Gritstone bio Inc (GRTS) 2023 Q1 法說會逐字稿

My name is Erica, and welcome to Gritstone bio's First Quarter 2023 Earnings Conference Call. Please note this event is being recorded. At this time, I'd like to introduce George MacDougall, Director, Investor Relations and Corporate Communications at Gritstone. Please go ahead, sir.

Thank you, Erica, and thank you, everyone, for joining us for Gritstone bio's conference call to discuss our financial results, clinical and business updates for the first quarter of 2023. With me on the call today from Gritstone Bio are Andrew Allen, Co-Founder, President and CEO; and Vassiliki Economides, Executive Vice President and Chief Financial Officer.

Today, after the market closed, we issued a press release providing our first quarter 2023 financial results as well as clinical and business updates. The press release is available on our website. I'd like to remind you that today's call is being webcast live via a link on Gritstone's Investor Relations website, where a replay will be available after its completion.

After our prepared remarks, we will open up the call for Q&A. During the course of this call, we will make forward-looking statements that are based on current expectations. These forward-looking statements are subject to a number of significant risks and uncertainties, and our actual results may differ materially from those described. We encourage you to review the risk factors in our most recent Form 10-K filed with the U.S. Securities and Exchange Commission and available on our website.

All statements on this call are made as of today based on information currently available to us. Except as required by law, we disclaim any obligation to update such statements even if our views change. With that, let me turn the call over to Andrew. Andrew?

Thanks, George, and good afternoon, everybody. Thanks for joining us for our first quarter 2023 conference call. This is a really exciting time for Gritstone. Significant momentum has been established for our personalized cancer vaccine, or PCV program called GRANITE, which is driving towards randomized clinical trial data. And with infectious disease, we're making great strides in realizing the untapped potential of self-amplifying mRNA. I'll begin today's call with a review of recent clinical and corporate developments, including the expansion of the GRANITE study we announced today. Celia will then present our financials, and I'll come back to share closing remarks. Okay, let's dive right in.

So most important to share with you today is the update regarding GRANITE, which is currently in the Phase II part of a Phase II/III study in patients with newly diagnosed metastatic microsatellite stable colorectal cancer. I'll refer to this as MSSCRC. This morning, we shared an enrollment update from the study. And I'm delighted to tell you that we're enrolling at a rapid pace and very close to achieving our initial target enrollment of 80 patients. We also announced that we've made the strategic decision to expand the Phase II portion of the Phase II/III study to 100 total patients. And in order to do this, we'll be shifting near-term funds from our off-the-shelf neoantigen program, SLATE.

SLATE is a key part of the future, but our planned randomized trial is not expected to deliver data near term, of course, and thus must play second fiddle to GRANITE as we maximize the potential of this high-value PCV opportunity. Now over the years, you've heard me speak of our strong belief that our PCV approach could be uniquely capable of unlocking the power of immunotherapy for patients with immunologically cold tumors.

Our Phase I/II data from GRANITE in advanced cancer patients that included immunologically cold tumors, published in Nature Medicine last summer, provided strong clinical data in support of the therapeutic hypothesis. Moderna's recent positive Phase IIb data presented at AACR in April provided further evidence supporting the PCV approach in Moderna's case, in the immunologically hot context of adjuvant melanoma. And then just yesterday in Nature, Vinod Balachandran and his team from Memorial Sloan Kettering, published data in the adjuvant therapy of pancreatic ductal adenocarcinoma with some provocative immunology and clinical data, potentially further supporting TCV, this time in the cold context of pancreatic cancer.

Our current GRANITE trial is how this all comes together, with our Phase II study potentially being the first to deliver randomized data supporting a neoantigen-based PCV against MSSCRC. And I think it's widely recognized that if GRANITE delivers benefit in metastatic MSSCRC and notoriously cold tumor that is difficult to treat with immunotherapy, the opportunity for further application is enormous and potentially encompasses all solid tumors, both hot and cold.

Now there are several reasons why we elected to expand the Phase II study. Let me detail them. Firstly, a larger number of patients will provide a better estimate of treatment effect size, which will, in turn, inform an optimal Phase III study sample size. So to drill down on that point, the current study is designed as a Phase II/III. And to remind you, the primary efficacy endpoint for the Phase II component is a molecular response or ctDNA response, with secondary endpoints, including progression-free survival, or PFS, and overall survival or OS. Now we anticipate discussing the Phase II data with regulatory authorities in the first half of 2024 and then moving into Phase III.

The Phase III primary efficacy endpoint is to be determined based on our interaction with regulators, but it is likely to be a traditional endpoint such as PFS or OS. Powering the Phase III study for a time-to-event endpoint will be done using observed Phase II data. Since the Phase II and III components are statistically separate, the greater the precision of the estimate of treatment effect size from Phase II data, the easier it is to power Phase III appropriately.

Secondly, momentum behind study enrollment has been very strong. As of May 10, 2023, we've randomized 71 of the initially planned 80 patients, and that number has already gone up, and approximately half of these were randomized in 2023. All sites are screening and consenting enthusiastically, and we expect this strong enrollment trend to continue for the foreseeable future. We anticipate completing enrollment of the full 100 subjects in the third quarter of 2023. Now sadly, worth reflecting on, this momentum arises in part from the high unmet clinical need. Colorectal cancer is the second leading cause of cancer death in the United States, and over 150,000 people are expected to be diagnosed with CRC this year alone. Median survival for patients with metastatic disease is around 2 years. Better therapeutics are desperately needed. Innovation is needed, and we aim to deliver both.

Thirdly, GRANITE is the tip of the spear for our neoantigen programs. The basic idea is that delivering many neoantigens to solid tumor patients in a potent vaccine that drives strong CD8 T cell responses will drive clinical benefit. Now Granite seeks to do exactly that, leveraging our neural network epitope identification platform, Edge, to accurately predict which subset of tumor DNA mutations forms true neoantigens that will stimulate an effective antitumor immune response. And our published clinical data show that patients mount strong T-cell responses to most of the administered neoantigens, with T-cells trafficking into their tumors and killing tumor cells, as evidenced by the approximately 50% molecular response rate that we've observed. And as you are well aware, our data further show that those patients with molecular responses experience extended overall survival versus those without.

However, a PCV approach will always take more time and cost than an off-the-shelf approach. And we're rapidly building out our off-the-shelf program, SLATE, which delivers shared tumor-specific antigens to each patient, but it is temporarily behind GRANITE in development. It takes work and time to identify high-quality shared tumor-specific antigens that can be combined into a single vaccine that delivers multiple relevant antigens to each patient. So where GRANITE leads, SLATE will likely follow. However, showing that GRANITE delivers efficacy in a randomized controlled trial is our highest near-term priority.

Now before we move on from GRANITE in oncology, I'd like to draw your attention to the value of circulating tumor DNA or ctDNA as an efficacy biomarker. Reduction in ctDNA referred to as a molecular response, is rapidly emerging as a likely superior surrogate biomarker of efficacy over radiology in the early assessment of solid tumor immunotherapy. It makes particular sense for us at Gritstone to assess molecular response in our studies because we aim to drive T cells into tumor lesions, where we have shown that they can meet cognate antigens and proliferate, potentially making the lesions appear bigger. This obviously makes radiology-based tools susceptible to errors of mislabeling.

lesions that increase or remain stable in size are labeled as progressive disease or stable disease, respectively, when in fact, the patient's immune system may be newly at the wall with their tumor, i.e., anything but stable. ctDNA dynamics are casting light on the key issue of what's actually going on in those complex lesions. Tumor cell destruction is reflected by a decrease in ctDNA over time. Now the molecular response is defined as a 30% reduction in ctDNA as the primary endpoint for the Phase II portion of our GRANITE study. We originally powered the study for this endpoint, and we've waited for mature clinical data from our Phase I/II study in advanced cancer patients to allow us to carefully define molecular response in a data-driven manner that is expected to best predict overall survival in colorectal cancer. Now to remind you, no one has done this before. So we've had to build the framework ourselves.

No one's ever generated strong de novo neoantigen-specific T cell responses in cold tumor patients before. No one's combined this with checkpoint inhibition. So there is no off-the-shelf answer to the question, what's the best surrogate endpoint for overall survival? We've had to find the answer and that's exactly what we've done. We've generated data in Phase I/II, and we're now deploying a data-driven definition to establish efficacy in a randomized controlled Phase II trial.

Once efficacy has been demonstrated in Phase II, will proceed into a pivotal Phase III. And as I've already noted, we anticipate using traditional efficacy endpoints such as PFS or OS in the pivotal trial, but there is a possibility that molecular response will have accrued enough support by 2024 to be used as a surrogate endpoint for accelerated approval. We would like that, but we are not dependent on it. We're very excited by the GRANITE Phase II trial, the momentum it's gained, and its potential to serve as a major advance in cancer immunotherapy as we seek to extend the emerging benefits of PCV to cold tumors. We will update you further on enrollment once complete, and we look forward to sharing preliminary data in the first quarter of 2024.

Now let's move over to infectious disease, where data demonstrating potential advantages of self-amplifying mRNA or SAM RNA, over first-generation mRNA vaccines continue to flow from our clinical-stage coral program centered on SARS-CoV-2. And the differentiation of SAM RNA from mRNA is now becoming clearer. The prolonged antigen expression associated with SAM RNA vaccines appears to be translating into a more durable neutralizing antibody response, potentially reducing the need for frequent boosts of vaccinations. And the replication of the RNA once in the cell appears to enable similar immunogenicity as nonamplifying mRNA at a fraction of the dose, offering a cost of goods benefit.

In April of this year, we presented new 6-month neutralizing antibody data from our ongoing Phase 1 coal boost and ColSEPI studies at the 33rd European Congress of Clinical Microbiology and Infectious Diseases, also known as ECMED. These presentations in Copenhagen demonstrated that the robust neutralizing antibody responses driven by our SAM RNA vaccine candidates persist for at least 6 months regardless of setting and across multiple subject populations.

Of note in the cola presentation, we reported high neutralizing antibody levels and persistence at 6 months in over 100 vaccine-naive subjects, a meaningful increase in subject numbers versus our prior data reports. SAM RNA is rapidly emerging as a tolerable immunogenic and flexible next-generation platform for vaccines against infectious diseases, including COVID-19 and the induction of broad, long-term and potentially variant-proof immune responses that we're seeing in our Phase I studies is highly promising for the field. We look forward to continuing to work with collaborators to demonstrate the full potential of our SAM RNA platform against both SARS-CoV-2 and other important viruses. We expect to share additional data from our core program this fall, and these data will relate to different immunogen designs, illustrating the flexibility of the platform to accommodate both B-cell and T-cell epitopes in efficient formats.

It is clear that there is still need for next-generation solutions against COVID-19 and the recent actions by the White House and BARDA are encouraging signals that the pursuit of enhanced breadth and durability of protection is not going to the wayside. Outside of our PCV and SARS-CoV-2 programs, our forward-looking efforts to identify and develop potentially transformative vaccines continues. Our partnership with Gilead to develop a vaccine-based curative immunotherapy treatment for HIV remains active and ongoing in a Phase I study. And promising results from a preclinical study in nonhuman primates within this program were presented at the conference on retroviruses and opportunistic infections in January. And our preclinical projects against human papillomavirus, influenza, and a new combination vaccine against multiple respiratory viruses continue to progress well. I'll now turn it over to Celia, who will provide more color on our financial results for the first quarter of 2023. Celia?

Thank you, Andrew. Good afternoon, everyone. We maintained a strong cash position in the first quarter of 2023 and have been able to extend our runway for multiple sources, adding this recent quarter with $153.2 million in cash, cash equivalents, marketable securities, and restricted cash. Our runway has now been extended into mid-2024. In addition to utilizing our ATM program from time to time, you'll recall we also have an $80 million credit facility with Hercules Capital and Silicon Valley Bank, a division of First Citizens Bank.

During the first quarter, we drew an additional $10 million from the facility, totaling $30 million to date. We also have the option to draw an additional $10 million before the end of 2023, and the remaining $40 million will be available through June 15, 2024, upon the achievement of certain milestones by Gritstone.

While we have taken steps to extend our runway so far in 2023, we also have multiple potential avenues to secure additional capital. These include but are not limited to, drawing down additional funds from our existing credit facility, establishing new or expanding existing collaborations, and other nondilutive funding sources such as a potential $40 million milestone payment from Gilead on our HIV cure partnership. Turning now to our first quarter 2023 operating results. We reported a net loss of $34 million compared with $28.9 million for the same period last year.

Our research and development expenses were $30.5 million for the 3 months ended March 31 and 2023 compared with $28.2 million for the same period in 2022. The increase in R&D costs was primarily due to increases in personnel-related costs and laboratory supplies. Specifically, for our GRANITE program-related expenses, we incurred $5.4 million in the first quarter of 2023 compared to $2.7 million for the same period last year.

The increase was primarily due to the continued execution of the ongoing Phase II/III trial in first-line maintenance, MSSCRC. For the period ending March 31, 2023, we were able to reduce general and administrative expenses to $6.7 million from $8 million for the same period in 2022 by managing our spend on outside services and facilities-related costs. You may recall that the establishment of collaborations and partnerships is a core part of our business strategy as we continue to leverage our unique platforms and capabilities. Gritstone's collaboration license and grant revenue for the first quarter ending March 31, 2023, totaled $2.4 million compared to $7.2 million for the same period in 2022. These also include collaboration revenue of $0.1 million related to the Gilead agreement and $0.4 million related to the 270 Bio agreement.

This also includes grant revenue of $1.5 million from the CEPI funding agreement and $0.4 million from the Gates Foundation. Finally, as of the end of the first quarter 2023, Gritstone had 87,848,417 shares of common stock outstanding and pre-funded warrants outstanding to purchase 13,573,704 shares of common stock at a nominal exercise price of $0.01 per share and 13,274,923 shares of common stock at a nominal exercise price of $100 per share.

In summary, we are pleased to have maintained a strong cash position and extended our runway during the first quarter of this year and are confident in our ability to execute on our strategic objectives. I'll now turn the call back over to Andrew for some closing remarks. Andrew?

Thanks, Celia. So as described today, happily, Gritstone is in a period of significant momentum. Robust enrollment within our PCV program, GRANITE, has enabled study expansion and the generation of additional clinical and scientific data to inform the future development of GRANITE and further validate our approach to solid tumors. Within infectious disease, we're pioneering a novel technology that could represent the next RNA platform approach against SARS-CoV-2 and beyond. We look forward to continuing to share our findings with you as progress continues. And with that, I'd like to thank you all for joining us today, and I'll now turn the call over to the operator for questions. Erica?

(Operator Instructions) Our first question comes from Marc from Cowen.

Congrats on all the progress and it sounds like some pretty exciting enrollment numbers that you're seeing these days. Maybe just to start off on the change in trial size. Can you talk through the kind of process that led to that decision? Was this initiated entirely by Gritstone or is this something maybe regulators ask you to think about? And then can you walk through the kind of the powering assumptions that are now built into the trial given the new trial size?

Yes. Thanks, Marc. This was a Gritstone-generated decision. Phase II basically is up to us. And obviously, what we're trying to do is optimize for Phase III now. And we are in the fortunate position that just a short period of extended open enrollment will put in a meaningful number of incremental patients. And it's really that combination that drove us to make this decision. Now the primary endpoint of Phase II, as you're aware, is molecular response. That's a binary outcome, and we are extremely well-powered to detect at least a 20% difference between the 2 arms of the study.

We're in the high 90s power for that endpoint. What we're now more interested in actually is looking beyond the primary efficacy endpoint for Phase II and starting to think about planning for Phase III. And of course, Phase II will all be about the size of the study. And so the current thinking, and let's recognize, we have not yet discussed this with the agency, but the way we approach this is to assume conservatively that we'll be using a traditional endpoint. The most traditional will be OS. Obviously, PFS has a good history in colorectal cancer. And we may be able to use PFS if we don't see too much pseudoprogression. So the data will help us make that decision.

But then you have to just think about the powering of the Phase III study. And obviously, more is always good from a power perspective, but more slower and more expensive. And so you want to size it appropriately. You don't want to overpower but you don't want to underpower Phase III. And the best way to gauge that is to have a good estimate of that endpoint from your Phase II study. And so really, that's what we're trying to accomplish here. By putting on this incremental 20 patients, we both increased the power against PFS and OS, although those are only secondary endpoints. But perhaps more importantly, we'll have a more accurate estimate of treatment effect size, which then feeds directly into the sample size for the Phase III study.

And the way to think about it, I think, is that you can make a reasonable assessment early on. And then as you launch the Phase III study as the Phase II data mature, you could then do a sample size reassessment based on the Phase II data to modify the Phase III sample size to more fully reflect mature Phase II data. And obviously, because these are statistically independent elements, there is no problem statistically speaking with doing that. And so this is the journey we're on now. So this is all really about optimizing the Phase III study, assuming we use a traditional endpoint.

Okay. That's all very helpful. And can you ask exactly how you have changed the definition of your response? Or is it still the same response you were using before? I know that there's a lot of debate in the field that's exactly where to set that threshold.

Yes. We basically had a straw man that we put in with very little data. Our intention was always to do the following: to collect samples from the ongoing Phase II study, but just bank them, so not analyze them. So we've not analyzed a single sample from our ongoing study for ctDNA. What we were waiting for was the data from our Phase I/II study to mature because obviously, what we're trying to do is use ctDNA response as a surrogate for overall survival. And to do that with our therapy in the context of colorectal cancer. And we have not a huge amount of data, but we have data from that Phase I/II study to help us optimize that definition of molecular response in service of the goal of best predicting overall survival. So we've done all that work now. We have mature data from the Phase I/II study, and we've set the definition, which is a 30% reduction in ctDNA from the baseline.

And using our data that is best associated with the prediction of overall survival. Obviously, you could set a lower threshold or you could set a higher threshold. We've been data-driven, and it's all about just optimizing the sensitivity and the specificity.

Our next question comes from Ted Tenthoff from Piper Sandler.

Thanks for the update on GRANITE and I'm very clear in terms of the increase in number of patients. I wanted to ask on SLATE and understand, as you said, the take-back seat. What is the take to really accelerate that or get that moving in the future? So in other words, what does the pause cost of?

Thanks, Ted. I think obviously, any delay pushes things back. Let's be clear-eyed about that. But the momentum is building in the field. We've obviously seen, as I mentioned in our prepared remarks, the Moderna data, yesterday, Nature published the data from the Memorial Sloan Kettering team, and that's using the BioNTech vaccine approach. Those data were presented at ASCO, but they were published yesterday. And then we have our own data published in Nature Medicine. And the biology, I think, actually holds together remarkably well. And so what we're seeing across all of these different platforms, we have a little bit less data with Moderna because we're waiting on T cell data from the donor, although we are told to expect that maybe at ASCO sometime around the middle of the year.

But what we can see from our data and from the BioNTech data, is this common biological theme, which is that if you drive good, strong neoantigen-specific T cell responses measured using a traditional not very sensitive assay, typically it's ELISPOT, then those patients that develop neoantigen-specific T cell responses of good number experience clinical benefit. That's the common thread that spans our disparate data sets, raised different products in different tumor types. But if you read the Nature paper carefully, which obviously I've had a chance to do, then you can see that the patients in whom they were able to generate a T cell response did then appear to derive clinical benefit, which is obviously consistent with our data.

Now we drive a good T cell response in just about everybody we've vaccinated and that's not true with the BioNTech data, they get a measurable T cell response in about half of their vaccinated subjects. They see a benefit in half. So there are some differences. And one can't tell from the published data, whether the differences relate to the neoantigen prediction or the power of the vaccine vector, or both, but the outcomes are highly consistent. And these are both cold tumors. So I think the notion that this platform, this approach has now matured to a point where it's really starting to deliver clinical benefit.

I think that theme is now becoming widely accepted. And that, of course, will help enrollment. So yes, we'll lose a bit of time on the front end. But as momentum continues to build behind these programs and this concept, I think a delay of a quarter or 2 quarters is not really insurmountable. That will shrink given, I think, the momentum benefits, and maybe you'll end up with data a quarter later. And of course, if we're able to secure additional capital, you just put a few more sites on, and you basically end up back where you started as it were because you'll just enrolled that certain number at a faster rate. So I think it's not a huge blow for the program. Obviously, today, we're in capital conservation mode like every biotech company pretty much. And this, for us, was a rational prioritization of our resources.

It makes a lot of sense. And I know that you guys were doing some optimization work on the antigen, is there any update there? Does that work continue? Or is that sort of put on hold now? Or have you even finished some of that optimization?

No, great question. The optimization continues. So our research team in Cambridge, Mass, this is a big effort for our company because, as I think most people anticipate, having multiple relevant antigens in an off-the-shelf vaccine is highly desirable. If you step back, we treat viral infection with combination drugs because the virus can mutate around a single drug. It has a problem mutating around 3 different drugs all being used in concept. And the same is true in cancer, of course. We use combination gemmotherapy for the same reasons. And so the general belief is that having a single antigen in a vaccine, while it may be effective, is likely to lead to some degree of acquired resistance where tumors can mutate around and edit that antigen.

And therefore, delivering multiple antigens is highly desirable. Now with GRANITE, we do that because these are personalized products. So it's 20 customized neoantigens for that patient. But with an off-the-shelf product, of course, you can't do that. And the truth is there aren't very many shared classical driver mutations. They are very few, in fact, and so with our KRAS product, if you're just delivering KRAS, then basically, you're giving a single neoantigen to each patient. And we saw clear evidence of efficacy in so doing, but if you just compare it with our GRANITE data, the extension in observed overall survival in molecular responders was of a lower magnitude than we saw in the GRANITE program. which is consistent with that notion I just articulated. So you want to put other antigens in. Well, it's not super obvious what those are, and that's where the interesting research comes in. So there are all kinds of classes of additional antigens. They're not what people think of as classical mutated neoantigens where a single amino acid changes between the wild type and the mutant protein.

These are things like fusions, atypical or nonchemical splicing that creates novel junctional sequences, inclusion of things that are normally excluded as introns suddenly get retained in tumor cells. These are things like reexpressed endogenous retroviruses. So there are lots of interesting things that tumors, because of their dysregulated genomes, display on the surface, which can act as targets for the immune system. And it's harvesting from that big library of novel candidate antigens. That's the work we're doing right now, and that will go into the next slate product. So we're absolutely using this time profitably, and we're excited to get going, but today is not the day.

Our next question comes from Jonathan Miller.

Maybe just a quick one before I get to me. When should we expect to hear something about the Gilead opt-in in HIV? That's something I had noticed in the press release. And then maybe more specifically about the GRANITE program. Can you talk a little bit more about the powering of that Phase III endpoint off of ctDNA from the Phase II portion? I mean are you going to have enough PFS or OS data by 1Q 24 from the Phase II portion to make powering assumptions on those traditional endpoints in Phase III? I'm just trying to get a sense for how that ctDNA data will get from Phase II is going to inform powering for a traditional endpoint in the Phase III.

John, I'll take the first part of your question with respect to Gilead. We haven't provided any specific guidance at this time as to when that option payment could occur. I will point to the fact that the partnership has done well for quite some time. They were very motivated. And we went from signing the deal to an IND cleared in about 11 months, and the program is ongoing and in Phase I, and we presented that data along with Gilead, the preclinical nonhuman primate data at core recently. So that's essentially the best I can guide you at this point, it is not very much.

Okay. Thanks, Celia. Let me take the second part, John. So I want to be very clear about this. The ctDNA endpoint, the molecular response endpoint, is used for proof of concept in Phase II. And that is the gateway for us to then say, this looks very effective. Let's move now into the Phase III setting. On the assumption that we are using a traditional time-to-event endpoint, such as PFS or OS, the powering obviously becomes the key issue as you well know. Now our PFS data will be reasonable by mid-next year because if you think about it, assuming we're putting on our last patient sometime around the middle of this year, the median PFS for the standard of care for the control arm is 11 months. And so even the very last patient we put on the study has a good chance of having achieved a progression event by mid-'24. So the PFS data, I think, will actually be pretty good as we begin that study.

The OS data, however, will be immature. And so if we're doing an OS primary endpoint for Phase III, we will begin with a placeholder powering. And then we will let the data from the Phase II component mature. And as those data mature, we will then do a sample size reassessment that's prespecified and obviously, legitimate statistically because we're using Phase II data to inform the Phase III sample size. We're not in any way interrogating Phase III data to modify the Phase III sample size. That obviously is a much harder thing to do. We're using Phase II data to modify Phase III. So that is a very legitimate and reasonable approach. And obviously, we'll discuss all of this with the regulatory authorities in the first part 2.

That makes sense. Maybe one more about that interim there. Given that you're going to have a little bit more data, you're getting more data already from Phase I and you'll have more data from the Phase II by next year, given the upside. Do you have any progress in developing a prospective biomarker to help you select responders from nonresponders?

Simple answer is no. We've looked at the baseline molecular genetics and phenotypes of all the patients we've treated in the Phase I/II study and the responders and the nonresponders look the same. So there's nothing obvious that we've yet identified. I will note that some of those patients progressed and died within a few weeks of starting the study. So they're included, obviously, in the data that we've put out there. But the thing about moving to the frontline setting, the newly diagnosed setting, is that those patients, obviously, will not exist. And so a 50% molecular response rate in late-line end-stage patients, I expect to go up as we move into a frontline setting. And once you're over 50%, I don't think anyone worries too much about trying to weed out whatever it is, the 30% that don't get immediate benefit from the study.

Obviously, we're always interested in that question because we're always looking to do better. But that would not be a strategic priority for us. If you ask me my thoughts around hypothesis, there are some data suggesting that T reg infiltration after vaccination, after treatment may be driving some of these disparate outcomes and that there are some people who, for reasons that we don't fully understand today do accumulate T regs in their tumors over time that are obviously antithetical to a good outcome. So that's probably the leading hypothesis today. Over time, I think we'll have some therapeutics that might be useful and interesting to add in. But as of today, that's just a scientific hypothesis and not something we're spending a lot of time or energy on.

Our next question comes from Kaveri Pohlman with BTIG.

This is [Christian] on for Kaveri. And our first question is in regard to GRANITE. Sorry, can you just talk about your rationale for selecting the IP FS as the primary endpoint for the Phase III part and how this endpoint has been used historically?

Yes. We have not selected IPFS as the primary efficacy endpoint for Phase III. That is a possibility, but we have not selected any endpoint yet for the Phase III part of the study. On the list would be in reverse order, OS, very traditional, widely accepted, no debates there, and actually, obviously, historically very good with vaccines. And I'll just note that the interesting thing about vaccines is that the benefits, those T cells that you elicit do not go away just because the patient has some form of progressive disease. And so it is possible, and we've certainly seen evidence of this in our earlier-stage studies that OS actually is a really good endpoint for a study with a vaccine such as ours.

Anyway, OS is the most traditional option. PFS would be next. If we see a lot of pseudoprogression, meaning lesions get bigger because of T cell infiltration, then typically, that's a good thing for the patient, but they will be labeled by RECIST criteria as a progressive disease, and they are mislabeled. That's obviously a problem. So if we see evidence of abundant pseudoprogression in the Phase II study, then obviously, we will not be using PFS as the endpoint for Phase III. And that's where i.e., PFS or immunologic PFS defined using the immune response PFS criteria potentially becomes interesting and valuable. And so we're evaluating both PFS and the IPFS in our ongoing Phase II study. So you'll see those data next year.

As I mentioned, there is a small chance that ctDNA response may have accrued enough evidence to be used as an accelerated approval surrogate endpoint. And obviously, because we're running a randomized study that will continue to deliver OS data with no crossover. This actually fits nicely with the FDA's recent guidance around accelerated approval trial designs where they're trying to move away from single-arm trials. So I think the study design is good. The question on the table is whether ctDNA has become sufficiently accepted as a useful surrogate reasonably likely to predict overall survival. That's the key question. Today, the FDA has clearly not made the determination that the answer is yes. That is obviously subject to change as data continue to accumulate.

Okay. That's very insightful. And just one more for KRAS. There's a lot of next-generation inhibitors for different mutations being tested with single agent and in combination. Where do you see the use of immunotherapies in this space?

Yes, it's a great question. So the earliest data that were published, and I'll go back 3 years ago, I think it was now when Amgen's sotorasib was published in nature with preclinical data, what they showed was very interesting. In murine mice, they have no functioning immune systems using standard xenograft, the efficacy of the drug was clear, but it was finite and the tumors typically actually grew through drug therapy. When they repeated similar experiments in immunologically competent mice using a syngeneic tumor system, there was a much better outcome for those mice that the tumors seem to respond much better, more durably to the small molecule KRAS G12C inhibitor. And then they actually did a study where they added in a PD-1 antibody and showed clear evidence of synergy, again, in a murine system.

Now then we took these drugs, obviously into humans and 2 have been approved. But what we've seen certainly from sotorasib is that clinical benefit is often relatively transient and acquired resistance is complex. This is not as simple as something like EGFR mutations where there is a dominant resistance mutation, T790M, and therefore, I develop a drug that binds T790M, the acquired resistance to sotorasib, much more complex, very heterogeneous and much harder to sort of simply drug around. So I think the combination of these drugs with immunotherapy is likely to be really important to drive the durability of clinical benefit. So it's been somewhat challenging for some of those companies to successfully and safely combine their small molecule drugs with PD-1 antibodies.

And I'm sure you're all very well aware of the data I'm referring to. Some of the newer drugs appear to be free of challenge when it comes to combination with PD-1 antibodies. And so I think there's a really interesting opportunity to combine those small molecule drugs with both PD-1 antibody and the vaccine because the preclinical data suggests that there is likely to be meaningful synergy. And there's a lot of evidence that KRAS mutation is relatively immunosuppressive. And by vaccinating against the Kras epitopes, we may essentially bypass that local immunosuppression and really recruit the immune response into the mix to help drive durable clinical benefit. So it's a long answer to your short question, but I think we're getting to a point where those studies will begin.

Our next question comes from Mayank Mamtani from B. Riley Securities.

So obviously, glad to see the sample size expanded here driven by organic investigator interest and also a strategic choice by Gritstone. I was just curious as you've seen a lot of the patients now in the study more than we thought. Any learnings or any protocol measures you might be taking to sort of manage the SUTA progressors in the study? And then I was also curious does this mean that your follow-up period may be longer than what you may have thought initially? And in terms of the percentage subjects that do get that 6-month at Boost, you have an estimate of what that would be now relative to what you may have initially projected for the interim analysis? And then I have a quick follow-up.

Yes. Thanks for the question, Max. So I'm not going to speak in any data-informed manner because I'm not aware of data. So let me be very clear. But what we observed in the Phase I/II was clear evidence of pseudo-progression. We saw that early on, and we also saw it in association with that second dose of the adenovirus as you astutely noted, whether IPFS adequately addresses that issue, I don't know. I think it's the best we have short of just saying, "You know what, PFS is just unreliable here, let's just go with OS." So I don't know the answer to your question, but obviously, we are generating those data.

In terms of managing the trial, right from the get-go, we knew that this would be all about physician education. And again, this is where it's very helpful to have our Phase I/II data, particularly now it's been published in Nature Medicine to show to investigators to say, look, this is what we're seeing.

So just be mindful of this, be aware of it. And if a patient has radiologic progression but they're otherwise doing very well, then we encourage you to consider leaving them on the study and performing a follow-up scan per the rules of iRECIST. And so per iRECIST, a patient is not labeled as PD if they're able to stay on study and have a follow-up scan. And then it's the outcome of that second scan that determines whether the patient then is labeled as confirmed PD if the lesions continue to enlarge or if actually the lesions stabilize or get smaller again, they revert to being a stable disease, and you then just continue to follow, of course.

So education is very important. Our team has done a beautiful job of obviously driving momentum and enthusiasm for the trial, and we're in very close contact with the physicians. But nothing I've said is informed by data from the study. So let me be clear about that.

Got it. And then on obviously, the comments you made about the Moderna data and also the BioNTech Nature publication yesterday I was just curious, based on everything that you're seeing, do you have any preliminary thoughts on where you might go next in terms of additional tumor types or disease settings? And do we expect all of your additional plans beyond the study to sort of be on pause until we hear from you on the GRANITE data set, including, obviously, you said the SLATE has also been pushed back? So is that kind of how we should think about the story?

No, you should not assume that SLATE won't do anything until GRANITE delivers data. That's not accurate. We are very pleased with the overall platform, the neoantigen-based approach. And so we are very happy to launch SLATE once we have operating capital to begin. We will not gate that on data from GRANITE. So those are dissociated. Today's decision is purely just about redeploying some capital this year without sort of driving our burn this year way north.

In answer to your second question about where do we go next. The question we get asked a lot is around adjuvant opportunities and high-risk adjuvant in particular, as Moderna did. And when you think about this for a little bit, a patient with a high-risk localized disease, who is not cured by surgery. That patient obviously has metastatic disease. It just can't be seen on a CT scan, often referred to as a micrometastatic disease. And so this idea that treating patients in the adjuvant setting is completely different from treating patients in the metastatic setting, I think, is specious because unless you believe that the fact that a lesion cannot be picked up by a CT scan means it behaves in one way. And as soon as it's detectable on the CT scan, the biology behaves in a completely different way, then that position just doesn't make any sense.

And I've never really thought that was likely to be operative. And again, you've now got data from Moderna, which is the first randomized data set showing that their program seems to deliver activity in a hot tumor. But even within that data set, as you are well aware, if you look at the PD-L1 negative population, the treatment effect with the Moderna PCV was actually stronger than the treatment effect in the overall population. So this notion that hot tumors only will respond to this or that adjuvant only will respond to a PCB approach, I think, is just wrong-headed. We obviously have shown very nice data, single-arm trial Phase I/II in advanced disease. Obviously, we're now doing the randomized study. BioNTech, obviously, again, cold tumor in the adjuvant setting. So we're sort of tempering the dark board as it were with these different studies. But the biology just looks really consistent, and it's all about driving large numbers of neoantigen-specific T cells.

You can do that, you seem to see the benefit. I think that's true in our study, and that seems to be true in the BioNTech study. Now with Moderna, we don't know because they haven't published any T cell data yet, but there's this big issue. Why the curve is not separate to 9 months? Well, maybe they don't see T cells until 9 months. Maybe it takes multiple vaccinations to actually drive that really strong T-cell response. I think that's a pretty good hypothesis and the data will obviously give us the answer.

That seems to be unifying. And so for us, the decision about where to go next is not really handed in by notions of hot, cold, adjuvant, or metastatic, I think we'll just look for the opportunities where we think there's meaningful unmet need where we can really add something to those patient populations where there's a large commercial opportunity, which, of course, also means trials will enroll quickly. Those are the things that we're excited about, contingent on the GRANITE data. So I'm not really answering your question, but I'm saying we're not going to be hemin or pending by these, I think, rather artificial and ill-thought-through boundaries that some people try to impose on us.

That's helpful. And just maybe a quick one on COVID vaccine. Now that the clinical profile is relatively clear in terms of durability, immunogenicity, safety, tolerability, all the things that you discussed. I was just curious how the feedback has been, if you stack this profile against the government funding framework that exists out there for next-generation vaccines. I don't know if you can comment on how you're sort of seeing your profile relative to what maybe the U.S. government RFI that exists out there is looking for.

Yes, thank you. We're obviously paying attention to that government initiative because durability protection was explicitly called out for something of interest. And we obviously have now clinical data demonstrating that. So clearly, we'd be remiss not to be paying attention.

Our next question comes from Sean Lee from H.C. Wainwright.

Most of my questions have already been answered. I was just wondering, would we see any additional updates from the Link program later this year based on the existing study?

No. We gave a good update at the last ESMO, and there's really nothing important or particularly novel to show from that program at this stage. So, no.

Okay. Great. My second question is also on the infectious diseases program. I was wondering, beyond COVID, what other diseases do you see the learning from Coral that could apply well going into the future?

Yes, there's a lot of interest in respiratory viruses and the notion of multiplexing pathogens into a single vaccine. And obviously, a population of subjects, people over 60 are frequently hospitalized with a distressingly high mortality associated with respiratory viral infections. And the ones that are often talked about are, of course, SARS-CoV-2, influenza, RSV, parainfluenza, and sometimes the human metapneumovirus. So those are all very interesting pathogens. Several of them have some similar structural features to their surface proteins that obviously, we've learned quite a lot about through the prism of SARS-CoV-2. And so that's an area of high interest for us currently in the research phase.

Our next question comes from Corinne Jenkins from Goldman Sachs.

This is Craig on for Corinne. So just a couple for us. First, how should we think about the median time of therapy for the first 50 patients featured within the 1Q '24 readout of GRANITE 1?

Yes. So obviously, we told you how many subjects have been randomized as of yesterday. And as I said earlier, the median progression-free survival is 11 months. So I think that will give you what you need to know about likely data in Q1 '24. So 50 subjects. Obviously, that antedates May and probably most of April, and you can do the math.

Got it. So just a follow-up here. Given that you're expanding the trial to feature more patients, how should we think about this having potential impacts to the initiation of the Phase III portion of study?

Yes. Of course, nothing changes in terms of the kinetics with which we accrue data. We'll get the same amount of data at the same time as we did before. We'll just get a bit more data that follows. And so as I said earlier, and I won't repeat the answer, what this enables us to do is to power the Phase III study more accurately, which obviously is always a good thing. So what we're going to get out of Phase II is a more precise estimate and a more accurate estimate of the treatment effect size. And that is what is used to power the Phase III study. But as I said earlier, you don't have to start this -- you don't have to wait for this data to mature to start the study.

You can start the study with a placeholder sample size. And then you can revise that based upon observed data from the maturing Phase II, and you can actually preprogram the whole thing. So literally, it's all baked in before we even start the study. If we ultimately observe a hazard ratio of this, then the sample size will be that, et cetera. So the whole thing can be prebaked you get going. So there's no delay in the program, you just end up optimizing sample size, which obviously is always a good thing. So the move we've taken, again, it's available to us because of the great momentum we've built without compromise really on time lines and it enables us to design and execute a better Phase III study. So this is the logic behind the change that we made today.

At this time, we have no further questions.

Great. Thank you, Erica. Thank you, everybody. Appreciate your time and attention today. If you have any further questions, obviously, you know where we are. Otherwise, we will update you upon completion of enrollment in the study. Thanks very much, and have a great evening.

This concludes today's conference call. Thank you for attending.