry, I meant non-inhibitors. And then final question, just on LUCENTIS. How sustainable is the return to growth from the new device? We're seeing good uptake from that. How sustainable is that?

With respect to OCREVUS and retention, I think you said -- you speculated the numbers were soft. Did you mean retention or you meant the sales results.

Sales. So I would say the sales result may be a little bit soft relative to expectations and potentially, maybe in the U.S. And then that leads, I think, to speculation over many years of retention for the products, so perhaps you could discuss that

Yes. I would say, first off, the -- I think we're beating estimates for a number of quarters. And -- but with respect to our forecasts and our projections and our plans internally, we were very much on target. And we didn't -- I mean, there was nothing amiss for us in Q4 in terms of patient retention, patients returning. The information, the data we have on that has held up very well. And as I said, the leading indicators are quite strong, leading indicators in terms of switchers, share. Also, we get -- we assist probably around 80% of patients with reimbursement in terms of helping them navigate the process. And so we get those requests directly from doctors for helping patients, and that's probably our best leading indicator and it's very strong and so far, in January, has been -- yes, held up very well. And so we -- we're quite confident in continued strong growth in OCREVUS.

PPMS

And then PPMS, it's certainly not declining. It's still increasing. It's just my comment was that when we first launched, there was a bit of a bolus of primary progressive patients who didn't have any options, who now had an option, so we got an immediate kick from that. But it's -- we've continued to penetrate PPMS, and there's no sign of any sort of backing off of that. And LUCENTIS? Yes. So LUCENTIS, I think there's 2 things going on. We -- I mean, I think one of the competitors had a bit of a challenge last year and led to some switching and some -- both switching of some patients, but also switching of some loyalty. And in addition, we launched a prefilled syringe, which has been very popular. And so that combination, I think, fueled the growth in 2018. In 2019, there's a potential launch of another competitor. I think there were some onetime gains that we had last year from that -- the event that I mentioned of the competitor. And so I think -- we think LUCENTIS will be quite a strong showing this year. But in terms of continued incremental gains, I'm not sure we're projecting a lot more.

And HEMLIBRA

And HEMLIBRA, yes, yes. This is a -- it's a very interesting one because HEMLIBRA is such a game-changer. I mean, when you're going from patients -- well, first off, let's say, patients that are severe hemophilia that are on prophylaxis, they're getting -- they're doing factor VIII infusions two, three, four times a week. Now they have an option for a once-a-month subcu that's going to give them equal or better protection to what they've had. And so that -- you'd say, "Well, that's great, everyone now ought to switch right away." On the other hand, a lot of these patients have been giving themselves factor VIII infusions for 5 years, 10 years, 20 years, 30 years, and so there's a fair amount of trying to guess at what rate will they actually switch if they're sort of used to the current system. And yes, I'd say, internally, we've been -- I don't want to say, I don't think our guessing has been particularly accurate. The uptake has certainly been faster than we expected. And I think we would exceed our, I guess, our faith in our own crystal ball if I was going to give you a guess as to how that's going to continue. But what we know is we get a lot of anecdotes of both. There are some physicians who are now saying to all their patients, like, "Hey, why aren't you switching to this product? I mean, why are you taking factor VIII?" We have some patients that are kind of going around telling all their friends because the community is quite tight, "Hey, hey, dummies, why are you still taking factor VIII?" So I think there's a lot of anecdotes. But anecdotes don't translate into forecasts, so I think we're quite optimistic based on the trends we're seeing. But I think it's going to take probably another couple of quarters before we think we've figured out what the real trend line is.

Thank you. Let's take the next question here. Yes.

It's Andrew Baum from Citi. 3 questions. The first was I'm struck on Page 12 when you go through your pivotal enemies just how few hail from Genentech versus the other research centers or externally. I also note that one of your former colleagues recently had a reorg inside his own organization and is effectively integrating his U.S.-based research hub. I'm thinking obviously of MedImmune. Now I wouldn't want to equate the productivity historically of Genentech with that of MedImmune, but I'm going to ask the question anyway. So under what circumstances would you think about integrating Genentech inside the large organization? And then second, assuming that the answer's going to be no, which I think it probably is, how do you look at the recent output relative to the other parts of the business? Is it just a function of time and cadence? Is it related to departure staff because of Boston and competing companies in the share price and options and so on? Or are there other factors? And so just more broadly, that general concept of Genentech productivity. Second, could you talk to the potential for accelerating the filing of one of your bispecifics against CD20 for DLBCL and FL through expanding a dose -- once you expand the dose in the Phase I, could you actually file that rather than waiting until you get through a registration file, given that's a treatment-refractory or a heavily refractory patient? And then finally, just some comments on your confidence in being able to work your way around the reimbursement and very P&L-savvy ophthalmologist to secure that they will use the port for LUCENTIS as a commercial barrier rather than a biosimilar?

Very good. Let me start with the question related to the structure and Genentech. And yes, you read me well. I continue to believe that one of the strengths of the culture of Roche is this more decentralized approach, that we really truly empower people, that we bring down the decision-making to the lowest level possible in the organization. And part of that is also to keep more autonomous units. And I'm deeply convinced that if you talk about research and if you talk about early development, it's really about the quality, it's about the cutting-edge science. It is not about scale. If you move into late-stage development, if you move into other functions, then we want to leverage the scale. And we won the global approach. But when it comes to research, when it comes to early development, the -- we think about it in terms of three engines, if you like; Chugai; pRED; and gRED, who feed, if you like, the global late-stage portfolio. Now it's true that it goes in waves. And if I look back 10 years ago, I got exactly the other questions. It was like, why don't you close pRED, why don't you integrate pRED into gRED, and what we have seen over the last 10 years is that, actually, pRED has delivered. And a lot of transitions, which we saw more recently, actually come from pRED. But still, gRED continues to be a very vibrant place. There are still very important molecules coming through. We had some of them on the slides today, which are also coming from gRED. It's true that there was a bit less coming through into late-stage over the last two years. If I look at the projection for this year and for the coming years, there is more to come actually.

Yes, I mean, and for example, of the 16 in pivotal today, there's 4 of them that are from gRED, and you can say, well, why isn't it 8? But I remember, a few years ago, we had the opposite conversation. People were saying, well, why wouldn't we integrate pRED into gRED because they're not delivering enough and these things don't come -- they come in bursts. And when I look at the early time line, it's very strong.

Time delays, et cetera, you know the time and you make the decision, you make the decision on something, which was kind of 10 years ago. This is the other element of it. But I stick to this approach, that the independence, this kind of diversity of thinking, eventually enhances innovation. Look at Chugai, for example. I mean, some of the molecules they bring through, I mean, we would never have pursued. HEMLIBRA, we would never have done. And if you talk with Mike and if you talk with William, they will say, we would never have done this. The Japanese were kind of -- I don't know how many thousand molecules they've tried out. Eventually, one worked. Never ever could this ever happen in our other sites around the world. We just have a different approach to that. And I remember many discussions, and I won't go into the details in terms of names and molecules, but I can tell you how often did the head of gRED say, now I mean, this will never work out. And it was in pRED. And how often that the pRED guys having this, it will never work out with the gRED. And those are now, a couple of years later, actually, the most promising molecules. So it is all important that you keep this diversity of thinking. And I rather have a little bit of duplication in the organization. You have mentioned one with the two bispecifics, where you then have the problem, that you have two of them, but I'd rather have two of them where I can choose based on data than having nothing because we have a big bureaucracy where people just can't put their ideas into action. So no, we won't integrate. Now in terms of retention of people, overall, voluntary fluctuation levers are pretty stable, but we do lose people from time to time. And it's also true that, in particular in South San Francisco and in Boston, there's a vibrant biotech community and it goes in cycles. So last year, a lot of money was going into San Francisco, in particular. They were all looking for people to lead these new startup companies. And they knocked on Genentech doors. There is no doubt. And by the way, the same is true with tech companies. They also start looking for scientific clinical talent, and they would also come to a company like Genentech. And I regard this actually as a good thing. I mean, I'd rather have people coming to us because we are known to be a great place, where great science is being done compared to a situation where I don't have the people and people wouldn't come to us and be interested in our people. And it's our obligation now to create an environment where they -- which they find exciting and where they are committed and where they continue to stay with us. And I think this decentralized approach is part of it. If we would integrate it and manage everything out of Basel, I don't think this would help on that respect either. So yes, there's a lot of competition and we always have to be sure that we are competitive, but it's not at the level where I would be worried. Right. I think there was a question on the bispecifics, right

And by the way, if you're still skeptical about the talent drain question, I think the day you stop hearing about a talent drain from Genentech, you should worry, right, because we've had a talent drain problem for 20 years. And again, if you stop hearing that, then the talent must have run out. But I think we're -- we -- yes, you don't hear as much about the talent we're bringing in, but we've brought in some real stars in the last few years, and we still have the ability to hire people that are running major research institutions and things and we're hiring them in as department heads in gRED. And I think that as long as we can continue to do that and really hire the best, I feel pretty good about it.

And it's such an important question. I have another thought, sorry. We'll get to the bispecifics. But it is really important, it's really, really important. For a company like ours, it's all about innovation, it's all about science. So we have to really, really take care of that. That is what is at the core of this company, and that will define our success for the future. So it's really, really important. And we have many discussions as you can imagine on what does it take to be the best environment for this type of people. And just a bit on that, this other element, I mean, if you look into the Bay Area and beyond, I mean, it's full of companies, which were founded or where Genentech people contributed to the success of those companies. And what it also means actually is we are very well-connected to the network. And it's not unimportant. I mean, if there are small companies, it's less about the money. For small biotech companies, it's much more about the right partner who helps you to bring a molecule to the next level. Now if it's a later-stage asset, it's a different story. It's only about money, it's a pure financial question. And to have this network out there of ex-Genentech people, who are part of many startup ventures, serves us well also in the longer term. And people talk with each other. And if we leave on good terms and if we maintain these networks, actually, it's a positive cycle where we also can attract again new people, people who potentially also think, 1 day, they will do something else. And if they are successful, then we have a natural pattern to speak with them as well. So it's not all bad. The important thing is that you do not only lose good people, but that you also keep a pipeline of the best talent joining your company. Anyway, we'll keep an eye on it. So bispecifics.

Yes. T-cell bispecifics. So a question about innovative approaches to sort of filing based on the early data. We're -- how to say? That never is far from our minds. So we're in Phase Ib with both molecules, and we will be examining all options with regulators on both sides of the Atlantic and elsewhere. So let's see. I mean, I think it all will depend on how compelling the data is and what other molecules are showing in the comparable patient categories. So -- but let's see. On LUCENTIS port reimbursement, I think there's a trade-up. So you're sort of asking, well, if actually, if a doctor is giving 12 shots a year and they're getting paid for that, are they really going to want to do something where they're giving two shots a year? And I think the reality is that patients aren't getting 12 shots a year. Patients are getting four or five shots a year in year two and after, and they're losing their vision. And in addition to that, the port delivery device actually requires a retinal surgeon. It actually requires a surgeon to operate, and we hear from a lot of the surgeons, they're like, hey, this is actually -- this is what I trained to do. I trained to do surgery, not to basically do 100 injections a day. And so I think there might be some of those economically minded ophthalmologists who would prefer to just keep giving lots of injections, but I think we'll have a really strong case, both from a patient standpoint and from a doctor practice standpoint, of why this is a good way to move forward. So

Yes. There is one question. We have the rep here from Graham Doyle, and he asked about the future of fixed dose combinations of Herceptin and -- thank you, all the fixed dose combinations of Herceptin and PERJETA and how you see the future in the United States and outside of the United States, how much you potentially could convert into that.

And so we're talking about the fixed dose of Herceptin plus PERJETA. And I think this is -- I mean, it's an attractive option, particularly for patients who aren't receiving chemo, which, for example, in the adjuvant setting, they're not getting chemo, and so it's a choice between coming in and getting an infusion of Herceptin and an infusion of PERJETA or coming in with a really simple injection with Halozyme. And we think, over time, it's going to make sense because it's just -- it's better. It's more convenient. And I think it may be the kind of thing that takes a while to convert if physicians are comfortable with the practice of doing the multiple infusions. But there's a lot of patients getting a lot of infusions there. I mean, Herceptin plus PERJETA is the standard of care in metastatic breast cancer, standard of care in early breast cancer. And it starts to get pretty compelling, even more so than the switch from, say, Herceptin IV to Herceptin SubQ because there, you're just replacing one infusion and now you've got an opportunity to replace two.

Okay. Thank you. So let's move on, on that side. If we can take here in that row, yes. Thank you.

It's Mark Purcell from Morgan Stanley. I was just going to follow up on the same subject there. I guess the life cycle management behind the CD20 franchise is very clear, that hurts your franchise slightly less, that a bispecific earlier in development better. Just in terms of the timing, it doesn't look like you're going to launch a subcut Herceptin in the U.S. You're going to go straight to the combination. And given you have visibility on three of the five players, and the first one, the biosimilar's going to be launched with very limited capacity for the first few years, I just wondered, obviously, you guide on risk rather than likely outcome, but if the decision to go with a 2020 launch on the combination is related to what could be a greater window than currently people are modeling. And just going back to the dynamics here, can you help us understand whether this is also partly related to the Part B pilot programs and the payer situation in the U.S. in terms of the choice to go over the combo as opposed to the single subcut Herceptin. The second one on risk of plan, could you help us understand the sort of positives, the negatives and pivot points here versus gene therapy? In SMA 1, 2 and 3, I guess, at the moment, you have parents who will do anything that they can to get the best outcome from their children, including combination therapy. But clearly, there's 2 muscles you're trying put out, the cutout muscle and the skeletal muscle. The skeletal muscle has huge amounts of turnover over the year -- the 20 years of these -- of these patient lives at the moment. So just some thoughts there will be very useful, if you can help us out. And then I was going to ask on gantenerumab. I've got a few questions in from clients in terms of when we see the future of the analysis, which was just showed with crenezumab, and where we should see the final data. And then the last question going about Andrew on the bispecifics. You have won the FAP-IL2, where, I guess, we should see some data in the first half of this year, maybe in Glasgow, and then the CAC3 program, since the -- [indiscernible] have given you work on how to dose patients, pre-dose them, and then move the dosing up on that asset. I think colorectal cancer was an original target for that bispecific. So if you can help us understand the position on those two assets, that will be fantastic.

Sorry, you're talking about the CD.

CAC3 and FAP-IL2 as well, where I think [indiscernible] has a sort of Q1 readout on that, so possibly.

Okay. That's quite a list. Actually, the first answer is there's a relatively straightforward answer. We are launching Herceptin SubQ in the U.S. It -- yes, it was a long journey because it wasn't clear the regulator was going to support it, and then we had to get it -- and then we had a priority on resources in EU because it was an ongoing launch. So basically, we were -- yes, we were delayed on time lines, but we are planning to bring it to market. I don't have an update on the timing, but -- and so there have been no decisions about prioritizing the fixed dose combination over SubQ Herceptin. We're going to bring SubQ Herceptin as fast as we can clear the regulatory process, and then we'll bring the fixed dose also as fast as we can bring it. On risdiplam versus gene therapy, this is a case where there's very little data on gene therapy and there's very little data on risdiplam. And so I'm not sure I can offer much beyond what any knowledgeable person in the field could say. I mean, gene therapy, you have some patients who have been exposed to the vector and will therefore be ineligible for the gene therapy. There will be questions about duration of effect that, again, we won't be able to answer for some time. And with risdiplam, we have a systemic agent, and so we're -- we think that we're likely to provide a benefit to all parts of the body affected, which has advantages over something that's administered, say, intrathecally. And it's really going to be about the efficacy data, and I think we're particularly interested in that efficacy data we're going to get on the newborns because that will really say what happens if you get in there from the very beginning. But again, based on what we've seen in type 1 and types 2 and 3, we're pretty encouraged. And in fact, I think, one of the things we're looking at is, we want to make sure that we've got a well-tolerated drug because, again, we don't know the effects of administering a splicing modifier over time in a developing child. So I think, for us, we're going to continue to be gathering a lot of safety data over the course of the year, but the results so far have been rather remarkable, and we look to have more of that.

I think what is well known with gene therapy is in general, and I think that would also apply here as a generic statement without having the data naturally because the medicines are so early in the development or in the clinics, is that, typically, patients develop antibodies against viruses as they -- as you grow up. And we know that for gene therapies, for an adult, typically, 30% would have an immunity against the common vectors, which are used with gene therapy. So what we know for sure is if you treat a patient with a gene therapy, who has a certain age, typically, 30% of the patients would not be eligible for a gene therapy. I say that because, as you know, our studies go not only for type 1, but also for type 2 and type 3. So what I would say is -- and it is a big part of the market potential. And I think it's likely that this is at least a complementary use next to gene therapies. Then the question is more how does it compare with the other molecule, which is already on the market. It's been an asset from Biogen. And here, the big difference is that [indiscernible] is provided into the central nervous system. So the effect is in the central nervous system, whereas, our molecule is given in the systemic way because it's a small molecule. So again, the data have to read out, but what we would hope is that this systemic effect would have advantages from a clinical efficacy point of view, especially over time. But again, that's a hypothesis at this stage. But clearly, you can see how the market might segment along these lines. And then, of course, the other, as Bill said, the big question, which nobody can answer, is how long does a gene therapy really work, and that is -- so whilst you have the advantage with babies who get a gene therapy because they are not yet immune against the vector compared to an adult, the problem with the baby is the turnover of the organs is very fast. And as you know, the gene therapies, they only work in the cells where they are given in, but when the cell divides, it's not in the cell, which is divided. So if the organ turns over, if the organ completely turns over, then, of course, you lose the effect of the gene therapy. And nobody has an experience. Now how many years does it take until you lose this effect simply because your organ has turned over. If you look at the liver, for example, there's literature out there which says that a diet turned their liver, which is typically an organ which absorbs viral vectors, turned over their liver, so everybody of us, three, four times in your life. Now how does this exactly play out in the case of SMA? What does this really mean, in duration, is a question which nobody can seriously answer because we simply do not have the data. But what we do believe is there is almost certainty that there is a place for molecules next to gene therapy. And what you could see is that either it's a separate segment because gene therapies simply don't work or because you could have a combination therapy over time. So if the effect of the gene therapy goes down, then you would add another medicine to complement the effect. So we'll see how it evolves. But it's really exciting. I mean, if you look at it from the families who are concerned, the parents, I mean, a couple of years, there was no option. I mean, the babies just die within half a year or a year. And soon, there will be three options out there for those families. I mean, what is success if you look at it from an industry and from a medical point of view. It's really fantastic to see how the field has moved on. Right. Now we still have two other questions, right

Yes. I'm adding here. So you asked about gantenerumab, when we'll have the data. I think we would hope to have our first meaningful data in 2021. And we're -- yes, I think that's probably -- that's -- internal plan is when we think we'll know something in 2021. FAP-IL2, we showed data at ESMO. We've got, I think, 3 other Phase Ibs that are ongoing. And we're going to see how that evolves as to when the next update would be and then the CAC3.

Okay. So basically, in the making. So open for recruitment, where patients are treated. And I have to ask the team, and they will give us an update.

Yes, and now we have a regimen that we feel -- that it works in terms of addressing the issues we had the first time around. But I think we're now -- it depends on how fast we recruit this -- we recruit the study. And so we remain optimistic that that's potentially a powerful mechanism in colorectal cancer, but it's one that -- it's not moving as fast as we thought because of the -- some of the nature of the molecule and the antidrug antibodies and how to overcome that. So I don't have a lot more clarity on that right now.

That's great. And just going back to the -- I apologize that I made a mistake in the subcut Herceptin file. It wasn't in the time lines. But just going back to the part of the question around sort of the Part B pilots and the biosimilar legislation that's coming in as [indiscernible]. There's some confusion about how a physician's incented or paid, whether it be LUCENTIS. To Andrew's point, whether [indiscernible] has subcut versus an infused Herceptin or an infused Herceptin biosimilar. It's not clear. Going forward, there will be a decentive to using subcut from the position of payers and the payment system.

Yes. Which Part B pilots are you thinking of?

I guess we're still unsure how, in the Part B pilot programs, how physicians will be paid. It's not going to be a potential link to an ASP-plus system, it may be linked to the complexity of the procedure and things like that.

Yes, yes, Well, we've been very involved with that, and I think, basically, there's a wide range of possibilities on the pricing front. I think it's a lot easier to say that physicians should be paid -- or shouldn't be paid an ASP-plus. It's a lot easier to say that than to do it. And the reason I say that is because -- think of it -- in fact, I've had this conversation relatively recently with the CMS administrator, and I think it was something where we were able to find agreement on that because, for example, if you have a doctor who's administering $100,000 therapy today and tomorrow, they're administering $1,000 therapy, you have to pay them differently because if something goes wrong with the $100,000 therapy, they've got a big problem. If the patient doesn't pay their co-pay, or if they break a vial or something, they have responsibility for that in the system. So there has to be a -- at least a part of the payment that is proportional to the value. And I think what reasonable people can agree on is we don't have a problem with ASP-plus-4, which is, on average, what doctors get from Medicare. We have a problem with ASP-plus-100 or ASP-plus-200, which is what a lot of hospitals are billing insurance companies. So -- well, there's been a lot of popular talk about getting rid of the ASP system. There's a reason that it exists, the mechanics of another system are very complicated. And my best guess is, is that ASP will remain and that the efforts are going to be how do you reign in the market power of the hospitals to put in these sort of outrageous markups.

Should we maybe take one question from the telephone?

Yes, please, yes. Why not.

On the second row, and then we go to the telephone, right, okay?

Yes.

Jack Scannell from UBS. 3 questions. One, I think probably for Bill, and then 2 broader ones, perhaps, for Severin and actually, one for Karl. Bill, for our technical question, you implied that triple-negative breast or maybe I misunderstood you to say, the triple-negative breast cancer will be filed and you're positive about the OS data. Now at ESMO, there was some discussion about the sort of technical glitch in the trial design in such that the PD positive group was a secondary endpoint, only to be tested if it worked in the ITT, and it was-not clear if it would work in the ITT despite the fact that any reasonable person could see it work in the PD-L1 positive. Has that issue been resolved, or is that still something we need to find out? The first question. The second question, again, Bill, I was struck by your slides. You said, well, revenues of pharma have grown at 7%. You said volumes have grown at 12%. What that says is the price mix globally has been a minus 5. Now the U.S. is half of that, right? Price mix in the U.S. is probably pretty positive because Part B drugs get a little bit more expensive every year, and you've launched OCREVUS and HEMLIBRA and Tecentriq and which are not inexpensive drugs, which suggest that, actually, the price mix inflation ex-U. S. has been very substantially negative. Now U.S. politicians concerned that drugs cost a very different amount in the U.S. and ex-U. S., and your results seem to be an example that this is widening. I just wonder if that -- how you guys think about that. And the third question, for Karl. The drug industry is discovering about twice as many drugs as it was five years ago, and arguably, there's about half as many sell-side analysts as there were five years ago. Do you think we're in an unusual position in systematic underestimation of drug industry pipelines by short-staff sell-side analysts?

Okay. While Karl thinks about the answer, let me hand over to Bill for triple-negative breast cancer first.

On the triple-negative breast cancer, that's right, so the study of Tecentriq plus nab-paclitaxel in triple-negative breast cancer was an all-comers study, where we also looked at the PD-L1 positive subset. The primary endpoint was in all-comers. And the -- it was a hierarchical analysis, where the question about OS in the PD-L1 positive population was downstream of at least 1 endpoint, where we didn't hit a P value less than 0.01. Nevertheless, the breast cancer experts and the regulators do see this as a meaningful result. So they were positive about us filing. And so we anticipate both that we will get approval based on that data, but also that they want to see the next study because we have another study with TMB -- or with Tecentriq and chemo in triple-negative breast cancer that we'll read out. I think it's at the end of this year. And so we think we'll -- they will be looking to see that that's also positive, but we should be able to proceed. The answer to your question about the global, it's actually really simple, it's all about China. Because in China, we had an average price reduction of 23% and we had an average sales increase of 28%. So the -- I mean, you have to get out your calculator to figure out exactly what happened on volume, but the volume did something like, yes, it went up by 70% or something. Somebody's done it. But anyway, so China had a disproportionate effect on that global equation.

And also, if I understand your question right, there's, of course, also the biosimilar effect because that is, of course, also in the volume picture. So as we -- I mean, there's a part of it where we lose share, but there's also a pricing part of it. And that, of course, is also only ex-U. S., and at some stage, of course, that will also occur in the U.S. So those 2 effects which explain the volume difference. Right. And then Karl. Come up here on stage.

No, I had hoped -- actually, I had hoped that I can skip that one, I really have to say. But no, I feel a lot of sympathy with you. I can't really say that because, I mean, we have a lot of budget restrictions, but I know that you have much more budget restrictions at the moment. And what we could see is that the number of sell-side analysts, I mean, you know it yourself, actually, is not increasing, to say the least. But I really mean it, I think you do a great job. I really have to say that. And while we have still the same -- a reduced number of people, it is true that the -- our plans in the industry, in general, actually went up, so the -- actually, the relative working we have to do on all the inquiries. And on top of it, it's not only about oncology anymore, you have to go into immunology, you have to go into CNS. So the whole diversity of network you have to cover actually also increased. It's not just the sheer number of assets, but also the sheer number of therapeutic areas you have to cover, and this is lots of work. So I think you really all do a great job. That's all I can say now. Thank you.

Karl, why don't we go to the telephone question? How does that -- how does this work?

The first question from the phone is from Steve Scala from Cowen. Please go ahead.

I have three, two for Bill and one for Alan. First, Bill, it appears as though the Humira -- or the HEMLIBRA EU non-inhibitor label will be more restrictive than the U.S. label based on what you said. Can you elaborate and compare the size of the eligible populations in both markets for HEMLIBRA? And secondly, I apologize for splitting hairs, but on LUCENTIS, in the prepared remarks, you said it will remain a very important franchise for years to come. But then in a response to a question, you said you were not expecting continued incremental gains. Are you implying flat to modestly down LUCENTIS sales over time? And then for Alan, at least one other major pharma will no longer include gains or losses from equity securities in the core results. I'm wondering what Roche's practice will be going forward.

Okay. Sure. On HEMLIBRA, I think this is mostly a difference between label language and actual use patterns. So for example, in the U.S., we don't have a restriction around severe, moderate or mild. However, the use is typically in severe patients because severe patients are the ones that have the lowest levels of factor VIII. They are the most subject to destructive bleeds. They're the most likely to be on prophylaxis. And so that's where -- the highest uptake will be in inhibitor patients. The next highest will be in severe. And then the relatively small part of the population that are either mild or moderate, there would be relatively little uptake of HEMLIBRA. In Europe, the regulators will just go ahead and say it's for severe. And we know about 70% of factor VIII use is in the severe population, and the remaining 30%, primarily, in on-demand use. So I think -- by the way, HEMLIBRA is still a product that's very young in its life cycle. I mean, imagine, we were only putting it in pivotal trials, what, two years ago, or I guess, maybe three years ago? And now it's been on the market for over a year. We will be exploring all kinds of questions about what is the use of HEMLIBRA in moderate and mild patients. Which kind of -- is it on-demand, is it -- there will be many of these questions to be worked out. But I think it's very much consistent -- the EU label is very much consistent with what we've been thinking about in terms of how the market shapes up. On LUCENTIS, maybe I confused things. But my comment was that LUCENTIS will be exposed to biosimilar competition in the coming years. It's hard to say exactly when. And we will be bringing forward this port delivery system in -- I don't know. What are we saying, 2021?

2021. Yeah.

Yes. So it kind of depends about -- I think with the port delivery system, there's a massive potential for benefiting patients and quite an attractive financial opportunity as well. And so then the question will be, will the biosimilars hit in, let's say, 1 year and then the next year, we're launching the port delivery system. So it could be that sales fluctuate, or sales could drop, depending on the timing of the loss of exclusivity and biosimilar launch relative to the launch of the port delivery system. But in any case, I think the -- when I said I think this is going to be a very important franchise to come because this port delivery system is a true innovation, and it will carry us well through a biosimilar launch. Does that make sense?

Yes, it does. Thank you.

Hi, Steve. I hope it's not too cold in Boston. Actually I think it is today.

Yes, I think so. Yes, I looked at that because when I got to New York, yes, it was minus 14 Celsius. So look, I think, first of all, let me say, I think we are pretty proud on our consistency that we show with the core results. I think that's really, perhaps, a difference maker in the industry because we have been pretty consistent or even a bit stubborn, what we've put into noncore and what we don't put into noncore. So I think when you get our results, I think you can always compare apples-to-apples. And I think that's something we put a lot of value in. And the other piece, I should say, is we gave it a little bit of thought, and I agree, I think when you now look at the fact that we have to account for unrealized gains here in equity securities for a list of companies, that brings a little bit more volatility into the core results, which is to a certain extent, a little bit unpleasant. But when I look clearly at what the [indiscernible] I think it's very well-connected to our core business. They work together with partnering. They work together with the research entities. And whenever we think, okay, there's something smaller, when we don't want to do the major step, we say, well, we take a small part and let's see, and we are a part of that and we are part of the whole story. And we could discuss Avexis, but that serves one element here that I would like to bring into the game. So we're engaged. We're part of that. I think it's really part of our core business, and therefore, I think it belongs to the core results, despite the fact that it could contribute a little bit more volatility to the results overall. But we gave it a little bit of thought, and we feel still well with it. Let's see what the future brings. But for the time being, I think it's the right setup.

Thank you.

Okay. Let's take other questions. Yes, I'm going a bit to the back. We have many questions still. Let's start at the very back in this corner. Right.

Marietta Miemietz, Primavenue. Just a follow-up question on HEMLIBRA, please. I'm just a little bit surprised that your spend now seems to be that we should really expect the biggest uptake in the patients who are actually already on prophylaxis and potentially switching from factor VIII. Because you had, in the past, shown us data suggesting that there's a lot of patients who are not on prophylaxis and the proportion is obviously higher the more you go down in terms of severity, but also, in the sort of moderate-to-severe patients in a large number of non-prophylaxed. And so I thought some of the thinking was that actually that segment is very much underserved because there's not really a good product out there for the people who don't desperately need to be on prophylaxis, and that could potentially be very, very big and also relatively a low-hanging fruit. So have you actually had any, like, real feedback from the market to the effect now that patients who are not prophylaxed are just basically completely recalcitrant to the whole idea of prophylaxis? Or are you just being cautious here, and that could potentially -- that, that non-prophylaxed segment could potentially be a blockbuster opportunity for you? Thank you.

Yes. I think where it gets a little complicated is because there's prophy and on-demand patients that are severe, and there's prophy and on-demand patients that are moderate. And then there's not a lot of prophy patients that are mild. But -- so you have this sort of mixing. And I think we see -- the biggest opportunities are in the prophy and on-demand patients in severe because they have the highest unmet need. The ones that are on prophy, they have the misery of prophy, and the ones that are on-demand, they're on-demand generally because they didn't want to be on prophy. So it's -- that's the clear opportunity with the highest unmet need with the most uncontrolled bleeds, joint damage, everything else. And that's why it's 70% of the factor market. And I think we've basically just said, we did a -- very much a faster market pivotal program. It's been very successful because we came incredibly rapidly from some small-scale Japanese first-in-human studies to very rapidly to the west with bigger studies. And we've achieved that. And now we've got Lifecycle program, where we're going to be figuring out what is the best way to serve the patients that are in those categories. So I think there still is an -- there's an opportunity in that moderate space, but we need to do our homework and figure out which group is that. Because you can imagine, by the way, in moderate, there's a range of 1 to 5. There's a big difference between a patient who has 5% of factor VIII and I. And so is there a -- it -- before, it didn't matter because there was no treatment options. But now we have to look at, okay, so which patients are we really targeting there?

And we take the question here in the middle. First, this one and then I'll come back to you, right? So please.

I'll try and be brief. It's Emmanuel Papadakis from Barclays. You alluded to a bit earlier [a falloff and buy some as] 340B as long as it's around, it's a big book of business for you in the U.S. Could you talk a bit about your base case pricing and the penetration assumptions within that channel relative to others and some of the commercial incentives that may affect that? Maybe a quick one for Michael. With that many years of sustained margin compression in diagnostics, does this represent a sustained inflection point? And are we any closer to seeing that elusive end-to-end NGS solution? And if so -- if not, when? And then maybe a quick one on balovaptan. Could you just talk about your levels of internal enthusiasm for forthcoming Phase II proof-of-concept data.

Okay. Michael, you want to go first?

Okay, yes. Emmanuel, thanks for the question. I mean, we discussed that already a couple months ago in Rotkreuz when we had our Investors Day. Well, I mean, you saw good numbers. And of course, we are working hard on keeping those good numbers. Can we call it already inflection point? I mean, we will -- we do our best to keep this good trajectory going forward. We are investing highly still, as I said at that time, in R&D, in innovation more than everybody else. And we are also doing our homework, I believe, in managing all the costs that allow us to get the flexibility to keep investing. Important is to maintain the sales trajectory. And this is important. And with the launch of all these new systems cobas pro coming now, the ALEX E8 or I that we launched a couple -- 16, 18 months ago. And the good pickup of those products -- and of course, at the end, this 20 billion test that we are performing and then -- and that being run on those analysts -- analyzed runs in the world. I mean, this will allow us for continuous growth in sales. And with the necessary measures that we are taking on all fronts internally, we should have a good trajectory going forward. And I hope we can, going forward, also show a more sustainable inflection point as I mentioned.

Yes. So 340B, is everyone familiar with the 340Bs

No.

Okay. All right. So 340B is the program in the U.S. which was initially designed to basically use pharmaceutical companies to subsidize hospitals that service a disproportionate share of low-income patients. And the idea was these are hospitals because they got a lot of nonpayment, that they have financial struggles. And so in a -- and it's genius, the American health system figured out the way -- a way to do that would be to allow these hospitals to buy drugs at a discounted price and then charge payers, including the government, the full price and then use the margin to subsidize their general fund. And what happened under the Affordable Care Act, aka Obamacare, was there was a massive expansion in the -- well, a significant expansion in the rate of discount, the level of discount they received and a loosening of the rules about what it took to qualify as one of these 340B institutions. And out of that, a whole cottage industry developed around helping health care institutions figure out how to become 340B and then also how to expand their footprint. So probably, the most bizarre example that I think of is there's a hospital that's a 340B hospital in the Los Angeles area that went around buying up rheumatology practices because the rheumatology practices were infusing drugs. And if the 340B hospital purchased the practice, they get to capture that business with the discounts. And so one of my favorites is Beverly Hills Rheumatology, that's actually a 340B center. And so yes, you can fill -- I'll let your imagination fill in the blanks. But -- so then the question comes, okay, with biosimilars, if these 340B centers are making lots of money on the branded drugs, are they going to want to sell biosimilars? And I'd say that's a valid question. And -- but I would also say a lot of policymakers are also asking that question. And so it wouldn't surprise me if that sort of loophole got fixed somehow. But this is -- I think this is -- you have to put this in the American health care blender a little bit because it's just one of -- I mean, you say, "How can this be?" and "What a strange thing." But there's a lot of things that are strange. And no -- and I say that we -- there's a -- the American health care system is a great health care system in terms of innovation, but it also -- so I don't know. Maybe the answer to your question is, I think that, that's an open question, and it's going get sorted out with this.

Okay, let's move to autism, right? That's my statement.

Yes, yes. Balovaptan, have you seen the Phase II data? Yes, that's right. But you saw that?

Yes I am looking.

Yes. Well, okay. So we think there's a large need in the pediatric space for a treatment that would help with development and potentially just help kids on the spectrum to have better functioning. And we think there's a need -- a large need with adults as well. We have basically Phase II data, proof-of-concept data in adults. And we are very hopeful about it. At the same time, I think we're quite judicious about it because it's Phase II data, it's a single study. And this is a -- it's a behavioral therapy. And we know that the annals of drug development are replete with things like that, that seemed very promising based on a study. And then when you go to repeat it, you don't repeat it. So we're going ahead full speed. We got a pediatric study, Phase III program in adults as well. And we're gaining further pediatric studies on the Phase II study. We're pressing on with the adult Phase III study, and I think it's a big maybe. Again, we hope, but we're being very realistic, I think.

Yes. I would also put it into the high-risk category. But the pediatric study will be important because if a second Phase II study shows the same signals, then of course, the confidence level will go up. Right. No, we were here, right.

Five more minutes.

Sam Fazeli from Bloomberg Intelligence. First one is -- they're all easy. Split with regards to product sales going forward. So thinking about PERJETA. How should we think about 2019 with regards to the adjuvant setting versus the -- prior to the AFFINITY data? And then the same for Tecentriq. You -- obviously, you've got new indications coming at some point. Have you had much use already off label? And obviously, you don't promote off label, but have you seen some? And how should we think about the existing indications that may be under pressure from competitors? Net price in the U.S., obviously you don't have biosimilars there yet. One of your peers spoke about oncology drugs seeing some net price declines in the U.S., which was a little bit of a surprise. Have you seen anything like that? Have you got any experience, any feelings with regards to that going forward? Biosimilars, have you become a little bit more -- or are you going to be less pessimistic when it comes to us modeling biosimilar models? I think Karl might know what I'm talking about in terms of the expectations in erosion rates going forward. And the very last thing is gantenerumab versus crenezumab. We had felt that there's potentially some reasons why one -- or the one that you -- that didn't work and you stopped should have worked. Is there -- can you just remind us how the trials are different? That might help us have more faith in the ongoing trials? Thanks.

Yeah. We have five minutes.

Okay, so I'll be really quick. Okay, PERJETA. So you said adjuvant versus what was previous? So PERJETA has been approved for metastatic and neoadjuvant. And then AFFINITY was the adjuvant. And I don't know if we have the precise breakout. But -- I mean, most of the growth is happening, obviously, in adjuvant because we were highly penetrated in metastatic and neoadjuvant. And I don't know if we can -- do we have it.

Perfect answer.

Okay. Tecentriq, off label, you asked whether we're seeing off-label uptake. The thing is, the data was just presented at ESMO and San Antonio. And we -- I think, as we said, Q4 sales results were strong, and -- but it's really too early to say what's driving that. We did get an NCCN guideline in small-cell lung cancer, which there could be some use happening there because there's a guideline. But we don't know what -- I mean, again, we have some anecdotes. I don't think we're seeing off-label use in breast cancer. I think that would -- there's not an NCCN listing and the data was just disclosed. Could it happen before approval? It could, but approval's going to be pretty soon. So I'm not sure there's really a long window for off label. Let's see. Oncology net price declines, no, we're not seeing that. Biosimilar, are we less pessimistic? No

Oh, yes, sorry. About what we've seen?

You mean on like what we experienced in Europe?

Yes, and what it means for '19.

Yes, that's fair. I mean, we saw in Europe, obviously, it was Rituxan that -- where we lost -- I think it was approximately half of the business in 1 year -- MabThera, sorry. And so we expect to lose less in 2019 because when you've lost half, you can't lose more. And with Herceptin, we lost -- was it 20-something percent by Q4? So I think we'll lose more because they're launching. And in the U.S…

Basically in Europe, we expect the impact to be rather similar to what you've seen in 2018.

Yes, the studies are similar, and we -- I think the 2 molecules, the MoAs are slightly different in terms of which species of a beta they're targeting. But also, the crenezumab was an IgG4 molecule, and it was designed to have less of an immune system participation in clearing the plaques. And so the -- I would say the reason for that was that the early anti-a beta molecules caused brain edema, which was dose limiting and treatment limiting. And so cren was designed to avoid edema. The question we always had was, will it work? Is that edema also a sign that they have activity? And that was always the key question. And we decided to -- because of the size of the unmet need and some data that looked somewhat promising that we would take it into Phase III. But I know that Sandra Horning, our Chief Medical Officer, and I had many, many conversations about is this going to work? Is it -- and it doesn't work. Or at least, it didn't in the population we studied, which I think was a well-designed study. Gantenerumab, very different. It -- we know it clears plaque, and so there, we're more testing the question of whether that clearing plaque will resolve Alzheimer's or at least make a dent in it. And so I think it is a pretty independent question. So I think cren and gante are pretty independent. I know the market reacted in terms of one of our competitors put -- probably put more of an association between those two than we believe, but we'll see.

Right. We have one more question, and I go back into this corner. I've gone this way, and now I still need to cover people over there. One more question.

Severin, you spent a large part of your annual report talking about artificial intelligence, big data. Just help us think realistically how you're actually using it across the 2 businesses today? Are you really seeing it improve productivity? If not, when do you actually think the investments you made in Flatiron, Foundation Medicine actually start paying up either from prospective diagnostics or improved property of successes, lower trial costs in Pharma?

Right. So we believe that the digitalization, which is happening in health care, as we speak, offers a great opportunity for research and development as well as for better targeted health care. I would say advanced analytics, artificial intelligence, et cetera, is probably the smaller issue. The bigger issues that you can collect the data that they are -- that there is electronic means to collect them. In big parts of Europe, data are not even available in electronic form. Then you need high-quality data. You get access to that. You have to aggregate it. So I think the challenge today is more on the data side than the advanced analytics. And that's where we focus on. We also need advanced analytics, but what we focus on is the data. And that is what Flatiron was about. This is what FMI was about. Now we try to bring these data sets together to have clinical genomic data sets rather than just clinical or just genomic data sets. And eventually, we try to enlarge this kind of data set with longitudinal clinical outcome data as well as clinical trial data where we have lots of them in-house. So that should provide us with new insights. That should help us to make clinical development more efficient, fast and eventually, also improved health care. But we use artificial intelligence actually across the value chain. I can perhaps just give you one little example. We are developing sensors in diagnostics. And with sensors, you have high wastage. If you look at flat sensors, for example, in industry, you lose 50% of your sensors and -- in the industry. And you don't know why you lose the sensors, right? There are small changes in this manufacturing process, which is a very complicated one, and you just lose the sensor. You know it's not working, but you don't know why. Now guess what? We took pictures of these sensors coming out of the production process, and we applied machine learning on the pictures, which came out of the sensors. And the machines could predict better which sensors are falling out than these experts could predict who have worked decades on improving the manufacturing process for these sensors. And we could increase the yield for those sensors. Now this is more an efficiency application. But I mention it because it shows that these kinds of things get relevant across the space. What you've seen with uPATH in pathology, for example -- now you have these pictures, and they're working here in the U.K. that a pathologist can change -- exchange the images to get the expertise and everything. But what we can also do is we can run algorithms on these digital slides if you like. And what we do see is the quality of the interpretation of the result goes up. I mean, you don't even want to know how often the results are wrong and how much. In this case also, clinical decision-making can be improved by this kind of tools. So there's huge potential here. I think -- I'm looking at Karl, we really have to close here.

Opportunity outside to get a drink and something to eat, so -- for people who want to stay.

I know many of you have to leave. For those who can stay, we have an apple cart out there. I thank you very much for your interest, and have a good week. Thank you very much.

Thank you.