Thank you for standing by. My name is Kate, and I will be your conference operator today. At this time, I would like to welcome everyone to the Summit Therapeutics ASCO 2026 Update Call. All lines have been placed on mute to prevent any background noise. After the speaker's remarks, there will be a question and answer session. If you would like to ask a question during this time, simply press star followed by the number one on your telephone keypad. If you would like to withdraw your question, press star one again. Thank you. I would now like to turn the call over to Dave Gancarz, Chief Business and Strategy Officer. Please go ahead.
Good morning. Thank you for joining us. Team Summit is pleased to be hosting this call from Chicago, where the oncology community has gathered for what has already been an exciting ASCO 2026 Congress. This meeting continues to showcase the rapid pace of innovation across cancer research and treatment, and we're energized to be a part of this important scientific and clinical discussion, directing the next wave of medicines for patients with cancer. We issued two press releases over the weekend relating to encouraging phase II data in metastatic colorectal cancer, and of course, the historic results of the phase III HARMONi-6 trial featuring ivonescimab, presented as part of yesterday's plenary session at ASCO.
The phase II CRC data is a global data set, including patients enrolled in the United States and China, and the HARMONi-6 study was conducted exclusively in China, sponsored by our partners at Akeso. All data in the HARMONi-6 study was exclusively generated, managed, and analyzed by Akeso. The press releases are available on our website, www.smmt.com. 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. Maky Zanganeh, our President and Co-Chief Executive Officer, Manmeet Soni, our Chief Operating Officer and Chief Financial Officer, Dr. Bilal Piperdi, our Chief Medical Officer, Dr. Allen Yang, our Chief R&D Strategy Officer, and Dr. Howard Jack West, our VP of Global Medical Affairs.
I am Dave Gancarz, our Chief Business and Strategy Officer. Before we get started with 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, so we'll be referring to slides being displayed on the webcast link. I'd encourage you to use the webcast link to see the slides being presented this morning. It will accompany our comments.
Following comments from our team, we will take questions. With that, I would like to hand it over to Dr. Jack West to walk through the beginning of the presentation.
Thank you, Dave. Yesterday, as you are aware, ivonescimab data were featured in a presentation as part of the plenary session of ASCO. The presentation, titled "Ivonescimab plus chemotherapy versus tislelizumab plus chemotherapy in previously untreated advanced squamous non-small cell lung cancer: Overall survival results of the phase III HARMONi-6 trial" was delivered by Dr. Shun Lu, M.D. Ph.D., Principal Investigator of HARMONi-6, Director of the Lung Cancer Center at Shanghai Chest Hospital, and Senior Professor. Prior to discussing the results of the study, we will reinforce the study design and baseline characteristics that were discussed at ESMO last fall, as well as published simultaneously in the first paper for the study in The Lancet. The study was highly significant in its primary endpoint of progression-free survival. Of course, the focus of the data release was the overall survival data from this study.
Revisiting the schema for the HARMONi-6 trial, this study evaluated ivonescimab in combination with platinum-based chemotherapy compared with tislelizumab, a PD-1 inhibitor, in combination with platinum-based chemotherapy in patients with locally advanced or metastatic squamous non-small cell lung cancer, irrespective of PD-L1 expression. HARMONi-6 is a single-region, multi-center, phase III study conducted in China and sponsored by Akeso, with all relevant data exclusively generated, managed, and analyzed by Akeso. Key eligibility criteria are shown here. Patients were randomized one-to-one and stratified by stage of cancer and PD-L1 tumor expression at baseline. Patients received either ivonescimab at 20 milligrams per kilogram plus carboplatin paclitaxel, or tislelizumab plus carboplatin paclitaxel for up to four cycles, then received either ivonescimab or tislelizumab as maintenance therapy for up to 24 months. Treatment was to be discontinued for intolerability, progressive disease, or initiation of new anti-tumor therapy.
The study's single primary endpoint was progression-free survival by independent radiologic review committee. The focus of today's progression will be overall survival. The trial included a total of 532 patients. Baseline characteristics show that this was a predominantly male population, nearly all with stage 4 disease. It's important to underscore that these patients with advanced squamous non-small cell lung cancer included patients with characteristics that have traditionally been considered as potentially associated with bleeding on anti-angiogenic therapy. Specifically, approximately two-thirds had a central tumor, 17% had encasement of major blood vessels in the chest, 9% had tumor cavitation, and nearly one in three had a history of some hemoptysis.
The breakdown of PD-L1 expression showed approximately 40% had a PD-L1 negative cancer, with the remaining 60% of PD-L1 positive cancers showing 40% in the low range of 1% to 49%, and about 20% with high PD-L1 expression of 50% or greater. The median follow-up time of this current data cut is 21.4 months. At the planned interim analysis of the study, the trial was positive for overall survival, with a hazard ratio of 0.66 and a corresponding key value of 0.0017, and representing at least a four-month difference in median survival, though this remains early at this time. The curves separate early at approximately six months after treatment begins and continues to expand with additional time. Landmark two-year overall survival rates were 64.7% for those patients receiving ivonescimab, compared to 48.6% for those receiving anti-PD-1 therapy with tislelizumab.
This represents the first randomized phase III study conducted in driver mutation negative frontline non-small cell lung cancer against another PD-1 therapy in combination with chemotherapy to achieve a statistically significant benefit in overall survival. When we look at various subgroups, it's important to highlight that the size of the subgroups is smaller by definition and not designed to show statistical significance on their own. The subgroup analyses for overall survival confirms the benefit in pre-planned subsets, as indicated by point estimates for each subgroup landing on the left side of the dividing line favoring ivonescimab. Showing that the study results were observed broadly and not driven by a specific subset or subsets of patients. These overall survival curves show benefit in patients with negative, low, and high PD-L1 expression, representing PD-L1 TPS scores less than 1%, 1%-49%, and 50% or more respectively at baseline.
Patients with negative PD-L1 tumor expression had a hazard ratio of 0.64. Those with low PD-L1 tumor expression had a hazard ratio of 0.67, and those with high PD-L1 tumor expression had a hazard ratio of 0.64. In other words, ivonescimab with chemotherapy demonstrated benefit compared to tislelizumab with chemotherapy across the entire spectrum of tumor PD-L1 expression. Ivonescimab continued to demonstrate an acceptable and manageable safety profile in the HARMONi-6 study, consistent with previous phase III studies of ivonescimab plus chemotherapy. No additional safety signals were noted in the HARMONi-6 study. Treatment-related serious adverse events occurred in 41.4% of patients receiving ivonescimab plus chemotherapy and 34.3% of patients receiving tislelizumab plus chemotherapy. Treatment-related adverse events leading to discontinuation in the study occurred in 5.3% of patients receiving ivonescimab plus chemotherapy, compared to 4.5% for those receiving tislelizumab plus chemotherapy.
The most common treatment-related adverse events in both arms were commonly associated with platinum doublet chemotherapy and included anemia, decreases in neutrophil counts, and white blood cell counts. Most of the possibly VEGF-related adverse events occurring in the ivonescimab plus chemotherapy arm were classified as Grade 1 or 2. Grade 3 or higher hemorrhage events were observed in 2.6% of patients in the ivonescimab plus chemotherapy arm, compared to 0.8% of patients in the tislelizumab plus chemotherapy arm of the study. In summary, ivonescimab provided a statistically significant and clinically meaningful overall survival benefit for patients with advanced squamous non-small cell lung cancer as first-line treatment in the HARMONi-6 trial, with a hazard ratio of 0.66, which was consistent across all key subgroups, including those with negative, low, or high PD-L1 tumor expression. Regardless of PD-L1 tumor expression, patients received consistent benefit.
ivonescimab's strong performance across all levels of PD-L1 expression is important as ivonescimab appears to provide clinically meaningful improvements to patients irrespective of PD-L1 expression. ivonescimab was well-tolerated with low rates of adverse events leading to discontinuation or death, both comparable to the tislelizumab plus chemotherapy arm. Prior to HARMONi-6, there were no known phase III clinical trials in non-small cell lung cancer that have shown a statistically significant improvement in both PFS and now OS compared to PD-1 or PD-L1 inhibitor therapy in combination with chemotherapy in a head-to-head comparison. Following the success of Akeso's HARMONi-2 study in China, where the PFS benefit was observed in a monotherapy setting for patients whose squamous or non-squamous tumors were positive for PD-L1 expression.
This is the second time in which ivonescimab-based regimens have become the first known investigational therapy to demonstrate a statistically significant benefit compared to standard of care PD-1 or PD-L1 inhibitor-based therapies. This underscores the specific value that ivonescimab in the HARMONi-6 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. We believe that ivonescimab has the potential to become a new standard of care for advanced squamous non-small cell lung cancer, and we look forward to the near-term readouts for our global phase III study, HARMONi-3, evaluating ivonescimab in first-line non-small cell lung cancer. With today's positive HARMONi-6 overall survival update, we gain even more confidence in the potential for ivonescimab to make a meaningful difference in the lives of patients with lung cancer.
On that note, we want to extend a heartfelt thank you to the patients and their families, the clinical site personnel, and the Akeso team for contributing to and conducting the HARMONi-6 trial. Additionally, it is worth highlighting that The Lancet published a full HARMONi-6 manuscript concurrent with yesterday's plenary session. With that, I'd like to turn it back to Dave.
Thanks, Jack. I'd like to echo Jack's comments around our gratitude to the patients who enrolled on HARMONi-6, their family members, trial site personnel, investigators, and of course, the Akeso team. We are highly encouraged by the read-through of this study to HARMONi-3 in frontline all-comers non-small cell lung cancer, as well as HARMONi-7 in non-small cell lung cancer with high PD-L1 expression. Anti-PD-1 therapy, with or without chemotherapy, depending on PD-L1 status, is the overwhelming standard of care in frontline lung cancer without driver mutations. ivonescimab, both as monotherapy and in combination with chemotherapy, now compares favorably in both settings in phase III studies conducted in China.
While additional HARMONi-3 data will be important to see in the coming quarters, the consistent, statistically significant, clinically meaningful results in progression-free survival and overall survival in HARMONi-6 and the PFS and OS data generated to date from HARMONi-2 are very encouraging when we look to our ongoing global frontline lung cancer studies and beyond. One additional point I'd like to ensure we go over. There were some commentary with respect to the subgroup analysis hazard ratios that were disclosed in the forest plot for overall survival related to results by age groups presented yesterday. As you may recall from ESMO Congress 2025, while patients under 65 and over 65 both showed a PFS benefit from ivonescimab plus chemotherapy as compared to tislelizumab plus chemotherapy, the effects were seen less in older patients.
An analysis was performed by Akeso to review differences in baseline characteristics between those over 65 in the ivo arm versus the tisle arm to understand this difference. In doing so, multiple imbalances were noted, including target lesion size and the presence of brain metastases. This was specifically addressed previously during the ESMO Congress 2025 presentation. Correcting for these baseline differences explained a substantive portion of the difference in hazard ratios by age for PFS, and it is important to see that written on the slide on the screen right now. If you are on the webcast, the slides from ESMO Congress 2025 note that when adjusting for these covariates, the adjusted PFS hazard ratio for those patients greater than 65 years old was 0.69. Note that for both PFS and OS, not even considering the imbalanced baseline characteristics, the older subgroup did not show any detriment.
Once adjusted for these baseline differences, the hazard ratio for PFS was normalized to the overall population. The difference noted appears to be most likely driven by the imbalanced baseline characteristics. These differences were seen in PFS and could be seen in the OS forest plots as well. Can happen in clinical trials. Each subgroup is not powered and is not necessarily balanced for baseline characteristics. If prognostic factors are not balanced, a subgroup can be impacted, as we described previously last October. Importantly, remember that HARMONi-6 is the fourth phase III study conducted evaluating ivonescimab. In the first three studies, results for patients under and over 65 were very comparable. In the global HARMONi study, the longer-term follow-up of OS showed a numerically better hazard ratio for those patients over 65.
HARMONi-6 is the only study where we see this difference, and this unpowered subgroup had imbalanced baseline characteristics. Let's move on and discuss the continued deep experience we have with ivonescimab. Ivonescimab has read out four phase III clinical studies to date, all four of which had positive data, leading to two approvals in China thus far. At this time, a total of 15 phase III clinical trials have either been announced, are currently ongoing, or have read out in multiple tumor types. 47 clinical trials have been initiated since 2019 between Summit and Akeso, evaluating ivonescimab in a variety of solid tumors. When considering investigator-initiated and collaborative studies, a total of 155 clinical trials are now listed on clinicaltrials.gov.
The enthusiasm demonstrated by investigators around the world to generate data and seek positive signals for patients facing high unmet medical needs really speaks to the opportunity and optimism surrounding ivonescimab. Together with our partner, Akeso, we have enrolled over 4,000 patients in either Summit-sponsored or Akeso-sponsored clinical trials around the world. Commercially, in China, over 70,000 patients have received ivonescimab. Turning to our pipeline and our global phase III HARMONi study. In April, we provided an update with respect to the squamous cohort of this study. We added to our protocol an interim analysis for PFS for the specific purpose of providing a potential opportunity for earlier regulatory discussions, specifically with the U.S. FDA. To achieve statistical significance at this early interim look, there was a meaningfully higher bar compared to the upcoming planned final PFS analysis based on minimal alpha spent on the interim analysis.
The interim PFS analysis was reviewed by an independent data monitoring committee, or an IDMC, and the committee recommended the study continue as planned. Importantly, no safety concerns were noted, and this study continues to be double-blinded. In line with prior guidance, we expect the final progression-free survival and interim overall survival analyses for the squamous cohort in the second half of this year. For the non-squamous cohort, we expect the final progression-free survival analysis to occur in the first half of 2027. The study continues to be double-blinded, and we do not have further information regarding the results of the study. We look forward to the results of the final PFS and interim OS in the second half of this year. With respect to the non-squamous cohort at HARMONi-3, we have completed screening for patient enrollment and expect to complete enrollment of the 1,000-patient cohort this month.
We continue to enroll in HARMONi-7 and HARMONi-GI3. Look forward to continuing our work with CorTech in head and neck cancer, as well as RevMed with novel RAS inhibitors, kicking off the GSK collaboration testing ivonescimab with ADCs, expanding the number of collaborations we have with other agents with novel mechanisms of action. With that, I'd like to turn it over to Allen to review the phase II CRC data in combination with FOLFOX chemotherapy, the backbone chemotherapy of our phase III HARMONi-GI3 clinical study. Allen?
Thanks, Dave. At this year's ASCO conference, ivonescimab data in colorectal cancer was also featured from the global phase II study, AK112-206, which evaluates two doses of ivonescimab, 20 milligrams per kilogram and 10 milligrams per kilogram, in combination with FOLFOX chemotherapy as first-line treatment in patients with microsatellite stable metastatic colorectal cancer. The study was conducted in China and the U.S. Akeso had previously disclosed encouraging phase II study data with ivonescimab in combination with different chemotherapy regimens in microsatellite stable CRC. The standard of care chemotherapy in first-line metastatic setting is most often FOLFOX, we expanded the phase II study to include two cohorts of ivonescimab at different doses in combination with FOLFOX. This was the basis for our phase III study, HARMONi-GI3, as Dave alluded to.
This also represents global data as a mix of patients who are enrolled from the U.S. and China for these two arms. In this analysis, all patients experienced a reduction in tumor burden compared to their baseline assessment. The addition of ivonescimab to FOLFOX delivered deep and durable response rates, and responses were consistent across the two dose levels. In the overall study population, ivonescimab plus FOLFOX chemotherapy achieved an objective response rate of 70.8% and a disease control rate of 100%. Treatment responses in the higher dose of ivonescimab plus FOLFOX arm were more durable than the lower dose of ivonescimab plus FOLFOX, with the duration response landmark estimate at nine months of 79.1% and 41.5%, respectively. For patients in the higher dose ivonescimab arm, the landmark PFS rate at nine months was 76.1%.
While progression-free survival remains immature, the high proportion of patients who are progression free at nine months is encouraging. The study demonstrated an acceptable and manageable safety profile for the ivonescimab regimen, and no new safety signals were observed. This was consistent with previous studies of ivonescimab, including phase II data in metastatic microsatellite stable CRC, and demonstrate the potential for a favorable benefit risk profile for ivonescimab plus FOLFOX in this setting. In total, including both arms, 20.4% of patients experienced serious treatment-related adverse events associated with either ivonescimab or chemotherapy. There were no ivonescimab-related deaths and one ivonescimab-related discontinuation, supporting the tolerability and ability to manage these adverse effects. These support continued expansion of ivonescimab clinical development in metastatic colorectal cancer. They also support the design of our global phase III HARMONi-GI3 study in microsatellite stable metastatic colorectal cancer that is currently enrolling.
We are pleased with these results at ASCO and are excited about the potential to help patients facing colorectal cancer. The disease has a particularly poor outcome in the metastatic disease setting, with a five-year survival rate between 13% and 18%, underscoring the urgent need for improved treatment options. Before concluding the call, I wanted to take the opportunity to remind everyone of our upcoming catalysts and recent accomplishments. As we've discussed on past calls, we will continue to provide further details on additional new global studies throughout the year as they are ready to share. To date, we have initiated global phase III studies in non-small cell lung cancer, colorectal cancer, and head and neck cancer through our partnership with Vortec. We look forward to continuing to grow our global pipelines and explore ivonescimab's potential to help more patients in new tumor settings.
For the HARMONi-3 trial evaluating ivonescimab with chemotherapy in frontline all-comers non-small cell lung cancer, we have completed enrollment in the squamous cohort. We are now completing enrollment for the non-squamous cohort and expect to conclude enrollment this month. As I mentioned earlier, the final PFS and interim OS analysis for the squamous cohort are expected in the second half of this year. For the non-squamous cohort, the final PFS analysis is expected in the first half of 2027. Turning to our BLA filing, based on our phase III HARMONi study seeking approval for ivonescimab plus chemotherapy in the EGFR mutated non-small cell lung cancer setting post-TKI, the submission is currently under review with the U.S. FDA. The agency has provided a target PDUFA date of November 14th later this year.
We continue to grow our commercial capabilities as we prepare for the anticipation of ivonescimab's potential first U.S. approval and potential launch later this year. Okay, I'd like to give it back to Dave Gancarz.
Thanks, Allen. On today's call, we've covered ivonescimab's accomplishments thus far in the clinic, but this is only the beginning of a vast potential market opportunity set for ivonescimab across solid tumors. We have discussed this several times before, but with each successful data set, the reality becomes closer. Ivonescimab has the potential to be a platform blockbuster drug. Ivonescimab is well-positioned to make a significant impact across the solid tumor treatment landscape between checkpoint inhibitors and anti-VEGF therapies, with an estimated total addressable market to be in excess of $100 billion globally. Looking only at the checkpoint inhibitor market for non-small cell lung cancer, market estimates for immunotherapy are expected to exceed $20 billion per year by 2028. On the slide you see on the screen now, many solid tumor settings where anti-PD-1 and anti-VEGF therapies are highlighted. PD-1 and PD-L1 indications in green, and anti-VEGF indications in purple.
The ones highlighted in yellow represent the indications where we are currently running phase III trials for ivonescimab globally, lung and colorectal cancers. Clearly, there are several additional opportunities for ivonescimab, including, in addition, novel settings where neither have been effective, such as EGFR mutant non-small cell lung cancer. Ivonescimab's differentiated profile, as we saw with HARMONi-6 achieving a statistically significant, highly clinically meaningful PFS and OS benefit, supports its platform potential across multiple indications, each of which could be blockbuster opportunities on their own. We have only just begun to see the potential of ivonescimab to help patients with solid tumors. These are exciting times at Summit, and we encourage you to stay tuned to our journey as we continue to expand the ivonescimab clinical development plan in new tumor settings.
We'll now turn the call back to the operator to see if there are any questions that our team can help answer. If you could please open the line for questions.
At this time, I would like to remind everyone, in order to ask a question, press star then the number one on your telephone keypad. Please limit yourself to one question and one follow-up. We will pause for just a moment to compile the Q&A roster. Your first question comes from the line of Tyler Van Buren with TD Cowen. Your line is open.
Hey, guys. Congratulations on the tremendous and potentially practice-changing data presented this weekend. Just, can you provide your latest thoughts on how you believe the PFS hazard ratio and now the impressive OS hazard ratio for HARMONi-6 will translate to the ongoing HARMONi-3 trial? As the follow-up to that, specifically, what is your confidence that ivo is performing similarly to HARMONi-6 in the ongoing HARMONi-3 trial, considering the fact that it passed on the recent interim PFS analysis?
Thanks, Tyler. I think with respect to the overall confidence on HARMONi-3, as we mentioned on the call, as well as we've mentioned a few times historically, the purpose for that interim analysis was specifically. As we look at the landscape overall competitively, there are multiple PD-1/VEGF in class of which we are significantly in the lead. So we looked at an opportunity to take the final PFS timing in the second half of this year and look to accelerate the timing by which we could speak to the U.S. health authorities, in effect, the regulators and in that case, specifically the FDA. So with the minimal alpha allocation, we looked to perform that interim analysis, as I mentioned on the call. As I said earlier, the IDMC recommended to continue and recommended no changes to the study.
With that, we have no change in terms of our overall confidence because what we did in that process was not change the final PFS thresholds in effect. It was extraordinarily minimal in terms of the cost on the final. As we look now at the data from HARMONi-6, we see two things. We see a highly statistically significant PFS and a highly statistically significant OS. That, in the same setting, gives us a lot of confidence as we look at the translation overall from one study to another in that same setting, that we are very well-positioned with respect to the second half of this year. I'd also remind, if we look back at the gold standard endpoint of OS, when we look at the HARMONi and the HARMONi-A studies, we had remarkably consistent median OS from the East and the West as well as highly consistent hazard ratios.
We've seen in the past consistency overall between data with ivonescimab in China as compared to globally. The final point I would make is overall, again, we've had four phase III readouts in total. All four have been positive. When we look at the trajectory of that data, it's all pointing in one direction at this point, and that's been positive. From that perspective, while no one has a clear crystal ball, it is very much as we look at the data, we see everything moving and aligning in the same direction from that perspective.
Dave, I'd like to add. Tyler, thanks for the question. Akeso has been a terrific partner. The HARMONi-6 and HARMONi-3 trials are very similar, if not almost identical. In addition, Akeso has helped us enroll patients from China for the HARMONi-3 study. About a third of those patients will be from overlapping sites and investigators with the HARMONi-3 and HARMONi-6 study. The one thing I'd also like to point out is we took a calculated risk trying to bring this drug to patients earlier. The enrollment for the squamous cohort just recently completed, therefore that data was immature. We understood that risk. We took it. Obviously, we were not happy with the results. We thought that was important for patients. As the data matures, we expect our data to strengthen, although we haven't seen it.
Your next question comes from the line with Salveen Richter with Goldman Sachs. Your line is open.
Good morning. Thanks for taking my questions and congratulations on the data here. Two from me. One is, how do you think the strong squamous results read through to the non-squamous portion of the study? Secondly, at the presentation, the presenter noted that the median OS was reached with the last patient event. Could you put that finding in context and expectations for OS upon additional maturation? Thank you.
Thanks, Salveen. Why don't we start with the first question that you asked there? Can you repeat the first question? I just want to make sure I've got full context on the first question.
Sure. Just how to think of the read-through from the squamous result to the non-squamous portion of the study.
Very good. I think it's important, and I'll give a little bit of history here, Salveen, because I think this is good context overall. When we look at the history of HARMONi-3, for example, we started in squamous, and that was generally because when we entered into our transaction with Akeso, the phase II data in squamous in combination with chemotherapy, was very strong and was maturing. We immediately, and I think Bob's alluded to this two times in the past, it was very important that when we entered into the transaction with Akeso, that we started very quickly in terms of establishing a presence with the PD-1/VEGF bispecific. The squamous data in phase II was mature. It was highly encouraging, and that's what prompted an immediate opening of that study.
As we then looked at the non-squamous portion, it was just simply less mature at that point in time, and so it was in our pipeline in terms of addressing. Fast-forward now to the middle of 2024. We see the positive data from HARMONi-2 that showed monotherapy Ivo performing well in both squamous and non-squamous histologies. At that same time in 2024, the phase II non-squamous data in combination with chemotherapy was also maturing, was highly encouraging. I say all that because when we get to now HARMONi-06, and your question with respect to the read-through, that HARMONi-06 data was highly validating to the phase II data that was seen in AK112-201, the phase II study. We saw highly encouraging phase II data in squamous. That then translated to the HARMONi-06 success that we're seeing now. We have highly encouraging phase II data in non-squamous as well.
That data becomes a little bit more validated as well with the phase III replicating the phase II data. When I say replicating, for example, the median PFS was nearly identical between the phase II and the phase III. What we're seeing is the data that we've relied on in terms of the encouraging opportunity in non-squamous, that phase II data has been, that same study has been shown to replicate phase II to phase III. Obviously, it's frontline lung cancer as well, and so highly encouraging opportunities are present, and the existing phase III data in squamous is highly encouraging. There's certainly nothing negative with respect to the non-squamous portion. I'll turn it over to Jack for the second half of that question with respect to-
Remind me?
Yeah. With respect to the last patient who passed away, leading to the median.
The highlight there is that at that part of the curve, it's a very small number of patients. A single patient death among four patients at that time led to a drop just below the median, but that's a very unstable finding. In fact, it is defining the lower limit of the confidence interval for that median, which means it can only improve from what that is, and that's exactly what Akeso, I believe, expects will be the case as the data mature and there's a much larger population to draw from who gets out to and beyond the median point.
That is the lowest number that would ever be seen and we might anticipate, based on the patterns, that that isn't truly a representative number and that it's going to end up with further follow-up as a stronger difference in terms of absolute number of months between the two arms.
Salveen, this is Allen. I'd add too, I understand the question about squamous translating over to non-squamous, being non-squamous is of greater unmet need in Western markets. I want to point out the HARMONi and HARMONi- A studies. Those were non-squamous histologies with the same chemo backbones, even though they are EGFR mutant non-small cell lung cancer. again, all the strong ivonescimab data in completely different tumor types like CRC, TNBC. the drug is active. The jump from squamous to non-squamous histology is small compared to the breadth of this molecule.
That's right. I think the other piece I would add on top of Jack West's explanation, when you look at the curve overall, what you are seeing is the curve separating at six months, but then continuing to expand. That's really what reflects the hazard ratio of 0.66, and that 34% relative improvement in reducing the risk of death.
Yeah. Obviously, the median is just one point in time, but that shape of the curve, as everyone highlighted, including at the plenary session, it is widening as the patients continue over time and the curve moves to the right. That median kind of underrepresents the benefits that was really seen.
Yeah. when you look at the baseline, tislelizumab plus chemotherapy, which performed either on par with or slightly above historical results with respect to the RATIONALE 307 study, which was the pivotal study in which tislelizumab plus chemo was approved. That relative improvement of 34% relative improvement on top of that 22, 23-month benchmark, you would certainly expect that to be a bit higher than the four months. Just the math would tell you it's closer to seven to eight. when you look at the median follow-up time of 21 months and change, what this really represents, because those curves are widening further with more time, that hazard ratio in respect to confidence interval, and thus the P value, becomes statistically significant, showing that there's clearly the benefit already with more opportunity to continue to look at longer term follow-ups, if you will.
This is the primary overall survival analysis. It is statistically significant and highly statistically significant at that. I think that 34% relative risk reduction is the important part to look at.
Your next question comes from the line of Yigal Nochomovitz with Citi. Your line is open.
Hi. Great, thanks and congrats on the very, very strong result at ASCO. Thanks for also mentioning the covariate analysis that you did at ESMO with regard to the PFS. I'm just curious, if you were to apply that same math and logic to OS, what the conclusions may be, and curious as to why the ASCO discussant didn't put forward that piece of analysis. Then also, could you just clarify as to why the interim was triggered at 204 versus the planned 225 events as defined in the protocol for the study? Thank you.
Thanks, Yigal. We appreciate the comment with respect to the covariate analysis. Importantly, there was not a separate covariate analysis run at this point with respect to overall survival. However, the baseline characteristics, to be very clear, don't change over the course of the study, right? These are the baseline characteristics of the patients entering in. The baseline characteristics, its PFS analysis remain, of course, consistent with those at the overall survival analysis. That's why we thought it was important today to remind everyone of this particular topic, which was discussed back in October of 2025, where those imbalanced characteristics, baseline characteristics within the greater than 65-year-old patients between Ivo and Tisley is very relevant.
As you saw, that translates from a PFS hazard ratio of about 0.88 down to about 0.69 when you took into account the fact that those in the ivo arm had higher rates of brain metastases, had larger target lesions at baseline. Those baseline characteristics, when adjusted, normalize the difference between those under and over 65. I think it is really important also to note that this is the fourth phase III study conducted with ivonescimab. The other three did not show any detriment when looking at those over 65 as compared to those under 65. this is where it's important to remember these are non-powered subgroups as well. you don't have necessarily balanced baseline characteristics amongst the subgroups, and you also aren't powering your study to reflect the results of those patients.
We see other studies, for example there's studies that have been run where highly statistically significant overall survival benefits, FLAURA2 is an example of this, where the results in first-line EGFR mutant non-small cell lung cancer, osimertinib plus chemo versus osimertinib monotherapy. About a third to 35% of those patients were enrolled in Asia ex China. That specific subgroup, that region, showed a hazard ratio of 1.00. Representing a third of the patients or more enrolled in that study. However, I don't think anybody would argue that osimertinib plus chemotherapy works in China, but not in the rest of Asia from that perspective, and then also works in the U.S. and in Europe.
I say all that just to provide an example of a place where when you look specifically at a subgroup and then you try to extract too much from that, it's very difficult to do because you don't balance for everything else that goes into a very well-designed and robust clinical study. as we think about that, there was no detriment to patients over 65, either from a PFS or an OS perspective. I think that's really important to point out.
I'll take the next question and you can add in. You've answered the question. The other question was about why they did the OS analysis slightly early. I think it was 204 events versus 220. It's not uncommon. It's a goal and it's often hard to predict how many patients you'll have at the time of the analysis, because there's a cleaning and sort of preparation of the data and the locking of the database. As long as you're well within 10%, that's not unusual. I can't comment. That question is probably best for our Akeso partners, but it's often, as part of a regulatory discussion, there may be a query from a regulatory agency. That's the reason I see that they would slightly vary from their targets.
Your next question comes fro m the line of Brad Canino with Guggenheim Partners. Your line is open.
Hey, thanks for the question and congrats to you, your team and your partner on those data. Now, the 0.66 hazard ratio and the at least four-month benefit, can you contextualize these results within the past frontline squamous lung trials that have changed practice? You always get a lot of questions on the translation of that benefit to HARMONi-3, but do you need to translate that magnitude, or is there some margin where even a lower result could be significant enough to still change practice? Thanks.
Yeah. This is Jack. I will field that. I spend a lot of my time talking with my colleagues who are clinical oncologists, and I would say that a huge amount of the excitement about HARMONi-2, I know I'm going back in time, was just that as impressive and as integral as pembrolizumab has been, a sense that it was an impenetrable plateau that we might never exceed in our careers. Just seeing the capacity to do better than that is remarkable, and all of the potential opportunities to exceed that with pembro or potentially other checkpoint inhibitors. Squamous in particular is a setting that has had a dearth of developments over time, as compared to non-squamous, where there's new driver mutations and targeted therapies that chip away. Squamous has really been left without many new opportunities, and that would make any developments especially welcome.
Seeing a PFS benefit is not negligible, and in many settings of oncology, that is eagerly accepted as enough to adopt a new therapy. I think that in lung cancer, non-small cell at least particularly, we often impose a higher bar of overall survival. That said, if it is a hazard ratio below about 0.75 or 0.8, that is enough for nearly every oncologist I speak with to consider that to be very clinically relevant, clinically meaningful, and worth adopting as a new treatment. I would also say that when I've spoken with and seen results of survey questions from colleagues about, well, if you see a PFS benefit, but the overall survival may or may not be statistically significant, but trending in the right direction, you get a splay, a mix of different opinions about whether that is clearly enough. About half of my colleagues really feel that-
PFS benefit and even a modest trend of improvement in overall survival is perfectly welcome, first to their patients, and secondly, themselves, because this is not a comparison against placebo or chemotherapy alone. This is the first time we're seeing an improvement against such a high bar, and it's not accompanied by a counterbalancing major liability and toxicity. People are saying, if this is compared to Pembro or tislelizumab, and you clearly exceed on PFS, and it's in the right direction for OS, and not accompanied by a problematic increase in toxicity, why would my patient not want to pursue that option? Yes, there's going to be a range in what clinicians require to adopt something. I would say the results we saw yesterday would be on the very far end of that range and really not controversial.
With a lot of room to see less than that and still be enough to motivate patients and most oncologists or many oncologists to readily adopt an improvement in some, even if not every parameter across the board.
Your next question comes from the line of Mohit Bansal with Wells Fargo. Your line is open.
Great. Thank you very much, and my congratulations as well. Would love to talk a little bit about the median follow-up at the point of interim in HARMONi-3 versus HARMONi-6. To what extent you can comment on that, and was there a difference between the median follow-up in HARMONi-3 interim versus six interim that could explain why you did not hit the interim at that time point? Thank you.
Yeah. This is Dave. I'll jump in and then I'll ask Bilal to provide any additional color. What I would say overall is enrollment patterns are a little bit different. As we know, in any study in China versus any study globally, the enrollment patterns are a little bit longer. What you will see in general when you look at enrollment in a global study is a little bit more back-ended enrollment, one, and two, longer period of time. I'm making very general comments with respect to enrollment in a global study as compared to China, where you tend to ramp up very quickly and enroll the duration of the study over a shorter period of time.
While Mohit, we're not going to give particular color in terms of here's median follow-ups and very specific related to HARMONi-3, as it would be incredibly unusual to provide that commentary on an interim look. What I would say is there are patterns that you can look at across global studies in general, across studies conducted in a single region, and particularly in China in general, where the distribution of events, the follow-up time, will be a little bit different just naturally, just based on the traditional patterns that you see for enrollment. I'll ask Bilal if there's anything else you'd like to add.
Great. Thanks for the question. I agree with Dave. I think more important rather than the median is the pattern of enrollment. Sometimes it's a little hard to translate one to the other. For the last interim for HARMONi-3, we took a calculated risk because we want to get this drug earlier to the patients. I think we're really fine now. We are very confident with how this is going towards the final analysis.
Very helpful. Thank you.
Your next question comes from the line of Corey Kasimov with Evercore ISI. Your line is open.
Hi, this is Josh on for Corey Kasimov. Thanks for taking our question. Given there's precedence of China-only trial allowing for approval in the EU, such as RATIONALE 307, wondering if there's any plans to use the HARMONi-6 data set to look for approval in EU, and if not, how could these results help with the HARMONi-3 application? Thank you.
Thanks for the question, Josh. I think I'll make a couple of points here. When we entered into the great partnership with Akeso, one of the things that we were pretty clear on with respect to the U.S. individually was that we'd never had an intent to take a single region study conducted outside of the U.S. and look to bring that specifically in China, and look to bring that to the FDA for approval. I think the FDA had made plenty of comments prior to that that was not what they were looking for. We, of course, opened the HARMONi-3 study. With respect to HARMONi-3, that'll be the primary study for which we would seek approval, especially in the United States. There's complementary considerations. There's another randomized phase III in the same setting, particularly in the squamous cohort of HARMONi-3.
From a package perspective, HARMONi-6 could be included to strengthen that overall package. The other piece is when we look at, to your point, Josh, questions and other geographic locations, those are things that we can continue to consider. Obviously, we're running HARMONi-3 in Europe as well. That's part of the calculus that needs to go into this, and part of this is also is we have, if you take the totality of events, we completed enrollment in the first quarter of this year for the squamous cohort. It was a late breaker at ASCO, so it's pretty fresh in terms of what we have here. Part of that, as we look at next steps, we'll consider multiple options, but it's important that we have the global HARMONi-3 data reading out in the second half of this year. We can look at combining packages in certain ways.
But this is also unlike the RATIONALE 307, there wasn't necessarily a study being run in those regions at the same time, which is where we've talked very consistently about our strong strategic and operational desires to bring ivonescimab around the world. With that, we can look at whether it makes sense to combine those packages based on what things look like when we get to the second half of this year as well, which make that a much more straightforward question to answer.
Thank you.
Your next question comes from the line of Reni Benjamin with Citizens. Your line is open.
Hey, good morning, guys. Thanks for taking the questions. Let me add my congratulations as well. I guess the question I have, given all that you know now for the HARMONi-06 study and the required amount of follow-up which occurred for the interim OS to hit statistical significance, can you maybe talk us through your thoughts as to why maybe you wouldn't delay the reporting of the interim OS for HARMONi-03? How do you maybe optimize so that you have the right amount of follow-up and the right amount of events that might occur so that both PFS and OS are potentially hitting in the second half of this year? Maybe as a follow-up, I think you mentioned the OS significance boundary was like 0.049 for this HARMONi-06 study.
Is it fair to apply the stat, since you've never specified the stat plan that we see in HARMONi-6 to HARMONi-3? Why you would choose a one-sided test over a two-sided test? Thank you.
You added a couple of questions in there.
I did. Sorry, Dave.
All good. Right. I think one of the points with respect to the one-sided test versus the two-sided test, I think there's a little bit of noise in one of the media coverage articles. I don't think that's a particularly relevant aspect here. The idea of both HARMONi-3 and HARMONi-6 is looking to show statistical superiority over PD-1 plus chemotherapy. We're not looking at it the other way around. There's really no functional difference between the two. It's just a different way to describe the results of the same test. With respect to the stat plan for HARMONi-3, I appreciate the question, but I think as I mentioned a couple of times, and as we've talked about, is a group and is common across nearly all interim analyses. We don't plan to provide individual details on the individual statistical plan for the study.
That includes both the previous interim PFS look as well as the upcoming interim OS look. I think the third piece, which is important is your question with respect to the bar in the interim OS at the end of this year, less the bar in terms of what the thresholds are, but the ability to compare the follow-up time in HARMONi-6 or other studies in looking at the timing of HARMONi-3. HARMONi-3, as we've talked about, this is an interim OS analysis, but it's also the final PFS analysis. Those are effectively tied together. We'll look at the final PFS analysis based on our pre-specified number of events. Those events look like they'll come in the second half of this year, and that's the end of the testing for PFS.
Traditionally, when you look at a final PFS analysis, you also, at that same time, have a look at interim OS, or you have a look at OS on an interim basis. I would say we should set no expectations that the idea is an interim OS look at the second half of this year is a make or break look. I think it will be very informative in terms of what that looks like from overall survival. That will also take place at a median follow-up time that's less than what the HARMONi-06 follow-up time was. HARMONi-06, I believe, completed enrollment in January 2025. This analysis had a data cutoff that was presented yesterday of the end of February 2026. You've got last patient in plus about 13 months there.
We talked about the fact that we completed enrollment in the first quarter of this year, and then we're saying that this final PFS and interim OS will take place in the second half of this year. While you don't know necessarily when in the second half, it's obviously going to be less than 13 months of median follow-up time. We think that's very informative in terms of directionally where OS looks, but I certainly wouldn't go into this with the expectations that we feel that OS as an endpoint is make or break in the second half of this year. We have a look, just like most trials do, at OS at the time of the final PFS, which is pre-specified, and that's what we're looking at second half of this year.
Within that interim OS gives us directional inference in terms of where we go from what that data looks like to describe the OS at this point.
Great. Thanks for taking the questions.
Your next question comes from the line of David Dai with UBS. Your line is open.
Great. Thanks for taking my questions. I also want to add my congratulations on the data. One thing you mentioned, Dave, was that there's imbalance between the 65 years of age and under 65 based on lesion size and brain mets. How do you think this might impact patient selection for HARMONi-3 trial? Do you expect to exclude some patients with these kind of covariates for the HARMONi-2 trial?
Yeah, I understand your question, David. I appreciate what you're asking. Importantly, HARMONi-3 remains completely blinded. If we step back, what you aren't able to define within subgroups as you look at the data is what different covariates will be unbalanced within the arms. I think it's a little bit of a speculative question of will different subgroups, which are not stratified in the trial, will they contain an imbalanced baseline characteristics? Again, I would say there's no way to know that before you go into trial. I would also say that's where looking at subgroups can be helpful, and I think as Jack mentioned in his comments. They're not powered for individual results, right?
What you're looking for is do you see a subgroup with a detriment, a hazard ratio over one, and particularly directionally well over one that gives you pause to say clinically, is there something here that may be more informative maybe than what we had thought hypothetically going into a study? I want to be very clear. All subgroups had a hazard ratio of less than one. All subgroups received benefit from ivonescimab plus chemotherapy in comparison to pembrolizumab plus chemotherapy. All subgroups were to the left of the line on the forest plots, right? There were no subgroups that were detrimental or had a detriment from ivonescimab as compared to PD-1. That's a really important context here. Even with the imbalance of these covariates, we did not see a detriment to any of these patients, right?
It was a smaller improvement, but an improvement nonetheless, right? I think that's the difference here that I want to make sure we don't lose sight of. This isn't an example where we're looking at a subgroup with a 1.3 hazard ratio and saying, "Hey, we really need to take a step back and understand if there's something there." There's no way to know a priori going into a study if the Western patients versus the Eastern patients or the PD-L1 high versus the PD-L1 low will ultimately have more brain metastases versus less or whatnot, because that's not really what you plan for. I think that's where we'll see that data as it comes through. It is very important that all subgroups were less than one, and therefore none of these imbalances actually led to a detriment.
Yeah, Dave.
Yeah.
Dave.
Go ahead.
I was just going to add one thing, too, to reassure you. There are criteria which we believe could impact the outcome. Things like brain metastases as well as PD-L1 expression levels are stratified for in the HARMONi-3 study.
Yeah, just to add on it here, this is Bilal here. If you need to look at the 95% confidence interval, I think they are overlapping between the two age groups. It just further strengthen that. This could be fairly random.
Your next question comes from the line of Dara Azar with Stifel. Your line is open.
Hi. Congrats on the data. Could you please characterize the growing separation of the OS curves? Do you attribute this to ivonescimab or the histology more? How strong is the case to see a similar phenomenon in non-squamous, non-small cell, considering HARMONi-2 OS hazard ratio translation from PFS was a little bit different in that mixed histology study? Thanks so much.
This is Jack. I think that the most noticeable thing, again, when you see the separation becomes clearer as the chemotherapy has been completed. Chemotherapy is kind of a normalizing effect. Once you're at the point where you're directly testing the ivonescimab against the other tier immunotherapy and pembrolizumab, the longer you go, the more that differential becomes apparent. I think that with non-squamous, non-small cell lung cancer, it's different because there's going to be ongoing maintenance pemetrexed, and so that may still have some kind of normalizing or effect that will change it from a paradigm where there's a prolonged period of just ivonescimab as monotherapy in the maintenance setting against pembrolizumab.
It'll be a little different, but you still see the separation occur, and I am optimistic that that's what we will see, especially as if patients are fortunate enough to be doing well without progression. In some cases, patients will have pemetrexed discontinued due to cumulative myelosuppression issues renal issues. Even though pemetrexed is generally considered a very well-tolerated therapy, it has cumulative toxicities if patients are on it for enough of a period that it really becomes longitudinal. There will still be patients who are just on ongoing, essentially monotherapy in HARMONi-3, ivonescimab against pembrolizumab. The separation I think will be there, but it'll be a little different question in the non-squamous population, both because of the underlying differences in biology and because there is not a maintenance chemo component in squamous.
Again, I think the only thing I would add to that is really the, again, you go back to the highly encouraging phase II data generated in this phase where we saw very some of this was disclosed as recently as ELCC 2024. obviously that's highly encouraging data as well that helps provide additional comfort on that front as well in terms of the applicability of ivonescimab plus chemotherapy in the non-squamous cohort.
I'd now like to turn the call over to Dave Gancarz for closing remarks.
I want to take the time to thank everybody for attending today's call. Obviously, the historic results of HARMONi-6, as we've gone over multiple times, is a day for celebration across both Summit team. I want to once again congratulate our partners at Akeso for a wonderful result. A hazard ratio of 0.66 is not something to shake a stick at. Not only that, this is the first trial to go head-to-head against a PD-1 plus chemotherapy in frontline driver mutation negative non-small cell lung cancer and show a statistically significant benefit, not only in PFS, but now overall survival. That is not, as Jack was very eloquently describing earlier, not a borderline case and not anything other than something to be fully celebrated. We want to take the time to thank you for attending today's call.
An archived version of the webcast will be available on our website. Enjoy the rest of your day. For those of you here in Chicago, we hope you have a wonderful rest of your ASCO.
Ladies and gentlemen, that concludes today's call. Thank you all for joining. You may now disconnect.