Summit Therapeutics Inc. (SMMT) Earnings Call Transcript & Summary

Thank you for standing by. My name is Tina, and I will be your conference operator today. At this time, I would like to welcome everyone to the Summit Therapeutics Update Call. [Operator Instructions] It is now my pleasure to turn today's call over to Dave Gancarz, Chief Business and Strategy Officer. Dave, please go ahead.

Good morning and thank you for joining us. We issued a press release yesterday morning relating to the Phase III HARMONi data featuring ivonescimab presented as a part of the Presidential Symposium at the International Association for the Study of Lung Cancer 2025 World Conference on Lung Cancer, otherwise known as World Lung or WCLC 2025. The press release is available on our website, www.smmttx.com. Our 8-K was filed today and is available on our website and via the SEC's website. Today's call is being simultaneously webcast, and an archived replay will also be made available later today on our website. Joining me on the call today is Bob Duggan, our Chairman of the Board and Co-Chief Executive Officer; Dr. Mahkam Zanganeh, our Co-Chief Executive Officer and President; Manmeet Soni, our Chief Operating Officer and Chief Financial Officer; Dr. Urte Gayko, our Chief Regulatory Quality and Safety Officer; Dr. Allen Yang; and Dr. Jack West, our VP of Clinical Development. Before we get started with the rest of the call, I would like to note that some statements made by our management team and some responses to questions that we make today may be considered forward-looking statements based on our current expectations. Summit cautions that these forward-looking statements are subject to risks and uncertainties that may cause actual results to differ materially from those indicated in the forward-looking statements. Please refer to our SEC filings for information about these risks and uncertainties. Summit undertakes no obligation to update these forward-looking statements, except as required by law. One item of note, this presentation is being webcast with slides, we'll be referring to the slides being displayed in the webcast link. I'd encourage you to use the website link to see the slides being presented this morning or otherwise to take a look at the slides that are on our website. Following comments from our team, we will take questions. Before I hand it over to Jack to walk through the details of the presentation, I'd like to give a little bit of an overview based on presentation yesterday and the details in our press release. HARMONi is a multiregional study, evaluating ivonescimab plus chemotherapy against standard of care chemotherapy alone. This study represented a continuation into a multiregional study -- into a multiregional setting of the original single-region HARMONi 8 study that was performed by our partners at Akeso in China. We in-licensed ivonescimab effective January 2023 and immediately went to the U.S. FDA in order to discuss bringing the open study from a single region to a multiregional setting. Once aligned, we began enrolling in mid-2023 and completed enrollment in the beginning of October 2024. This resulted in sequential enrollment of patients first in China and then in the Western regions of North America and Europe. As a result, at the time of our prespecified primary analyses of both primary endpoints, PFS and OS, the majority of events were driven by China. Let's discuss the timing of the analysis performance. The primary PFS analysis was performed in July 2024 as per the protocol as a prespecified number of events. Due to the sequential nature of the enrollment, as I mentioned, these events were reached prior to completing enrollment and hence, there are fewer patients in that analysis in the 438 patients noted in the chemo, mostly coming from Asia. As a result, we performed a total PFS analysis when we perform the primary overall survival analysis in April 2025. This allows for the inclusion of all Western patients and more maturity of these patients, roughly 9 months of median follow-up time to evaluate the global view or multiregional view of progression-free survival in the multiregional set. However, for the primary OS analysis taking place in April 2025, which we noted in our press release, the prespecified number of events was reached as per our protocol, but the follow-up time for Western patients, again, that was approximately 9 months was not mature as it was far short of the median survival in either arm. We, therefore, noted at that time in our May press release, that we would continue to follow Western patients for additional time on study. We launched the database in September 2025 when there was a longer follow-up and 76 events in Western patients. The number of OS events in Western patients at this point in time was in alignment with the estimation of events for Western patients expected to be included in the primary analysis for our protocol, and to be included in yesterday's presentation at WCLC 2025, the updated OS analysis that we just completed that included a longer-term follow-up of Western patients. When we in-licensed ivonescimab in the beginning of 2023, we sought to bring ivonescimab to patients, one as quickly as possible; and two, in particular, with respect to the setting those patients with EGFR mutant non-small cell lung cancer who progressed after targeted therapy where there are very limited options and many drugs have failed to show a benefit here. We sought to bring an opportunity for those patients to have an additional therapy option. As I mentioned and as we've discussed over the past 2.5 years, the setup of this study was a bit unique in continuing a single reason study into a multiregional study. And therefore, the follow-up post-hoc analysis are uniquely critical to evaluating the study with sufficient time line study for Western patients. So as Jack walked through the details of the presentation, we felt it would be appropriate to provide this background and understanding prior to Jack's taking you through the results. And with that, I'll hand it over to Jack to walk through the presentation.

Thank you, Dave. Yesterday, as you're aware, ivonescimab data were featured in a presentation at WCLC 2025 as part of the presidential symposium for the second straight year. The presentation titled ivonescimab versus placebo plus chemo, Phase III in patients with EGFR-positive non-small cell lung cancer progressed with third-generation EGFR TKI treatment, HARMONi. It was given by Dr. Jonathan Goldman MD, Professor of Medicine at UCLA in the Hematology Oncology Division as the UCLA Director of Clinical Trials in Thoracic Oncology and also Associate Director of Early Drug Development, and Chair of the University of California Lung Cancer Consortium. Revisiting the schema for the HARMONi trial. HARMONi evaluated ivonescimab in combination with platinum double chemotherapy compared to placebo plus platinum double chemotherapy in patients with epidermal growth factor receptor for EGFR-mutated locally advanced or metastatic non-small cell lung cancer, nonsquamous non-small cell lung cancer, who has progressed after treatment with a third-generation EGFR tyrosine kinase inhibitor, or TKI. This was a global Phase III clinical trial conducted and sponsored by Summit for which top line results were announced in May of this year. Key eligibility criteria are shown here. Patients were randomized 1:1 and stratified by geographic region and presence of brain metastases at baseline. Patients received either ivonescimab at 20 milligrams per kilogram plus pemetrexed carboplatin or placebo plus pemetrexed carboplatin, then received ivonescimab or placebo in combination with pemetrexed for up to 24 months. Treatment was to be discontinued for intolerability, progressive disease, or initiation of new antitumor therapy. The study had dual primary endpoints of overall survival and progression-free survival by Independent Radiologic Review Committee. Secondary endpoints included response rate, duration of response and safety. The dual primary end point is important. As you know, the study was a positive study based on achieving its PFS primary endpoint. The baseline characteristics showed ballast arms for age, gender, region, race, smoking status and the presence or absence of liver and brain metastases. Patients in this study received a prior third-generation TKI either in the frontline or after an early generation agent. Approximately 60% had a deletion 19 mutation, 35% had L858R and about 5% had a noncanonical mutation. Brain metastases were present at baseline and 25% of the patients on HARMONi. The trial included a total of 438 patients, including 165 from North America or Europe, representing 38% of patients in the study. In several EGFR mutation positive global Phase III lung cancer study, over half of the patients who are randomized coming from Asian territories. And so this enrollment breakdown is in line with other multiregional studies of this tumor type. The progression-free survival curves show the trial met the primary endpoint with a hazard ratio of 0.52 with a corresponding p-value of less than 0.0001. Median progression-free survival was 6.8 months for the ivonescimab with chemotherapy arm compared to 4.4 months for those patients receiving chemotherapy alone, a difference of 2.4 months favoring the ivonescimab arm. At 6 and 12 months, almost 20% more patients taking ivonescimab in combination with chemotherapy were progression-free. When we look at various subgroups, it's important to highlight that the size of the subgroups is smaller. And by definition, it's not designed to be statistically significant on their own, but the subgroup analyses can be very informative, nonetheless. The subgroup analysis for progression-free survival confirmed benefit in all preplanned subsets as indicated by plots for each subgroup landing on the left side of the dividing line favoring ivonescimab. So there's really not a situation where one subgroup is an outlier of the study results were driven by a specific subset or a subset of patients. Ivonescimab benefit was demonstrated across all clinical and molecular subgroups that were evaluated. These progression-free survival curves showed benefit in both patients with and without brain metastases at baseline. Patients with no brain metastases at baseline had a hazard ratio of 0.59, which is very good and clinically relevant. But a particular importance in the EGFR mutant lung cancer setting is the results in those patients with brain metastases, with a hazard ratio of 0.34. Ivonescimab showed strong clinical performance against brain metastases, and in this setting, it's a cornerstone importance, particularly post TKI because ivonescimab appears to overcome the poor prognostic implications of brain metastases. As Dave noted earlier, we performed a total PFS to include all Western patients. This analysis shows the consistency of patients in Western regions versus those in Asia. These curves show the consistency of data with additional time for Western patients from the primary progression-free analysis, [indiscernible] free survival analysis in the solid line to a later analysis at the time of the predetermined overall survival readout in the dotted line. The highlighting -- I'm sorry, overlapping curves and the consistent hazard ratio between the primary analysis and the total PFS analysis demonstrates that the multiregional benefit holds with more mature data. The hazard ratios were consistent and favorable and highly clinically meaningful in both Asian and Western patients. Recall from the baseline characteristics of the study was approximately 62% Asian patients, which is in line with historical multiregional studies conducted to patients, the EGFR mutant lung cancer and is representative of a multiregional study from a breakdown perspective. We specifically overlaid the curves to allow for visual to show the primary analysis conducted with a prespecified number of PFS events occurred for the protocol, primarily driven by events in Asia and the total PFS when all Western patients were enrolled and the median follow-up time for Western patients was over 9 months. You can see highly concordant curves here, showing you visually that the data are highly consistent when including all Western patients. The hazard ratios are highly consistent overall from the 2 time periods, 0.52 to 0.57. And the regional differences, both of the regions are within 0.10 or 10 basis points of the global hazard ratio. These are consistent with highly overlapping confidence intervals. We wouldn't and don't expect the exact same numbers for hazard ratios for subgroups, which are smaller subsets of the total population. The overall survival curve separate with the difference at the median of 16.8 months versus 14.0, with a hazard ratio of 0.79 and a p-value of 0.0570. This was a strong result in this patient population. To date, there are no drugs approved in this setting that have demonstrated a statistically significant overall survival benefit. Our data showed a hazard ratio of less than 0.80 and clear separation of the curves. Note that the curves begin to separate at approximately 9-month follow-up. As well, you can see Western patients' median follow-up time just over 9 months. Based on the follow-up time of these patients, as we stated in May, our top line press release, we noted that we would continue to follow Western patients for additional time on trial as we felt that, that was clinically relevant to the results of this study. Because Asian patients had been followed for over 30 months and exhausted complete follow-up at the time of the primary OS analysis, we used the agent patient data from the April 2025 data cut and an updated -- and updated the follow-up of the Western patients. We launched the database in September 2025 when the median follow-up for Western patients was approximately 13.7 months, and with approximately 4.5 months of additional follow-up for western patients. There were a total of 76 OS events in the Western patients, which is in alignment with the original expected number of events in the primary analysis for Western patients when we began the study. Hence, we included in yesterday's presentation, the updated OS analysis that included longer follow-up of Western patients. The long-term hazard ratio was consistent with the primary analysis, at 0.78 with a nominal p-value of 0.0332. Median OS for this analysis remain the same in both arms from the primary analysis. Median OS in Western patients receiving ivonescimab of 17.0 months compared to 14.0 months for those receiving placebo with chemotherapy. Median OS and North American patients specifically had not yet been reached in the ivonescimab arm compared to 14.0 months in the placebo arm. Median overall survival in this longer-term analysis is highly consistent here. With the 2 regions, Asian and Western patients, both showing highly consistent results in both the ivonescimab arm and showing identical results in the control arm. As we noted in our press release yesterday, the statistical analysis plan for the study required a p-value of 0.0448 or less in order to achieve statistical significance at the time of the primary analysis of overall survival. These results show the impact ivonescimab had versus standard of care in this setting. With the curves maintaining their shape and their separation of approximately 9 months. The additional median follow-up time of Western patients demonstrates the consistency in the data between Asian and Western patients. Of course, the nearly identical median -- in the median overall survival also drives the regional consistency point home and punctuates the level -- the value that ivonescimab can bring to patients globally after its success thus far in China. This demonstrates both the consistency in the results across regions and the strong performance of ivonescimab in this second, a setting where no drug has shown an overall survival benefit. Important in this setting as well is the safety and tolerability, where ivonescimab can play an important role as an option for physicians and patients facing this disease and the need for treatment. The overall survival force plot shows consistency across the trial subgroups. In particular, as it is important to acknowledge that results in North America are of increased focus in today's environment, it's encouraging to note a hazard ratio of 0.70 amongst North American patients. There were no patients that -- there was no patient that was a detrimental impact on survival, and we see stability over time on primary analysis. Both objective response rate and disease control rate were improved for ivonescimab by an absolute difference of 11%. The median duration of response was improved from 4.2 to 7.6 months. I'd like to pause here to reemphasize what we have just seen. The primary analysis for PFS, which was strongly positive in July 2024 was backed up by the consistent results with Western patients having more maturity in the total PFS analysis that was performed. The overlap in Kaplan-Meier curves, the curve saw a moment ago where we overlaid the primary PFS analysis and the total PFS analysis, showed a high level of consistency between the data with mature global PFS data included. The primary overall survival analysis was performed in April 2025 and with a follow-up time for Western patients that was less than the median survival. Given the clinical relevance for following these patients, we did a recent analysis that lets us evaluate the results in the Western patients where there -- when there were the number of Western events expected in the protocol in the primary analysis. These results were highly concordant in hazard ratios and medians in overall survival, including Western versus Asian patients, demonstrating the consistency of the data. When we look at HARMONi-3 and HARMONi-7 -- sorry, when we look at HARMONi-3 and HARMONi-7 in the front line, driver mutation negative non-small cell lung cancer setting, the importance of overall survival is only further highlighted. The regional consistency seen in HARMONi is very encouraging in that light. Treatment-related adverse events showed small increases in the ivonescimab group. Any grade events were increased from 93% to 95% and Grade 3 events from 42% to 50%. Treatment-related adverse events leading to discontinuation of ivonescimab or placebo occurred in 7.3% versus 5%, respectively. Based on the dual targeting of ivonescimab, we analyzed adverse events that were thought to be immune related or VEGF related. These were increased from to 6% to 9.6% for Grade 3 and higher potentially immune-related events and from 3.2% to 7.3% for potentially VEGF-related events and will be detailed in the next slide. On the right, events are split out by specific terms. The most common events were lab abnormalities, nausea and decreased appetite consistent with the chemotherapy backlog. As such, there is a little difference between the arms. Focusing further on immune and VEGF-related events, we see that most events were low grade. The most common immune-related events were thyroid abnormality, and other events were at low rates and similar whether the patient received immunotherapy or not. Among the VEGF-related events, proteinuria, hypertension and hemorrhage, each was seen in 10% to 14% of patients, but were not general -- I'm sorry, but were generally low grade and the Grade 3 rate of hemorrhage was under 1%. Venous and arterial thrombotic events were similar in the 2 groups. I do want to pause here for a moment and speak to the highly validating and strong safety profile seen in the Asian patients that was also validated globally with the results of the study, where we saw generally consistent results. This again differentiates ivonescimab from the results seen in prior studies of PD-1 plus VEGF monoclonal antibodies being administered together and the difference in VEGF associated toxicities from previous studies with anti-angiogenic therapy. In summary, ivonescimab provided a significant and clinically meaningful progression-free survival benefit in patients with EGFR-mutated non-small cell lung cancer after progression on a third-generation TKI, with a hazard ratio of 0.52, which was consistent across predefined subgroups, including those with brain metastases and consistent with the total PFS analysis that included all patients, including all western patients. Overall survival showed a favorable trend with a hazard ratio of 0.79, response rate and duration of response were also increased. Ivonescimab was well tolerated with less than 1% Grade 3 or higher bleeding events and low grade -- sorry, low [ grades ] of adverse events leading to discontinuation or death. Most adverse events incidences were similar between the 2 arms. One additional point that I would make with respect to this particular setting and that Dave acknowledged earlier, is that there are very limited treatment options available for these patients. With such limited options available to patients with EGFR mutation positive non-small cell who progress after a third-generation TKI with only a single approved regimen in the United States of ivonescimab plus chemotherapy, that regimen was approved based on the PFS benefit, reasonably similar to what we've seen here in a cross-trial comparison admittedly. And with a favorable trend in overall survival without showing a statistically significant benefit. I'd also note that there may be meaningful tolerability differences between these regimens based on data, not only in this specific setting, but data presented as a whole, considering both compounds, ivonescimab and amivantamab, this underscores the specific value that ivonescimab an HARMONi regimen could bring to patients in this setting with the opportunity to include a differentiated mechanism of action for physicians and patients to choose from. Importantly, we want to thank patients and their families, the site personnel and the Summit and Akeso team for the many roles they collectively played in carrying out the HARMONi trial. And with that, I'd like to turn it back to Dave.

Thanks, Jack. Finally, there's one additional point that I would like to emphasize. If we look at the read-through of this study, of our other ongoing Phase III studies in frontline non-small cell lung cancer. Just about everyone would agree that EGFR-mutated non-small cell lung cancer post targeted therapy is biologically quite different than frontline non-small cell lung cancer that is driver mutation negative. The way to show that the existing treatment options is the PD-1 therapy is the overwhelming standard of care with or without chemotherapy, depending on PD-L1 status in the frontline driver mutation negative lung cancer setting. Those same PD-1 therapies did not show a PFS or an OS benefit in multiple multiregional studies in the post-TKI EGFR mutant setting. Ivonescimab clearly outperformed the results of the PD-1 monoclonal antibodies previous studies. From a market opportunity perspective, there are fewer patients suffering from EGFR-mutated lung cancer. In this HARMONi setting represents much less than 5% of the potential market opportunity for ivonescimab long-term. But when we look at these much larger frontline settings, overall survival will be important. The highly consistent results in overall survival in the HARMONi study, taking a look at the North American results based on today's -- including taking a look at the North American results based on today's environment and seeking consistency there specifically. But seeing regionally consistent results and seeing consistent hazard ratios and medians is very encouraging when we look at the HARMONi-3 and HARMONi-7 and beyond where overall survival will be highly important. These results presented yesterday are highly encouraging and showing the regional consistent performance of ivonescimab and ivonescimab's ability to break into a setting where historically immunotherapy has struggled. Anti-VEGF therapy has not yielded much in terms of overall survival here in previous studies, and continues to validate the opportunity presented by ivonescimab to make a significant difference across a vast number of patients facing solid tumor diagnosis. Thank you, team. And with that, we will now see if there are any questions that our team can help answer. Tina, if you could please open the line for questions.

[Operator Instructions] And our first question comes from the line of Yigal Nochomovitz with Citigroup.

Could you comment, please, on the path forward with respect to our regulatory interaction on HARMONi? Presumably, you're going to have a meeting with the FDA to discuss these results. Are you going to do that with the current OS cut, or will you wait to do potentially another cut? And could you comment at all on what FDA has said with respect to the level of evidence that they want to see on the U.S. patients in addition to the overall statistical significance on OS to support a U.S. approval?

Thank you, Yigal. This is Dave. So we believe the data is very encouraging and is fileable. We previously disclosed that the FDA is looking for a statistically significant overall survival benefit in the study. And what I would say is the data cutoff for this analysis that we presented yesterday is September 2025. And so obviously, given today's date is only days old at this point. So we're highly encouraged by this data. But we do need to plan the appropriate next steps in terms of ensuring that we provide ivonescimab with the best opportunity to reach patients as quickly as possible. So we plan to finalize that strategy in the coming days and weeks.

Okay. And then the other question I had...

Our next question comes from the line of Salveen Richter with Goldman Sachs.

Any hypothesis here as to what is driving the differences in PFS between the regions when you look at China versus ex China?

So I'll make a comment, and then I'll look to anybody else. This is Dave. But with the total PFS analysis, you can see that the curves are highly concordant when you overlay those on top of the primary PFS analysis. And that's usually striking way to show that they're highly consistent when you include the Western patients. So the hazard ratios are highly consistent overall between the 2 time periods in total. The 0.52 to 0.57 and the regional differences are highly consistent with highly overlapping confidence intervals is being included in this line. So we wouldn't expect, and we don't expect the exact same numbers of hazard ratios for subgroups. And in particular, when you have small subsets of patients or smaller subgroups that you'll have some variability in the exact numbers. But as Jack mentioned on the discussion from a PFS perspective, both in the longer-term follow-up, both the Asian patients as well as the Western patients were within 0.1 of the median for the study, as a whole. And so those are statistically highly consistent results. And of course, for overall survival, you see overwhelming clear consistency within the median, high consistency across the board in total and the median overall survival is nearly identical Western region to the Asian groups. They were exactly the same in the control arm. Every none -- these are mutually dependent patients who each showed a control arm 14 months at the median for overall survival and between 16.7 and 17 months of overall survival in the ivonescimab arm. So highly consistent across the board.

I would just add that in my earlier comments, we emphasize not to overread too much in a subgroup analysis in terms of you can't evaluate each one separately, is hypothesis generating. But I think we should be reassured that when we look at the subsets based on geography for both PFS and overall survival, those results address the question in a reassuring way that say, this is not a China-specific benefit that it's overall looking very strong, particularly in the North American population that is right at this moment, a particular priority. And that even if we don't want to say, oh, this is better, there's certainly no suggestion that it's worse and I think is very reassuring that the impressive observations that we've seen with this out of China, not just here, but in other studies, are not remotely suggestive of this being a benefit that stops at the border of China.

Our next question comes from the line of Tyler Van Buren with TD Cowen.

This is Greg on for Tyler from Cowen. Congrats on this data. I'd be interested to hear what feedback are you hearing from physicians at the conference following this data set?

Yes. I -- this is Jack West. I have had very little time to sleep in between all the conversations I've had with my colleagues from the U.S., from all over the world, even some in Asia. And I would say that it has been interesting as people are well aware, there have been kind of mixed messages and responses to the results in what's been written about so far, largely in the financial settings. But my colleagues have been congratulatory has felt that the characterization of the results as negative for OS is something that maybe you should have an asterisk next to it or with caveats that this had many practical challenges with it. We have already acknowledged the sequential nature and the -- what we would now with hindsight call in opportune timing of the final OS analysis. But when you actually look at what was achieved a hazard ratio of 0.79 a p-value where it was, this is not a hazard ratio of 0.89 and a p-value of 0.62. And so we would not want to have the risk of this being potentially a false negative almost. And so to my colleagues who are treating patients going back to their clinics this week and seeing patients with EGFR mutation positive disease. They say that they would be very eager to use the HARMONi regimen. They would prefer it over the alternative. Even some of the people who have been more critical of the trial still acknowledge that this would very likely be their preferred choice over what's currently available. And so when you look just in absolute terms that what was achieved in efficacy, the safety profile and the remarkably favorable, I would say, tolerability, particularly in the context of what else we have. This is something that I believe patients and physicians would eagerly welcome to have available. And they have also said they consider these results overall to be very favorable and don't discount the potential utility of ivonescimab in the first-line trials that are ongoing.

And our next question is from the line of Brad Canino with Guggenheim.

Great to hear the updates from you all. One of the top investor questions I get is, will the HARMONi-2 interim OS of 0.777 hold up over time. I noticed that this analysis included 15 extra months of follow-up for the Asian subset since the last time Akeso cut the HARMONi-A OS data, I think it was December 2023. Is there anything to learn about the ability for PD-1 VEGF to maintain OS from this analysis?

Thanks, Brad. And I think that's a good question. So a little bit on the specific point here, and then I think we can further from that. And so if we look at HARMONi-A, which was the study sponsored by Akeso. And that was highly overlapping, but not the exact same patient population. So about 86% of the patients from that study included here. And those who received third-generation EGFR TKI are the ones that are included here. And so the -- but we saw the overall survival analysis. The Chinese health authorities, the NMPA has requested during the approval process for ivonescimab in China back in 2023 and 2024. And so a 30% data maturity, we saw a hazard ratio of 0.72, and at 52% data maturity, and when I say data maturity, I'm talking total events divided by the number of total patients in the study, just to be clear. We saw at 52% data maturity; we saw a hazard ratio of 0.80. And so a little bit of commentary has taken place with respect to, therefore, as sometimes had been seen in historical studies associated with anti-VEGF therapy, a degradation of OS was a concern. And so we have maintained from the beginning that the 30% data maturity was a very early cut. And that's not exclusive ivonescimab to be clear. That's much more of a general comment. But when you get to 52% data maturity, you have stabilization that takes place, if anything else. And so now what we're seeing -- you see within our forest plot, you see the Asian cohort, right, it's not exactly the same patients. It's highly overlapping, but not exactly the same patients, right, who show a hazard ratio now for overall survival of 0.76. And so what we're seeing here is not a degradation of overall survival, but actually comparable, consistent, it's a tick towards favorability. I'm not going to make a larger, broader statement on that. But that is a piece of evidence now that we have a long-term follow-up on a randomized Phase III setting with ivonescimab that shows no degradation in overall survival. The other piece I would say is that that's really addressing the VEGF component. If we look at this study and on the slides that were presented yesterday as well as are available on our website, you look at the long-term overall -- sorry, the longer-term follow-up of Western patients and you look at the overall survival Kaplan-Meier curves. And what we see from our perspective, in our opinion, is an IO tail. And that tail seems to be clear present to us, and that represents what we see in other immunotherapies. And so what we are seeing as a result of taking a look at this data is both the PD-1 and the VEGF components of ivonescimab coming into play as well as a differentiated efficacy profile from what has been seen historically with anti-VEGF treatments. And we see a tick -- again, a tick favorable. I'm not making a larger conclusion there, but a tick favorable in terms of the hazard ratio associated with patients in Asia. While undisclosed, our colleagues at Akeso have also announced that the HARMONi-A study showed significance in OS. That's their announcement, and they haven't given any further data, so I won't speak any further to that. But that also backs up that same point with respect to the utility of ivonescimab and what it can -- what it has the potential to do and what it can be in that setting.

And our next question comes from the line of Asthika Goonewardene with Truist Securities.

Good additional color on the call today. I've got a 2-part question here. So the discussion at Lung took a very conservative view and some form of KOLs called it a little harsh. Jack, I was wondering if you could tell us what is your biggest disagreement with Raj's comments -- sorry, Ram's comments? And then second, the OS curves do not -- that were presented do not lose separate until maybe about 9 months in. And this certainly underlines the importance of data maturity and suggest that the ex China component, the data can continue to improve the follow-up in time. But any hypothesis as to why you're seeing that early overlap?

Okay. I didn't quite hear the second part, but let me get into this and we can see. I'll preface this by saying that Dr. Ramalingam and I have go back 25 years as respected colleagues and friends, we are not any -- no, I'm just kidding. We are still, friends. But the -- I would say that to me, the biggest issue is that he really minimized, I would say, the foreseeable understandable issue of the overall survival cut, the relative immaturity of the OS data and our ability to -- we had called that we would -- what would loom large is how do things look as we get more data from longer follow-up of the Western cohort and knew this was going to be a problem. And I would say that the issue is just potentially throwing out the baby with the bathwater. And to say none of us is claiming that this is statistically significant. It will not be. That is the way it went. But as I said, I think we are always, as clinicians looking for truth within the reality that the trials don't always provide clear answers. And I think that this one was fraught with challenges in how it was developed and how the data ended up getting analyzed. But -- so we don't want to dismiss truly beneficial therapies inappropriately. And my concern is that -- and I'd also would say this is based on conversations with many other colleagues in the last 28, 30 hours. I think a lot of us come into it with biases about whether we believe that it's all going to be anti-angiogenesis anyway. And so if that's your view, it's very hard to overcome that bias. I believe that was the case here. So I would also say that the reality is that Dr. Ramalingam is not coming in with experience with ivonescimab directly or with bispecifics. Some of the conversations I've had just in the last few hours are with people who have actually used ivonescimab who are extremely committed to it, who have participated in this trial or others, and are strong believers because, first, they've put patients on trials and said, you cannot tell the difference in who's randomized to placebo or ivo, and that's something you really like in the drug. And they've seen what they've come to conclude are really good results as well. And so there's nothing like that clinical experience. So that's my view. I would say that no one would ever say that Dr. Ramalingam is not knowledgeable or unfair necessarily. But I think that his view was about as far in the negative range of the confidence intervals of fairness as you could get within the admittedly ambiguous setting of the trial. This couldn't be said to be a slam dunk. So it's subject to interpretation. But I think it was a quite uncharitable one and more important than what I think because of where I'm sitting, colleagues have unsolicited come to me saying that they're talking with each other. And without having a horse in the race, they really feel that, that was not a really balanced and appropriate interpretation of admittedly still somewhat ambiguous data.

Our next question comes from the line of Cory Kasimov with Evercore.

I recognize this isn't the subject of today's update, but is there anything you can share with regard to your latest strategic thinking with ivo in terms of both interest in a potential partnership as well as plans to move ivo into Phase III trials in additional tumor types like your partner has been doing?

Cory, this is Manmeet. To take that and give you an update. First of all, yes, I would say we have full intention and plans to initiate a few more Phase III outside lung and also export more in lung in coming months, and you'll hear that in the coming months from us. On related to strategic activities, we've always said, we are always looking to expand the portfolio as soon as possible and as appropriate and we have also mentioned earlier, we are open to partnerships, right, to do and collaborate, which is -- which should be and would be beneficial for both partners for ivonescimab for all the patients to ensure that the drug can be got accessible and can accelerate development. And I would say that what I could see over here right now, and we are always open to do all those strategic developments.

Okay.

Yes, I will just reiterate, Cory, the comments that Manmeet made with respect to continuing the development of ivonescimab as well, as we said in the beginning of this year, fully intend to continue and more to come on that point. So we look forward to that later on as well.

And Cory, I can only add one more thing, right? Every month, we are seeing more and more positive news on a ivonescimab. You just heard, right, a couple of weeks ago, from our partner, Akeso, we announced the first OS for HARMONi-A, right, which was an EGFR, right? That was in China population, and they will provide more update in the coming months at appropriate conferences. You already know, right, HARMONi-6 a couple of months ago had -- was talked earlier, and that data will be announced in upcoming conference pretty soon. And our partner also mentioned ivo has HARMONi-2 updated OS, if you recall me this 0.77, which was a very earlier cut the more mature OS data would be coming in the next, I would say, 6 months or before, which our partner has said. So all of those catalysts are upcoming and near-term catalysts, which just open up the value of amivantamab and would allow us to expand our portfolio pretty significantly.

And our next question comes from the line of Mohit Bansal with Wells Fargo.

Great. Thank you very much for this update. I would like to go back to the discussant topic from the presentation yesterday. So I fully appreciate the maturity of the OS cuts and all that. But I mean -- [indiscernible] made a comment that it does seem like Western patient population does dilute the impact or effect of ivonescimab overall even from ORR point of view, it does seem like lower response rates. So how do you respond to that? And then, I mean, I'm still going back to the fact that between a lot of the investment thesis is dependent on Chinese data being translatable to Western patient population. So why do you think this is not -- this should not be read -- this should not be a read for your broader lung cancer development program.

I don't have the response rate. My understanding is the response rate in Western patients actually is -- at least shows a difference that is stronger between the ivonescimab and placebo arms than was seen in Asia, even if the absolute numbers are different. But the relative difference is the delta between them is actually greater. I don't think we would differ in -- the general conclusion that the Western patients diluted the effect of the Asian patients, but that's a time-dependent issue. It's just that, yes, if you -- if you have immature short follow-up in Western patients added to the much, much more mature data out of China, you dampen the effect short term, but that doesn't mean that long-term, that's going to be the case. That's why I would say it was a very inopportune time to do that final cut. But if the timing based on the number of events in the West would have been corrected for the actual delay and time it took to get the trial up and running in the West and it was done now, we would be talking about it differently. That's also something that some of my colleagues have said, when we talk about this as a negative trial, you're talking about the outcomes of a few patients. And if it had broken the other way, this would be hailed as a triumph in every way. And so I think it's really worth bearing in mind that this is a very time-dependent effect, and it was just very unfortunate that that's when the official final primary -- the primary OS was done. Yes, it did end up diluting that effect at that time. But we have seen and expect that we are prone to see in the future that any follow-up if subsequent work is done, will demonstrate at least a stable, if not growing improvement in OS as time goes by.

Yes. The only thing I would add to that, Mohit, in Jack's first comment that I made with respect to response rates. It's not necessarily a direct one-to-one to take the HARMONi-A overall response rate. Keep in mind the 14% of patients who did not come on to this study, also going to receive a third-generation TKI. So it's not necessarily that those numbers are identical. So just -- I think that's a conclusion a step too far in terms of speaking of the dilution of response here.

Because that's not the exact same population that actually entered into HARMONi.

That's right.

Our next question comes from the line of Eric Schmidt with Cantor Fitzgerald.

This is [indiscernible] on for Eric. Just curious as to if you have an explanation for why the EU patients are a little bit more poorly in the study and what the PFS hazard ratio is in North American patients?

The question was what about -- what happened with the European patients and why they may have fared less well.

Got it. Yes. So this is [indiscernible], right. So from the analysis point of view, right, we have a subset of subgroups that we are prespecified looking into. For this particular study, the study was stratified by region combining both Europe and North America together. So the most important point is to look at that as the Western patients in terms of what it tells us and the results we see there for both PFS and OS at the time. When we look specifically into Europe, it's a little bit what Jack also has talked about. So Europe has 2 things so to speak against it. One is it's actually the smallest subgroup within the region. Europe is smaller than the North American region. And on top of that, it started to enroll last, right? Because the study was, we started enrollment in U.S., specifically Canada right away and only then a few months later in Europe. So we -- I don't have any concerns about the Europeans. I know in this particular study; we talk so much about subgroups and trying to take each subgroup of itself. But I think what we can say is there is no subgroup here that has shown any detriment in terms of look and prespecified. And there's also not one subgroup that overwhelmingly drives the benefit. And that's really the purpose of looking at subgroups. We can take every individual number and go into a clinical study that has variance and noise and think that every number comes up exactly perfect, being exactly in alignment with each other, identical et cetera. So that's what I can add to that question.

Yes. And what I would say is I'll look at a different study that we don't have a horse in the race with that we're not in the same setting. But yesterday, the overall survival results were announced for osimertinib plus chemotherapy versus osimertinib, the FLAURA2 study in frontline EGFR mutant non-small cell lung cancer. It was a statistically significant benefit that was announced, and it should be treated as such and a win for patients. Underneath that in their disclosure of forest plots in regional differences, one of their defined regional subgroups had an overall survival hazard ratio of 1.00, right? And that is not something that people have talked about. And I agree with that, they shouldn't see because everything that Urte just said, looking that far into individual subgroups starts to make a bit of noise. And so I think that's an important context. Some of the analyses looking at like the very small differences in certain points here are a bit more granular than statistics are intended to take into play.

Yes. Subgroups of subgroups. And then the other thing is that I think clinicians and the broader world has been looking to see, does ivonescimab and some of these other agents that have looked promising in China, do they work outside of China. But if you see that there's a good result in a Western population, it is biologically implausible that you would see an effect that is very good in North America, but not a very similar biology in Europe. So I think that is very hard to tease apart how you could have China and North America look singularly similar, but different from the European population.

Our next question comes from the line of Kelsey Goodwin with Piper Sandler.

In terms of locking the Asian population for OS, but then continuing to follow the North American and the European, was this agreed upon with the FDA at any point? And is there any precedent for this?

What I can say is we did not go into detail with this additional longer-term follow-up with FDA. We discussed that it would be valuable to look at it because the maturity of the Western patients itself is informative. The maturity of the Asian patients have completely been reached and exhausted at the time of the primary OS follow-up. So it's way beyond the median for both of the arms. And it is very hard over a long period of time to actually follow patients and everybody dies. I don't actually think any clinical study has accomplished that. So this is an appropriate way of analyzing that, and we can discuss that with statisticians and use of it, but this is an appropriate way. And we admitted from the very beginning that our intention is to follow-up long-term the Western patients after the primary analysis.

Yes. And I think that was clear in our press release in May as well.

Our next question is from Daina Graybosch with Leerink Partners.

Let's say that you don't receive FDA approval ultimately based on the data. In that scenario, will you do another trial in this setting? And if yes, how might you design that trial?

Daina, thanks for the question. This is Dave. I think I'll probably go back to my earlier comments from a moment ago with respect to the fact that the data cutoff for this study ultimately being in September has not given us a lot of opportunity in terms of planning exact next steps with respect to agency interaction, how we will take this forward, what the next points become. And so I'm not sure we're at a point of going any further than that at this point.

Our next question comes from the line of Clara Dong with Jefferies.

So for the HARMONi study, is there any additional analysis you plan to provide to us, especially overall survival data from the Western populations at a later data cut down the road or it will kind of depend on your regulatory interactions later on? And then also for the forthcoming data readout from HARMONi-6 for the combo of ivo with chemo in frontline squamous non-small cell. Would you be able to share the scope of data we should expect to see there? And what do you believe is the bar for success?

Thanks, Clara. I think on the first question, yes to the -- your implied answer, which was that would -- I think that would have to do with further interactions. I think on the second point, HARMONi-6 I just want to be careful in the sense of that is the study that is sponsored by and conducted by our partners, Akeso. Very excited for that data to be shared. However, in terms of bar for success, I think these are predefined statistical analyses. The comparator arm, obviously, is extraordinarily high as a bar with PD-1 plus chemotherapy. And importantly, from a timing perspective, this was a success on an interim PFS analysis, right? And so hopefully, that -- those bars are high across the board, high comparator, high-quality comparator, patient population that previously wasn't able to be treated with anti-angiogenic therapy. And so that's an important component here with the PD-1 VEGF bispecific of ivonescimab. And so...

You're talking about treating 266 or so patients with advanced squamous lung cancer with a myelosuppressive doublet combined with ivonescimab and being able to complete that, not stop the trial, I believe no new safety signals, we'll see the data. But along with the efficacy, the safety will be very reassuring and important to the clinical community.

Yes. We're very excited for our next steps on that data. But given that, that is a study sponsored and conducted by Akeso. I'll leave it at that.

Our next question comes from the line of Reni Benjamin with Citizens.

Congrats on the data. Maybe just based on these data, can you talk a little bit about how this might be -- the slight differences might impact your statistical plan kind of going forward with your Phase III trials, could we expect any sort of changes maybe in the number of patients being enrolled just to make sure that statistical significance has helped? And then does -- have you -- are you doing any other sort of additional evaluations to look at what these patients might be getting post progression? And does that vary based on geography?

Let me start with the second one and then we go back to the first one. We have looked at the next time therapy for the HARMONi study and obviously, because of progressive disease was much further in the control arm compared to the IO arm there is already much more treatment in the control arm in terms of percentage of patients getting it, but they are getting it. In the control arm, there was 50% or more that have already started to get second-line treatment. What we have seen in the second-line treatment is very much what we would expect. Majority of treatments was with additional chemotherapy, but there was a meaningful cohort, approximately 20% of those patients that also get further EGFR-targeted therapy. So normal distribution would suggest to us there is also normal follow-through for the patients that can still get next-line therapy.

Very good. And I think on the first question, I'd reiterate some of the earlier comments with respect to the idea that, again, with the September 2025 data cut off, and I think it's important to dig a little bit of time to take a look at, given the [indiscernible] at this point.

And our final question comes from the line of David Dai with UBS.

So just regarding the nominal p-value based on the ad hoc analysis. I'm curious hat kind of statistics analysis have you done to drive the nominal p-value? And then at the same time, just wondering what do you -- how do you think the FDA will view this nominal p-value given that's less than 0.05 now. Is there a potential for them to view this as favorable and can potentially get approval based on?

Let me take the first one because the first [indiscernible] I am not the biostatistician. So I'm just going to comment my ability to understand this, but the general methodology in terms of assessing the p-value based on the Kaplan-Meier methodology is the same or similar between the primary analysis and this analysis. But we have been very direct, and Jack made a couple of comments in this regard that we know. And of course, we would have loved to that we know that we did not reach significance. So we are not explaining that. But the tool of assessing a p-value tells you still how much variability or lack of variability or uncertainty is in a given data set that you look at, because we are not claiming significance here, it is more about the descriptive that there is not a lot of variability and tightness of the data at this long-term OS analysis.

And the directionality of it.

And the directionality of it and being stable, right, or stable for us is good. That's what we are looking for and bringing in obviously the follow-up for the Western patients. What was the -- sorry, what was the third part of the question?

The FDA.

Yes. So that's right. And so with respect to the FDA, I mean we have a high amount of respect and appreciation for our work with the FDA, including our appreciation of they're working with us from the beginning of this process, as we said during this call, we in-licensed ivonescimab in January 2023. One of the first things we did was go to the FDA with respect to looking for a way to expand a single reason study that was opened at the time into a multiregional study. So we have an immense amount of respect and appreciation for the work that the FDA does as well as their collaboration efforts as a whole. And so I think just a comment on how we think they will think about it something just doesn't feel appropriate to us. And so we'll kind of leave that piece there. Our -- we are encouraged, as we said, by this data, but I'm not going to comment with respect to how we think the FDA will feel it. It's very important that we walk through more details if that were to be the case.

Dave, this is Bob Duggan. Can I expand on your comments a couple of things to point out. At the time the trial was launched, we were very, very appreciative that the FDA leaned in and allowed us to include China data and to do this -- to pull this trial off. That was something that people did not expect and we, in retrospect, they obviously made the correct decision. But at that time, I believe all of us underestimated the level of conservatism and caution amongst North American physicians with regard not only to bispecifics here, that this was going to be novel, but also with regard to China data. So this was the first and it took a lot -- took quite a bit of time to gather momentum to get off the tarmac, so to speak, and into the air. And that's why the estimated time for completion of the trial then came in a bit shorter than had we been able to evaluate that properly, it would. And that's the only reason why there's a -- the only reason why we have -- we extended and allowed the maturity of the data to be equalized so to speak, so that one could get a fair analysis as to the response. And you can clearly see it from the statistics it was a very good idea that we did. The last comment here, I know Yigal was cut off on his second question. Many of you appropriately were able to ask a second, Yigal if your second question is still of interest to you and of us or has it been answered, feel free to jump in.

And with no further questions, I will now turn the call back over to Dave Gancarz for closing remarks.

Thanks very much. We're just checking to see if Yigal was still there. But in summary, I think we're -- appreciate Bob's additional comments and appreciate the team's participation here. We're very happy as we currently stand and with respect to this updated data analysis, the presentation from yesterday, and I want to take this time now to thank each of you for attending today's call and your interest. And reminder that an archived version of the webcast will be available later today on our website. And with that, thank you very much, and enjoy the rest of your day.

Thank you for joining us today. This does conclude today's presentation. You may now disconnect.