AstraZeneca PLC (AZN) Earnings Call Transcript & Summary

Welcome, ladies and gentlemen, to AstraZeneca's Investor Conference Call ESMO 2021. Before I hand over to AstraZeneca, I'd like to read the safe harbor statement. The company intends to utilize the safe harbor provisions of the United States Private Securities Litigation Reform Act of 1995. Participants on this call may make forward-looking statements with respect to the operations and financial performance of AstraZeneca. Although we believe our expectations are based on reasonable assumptions, by their very nature, forward-looking statements involve risks and uncertainties and may be influenced by factors that could cause actual results to differ materially from those expressed or implied by these forward-looking statements. Any forward-looking statements made on this call reflect the knowledge and information available at the time of this call. The company undertakes no obligation to update forward-looking statements. Please also carefully review the forward-looking statements disclaimer in the slide presentation that accompanies this conference call. [Operator Instructions] And with that, I will now hand you over to AstraZeneca.

Thank you. Good afternoon. Good evening, everybody. And it's a pleasure to host this investor and analyst science conference call coinciding with the ESMO 2020 on Congress. It's great to be with you all. So if you could move to Slide 3, please. Thank you. Today, I'm joined by Dr. Sara Hurvitz, senior investigator in HER2 Phase III DESTINY-Breast03 study and professor at David Geffen School of Medicine at UCLA. I'm also joined by AstraZeneca colleagues who will present this evening. Susan Galbraith, our EVP of Oncology R&D; and Dave Fredrickson, our EVP of Oncology Business Unit. For Q&A, we'll be joined by Cristian Massacesi, our Chief Medical Officer and Oncology Chief Development Officer; and Sunil Verma, our Vice President of Late-Stage Breast Cancer Development. If you could move to Slide 4, please. Here's the agenda. Susan will kick us off first with an overview of our pipeline and scientific pillars that support it and a recap of ESMO 2021. Dr. Hurvitz will then walk us through the DESTINY-Breast03 data. Dave will then talk to us about what this means for Enhertu and what's next as we broaden out the development of this potentially transformative medicine. Then back to Susan for a recap of AstraZeneca's other ESMO highlights, including Enhertu beyond breast cancer plus Imfinzi in combination with 2 novel mechanisms of action. As is tradition, we'll leave plenty of time for Q&A, such as being the rich flow of data at this ESMO. This will be chaired by Dave. I will now pass you over to Susan and if you could move to Slide 5, please.

Thank you, Chris. I'm delighted to be able to introduce you to what has been a fabulous ESMO meeting for AstraZeneca. If you move to Slide 6, I'll just remind you that this is in the context of the comprehensive portfolio we have at AstraZeneca with our approved drugs on the left-hand side, and on the right-hand side, a reminder of the mechanisms that we're investing in discovery and early development. On the right-hand side of this wheel, you can see the mechanisms that kill cancer cells directly through tumor driver and resistance mechanisms, DNA damage response, antibody-drug conjugates, epigenetics. And on the left-hand side, you can see that we're also very active in trying to reactivate the immune system through classic checkpoint inhibition, but also through novel approaches, such oncolytic viruses, cell therapy and immune cell engagers. So that's the context in the background for this ESMO 2021 meeting. If you can move to Slide 7. Delighted that this ESMO is the fifth year in a row that we've had a presidential symposium presentation. And beyond our presidential data with the DESTINY-Breast03, which we'll come to shortly, we've had 65 abstracts in total and 20 oral presentations, as you can see listed here, with highlights, as Chris has already spoken to, beyond D-B03 with Enhertu in other cancers, including the DESTINY-Gastric02 data, the DESTINY-Lung01 data in HER2 mutant lung cancer. And then exciting data with Imfinzi with updates from the real world in the PACIFIC-R study, showing the importance of durvalumab in that Stage III non-small cell lung cancer setting; long-term updates for outcomes in CASPIAN, which has been unprecedented in small cell lung cancer; and then again, the exciting Phase II data that we've got with COAST building on our presence with durvalumab in the PACIFIC setting. And of course, there's more data from the rest of our portfolio as well, which we'll be happy to answer any questions on. But now let me turn over to a guest that we have. Dr. Sara Hurvitz, who's the Senior Investigator of DESTINY-Breast03 study, who's going to walk you through the very exciting data from this study. Dr. Hurvitz, over to you.

Thank you so much. Please go to the next slide and the next one. So here are the study design for the DESTINY-Breast03 clinical trial. In this trial, patients who had metastatic or advanced unresectable HER2-positive breast cancer had to have previously received trastuzumab and a taxane in the advanced or metastatic setting or were randomly assigned in a one-to-one fashion to T-DXd or T-DM1. I want to point out that patients were allowed on this trial as they had clinically stable treated brain metastases as well. About 20% of the patient population fell into that category. And stratification factors included the hormone receptor status, prior treatment with pertuzumab, which was not required that allowed, and a history of visceral disease. I would also call out the interim efficacy analysis for progression-free survival with the boundary of superiority of less than 0.000204 based on 245 events. And based on the IDMC's viewing of data, it was recommended to unblind the study at the end of July. So the PFS by blinded independent review was the primary analysis, and we'll go through that, including -- as well as the investigator assessment PFS. And I would just call out as well that at the time of this reporting, at the time that this study was started and for really the past 9 years, T-DM1 has been the gold standard in this second-line setting or in the setting after trastuzumab and taxane. It's really not been bumped. So these data are all that more important because of that. Next slide. Here is the patient disposition. There were 524 patients who were randomized to each of the arms. About 257 were treated on T-DXd, 261 treated with T-DM1. And you can see that the 132 are ongoing with T-DXd, only 47 with T-DM1. The median follow-up for T-DXd was 16.2 months and for T-DM1 was 15.3 months. Next slide. Here are the waterfall plot. So I just want to take a minute here and orient everyone to these plots. First of all, you can see that each one of the little columns represents one patient's best response on therapy. So for T-DXd, on the far left, you see that there are only a couple, maybe 4, 5 patients, whose best response was to have growth of their tumor. The rest of the patients enrolled and treated on a T-DXd arm had some level of shrinkage or disappearance of their disease. And in fact, the confirmed objective response rate with T-DXd was almost 80%. Keeping in mind, this is a study where patients were second line and beyond. These objective response rates are pretty phenomenal. When you compare that to T-DM1, it's 34%. And if you think of our current gold standard frontline treatment, meaning patients previously treatment naive from the CLEOPATRA study with THP, these numbers are comparative to those numbers in terms of objective response rate. Moreover, the complete response rate in this study was 16.1%, again, in the second-line setting or beyond. And if you compare that to CLEOPATRA, THP in the frontline setting where patients were primarily treatment naive, this compares very favorably to that. And of course, this study beats T-DM1. Next slide. These data speak for themselves and really not only are these impressive data and important data for our patients, but they are historic data now. I have never seen a hazard ratio progression-free survival of 0.28 in any paper in my career relating to breast cancer or historically. The separation of the curves that occurs with the first set of scans is equally impressive and stunning. The 12-month PFS rate at landmark was 76% of patients, and T-DXd were still progression-free at 12 months. Only 34% of patients treated with T-DM1 were progression-free at 12 months. The median PFS wasn't yet met for T-DXd by the blinded independent review. It was 6.8 months for T-DM1. Next slide. Now looking at these data by investigator assessment. First of all, it's reassuring the curves look similar. When you look at blinded independent radiology review and investigator assessment, there is a clear separation of the curves, a very similar hazard ratio, a very similar 12-month PFS rate for each of the arms. But now there is a median PFS that's estimated based on this assessment of around 25 months with T-DXd versus 7.2 with T-DM1. I would like to call out that this median PFS of 25 months in the second-line setting and beyond is better than what was seen in CLEOPATRA in the frontline setting where the median PFS was 18 months. 90% of the patients enrolled in CLEOPATRA had never previously received trastuzumab. One would expect that PFS to be even higher based on these data here, but this is not just any drug. This is really phenomenal PFS data that we are seeing here. Next slide. Overall survival is immature, of course. This is an early glimpse of the data but with very few events, just over 80 events having occurred. So it is trending in the right direction with T-DXd being in the lead. The 12-month overall survival rate with T-DXd was 94% compared to 86% with T-DM1. Hazard ratio, 0.56, very, very encouraging. However, the p-value, which does look incredibly impressive, did not meet that prespecified boundary of this analysis. These data are ongoing, and I would not be surprised if we're going to see overall survival meet significance when the data are more mature. Next slide. Drug-related treatment-emergent adverse events that are seen in 20% of patients or more are shown here. You can see the rates of Grade 3/4 neutropenia a bit higher with T-DXd 19% compared to 3%. But other TEAEs at the Grade 3/4 rate are fairly similar, a little bit more GI toxicity at Grade 3/4 with T-DXd. But generally, the drug has fairly similar outcomes when you compare to T-DM1. Alopecia is experienced more often in patients treated with T-DXd. I've seen that in my own patients. I've never seen full alopecia but certainly hair thinning. Next slide. Now this -- although the PFS and waterfall plot data are absolutely beautiful, this is probably the most important slide. Because if you have a drug with phenomenal, phenomenal activity, but it has phenomenal safety issues that are leading to a relatively high risk of patient deaths, then your therapeutic index is not going to be favorable, and it's going to be nearly impossible to take it into the curative setting where we have safe and effective therapies available. From DESTINY-Breast01, we had about a rate of close to about 15% of all grade interstitial lung disease with 2.7% of patients having Grade 5 events, which is death. We were all a little bit nervous. I myself was biting my nails, waiting to see this slide when I first saw the DESTINY-Breast03 data and was very relieved to see no Grade 4 or Grade 5 events with ILD, a lower rate of all grade ILD at 10.5% than previously reported on only 2 Grade 3 events. The protocol was written in a relatively aggressive manner to very aggressively mitigate or to diagnose and manage any patient even with asymptomatic, that means just imaging evidence of ILD. And in asymptomatic patients, treatment have to be held, and treatment was permanently discontinued with Grade 2 or greater ILD. So I think these -- this sort of very aggressive monitoring and management has led to the improved safety profile, which I think is very encouraging for us and supports the valuation of this agent in the early-stage setting. And then LVEF is shown at the bottom there, and you can see no Grade 3, 4 or 5 LVEF events, which is nice. There's numerically a few more events of Grade 2 with T-DXd, might be related to the fact that patients are being treated much longer with T-DXd, but again, very encouraging data. Next slide. So in conclusion, in the very first reported randomized trial evaluating T-DXd, in this case, compared to T-DM1, a highly clinically meaningful and statistically significant improvement in PFS was seen compared to T-DM1, which was the then standard as of yesterday or 2 days ago. There was a consistent benefit across key subgroups and efficacy endpoints, including patients with CNS metastases, and a confirmed objective response rate of close to 80% versus 34% with T-DM1 and complete responses in 16%. There was also an encouraging overall survival trend at this very first analysis, which is highly immature. And the safety profile is similar to what's been reported previously, although the rates of pneumonitis or ILD have improved with this large randomized trial. So these data, I believe, do support T-DXd becoming the standard of care for second-line HER2-positive metastatic breast cancer therapy. Next slide. And as my last slide, here is a picture of the treatment paradigm, the way we approach a patient who is diagnosed with metastatic HER2-positive breast cancer and the outcomes that we can expect based on the published evidence. CLEOPATRA remains our standard therapy in the frontline setting with a median PFS of just under 19 months. The T-DM1 pertuzumab study did not beat trastuzumab, taxane and is not considered a standard first-line therapy. T-DXd, which was tested in the second-line setting against T-DM1, the standard of care, has demonstrated a significantly improved progression-free survival with an investigator-assessed PFS of 25 months, which, as you can see here, is beating the data in the first-line setting, although we all know the caveats about cross-trial comparisons not being fair. And then now we have a number of agents available in the third-line setting and beyond. We have data to support the use of T-DXd in patients whose disease has been heavily pretreated from the DESTINY-Breast01 study. And I think the question comes up now is moving T-DXd up, certainly going head-to-head against THP. But I think that the safety data and efficacy data that have been demonstrated from DESTINY-Breast03 here strongly support the evaluation of this agent in the curative setting. Thanks very much.

Thank you very much, Dr. Hurvitz. I appreciate very much that review of the data. If we could go on to my next slide, and I'll seek to put into -- one more slide, please. I'll seek to put into contexts for -- in the real world and how we see this likely unfolding. So just by way of review, today, we're approved in 35 markets for the third-line-plus setting. U.S., Europe, Japan, U.K., Israel are among some of those. In each of the markets where we have launched, Enhertu has rapidly moved to a leadership role. In the U.S., we've got about a 35% to 40% share in the third-line-plus setting. And we've got now experience of over 7,000 patients that have been treated to date. It is still, I would say, a reasonably fragmented third-line-plus set of decision making. There are many options for patients with HER2-positive disease. And physicians intend and do cycle through their sets of options as they continue to work to try to find something for which patients will respond and continue to respond over time. We've been working diligently in the markets that we've been in around ILD management. I'm pleased to say that the ILD management in the real world has been, I would say, better on the whole through our awareness, monitoring and management than what we've seen in some of the earlier studies. That's why we're also particularly encouraged to see some of the same coming through in 03. With that as the backdrop, as we then look towards health authority conversations that we hope will result in approvals in a matter of single-digit months as opposed to anything longer than that, Dr. Hurvitz, I think, quite nicely articulated the unprecedented benefit and put it in the context of not only other second-line studies, but indeed, what we're seeing in frontline studies through multiple combinations and again, in studies that did not have a backdrop of adjuvant pertuzumab and sort of latest cutting-edge adjuvant therapies. Also really important consistency and efficacy across all subgroups. Some of the ones that were particularly noteworthy for us is irrespective of whether or not pertuzumab was used as a prior therapy. This is relevant as you think across the globe and the availability and incorporation of pertuzumab in previous lines, lines of therapy and CNS mets. And then also, the safety profile supports extended duration. I think just very, very quickly on this, Kadcyla, we know, is at a run rate of about $1 billion in sales for the first half of this year. We certainly see an opportunity that is very clear for Enhertu to quickly become the standard of care in second line. Kadcyla has about a 50% to 60% market share in the U.S. within the second-line context. The balance is coming from Herceptin and some TKIs. We really clearly see that there's an opportunity to very rapidly establish Enhertu within this setting. The median duration of treatment is 14 months in this study. We certainly see that, that's something that's likely to increase with time, and we'll certainly see that as something that's going to be an important part of understanding the utilization in the real world. And then lastly, as Dr. Hurvitz said, she pointed out that TEAEs was the most important slide. And the reason for that is that while today, we see the opportunity to change the standard of care in the early line -- or sorry, in the second-line setting, it is tomorrow and really this point that Dr. Hurvitz made about the possibility of potentially putting cure on the table for some women by being able to move into the earlier settings. And it's really the combination of the efficacy and the safety that allows us to do that. The program now moves towards how do we ensure that Enhertu is an option for every single patient with HER2-positive disease. With that, if you just quickly go to my next slide, we've got a program looking across the 3 major pillars. Susan will speak in some greater detail about this. The top pillar is obviously expanding in HER2-positive breast cancer. We have studies already underway in the frontline setting with D-B09, also in replacing T-DM1 in the postneoadjuvant setting with DESTINY-Breast05. The data presented on Saturday will certainly galvanize those efforts. We're awaiting for the DESTINY-Breast04 data in the second-line setting for HER2-low. We're expecting those in the first half of the year. Obviously, pending those results, it allows us to be able to understand if we also can move forward more rapidly in those settings. It's important to note also, remember, that this is a mixed population of both hormone receptor positive as well as triple-negative patients. And then finally, of course, going into other tumor types where HER2 positivity in overexpression and mutation can potentially play a role with gastric, lung and colon cancers. So with that, I'm going to turn it over to Susan, that's going to talk about where we also had exciting news not just with 03. Even though 03 was the bell of the ball, there was actually quite a lot of data that we had at ESMO.

Thank you, Dave. Can you please move on to the next slide again? And so that's Slide 23 for those of you following online. Okay. So again, as David alluded to, we had exciting data in other tumor types beyond breast cancer. So first on the left-hand side of this slide, the Phase II study, patients with HER2 mutant non-small cell lung cancer, which is a segment that's about 3% of non-small cell lung cancer that have a mutation. And when you have a mutation, you actually have a rapid internalization rate. And this, I think, is one of the reasons of the receptor on the self service. This is one of the reasons why I think we're seeing very robust activity within HER2 in this setting with a 55% response rate. And here, you can see the median overall survival of 17.8 months. This is in a segment that there's no targeted therapy is yet approved. So really exciting data in that setting. On the right-hand side, you can see the Phase II data from the DESTINY-Gastric02 study, which was a recapitulation of activity that has been seen in the Asian patient population for which we've already got approval in gastric cancer. And what's encouraging here is that we continue to see consistent activity in this western patient population with a 38% response rate and again, an impressive duration of response in that setting and good safety profile. The next slide, Slide 24, just gives you a little bit more detail about the extensive development plan that Dave has already highlighted. Again, just for clarity here, in the yellow boxes are the studies that we have in combination, while the maroon boxes or the studies we have in monotherapy, and you can see broken out by line of therapy in the different tumor types, breast cancer at the top, gastric cancer in the middle and lung, colorectal and other cancers at the bottom. Looking at those in that case, HER2-mutated as well as HER2 overexpressing. So you can see already that we've got an extensive development program. And I do think that there is great potential for the combination studies to further improve on what we've already seen as impressive monotherapy activity. And I'll draw your attention in a minute to some of the [indiscernible] data that was actually presented at ASCO earlier this year. Go to the next slide. And now I want to move on from Enhertu and just draw your attention to some of the data we presented on Imfinzi that I highlighted at the beginning. So firstly, on the left-hand side, real-world data from the PACIFIC setting in Stage III non-small cell lung cancer is unresectable. And again, very encouraging to see that in this PACIFIC-R study, the median progression-free survival in this setting was nearly 22 months, which compares very favorably with the trial data where we had around a 17-month median progression-free survival. And then on the right-hand side, again, here's the data from a randomized Phase II study, 189 patients randomized between 3 arms, looking at durvalumab versus durvalumab plus oleclumab, CD73 inhibitor, versus durvalumab plus monalizumab, the NKG2A inhibitor. Again, this was in a patient population following chemoradiotherapy. And again, one of the things that I think is important to highlight here is that this is a worst prognostic patient group that was enrolled in the study and was in the PACIFIC study. And that's based on tumor size, prior platinum, the number of -- the time period between the end of radiotherapy and the initiation of study. But what we did see is a substantial improvement in the landmark progression-free survival you can see highlighted here at 10 months from 39% in the durvalumab arm to 65% and 73% in the oleclumab and monalizumab combination arms, respectively. So this is very encouraging data, and we have already announced that we are moving these agents into Phase III trials. So just my final slide now is just to -- if you just move on one more slide, is to highlight that we have with Imfinzi now multiple examples of where we're looking at combinations that can extend the presence in the immuno-oncology space. Again, I'll show you the data from the BEGONIA study in triple-negative breast cancer looking at Enhertu plus durvalumab in that setting with an impressive response rate and good durability of response. I've shown you the COAST data. We've got the MEDIOLA study looking at the PARP inhibitor combination. The presiding data that was presented at World Congress on Lung Cancer, looking at the CTLA-4 combination in the non-small cell lung cancer setting. Now I would draw your attention to the fact we also have combinations with TIGIT ongoing in combination with Arcus Biosciences. We have a collaboration with domvanalimab in Phase III. That's the PACIFIC-8 study. And we also have -- we are starting Phase I studies with bispecific PD-1 TIGIT bispecific antibody AZD2936. And on the far right, you can also see that we've got combinations with our ATR inhibitor being studied in the HUDSON study, which is post checkpoint inhibitor in non-small cell lung cancer, where we're seeing an improvement in progression-free survival and overall survival for that cohort there. So really -- and we have an extensive program of ways in which we can improve on the activity we've already seen with Imfinzi. So I'm going to close here and hand over to Chris Sheldon so that we can move on to your questions. Chris, over to you.

Thank you, Susan. Now for the customary Q&A. And thank you, of course, to yourself, to Dave and our external guest, Senior Investigator, Dr. Sara Hurvitz. [Operator Instructions] With that, I'm going to hand back to Dave, who's going to triage the questions for us.

Thanks, Chris. So our first question is from Emmanuel Papadakis from Deutsche Bank.

Maybe I'll kick off with a question on D-B03, perhaps one for Dr. Hurvitz. CNS met baselines in this population of D-B03 was around 20%. Is that representative? And I'm asking in context of some interesting comments from the discussion she still viewed tucatinib as a preferable choice in patients with known CNS mets ahead of Enhertu despite the highly impressive D-B03 data. So I would love to hear your perspectives on that, whether you would agree with that assessment and whether you think it's going to be changed. And then perhaps a question on Dave, for you or one of the team. The updated perspective in terms of route to regulatory assessment approval in lung and colorectal. In lung, I mean both in the sense of the HER2 -- smaller HER2 mutated but also the overexpressing population. If you could just give us an update on where we are in terms of potential time lines improvement, that would be very helpful.

Thanks, Emmanuel. So Dr. Hurvitz, I -- why don't I suggest that you can answer the first part of that question. And then, Sunil, if you could also maybe expand by the work that we're doing in brain mets to build on top of the D-B03 data, that would be great. And then we'll move on to the next question. Dr. Hurvitz, to you.

Yes. So if you saw the full slide presentation, the patients who had brain metastases benefited similarly from T-DXd -- similarly to patients who did not have a history of brain metastases. The median PFS was about 15 months in patients with brain metastasis with T-DXd and under 6 months in those treated with T-DM1. This study is unique compared to the tucatinib HER2CLIMB study where patients had all been previously treated with T-DM1. So there are different patient populations we're talking about with the tucatinib data being in a T-DM1-naive patient population -- or T-DM1 pretreated population, T-DXd -- our DESTINY study being in patients who had previously received -- not received T-DM1. I think the cop-out answer is for me to say it's always a patient-to-patient decision. When I have a patient sitting before me who has a couple of brain metastases and visceral disease that I need to get under control, I am going to be very compelled by the T-DXd data and give that treatment to them over tucatinib because I still am able to use stereotactic radio surgery to manage small volume disease in the brain. If, however, I have a patient who's already had stereotactic radiosurgery, has multiple metastases in the brain and relatively light extracranial disease, I may go for tucatinib in that particular situation, given it's compelling activity from that large randomized study in the brain area, that we actually see what the responses are in the intracranial area with tucatinib from that study. So I think it really is patient-by-patient decision-making, but we are talking about different patient population studied in those 2 trials. I'm looking forward to data AZ may not be interested, but there is an ongoing study of T-DXd with tucatinib, and maybe that's going to be a therapeutic combination that people will go for in those patients who have CNS metastases. But I just think that these PFS and early OS data, combined with the response rate data, are so compelling for T-DXd. It's nothing like what we saw with the tucatinib base combination. So I would probably lean more toward T-DXd.

Thanks, Dr. Hurvitz. Sunil, any kind of other areas of context, but then also expanding on what we're doing around evidence generation in the brain mets population?

Thank you, Sara. So 2 points. So one of all is the -- putting into context the 15-month PFS that Sara shared with us in this trial in patients with stable and brain mets is the highest reported PFS we have seen in patients with brain mets. The tucatinib combinations with chemotherapy and trastuzumab, as per HER2CLIMB, the CNS PFS was in the 7.6-month range and again in the T-DM1 pretreated. So I think this gives us a strong rationale to consider this in earlier line. And secondly, we recognize that there is an opportunity for learning more going beyond stable brain mets to look at patients who have progressing brain mets. So the TUXEDO single-arm study was shared at the ESMO meeting. This was a investigator-initiated trial where they looked at patients who were pretreated. And 5 of the 6 patients who were pretreated receiving T-DXd had intracranial response or response rate of 86%, and we are generating that data within our program as well. So we have high confidence that this benefit will extend beyond the stable brain mets to active brain mets, including seeing intracranial responses.

Thank you, Sunil. Emmanuel's second question was around regulatory path and health authority interactions for the lung program. And also probably worth also just making a quick comment on just with all the studies that we've seen. Cristian, please?

I think we started the conversation, of course, for all these regulatory interaction and submission. You can imagine that for D-B03, let me start with D-B03, this is -- we are seeking for RTOR with FDA and Orbis, of course, to give -- the opportunity to give access to this drug as many countries as we can as soon as possible. There is a lot of interest in this preliminary conversation we had. And together with Daiichi Sankyo, we started this interaction. For lung cancer, the data are quite solid. And also, we started this conversation, and we will update you based on the evolving discussion [indiscernible].

Thank you, Cristian. Now we'll move to the next question from Simon Baker at Redburn. Simon?

Firstly, for Dr. Hurvitz, just really continuing on from Emmanuel's question. You highlighted one area where you may use tucatinib instead of Enhertu. I just wonder if you could give us other areas where you would not use Enhertu given the overwhelming overall benefit that it's shown over T-DM1. And then another question, really on trying to get to why it was so good. And I just wondered if you had any mechanistic data on what exactly is driving this vastly superior performance and how that reads into your levels of confidence for Breast09 versus pertuzumab and 06 Enhertu low.

Excellent. So again, Dr. Hurvitz, if you can offer your thoughts on the first part of Simon's question in terms of kind of who are the patients in that second-line setting that you might think to use something else. And then Susan can take the question on mechanism.

Absolutely. So really, the only patient where I would not turn to T-DXd is a patient who has underlying significant pulmonary fibrosis, is on steroids for an underlying interstitial lung disease, which I must say is fairly rare. I would also just add to that, that the presence of pleural carcinomatosis or lung metastases themselves does not appear to be predictive of the development of ILD. There is a retrospective data with T-DXd indicating that presence or absence lung metastases does not correlate with the development of ILD. And indeed, patients with lung metastases were enrolled on DESTINY-Breast03 and DESTINY-Breast01 and had responses. My own patients had lung metastases and had very nice responses in the lung. So that wouldn't preclude me from using it, but the presence of a clinically significant interstitial lung disease would make me nervous about using this agent. Outside of that, I don't think we've seen this type of activity with any other agent. And my feeling is with the profound PFS and early OS benefits we are seeing, I'm going to want to use this agent as early as possible.

Okay. Thank you. So let me take the second part of the question about the mechanism. And again, I just want to point out that the drug was designed by the Daiichi Sankyo protein engineers and chemists, and they did a fabulous job in the design of the molecule. So we know the antibody works because of just trastuzumab and T-DM1. What is different about this molecule is that you've got a stable but cleavable linker that enables a bystander effect. And we know that heterogeneity in the macro environment about the level of HER2 expression, even in HER2 high population, can affect the efficacy across the tumor. And the progression-free survival is driven by the parts of the tumor that behave the least well, if you like. So overcoming heterogeneity, I think, is a really important part of the design of the molecule. Secondly, you have a different orthogonal cytotoxic mechanism in the warhead. And again, remember that it's in a post-taxane situation for most of these patients. So with a topoisomerase warhead, you aren't going to get cell kill without any resistance to the inherent taxane mechanism. And then you've got a high drug antibody ratio. So for every molecule that's delivered, you're delivering a relatively high dose, and that was the differential therapeutic index. So it's fundamentally understanding of the biology and a better design of the molecule that's driving the mechanism.

Thank you, Susan. Okay. Now next question is Andrew Baum at Citi.

A couple of question. Should be fast. First, could you talk to, and I appreciate it's early days, the mechanisms behind acquired resistance to Enhertu, given that's been somewhat done with acquired resistance to T-DM1, thinking about the comments of potentially whether it could be curative in the frontline? Second, I'd be interested in Dr. Hurvitz' view on whether pertuzumab is actually going to be additive to Enhertu in the first-line DESTINY-09 trial, whether you think it's going to be redundant given the data you've seen? And then finally, thinking about DESTINY-04, low HER2. I was struck by the fact that you reported responses in your non-small cell lung cancer trial in patients who didn't have HER2 detected by IHC. I'm assuming that's because there may be heterogeneity of expression of HER2 among those patients even if it wasn't detected on the biopsy. Should we believe that among the low HER2 patients, there may be cells with high HER2, and by binding Enhertu and plus the bystander effect, that can drive the treatment benefit? So how important do you think is heterogenetic expression among defined low HER2 patients?

Super. Susan, can I suggest that you start with the acquired resistance, Dr. Hurvitz on pertuzumab's role in frontline, and Sunil, the D-BO4 question that was asked?

Okay. So at the moment, we don't yet have a good enough understanding of the mechanisms of acquired resistance. This is obviously something that's very important, and we will be looking at that in more detail. Again, you can speculate about what they might be. So you might have resistance due to a lower expression of the target, some inherent resistance to the mechanism of action of the warhead. But we need to characterize those and we will be doing that during the course of the program.

With respect to the question on pertuzumab and whether it's really necessary, I am -- perhaps I don't represent all oncologists out there, but I've been, over the years, a little bit dubious about the amazing effects seen in the CLEOPATRA study in that delta between the THP and TH arm. When we look at APHINITY adding a year of pertuzumab to trastuzumab is really marginally beneficial to patients and really only those patients who have node-positive disease for the highest risk patients. So the synergy between pertuzumab and trastuzumab or pertuzumab with other HER2-targeted therapies, I think we still need to interrogate that question regarding whether or not it's necessary. I think it's an important question to look at, though, whether we can extend the benefits of T-DXd in the curative setting. I'm not concerned about it adding much in the way of toxicity. So I'm not worried that the therapeutic index would change by doing a combination strategy. But whether or not it's absolutely necessary, the data will have to show that. I'm not sure that we're going to see that.

Thanks, Dr. Hurvitz. Thank you, Sara. So just building off on pertuzumab. So in DESTINY-Breast09, we have the 3 arms. So we have monotherapy and monotherapy and pertuzumab versus THP. So I think that cushion is going to be answered with the additional value of pertuzumab. And coming to the third question you had, Andrew, about HER2-low in heterogeneity, it is a big cushion, right? So we have high confidence that in HER2-low breast cancer, based on the Phase I trial and activity that we've seen, and of course, other investigator-initiated trials that are ongoing, that there is activity of Enhertu in those patient populations. And important to note, with chemotherapy alone, we get the subpar performance in second, third line in DESTINY-Breast04 with a PFS of only 3 to 4 months and a response rate at less than 20%. What we have seen in the Phase I is a response rate that is in the mid-30% to low-40% range. So we have confidence. Now is that driven by HER2 expression, HER2-low? Is that HER2 heterogeneity? Is it sort of a mix of both? We will certainly be doing the translational work to better understand that and also understanding how low can we go with HER2 expression from DESTINY-Breast04.

Thank you. Thank you, Sunil. Let's move on to the next question, Jo Walton at Credit Suisse.

My 2 questions are related to safety. I'd be intrigued with Dr. Hurvitz' view as to how early you would be able to go. Interstitial lung disease has been managed by a very aggressive protocol. In a real-world setting where patients are in an adjuvant setting, do you think it's actually possible to use this? Because -- will patients come back and be scanned so they can be removed before any symptoms develop? And can I also ask about the safety in the COAST study? Imfinzi plus the other 2 agents, fantastic results, but is there any safety offset?

Great. Maybe I'll suggest, before Dr. Hurvitz, you answer the question on the adjuvants, Sunil, can you comment on the ILD protocol just in terms of Jo had referred to, it is aggressive? But I think we ought to actually just -- let's state what it actually is. And then Dr. Hurvitz, I think you can comment on the question about kind of how do you see that as you move early. And then suggest that maybe we move to Cristian on the COAST question, the safety within that.

So pneumonitis is seen very commonly with many breast cancer therapies, mTOR inhibitors, CDK4/6 inhibitors, immune therapies. So it's not an unusual side effect. What we did was recognize that this is happening and then implemented 3 things. One was education for the [ sites ] investigators that this is something associated for them to be aware. Secondly, we put in guidance related to what is the exact steroid management that's needed so that there is standardization on how best to manage if pneumonitis was to occur. And the third is to, as Dr. Hurvitz mentioned, ensure that patients who have ILD previously not to receive this therapy. And then the management itself and the recognition is as per standard guidelines. So I think all of those are implementable and always recognizable by the clinical community. Now they have the tools to be able to watch and monitor patients. So I wouldn't really term that aggressive. I would say it's information that's useful and awareness for the clinical community.

Sara, please go ahead.

Yes. So in the adjuvant setting, when you're going to be receiving an IV medication, you will be having frequent visits with the clinician, the physician or advanced practitioner will be meeting with the patient regularly. This would not be a therapy like endocrine therapy where patients would be on it for 5 to 10 years, and there would be a risk of losing touch with them and what symptoms they're having. During every 3-week IV therapy, you're going to have a very close monitoring of the patients enabled by the frequent visits to the infusion room into the clinic. So I don't think that it's overly burdensome to patients or clinicians to monitor for symptoms. I don't think we need to be scanning patients per se. I think we just need to be in tune to new symptoms and side effects that they're experiencing. And if a patient were to develop shortness of breath or cough that wasn't associated with the simple cold, for example, that it'd be evaluated and managed immediately.

And Jo, with regards to the question on COAST, that in addition to the excellent outcome in terms of PFS that you have seen from the study, so in terms of efficacy, we were actually very pleased on the safety findings. Because when you look at the overall grade that we observed across the 3 arms, there was not a difference. Fundamentally adding both monalizumab or oleclumab on top of durvalumab does not seem to increase what we would expect with durvalumab alone. There was some numerically increase in early grades, Grade 1 and 2 for rash. But when you look at the adverse event of specific interest, specifically for durvalumab, nothing is standing out, especially the severe events that you would have a look at like [indiscernible] or other more events like hypothyroid, hyperthyroid. So these events, we are quite balanced with the 3 arms. These give us confidence that these doublets actually can be moved even in a larger setting. And as we announced, we are now testing them in the context of a Phase III setting.

Thank you, Cristian. Next question from Seamus Fernandez at Guggenheim.

Great. So I think a number of my questions have been answered on Enhertu. I wanted to actually maybe take a step back. We haven't had an opportunity to go through some of the data presented at World Lung. So I'm going to take a little bit of a right turn here and go to the 1062 data. Dave, really just trying to get a better sense of how the team is thinking about the data presented at World Lung. I know this was follow-up data. And the response rate in the treatment refractory lung setting really does imply a response that was as robust as we've seen with second-line PD-1's postchemotherapy head-to-head versus chemotherapy. So just love to know how you're thinking about that opportunity for 1062. And then separately for Susan, just hoping you could help us understand a little bit better the results from the COAST study. I think some of us were a bit confused by the poor performance or the seemingly poor performance of durvalumab in the control arm. Just wanted to get a better sense of that if it really was baseline characteristics and if those baseline characteristics were balanced across the other arms such that we're quite confident that the stepped-up response that we're seeing is likely to play through as you move into Phase III.

Cristian, do you want to start first on 1062 and then we'll move to Susan for the question on COAST?

The data that you've seen in World Lung are summarizing what we presented and confirming with a longer follow-up of what we presented at ASCO. And in, specifically, the population without tumor drivers, so you see that the level of response rate, the duration of response and median PFS is holding specifically at 6-milligram dose. Of course, the data will be constantly revised and updated. But this is the basis on -- of our current ongoing Phase III studies in second line in patient that failed checkpoint inhibitors and chemotherapy versus docetaxel. So I think this is recapitulated with this update. Then I think the new piece of data that we shared are actually in those patients that are very heavily pretreated and have tumors occurring driver mutations, eGFR, ALK and other driver mutations. These are patients that usually failed, most of them -- all of them, TKIs and chemotherapy. And I think that T-DXd showing in this -- also in this population, quite preliminary but very interesting activity in terms of response rate and in terms of duration of response. So this is another pillar for the development of Dato-DXd that is quite important. And actually, there is an ongoing trial, a large Phase II, in this specific population of non-small cell lung cancer pretreated patients that eventually can provide for Dato-DXd and upside in non-small cell lung cancer in addition to the ongoing program and, of course, the plans that we have in combination with checkpoint inhibitors.

So thanks, Seamus, for the question about COAST. If I could just ask you to show Slide 33 from the deck that we presented. I just want to go through the -- we did this propensity matching because it's a very important question that you asked. And of course, we asked it ourselves when we first saw the data, which is how does the performance of durvalumab in the COAST study compare with what you have expected if you were matched with the prognostic variables. So you can see here that the variables that we matched on were age, race, prior therapy, whether it was carbo or cisplatin. Very importantly, the time from the last radiation to randomization, whether that's less than or more than 14 days, which we found in the PACIFIC study to be an important [indiscernible] factor, best response to prior therapies and the disease stage at entry, IIIA, IIIB or IIIC. And what you can see is, first of all, that if you look at the numbers underneath, the data are only really mature to around the first 10 to 12 months, which is why we presented the 10-month landmark analysis. And if you look at this, you can see that actually the durva arm in the COAST study is roughly in line with expectations with a similar landmark analysis at PFS and a similar response rate. Yes, we did look at whether or not there was any imbalance of these factors across the other arms, and we're satisfied that the effect size that we're seeing was robust even when taking account of all of these factors. That's an important question. Thank you.

Thank you, Susan. And with the slide, let's see. What do we have next? Yes. Luisa Hector from Berenberg.

Thank you, Dave, and thank you, everyone, in making data. Well done. So I wanted to go back to D-B03. So Dave, you mentioned this, the treatment duration of 14 months, quite a lot shorter than the PFS. So I'm just wondering if you can discuss the reason for stopping treatment. And Dave, you said you hope that will increase with time, the duration. So why is that? Is that linked to the safety monitoring? So just some clarity around that. And what would the treatment duration expectation be in the first-line study? And then just whether you have any sense of who these patients are with the complete responses or the patients who do experience the ILD despite -- I don't think there are any biomarkers or whether you have any views on who these patients are.

Perfect. Thanks, Luisa, for the question. I'm going to ask Sunil for you to comment on this. And then Dr. Hurvitz will also take the opportunity. I know that we have you until the top of the hour, which is about 2 more minutes. So just let me take the opportunity to thank you so very much for joining us. I trust you'll drop off when you need to. We will stay on for an extra 10 minutes because we've got a half dozen more questions that we're going to endeavor to get through. But again, thank you, Dr. Hurvitz, very much for joining us. Just -- so again, on D-B03, maybe, Luisa, the one thing that I'd offer on this, remember that this is a very early look with 14 months of follow-up. And so actually, I think that it's important to have that as we take into consideration where -- how immature the median [indiscernible]. With that, though, Sunil, please?

Yes. No, thank you. And thank you, Luisa. So you're absolutely right. Given the median duration of study is only 15 or so months, that's why we're seeing a median duration of treatment that is also very short. I think with longer duration of follow-up, we will see that to be much longer. Off note, in DESTINY-Breast01, our single-arm study, many patients, we had more than 20% in a very late-line setting, patients who are on treatment for more than 2 years and many more than 3 years. So patients can stay on treatment. And I think we will see a more mature median duration of treatment when we have a more mature duration of follow-up in the study. The 3 other questions that you asked is, do we know the risk factors of ILD? So this is the first randomized data set that we have. So certainly, we'll be looking at that. As Dr. Hurvitz shared with us, the frequency of ILDs is lower. We do know that majority of ILD, when it occurs, occurs within the first year. So with longer follow-up, we don't expect that ILD rate to substantially change, but we will learn more about potential risk factors, and we'll be certainly looking at that. The second part was complete responders, who they are. We did note that earlier we go, the higher the rate of complete response. And we noted that from DESTINY-Breast01. And we'll be similarly looking at that from DESTINY-Breast03 to see, is there a response rate -- complete response rate that is even higher for some of those patients who entered the study in the first line and those 50% of patients in the second-line setting. And finally, we also need to figure out who are those 3 patients who did not have a response and who had progressive disease. As was noted, there were 3 patients who had PD, and we need to know why this drug doesn't work. Because so far, it works for nearly everyone.

Thanks, Sunil. Next, we'll move on to Mattias Häggblom at Handelsbanken.

Two questions, please. Firstly, on the issue of ILD that 26% of patients experienced in DESTINY-Lung01. Help us understand to what extent do you think that too can be better handled with a more aggressive look at protocol that Dr. Hurvitz referenced in the DESTINY-Breast03. Or given the indication, should this be seen as a nontarget effect that we need to expect to remain at a higher rate compared with, for instance, breast? Secondly, back to COAST. On your in-house MedImmune drug antibody, oleclumab, targeting CD73, talk about how you think it's differentiated from competition, but also include small molecule inhibitors.

Thanks, Mattias. Again, I think, Sunil, let's go to you on the ILD question. Again, that 26%, just as a reminder, the majority of that is Grade 1, Grade 2. 0.8% is Grade 3. I think that's quite important in the context of the severity of what's being managed.

So just to sort of confirm, so there's 2 different data sets. In DESTINY-Breast03, the ILD rate was only 10%. I think if you're mentioning DESTINY-Lung, then certainly, there was a higher rate of ILD. And how much of that was dose related and could have contributed towards that, I think, is something that we need to address as we're learning more about there may be a dose relationship at least within lung cancer with ILD rates. And breast cancer at the 5.4 milligram, we are hovering around 10%, and we'll gain majority of that within Grade 1, Grade 2.

Susan, do you want to comment on the differentiation point on the CD73?

Yes. Well, the CD73 antibody is highly specific, as you would expect, and potent. And we have very good receptor occupancy at the dose that we're moving forward into the Phase III. So I think it's -- the safety profile, as Cristian has already alluded to, is clean. We've got potent inhibition of the target and a good safety profile. So I think -- and we're the only ones with data in this setting in combination with the PACIFIC data. So that's what I would have to say about that.

Great. Thanks, Susan. I think actually, Mattias, one of the things that I think you might have done, I could be wrong, I'm just doing this based on the maths that I do, it's possible that the percentages, you were looking at the number of events in the T-DXd arm, which gets to that 26 events. Actually, it's 25 events. But anyway, to Sunil's point, the -- it was 2.7% Grade 1, 7% Grade 2, 0.8% for Grade 3, and no Grade 4s and Grade 5s, for a total of 10.5 for any grade. So with that, Richard Parkes, over to you.

Congratulations on some incredible datas, growth for patients as well. So 2 questions. Firstly, on -- I wondered if you could discuss the selection of the drug antibody ratio with Dato-DXd, which I think is slightly lower than with trastuzumab drugs to count at 4:1 versus 8:1, given the high ratio for Enhertu speculated partly account for its compelling efficacy. I wonder why that was chosen. Was that just the ceiling of what could be achieved? Or was there another reason for that being selected? And then secondly, there's been an interesting study presented on rechallenge with PARP inhibitor in olaparib at ESMO. And I wondered if you could talk about your interpretation of that data and whether there are any plans to conduct clinical studies in retreatment with Lynparza.

Super. Cristian, do you want to address both of those questions in terms of Lynparza retreatment as well as the drug antibody ratio for Dato-DX?

Yes. Let me start with the first one. I think the target is very different. HER2 is a different target than TROP2. And TROP2 is sometimes stressed also in normal tissues. So actually, in several years that there are efforts to try to develop TROP2 ADCs and toxicity, there's always been a major limiting step in finding the right therapeutic window. I think Daiichi Sankyo, when developed Dato-DXd, tried to find the right balance between safety, toxicity. And this data is coming from that specific balance to be kept. That said, I think Dato-DXd safety and the dose that's been selected to move in Phase III is quite manageable and acceptable. And I think also the preliminary data that we have across indication in non-small cell lung cancer and breast cancer, specifically, triple-negative breast cancer, are supportive of a very interesting and exciting preliminary activity. In terms of the second question, Lynparza, this is a rechallenge study that was still an open question. Okay, what can we achieve if we give back olaparib PARP inhibitor to patients that have already been exposed to a PARP inhibitor? I think the OReO/ENGOT study is telling us that actually, this is a strategy that can be pursued in the clinical practice. Because what is emerging from the study in this patient population that was heavily pretreated is that when you rechallenge olaparib, either in patient with BRCA -- with tumors with BRCA mutation or without BRCA mutations, you have a very important reduction in terms of disease progression. The other ratio are quite relevant, a 43% reduction in BRCA mutant cohort and a 47% reduction in non-BRCA mutant cohort. And even more interesting is the exploratory analysis in non-BRCA mutant cohort, in which you have the same benefit in HRD-positive tumor, so in HRD-negative tumors, the median PFS is similar, 5.3 months versus 2.8 in HRD-positive and 5.4 versus 2.8 in HRD-negative tumors. I think this data set is telling us that, yes, this is a strategy that eventually can be put in place for those few patients that still did not receive olaparib as maintenance treatment after the first line.

Thanks, Cristian. We've got 3 more questions. If you all can ask one, we'll endeavor to get through all 3 of them and make it happen quickly. Peter Welford at Jefferies; then we'll go to Christopher Uhde at SEB; and then Steve Scala at Cowen.

Let me go back to COAST then, if I can. So just curious, when you look into the details, are there any differences between the patient groups that could tease out for the Phase III's oleclumab versus monalizumab combinations or is the strategy here to produce to, I guess, twinned Phase III trials and let the best man win, so to speak? So curious, really, how you think about developing both those combinations in the future of combinations with Imfinzi.

So I'm happy to take that question. Thank you. So one of the challenges that we've got from the COAST study is that we haven't got baseline biopsies from all of the patients to really understand in depth the patient selection criteria for the different molecules. In addition, both these targets are -- have increased expression of CD73 and NKG to increased expression in the presence of radiation treatment. And so even a baseline diagnostic biopsy might not tell you the expression level that's actually likely to predict for a response in this particular setting. We will get more data from the neo COAST study. And what we are encouraged by is that we're seeing activity across different PD-L1 backgrounds, even though the data were more limited. We are certainly seeing activity for these combinations in the PD-L1 low setting as well. So there's more we're going to learn as we develop the agents, and we are going to do another neoadjuvant study, which will help answer some of the biomarker selection patient questions that you have. But the intent is to go into Phase III trials with both, and we will see, obviously, how they'll perform in Phase III.

Thank you, Susan. Uhde, please.

So isn't HER2 in the protocol for the D-B09 as a potential subsequent therapy in the comparator arm? And then very quickly, in HER2 NSCLC, the initial data had PFS at roughly 2x as long as the cut this year. Should we ascribe that to the limitations of small sample size? Or are other factors at play? Is HER2-low contributing?

Sunil, would you like to...

Yes. For D-B09, we do not have HER2 sequential post THP progression.

Okay. And then Susan or Cristian on the non-small cell question?

I mean do you want to -- I'll go ahead, Susan. Eventually, you can chime in. I think the cohort to DESTINY-Lung01 and HER2-mutant cohort that we presented, you refer the lung cancer cohort, correct?

Yes.

Yes.

Yes. Okay. I think the data are quite consistent with the first report. Of course, with longer follow-up, you can have some fine-tuning of the data. But 55% response rate with the median duration of 9 months and PFS of 8.2 months in this setting, considering how pretreated are -- these patients are quite solid and unprecedented level of activity in these HER2-mutant patient population and not reported so far by any other targeting and on targeted therapy. So we believe this data still represent really outlier data in this specific operation. This is why we will continue to develop the drug in this specific setting, in an earlier setting, of course.

Thank you, Cristian. Finally, Steve Scala from Cowen. Please, Steve.

Just a simple question. How will AstraZeneca market the POSEIDON regimen? And what will you point to as differentiation?

So Steve, maybe I'll take that one to close. So obviously, the POSEIDON data is a data set where in an all-comers population, we saw the benefit within the study. At the same time, we know that there are many patients despite the advances that have been made with IO-chemo-based therapies who do not get response and don't get adequate response. We think that there's some subsets within the forest plots that are particularly interesting. Certainly, patients that are PD-L1 less than 1s is a group where we don't see as much benefit from the IO-chemo regimen and certainly not from an IO-mono regimen. We think that there's opportunities to explore a triplet here. Steve, I think the other thing I'd say is that we already today see somewhere between 5% and 10% share of CTLA-4-based regimens in that lung cancer setting. So clearly, there's also people voting with their feet that there's a need in this population. So step 1 is health authority interactions. And then step 2 is just as I had said and working through it. So with that, we're really deeply appreciative of everybody joining on this call. I want to just take 10 seconds to thank and acknowledge our great partners at Daiichi Sankyo. This is a drug that they discovered. We get the pleasure to codevelop it with them, but certainly, D-B03 is a great example of courage and vision. And it's a pleasure to be able to have an opportunity to discuss it today. We hope that this is just the first or maybe second chapter of what will be a very, very important book that gets written on Enhertu. And then also, we're really pleased that we had an opportunity to talk about the entire AstraZeneca portfolio or at least elements of it throughout the day. I want to thank Dr. Hurvitz, my colleagues at AstraZeneca, and thank you all very, very much for the time on this call. Good evening. Take care.