[Operator Instructions] One last remark. If you would like to follow the presented slides on your end as well, please feel free to go to roche.com/investors to download the presentation. At this time, it's my pleasure to introduce you to Severin Schwan, CEO, Roche Group. Mr. Schwan, the stage is yours.

Thank you. And also a warm welcome from my side. Welcome to our Q3 Sales Briefing. Let's get right into it. Here we go. As you have seen, group sales are up by 2% at constant exchange rate. And I should say, in spite of the sharp decline we have seen with our COVID-related sales, both our COVID medicines, Ronapreve and Actemra on the Pharma side; as well as the decline on the Diagnostics side. And all of that was only possible because of the strong growth of the underlying business of our newer medicines and the strong performance of the base business in Diagnostics. And Bill and Thomas will comment on that in more detail in a moment. So if you look at the numbers again, stable business in Pharmaceuticals, 6% up in Diagnostics, which results in the overall growth on a group level. So really here, on this slide, you see the impact of COVID and the base effect versus last year. We had very strong COVID sales in Q3 last year, and you see here now the overall business coming down by 6% in the third quarter. Just to put it into perspective. In Pharma alone, we lost CHF 900 million versus previous year, just in the quarter, due to Ronapreve and Actemra. And likewise, on the Diagnostics side, even to a lesser extent, but still we've seen now declining sales, minus CHF 400 million versus the previous year. Now if we look forward into the fourth quarter, actually, we expect that dynamic to change. And the reason is that we -- I mean on the COVID side. And the reason for that is that we have an outstanding order for Ronapreve in Japan in excess of CHF 1 billion. And combined with the continued strong growth of the underlying business, that gave us the confidence also to confirm our outlook for the full year. If we turn to the next slide. You can, again, see here the strong growth of the Diagnostics base business as well as the underlying Pharma business. You see here in more detail again, the effect of COVID-related sales with Ronapreve and Actemra. And you also see that we had a negative impact from the entry of biosimilars in terms of Avastin, Herceptin and MabThera of roughly CHF 1.5 billion. The guidance we gave you at the beginning of the year was we expect a full year effect of about CHF 2.5 billion. From where we stand today, just for Avastin, Herceptin and MabThera, it will probably be a bit less, a bit above CHF 2 billion, for the full year. On the other hand, we have seen a more pronounced effect in the erosion of Lucentis and Esbriet, where we have seen generics coming in. So overall, actually, even though the composition is slightly different, the CHF 2.5 billion erosion still holds true. If you turn to the right-hand side of that slide, I think it's a pretty impressive structural change of our portfolio just over the last 5 years. And you can see 2 things here. One is Avastin, Herceptin and Rituxan has now become a relatively small part of our portfolio. It's much more diversified. We have entered new areas, like multiple sclerosis, hemophilia, et cetera. But what you also see is that in spite of the erosion, the massive erosion we have seen for those 3 top cancer medicines over the last 5 years, we could still grow the business, from roughly CHF 40 billion to CHF 47 billion. So that's an increase of 20% in a period where we had to digest the entry of biosimilars for those 3 medicines. On the next slide, you can again see the underlying business if we correct for the COVID-related sales. So you see Pharma in the third quarter at 2%. Also Diagnostics, robust sales with 7% versus the negative evolution, including COVID. Right. So if we look a bit into the future into our product portfolio, I think we can say that we have a new record of new molecular entities in our clinical stage portfolio. Overall, 85 new molecular entities. You see here 13 projects in Phase III and registration. Now you see the decline versus half year. That's entirely due to the fact that some of those medicines have been launched in the meantime, and there is a lot coming through over the next quarters. We still have important readouts in the fourth quarter. Of course, we're all eagerly awaiting the results for gantenerumab in Alzheimer. But beyond that, there is also important readouts for Vabysmo in an additional indication, RVO. And I should also mention Venclexta in multiple myeloma. Into '23, we have a whole range of important readouts, in particular in oncology. And I should highlight here the many readouts with Tecentriq in the adjuvant setting. Now coming back to Alzheimer. We will present results for gantenerumab end of November at CTAD. I just want to reemphasize, we do not yet have any results in-house, so we have to wait for the results. We expect them in November, and we should be able to communicate more by the end of November. What I should also say is that we got a breakthrough designation for our blood-based diagnostic test, which obviously will be of high relevance, not only for us, but also for the field overall and for potential other medicines in this field. Right. I think with this, I'd like to close. As you have seen, we confirm the outlook for the full year with stable to low single-digit sales growth, EPS to grow in the low to mid-single digits. And on that basis, we should also be able to increase our dividend in Swiss francs again. And with this, I'd like to hand over to Bill. Bill, over to you.

Thanks, Severin. Hi, everyone. Great to have the opportunity to share our latest results with you. It's been a very dynamic quarter. And yes, looking forward to more good things to come. So this is the geographic picture to start with. Overall in Pharma, we had stable sales year-over-year, year-to-date. And this really reflects the competing forces, very strong delivery from our newer medicines, but at the same time, offset by declines in AH&R; in other medicines affected by loss of exclusivity, like Esbriet and Lucentis; and then also the decrease in the COVID sales. And this happened pretty much the same in every area except for Japan, which actually had a small increase in COVID sales, and you can see the difference on the results. This is the product view. And again, you can see what we're really pleased with is that excellent performance on our newer medicines, OCREVUS, HEMLIBRA, Evrysdi, Phesgo, Vabysmo, Tecentriq, all of them with sales in -- right across the regions around the world, very strong performance. Down at the bottom of the page, on the other hand, you can see the impact of AH&R declines, but also Actemra and Ronapreve with big reductions in COVID use. And then you can see Esbriet and Lucentis with the impacts of biosimilars or generics. And so I think what's very encouraging to us is the fact that the growth drivers will stay, whereas the things that are dragging down sales are on their way out. And so this really portends well for future quarters. Looking to oncology. The overall oncology sales for -- year-to-date are down 1%, which is pretty remarkable given the drop in AH&R. And this is really accomplished by strong growth in the HER2 franchise, 4% growth with Perjeta, Kadcyla and Phesgo, offsetting the declines in Herceptin. Also strong performance you can see in Tecentriq. In the hematology franchise, we have several things to point out. Polivy, which is now just starting to get uptake in first-line setting in Europe, so 79% growth year-over-year. Now we have a PDUFA date in April for the U.S. for Polivy. So we look forward to more growth from Polivy. And then also Lunsumio, which is just launching in the EU, and we have a PDUFA date at the end of the year in the U.S. So we look forward to more gains from Lunsumio in the months ahead. And then finally, Alecensa with 16% growth. It's the leader and continuing to be the new patient share leader in all the major markets. So looking a little more closely at the HER2 franchise. So you can see Phesgo very strong uptake, now at a CHF 900 million annual run rate. This is the fixed-dose combination of Herceptin plus Perjeta. It's rapidly becoming the lead in many of the major markets. We also now have a Phase III study in first-line dual-positive MBC with Phesgo and giredestrant. So I think Phesgo is going to continue to be a very important medicine for many years to come. And then Kadcyla, which remains a leader in certain early cancer settings as well as first-line MBC. But we also have 2 Phase III studies, 1 in first-line MBC and 1 on adjuvant, with Kadcyla and Tecentriq. And so we're taking the power of an ADC combined with cancer immunotherapy. We think that, that as a proof of concept has already been demonstrated, and we're looking forward to bringing more advances to breast cancer patients with Kadcyla and Tecentriq. This slide have a couple of the graphs from studies we've shared this year at ESMO and ASCO, which really explains why we have confidence in giredestrant as a SERD. We think it's a best-in-class molecule, based on its drug properties, based on its affinity for the target. And you can see that comes through, for example, in the later line setting, in second and third line, in patients that have an ESR1 mutation, which is a subset of patients that are still estrogen-sensitive. And you can see in these patients, we beat standard of care with a hazard ratio of 0.6. And then at the other end of the spectrum, in the adjuvant setting, with the coopERA study in neoadjuvant. And there, we saw a strong impact in Ki-67, which is a biomarker for proliferation, and it's associated with improved long-term efficacy. And so we feel like we've got strong signals for why this molecule could be very important in the future. And the studies in adjuvant and first-line metastatic breast cancer are continuing to enroll. Turning to Tecentriq. So in Q3, we had 9% year-over-year growth. And this is basically Strong growth in the U.S., in Europe and international markets, somewhat offset by a small decline in Japan, which was a result of a price cut in Japan. We're mostly through the price cut dynamic, and we should see a return to growth for Tecentriq in Japan as well. And I'd also like to point out, we've got important adjuvant studies still to read out in Tecentriq. The uptake is strong and growing around the world in the non-small cell lung adjuvant setting, but we're not done. We have more to come and look forward to more growth on Tecentriq. Turning to hemophilia. HEMLIBRA, another really strong quarter, broad-based growth. You can see it's virtually linear. We now have 18,000 patients treated around the world. We've opted in now on the second generation bispecific, which is called NXT007, and we'll be developing that in partnership with Chugai now around the world. What we're looking forward to in the near term with HEMLIBRA is the expansion to include mild to moderate patients in Europe as well as the presentation of HAVEN 7, which is the infant study for babies. And we'll be presenting that at ASH in December. So hopefully, you'll get to see it there. Turning to immunology. Here, you can see the impact primarily of lower sales of Actemra due to less COVID, which we -- I think we're all happy that there's a lot less COVID in the world. And hopefully, that trend will persist. Also unfortunately, Esbriet now affected by generic competition, so down about 48% in Q3. But these are offset by strong performance by Xolair at 8%, and Xolair remains the market leader in asthma in biologics. And also, we have strong growth, continues, in urticaria. So turning to the MS franchise. So I think we're seeing quite a strong year for MS therapies. We've maintained our market share, for example, in the U.S. at 35% new-to-brand share, and that's consistent. But yet, you can see very strong sales volume. And I think, again, it looks like more growth in store for OCREVUS. I'll talk a little bit more about that. I wanted to mention the fenebrutinib Phase III program, where we have studies in RMS and PPMS. They're accruing well. And yes, we believe this is really going to be an important molecule for us in the future as well. So far, we've had studies in multiple indications, and it seems to be well tolerated and safe medicine. And look forward to getting those studies enrolled and getting results for people with MS. I want to say a little more about the OCREVUS franchise, a couple of the major things going on there. One is our OCREVUS subcutaneous program. So we'll still -- now we have IV 6-month dosing. And now we're going to be moving to subcu, but with 6-month dosing. And really, we have the opportunity to cut the infusion time down from hours to minutes. And we think this will be very attractive to patients, to be able to receive a therapy for disease as significant as MS, to get that in minutes a year in 2 convenient injections. And so we expect data in 2023, and we will file rapidly after that. This will allow that patients not only could get OCREVUS in a doctor's office, but potentially in their home as well. And then we're still studying OCREVUS in higher dose. We think there's a potential to have greater efficacy, even than the standard dose OCREVUS. And we have those patients on -- or those studies ongoing in RMS and PPMS. Turning now to spinal muscular atrophy. Another really excellent quarter for Evrysdi. And we have more than 7,000 patients now. We're the market leader in all the major markets. Growth is driven -- continues to be driven by both switches and naive patient starts. We also shared the Phase II JEWELFISH 2-year data at WMS. This is really exciting. It's the largest SMA study in patients who've been previously treated with something else. And so that includes patients who have been previously treated with Spinraza or with gene therapy. And we saw in both groups, a doubling of SMN protein, which is pretty exciting to see, including patients previously treated with gene therapy, that we could increase their SMN protein. I think this is really promising development for people with SMA and for Evrysdi. We look forward to label expansion in Europe, getting into the newborns, and more opportunities to grow and make sure we can treat patients in every country in the world. So turning on to ophthalmology. Another very exciting quarter for Vabysmo. The uptake in the U.S. has been very strong. October 1, we got our permanent J code, which is really excellent because this means sort of the last remaining reimbursement obstacle has been cleared. And so now we really have broad coverage for Vabysmo for virtually every type of patient in the U.S. We now have approval in Europe and we've launched in a number of countries in Europe, and the initial response from physicians has been very strong. And also Japan, you can see the bar there. Actually, Japan is the #2 country already and with very strong uptake of Vabysmo. Susvimo, I want to mention, unfortunately, we've had to do a voluntary recall due to a manufacturing issue. We noticed in our laboratory testing, sort of reliability testing, that in certain cases, the septum, which is the sort of the seal on the port delivery device that prevents the medicine from leaking out once it's been injected in, that, that seal could fail after repeated dosing. And so we decided -- because it didn't meet our performance standards, and we want to make sure that we have high reliability, we decided to voluntarily stop distribution of the port delivery system. So patients who've already had the implant continue to receive their refills. There's no problem with that. But we want to make some corrections to the manufacturing process that we hope can assure a much greater reliability. And we hope to have new port delivery systems available in the market as soon as possible. So that's an update on Susvimo. And as Severin mentioned, we have a number of important studies reading out in Q4 for Vabysmo in RVO and for DME and diabetic retinopathy for Susvimo. Now I did want to take a minute and just say a little bit more about the studies that we have for Vabysmo, the Phase IIIs in DME and AMD, because this is an area of intense interest. There's competition here. There's a lot going on. And so I thought it would be a good time to remind folks of the data we have and what we're able to bring to patients and physicians. First off, I think it's really noteworthy we have 2-year results from Vabysmo in both these disease settings. That's really important because this is a chronic disease, and physicians are really looking for that longer-term data to understand what happens over time. The DME study was really the first time we had these treat-and-extend principles applied in a Phase III setting. And that's been -- I think doctors really appreciate that because that very much approximates what they try to do in the practice setting. We -- using that protocol, we were able to show 78% of patients in year 2, getting dosing at Q12 or Q16, which is a big advance. And I think what's especially important. And when you look at the results, the efficacy results on the right, is that we did that at no sacrifice to patient efficacy. And you can see these lines are really super imposable, whether you look at the less frequent dosing or the more frequent dosing of faricimab, or the Q8 week aflibercept. And then I think also, what we hear a lot from physicians, they appreciate about faricimab that has not been seen before with any higher dose of anti-VEGF therapy alone, is a greater drying. And you could see that in terms of the statistically significant greater level of anatomical improvement on the graph on the bottom. We've not seen that with higher-dose Lucentis when we tried that in the past. And so this is certainly an important indicator, for physicians, of disease control. Looking at the AMD study -- or studies, you see a similar thing, which is that we did. Here, we did a treat and extend in year 2. In year 1, we evaluated all the patients. After a loading period, we evaluated them at week 20 and week 24 to see if they would be eligible for less frequent dosing. And we used, I would say, a very stringent criteria for allowing patients to get less frequent dosing. The whole point of this was to have the greatest opportunity for every patient to get maximum benefit. And I think why that's important, I'm going to demonstrate on the next slide. But what we did was we said there were 5 different criteria that patients were evaluated at, at week 20 and at week 24 and they had to pass all 5. If they failed on any 1 of the 5, then they were not eligible for less frequent dosing. And you can see this here illustrated on the bar along the top with the 5 criteria. And what's key is the or. So if you failed on any of these criteria, you were ineligible for less frequent dosing. And using this very stringent criteria, basically 22% of patients failed 1 or more of those and we're not eligible. 78% though we're able to get Q12 or Q16 week dosing. Now an alternate way to do that would be to say, "Oh. No, patients actually have to fail more than one criteria." And a competitor study recently had that approach, where you had to fail both a visual test and an anatomical test. And if you didn't fail both, then you could be eligible for less frequent dosing. Well, if we applied that criteria to the patient population in the TENAYA and LUCERNE studies, so in the faricimab studies. If we applied that criteria, those criteria instead of the stringent criteria we used, then only 4% of patients would fail and be ineligible for the less frequent dosing. And so I think the main point of this is to say, hey. You have to be very careful about making comparisons across these trials because there's a whole lot in the word and or the word or. And so you really have to look at the criteria that are used and make sure that's understood. Okay. Let me just close with an update on the key late-stage news flow. I think probably the most significant thing to mention here is a number of the adjuvant studies that we had planned in 2022 to read out on Tecentriq or Alecensa have shifted by events, either to the very end of the year or early 2023, in the case of our periadjuvant study in non-small cell lung cancer and our liver cancer studies, adjuvant studies for Tecentriq; or into mid or later 2023, as is the case for some of the other adjuvant studies. Other than that, we're pleased that we've been able to bring a number of important developments to patients around the world. And I look forward to answering other questions you might have in a bit. And with that, I'll turn it over to Thomas.

Thank you very much, Bill, and good morning, good afternoon, everyone. With sales of more than CHF 13.8 billion, we had a good growth of plus 6% year-to-date. And now this growth is driven by the base business across all of our customer areas, and I'll go into more detail on the next slide. Now the COVID-19 testing sales were at CHF 3.7 billion, and they were no longer a driver of growth in these numbers. Now let's take a deeper look. And first, Core Lab. Now Core Lab grew at 5% year-to-date. This is the customer area that was mostly impacted by the China lockdowns in Q2, but have -- this customer area has recovered extremely well in the third quarter. Now if we take the third quarter numbers, Core Lab grew 7%. And if you take out the COVID-related sales, which are the antibody sales, actually, Core Lab grew at 9%, which is very strong. And this is despite also a decline in custom biotech. Now Point of Care is growing still very strongly with 30%. This is driven by rapid antigen tests, specifically in the U.S. and also in APAC. But even the underlying business is doing really well with 8% growth. Molecular, minus 8%, entirely driven by less testing versus previous year. And if you take the underlying business here, the base business is growing 11%. Diabetes Care is showing minus 3%. As I mentioned in the last calls, always, we did have a onetime effect in Q1 2021, which was a resolution of a dispute over a rebate. If we take that out, we're actually flat. And you see Pathology growing very strongly at a double-digit range. Now taking a closer look at the different regions. As I mentioned, North America and the APAC. Here, we have still a very good growth in COVID-19-related sales and also in the base business. But also here in these 2 regions, in Q3, we did see a reduction in COVID-19 testing. EMEA and Latin America were heavily impacted by less COVID-19 testing. Now if you just look at the base business in EMEA, it was growing 5%. And you look at Latin America, Latin America is actually growing at 15% in the base business, so doing extremely well. Now let's take a closer look at the split between the base business and the overall sales. As I mentioned, we had, in total, CHF 3.7 billion in COVID-related sales over this year. But if you look at Q3 and you compare it to Q3 in the previous year, with a decline of CHF 400 million. And so if you look at the underlying base business, we're actually growing more than 7%, which is really a fantastic result. Specifically also looking at Q2, where we were at 3%, which was obviously impacted very much by the China lockdowns. But even in Q3, we do have still sporadic lockdowns in China to a little more limited extent, but it's not completely back to normal. Now let's take a look at some of the recent launches. First, let me highlight that we've now delivered more than 1.8 billion COVID-19 tests to the world in more than 20 solutions. We've always acted very responsibly in terms of pricing to enable access. And we've placed more than 2,000 6800/8800s. And in the meantime, with the launch of 5800 end of last year outside of the U.S., we're still waiting for the U.S. approval, we've had more than 250 placements there as well. Now we launched an update to the rapid antigen tests, both for professional use and home use, with a new antibody which increases sensitivity even more. We've also launched a new antibody test with improved protein to detect or pull down the antibody. And we've also launched the interferon gamma release assay SARS-CoV-2 test, which basically measures the T cell response and T cell memory against this virus by measuring interferon gamma release or cytokine release after treating the sample with different antigens. So really looking at whether or not the T cell can recognize these antigens. We're still waiting. We've submitted to the FDA the 510(k) approval for the SARS-CoV-2 test on the 6800/8800s. So we would be one of the first companies to deliver its basically full -- with full regulatory approval to the U.S. Now let's take a look at Digital LightCycler. The Digital LightCycler is the next generation in digital PCR system. It's based on plates, and you basically have 3 different plates, depending on the application that you want to use. So it's not using emulsion as some of the other players are. And with that, we have a superior performance of the system versus other systems on the market. There are a number of different applications for the Digital LightCycler. One is around cancer treatment monitoring because you can have extremely high sensitivity. It's also looking at transplant rejection monitoring. And you can also monitor infectious disease spread in the environment, example: Wastewater, using this kind of system. We're also very excited to launch PRAME, which is an immunohistochemistry assay. And this assay helps to differentiate between benign and malignant -- or basically melanoma. There are 300,000 new cases every year and 60,000 deaths caused by melanoma cancer. Now if you look at a potential melanoma, sometimes it's very clear. But at times, you will really need to use a test to really understand, is it benign or is it malignant? And actually, if you detect it early before it becomes regional or distant, if it's still localized, give a survival rate of more than 99%. And with that, we extend our dermatology portfolio, which now includes more than 50 different biomarkers. As Severin mentioned, we did receive FDA breakthrough device designation for amyloid plasma panel. And what this plasma panel does, it actually reduces the amount of patients that would have to go for confirmatory testing. So if you look at it today and people would have to do PET scans, there are not enough PET machines to actually make sure that people get the right diagnosis, and with that, the right treatment. So if you can reduce the amount of people that potentially have Alzheimer's, you can also make sure that people have better access to medicines. So in the first step, you would take this plasma panel. With that, you reduce very much the people that are suspected of Alzheimer's but don't have Alzheimer's. And then you go for confirmation testing. You can do that through our CSF assays, but you can also do that through a PET scan. And with that, then you can then identify those people that up for therapy. And we believe this is a very good sequence and it's going to be very helpful to give people the opportunity to get, then, access to the medicines. Now finally, let me say I'm very excited with the launches that we've had already this year. We are on track to launch the rest in Q4. And also, we have a number of really exciting launches ahead of us in 2023. And with that, I hand over to Alan to take us through the financials.

Yes. Thanks, Thomas. And I will be very brief today because we have a sales call, as you know. I think really on the highlight side, really nothing to mention here. When you look really at the portfolio rejuvenation, which is really ongoing, I think Severin went really through it in detail. I think really you see in the Pharma underlying business. Let me call out here that Esbriet really showed a reduction of roughly minus CHF 200 million, and Lucentis with roughly minus CHF 250 million. And then showing really that underlying growth, I think, is really a nice achievement. When we go really through the regional sales development, I think really when you look at the Pharma Division in total, pretty flattish. You see the reduction in the U.S. and in Europe, which is predominantly Actemra-and Ronapreve-driven. Whereby in Chugai, really, the growth here is very much driven by Ronapreve itself. And you see the great momentum in the Dia Division, which evidently comes from the underlying business as well as from the COVID sales. So up 2% in constant rates. You see then the currency impact, which is solely 1 percentage point, which brings us to group growth in sales in Swiss francs of plus 1%. So let me talk quickly about the currency effect which is outlined on the next slide. And what you see on the left-hand side is really the growth of sales in concentrates, rounded, 2%. Here more precise, was a plus 1.7%. And you see on the right-hand side, the growth in Swiss francs was plus 0.8%, rounded, 1%. So that's the difference of 1 percentage points that I'm talking about. And basically, what you're seeing is based on the weakening euro and the strengthening of the U.S. dollar against the Swiss franc. Yes. With that, I get to the expected currency impact of 2022. And the nice thing is, it's so -- it's much more robust certainly after Q3 than it is at the beginning of the year. So I think a high likelihood that we really will see that. Nevertheless, there will be certainly volatility in Q4. And you know that this projection that you see on the right-hand side is based on the assumption that the currency rates at September 30 remains stable at the end of the year, which they don't at the very end. But nevertheless -- but what you're seeing is really a sales impact of minus 1 percentage point. You see a core operating profit impact of minus 2 percentage points. And on core EPS, same impact of minus 2 percentage points. So a relatively robust picture in a pretty volatile environment. And let me close with the guidance. I think we confirmed the guidance, which I think in the current environment, is quite an achievement. Really -- I think really also, when you look at the dividend here, let me clarify the points around what do we expect from the biosimilar impact, and it was mentioned by Severin already. We think that we will have a biosimilar impact for AH&R of roughly CHF 2 billion, which is a little bit better compared to what we started the year with. But nevertheless, I think it will be offset by losses on Esbriet and Lucentis. And that's why I also called it out. So I think at least, hopefully, if you like, recognizing we end up, when you put everything together, with the roughly CHF 2.5 billion we started the year with. And then furthermore on the COVID sales. Let me say I think we projected for the year, is that we will come from roughly CHF 7 billion in 2021 to roughly CHF 5 billion in 2022. And now it looks like more like a roughly CHF 6 billion. So let's see what Q4 brings. But I think really when you take everything together, we confirm the guidance. Good. And with that, I think we're looking forward to your questions.

Thanks, Alan. So then we will go now into the Q&A session. I think we have 50 minutes left, which is good. We have 15 people currently in the queue. [Operator Instructions] And the first question will come from Peter Welford, Jefferies. Peter, please.

Can you hear me now? All right?

Peter, we hear you, yes.

Yes. Sorry. I gave 2 questions. I'm afraid I'm going to start with, the first one, on gantenerumab, which I hope that you've already anticipate. So thanks for confirming that you haven't got any data in-house at the moment. But I wonder if we could just hear perhaps from Severin and Bill regarding how your confidence has changed following the lecanemab readout. And importantly as well, how you potentially think about the 2 studies at this point in time. I think you've mentioned in the past that Roche would really like to see 2 positive trials to move forward. Has your view on this, or indeed, a decision with regulators changed at all? And if you could possibly give us an update on your thinking with regards to any prespecified pooled analysis that there may be. And then just secondly, I wondered if on Vabysmo, could you just give us an update on the U.S. rollout there? I think 165,000 vials was slated for the U.K. Could you just update us perhaps on what you're seeing in terms of the latest number there? Which type of patients you're usually seeing this used in? And if possible, whether or not you are seeing at the moment any sort of impact at all from the Lucentis biosimilar on adoption.

All right. Bill?

Yes. Thanks, Peter. Let's see. So first on gantenerumab and the effect of the positive readout from lecanemab on our thinking. I think it's pretty straightforward. I think before that readout, there was quite a raging debate as to whether targeting A-Beta does anything whatsoever to clinical signs and symptoms of Alzheimer's. And I think after that data, I think that debate surely has to take on a different character and certainly reassures us in the validity of the pathway. Obviously, gantenerumab is a different molecule. There's different affinities for different species of A Beta and oligomers and fibrils and all that. It's a different dosing, different regimen, subcu versus IV. So there's all kinds of differences. That being said, we've worked hard to deliver a regimen that we thought would clear plaque and hopefully deliver a clinical benefit, and we look forward to seeing the outcome of that very soon. We have 2 large well-controlled studies. Each of them, they're powered to independently show a 20% effect. There is a pooled analysis planned as well. And I think beyond that, we just have to wait for the data. And we'll know it when we see it. We hope to have positive -- 2 positive studies, and we'll go from there. Vabysmo. Let's see. Yes, I think we've really provided the latest numbers. I think the run rate is up over CHF 750 million worldwide as of September. And we're seeing both switches from patients who've been on another anti-VEGF. We're also seeing naive patients being put on Vabysmo, which makes sense. We think we'll see more of that with the permanent J code in the U.S. And I think with Lucentis, it's been a combination of factors, including some contracting effects, some order pattern effects, some switching to Vabysmo and then the entry of the biosimilars which has happened more recently. So it's a combination of those facts. Thanks for the question, Peter.

Severin, do you want to add maybe on what you think about gantenerumab?

No, I think we're all very aligned. Again, we don't have the data in-house yet. So we have to wait for that. We are eagerly waiting the data. We look forward to communicating those data with all of you end of November. So let's keep fingers crossed.

Okay. Very good. Next question will come from Andrew Baum, Citi. Andrew, please.

Just picking up on that previous question. Severin, you stated very clearly that, prior to the lecanemab data, you required 2 positive trials. Paulo, when we saw him the morning actually of lecanemab release, stated that one positive trial and directional trend, you'd file. And especially if there's a prespecified planned pooled analysis, as you outlined. My assumption was that, that was always the case, that as a sponsor, if you had that data set, given the unmet medical need, you would always proceed. And your commentary about having 2 positive trials was more expectation management than there's been a sudden shift in the filing strategy between then and now, given lecanemab data, but correct me if I'm wrong. And then the second question related to the impact of the FDA from the lecanemab data. They obviously took a lot of criticism for approving Aduhelm on A-Beta lowering as a surrogate endpoint. Now lecanemab data seemingly validated that decision and more broadly accelerated endpoints. What does that mean for you, as being a major participant in CNS drug development, and your willingness to pursue targets as well as companies using accelerated approval on surrogate endpoints as a way of bringing them to market? Including things like neurofilament light chain, and there are a bunch of other experimental surrogates. Does this open up and give some incentive for you to be more aggressive in the CNS space and the industry more broadly?

Interesting. Thank you, Andrew, for the questions. Whilst I comment on the first one, perhaps, Bill, you can reflect on the second one, on how we deal with accelerated approvals and surrogate biomarkers. Yes, I mean, you are right. When I spoke about 2 positive trials, then essentially, what I meant is you can't have a costly negative trial and a positive trial and then expect the regulators to approve. But what you do in fact is you pool the data, right? And on the pooled data, you have a holistic look at the entirety of the data, and that is the basis for the filing. So there is no change to the principles of how we file data, not only for potentially gantenerumab, but for any new medicines. So you referred to Paulo Fontoura, so I can confirm what he stated in that respective meeting. I hope that clarifies a bit the way forward on the filing side. Bill, what's your perspective on accelerated approvals in CNS?

Yes. I mean, it's a good question, Andrew. I guess I've been working for 25 years on development of disease -- medicines for neuro diseases and I don't think there's this huge change, okay? I think there's a risk of people over-interpreting the FDA's actions. They were faced with a situation where they had a drug with a positive Phase III study and a negative Phase III study and some fairly compelling pharmacodynamic data. And in the face of that and their desire to provide alternatives with -- people with a serious disease, they found their way to an accelerated approval pathway based on that pharmacodynamic biomarker with confirmation by clinical data. I've not seen any inclination from the neuro division or any other division at FDA that there's some new emphasis on nonclinical end points for -- as bases for approval. So I don't think this has fundamentally changed our view. And frankly, even if it changed our view about the regulator's bar, I think the lesson of Aduhelm is probably that it's not enough just to win over the regulators. If you have data where the physicians and the payers are not confident in it, then they're not going to use it. And so Roche is not in the business of getting things approved, we're in the business of making great medicines that all stakeholders could agree about their impact and their benefit. And so yes, frankly, I don't think it changes a thing for us.

Thanks, Bill. Next question would come from Tim Anderson, Wolfe Research. Tim, please.

A couple of questions on Lucentis in the U.S., and it's just as I'm trying to think about biosimilar erosion in the future with this category. So my questions, two of them. How much of the decline in Q3 Lucentis was due to Vabysmo cannibalization versus the impact of biosimilar Lucentis? And more generally, how closely is the ophtha office space in U.S. managed by payers? Can they kind of force use of the biosimilar ahead of brands? I know it's a buy-and-bill market, at least in oncology. That usually means it's not heavily managed, it's more of the physicians. Is that the same in ophthalmology?

Yes. It's a good question. And we don't have a lot of detailed data on this. This is obviously very new developments. And we don't have data on the biosimilar uptake, what their volumes are. So we can't really square the circle. I would say, don't forget that the biosimilar companies can offer discounts as well. And so they can offer financial incentives to physicians. So it's not just a matter of whether payers can force certain things to happen, but the biosimilar manufacturers, certainly in other fields, have given financial incentives to physicians to switch. So I wouldn't ignore that. And honestly, I can't give you a precise breakdown. I think, as I said, there's a combination of factors, including the discount programs and such as well as erosion by biosimilars and cannibalization by Vabysmo.

Okay. Next question would come from Matthew Weston. Matthew, please?

Two questions, please. The first is back to gantenerumab, I'm afraid. You may have been aware that there was some discussion that there will be a period when you know one study results, but you don't press release because you said you're going to wait for both. That seems relatively unlikely to us, and we would assume that you'd actually just unblind both studies simultaneously. So can you please clarify how you're going to deal with the 2 data sets? And if there will ever be a period when essentially Roche -- a meaningful period when Roche knows something before they issue a press release to the market? And then the second question is around COVID orders going into 2023. Severin, you laid out in your introductory comments that you're expecting a significant government order at the end of the year. I assume other discussions with suppliers and governments are also ongoing. Anything that you can tell us to try and shape how next year looks? And the magnitude of the drop off, I think, would be very helpful.

Okay. So first on gantenerumab, I mean, my understanding is that this will all come in basically simultaneously. So there will not be a period where we kind of sit on more information than we would communicate to the outside. On the second question, just to clarify again. So as far as COVID is concerned for the fourth quarter, we expect, in our planning -- I mean, we never know how the virus will develop and whether there is suddenly demand. But for the time being, we don't see any demand, and we expect that the demand will remain low. But we do have a pending order from Japan for Ronapreve, and that's what I referred to, which will reverse the trend which we have seen in Q3. But this is really a one-off. So our planning assumption for next year is that it remains on a lower level, that we move into a kind of endemic state. And we'll still have some residual sales, both in Diagnostics and in Pharma, but at a much, much lower level, right? I think that was your 2 questions. Thank you.

Next questions would come from Michael Leuchten, UBS. Michael?

Two questions, please. Just going back to Vabysmo, just wondering if you could give us an idea of sort of business for the product. And was the revenue number in the third quarter in-market sales? Or is there any inventory fill in that number? And then, Bill, I just wondered if you could help me understand how you think about Tecentriq subcu. It's something that keeps coming up, a little bit more often now in discussions. Like how do I think about this in terms of addressable market? Where would it sit in terms of positioning within the disease settings?

Bill.

Yes. Thanks, Michael. So let's see, Vabysmo, let's see. The latest data I have is primarily switches, so about 85% to 95% of the business is switches from other products, of which, approximately 70% is aflibercept and 15% to 20% Lucentis and the remaining being naive starts. Let's see. What was -- sorry, Michael, what was your other question?

Oh. Is it all in-market revenue? So is there inventory fill in the third quarter?

Right. Right. No. I mean, there was a little bit of channel fill back in February. But since that time, we think we're largely seeing the market revenue. We don't think there's any meaningful inventory build in the channel. Tecentriq subcu. So the way we designed that program, while it's in lung cancer, we designed it so that it would basically show PK results and PK comparability to the IV formulation, which is what we did see. There's another study, which is a patient preference study, which just supports the idea that patients would prefer a subcu to an IV. That data will be in, in Q1. We'll be taking all the data to regulators. And our expectation would be that this data would be implementable across multiple tumor types, that it wouldn't be limited to lung cancer. Hopefully, that makes sense.

Okay. Very good. Next question would come from Luisa Hector from Berenberg. Luisa?

I just wanted to check on a couple more items as we look out to 2023. Actemra both trying to -- just to understand. Are we sort of back to a routine level? Or do you think there's some kind of COVID stocking that's being unwound into more of the routine use? And any thoughts on biosimilar entry next year? Also for Kadcyla, obviously, we have some competitor dynamics here, but your growth in the early stage setting. So how are you thinking about the outlook into 2023? And I wonder with Diagnostics, Severin, I hear your comments on the endemic levels. Could we say that Q4 for Diagnostics is more a kind of endemic level?

Right. Thank you for your questions. Perhaps, Thomas, if you could just directly comment on the Diagnostics questions. And then Bill, follow up on Kadcyla and Actemra.

Yes. Sure. I mean, what we see at the moment is that the decline that we've seen in Q3, now if we would compare Q4 start to Q3, that we see a further decline, right? So we're moving more and more into this endemic stage. But we're still only in the middle of October, right? And so we have to see what happens in November, December. But at this point in time, I mean, that's the view that we have because we don't have many government orders at this point in time. Now if you go into next year, we've always said that an endemic level of COVID testing could be in the low to mid single-digit billion area. And I think that's what we'll probably see as a market, right? Not as Roche, as a market, for COVID testing in next year.

And then the question is what is our potential market share in such a market? What would you say, Thomas?

So we always said around 20% of that would probably come to us.

Right. So as a result, of course, you can immediately calculate that COVID testing would be expected to be on a much lower level next year than we had still this year.

There were a couple of questions on Actemra.

Right. Please.

Yes. So let's see. You asked about whether we have now sort of a routine level of Actemra use. I think there's still some Actemra being used for COVID in the hospitalized setting. It's approved in most major markets in the world. We're guessing this year, it's maybe CHF 300 million or CHF 400 million worth of COVID sales, and then the rest being sort of the baseline. And next year, it's a little bit anyone's guess. We think there may still be some COVID use to the extent there's still people being hospitalized with COVID. But we would expect that year-over-year, 2023 would be rather similar to 2022, maybe down a bit. But we'll see. And in terms of biosimilar impact, it's not clear that there would be much meaningful biosimilar impact in 2023. But that will depend on when various approvals happen, but that's sort of our outlook right now. And then Kadcyla. As I mentioned, we have double-digit growth year-to-date on Kadcyla. We would expect to continue to see some growth next year. The fact that there's competition in later lines, notwithstanding, we still are penetrating in the early indications and in many markets in the world. Kadcyla is growing rapidly in China. We have a potential to be an NRDL in China with Kadcyla, so as examples. So we think -- we're hopeful that we'll have continued growth in Kadcyla still for some time.

Very good.

Next question would come from Richard Parkes from Exane. Richard?

Yes, a couple of questions. First, just moving back to gantenerumab. There's obviously been a lot of discussion about how the CLARITY AD trials set quite a high bar in terms of efficacy, and in terms of safety, in terms of ARIA incidents. Obviously, we'll have to wait for the efficacy data. But I assume that, usually, companies get access to blinded safety data across ongoing trials. So I'm just wondering whether that's giving you kind of insight into the rates of ARIA across the study, and whether that impacts your confidence that the dose titration schedule has been successful in keeping overall rates low. It seems like you're quite optimistic that gantenerumab could have a competitive profile. So I'm assuming that's the case. Whether you could confirm that, that would be great. Then secondly on Vabysmo. Just wondered if you could discuss the impact of the J code, permanent J code awards. Is that just an incremental smoothing of the reimbursement access? Or does that help to unlock a population of patients that physicians have previously had challenges in prescribing the drug to?

Bill.

Yes. Thanks, Richard. So yes, we think that efficacy and safety, the bars are going to be high for both of those in Alzheimer's, and especially safety because the nature of the disease, the fact that people can live with this disease for many years, and you can't have a medicine that's putting patients in -- at harm. And so that's why we designed the titration schedule that we did for gantenerumab. We don't think there's a lot of mystery around that because we had hundreds of patients with that regimen in the open-label setting before we commenced the Phase III program. So we don't expect anything out of that. I mean, you asked about blinded data. I mean there's an IDMC that's been monitoring it closely. I mean, we think, if there was something surprising or significant development, then we might have heard about that. And so we're looking forward to seeing the data shortly. With respect to Vabysmo and J code impact. I think there's probably 2 ways to think about it. One is that, for particularly smaller practices that have maybe less capacity or capability on the insurance and billing, that those practices would tend to wait more for using a new drug until there's a permanent J code, that the system is sort of automatized and it kind of takes the mystery out of it. And so that we could expect to see some increase in breadth based on the permanent J code. And then also in those practices who were already using Vabysmo, certainly, they would have started with the patients that they perceived had the highest risk, the highest unmet need, which would have been patients who have inadequate disease control on existing therapies. And so with the advent of the permanent J code, we would hope to see an increase in naive patients being treated with Vabysmo. And it's only been 2 weeks since we got it, so it's hard to say whether that's happened yet. But certainly, that should play out over the next few months.

Very good. Next question would come from Sachin Jain from Bank of America. Sachin?

I have some of my questions -- apologies, one back on gant, if I may. So you've been clear on filing strategy, but I just wanted to get your commercial perspectives. Specifically, how important is the 27% benchmark from lecanemab important to you? You've mentioned your studies are powered at 20% minimal. So for example, if both studies hit at 20%, do you believe you have a competitive asset? And if so, why? And then if you could just clarify where you sit on clinical relevance. I think before you, Paulo, have mentioned 25% as your clinically relevant threshold versus studies powered at 20%, so I just wanted to clarify that. And then second question very quickly. The next-gen bispecific in heme, I think it was NXT007. If you could just provide target profile related to HEMLIBRA and your time lines there.

Right. Bill?

Yes. Thanks, Sachin. Let's see. So gante, the commercial outlook, I mean, it's all going to be about the data, but reminding that also the mode of administration definitely matters. And you can look at many other disease areas, things like rheumatoid arthritis, where there were IVs and then subcus, and how that dynamic played out. Yes, many other areas. And I think it's very worth noting that the people who treat people with Alzheimer's disease are not people who are giving IV infusions. These are -- they're usually diagnosed by a neurologist, but they're often under the care of a primary care physician, doctor of internal medicine, people who are not giving IV infusions. And when we talk about the numbers of patients involved being in the millions, the idea of millions of old people needing to get IV infusions every 2 weeks or something like that, that's not an easy thing. And so that's one of the reasons we put a lot of energy in gantenerumab in developing a subcu formulation. So I think that, that's going to matter. Efficacy. People care about efficacy because that's why they're taking a medicine. So we are intent on delivering strong efficacy, but as I mentioned before, a lot of this is going to depend on what time frame we're talking about. So for example, the lecanemab study had 18-month end points. We have a 27-month endpoint. We will be looking at the effect of gantenerumab over time. What happens during the initial titration period? What happens in subsequent time? And I think all these things are going to play into what is the ultimate profile of the medicine. And we'll have answers very soon, I hope. In terms of NXT007, the next-generation bispecific antibody in hemophilia. The target profile, I can't give you the specifics, but it needs to be better than HEMLIBRA, which is no small feat, which is one of the reasons why -- yes, we think the bar is very high. That's been true for gene therapies, that's certainly going to be true for our next-generation medicine. With HEMLIBRA, we have 80% to 90% of patients with 0 bleeds and most patients getting dosing every 2 weeks or even every 4 weeks. And so that's a very strong profile. But I think some of the things we'll be looking at are things like what are the peak and trough levels of protection and looking at dosing intervals and other things. But more to come on that, Sachin.

Sachin, I hope we were able to answer your questions.

Yes.

Okay. Next question would come from Wimal Kapadia from Bernstein. Wimal, please.

Great. Just starting with Vabysmo. So it's quite interesting that, as time progresses, the number of people on every 16 week dosing increases in DME, whilst in -- for Eylea high-dose, the number of patients moving off 16-week dosing increases. So I'm just curious, I know you've seen the 2-year data for Vabysmo, but how do you think that would continue to trend for Vabysmo beyond the 2-year period? And do you think there's a peak number here, really, where we reach a steady state for number of people on every 16 week dosing for Vabysmo? And then my second question is just sticking with ophthalmology with Susvimo. I appreciate the color earlier. But could you just maybe give a bit more context? What's your level of conviction in fixing the septum issue? And over what time frame should we really be looking at? And then just tied to that, is this not a major problem, given the caution on the device and the surgical procedure to begin with?

Yes. Thanks, Wimal. Let's see -- I'm just taking notes here. Yes. So as to the trend on the Vabysmo and the longer intervals, I mean, we were certainly encouraged to see that over time. It's hard to say. Does it have to do with that, once you get disease control, that you're able to extend further? Very difficult to speculate. At the end, we have to go on the data. I mean, fortunately, because the treat and extend principles are applicable not only in clinical trials, but certainly in the real world setting, doctors will be able to observe this in their patients. And basically, that same protocol which gives the patients the best opportunity for a maximum benefit, which is to say that they're achieving both control anatomically, but also they're maintaining their vision, that they can put that into practice out beyond year 2 for years and years and have a level of confidence that they're doing their best for their patients. So let's see how that trends over time. In terms of Susvimo. I mean, we definitely take this very seriously. I mean, our intent is to provide medicines that are both highly efficacious and safe, but also reliable over time. And so we don't like that we saw a level of failures in these laboratory tests that exceeded what our projections were. But we're -- we've already been working at this. And we have a better understanding now of what we think is happening, and we're making modifications and we're seeing impact in terms of our lab experiments. And so we hope to be able to resolve this in some matter of months. It's not going to be days, but we hope it's months. And that we would be back with a device in the market within a year or so. But that's just our best estimate now.

Does that answer your question?

Perfect.

Very good. Next question will come from Emmanuel Papadakis from Deutsche Bank. Emmanuel?

Maybe I'll take one on Tecentriq, please. You did kindly highlight the headwind on pricing in Japan. But the question is really on the outlook in the U.S. Q3 slowed, I think, to just 3% CER growth. So perhaps you could talk a little bit about the growth potential in that region over the next couple of years, and to what extent it is or isn't entirely dependent on the adjuvant readouts that are now due 2023. And then maybe a question on TIGIT, given this is the quarter you formally confirmed the SKYSCRAPER-06 upgrade to Phase III. Any color you can give us on whether that was on the basis of a blinded interim or futility analysis by the IDMC? What's the criteria that you set for that? Or to what extent it was informed by the OS trend of SKYSCRAPER-01? Or indeed both of those factors? Any color would be very helpful. And maybe just a follow-on on TIGIT. I think you've said you'll show us the Phase II cervical data, SKY-04, first half next year, but I believe you already have that in-house. So any comment you can provide us on the nature of that data pending the details would also be helpful.

Great. So let's see. Tecentriq in the U.S. There were -- I think there's a couple of factors. For sure, there was some buying pattern that we saw from June to July, a shift. And so there was probably some end-user inventory build that -- not in the channel, but some hospitals were stocking or something because we noted that. Beyond that, I think we continue to see good uptake of the adjuvant launch. Now 90% of adjuvant patients are getting tested, and we have 55% share in the adjuvant space. We think there's an opportunity to continue to work on the share part. But also, it's true that the U.S. market is more mature for Tecentriq, and we are very much anticipating those additional readouts. And since we have 4 of them coming in the next year, I think a lot to look forward to. In terms of TIGIT, let's see. SKY-01, I think we've already mentioned that we're hopeful to still see a positive readout on OS. That could happen in Q1 with the next interim OS or about 6 months later with the final OS assessment. SKY-06, the decision to convert that to a full Phase III was based on predefined criteria that relate to SKY-06 only. There's no tie-in between SKY-01 and SKY-06. And SKY-06, for those of you who aren't familiar, is the study of TIGIT, Tecentriq plus chemo in non-small cell lung cancer, first-line metastatic. And so we had a set of criteria that were predefined. And that data was looked at by an independent monitoring group, and it met the criteria and so was expanded. And then finally, let's see. You asked about TIGIT in cervical. So that was a China study, and it -- we had the interim readout, but it's continuing to the final OS readout. And we'll have the data -- full data from that in 2023.

Was this your question, Emmanuel? Or have you meant -- also you meant -- there was like -- I think i this was the esophageal study, Bill, and the cervical cancer study, this was the Phase II nonpivotal study where we just announced that we will have data in the first half of next year. Just to confirm.

And any color you can provide us on the directional nature of that data?

No, not really at this point. You'll have to wait, unfortunately. I think our competitors would like to have this answer as well. Okay. Let's go on. Next one on the row would be Keyur Parekh from Goldman Sachs. Keyur, please, 2 questions?

Two, if I may, please. The first one, going back to gantenerumab and the pooled analysis that you've been talking about Will you be in a position to confirm what the results of the pooled analysis are by the time you issued the press release for GRADUATE 1 and 2? But I suspect that will be needed for us to be able -- or you to decide whether you can proceed with the filing or not, assuming there is mixed results on those 2 studies. So just a confirmation of whether we will get the data on the pooled analysis at the same time as the GRADUATE 1 and 2 studies. And then separately, as we look at the diagnostics business longer term, Thomas, you've been talking about opportunities both for mass spectrometry, but also on next-generation sequencing. So just wondering, kind of from a time line perspective, how should we think about both of those.

Bill, please.

Yes. So we -- our intent is that we have clear-enough understanding of the data at the unblinding, that we will have at least a directional statement in the press release. And yes -- but more than that, I won't say.

Yes. So let me answer the questions to Diagnostics. On the mass spectrometry, we have given the launch date externally already. So that's in about 2 years' time. And so we're very excited about that potential of that product. Going into in a market where we're not present today. And really, this opportunity to combine mass-spec with clin-chem and immunology will be very, very positive for the market because, first of all, this high level of automation in a setting where, today, you need a lot of people to do this. That's a saving for customers in terms of personnel, but also they don't even have the people anymore. So they actually need to automate. And we do believe that there will be an effect across the different disciplines because you'll be able to offer one solution. Beyond that, normally, we don't give out any time lines beyond next year. And I can say we won't launch NGS next year, but we are progressing as planned in that area as well.

Next question -- next 2 questions would come from Richard Vosser, JPMorgan. Richard?

One quick question, just a clarification on gantenerumab again. I'm sorry. I completely understand a pooled analysis that might be sufficient for the U.S. regulators given their previous historical approach. But in Europe, I think it's been traditional across many indications for 2 trials to be statistically significant. So how are you thinking about the European regulatory framework for Alzheimer's? And then second question, please, just a quick one on the COVID split between antigen testing and PCR testing in that 0.6 billion that was reported this quarter, or even the 9 months. It would just be helpful to understand that dynamic.

Thanks, Richard. Bill?

Yes, I'm trying to figure out how to answer these gantenerumab studies without repeating myself too much. But I give you credit, Richard, for thinking of another angle. So that's good. Look, our plan -- our primary analysis is the CDR Sum of Boxes in each of 2 independent Phase III studies. And then there's also a pooled analysis planned. I think it's all going to depend on the data, right? I mean, for example, if you had 1 study that had a 40% impact that was statistically significant, and you had the other that had a 20% impact with a p-value of 0.06, and you have a pooled analysis of 30% impact with a p-value of 0.00001, then my guess is that would probably work for regulators in the EU. Whereas if you have 1 study that just sort of limps over the line with statistical significance, and the other 1 is failing, and -- I think that's another story. So I think let's see what the data brings. These are 2 large, well-controlled studies, and we hope to have a convincing answer.

Good. So regarding your question on PCR and antigen. We had more PCR sales than antigen in Q3. So it's about 60%, 40% split. But again, as I mentioned in previous calls, so in the antigen, the sales we had was predominantly from one customer. Now it was a different customer than in Q2. And so maybe we have another customer in Q4 that we don't know anything about today, let's see. But on the PCR side, we have thousands of customers. So what we really see is a more stable base on the PCR and the antigens are basically ordered, to a very high degree, more than 90% plus, from governments. And I would say the focus of governments have been more on energy than at the moment on COVID. So it depends on how the situation evolves. If we have again a new variant coming that will increase the amount of severe infections, we may see a different situation. But at this point in time, with the low levels of hospitalization and the fact that the hospitals are not overwhelmed, we don't see that. But we'll see how the virus continues to evolve, and we'll be prepared for all optionalities.

Next 2 questions which come from Emily Field from Barclays. Emily?

A couple on OCREVUS, actually. Just on the subcutaneous and high dose studies. Would the filings of either of these be NDAs, like new drug filings, versus supplemental? And then for the subcutaneous setting, you mentioned the potential for at-home administration. Do you expect that there would be a change in Medicare coverage, i.e., moving to Part D? And kind of for either of those, if those would impact whether the asset would be on any Medicare drug negotiation list in the future. And then just a quick follow-up on the Tecentriq adjuvant studies expected next year. Just if you could provide any color on which of those 4 you expect would be the most significant commercial opportunity.

Thank you. Bill.

Yes. Thanks, Emily. Let's see. Subcu -- yes, high dose -- okay. High dose, we think, would be an sBLA. Subcu, I think, is likely to be an sBLA, but I think that one is -- it's something where it's possible that it could be something different. But we have the precedent with P+H in Phesgo. And I think also the question about at-home and Part D, it will be determined in part by whether it's approved for use by health care professional or for self-administration. And so it's possible you could have a situation where it's approved for use with a -- by a health care professional, but it could be done by a nurse in -- like a visiting nurse. In which case, it could be Part B, but still be at home. These are all things that are to be worked out, and likewise any possibility of it being included in a negotiation list. In terms of Tecentriq and which studies we have the most hope in. I have to say, I think 2 that are -- to me, that strike me is particularly important. One is the liver cancer study because it's with Avastin. This is a really powerful regimen in the metastatic setting. And we would hope that in the adjuvant setting, we would get a very strong result. And then the other would be the non-small cell lung cancer, which is actually a periadjuvant study, so it involves treatment before and after surgery. And we think that's probably giving patients the best possible chance at a long-term kind of no cancer over time. And so I'd say those are the 2 I probably have my eye on the closest.

But maybe to add on, Bill, from my side. I think we had communicated before that in terms of opportunities and size, it's head and neck, it's liver, it's the lung studies, which are biggest. Okay. Next questions would come from Mark Purcell. Mark?

Yes. So just 2 questions for me. Firstly, on Diagnostics. Obviously, high single-digit growth on the underlying basis, we're seeing coming through. Should high single digit be the new paradigm when we think about the forward outlook for Diagnostics? And can you help us understand the size of the opportunity behind the blood-based Alzheimer's screening? If we would say that would be based on lecanemab's top line profile, can you sort of comment on what you'd expect to see in terms of uptake, which is obviously linked to the perception of the clinical benefit, in the U.S. and also ex-U.S., on that lecanemab profile? That's the first question. And then the second question, going back to one from Michael Leuchten, Tecentriq subcu. Could you comment on the importance of shortened infusion times for Tecentriq? Obviously, this partly depends on the partner medicines that Tecentriq is used in. And so I'm just wondering how broad this subcu focus could become. So for example, is it possible to do a subcu version of Tecentriq plus tiragolumab? Which you could potentially then co-formulate as well.

So let me take the first question. I think, looking at the 7% growth that we had in the base business in Q3 was a very, very strong growth, if you ask me. So can we just make a line going forward at 7%? Optimistically, yes, but it could be a bit lower than that. But of course, looking into the midterm, now we talked about certain opportunities like mass spectrometry, our point-of-care system that we have in the pipeline which can do lab-like performance out of whole bloods. And also other opportunities, like Keyur mentioned with next generation sequencing, those are obviously very attractive markets that we're not in today, or not at a high degree, at least. So these are really opportunities that I see that could change the trajectory. Now when it comes to blood-based biomarkers, it really depends on how many people will go for the testing. And I think one of the things that also Bill mentioned is where I do believe that we do have a advantage is the subcutaneous formulation because I think at the end, there will not be enough opportunity to treat people on the IV side. So I think that will be a limit on how many people can be handled by the health care system at one point in time. And certainly, with the subcutaneous formulation, we do have an opportunity here to accelerate that. And that would have an impact on the blood-based markers.

And then in terms of your question about the shorter infusion time of Tecentriq subcu and how much of a difference that makes. I think it really -- it tends to matter whether it's being combined with chemo or not, right? So for example, in non-small cell lung cancer, where we have monotherapy on the PD-L1 positive patients, that would be a clear advantage. For the chemo combo settings in non-small cell lung cancer and small cell lung cancer, post the chemo course, then sometimes, there's continuation of monotherapy, so that could be important there. And so I think it kind of tends to break along those lines. You asked about the possibility of Tecentriq plus TIGIT subcu, and that's definitely a possibility. We'll see after we get our SKY-01 next interim or final results, and maybe we'll be talking about it then.

Very good. I think we have 2 more people in the queue, and I just would recommend that we have one question each. So the next question will come from Stephen Scala from Cowen. Stephen?

Hi. Can you hear me?

Yes. Clear.

So I apologize for splitting hairs, but it was stated this morning that gantenerumab data would be available in November. But previously, this company had said that a top line was possible in October. Is a November top line press release now more likely? And if yes, why was there a delay? The company previously had said the data analysis would be fairly quick. And will the release contain detailed data or just directional statements?

Okay. I mean, I can confirm that we will be able to communicate in November, but not in October. So I'm personally not aware of earlier communication of that kind. Perhaps we said fourth quarter or something, but I don't remember that we ever said specifically October. But anyway, I mean, I can confirm that it's November, and there have been no delays with the study as such. So it's running its normal course. That's when we expect the data to read out, and then we will communicate accordingly. As far as the information is concerned, which we'll share with the top line results. Bill, any additional comments?

No. I think we'll make the call when we see the data in terms of what's appropriate to share. But Steve, I promise you won't have to wait long because we're planning on being at CTAD. And so the time between the press release and the conference is going to be short. There's 30 days in November, I think that's still true. A lot of things have changed in the world, but I think we're good on that. So there won't be waiting any longer than 30 days -- or 29, I guess, if CTAD's on the 29th.

Okay. The final question for today would come from Eric Le Berrigaud from Stifel. Eric?

Yes. So one question maybe to try to get some kind of information we had the last couple of years, at the same time with Q3 call, about by biosimilar and COVID impact for the opening year. So if I get pieces of what you said previously, you were expecting COVID sales to come down from CHF 7 billion to CHF 6 billion -- CHF 5 billion, and actually likely to be around CHF 6 billion. And I would split that CHF 4.5 billion in Diagnostics and about CHF 1.5 billion in Pharma. So starting with those 2, you said also that routinely COVID in Diagnostic could trend to 20% share of CHF 1.5 billion number in full for the market. So half -- a single digit in billion, make CHF 5 billion. 20% of CHF 5 billion, make CHF 1 billion. So Diagnostics might come down from CHF 4.5 billion to CHF 1 billion, maybe not in a single year. But maybe first answer from you here, whether we have to wait for more than a year to get to that level. And then on the CHF 1.5 billion for Ronapreve and Actemra, how much should we get out for next year in AH&R? If you take 3Q into the next year, probably there another CHF 1 billion or so to lose. Is it virtually the same direction? The right direction for '23 for those 3 parts of COVID plus biosimilars?

Right. So I mean, it's a bit too early to give a more specific guidance for next year. We would, as usual, do that in January, February together with the year-end results. But directionally, what you suggest is a sharp decline for next year. And that's actually also our base assumption. Now the reality is, of course, in the past, we had many surprises with this COVID virus. And I remember personally that I declared this is over now, and then it started 2 weeks later again. So we'll see what really happens next year. But indeed, our planning assumption is a sharp decline on COVID-related sales for next year. So directionally, I would confirm what you are saying, but it's too early for a specific guidance.

Okay. I think with that, we are at the end of our call. I would like to thank the speakers and also the IR team for the work up front and preparing the event. And thanks for your interest in Roche, and have a good day.