HUTCHMED (China) Ltd (HCM) 2017 Q4 法說會逐字稿

Hello, ladies and gentlemen, and welcome to the Chi-Med 2017 Full Year Results. My name is Josh, and I'll be your coordinator for today's event. (Operator Instructions) I am now handing you over to your host, Mr. Christian Hogg, CEO, to begin today's conference. Thank you.

Okay. Thank you, Josh. Welcome, everybody, to our 2017 full year results call. I'm going to take you through the presentation. I'll try and do that in 25, 30 minutes, and then open it up for Q&A at the end of it. I'm assuming everybody has the presentation in front of them that's been published today on our website.

So if you go to Page 3, you can see the highlights. So during 2017, Chi-Med made a lot of progress. The pipeline is just deep and approaching approvals in certain areas. You can see on the top of that page, we submitted our first NDA on fruquintinib in third-line colorectal cancer. The -- all indications are that we should be able to launch it in 2018. Specific timing of it is difficult to judge, but we are well through the process on the NDA inspection and approvals process, and I'll give you an update on that later.

On savolitinib, we've announced today in our detailed announcement that AstraZeneca has agreed to proceed into final development on the savolitinib/Tagrisso combination in second-line nonsmall cell lung cancer. I'll talk more about that later, but it's a big program that we have planned and very complex. But it's an indication, as we presented at the World Conference on Lung Cancer in October, that we have a very strong signal of efficacy for savolitinib/Tagrisso and good safety as well. So I'll get into that in a bit more detail later.

Six Phase III trials underway. The FALUCA study in third-line nonsmall cell lung cancer. We completed enrollment of 527 patients about a month ago. Expect the top lines to report at the end of the year, but that's a big pipeline of pivotal studies that we have running, and it's only expected to expand. And then 22 proof-of-concept studies on our 8 drug candidates, all enrolling in parallel. So we currently are running studies on 36 target patient populations around the world at the moment, and that's up from about 30 a year ago.

So the pipeline is broad and moving. The discovery engine is -- under our Chief Scientific Officer, Dr. Weiguo Su, is continuing to produce. We are getting closer with our second wave of innovations -- sorry, our third wave of innovations, sort of second-generation IO targets that we expect to start seeing entering in the clinic over the next year or 2. And everything's in-house. So pretty much everything we've produced, everything we have in the clinic today is out of our own shop.

And then the commercial organization is large and is ready to launch our products. We'll talk about our commercial organization as I go through the results, but it's a really powerful group of people that's delivered some tremendous results in 2017 on our own products as well as third-party products. And I have a great deal of confidence that when we do get to the point of launching our oncology assets into China, our team is going to be able to deliver some terrific commercial execution.

So going over to Page 4. You can see the results for the year is in line with guidance. On the right-hand side of the page, 2017 guidance was $13 million to $28 million. Net loss, we came in at $26.7 million on the sort of higher end of it. But the -- if you look into the detail, you'll see on the innovation side, we had a net loss of about USD 51.9 million. That's including investment of about $88 million on our clinical programs. Our commercial platform did very well in 2017 with net income after tax increasing 25% over 2016. And so net, everything pretty much fell in the guidance and as we'd predicted.

If you look at Page 5, you can see what I've just said: $88 million in innovation investment, about $36 million of revenue in 2017 from our partners, AstraZeneca, Lilly, Nestlé. And then the commercial side, you can see the $37.5 million after tax, so a 25% increase.

I mean, in the context of what we've done in 2017, that result on the commercial side is terrific. I mean, we've basically moved both our major manufacturing facilities to new sites, built new manufacturing facilities affecting 1,000 manufacturing people, and we've done that while growing our business 25% in the context of net profit. So it's just a great team of people we have on the commercial side in China, and they're waiting for our oncology assets to come to market.

On Page 6, you can see, as a result of the follow-on offer in October, we are sitting in a good position from a cash standpoint, $358 million in cash and cash equivalents, short-term investments and another $120 million-or-so in unutilized banking facilities. So we're in a strong cash position. Minimal borrowing, $30 million in borrowing. And another $60-or-so million -- $67 million in cash in our joint ventures.

So looking forward into guidance for 2018. You can see that the commercial platform will continue to grow net income. We're guiding to $41 million to $43 million in '18, up from $37.5 million in 2017. So that's sort of 10% to 15% growth.

There are some adjustments that are being made in the context of how much sales we can recognize on the commercial platform because the new Chinese SFDA Two-Invoice policy now means that we cannot consolidate the sales of Seroquel, for example, third-party products. Instead, we will only be able to book the fees that we are paid for, for delivering those commercial services.

So from a profit standpoint, there is no change in the profitability of our commercial platform. But from an optics standpoint, we'll be consolidating less sales as a result of that Two-Invoice policy. But it doesn't affect the operation of the business and how we commercialize Seroquel in China, for example. Nothing changes. We still run a team of 120 reps, and we're still doing a good job, which we'll talk about later.

So -- but when you add it all up, we also have -- we have quite a big onetime gain range for 2018 of 0 to $20 million, and that's around the property in Guangzhou. That may or may not come in 2018. We hope it will. But we're sort of at the mercy of the Guangzhou municipal government on that.

If it comes, it'll be very helpful. If it doesn't, it'll come next year. And the good thing is that the price of property in Guangzhou is increasing every day. So frankly, we hope it comes this year. But if it comes next year, it's not the end of the world.

So the range of guidance for 2018 is somewhere between $19 million and $52 million. It's quite a broad range. But in the context of running a portfolio of clinical programs that is going to invest between USD 110 million and USD 120 million in 2018 in clinical development, it's all very sensible. And in the context of our cash resources that we have at the moment, this is why we're saying in our guidance that the cash that we have at the moment is sufficient to see our clinical programs through 2020 without much of a problem.

Okay. So moving on to Page 7, the milestones for 2018. You can see a lot going on. Savolitinib, it's starting this global randomized, chemo-doublet controlled study of savolitinib and Tagrisso in second-line nonsmall cell lung cancer. It will start out as a Phase II. The reason for that is we have not yet engaged in discussions with the regulatory authorities on the design of the phase -- of this study. And the reason for that is because we're still working on the dose.

Now I'll go into that in a lot more detail later, but this -- the dose work that's going on at the moment is fine tuning, I would call it. And we expect that to be completed probably around the end of the year, at which point we'll go talk to -- well, AstraZeneca will go talk to the regulatory authorities with regards to a potential Breakthrough Therapy designation and this particular study.

What we didn't want to do is wait until the end of the year to start getting ready for this study. It's going to be a global study. It's powered to -- for ORR and for PFS. So it's a good signal from Astra that they're aggressively moving ahead even with some fine tuning on dose continuing. And so our hope is that, by around the end of the year, we'll have all the dose work done. We'll have the interaction with the regulatory authorities. We will have already started this global randomized study with multiple arms, and we can drop off the arms we don't want at that point. So it's a very aggressive program of -- moving in parallel as opposed to in series.

Number two, the Tagrisso combination in third-line nonsmall cell lung cancer. This one, again, will depend on the dosing, what we're doing and the interactions with the regulatory authorities on second line. But for us, it's inevitable that we will move ahead in this. I think with AstraZeneca, the great opportunity here is to broaden their Tagrisso and their EGFR franchise, and savolitinib is the perfect partner for this.

On number three, the savolitinib/Iressa combination in China. This one is more debatable. Astra, their view on the savolitinib/Iressa combination is that they are concerned it's going to cannibalize Tagrisso. My view on that is that I'm more concerned about maximizing the sales of savolitinib as opposed to cannibalization of Tagrisso. So our view is that, in China, the Iressa/savolitinib combination will have great potential given the large patient population. Almost 50% of non-small cell lung cancer patients in China EGFR driven. Iressa is now off-patent in China. It's generic -- multiple generics, and the costs of a month of treatment of Iressa has gone from, 2 years ago, around USD 2,500 a cycle, now down to around probably $400 a cycle.

So it's -- the combination of savolitinib and Iressa is, by far, the most likely to succeed in China in the context of patients being able to afford it. So we're pushing it, but it's a discussion that's ongoing with AstraZeneca. And I'm sure, over this next 6 to 9 months, we will conclude a sensible way of proceeding in that. And I can talk about that more later.

Finally, the molecular epidemiology study on PRCC is set to read out around the end of the year. We hope that, that will lead to a compelling dataset showing that MET is a major driver of negative outcome in PRCC and hoping that we can take that molecular epidemiology study data, combine it with our Phase II data and then go talk to the U.S. FDA.

But that Phase III in PRCC is enrolling now. There's probably 60, 70 sites up and running around the world. And we expect that Phase III to complete enrollment around the end of next year. So the molecular epidemiology study, we feel that, that is something that can potentially accelerate what is already a global Phase III program.

Fruquintinib. It's about the NDA approval and the launch in China, and then the release of the FALUCA data probably Q4 of this year.

Epitinib. We've been working on dose on epitinib. The data we presented at World Conference on Lung Cancer in late 2016 was 160 milligrams. We spent '17 working on 120 milligrams. I'll take you through why we've been doing that when we get to epitinib. But we're now pretty confident that 120 is likely the dose we want to take into Phase III, and we have to decide that within 2018.

Sulfatinib is moving along with a Phase II set to start, hopefully, before the end of the year in the U.S. The Syk inhibitor, we're enrolling patients every single day, hopefully, and expanding sites in both Australia and China. And we hope that by the time we get to ASH by the end of this year, we'll be able to present some pretty compelling data. And the 689, the PI3K delta is moving along nicely as well, seeing some encouraging early activity.

So moving on to Page 9. I won't go through these in a lot of detail, but the scale of the organization continues to build, and the investment is now well over $500 million.

Page 10 and 11, the selectivity of our programs. I won't go through that in detail as everybody on this call is aware of our strategy here. But we've added a chart on Page 11 that shows discontinuations -- total discontinuations of first-generation small molecule-targeted therapies and compares them to our highly selective small molecule-targeted therapies.

And you can see, taking Sutent as an example, the total percentage of discontinuations on Sutent -- and this is in ccRCC. It's sort of 35%, 38% discontinuations. It's a $1 billion drug. So that's kind of a -- and if you look through the rest of them, looking at cabozantinib, looking at lenvatinib, looking at Stivarga, sort of talking 20% to 30% discontinuations, whether that's for AEs or for any other reasons. So that's what small molecule -- first-generation small molecule multi-kinase inhibitors tend to see.

If you look at our monotherapy, savolitinib, fruquintinib, sulfatinib, epitinib, you're seeing sort of low teens for the most part discontinuation rates. And we put that down to better selectivity, less off-target toxicity and better tolerability. And I will come back to this later when we start talking about savolitinib in combination with Tagrisso, but these reference points are quite important.

Moving to Page 12 and 13. I won't go through it in detail, but you can see, we've -- we have -- on this chart as well as in the full announcement that we've put out today, we've given a full update on each of those 36 target patient populations that we're studying at the moment. You can see savolitinib just continues to expand with a prostate cancer study, a Phase II study, investigator-led study in Canada that just kicked off recently. So we're doing more and more around our initial indications that we're going full on -- at full speed on. We're doing a lot more exploratory around the edges to look for what are the life cycle indications of the future.

Page 14 lays out the 8 to 10 shots on pivotal success. I want to talk a little bit about savolitinib in combination with Tagrisso in second line. It's the first circled study there -- randomized chemo-doublet controlled study to initiate in 2018. This is a study that's powered for PFS and ORR. We'll start it out with 3 arms.

So basically, what we're doing when it comes to dosage is the World Conference on Lung Cancer data was presented with data for 300 -- sorry 600 milligram QD dosage. And what we saw with that 600 milligram QD dose was sort of 60% -- around 60% objective response rate in these patients, second-line patients; 55% on MET-positive T790M positive and 61% on T790M negative patients. We've been able -- since we published that data to watch how PFS has evolved. We have not published it. But it's evolved as we expected, and it'll be presented at a future scientific conference.

So what we're now doing here is getting ready to initiate this study, and we're going to do it with a couple of dose regimens that, we believe, both of which will likely be successful. The first is 60 milligrams -- sorry, 600 milligram weight-based algorithm. So as we use in papillary renal cell carcinoma, patients over 50 kilograms get 600 milligrams of savolitinib. Patients under 50 kilograms get 400 milligrams of savolitinib. That's for the monotherapy.

So what we're doing with the combination here with Tagrisso is starting out with a weight-based algorithm off of 600, likely to be -- we haven't published what that lower dose will be for the lower-weight patients, but we'll figure that out shortly and design our study around it. Separately, we're also going to have a 300 milligram arm in this randomized study. And then a third arm is the chemo-doublet controlled arm.

Now that study will prepare, and it will initiate in the second half of 2018. In parallel with that study, we have continued to enroll patients in the TATTON D study of 300 milligram QD. What we're looking for is, are those patients seeing the same levels of response, so the 55%, 60% response that we saw at 600 milligrams? And what are the duration and response of the PFS data for 300 milligrams? If we can get patients to deliver the -- or we can deliver the same levels of response and the same levels of PFS with 300 milligrams, that's obviously going to be more tolerable. And I'll go into tolerability of the combination in a moment.

But -- so what we intend to do is answer that question through TATTON D over the balance of the year. If we need to go with a 600 weight algorithm dosage, that's fine. If 300 is delivering the same levels of efficacy, then we'll go with that. We'll -- either way, we'll drop one arm come the end of the year or early next year and continue that on. We'll take that data from TATTON D to the -- or AstraZeneca will take that data to the regulatory authorities, talk about Breakthrough Therapy designation, talk about registration pathway. And as I said, this study is powered for PFS and ORR. And I think we all expect that, that should be sufficient for a submission if we play this study out to its endpoint.

I want to take you briefly to a chart in -- on Page 24 to talk about the work we're doing around dosing on savolitinib/Tagrisso. If you look at just nonsmall cell lung cancer, you can see, on Page 24, Tagrisso in second-line EGFR refractory T790M positive nonsmall cell lung cancer, the total discontinuation rate of that monotherapy is around 13%. But you look at ceritinib, you look at Keytruda, you look at the other IO agents as well as Cyramza, the VEGFR antibody from Lilly, you've got your 20% to 30% discontinuation rate mostly from monotherapies there. So that's kind of normal, and these are all approved therapies -- mono -- approved monotherapies. Actually, Tagrisso is uniquely low just because it's such a clean, selective compound.

If you look at the second section of this chart, these are the control arms, the chemo doublet and Taxotere monotherapy. And you can see, total discontinuation rates of those therapies are sort of high 20s all the way to high 30s, so similar to what you would see with Sutent, for example, in -- from the previous page.

And if you go down to the bottom, you see savolitinib. Savolitinib monotherapy in PRCC is a reference point, 14% discontinuation rate. So similar to Tagrisso monotherapy. It's just a very low level of discontinuation with the monotherapy. But when you put Tagrisso and savolitinib together in combination, you see about a 33% discontinuation rate.

So that is more consistent with either first-generation multi-kinase inhibitors or the top section of the chart, the IO, et cetera. So it's not bad. It's consistent with what's out there at the moment. But if we want savolitinib to be a multibillion drug as opposed to a $1 billion drug, our desire is to fine-tune that tolerability and to reduce that discontinuation rate.

So this -- one thing you should realize also, in the bottom 3 lines of this chart, the Iressa combo and the 2 Tagrisso combos, these are greater than second-line and greater than third-line patients. Many of these patients have been on many lines of treatment prior to savolitinib and Tagrisso, so generally sicker, further along in their disease and more difficult to treat. So that can lead to slightly higher discontinuation rates as you have patients that are sick or have been on the drug longer -- or have been on multiple therapies longer.

Also, you should remember that these are at 600 milligram QDs, so there's no weight-based algorithm in play here. This is just 600 milligrams until either progression or discontinuation or death.

So we feel that the weight-based algorithm of 600 above -- sort of at a high weight and a lower dose at a lower weight will allow for better tolerability and for longer-term usage. Ultimately, it's about keeping these patients on this combination for as long as possible.

Also, that's why we've gone already enrolling TATTON D and dosing patients with 300 milligrams because if we're able, on the combination, to deliver the same level of efficacy that we've been seeing here and same level of duration -- response of PFS at the lower dose, and then that dose, 300 milligrams, is the dose that we've seen a lot of efficacy at in our early development, then that would also potentially lead to better tolerability for long-term usage.

But either way, in the worst case, we go with the 600 milligram weight-based dosing algorithm, it's going to be better than what's on this piece of paper and in a specific target patient population of second line, not second line and above, you're going to see better tolerability. So this is why AstraZeneca has moved rapidly into planning for this randomized -- global randomized chemo doublet controlled study and recognizing that there may be some throwaway as we drop off an arm as we go along the way. But so be it because we don't want to waste any more time on this, and we want to go as fast as possible.

Okay. So back to Page 14, the pipeline chart -- sorry, the 8 to 10 shots on goal. It's all basically self-explanatory there. Fruquintinib, the top lines will come at the end of the year. And I'll talk a little bit about epitinib as we get going and what we're going to do on epitinib.

So 16, 17, I won't go into any more detail. Most of you know and are very familiar with savolitinib's scientific rationale in lung cancer.

Page 20, you see the Tagrisso combo data in second-line setting. We feel strongly that when we can go to the regulatory authorities with a clear picture on dose, whether that's the 600 milligram weight-based algorithm or a 300 milligram QD dose, this kind of data should support Breakthrough Therapy designation. And we hope that, that discussion with the regulatory authorities can happen later this year or very early next year.

Page 21 shows why we are pushing AstraZeneca on the savolitinib/Iressa combination because in those T790M negative patients, savolitinib and Iressa is a very effective treatment regime with potent efficacy. And you can see here, good durability of response. So we hope that we can get that agreement to proceed there, sorted out over the next few months.

Moving on to Page 22, 23. In third-line nonsmall cell lung cancer, the combination with Tagrisso again, increasingly large patient population is MET-driven as you go through Tagrisso failure. So that is why Astra is so keen on that area, and we will push that as well. I think the third line is -- to some extent, we've chosen -- or Astra has chosen to go into this global study on second line just because it's the stronger of the signals. It's sort of 55%, 60%. The third-line setting, the signal is sort of 30%, 33%. But still, it's a major patient population, and we intend to keep pushing this. And I expect we'll start pivotal studies on this globally next year some time or after we've had our regulatory discussion.

I won't talk any more on savolitinib. Moving through Page 26, 27 into papillary renal cell carcinoma. That study, as I said, is enrolling globally at the moment, the Phase III. And we're waiting for the molecular epidemiology data so that we can go back to the regulatory authorities and talk Breakthrough. Hopefully, that molecular epidemiology study data will show how bad MET is in the context of outcome in PRCC. The same -- as shown on Page 29, gastric cancer patients, lung cancer patients, if they're MET-driven, they're seeing a median overall survival of only 2 years from -- for after surgery. Whereas if they're MET-independent, they're living -- they're seeing a median overall survival past 10 years. So MET is a terrible driver of these types of solid tumor. And we believe the same is to be said for papillary. And once we can show that in the molecular epidemiology study, we will then reengage with the regulatory authorities.

Page 30 is the safety data and a bit more detail of the monotherapy in PRCC. Very clean. Page 31 shows gastric cancer. And Page 32, also the example that many of you have seen already.

We screened, over the last 2 years, 850-plus gastric cancer patients for MET-driven disease. We have concluded it's around 5% of patients that are -- that have MET-driven disease in gastric cancer. And we've presented data to show that when exposed to savolitinib monotherapy, these MET-driven gastric cancer patients do very well. So it's an example of what's the -- what exploratories are we doing and what would be the life cycle indications. Once we get kidney cancer approved and once we get lung cancer approved, we will be going after multiple indications, including prostate cancer, as I -- was we wrote in our document here showing that we started a Phase II study now in MET-driven prostate cancer in Canada.

Fruquintinib, Page 34, 35, just the data that we presented at ASCO last year. And with the NDA submitted, we're now waiting for that approval.

I think 36, 37 shows the differentiation versus Stivarga, particularly in the area of safety. But most of the people on this call are well aware of that.

Page 38 is the FALUCA Phase II data. It was only powered for PFS, and you can see a very clear statistical significance in PFS and a directional benefit in PFS in the EGFR mutation-positive patients. But this 527-patient study is now complete and will read out likely in Q4 some time.

And then Page 39, you see fruquintinib in first-line nonsmall cell lung cancer, combo with Iressa. That's moving really well at the moment. 76% response rate. The tolerability of the combination is very good. The flexibility to take the small molecule VEGFR inhibitor off if toxicity has emerged is there because fruquintinib has such a short half-life that it can be taken off easily and patients can have their AEs resolved if they do occur, whereas, with Avastin or ramucirumab with a 3-, 4-week half-life, it's very difficult to take patients off quickly. So we think a small molecule in this setting is very important.

And then Page 40 is the gastric cancer Phase III we have enrolling at the moment in China. It's a study that's going to end up being about 540 patients. We have an interim analysis for futility sometime late this year, maybe very early next year. And that'll be important point for us.

Sulfatinib. Page 43, there's a bit more detail on Page 43 than before showing which patients had failed on Sutent and Afinitor. And you can see very good response on these pancreatic NET patients that fail on Sutent and Afinitor, good response to sulfatinib. So we're moving now into a Phase II expansion study in these patients in the U.S., which, hopefully, will start before the end of the year.

Epitinib. I won't repeat the data on 46 and 47. But if we go to 48, I can highlight what we've been working on this last year on the dose. I mean, if you look at the AE table on the right-hand side, you can see hyperpigmentation on epitinib of about 48% of patients. Now what this is, it's a cosmetic pigmentation of skin, basically getting freckling on the skin, sort of brown freckling.

Now with most of these patients in China in EGFR mutation-driven nonsmall cell lung cancer being women, pigmentation of the skin is something that they don't like. It's cosmetic. And in the context of having brain metastases, you would imagine that patients wouldn't worry too much about it. But in China, skin color and skin tone is very important. In particular, these patients, late in their lives, they don't want to look bad. So what we've done over the last year is we've been working on enrolling patients at 120 milligrams. A case study of one of those patients is on the right-hand side. You can see very compelling antitumor effect in the brain at 120 milligrams. In this case, it's a man, 46-year-old male.

Both the top and bottom CT scans, it's the same brain. It's just a different cut of that brain. And you can see great efficacy at 120 milligrams. So epitinib at 120 milligrams gives the sufficient drug exposure to deliver the benefits we need. And so most likely, we'll start the Phase III at 120 milligram. And what we have seen also is at 120 milligram, the pigmentation in the skin is nowhere near as common. So that's a good thing. So we'd hope to move that Phase III into initiation this year.

I won't go through the Syk inhibitor. The balance of the pipeline, just to say, as I said earlier, we continue to build the dataset on the Syk inhibitor. And hopefully, by the time of ASH, we'll have some pretty material data to share.

Page 57 on the commercial business. You can see 25% net income growth after tax attributable to Chi-Med, $37.5 million. It's been a tremendous year for our business in China. Page 58, you can see the scale of our operation in China is enormous, and it's sitting there ready to take on our oncology assets when they're ready to launch.

Page 59 shows the performance of Seroquel and Concor over the last year has been terrific. Ever since we took over Seroquel from AstraZeneca back in very early 2015 -- or April 2015, it's been great performance. This year in 2018, we're off to a very strong start. And Concor as well. The better we do, the more provinces that Merck Serono gives us.

So I'll leave it there. Page 60 just sort of summarizes the main products that we're commercializing at the moment. But I'll leave it there and open it up for questions from everybody on the line.

(Operator Instructions) So our first question comes from the line of Ying Huang from Bank of America Merrill Lynch.

Maybe, Christian, just the first one, quick one on savolitinib. Have you guys and also AstraZeneca looked at the human PK of 300 milligram dose? Will that give you similar exposure -- or sufficient exposure for efficacy purpose? And then secondly, under which condition would this Phase II trial in second-line nonsmall cell lung cancer become a pivotal trial?

So yes, thanks, Ying. So we have extensive human PK data on savolitinib and Tagrisso combination. And the reason that we're looking at the 300 milligram QD dose is because we think the target coverage should be sufficient at that level. But we want to get some more patients in -- treated and look at that data. So I think we've already actually enrolled several 300 milligram QD patients. It's a huge priority of AstraZeneca to answer this question as quickly as possible so that, hopefully, if we can answer this very rapidly, when we actually initiate the study, we might also be able to initiate that global study at that dose. It's unlikely to happen that way. I think we'll start with -- the 600 milligram weight algorithm dose plus the 300 milligram QD dose in 2 separate arms and a third arm being the chemo doublet. And maybe in early '19, we'll drop off one of those arms. That's the most likely scenario. But as soon as we have sufficient information on that question, we will go straight to the regulatory authorities in the U.S. because I think we're all quite keen to put this data in front of them, to talk about Breakthrough Therapy designation and to get alignment on what we're doing. I think once that alignment is forthcoming, then it would make total sense to have this study that starts off as a Phase II then evolve into a Phase III. But to be clear, it's powered for registration, so for PFS and for ORR. So assuming that the dose discussion is resolved, and based on the TATTON B data of efficacy, this study will be powered for submission based on its completion. Does that answer your question?

Absolutely.

Our next question comes from the line of Alex Schwartz from Stifel New York.

I had 2 fruquintinib-focused questions. First off, with your Phase I bridging trial that just began in the U.S., kind of pending success of PK and safety tolerability as well as factoring in the U.S. commercial landscape, are there any indications or -- that are particularly attractive that you'd pursue in development first?

Okay. Thanks, Alex. Yes, I mean, what we're not looking at on fruquintinib is to go head-to-head with Stivarga in third-line non-small cell -- sorry, third-line colorectal cancer. We don't see that as a big opportunity. We see the big opportunity for fruquintinib is innovative combinations playing off of the strong tolerability of that drug or fruquintinib. And so the kinds of things that we'll be looking at are, for example, combination of fruquintinib and savolitinib in clear cell renal cell carcinoma, where VEGFR and c-MET are both very important targets. The combination of fruquintinib plus Iressa -- or potentially fruquintinib plus Tagrisso in first-line nonsmall cell lung cancer, I mean, that's a long study because the PFS there should be pretty long. But it's those innovative combinations that we're interested in, also fruquintinib plus various IO programs in various solid tumor types. So that's what we'll be looking for is innovative combinations. And as a result of that -- I mean, we expect this to be -- the Phase I to be ready by the -- complete by the end of this year and for us to start then looking into these combinations. So this year, we're already beginning to build up our clinical regulatory team in the U.S. in readiness for it to take on more of these later-stage proof-of-concept type studies. And we're -- so that's the process of building the team there that's going on at the moment.

Okay, excellent. And then my second question, with your 2018 guidance, what estimates do you include of revenue or OpEx in regards to fruquintinib's launch? And then with your GMP facility that's built, what steps, if any, remain to manufacture commercial quantities of the drug?

Okay. So in 2018, if we're launch -- if it's -- just for argument's sake, we get an approval sometime around the middle of the year, we will then take 2, 2.5 months to actually produce commercial supplies and then launch probably in Q3 some time. Now -- so the revenue in 2018 will be relatively low, I would imagine. We will be paid a $15 million milestone on approval by Lilly, so that'll hit in 2018. So -- I mean, that can give you a sense of what to expect in 2018 subject to an approval middle of the year. Hopefully, it'll come a little bit sooner than that, but we'll have to wait and see. But I don't foresee it being really that material on the broad scope of the revenues that we book at Chi-Med. The GMP certification is set to allow us to produce commercial supplies of fruquintinib for China. So once we -- and that's the stage of NDA approvals that we're at, at the moment, is we're into this sort of the last phase, which is the GMP certification of our facility and the inspection of all our API site -- or the AP -- the site that we produce our API at. So once those GMP certifications and inspections are complete, we'll just be able to produce commercial supplies at those sites. I don't know whether that answers your question, but it's -- that's the situation.

Our next question comes from the line of Andrew Peters from Deutsche Bank, San Francisco.

Another follow-up from me on savolitinib. Just wanted to understand your comments around potential trial design and the path forward towards a potential registration for the study. As you think about your powering comment, is the study powered right now sufficiently with enough patients that if you drop one of the arms in the study? Or how do you think about kind of that adaptive-type trial design, where you're looking at 2 doses right now but potentially only one for the primary endpoint? Is that something that's fairly standard for potentially pivotal studies? And in your conversations with your partner, are they kind of on the same board in terms of ability to view the study as potentially pivotal at some point?

Thanks, Andrew. So as far as the powering and the number of patients, what we have here is a discussion and a planning process with AstraZeneca that has been going on for the last 3, 4 months. Hugely detailed. The -- I think the belief is that this dosing work TATTON D that's going to play out as well as the maturing TATTON B data will be sufficient for us to make this decision around later in this year, very early next year, just depending on scenarios. So the randomized global study is designed taking that into account. So we are beefing up the numbers of -- if we had a single dose and we were starting this study, often it was a single dose compared to the chemo doublet, it would be a slightly smaller study. But we're beefing up the study knowing that we're likely going to drop one arm. So the study has been designed in that fashion to basically upsize it knowing that we'll drop an arm probably early next year some time, okay. So that's fully designed for taking us through to registration, taking into account dropping one of these arms. We won't design it, drop an arm and then find ourselves in a situation that's not powered enough. That's very clearly understood and managed. On the -- or your question about the conversation with AstraZeneca, what was the question again?

Well, I guess, just understanding their commitment to the program right now and their understanding, their time lines, I guess, and sense of urgency around the program.

Got it. Well, I mean, I think this is a good question because designing a program like this is quite -- it's not common that you would actually kick off a study aimed at being a study that's ultimately going to be used for registration before you finalize the dose. It's very uncommon. But the reason they're doing it is because they don't want to lose any time because they feel very strongly about the savolitinib/Tagrisso combination. Actually, there's a statement in our announcement that actually AstraZeneca wrote that talked about -- encouraged by the results of TATTON B, they've decided to progress into this global chemotherapy-doublet study. Those are their words, not mine. So they are clearly, through their actions, showing that they -- regardless, they feel that this dose question is more fine tuning rather than anything beyond that. And they don't want to wait until that fine-tuning decision is made before they start planning this global study. So they want to start the study -- planning now. It's going to take 6 to 9 months to kick off a global study of this scale. Anyway, so they don't want to have to wait till the end of the year to kick that off. So I think that's a good sign of their commitment, that it's all just moving ahead in parallel. And everything is very clear, designed in that fashion. And frankly, the difference between a 600 milligram weight-based algorithm dose and the 300 milligram QD dose, whichever one turns out to be better is going to be fine. Even the one that was worse would be fine because what you're seeing from the charts I've shown you is that the percentage of patients that are dropping off of this combination dose is about 33%, which is very consistent with pretty much everything else other than fantastic drugs like Tagrisso monotherapy. So I think they know we've got a winner here. We're trying to fine tune it to optimize long-term use of this combination so the patients can ultimately stay on it longer. And that will really benefit both patients as well as the market potential of the combination. Hopefully, that answers your question.

It does. And I guess a quick follow-up on that. Can you remind me the time lines then for savolitinib in Tagrisso failures? And I guess, what's the driver of the slight delay versus the Tagrisso combination?

Yes. It's -- there's no delay, really. The -- so what you're seeing on the second-line setting is a very aggressive program that's going forward even before the final dose is figured out. On the Tagrisso failure patients, so for the third-line patients, I can say from my interaction with AstraZeneca, they are very bullish on that indication. It's clearly a lower signal. The strongest signal is in the second-line setting, where you see the 55%, 61% objective response rate. So in the third-line setting, you're seeing more like 30%, 33% response rates, albeit -- I won't go to a lot of detail, but they're very encouraged by that. In particular, I'll raise the point that those third-line patients that failed on Tagrisso, many of those patients in TATTON B had been on multiple lines of chemo, multiple lines of treatment. Some of them were 5- or 6-line patients. So that 33% in that patient population that's been exposed to a lot of treatment is actually very encouraging. So they're encouraged by it. They're aggressive to pursue it. But first things first, let's get the second-line study going, planned, get through this dose fine tuning. And once we've got this interaction with the regulatory authorities around what we find out of TATTON D, the dose finding -- the dose work, then the Tagrisso failure patient global study will be easy to do. And we'll focus on that early next year, getting it going.

I think TATTON B is still continuing now. I mean, they can -- if we have a large-enough database (inaudible) we'll figure out discussing with FDA on Breakthrough Therapy designation. Right now, it's just -- the sample size is not big enough.

Yes, that's true. Weiguo is right. The TATTON B study continues to enroll, and you'll have more patients as well. So bottom line is, Andrew, they're extremely aggressive on that. And actually, it was an ongoing discussion as to which one to pick to go super aggressive on, and they ended up going for the one with the strongest signal, basically.

We currently have no further questions on the line. (Operator Instructions) Our next question comes from the line of John Newman from Canaccord New York.

Every time we have a call, it's just hard for me to realize how much work you have going on and how much success you had with multiple drugs, so that's great. My question is pretty straightforward, just a quick check. The Phase II portion of savolitinib plus Tagrisso combination that you're looking to start, I'm just wondering about the geography there. Would the geography there also be worldwide? Or would it be a bit more specific?

No, no, its global. It's global, John. So that study, I imagine, we -- I -- we haven't got all of the countries that we're going to do it in laid out just yet, but it's -- it'll be global. So North America, Europe, Asia.

So we have no further questions on the line. (Operator Instructions) Okay. So there are no further questions on the line, so I'll hand back over to your hosts.

Okay. Thanks very much. Thanks, everyone, for calling in.

Thank you very much for joining today's call. You may now replace your handsets.