Transgene SA (TNG) Earnings Call Transcript & Summary

Good afternoon, and welcome to the Transgene KL event. [Operator Instructions] As a reminder, this call is being recorded and, a replay will be made available on the Transgene website following the conclusion of the event. I'd now like to turn the call over to your host, Hedi Ben Brahim. Please go ahead, Hedi.

Thanks, Sara. Thanks, everybody, for taking the time to discuss about our program, TG4050. I'm very happy to welcome today Professor Christian Ottensmeier from the University of Liverpool and La Jolla Institute for Immunology. We also have the chance of having our colleagues from Maud, our CMO; Eric, our CSO; and [ Kaïdre ], our Program Director. Today, we want to focus on our individualized therapeutic vaccine TG4050. We'll also take a time to discuss to review our other programs. We will start with a brief introduction. Then we listen to Professor Ottensmeier, who is really specialized in these kind of approaches. Then we'll have a global overview of what Transgene is doing right now and our projects for the future. And then we'll move to Q&A. So first, I think it's clear for patients, for the industry and for investors that immunotherapy is really the fourth pillar of cancer treatment on top of surgery, chemotherapy and radiotherapy. Nevertheless, the number of responders can still be improved, and the duration of the response still needs to be significantly improved. We see, for example, ICI bringing a lot, but being far from achieving 100% of response rates. We think that treatment resistance could be overcome by better understanding of the tumor macro environment and by offering the right combination of immunotherapy. And typically, virus-based cancer vaccine are perfect for that. Because of the great profit -- the great safety profile, because of their mechanism of action, they are perfect to combine with other immunotherapy and with other kind of treatment. So they can become potential game changers in oncology. [ Change ] is at the heart of this revolution. We have just presented 8 posters at AACR, and we are very proud of that. And it's highlighted the quality of our portfolio from R&D program to product already in clinic. We will talk today about TG4050, our [ unutilized ] cancer immunotherapy. It's a perfect example of a global approach of using the right vector with the right antigen in the right indication. Here, we are using the MVA, Modified Vaccinia Ankara, which is a virus that we have proven with a great safety profile, has the ability to use the [ broad ] and specifically new response, growth from innate and the adaptive immunity, and that strongly differentiates against [indiscernible]. What antigens are we putting in our MVA to make it TG4050? Here, we are putting individualized antigens that are selected by the artificial intelligence designed by our partner NEC, and this is very committed in the program, that they have [ developed ] this artificial intelligence, and they're financing half of the 2 clinical trials that we are in right now. When we have designed this product, we then selected an indication that is the head and neck patients, where today, patients have no solution to reduce the risk of relapse. So we want to address that unmet medical need, and we are talking about [ 70,000 ] patients per year. Until now TG4050, part of the myvac platform, has proved its ability to induce an immune response, leading to the clinical benefits. And based on these exciting results, we are already preparing for a potential regional trial that will start before the end of the year. Now I would like to hand over to Professor Christian Ottensmeier.

Hedi, thank you so much. Yes, I think this is a really exciting offer and opportunity for preventing recurrence in high-risk head and neck cancer, where currently, after standard surgery and chemo radiotherapy, really, there is very little we have to offer these patients. And just contextually, we need to recognize that about half of these patients will relapse and at relapse, the majority of the patients will die. So this is a really important clinical puzzle. But I think it will be really exciting to share with you some data why I also think it's a really important immunological puzzle that this program is addressing. So conceptually, then [ cancer ] immunotherapy is an intriguing challenge because if the cancer develops, then the immune system must have failed. So following from that, the puzzle then becomes, can we reestablish immune control, what would that look like? And while at the patient level, that's simple, cancer disappears. At the cellular level, it's much harder. We don't really know what the optimal immune response is, now what the targets should be that T cells, if they're then driving this, we should see. And then, of course, this knowledge is critical to understand how do we get there. Cancer immunotherapy has been around for a long time. And really the best evidence that the immune system can kill malignancy comes from allogeneic transplantation, where in hematological malignancies that were meaningful to curate the first patients who had relapsed disease and had no other treatment options, could be offered a chance at disease-free survival. But the toxicity was significant. And in solid tumors, this whole field changed in 2011 since when cancer immunotherapy has really become a standard of care with stunning benefit in responders. But as we all know, only about 25% of patients will benefit across multiple tumor types with an anti-PD-1 or [ anti-PD-L1 ] antibody. And while there are some examples how combination with chemotherapy in lung cancer, anti-CTLA-4, anti-LAG-3 in melanoma might alter this a little bit, nonetheless, even those responders, by and large, of many of those patients will relapse. So then the question is, how might you overcome this lack of T cells that are recognizing and then attacking and ultimately eliminating cancer tumor cells? And I think the platforms that are being exploited with most detail are really aligned at the bottom of viral vaccines, which is, of course, what we're talking about here; RNA vaccines, which have set the world in light of coronavirus, but actually, that's a really low bar for delivery of an immunogen against the virus that we really haven't experienced before; and peptide vaccines, which have been developed by a number of groups. And I think from our understanding of what viral vaccines or vaccine backbones that are viruses encoding interesting cargo can deliver is the delivery of an innate immune response, identified here on the left-hand side, which is, of course, very rapid because it's preexisting as well as the developing an anti-antigen CD8 and CD4 T cell response to an adaptive response, which will ultimately be the key feature for protection. And from the data that are in the literature, we believe that the viral vaccines have the likelihood of being the best-in-class in that context. And that would be really quite unsurprising because, of course, our immune system is geared up to eliminate viruses. So viruses contain all the signals that make the immune system really exciting. So it's, from my perspective, as a clinical immunologist, is not really a surprise that this would be a highly effective platform. The more challenging piece, I think, is what to put into the vaccine. And this is a figure that I made from [ CAD ], really identifying that not all antigens that might be recognized by T cells, specifically antigens that are present in tumors, are born alike. And that is because we have this built-in mechanism, which tries and direct our immune system away from seeing self-antigens. So this is a process of central tolerance in which our T cells go to school and they are educated against a curriculum of what not to recognize in the ]. So the more interesting antigens that might not be subject to self tolerance are shared antigens such as on [ some cancer testis ] antigens. But we already know from many bitterly disappointing trials that these have been less effective. With perhaps some really intriguing targets such as [ MAKO ] 1, which biologically, immunologically, really stand out, as well as oncoviral antigens, and I'm sure the Transgene team could spend a whole hour talking about those, because these are, of course, antigens that are new to the patient's immune system, but are mandatory for the tumor cells to retain them. But the field, overall, is now converging on the group of antigens that as a result of the genetic manipulations that the tumors undergo as a part of process of the cancer becoming malignant. And while we know that from the genetic abnormality, if you were to count them up, say, you've got 1,000 mutations that are naturally occurring in the cancer cells, only a small percentage of these will be ultimately visible to the immune system. There are many immunological features that makes them really interesting. So there's no natural tolerance, they're most different from this -- from the bar -- from the cells, and therefore, they're most likely to be immunologically visible and therefore, conceptually offer a low hurdle. So the key in this trial really is to try and translate these immunological concepts into patient outcomes. And the background for choosing the vaccine design, as ultimately delivered, was to try and set this up in a way where you would have the least non-immunological barriers to the vaccine per se and -- that the patient might bring to the table. So we argued that in a minimal residual disease setting, there will be minimal immunosuppression, have already identified that head and neck cancer, there is a high unmet clinical need. And the program, therefore, went from safety evaluation to feasibility to now identifying whether there is clinical benefit and ultimately, whether this maps onto immunological features. So the head and neck cancer patients, in terms of their clinical outcome and their link to the number of T cells, are relatively well understood. So here is a bit set of our own, in which I've split the tumors into blue, these are HPV-driven tumors; and black, HPV-negative tumors. And we know that the viral antigens in HPV-positive disease confirm a significant benefit. So most patients who have HPV-positive disease do not actually relapse. Conversely, and in HPV-negative disease, the majority of patients do relapse in an immune cell-related manner. So if you have lots of T cells, then the patients do better than if you have a few T cells. And please bear this in mind because I think it's really intriguing to take this concept into the data as we're going to present them. We know that anti-PD-1 antibodies have minimal benefits, or 10%, 15% of patients have durable benefit. So that's no doubt related to the fact that checkpoint inhibitors activate what's already there, and the majority of patients don't have many T cells. And therefore, clinically, this is the -- HPV-negative tumors are the bad end of the spectrum of these low cell cancers, with a median survival of about -- progression-free survival of about a year. So the design of the trial was to try and unpick both immunogenicity as well as safety and, of course, visibility because making a new virus per patient was a tall order. But also, we aim to set this up in a way that it might be able to allow us to get some efficacy data in a relatively small trial. So we argued that early randomization at the completion of standard-of-care treatment would allow us to make a vaccine for everybody, so entry criteria to the randomized arm is successful vaccine production, but would then allow us to have the vaccine off the shelves and give the vaccine to half of the patient and see what happens, and at relapse, are in standard-of-care treatment. And increasingly, that's becoming checkpoint inhibitor treatment, all follow the patients only and then add relapse, rebiopsy, resequence, but deliver the vaccine that already is in the fridge or the freezer, more likely, in combination with standard-of-care treatment. And we calculated that from the numbers of patients at risk in this condition from the setting that we had chosen, very high-risk disease, extracapsular spread, large tumors, nodal involvement, we would need about 30 patients to detect the clinical [ site ]. And what has emerged since we started dreaming up this trial? And it's really that transcriptomic analysis has allowed us to categorize patients, in this case with [ headaches ]. And these are data not of my own, but these are taken from [indiscernible] the slides together. But what we really identified is that there are 4 sort of big pots, in this case, in head and neck cancer. So one is that the tumors that have lots of immune cells. So there are these ones here. So these would be, in my previous graph, that TIL-high tumors. There are no other obvious immune escape mechanism that are detectable from this kind of transcriptomic analysis. Then there are other tumors that have no T cells, and then there are the tumors that don't have very many T cells or no T cells at all, that are characterized by fibrotic stromas. And then there is a sort of mixed bag that has fibrotic stroma and T cells, an all-in-one. And you would predict that the failure of checkpoint inhibitor treatments would be dominated in these 2 groups and so the tumors that don't have immune cells or that are kind of fibrotic -- fibrosis driven, and we know it had neck cancer -- cancers associated with [ fibrosis ], so the former hedge around the cancer cells, protecting them against the immune attack. And I think that particular feature I've just mentioned, this immune exclusion is mainly driven in this particular group on the left. So we would predict then that the group of patients who benefit from standard immunotherapy, would sit at this spot, but these are the groups that we really need to do something about. So from that background then, it is most intriguing that just by happenstance, we seem to have done just that. So these are data that the Transgene team have assembled looking at transcriptomic data, so essentially identifying this kind of transcriptomic information and allocating the patients to the expression of [ PD-L1 ] on the one hand, but also the -- one of these 4 categories. And you can see that the majority, more than half of the patients in both groups, are really in this immune-deserted category. So any immunological event that might be driven, must be a direct effect of the toxin. And if it turns out that there is a clinical benefit, that would really make a very strong case that we found a way to overcome this particularly immunologically challenging hurdle that so far, no checkpoint inhibitors can actually get. So the data look really promising. We are still in early days. And clearly, this needs to mature. So I'm always a little bit hesitant to review data of an ongoing randomized trial. But so far, we know that there have been no relapses in the treatment group, while there have been relapses in the patients who have been observed only. And so, so far, so good. It looks as though we are ticking all the right boxes. Additionally, we know that the vaccine has been super immunogenic. So of the 30 new epitopes that were incurred in the vaccine, we see a median of 9 responders. So in other words, over 1/3 of the targets are actually able to activate T cells. So this is, I think, really critical because it identifies that we've picked the right clinical setting, patients whose immune systems are able to handle the vaccine onto [ multi-immune ] responses. And the immunological data, of course, compare what was there before to what happens after vaccination. And it's been most intriguing to see that there is a significant group of new T cell responses, which have no preexisting counterpart by immunological [ elapses ] in the blood and that we also amplify responses that are present. So I think this is a super exciting data set, where everything seems to the hang together in the particular immunological group of the highest clinical unmet need. So I think we know that T cells are the critical condition, the [ zenith ] condition, without which nothing can happen. No T cells, you can't do immunotherapy with clinical success. And we've already identified the checkpoint inhibitors awaken in what's there before. And that made, I think, a case for why this vaccine enables in T cell infiltration at a time point when the tumors per se don't actually have any T cells of their own. So overcoming a really big gap in our current . And of course, we know that vaccines are the only treatment that we currently have to train T cells because, of course, that's how we've used the vaccines over many, many decades, and we know that is safe and highly immunogenic. And I think I would like to reiterate that the viral delivery mix biologically, so much sense because that's exactly what our immune system is good at. So I think we've ticked the boxes that we can also induce T cell responses of the right flavor. We know that is now -- that this kind of approach is clinically feasible. So it's really quite exciting to see this program developing. And of course, we haven't talked about the parallel study in ovarian cancer, which really backs up the data in a way that is entirely coherent with what we are seeing in the TG4050 study. So clearly, the next step will need to demonstrate the efficacy in a larger study. But I think this -- the data comes at a time when there is enormous enthusiasm that is filling everybody that counter vaccine can play a role, except that I think in our data here demonstrate that vaccines can do this without the help of anti-PD-1 in contrast to what we've learned from the Moderna study, in which, of course, if you were to top up the numbers, only about 10% of all the patients vaccinated actually benefit from the vaccine. And so I think we're now really set to go for a larger multicenter trial that reduces all the problems that are associated with a [ 4 ] center randomized, admittedly, an international trial. But I think to persuade the community that this is really -- this really has legs, that will make a big difference. So if I summarize this, the vaccine seems to be safe, it's immunogenic. It doesn't require anti-PD-1 to see effect. The early data are consistent with, but don't yet prove clinical benefit. The fact that this is visible in a small data size, [ visible ] in small sample size argues that we are observing a large effect. Therefore, I'm very optimistic that the next study will confirm this. And I think it will be really exciting to see whether at recurrence in those unfortunate patients, vaccination is able to rescue the recurrence and of course, coherent with the ovary trial, they'll be very careful with [immunological ] data, something that I think I've been astounded Moderna hasn't been able to offer us so far. So at the moment, we are, I think, sitting really rather sweet, and it looks very beautiful. So I'm super excited to be part of this program. And with that, I think I said all of these things already. I think the bit that I would really like to finish on is I genuinely believe that this kind of approach offers us hope to cancers that are not only devastating to the patient, but also to their families and often terribly -- rapidly causes of death of the patients who are suffering from them. Thank you very much. I'll hand back to Hedi.

Thanks, Ottensmeier, for sharing your analysis about the program and giving also your view on the whole field. And we share the same enthusiasm about the fact that this field is moving really in the right direction. There is a strong momentum around cancer vaccine, and you mentioned Moderna, of course, and we think that we compare favorably with what we bring in terms of safety profile, immunological response rate and so, of course, the first line of [ immunotherapies ]. So, so far, TG4050, and we've shown that in the poster that we had this week at the AACR, has shown a great safety profile. We have sign of clinical activity in high-risk patients, and we show that in monotherapy without the support of ICI by delivering this broad and specific response. That's why we are already preparing the future of the program, which will be a Phase II, potentially targeting registration that will start in the second half of 2023. We are addressing a total need of more than [ 70,000 ] patients per year that could represent a market of more than $1 billion. And this is talking only about head and neck in adjuvant setting. . But it's obvious that this program, because it's patients, the specific, can address a huge number of indications, especially in adjuvant or maintenance setting, but maybe also beyond that. And it can be a number of cancers listed here, the [indiscernible] or many others. . TG4050 is very exciting, and we have a great posture on AACR, that was not only one. We had a lot of news on other programs. We talked about [indiscernible], our other therapeutic vaccine with fixed antigen, supporting patients suffering for -- from HPV-positive [ anogenital ] cancer. We are very happy to announce that we will have a new transactional data presented at ASCO, so just a few months from now. And we target to end development in the first half of [ 2024 ]. If we look at the oncolytics part with statistics in the new poster, we have further demonstrated the great safety, the selective [ replication ] and the expression of the payload in the tumor. We have the evidence of T cell activation and immune checkpoint receptor expression. Also, we've been able to share more about our latest antivirus TG6050 that has shown in preclinical model outstanding antitumor activity. We expect the first patient to occur in the coming weeks. We talked at [ ASCO ] not only on our existing program mechanics, but also our projects for the future. And we are happy to share news covering hot topics in our industry such as novel payloads, new venous backbones, tumor aids and also extracellular vesicles. Last, the overview of our portfolio, and you know that Transgene has multiple shots on goal. We have 2 therapy vaccines. We've talked a lot about them, TG4050 and TG4001, and that's also now 3 oncolytics in clinic and also the very important partnership with AstraZeneca around oncolytics. If we look at 2022, we have achieved great results on all programs, which gives us a lot of confidence for the milestone that will happen in the coming 18 months. TG4050, we have shown repeatedly last year, and each time we've confirmed, again at AACR, that we have promising data as a single agent. We expect to start the treatment of the last patient in the head and neck trial very soon and to start the Phase II before the end of the year. TG4001, we had a positive interim analysis, end of 2022. Now we are looking at the end of randomization in 2024, and we intend to rapidly start registrational-directed trial. On the oncolytics part, we have proven the feasibility of the intravenous [ administration ], which is the key difference we want to bring in the oncolytics. That has been translated with a rival of 6050 that will be injected intravenously in [ lung ] cancer with the first patient [ via medicine ] is all right in clinic. We will finish the Part Ia as a single agent very soon. Then it will move into the combination with [ pembrolizumab ] in partnership with . So I think we've covered all what Transgene is doing right now with 5 products in clinic, based on our deep expertise of bioengineering and our approach of using the right vector with the right additions in the right indications. This technology is bringing now great news to clinic, and it's supported by various partners especially AstraZeneca in the oncolytics. Merck and Pfizer was supporting TG4001, for example. NEC, which is really central for TG4050, and BioInvent for BT-001. Today, we have -- I remind you that we have the financial visibility until early 2024. Thanks again for listening to this discussion. Now we're very happy to turn to questions and answers.

So our first question comes from Brandon Folkes from Cantor Fitzgerald.

Congratulations on the data. Maybe just firstly on the data, any color on the patient background and characteristics of the patients who relapsed in the control arm? And then just any commentary in terms of are those patient characteristics similar to some of the patients or [ poly ] patients with the same characteristics in the TG4050 arm currently, who are responding well.

[Technical Difficulty] So to answer your questions, what we -- to remind you, we only review patients with [indiscernible] T3 or T4 disease. And to answer very precisely to your question on the 2 patients who relapsed in the observation arm, actually one was Stage 3 with a new special [indiscernible] on these features. And the other one was Stage 3 -- 4, sorry, is extra-capsular lymph node invasion. So it's normal behavior of patients. There were no specially [ mode ] and -- especially the one with Stage 3, we not especially -- we were not especially expecting that patients will relapse. Are you happy with my answer? Or do you want more details?

No, that's fantastic. And maybe if I just follow up with a second trial granted, you sort of -- the Phase II is not designed yet. But just maybe any thoughts in terms of how large and how many patients are needed in Phase II, especially from a safety perspective. . And then secondly, you presented very promising single-agent data, but do you expect to run a combination arm in the Phase II?

Regarding the center side, we have some precise ID and especially in light of what has been done by Moderna. You all know the controversy around the statistic and the fact that Moderna trial conducted the one-sided statistical design. And therefore, we are working on the two sides. We believe that it's a thing which is evaluable, if we want really to have the trial which persuade the community, the scientific community that we are going in a good direction. So obviously, it will be a trial of more than 100 patients. And the important feature I want to underline is that thanks to the design of Phase I, our Phase I trial, which is already a randomized trial, we can benefit from the -- now we have 32 patients who have been randomized Phase 1, we can benefit off of those patients and analysis with the patients who are going to be on Phase II. That's for your first question. Now your second question, as you know and as Christian mentioned, so far the PD-1 or PD-L1 has been very disappointed in the adjuvant setting. There is 2 large Phase III trials, one with pembrolizumab and the other one is avelumab [ trial ] negative. And so there is an ongoing trial testing perioperative pembrolizumab, which data are expected by mid '25. So for the patent, we are not considering a combination because of the lack of efficacy of the ICI. So why to increase the toxicity burden of the patient if there is no benefit? Having said that, depending on what can be released from the ongoing trial of perioperative pembrolizumab, we are open to the possibility of combining with pembrolizumab, by the way.

Congratulations, again on data.

Thanks for the questions, Brandon. So our next question comes from Olga Smolentseva from Bryan Garnier.

Maybe firstly, just I was wondering acknowledging that you've seen preliminary responses or relapse 3, I guess, [ ranges ] in PD-L1-negative patients, but it seems that there is a slight imbalance between PD-L1 status in TG4050 and control arm. Do you expect that to somewhat impact the final data?

Yes. Sorry, you expect answer to the question?

Yes. So yes, there is indeed, somehow, an imbalance, as you put it into -- across the 2 arms of the study. The reason is that the [ PD-L1 ] staining is an exploratory parameter and was not intended and the study was not designed to be calibrated with respect to the 2 arms because the 2 arms are not actually comparative arms in this study. Nevertheless, all data is not available. As you could have seen on the poster, there still analysis ongoing, and I don't know to which extent this is going to be [ flattened ] with the additional data. And last, even if you see it as the imbalance and you want absolutely to keep the idea of having a comparative view across the 2 arms, the difference between the 2 arms is not really going in a sense that is favorable to the vaccine. So when you look at the overall data, you can see that even in the vaccine arms that pretty much an enrichment for very desert to more microenvironment and many patients with also low or very medium expression of [ gene ]. So we're not -- and last, we're also giving the vaccine in monotherapy. So it's not an [ ECA ] combination. So the expression of PD-L1 and its impact on the activities is not something that we can straightforward interpret. So yes, yes, there is there is difference, but we don't -- today, it's hard to make an assessment on how impactful is that and how this is meaningful on the vaccine.

Thank you, Kaidre. [indiscernible] wanted to comment.

Yes, please. I think it's really important to note, first of all, the data are randomized. So there is no choices made on the [ PD-L1 ] upfront. And of course, with an end of 30, you just have to live with what the randomization ultimately reveals. That's point 1. So point 2 is the data on on [ complete ]. So about 1/3 of the patients in the randomized observation alone have not yet been evaluated in this way. So I think we need the full data set before we can come to a conclusion that there is an imbalance. And I think we have no -- there's no suggestion from any of the data that vaccines would work better or worse in [ PD-L1 ] low, moderate or high disease. And I think the -- therefore, at the moment, we can only note that in the incomplete data set, where 30% of the data are still missing, there appears to be an early imbalance, but that may very well balance itself out once the whole data set is available.

That's very helpful. And maybe just the last one. I was curious as of the net antigen-positive T cells, what percent of CD8-positive CD4-positive cells you've seen, if you have such data available?

So most of the -- it's -- yes, thank you for the question. It's hard to answer because actually, 70% of the CD8 T cell response are also associated with the CD4 response for the same target. And as we are essentially targeting CD8 epitopes, we put 25 CD8 epitopes and up to 5 Class II epitopes in the vaccine. But still, most of the Class I epitope, because of the design of the vaccine, are also associated with the Class I epitope. So it's hard to give you like a percentage of Class I and Class II. They are essentially Class I, but most of those Class I are also associated with the Class II. And I would say that we have a calibrated response between CD8 and CD4 count, which is a classical feature of the MVA vector.

Thanks for the questions, Olga. This concludes the verbal portion of the Q&A. I'll now turn it over to Lucie to read the remainder of the questions from the web.

Thank you,. So we have a first question here. Do you plan it with the FDA ahead of initiating a registrational trial with CD4050? And if so, will you be communicating to investors the feedback from this program? I hand over to Maud.

Yes, Of course, so very important question. And we do tend to ] the FDA to discuss all the details, not only on the clinical aspects aspect, but to discuss that with the FDA. So this is something which is quite important. I am not sure that it will be of great interest for the investors to share the conclusion of the FDA. But for sure, this will translate into some adjustment of some slight modification of the[ quarter ] we are at the moment, designing. So this is something which will come in the coming months.

Thank you very much. [Technical Difficulty] the first patients to refine the targets for the following patients so that a self-improving process. So this is a question I will hand over to Kaidre to answer.

Sorry, I couldn't hear the beginning of the question.

I think the question is to know whether we use the date -- what we learned from the ongoing trial patients to feed the database and the artificial intelligence we're using to select new antigens.

Yes. So the principle of the machine learning system is that it will improve over time by being looked with the data we generate from the patients. So up to -- no need to change slide, I think. Up till now, we have been using the [ Basel ] system. It has been updated once throughout the Phase I. And it will be updated then regularly. And we had the recent exchanges with FDA, where although they do not have a process for those adaptive machine learning system, they are kind of putting in place good practice on the go and had exchanged with them recently on how we should be updating them and informing them of when and how we are adding data to the system. So we're moving from a fixed system to what's called a semi-adaptive system with time points, where we'll update the system at a given time point and inform the agency of those evolutions. And we're pushing forward and moving forward, fully adaptive systems, where we have a continuous improvement and the continuous feedback group on the machine learning and the improvement of prediction, starting from the data we're generating from past patients. So it's by design, it's on the go, and we're working with the regulator to make sure we're implementing these new approaches into clinical setting.

Thank you very much. We have a final question, which I will direct to Hedi. The initiation of our registration-driven trial of TG4050 will be contingent on the company raising additional capital to fund this program?

So you know that today, we have financial visibility until early 2024. We will start the trial before -- at the end of the year. Of course, the duration of the trial will be much longer than existing current financial visibility. And I think that considering the momentum that is in our field of therapeutic vaccine and especially individualized therapeutic vaccines and the data which brought the new milestone we have in this program and the other programs, we will have support from our existing shareholders and also new shareholders in the future.

[Operator Instructions] I think I don't have any questions arriving, so we can turn to closing remarks.

Thank you, Lucie, and thank you, Professor Ottensmeier. Thank you, Eric, Maud and Kaidre for the organization and being part of this session. I think you've been able to hear our enthusiasm about this program and the analysis [ properties ] about the field and TG4050. We will continue to bring new data on this program on the existing trials. But without waiting, we'll move into a new trial before the end of the year because we think it can be a key solution for patients in this adjuvant setting in [indiscernible]. Thank you very much for your time, and have a great end of the day.