Good morning, everyone, and welcome to the NovaBridge Biosciences webcast titled "Givastomig: Mature Clinical Data and Path to Registrational Development." At this time, all attendees are in a listen-only mode, and a question-and-answer session will follow the formal presentation. As a reminder, this call is being recorded, and a replay will be made available on the NovaBridge website following the conclusion of the event. Some of the statements made on this call will include forward-looking statements within the meaning of the U.S. Private Securities Litigation Reform Act of 1995. Actual results could differ materially from those stated or implied by these forward-looking statements due to risks and uncertainties associated with the company's business, including those set forth in the risk factors section of NovaBridge's annual report on Form 20-F for the year ended December 31st, 2025, and any other filings that are or may be made within the SEC.
Please also carefully review the forward-looking statements disclaimer in the slide deck that accompanies this presentation. In addition, our statements are based on information available to us as of today, November 2nd, 2026. NovaBridge undertakes no obligation to revise or update any statements or information to reflect events or circumstances after the date of this conference call. I'd now like to turn the call over to Dr. Srishti Gupta, Chief Executive Officer of NovaBridge Biosciences. Please go ahead, Srishti.
Welcome, and thank you for joining today's webcast to discuss the updated phase I-B givastomig data presented at ESMO. A year ago, we were encouraged by the early activity we were seeing with givastomig. Today, with substantially more follow-up, we have a much deeper understanding of the durability, safety, and overall clinical profile of the program. Today's discussion is about more than an updated data cut. It's about how a much more mature dataset has changed our understanding of the program and increased our confidence in the future development. With additional follow-up, the key objectives of the phase I-B study have now largely been accomplished. We have characterized the durability and consistency of the clinical benefit, selected the 8 mg per kg dose for the registrational development, and used the data generated through this study to inform our phase III strategy.
What is particularly encouraging is that our confidence in the clinical profile of givastomig has increased as the dataset has matured. We continue to see evidence of durable clinical benefit with responses that can persist and, in some patients, deepen over time. The data also continue to demonstrate clinical activity across the CLDN18.2 expression levels, including lower expression tumors, supporting the broad potential of the program. To summarize, this mature dataset has allowed us to better characterize durability, establish the registrational dose, and support a global registrational study protocol that has been approved by the FDA. With that, I'd like to introduce our Chief Medical Officer, Dr. Phillip Dennis, who will review the updated clinical results, discuss the key findings from the mature dataset, and explain how these data inform selection of the 8 mg per kg registrational dose and support our planned registrational development strategy.
Following Phil's presentation, he will be joined by Dr. Sam Klempner, Associate Professor of Medicine at Massachusetts General Hospital and lead investigator of our Phase I-B study, who will provide an independent clinical perspective on the durability of the findings, the relevance for patients with gastric cancer, and the potential implications for the treatment landscape. Phil, over to you.
Thank you, Srishti. Before getting into the data, it is helpful to step back and consider what a successful therapy needs to deliver in frontline gastric cancer. Efficacy is critical, but so are broad patient applicability, manageable safety, and the ability to integrate into routine clinical practice. Many CLDN18.2 targeted approaches involve trade-offs between these dimensions, and this slide helps put that landscape into context. Against that backdrop, let us turn to givastomig. With the additional follow-up we are presenting today, we can now assess the profile of givastomig with greater maturity, including the depth and durability of response, safety, and activity across CLDN18.2 expression levels. Let us walk through what the updated dataset tells us. Next slide, please. When we initiated this Phase I-B study, we believed that a successful frontline therapy would need to demonstrate four things. First, broad patient applicability. Second, durable clinical benefit.
Third, a manageable safety profile. And fourth, practicality for routine clinical use. Earlier in development, many of these questions remained unanswered because the dataset was still relatively immature. What makes today's update different is the amount of follow-up we now have available. With nearly 18 months of median follow-up in the selected 8 mg per kg cohort, we can now move beyond simply asking whether the drug is active. We can now begin to assess whether responses are durable, whether safety remains manageable over prolonged treatment, whether activity extends across CLDN18.2 expression levels, and whether the overall benefit-risk profile supports broad frontline use. As I walk through the data, I will come back to each of these four attributes because they provide a useful framework for understanding how the program has evolved as the dataset has matured.
But first, it is worth briefly revisiting the mechanism and why we believe this approach had the potential to address those four attributes. Next slide, please. givastomig combines 18.2 targeting with conditional 4-1BB T-cell activation. This design seeks to amplify anti-tumor immunity within the tumor microenvironment while maintaining a safety profile suitable for frontline use. The key question has always been whether that design translates into durable clinical benefit and manageable safety. The updated Phase I-B dataset provides increasing confidence that it does. Of course, mechanism alone does not determine clinical impact. Ultimately, that needed to be evaluated in the Phase I-B study, which was designed to answer several key development questions. Next slide, please. The Phase I-B study was designed to not only establish safety and characterize activity, but ultimately to answer three questions that would be critical for registrational development. First, what dose should advance into registrational development?
Second, is activity observed across a broad CLDN18.2-positive population? Third, does clinical benefit remain durable with longer follow-up? Importantly, the study designed patients across the CLDN18.2-positive spectrum rather than focus exclusively on the high expression cohort. The updated dataset provides important information on dose selection, durability, safety, and patient population that helped inform registrational development planning. With that framework in mind, let's turn to what we've observed clinically, beginning with the breadth of activity across the CLDN18.2-positive population. Next slide, please. One of the most important findings in this dataset is the consistency of activity across CLDN18.2 expression levels. Strong activity was observed in patients with CLDN18.2 expression in greater than 75%, which is where currently approved 18.2 targeted therapy is primarily focused. Importantly, activity was maintained in lower expression tumors.
At the 8 mg per kg dose, which we have selected for registrational development, objective response rates were 82% in the claudin high population and 75% in patients with CLDN18.2 expression between 1% and 74%. Disease control approached 100%, and nearly every efficacy-evaluable patient experienced some degree of tumor shrinkage. Activity was also observed across PD-L1 subgroups. Importantly, these findings have persisted as the dataset has matured, with nearly 18 months of median follow-up in the 8 mg per kg dose. While these are early phase data, the consistency of activity across expression levels gives us increasing confidence that the opportunity may extend beyond the currently defined high expression treatment segment. Beyond response rates alone, we also wanted to understand the consistency and magnitude of these responses. Next slide, please.
Response rate is important, but we also wanted to understand the breadth and depth of activity across the study population. What we observed was evidence of activity across CLDN18.2 expression levels, as well as PD-L1 subgroups. Beyond the responses themselves, we saw substantial depth of response at the selected 8 mg per kg dose, with a median tumor reduction of 62%. Most notably, 68 of 69 efficacy-evaluable patients experienced some degree of tumor shrinkage. That level of consistency suggests the activity we're observing is broadly distributed across the population rather than driven by small subsets of patients. In our view, these findings strengthen the overall response rate signal and provide important context for what comes next. The key question then becomes whether those responses translate into durable clinical benefit over time.
With nearly 18 months of median follow-up in the selected phase III dose cohort, we are now in a position to begin answering that question. The key question then becomes whether those responses translate into durable clinical benefit over time. Next slide, please. If I had to identify the single most important takeaway from this update, it would be durability. Earlier in development, we knew givastomig was generating responses. The key question was whether those responses would translate into durable clinical benefit over time. With nearly 18 months of median follow-up at the selected 8 mg per kg dose, we now have much greater visibility into that question. What we're seeing is not merely a response rate story. The median duration of response is now 15.2 months. The median progression-free survival has reached 16.9 months. 14 patients remained on treatment at the data cutoff.
Importantly, responses often continue to deepen after an initial response, and today's dataset looks better than it did a year ago. That is exactly what you hope to see from a therapy intended for prolonged frontline treatment and benefit. What gives us confidence is not simply that we have observed the responses, it is that those responses translated into durable disease control as the dataset has matured. It is also worth noting that despite nearly 18 months of median follow-up, only 13 progression events had occurred, while 18 patients remained censored and 14 patients were still receiving treatment at the time of the analysis. In frontline metastatic gastroesophageal cancer, generating responses is important, but sustaining responses over time is what ultimately matters most to patients and physicians. For us, that is one of the most encouraging aspects of this dataset. The signal became stronger rather than weaker as follow-up matured.
That additional follow-up significantly increased our confidence in the durability of benefit, supported selection of the 8 mg per kg dose for registrational development, and helped establish the foundation for the FDA-agreed global registrational study protocol. But durability is only meaningful if patients can remain on treatment long enough to realize that benefit. That brings us to the safety profile, which I will cover on slide number 8. Greater follow-up also means substantially more drug exposure. We were pleased to see that as exposure increased, no new safety signals emerged, and we gained a much clearer understanding of the safety profile. No Grade 4, no Grade 5 givastomig-related adverse events were observed, and rates of immune-related adverse events generally appeared comparable to what would be expected with immunotherapy plus chemotherapy as a backbone for frontline metastatic gastric cancer.
The most common treatment-related adverse events include nausea, vomiting, and fatigue, all of which occurred without a need for titration or active infusion monitoring. No Grade 5 treatment-related adverse events were observed. Importantly, the safety profile needs to be viewed alongside the durability data we just reviewed. Patients remained on treatment for extended periods of time, providing meaningful insight into the long-term tolerability in a frontline setting in metastatic gastric cancer. The only increased immune-related toxicities compared to nivolumab plus chemotherapy is gastritis. Gastritis is a biological consequence of combining a checkpoint inhibitor, nivolumab, with givastomig, a claudin 18.2/4-1BB immune amplifier. Most gastritis events are less than or equal to Grade 2. Grade 3 events are rare. Gastritis typically develops after a tumor response is already documented.
The 8 mg per kg dose is associated with lower overall and lower Grade 3 incidents while maintaining a compelling efficacy and durability profile. The majority of events for gastritis are manageable with established intervention strategies, including temporary cessation of the immunotherapy assets, institution of high-dose steroids followed by a slow taper. This approach commonly leads to resolution of symptoms to where patients can be re-challenged with givastomig. Standardized management algorithm were implemented and refined during the phase I-B trial. As a result, the duration of gastritis has decreased over the course of the phase I-B trial. As the dataset has matured, 8 mg per kg emerged as a dose that best balanced efficacy, durability, and safety. Next slide, please. As we bring the data review together, we believe the updated dataset accomplished the key objectives of the phase I-B program.
We observed compelling clinical activity across tumors with CLDN18.2 high and CLDN18.2 low expression. With longer follow-up, we gained increasing confidence in the durability of benefit, with responses that persisted and often deepened over time. We established a manageable safety profile that supported prolonged treatment and did not reveal new concerns as exposure increased. Collectively, those finding enabled us to confidently select the 8 mg per kg dose for registrational development. Importantly, the maturity of the dataset also supported agreement with the FDA on a global registrational study protocol, including the patient population, study design, comparator arm, endpoints, as well as statistical powering. Taken together, we believe the key objectives of the phase I-B study have now largely been accomplished. We've spent the last several slides discussing the data from the company's perspective.
To help put these findings into clinical context, I'd now like to welcome Dr. Samuel Klempner, Associate Professor of Medicine at the Massachusetts General Hospital and the lead investigator of the phase I-B study. Dr. Klempner is a gastrointestinal medical oncologist with a clinical and research focus in upper GI cancers, specifically gastroesophageal cancers. He has been closely involved with the development of givastomig and has a unique perspective on what these results mean for physicians treating patients with metastatic gastroesophageal cancer, the importance of the durability signal, and where givastomig may fit in with the evolving treatment landscape. Sam, thank you for joining us. For the first question, Sam, you've been involved with this study since early in development and have seen the data evolve over time. Let me start with a very broad question.
As you've seen the data mature, what has stood out to you most about the updated results?
Yeah. Thanks, Phil. I would say a few things have stood out, and it has been nice to see longer follow-up and this data mature. One, it's nice to see the initial reporting of the response rate continue to hold. But really, in the world of effective frontline therapies and increasingly effective therapies, what we want to see is durability. I think that to see the response rate translate to durable disease control and durable responses, and translating into PFS and early OS reporting, is really the things that have continued the enthusiasm for me as an investigator treating these patients in front of me. Everybody likes to see their tumor getting smaller, but it's really the durability that tends to move the needle in this disease.
That's been, I'd say, at the very high level, some of the things that have stood out to me with the more mature data set.
In terms of the durability, have you seen patient subgroups benefit? What patients seem to benefit most, or do you think all benefits are obtaining some benefit?
Yeah, I think it ultimately will take a very large number of patients to tease apart, but it's encouraging in some respect that there is not a biomarker group that's clearly driving the benefit. For example, I think it's pretty clear from looking at the waterfall plot that the activity is across the entire claudin spectrum, not only seen in the very top expressors. Similar to PD-L1 expression, it's not like all of the responses are in the CPS high patients. The more encouraging story, I think, is that the early data suggests that the combination is active across the entire spectrum of PD-L1 and an extremely broad spectrum of claudin expression, going from 1+ in 1% or more of tumor cells is basically 80% of all gastroesophageal cancers.
In some respect, it's been the lack of a clear cluster that's driving the benefit that is one of the more encouraging things to me. Of course, much larger numbers may be required to see some trends, but this has been one of the more encouraging findings, I would say.
As the PFS data has matured, with now almost 1.5 years of follow-up, what are your expectations for overall survival given the current data with ORR and PFS? I know it's a projection, but given the association of these endpoints with OS, what's your sense of what overall survival might end up looking like?
Yeah. I mean, I'm optimistic. I think when we start to see PFS pushing past a year easily, and in some cases here, closer to 1.5 years , we start to get optimistic about overall survival, especially as we've seen some later-line claudin therapies demonstrate activity as well. So there's the potential for sequential claudin targeting like we do with HER2. So we may really start to push this into the two-year range. I mean, 20 months, 18 months, 20 months and beyond, I think, would get someone really excited in a larger data set for an overall survival. But maybe, who knows, maybe even better. But with PFS that's beyond a year, we know what we get out of later-line therapies in general. So, I would be looking hopefully well into the 18-month, 20-month kind of range, optimistically.
Thanks. Earlier in the presentation, I spent some time discussing safety as well as efficacy and the rationale for selecting 8 mg per kg for a registrational study. Now, what's your perspective on the benefit-risk profile of this selected dose?
Yeah. I think, as the data you presented, there is experience at 8 mg per kg and 12 mg per kg , and certainly the shorter follow-up at 12 is not something that particularly bothers me in terms of assessing the overall kind of totality of the data. It is a balance. These are patients who are having very good responses, and it's perhaps slightly higher response rates in the higher doses. But certainly, we also see slightly higher numerical rates of toxicity, including gastritis at 12 mg per kg. So, although in my own experience, having treated more than 30 patients on this trial, I would say the manageable toxicity has been similar at 8 mg per kg and 12 mg per kg in my own experience.
So we're not always looking for the highest dose, especially with agents that have an immunologic mechanism component to their action. It's really the durability. We have to have a balance of tox and efficacy that allows patients to stay on therapy for extended periods of time. To me, I think it is very reasonable thinking along the 8 mg per kg as sort of the dose that perhaps offers the best balance from a patient perspective. It is encouraging to see that both doses are clearly active, and I would agree that the 8 mg per kg is worth continued development.
If you look at the totality of the toxicity profile and think about what you would expect with nivolumab and FOLFOX alone, what is your overall assessment of what the addition of givastomig does to that profile?
Yeah. I think one of the presidential symposiums at ESMO, a couple of the discussants made some really good points about the limitations of CTCAE capturing events, and I would agree that it is an imperfect system, but it is the system we have. Actually looking ahead at CTCAE version 6, there are some meaningful improvements that provide some chronicity along to the grading. I think it will actually help the grading be more reflective of the patient experience, and ultimately, that is what matters in the real world. I would say that, to answer your question, chemo has a known toxicity, immunotherapy with nivolumab has a known toxicity, and givastomig in the monotherapy trial has some single-agent adverse events, as described in the original phase I. When we put them all together, I think as an investigator, we look to see if there is cooperative or synergistic toxicity?
Are we enhancing the toxicity profile of one of the existing components? I would say in my experience, we do not see any worsening chemotherapy-related toxicity from the addition of givastomig, and we have not seen a clear signal in emergence of other existing immune-related adverse events from nivolumab. For example, I do not see significantly more hypothyroidism or skin-related toxicity, things that we know from nivolumab. The one that stands out, and I think is no surprise, is really the GI side effects, the gastritis in particular. I would say as the data has matured, most of the events continue to be low grade in the current system, so low Grade 2 or less. There is not a new safety signal, but it is the gastritis that continues to be the one side effect with this combination regimen that just requires attention and education and monitoring.
I think it can occur across a wide spectrum of clinical manifestations, the majority of which are manageable in my own experience.
What do you, just personally, in terms of, again, given your vast experience with this combination, can you comment on how the management has changed, or do you think it's improved over time? Where are we now as compared to the beginning of the phase I study?
Yeah, I think we're understanding sort of the kinetics of side effects and having greater confidence in sort of these early descriptions. I guess I'll add a little bit of color to that. With zolbetuximab, for example, with larger experience, we learned very clearly that the toxicity is very early, often with the very first dose or the first couple doses of drug. That can be a barrier and deterrent for some patients because they have a toxicity before they know if they're having a benefit. Sometimes it's hard to have that discussion with people if they're not feeling well. The gastritis, with more mature data, in my own experience, the time to onset is well after the patients have had at least one or two scans.
It's very helpful in the discussions because patients have a sense of whether they're benefiting from the combination. I think when patients understand or have some objective evidence that they're responding to the combination, they're certainly invested and inclined to try to adjust around any side effects that may be upcoming. In some ways, this has been a nice thing to see, sort of the greater understanding of the timing and sort of onset of this adverse event profile in some ways itself is a bit favorable in terms of having patient discussions. That is one piece that I think has really started to come out more clearly with longer follow-up.
We also have a greater experience in managing the side effects because we have followed these patients for longer, so we have an experience of re-challenging and continuation of other components of the therapy during management of gastritis, which I think is another nice component to the combination, is that it's modular, meaning that you have chemo, checkpoint inhibitor, and givastomig. This is different than some other agents like the ADCs, which are either given as single agent or, in some of the ongoing trials, combination, but the ADC is being asked to do a lot of the work. If you hold the ADC in some of the later line studies, the patients are getting no therapy at all. Here, if we have to hold givastomig for a period to manage gastritis, for example, there is still ability to continue components of therapy.
I think that's important for practical management of this disease.
Great. Thanks. I just want to touch on the topic of re-challenge. Obviously, that's an optimal outcome, that if a patient develops gastritis and they have the immunotherapy components withheld for a short time while their symptoms improve and they're treated appropriately, how do you see re-challenge evolving over time? Right now, we have slightly more than half the patients that can be re-challenged. Do you see that improving over time with better management?
Yeah, I think so. I would say it's a discussion with patients. Obviously, they're going to bring their own perceptions and how the toxicity may have impacted them to the discussion, and that's very important. We have discussions of re-challenge with chemotherapy components that also have similar themes. I would say, in general, the re-challenge has been at 5 mg per kg, although I have re-challenged patients at full dose as well. My own experience with, I think, is probably three to five patients with re-challenge. I think maybe I've had one person who developed recurrent gastritis that was manageable with hold. I think, again, we need larger numbers of patients to sort of get a little more granular data around the chances of recurrent gastritis.
But thus far, in my own experience, re-challenge is something that if I think the patient is clearly benefiting and the toxicity has improved to Grade 1, like how we would think for other immunotherapy-related toxicities, I feel pretty comfortable having that discussion and thinking towards re-challenge. Particularly if it's relatively early on, like the first six months of someone's therapy, where the ideal activation of the sort of immunologic component of this combination may still drive a lot of the benefit, I would seriously consider it in patient discussions.
Great. Thanks. Maybe as a closing, again, finishing with a broader question, knowing your involvement and leadership in the study, if you compare where the program is today versus a year ago, what's changed most in your own thinking about Giva and its future potential?
Yeah. It's a good question, and it'd be nice to play this back in like five years and see how close I was to feeling the same way. I think the longer follow-up has been quite nice, and I've been involved in the program for a few years now. I think we have really seen no major changes in terms of particularly toxicity, and I'm feeling more confident across sort of the spectrum. But I don't think there's a specific result that drives this. But if I look at the entire space as a whole, I think there is some clear positive differentiating factors. One, as I mentioned, is that the modular components make this very manageable over the course of someone's frontline therapy, which is something that some other agents and approaches can't really offer.
I think the ability to have data all the way down to the very low claudin expression and PD-L1 negative patients is something that we have not really seen yet. And even the ILUSTRO data set, really the benefit was primarily in the high claudin and PD-L1 positive, so that's quite different than the givastomig data set, where the activity looks consistent across the entire claudin and PD-L1 spectrum. So that is a nice thing. And then the sort of continued validation of claudin as a target has been something that is very, very clear. I mean, at ESMO, we saw two late-line phase III trials from ADCs that continue to validate the idea of pushing on claudin, and then we will see data from additional phase IIIs. But the mechanism is also different, and I think that that is important.
I view givastomig as a potential way of solving the 4-1BB agonism question that has been a very active area of science and has very good scientific rationale, but hasn't been a problem that we have solved clinically. Meaning the naked 4-1BB agents were limited in toxicity, so a spatially delivered 4-1BB agonist seems to get around that. I forgot to mention in the toxicity that one thing that doesn't happen, at least in my experience, is really hepatotoxicity. We don't seem to be plagued by some of the problems of the naked 4-1BBs. I think that is probably reflected in some other companies trying bispecific 4-1BB formulations such as PD-L1 by 4-1BB and others.
I think that's been a nice thing to see in the space, is that we may, or you guys may have partly solved, if not solved, the ability to deliver 4-1BB, because it's always been an attractive immunologic approach. It's just been limited by toxicity. Maybe I got a little bit off on a tangent, but I think that that is something that has been a nice thing to see.
Well, I appreciate it, Sam. This has been a great conversation. We appreciate your insights and your leadership on the study. Thank you.
Thanks a lot.
As we discussed, a key takeaway is that this increasing mature dataset has strengthened our confidence in the attributes we set out to achieve: broad activity, durable clinical benefit, a manageable safety profile, and an optimized benefit-risk profile at the selected 8 mg per kg dose. With that context in mind, I'd like to turn the discussion back to Srishti to talk about the strategic implications of these findings and how we think about the opportunity ahead for Giva. Srishti, over to you.
Thank you, Phil, and thank you, Sam, for helping put these findings into clinical context. Before we move to Q&A, I'd like to leave you with a few takeaways from today's discussion. A year ago, we were encouraged by the early activity we were seeing with givastomig, and today, we have substantially more follow-up and a much deeper understanding of the durability, safety, and overall clinical profile of the program. Perhaps most importantly, the dataset has strengthened as it has matured. With longer follow-up, we continue to see durable clinical benefit, a manageable safety profile, and activity across the CLDN18.2 positive spectrum. The updated dataset has also given us the information needed to select the 8 mg per kg dose and continue advancing our registrational development plans. As the CLDN18.2 treatment landscape continues to evolve, we believe that the profile emerging for givastomig is increasingly important.
The combination of durable clinical benefit, activity across the CLDN18.2 expression levels, and a favorable benefit-risk profile gives us confidence that givastomig has the potential to differentiate itself and play a meaningful role in the treatment landscape. There is still important work ahead, but we believe the program is increasingly demonstrating the attributes we hoped to see when the study was first initiated. For NovaBridge, this ESMO update represents an important milestone. We have a more mature dataset, a clearer understanding of givastomig's clinical profile, and greater confidence in the path forward. With that, Tara, we'd be happy to take questions.
Great. Thank you, Srishti. Yes, at this time, we'll be conducting a question- and- answer session with our speakers. Please hold for a brief moment while we poll for questions. Our first question comes from Daina Graybosch at Leerink. Please go ahead, Daina.
Hello. A few questions from me, but maybe we can start on toxicity, maybe a few there. On the poster, if I look at the 8 mg per kg, the numbers of Giva delay, interruption, and discontinuation, you reported at 55%, 23%, and 26%. Three questions there for the company and Dr. Klempner. One, are those the same patients that we are seeing interrupt and discontinue, or do we tend to have one of those happenings, you delay or you discontinue? Just trying to understand the totality of the number of patients that have any of these treatment-related changes to their dose. Two, I think you talked about two things make you encouraged about this toxicity, and I wonder if you can quantify it.
One, how this number changed, let us say, from the first 15 patients enrolled to the second 15, or however you want to cut it, to help us quantify that the management improved these numbers. And three, you talked about these coming late, but you do not have on here sort of a duration of treatment or exposure. Do you have any quantification about sort of when these dose changes happened in the treatment course? Thank you for that.
Thanks, Daina. Thanks for the questions. Maybe I will hand it over to Phil for the first part of the first question around understanding sort of the definitions and the totality and the patients, how we define them in the toxicity table for the poster. Phil, do you want to jump in on that?
Sure. Thanks. I will start with some definitions. Dose interruption is defined literally as an interruption of the infusion, right? Patients start their infusion of their different drugs, right? As Sam mentioned, several modalities. In the course of that infusion, if they develop an infusion-related reaction, then the infusion could be interrupted, medications can be provided, and then the infusion can continue, often at a slower rate. But the clinical impact of those types of interruptions is typically quite low. Dose delay indicates that a patient who was intended to get a cycle of treatment has a delay in that cycle of treatment. And this is not uncommon, again, given two different immunotherapy agents combined with a chemotherapy regimen as well that is quite complex and has its own set of toxicities.
For perspective, in terms of dose delay, CheckMate 649, which is a study that led to the approval of nivolumab and chemotherapy, had a dose delay incidence of 67%. It is actually quite common to really experience that. Maybe I want to, since Sam has lived through this by treating the patients, I would like to know, Sam, if you could just comment on that experience of dose delay and dose interruption.
Yeah, sure. Happy to. Few things. One, this is a problem in line of sight for all clinical data sets. It would be wonderful to see by cycle the rates of one component being held. We almost never see that, even in big registrational phase III, even though the data exists. This is partly the way that AEs are captured in the CTCAE, and partly just management of the data sets. I would say clinically, obviously, I have treated a large number of patients with this drug, at least 30 or 40 people. Yes, there are some of the same patients who have delayed, ultimately leading to a discontinuation, but many patients, that is definitely not the case. I think I cannot answer fully from the data if it is the exact same patients who interrupted that discontinued, but I agree, interruptions happen all the time.
Even if a patient is going away for the July 4th to have family barbecue, sometimes these things get recorded as an interruption, which is not held for a toxicity reason. Although, in the table, this is treatment-related adverse event, but those can be really a huge spectrum. I think it is incomplete to try to draw conclusions from that. To answer your earlier question about the evolution of the toxicity and any formal metrics to understand that, I will let Phil comment, but in my own experience on this trial, I think the recognition and management has led to a materially different ability to maintain people on components of the therapy, including givastomig. I think I maybe only discontinued completely one or two people out of my entire cohort, and some of that is experience with these types of drugs.
Some of that is the excellent support that we have at big tertiary centers. I do think it is materially different from the initial 15 to this 71 patient data set.
If I could just add to that. The discontinuation rate that we show in our poster is accurate, and it reflects, for example, three incidents of infusion-related reaction that happened within the first 8 months of the combination study. We have not had an occurrence of an infusion-related reaction that has led to discontinuation since then, due to the simple fact that we now have premedications and we did not have them at the time that those events led to discontinuation.
The way, if you exclude those patients, our discontinuation rate is 16%, and if we look at patients who discontinue because of gastritis, it is 13%. I will emphasize also that the timing of gastritis is getting shorter and shorter. If we look at Grade 3 events that have happened over the course of the last year, the average time of Grade 3 events is now less than 10 days.
We are decreasing the If patients experience gastritis, especially Grade 3, which are very rare, right? The clear majority of cases are Grade 2 or less. Our Grade 3 incidents at the dose that we are taking forward registrationally, 8 mg per kg, is 6% Grade 3 events, and we know that those very rare events now, luckily, are happening for a much shorter period of time. That is due to better management, early diagnosis, early intervention, slow steroid taper, and really just judicious management.
Thanks, Phil. Then maybe to answer Daina's last question, from the phase I data, what are we gleaning or how can we sort of make a comment on the duration of treatment and the direction of that? Even though we may not have an exact, obviously, the phase III will be important for that, but how are you thinking about the duration of treatment?
I guess one of the ways to think about this practically is that we know that for patients who do have dose delay, we know that the impact of that now is less, right? We know that the impact is arguably not It does not cause patient detriment, right? In other words, we know that patients who develop gastritis do not do worse. We know that treatment, delaying for a couple of cycles, letting gastritis quiet down over a matter of a few weeks does not adversely affect outcomes. In the phase III study, we recognize the burden of really the safety issue for patients in a larger geographically diverse network. But again, our phase I experience has informed earlier recognition, diagnostic criteria, and intervention guidelines.
We will plan on investigator education and standardized treatment management protocols that are now included in our phase III study as we get ready for that. We want to ensure that patients maintain treatment, but if they do have to withhold treatment due to gastritis, which again, is a very important point, it develops four to five months in after a tumor response is typically initiated, that those patients can have that resolve. As we now have experience in Sam's institution, actually, I believe can restart 67% of patients now on givastomig after an episode of gastritis. We think that as with this implementation, especially in our phase III, gastritis will have a very minimal burden in such a registrational study.
Thanks, Phil. Tara, can we go to another analyst and then maybe come back if Daina has further questions?
Yes, absolutely. Our next analyst that we have on is Andres Maldonado at H.C. Wainwright. Andres, please go ahead.
Hi, guys. Congratulations on the progress, and thanks for taking my question. I will start with two, one for the company, one for Dr. Klempner. I guess let's start with Dr. Klempner. If the randomized data holds up, could you paint a picture of which patients would you reach for givastomig first, and maybe which would you still favor zolbetuximab or another kind of chemoimmunotherapy option alone? Then for management, just curious on how many patients with CLDN18.2 low and PD-L1 low, just curious if you can give us a little bit more color on how they are doing. Essentially, that seems to be a particularly important population for givastomig. Essentially, the question is, how will the phase III really establish that benefit in those low patient expressors rather than just being driven by that overall result by the high expressors? Thank you.
Thank you, Andres. Dr. Klempner, handing it over to you for the first question.
Sure. Multi-part question. Who will I give the drug to if it was approved? I think, in some ways, the market is speaking a bit. The adoption of zolbetuximab and implementation in the U.S. has been dramatically smaller than the predicted patient population just from prevalence of the biomarkers. I really do think that that is driven by the early toxicity and the perception of patients. It's just unlucky for them that the very high rates of nausea actually come with vomiting during the infusion. It is unfortunately a big turnoff for patients to want to come back.
As Phil alluded to, I think while givastomig certainly has a known toxicity profile and it can impact patients, the timing is a big deal because if you have a patient who's been on therapy for six weeks, three months, and they've had two scans, and they know that their tumor is shrinking, as most patients have, they are very incentivized and motivated to ride out and manage actively any side effects. That just doesn't happen with zolbetuximab. I suspect LUCERNA will be a positive phase III trial in the PD-L1 positive claudin high population that they're targeting. Even if that is the case, if givastomig were also a positive phase III trial, I think that it would cannibalize a large portion of the zolbetuximab market.
I've participated in the phase III zolbetuximab trials, participated in the AZD0901, sonesitatug vedotin studies, and those ADCs may come to the first line. I think they're problematic to combine with chemotherapy in first line, and asking an ADC to completely replace oxaliplatin is still a relatively high bar. This was my comment about sort of the modular features of having three drugs and a standard backbone chemo that everybody's comfortable with. Short answer, I think it would claim most of the frontline patients because the prevalence of claudin 1+ and 1% or more is like 80% of every patients that walk in the door. Your question about splitting the data by claudin levels and PD-L1 levels, I mean, as you know, the numbers are a little bit limited.
There is data from the 71 patients presented at ESMO looking at claudin above 75% and 1%- 74%, but it is limited by small numbers and some overlap, which perhaps is encouraging that there doesn't appear to be a big difference there. The other thing is that there's actually not that many PD-L1 negative patients in my experience, about 15%-20% of gastric cancer. It's always a hard subgroup to pull apart, but I don't think that there's any suggestion so far that says there's a clear difference in activity or there's a clear population that is driving the benefit. Ultimately, that could be true in the setting of phase III, and there's always some patients that tend to do better, but it's not always limited to biomarker. Sometimes GE junction does worse than gastric in many of our data sets, and some of these are not clear.
biology or whether it's other features that tend to co-migrate with those anatomic restrictions. I'd say short answer, hard to answer, hard to give you a clear answer on a subgroup that's driving the benefit, but a phase III trial would have the numbers to give a little bit more confidence around point estimates for some of those subgroups.
If I can add just some-
Phil, I think you can add on. Yeah, just a couple quick add-ons.
Yeah, just two very quick add-ons, Andres. The first is that 55% of our patients at 8 mg per kg had peritoneal metastasis, a notoriously difficult group of patients to manage. I think that's an important prism to also look through our current data. Finally, we do recognize those double lows, if you will, even though they're very rare, to Sam's point, low for PD-L1, low for claudin. We have a cohort in our ongoing phase I combining givastomig with chemo alone with no checkpoint inhibitor because a checkpoint inhibitor is not indicated in that group of patients, and we're seeing very promising early data. We hope to, again, as a life cycle management, we hope to be able to move into a PD-L1 negative group as well.
Thanks, Phil. Tara, are there other live questions?
Yes. We have one more from Daina Graybosch at Leerink. Daina, you can go ahead.
Yeah. Thanks for another follow-up. I wonder if you could just give us some more context on the upcoming registration trial, where you are in finalizing that protocol with FDA and global regulators, and when you expect to begin the study.
Thanks. Phil, do you want to jump in on some of the agreed-upon points with the FDA so far?
Sure. We've had several very productive meetings with the FDA. We've got alignment on key portions of the design. We've stated publicly we have two paths for accelerated approval for ORR and PFS. It's clear that the FDA wants overall survival as an endpoint that will be required for full approval. Again, we've gotten full endorsement of our regional distribution, the statistical details of the study, so we're very encouraged by that, and we're in the process now of finalizing some final features to really ensure the study is built for success.
Today's focus, Daina, was really on the dataset. We're finalizing the execution plan based on where we have feedback from the FDA, and we'll update on the trial start as soon as possible. But as Phil says, given Giva's long-term clinical and commercial value, we want to make sure that we design an execution plan that maximizes its ability to get to patients. We look forward to updating on the timeline. Tara, I think if that covers live questions, were there any questions that came in via the platform?
Nope, that's all we have time for, so I'll turn it back to you for some closing remarks.
Perfect. Thank you so much. Thank you, Phil, and thank you, Sam. Also, thank you everyone for joining, both in the evening as well as in the morning. If there's one thing I hope you take away from today, it's that givastomig has reached an important point in its development. We now have a substantially more mature clinical dataset. We have a greater understanding of the benefit-risk profile and the regulatory feedback that we'll need to inform the path forward. This gives us a very strong clinical development and regulatory foundation to advance givastomig into a registrational path forward. We're really excited about what lies ahead for the program, and we look forward to keeping you updated on the progress. Thank you again.