 next panel, which is another one of the panels that we've been doing every year at the USAIC Health Care Summit, and that's our oncology panel. And this year we have an amazing moderator in Jay Bradner. Jay Bradner is a prototype physician-scientist, entrepreneur. Jay is just stepping down after a very successful multi-year tenure as the head of the Hollowed Institute at Novartis, Niber, where he led the R&D organization. Jay has started five biotech companies. He continues to see patients. He teaches in students and residents, and he also teaches all of us in many ways. And Jay, as many of us in the industry, has demonstrated immense resiliency, but none more so than this week when his Bruins, who are the best hockey team in the history of the NHL, lost in Game 7 in the first round of the playoffs. So, Jay, I'm glad to see that you've bounced back after that loss. And with that, I will hand it over to you. Well, thanks, Andy. We can commiserate with your Islanders at the appropriate time. Bruin, Andy, and colleagues, thanks so much for another totally spirited and incisive summit, our 17th, just blown away all day today by the engagement, participation, and the attendance of such a meeting. Welcome to the Oncology panel. We're joined here today by real experts advancing the bleeding edge of science and industry in the academy, in pharma and in biotech. We have with us today Elise Reisen, the chief executive officer of tectonic therapeutics, Frank Nestle, the global head of research, and the chief scientific officer of Sanofi, Kathy Seidel, the vice president and head of the oncology drug discovery unit at Takeda, but Katie Rizvani, professor of medicine, chief of the section of cellular therapy, director of translational research at America's largest cancer center, the M.D. Anderson Cancer Center, Maria Verzi, the co-founder and CEO of Proceed Biosciences, a new co-focusing on enabling precision medicine through a next generation liquid biopsy platform technology that I know there's just can't wait to hear about, hear more about. Welcome everybody and thank you for being with us here today. Let's open this panel up with a softball over the middle of the plate. As you reflect on all that you've seen in your own shop at the ACR, out on the savannah of biotechnology, to your mind, what is the most promising investigational therapy or new mechanism or new technology? And we'll go around and ask everybody, Katie, let's start with you. Okay, well, of course, I'm biased, so I am going to say for me of cell therapy certainly seems to have the one of the greatest potentials that I see for treatment of cancer that potential has been fulfilled, at least somewhat in hematologic malignancies, but I do believe that that potential will also be fulfilled in the setting of solid tumors as we understand more the biology and use the tools that are available to us to manufacture the next generation of cell therapies. All right, well, we'll double click on that in a moment. Frank, what's your perspective? Yeah, thanks for having me, Jay, greetings from Paris. I would like to pick smart biologics. It addresses the question, what if you turn a biologic like an antibody from a mono-specific binder with a passive distribution pharmacology into an intentionally designed molecule with attributes like multispecificity, designed PK, targeted distribution. One of those molecules we have currently in early clinical trials is a precision-activated T-cell engager. It's a technology platform where you engage T-cells, but you do it with a mask attached with a protease cleavable linker, and this mask comes off only in the tumor and that avoids CRS, which is obviously a major blockade for appropriate dosing in cancer patients, and hopefully gets us to a better safety and efficacy profile. Thank you. Kathy? So thanks for having me, Jay. It's really great to be on this esteemed panel. I have to say a similar answer to Frank in terms of biologics. We have some very interesting T-cell engager biologics that we're pretty excited about that are conditional CD3-activated molecules, and these, again, we're hoping for that better therapeutic index, which in preclinical studies held out to be the case, and so these kinds of drugs with a better safety profile should allow us to have less CRS and be safer for patients and be able to go after targets that have been a bit more challenging because of their broader expression. All right. A growing consensus. Raham, what do you think? I'm going to deviate and be also a little bit self-interested here, which is probably a good thing, I guess, if we should all be working on things that we're sort of innately passionate about, gives it the best chance of being successful. I'm really interested, honestly, in a lot of the technology that's emerging that looks like it's going to enable us to detect disease much earlier, better define, I think, the disease biology minimally invasively, and perhaps track, I think, the evolution of disease biology more precisely, so that we can take wonderful medicines and rather than testing them empirically, actually get them to the right patients at the time when you can sort of achieve potentially the best outcome. And I think the last decade has been so much progress. There's obviously a lot more to come, I hope, with this really interesting technology out there that's sort of developing at this stage. Great. More on that for sure in a moment. Elise, what have you seen out there? I wish I was going to be a contrarian, but I think it sounds like we're going to have a lot on the immune system, and in my second choice would have been where Rihann went. So I also, I think we're in the early stages of cellular therapy, and we're just starting to see the potential. Yesterday there was a press release about TCR approach to solid tumors. There was over 60% response rate unclear if you had durability, but even getting that was super exciting. And then I think you start to think about access, because that's an issue. And where are we with the ALO gene, ALO CAR-T, certainly getting high response rates for the persistence isn't there, but I'm getting starting to get excited about second generation that hopefully will overcome the persistence piece. And then I know it's an oncology panel, but I have to mention CAR-T and the immunology space because there's super exciting preliminary data there as well. Well, thanks. Just so that there's fair representation as a chemist, I would add that there's been just an absolute revolution in our capacity to engage, modulate, degrade, selectively inhibit, allele specifically inhibit targets that for 40 years have been beyond the reach of therapeutic discovery, whether it's, you know, covalent chemoproteomics or molecular glues. And secondarily, I really start to admire the way a very small group of drug developers in our field are going earlier with precision prevention, because that unfortunately given the heterogeneity of this disease and most patients is where we're going to need to be. All right. Let's move on. So let's open the discussion with immuno-oncology. I apologize to anybody if this, these opinions seem provocative, but I'm just back in the clinic again after seven years being away. And I guess I had expected more from the contribution of the amazing medicines we read about every morning and fierce biotech and endpoints. Why is the next generation of immuno-oncology medicines taking so damn long? Why are these medicines so underperforming? It was Coley in 1891 until CTLA-4 in 2011, PD-1 in 2014. And then it went quiet and has for a while. And it's not for lack of investment. And it's not for lack of trying. Massive biotech and NIH investment with actually very few emerging agents. And I can only think of one or two that have a shred of single agent activity. So, Kathy, let's please start with you. How have on-treatment biomarkers that we've all systematically obtained on these studies failed to yield actionable insights? And what can we do is correct? Yeah, absolutely. Thanks for the question. Of course, biomarkers have led to many improved response rates and more durable progression-free survival. And even in some cases, better overall survival. But yes, there are many trials with maybe one of the most famous recent ones in the anti-tidget phase three disappointing result from Roche Genentech. And I think where we're left is that we don't really know in the IO space how to select our patients, right? And even in PD-1, anti-PD-1 treatment where we can enrich for patients that respond. We still have some patients that in the absence of the ligand PD-1 still respond to anti-PD-1. And we don't really understand why, despite, as you say, many, many efforts to try to understand different kinds of approaches, sequencing everything under the sun and looking at different proteomics in every possible measure. So it's the lack of translatability from mouse to human. A lot of this work is done in mice. And of course, there's many examples of that not working out. And then also, I think when we are selecting our patients, we don't have enough mechanistic understanding of the therapeutics that we're using to be able to then pick the patients that are going to respond. And another maybe provocative thought that I've been considering with our teams here, Takeda, is, is there any way that we could use exploratory biomarkers to select patients that are set within the strong hypothesis of the particular therapeutic that we're going after? And that's in the broad IO kind of approach. Of course, if you have something with a target, a cell surface target or a TKI or some clear handle that you can select patients based on, we should be doing that. And even there, I think we could be doing a much better job at taking a look at the overall pathways that we're going after to make sure that we are encompassing the broad mechanism of a particular therapeutic. You know, it's well said, all these years later, we're still using PDL 1% probability for agents that don't really even modulate that pathway. So we have so much more room to move in biomarker space. Round to put you on the spot, what sort of biomarkers might we create that would better stratify patients for next generation iotherapeutics? Yeah, I think, you know, Kathy mentioned a really interesting trial, you know, the PDL 1-Taget trial in small cell lung cancer. And when they presented that trial, you know, the complete failure and disappointment that it was, there was a sub bullet that said, please note that transcriptional subtyping was not done in these patients. And what's really interesting is, you know, transcriptional subtyping of small cell lung cancer, you know, has been done by numerous groups, tissue-based RNA-seq, and they've all converged on a thesis that there is a group that might be driving the response to immunotherapy in patients with small cell lung cancer. And AstraZeneca sort of retrospectively validated that thesis in an analysis they did of their own positive trial. So I would argue, and it gets back to I think what Kathy said is, are we using all of the exploratory exploratory tools at our disposal when we're actually doing these trials? Because if we're not, the risk is we run up really, really quickly, we fail, and then we have no good explanation for why we failed. And the consequences, we haven't really moved the field forward other than to say, don't do the exact experiment that we did. So I think to me, there's a really big opportunity to use, you know, things like that. And, you know, Jay, I was very inspired by one of the companies that you founded, SIROS, where, you know, one of the strategies was to look at transcriptional dependencies. And, you know, and that seemed to work. But if you think about it, it's not really an approach that's used kind of broadly. So I guess in my view, I feel sometimes the competitive pressure, the urgency to act, the focus on costs, the difficulty of getting tumor biopsies has really basically posed a pretty big challenge in terms of really doing a thorough scan on exploratory biomarkers that might inform success or failure and enable us to move forward with much more clear learnings from either case. Well, it's quite an insight into the state of our understanding of immuno-oncology, just as a fundamental biology that we might hear invoke such holistic measures, new microscope lenses, to look at this same age as old problem. Let's pivot in immuno-oncology to sort of new mechanisms and new approaches. And Frank, you had invited a consideration of modular biotherapeutics, which indeed are very exciting. They're often not trivial to manufacture. You've extensively studied NK cell engaging therapeutics, modulating therapeutics. How might these sorts of therapies complement or orthogonally accompany checkpoint response and resistance? And will they have single agent activity? That's a great question today. And, you know, we've just discussed how T cell checkpoint inhibitors might have exhausted their capability in patient population to achieve responses. And if you think about T cells as an effector cell in the spirit of Coley, you just should remember that there are other, and he was the father of cancer immunotherapy, as we probably all agree on, that there are other important innate immune effectors. And it's really interesting that T cells, why they are super effective and they're endorsed with memory capabilities. Some of the few cells in the immune system, which can remember what happens, they have an Achilles heel. The Achilles heel is that they have to recognize a cognitive antigen and they have to recognize in the context of MHC. And that makes them susceptible to antigen loss and MHC loss. And to all the failures we're seeing, especially then also in the post checkpoint space. Now, in case cells have been put into place, by evolution to circumvent that Achilles heel, they recognize cells that can and including cancer cells in an MHC independent manner. And that also gives them a very practical attribute if you use them as a cell therapy, that you can use them as a universal donor of the shelf cell therapy product. So at Sanofi, we decided a few years back to have multiple shots on goal in this space. And again, coming back to the smart biologic concept, I'm most excited about what NK cell engages can do because they have excellent drug-like properties. They could replicate what T cell engages have done in blazing the trail and they might have better therapeutic indices. But you have to be humble because your NK cell compartment might not be efficient. And that's why you have to handle NK cell therapies. And we do this with a universal donor approach. And then you have to be also humble about what it takes to activate an NK cell. And you might need the appropriate cytokines, which we have also in our toolbox. So really getting at NK cell therapy from multiple angles, NK cell therapies, engagers, and cytokines is probably what it takes to make it successful. Now, your question, single agent activity? Yes, we showed in at Ash two years back that in two thirds of patients with relapse refractory AML, you get CRs or CRIs with a donor derived NK cell product. But the big question will be, and there's a race to now get this to approval in liquid cancer, can we move into solid cancers? And this is a whole different ball game. You have to get NK cells into solid cancers. You have to survive the hostile tumor microenvironment. And that's a big, big front share where we're all exploring and what we need to do to really make NK cells the effective cell it serves to be. Cathy might have some additional comments, obviously. Well, let's pivot right to that. Thanks for the segue to cellular therapies. You know, one thing that Glenn Dranoff used to always say to me in our one on ones was this current wave of so-called next generation IO medicines are really a bunch of targets also discovered in the 80s and 90s. And it will take the discovery of new targets from on treatment biopsy. I believe to prompt a really effective next generation of IO medicines with just exactly the kind of measurements Rahan that you mentioned. All right, let's pivot to next generation cell therapies. And Katie, you're up next. So, yes, CAR T cell therapy is now firmly established as disruptive as to many patients definitive just celebrated another of Emily Whitehead's birthdays. It's also intensely challenging. And the field has sort of lacked creativity of late 200 thinly differentiated CD 19 like therapies. And it's not proven plug and play in solid tumors, but for neuroblastoma, which I just couldn't believe it was so exciting to read. So out of the box and taking all the information that we have available to us to design the next generation of cell therapy. So for instance, looking at what are the biomarkers of which type of patients are likely to respond using the tools that are available like CRISPR gene editing in addition to the various viral vector manufacturing and engineering strategies that are available. And also using all the single cell omics data that are now coming through that tell us a lot about mechanisms of resistance and mechanisms of relapse. Ultimately, the product that would probably make really self therapies the mainstay of treatment of cancer will have to be a product that is safe that can be delivered to patients as a point of care or at least available beyond highly specialized centers. A product that will also result in durable responses to patients. For that we will end up having to use a multi, I think a multimodal approach of a product that's engineered potentially combined with biologics potentially combined with checkpoint inhibitors and other immunomodulators. And I don't think we will ever have a product that will fit all because for each cancer, for each disease, those mechanisms of resistance, mechanisms of relapse are going to be very disease specific and even disease subtype specific. So we've put a lot of our effort into natural killer cells and Frank very eloquently described the advantages. I don't want to belabor that point, but I think at least to my mind having a product that is and I know we're going to talk about access, but ultimately you want to have a product that can be scalable, that will be affordable, that will be safe, both in terms of the patient side effects, but also in terms of giving confidence to an oncologist, maybe in a smaller clinic to administer it to the patient, again, reducing cost, increasing accessibility, but then resulting in inducing profound responses that will be long lived, so more so than just for instance a bridge to transplant something that could potentially result in a cure in a patient. So again, looking at how can we engineer our natural killer cells using CRISPR gene editing to make them resistant to the tumor microenvironment, putting a car in there, engineering them to secrete cytokines, combining them with bispecific antibodies, we presented our data at ASH with a bispecific Engager targeting CD30, where we observe responses of greater than 90% in multiply relapse patients with heart disease. So I think all of this is feasible, but a lot of science will have to go into the development, both in terms of engineering the cells, but also the biomarkers of who is going to respond. Just one last point I would like to stress, especially for an allogeneic cell therapy product, even though these cell therapy products are manufactured from a healthy donor, not all donors are the same, not all healthy people are the same, we all have different susceptibilities, for instance to viral infections, we have different immune systems, so also identifying the right donor, allogeneic donor that you will be using to manufacture 100 doses or 1000 doses of your cell therapy product is extremely important and I think that's something that more and more people are thinking about and also we're focusing on. Love that thought about donors, you could imagine we're using a patient's T cells with clonal hematopoiesis of indeterminative potential will be a disaster or a really good idea, that's a really key insight. Elise, I know you spent a lot of time thinking about this too, feel free to share reflections on next generation because I think they're linked, but I'd love to ask you about access, you've brought medicines forward in development that have reached patients all around the world and now as a CEO you must be working backwards from what does it look like to have a medicine that reaches patients, how do we democratize access, because we have to balance the cost of preparing this medicine with its impact, but we also for it to be a sustainable business have to ensure that there are margins to this that don't dilute the productivity of the whole rest of the oncology portfolio, this matters not just for patients accessing medicines, but also especially in underserved geographies, but also for drug development. Katie, you just mentioned combination of studies, we did a combination of Kim Raya with a Brutonib, it was so hard, it's hard enough to develop these as single agents, so we need to fix access for so many reasons, how will we do this, Elise? God, I wish I could have the magic bullet to fixing access, but I think in the cellular therapy space it starts getting away from auto to ALO or to NK, I think that's really the only way you're going to significantly reduce the cost of making these, now they're still not going to be as cheap to make as a small molecule, I think in vivo is the other place that could reduce the cost, but depending on what you're using, using lentiviruses for in vivo, there's still the cost of making the lentiviruses, so I think it gets you a significant amount of the way there, they're still going to be very expensive to make, the better we can make them, and the more they lead to cure, the more the value proposition is there, so if all you're getting is two or three months of progression-free survival from them, it gets a little bit harder potentially to justify the cost, but if you can really do the cure with them or get to really prolonged maintenance of disease, I think you start to get there, and maybe it gets back to, Katie started to talk about a lot of combinations, I come from an ID background initially many, many decades ago, so it always made sense to me that you're going to have to treat oncology with multiple modalities, and maybe even in the CAR-T space, in the cellular serotonin space, one of the mechanisms is you stop the tumor stop expressing the antigen, you've got to be thinking about making cellular therapy that goes across two antigens on the tumor type, and maybe that's the way to go. Just agree with everything you said and the framing. All right, let's pivot to targeted therapy if it's okay. One of the benefits of having this magical non-compete clinical part of my life is I'm reading a ton, and I just read a stunning paper on men in inhibition from Gaius Issa at MD Anderson, a fellow who trained in my lab as a postdoctoral student, and it just inspired me the way I first reading about a matinee. Now our field has turned over every oncoprotein stone, and we've almost drugged them all. We'll get there eventually with new chemistry and biotherapeutic approaches, but every time we encounter immediate or eventual resistance, chronic myelid leukemia is just the exception to the now rule defined by the experience of developing targeted therapies in cancer, and patients rightly expect better. We all do. So, Kathy, how are we going to improve the performance of targeted therapy amidst all the challenges of tumor heterogeneity, advanced disease, resistance just seems inevitable. Yeah. How have we got there? Yeah, no, absolutely. That is a big, big challenge for our field, and one of my favorite papers that came out in 2022 by Merida et al. It was a beautiful AML single cell analysis paper looking longitudinally in patients before, during, after different therapies. And basically the results are exactly as we would have predicted, right, that the tumors respond in response to the drug that's targeting certain cell types and then other cell types that are insensitive to the drug expand. And we know from many years of work by Irv Weissman and many others that those mutations are already there. They're just at very low levels, and now we've created this environment in the presence of the drug for them to be able to expand and take over. And so I'm quite hopeful with some of the new techniques, looking at the chromosomal instability, the chromostripsis, say that word 10 times fast, as well as other kinds of chromosomal deviations that we will get better at detecting and you have to be more sensitive at detecting those mutations that exist before the therapy and are likely to grow out after the therapy. Now the other thing that this paper showed, which is terrifying, is how unique every single patient, every single tumor is. And so one patient's response is not similar to another patient's response. And so it's very challenging to predict what that resistance is going to look like in everybody. But we have to start somewhere. And of course in AML, it's a lot easier to get this kinds of longitudinal, very informative data. But recently at ACR, people are starting to collect this kind of data in solid tumors. And so we can start to build a map of here are the most likely resistance mechanisms that will happen on a particular therapy. And that is then something we should be already working on in the preclinical space, right? So that we have these different options available for patients knowing that each one is just going to be another option to keep and prolonging that patient's life with a good quality of life. So I think it's a combination of a lot of things as we learn more about how these tumors respond to the therapies we're giving. Now, of course, IO is the promise of being able to cope with lots of heterogeneity. And so one of the things that we've put a big bet on in Takeda is focusing on leveraging the innate immune system we've already heard about in K cells. We're also looking at gamma delta cells, both from cell therapy as well as the cell engager point of view. And we're counting on that innate ability to recognize tumor versus normal that this is an abnormal cell that should be killed and gotten rid of to kind of deal with some of that natural heterogeneity that we don't really understand and don't have a way to go after. So that's where we're, you know, focusing a lot of our efforts right now. But I think a lot of that other kind of analysis over time for how patients respond to therapies is what we're going to have to as a field work on. Well, thank you. I heard think truncle, think early, think in combination and really know the patient. And so it's not to only focus on small molecules. I wonder if I could ask you a question about another modality, antibody drug conjugates. Here is a modality that I find very interesting because it has a truly massive denominator and a very small but very important numerator of successes. I mean, they're finally emerging as first world therapeutics. I can only imagine how many her to ADCs genetic made in the day. What's your forecast for these medicines and this platform and and how could we better develop them with data? Yeah, that's great, Jay. And just before I kick off a little bit, I just to comment on one thing Kathy said, I agree with it. And I think there's also this need to understand non genetic mechanisms of resistance and better understand because they tend to predominate, you know, particularly when we're going off to very specific targets. So with that said, I think the ADC is really interesting. I think three years ago, when I was on this panel, we were asked what our prediction was in terms of the most, you know, the technology that's going to make the biggest bang over the next few years. And luckily, I said ADCs. And I think I think I'm probably right that it's made a tremendous impact. I think in the last two to three years. So look, I think the general view is very positive. Obviously, at the moment, given, you know, of course, you know, significant M&A and deal activity, but of course, you know, also phenomenal efficacy results, you know, coming out, for example, last year's Askel and her to standing ovation, it's amazing. So, you know, it's sort of, and I'd be interested in what what other people view is here. You know, the question is, do we have the formula really today now to understand what makes a great ADC? And as I've spoken to folks, you know, it's pretty clear that it's, it's not that clear cut, right? You know, we've definitely made tremendous progress and understanding, you know, all of the various different elements of an ADC, you know, linkers, conjugates, you know, toxins, exactly what antibody right, you know, construct you actually want. But there isn't a winning formula. It's very nuanced, very context specific. And so what you're now seeing, which I think is a good thing, is intense activity in the field, a lot of activity against some pretty common targets, but where people are actually modulating very selective components of the constructs. And it's going to be really interesting, I think, to see how the landscape plays out. I think for sure it's going to grow significantly. But what I think we're probably going to end up with is some new targets that yield to some new ADC constructs. We're going to have some existing targets yield, I think, to better constructs. And net net, I think we're going to see a lot more ADCs in pretty much every major indication. And you know, as I've spoken to investigators, I think it's going to open up a little bit of a problem and a challenge. Because with these very targeted agents, the question is going to be how do you pick patients for each agent if they don't have an upfront means of selecting. So if you're a lung cancer trialist today, and you have her two ADCs, her three ADCs, you know, ELL three ADCs, all of these ADCs coming at you, all of them say we work broadly in a pretty big selection of patients. You know, how do you triage patients to the appropriate drug? And I think this is going to put a lot of pressure on our ability to come up with ways, both invasively and minimally invasively, you know, ways that are more accurate to require less tissue than today's current means to get an assessment of patients upfront, and then triage them to the antibody that they're most likely going to respond to. Thank you. All right, let's take a couple questions from our audience. I see a question here from Christina Trial Hansen. Christina, can you turn on your camera and bring your question forward? Absolutely. Thank you for a great panel here and a great job, Jay. So my question is a little bit in what we just heard from Kathy and Rehan. So I've been really excited about all the developments in understanding, you know, spatial transcriptomics, chromatin structure, epigenetics, and how we use all of these, you can say, informations on top of each other to optimize our cell products or understand biology more. But my question is more, when are we going to, you know, see the transition to also use that in patient selection? And how likely is that we can actually implement these technologies in what I will call precision immuno-oncology? I love that. At least I can't help but look at you. I mean, you guys took out the big shotgun to develop PD1. If you were to do it all over again with precision in mind and measurements, what does precision immuno-oncology look like? And Frank, I'd love to hear your thoughts as a I'm ready to go. Let this start. Okay. I agree with everyone else. I think the problem with the next generation immuno-oncology has been we haven't been able to select the right patients. And part of me wonders whether we've tried hard enough at it. You know, we put a huge amount of resources into the PD1 biomarker. I can't even tell you how much work went into that. It's not a perfect biomarker. There's no question about it. But it's also clear that there were indications where if we didn't have that biomarker, we probably wouldn't have gotten efficacy or known how to select those patients. And part of me wonders, I actually think and Jay and Frank, you guys know this better than I do now, that in large pharma, when we do those early studies, we're collecting all of this this biomarker data. The question is, if you're not seeing a lot of responses in phase one, do you really spend the time going through that data? And then do you have the wherewithal to go check the next hypothesis, knowing that chances are that hypothesis might not be right. It might be by chance. And so I think we've really got to think hard about the effort we put into the biomarkers and go in with some pre, you know, hypotheses up front, try to enrich for those patients. And maybe it'll help us in some way get to a better place. Well said. Frank, thoughts? Yeah, just creating molecular disease maps is sort of the other side of the coin of the modality explosion we are witnessing, because this is basically what it means to unlock undruggable biology in real time, ideally in patients. And, you know, we've started with single cell immunology disease maps and applied that to IO and INI and neuro diseases. And the next big frontier is to go into other cell types, including neoplastic cells, and to create these cell maps where a cell is the smallest physiological unit of what the pathology looks like. And, you know, we've called it, you know, atomic resolution, because that's what an atom of physiology looks like. It's a cell. And that's incredibly exciting when we built these setups. The big question is, can you get this into the clinical and do this on a routine basis before you get it to a patient hospital near you? And that requires, and I think, as you mentioned that, a change in culture and paradigm in the early development space, put these single cell maps as a must have and not a nice to have. And think about early clinical development as a learning experience as much as a faster, cheaper, better, cheaper experience. And that is a whole cultural change we need to do. Like 20% of the budget needs to go into these types of efforts. It's a bit like your Amazon, and you're the best logistics enterprise and you own e-commerce. But you're not interested what your customer thinks about it. You don't want that feedback. You're not really, you're not investigating. That's what I'm sometimes seeing when we get into the clinic and we just move these molecules forward, just simply thinking about efficacy safety and maybe one or two pk and pd biomarkers. And then there's this big freezer full of stuff, and there's no clear agenda and how to use this and learn and close the cycle. Only when we do this as a pharmaceutical industry, then we will have also the tools to ultimately do what we have to do is get it to patients with a prickle, finger, and in a hospital near you. If we don't do this in our very bespoke and very expensive clinical trial environments, how will this ever go to a patient in a hospital near you? So key to know where to invest in subsequent development as well. And we have one last question from the audience. Dr. Sundeep Gupta, Sundeep, can you get your camera on and kindly ask your question? It will be the last question of the session. Sure. Actually, the question I had in mind was more on acquired resistance, which I think has been covered by Kathy and Katie both as well as UJ. So I'll switch my question a little bit more and talk about, I mean, we have been very successful in discovering novel therapies, targeted therapies for sure. But I think the clinical development, early clinical development has always been a challenge because we need to work on initially in patients who have failed multiple lines of treatments. So what are the strategies and what are we thinking about? How to enhance our early drug development, clinical development experience so that it doesn't take that long and if we fail, we fail faster rather than have to deal with patients who have failed multiple lines of therapies. So that's the resistance becomes even more of an issue. I'd love to hear what you think on this because actually more early clinical drug development is done in academia through investigator initiated studies and we don't often listen to that community in industry, not enough at least. Do you have any thoughts about how to better direct first into human to phase 2a clinical development? Well, you know, actually our cell therapy programs are all in multiply relapse and very they're patients with highly progressive disease. We face the same challenges in that FDA also at least in the setting of cell therapy mandates that we treat patients that have failed two or three lines of standard of care treatment. But for instance, our recent current case study where we wanted to include patients with acute malo leukemia, we were specifically told to remove those from our cohort because patients with AML have other alternative treatments despite the fact that we already said only we will recruit patients that have failed two different lines of treatment and FDA said you can only then go and amend the protocol once you've had three complete remissions in other disease settings. So it's hard. I can see FDA's point. Of course, if it's something that's never been tested in patients, they want to and the safety is improved. They want to have that kind of safeguard of patients. But then the other hand, when you have a patient who's found 11 line different lines of treatment, including two transplants, which is actually the kind of patients that we are currently treating on your in case of therapies, it makes the whole thing more challenging. I think this needs dialogue with the FDA. Also, I can't imagine anyone on this call Sunday not having a strong opinion about how we can do early development better because it's an important platform, not just to pressure test the medicine, but to set up success in phase two and phase two success rates in cancer are woefully underperforming. All right. Well, I regret to say that we're out of time. I'm going to turn this back over to our fearless and handsome moderator, Andy Plump. Thank you all for being with us today. Thank you to the panelists for joining and for sharing your thoughts so freely. Thanks so much. Thanks to the whole panel. Really, I think a good balance of the excitement of what's happening in oncology with the reality. And if I just go and reverse to three or four years ago for the same panel, it was all just hype around what was emerging in cell therapy and immuno oncology. And I think we've seen over the years some reset, but still a lot of potential and great, great, really great panel. So let's open it up to our audience. And if we could please bring up the polling question. So what will cause the biggest change in clinical trials in oncology? A, dose optimization guidance? B, accelerated approval guidance? Or C, FDA guidelines regarding clinical trial diversity? So we'll look forward to the audience engaging in this in this great dialogue. Thank you. The next panel, let's pull up the results of our poll, please. Oh, everything is about even here. So what will cause the biggest change in clinical trials in oncology? And we have a photo finish. Dose optimization guidance, accelerated approval guidance and FDA guidelines regarding clinical trial diversity. So thank you for sharing your feedback with us.