Conference Transcript - FTH 2026-04-16

Gil Blum, Analyst — Needham & Company

Good afternoon, everyone, and thank you for joining us on the fourth day of the Needham Health Care Conference. My name is Gil Blum, and I'm a senior biotech analyst here at Needham & Company. My coverage also spans this field of oncology. It is my pleasure to have with me today Anand Perk from FITE. Is that the right way to say it? FITE Therapeutics? FITE Therapeutics, yeah. Anand, you have the floor.

Anand Parikh, COO

Thank you, Gil, and thank you, Needham, for the platform. Fythe is a company focused on cancer metabolism and crucial multi-node pathways in cancer. Before I start, I just want to mention that today's presentation will contain forward-looking statements subject to risks and uncertainties detailed in our 10K and other SEC filings. Fythe recently went public in a reverse merger transaction with Sensei Biotherapeutics. This was structured as a stock for stock transaction and concurrent with the deal, we raised $200 million in a private placement from institutional investors. The Fythe leadership team has been merged with the Sensei leadership team, and we're very excited that the capital that we've raised is primarily dedicated to advancing Fythe's lead program, PICTOR, which is expected to have phase two top line data in endometrial cancer and initiate a phase one B in HR positive HER2 negative breast cancer, both expected by year and 20. A little bit of an overview on Fythe. Fythe is a company focused on multinode pathways, complex pathways in cancer, where multinode inhibition is going to lead to not only hopefully better efficacy, but also less toxicity. We're beginning our work with the PI3 kinase pathway, the most frequently mutated pathway in solid tumors, and we are engaging in vertical blockade with two oral small molecules. That vertical blockade inhibits PI3 kinase alpha, TORC1, and TORC2. Compared to other pathway inhibitors, we believe that not only are we oral compared to intravenous, that we have similar rates of all-grade hyperglycemia, we have lower rates of stomatitis with no prophylaxis, and that we have greater exposure above common efficacy thresholds like IC90, staying above those thresholds for longer than the competition, which should translate to greater efficacy. In terms of recent and anticipated readouts, we had a phase two recently readout at ESMO, which met its primary endpoint. That was with one half of the PICTOR intervention, sapinocertib, with paclitaxel. That's a randomized phase two. We also have a phase two with the entire PICTOR intervention in endometrial cancer. Now we'll readout in the second half of 2026. As I mentioned, the money we raised will also fund, in addition to the phase two and endometrial cancer, a phase 1B trial in HR positive HER2 negative advanced breast cancer. That trial will initiate in the first half of this year. We have a pipeline beyond Pictor, BitPictor is the focus of the company. And one thing I would note about the origins of Fythe, Fythe has really been, from the beginning, focused on multi-node inhibition of crucial cancer pathways. Our team and our scientific co-founders really reflect that, with Lou Cantley in particular being notable as the discoverer of the PI3 kinase pathway. We're glad to have Lou and our other esteemed scientific co-founders on our side as we attack this very important problem for patients worldwide. So as I mentioned, the PI3 kinase AKT mTOR pathway, also colloquially known as the PAM pathway, is the most frequently mutated genomic driver across all solid tumors. The way we are attacking this is with two oral small molecules that target what we believe to be the most crucial nodes in this pathway. Cerebulisib, which inhibits PI3 kinase alpha, and sapinocertib, which inhibits mTORC1 and mTORC2. Together, we call this PICTOR. One of the major benefits of targeting this pathway at multiple nodes is you're actually able to shut the entire pathway down. We believe that single node inhibitors which target components of this pathway are partial pathway inhibitors. When you actually inhibit this highly correlated and interdependent pathway, you are able to decrease the dose of any individual agent, which sets a firm lower bound to your therapeutic window and allows for much greater tolerability. Another benefit that arises from actually setting that firm lower bound on the therapeutic window is that evolved resistance is far less likely. Evolutionarily derived mutations, resistance mutations that occur as a result of pressure being placed on an individual point of this pathway, escape mutations can be avoided by multinode inhibition. The other benefit of multinode inhibition is it's a much larger addressable market. Of course, if you're focused on single point mutations in the pathway, then you're limited to that market. We, on the other hand, have shown the ability to target patients with mutations throughout the pathway and actually outside of the pathway as well, as our phase 1b shows. Finally, one of the benefits of FITE's method of multi-node inhibition is that we are orally administered. And while oral administration is, of course, a convenience advantage for patients and physicians alike, it also provides important PK advantages, which I'll detail throughout the course of this presentation. So I want to walk through some science here that I think is crucial to understanding the benefits of multinode inhibition. Whilst targeted therapies focus on shutting down the genomic drivers of cancer, the actual translational readouts at the bottom of these pathways are the key to shutting down the activity of these pathways in cancer cells. And so for the PI3 kinase AKT-MTOR pathway, the key translational readouts are phosphorylated S6 and phosphorylated 4-EBP1. As you can see from these Westin blots, single-node inhibitors do a good job of shutting down phospho-S6, but not such a good job shutting down phospho-4-EBP1. Our agents in combination, however, are able to shut down both S6 and 4-EBP1. And as has been demonstrated in the literature, it's 4-EBP1 that correlates far more closely with drug potency. This is true not only of the approved therapeutics in the class, but also the investigational therapeutics, including mutant-specific inhibitors, as detailed on the western blots to the right. We shut down phospho-EKT, phospho-S6, and phospho-4-EBP1 in a far more complete way than any single-node inhibitor. As it relates to the multinode inhibitors, we think we have some crucial and important advantages. The historical development of multinode and pan-PI3 kinase inhibitors has shown that alpha selectivity is important. Pan-PI3 kinase inhibitors have traditionally been dogged with less tolerability and also potential immunological toxicities due to the fact that PI3 kinase gamma and PI3 kinase delta are largely expressed in immune cells. In solid tumors, it is PI3 kinase alpha that predominates, and any limited signaling that may occur through beta or other PI3 kinase isoforms is largely inhibited by the fact that we are also shutting down downstream mTORC1 and mTORC2. So the initial clinical data for our combination was really demonstrated in this phase 1b study. This was the first time that PICTOR was put together with another mechanism of action, paclitaxel, in this case, although we also believe and are undertaking trials with other cytotoxic agents, including CDK4-6 and potentially hormone in therapy. In this trial, however, Pactlitaxel was the partner of choice. Within these advanced solid tumors, you can see that the doses tested, the sapinacertib and cerebolicid, were significantly lower than the monotherapy RP2D of these agents. These agents, when given in monotherapy, the RP2Ds were determined to be 900 milligrams and 9 milligrams. Our recommended phase 2 dose was 3 milligrams of sapinacertib, so three-fold lower than the RP2D, and 200 milligrams of cerabilisib, so over four-fold lower of its RP2D. That's an important fact because when we see the activity that we've seen and which I will further demonstrate, this explains why the tolerability is also so beneficial for patients and physicians. So in that phase 1b, we saw a 47% overall response rate with three complete responses, four partial responses, and four patients with stable disease. When we look at those patients who had stable disease, on the swimmer plot, we can see that that stable disease was sufficient to qualify also for clinical benefit. These patients had endometrial, ovarian, and breast malignancies, and on average had four prior lines of therapy. It's notable that all of these patients had prior, except for one, had prior taxane. So this was taxane re-challenge. In that setting, taxane re-challenge in advanced solid tumors, we would expect a 10% to 20% overall response rate with about four months PFS. So this data appears to be outpacing that significantly. I also want to discuss the mutational landscape. When we look at the mutational landscape here, you can see classical PIK3CA mutations in green, PI3 kinase pathway mutations without PI3 kinase, PIK3CA mutations in red. So mTOR, AKT, PIK3R1, those kinds of mutations, those are in red. And then patients without any pathway mutation in block. It's notable to see that we had activity across all three different mutational subtypes. And also, if we look to the Victoria 1 study, which was undertaken by Cellcuity, the data that got a positive market response was actually in wild-type patients. That means patients who had an absence of a PIK3CA mutation. In this study, that would be all patients denoted in the red or the black. All those patients would be considered PIK3CA wild-type under the Victoria 1 stratification. And importantly, none of these patients would be eligible for a mutant-specific inhibitor because they have co-occurring mutations in P10. Given that fact, we thought the safety profile was very favorable with grade 3 AEs in 58% of patients, which compares favorably with other comparable therapies and discontinuation in about 5%. Common AEs were GI and low grade. We did see some neutropenia neutropenia, and anemia, but we think that was potentially due to patlitaxel, given that that's a signal that hasn't really historically been seen with these agents. When we move to the adverse events of special interest, we see a Monk's best-in-class hyperglycemia profile. This data is from our ongoing phase two in endometrial cancer, which is given all at that 200 milligram, 3 milligram dose that was determined in the phase 1b. Here we see a 14% overall hyperglycemia rate, which compares very favorably to inavilisib, the relay compounds, or the scorpion compound, and quite favorably to gadatilisib as well, especially considering that we have amongst the most lenient HbA1c and fasting glucose criteria in the field today. When we look at stomatitis, we have low-grade stomatitis and low overall rates of stomatitis with no prophylaxis. This is unlike gadatalisib, which has prophylaxis and at much higher rates of stomatitis. We have received questions around, well, why do you see that given that it's an on-target mTOR effect? This is something that has been seen with everliners. And the reason is oral dosing. Oral dosing is not just a convenience advantage, but it confers some very particular key PK benefits. In particular, our dosing regimen of three days per week, every week, with no weeks off, means that we achieve greater exposure above critical efficacy thresholds, like IC90, compared to the competition, which are administered intravenously. We see about 150 to 190 hours per month, and that's above IC90, and that's as a result of dosing in above that efficacy threshold three separate times per week with no weeks off. If we compare that with the competition, we see 60 to 70 hours per month above that efficacious threshold. Additionally, as it relates to CMAX, because we are oral, we can continually dose into that efficacious range rather than intravenous, where the entire bolus of drug is given at the beginning of the week, and then it falls below that efficacious range. Additionally, with intravenous drugs, you must have a very high CMAX. Because we are able to continuously dose, we don't need such a high CMAX to IC90 differential. In fact, we never reach more than about 2x our IC90. This is crucial because it relates to stomatitis. We have indeed seen in earlier studies of sapin assertive, our TORC1-2 inhibitor, that when given in large single weekly boluses akin to an IV dosing schedule, we see increased stomatitis.

Gil Blum, Analyst — Needham & Company

So it is indeed a class effect, but by breaking the dosing up into this three-day-per-week regimen, we see far lower rates of stomatitis.

Anand Parikh, COO

Moving on to clinical development. So endometrial cancer is where we've begun our development, and given the data from the endometrial patients in the phase 1b study that I outlined, we think that was a logical place to start. Three of those patients had complete responses. Four of those patients had responses with a median PFS of 11 months. Several of these patients had previously received IO as well. Now, given that paclitaxel or doxorubicin is the prevailing second-line therapy in endometrial cancer, with PFS of about four months and a response rate of between 10 and 20%, we think that if we're able to carry on, this level of efficacy would have a significant place in the firmament of physicians. And I will note that endometrial cancer is a disease and one of the few cancers that is growing in both incidence and prevalence with very few approved therapies. In terms of our phase two trial design, we have the single dose established in the phase 1b combined with pachytaxel, again, as established in the phase 1b. It will be 40 patients, a single-arm study, and we hope to read that out, at least response rate and some preliminary safety data, by the end of this year. It will be an endometrial endometrioid cancer. All patients will be post-carboplatin, post-IO. Moving on to breast cancer. We, of course, monitor this acuity data with some attention, and we're very pleased to see that succeed for both patients and the field, as we think it provides a really good validation of the multinode hypothesis. We believe that our advantages due to oral administration and PI3K alpha specificity, as well as increased tolerability, offer some unique advantages over the cell curative regimen. And indeed, our agents have been developed in breast cancer previously. Sapinacinotib plus fulvestrant was randomized against fulvestrant alone in HR positive HER2 negative breast cancer patients and showed an intriguing PFS and DCR rate when compared to the Victoria 1 study. This was in all comers, similar to the Victoria 1 study. Cerebilis of itself, the other half of the pictor regimen, has shown encouraging efficacy as monotherapy in breast cancer as well. We, of course, believe that putting the two agents together would be superior than either one alone. And as we look at the landscape of breast cancer, we believe that not only has CDK4-6 or CDK generally and hormone therapy been established as the backbone of breast cancer, But with the Victoria 1 data, that increasingly PI3 kinase agents, multi-node agents specifically, are creating a third pillar of breast cancer. And as we see that come to bear, not only in second line, but looking forward, also potentially in first line, that it will be important for patients to have an all oral regimen that has really good tolerability. And that's what we hope to bring to first and second line settings. We will also be examining future standards of care as the treatment options in breast cancer continue to evolve beyond CDK4-6 and aromatase inhibitors plus or fulvestrant. We will be doing work with some of the novel mechanisms of action in this space to determine tolerability and safety in the first instance, and then efficacy as well. As we move on to other indications, ovarian cancer, we had a positive readout in a randomized phase two in platinum-resistant ovarian cancer. This was in patients who had greater than one line, on average about four lines, and this randomized phase two, which had 124 patients across about 25 sites, the control arm was paclitaxel, the experimental arm was paclitaxel plus sapinassertib, which is our TORQ12 inhibitor. It did read out positively in terms of PFS with a low hazard ratio. Grade 3, 4 AEs across both arms were similar, and the intervention was well tolerated. In terms of next steps, we await OS, ORR, and detailed safety data, and we look forward to an FDA interaction on this program in the future. As you can see from the Kaplan-Meier curves here, the PFS separates and does not come back together for the two regiments. We think this is encouraging and look forward in future development of this to add cerebulicid to this intervention as well. In terms of our regulatory positioning, we have inherited a large safety database from Takeda, which we're very pleased with, and we've also seen a lot of prior dosing work with these agents. That sets us in good stead for future development, and we will need to round out some contribution of components. That contribution of components has been done with sapinacertib and paclitaxel and sapinacertib and fulvestrant, but we will likely need to round it out with cerebilisib and fulvestrant and cerebilisib and paclitaxel. It is our belief that those studies will be qualitative. They will not require statistical significance, and we will provide further updates after greater engagement with the FDA. In terms of CMC, we are well positioned with ample drug supply and drug product for ongoing and proposed phase 1b and phase 2 studies, including extensions thereof. We have a healthy pattern life stretching into the late 2030s and potentially into the 2040s, with also the ability for a fixed-dose combination for our two oral small molecule agents. We're very excited about the catalyst calendar we have coming up. We have the Sancuity mutant data in the first half of 26. We have our own endometrial data in the back half of 26. We will have the breast cancer data safety initially in the first half of 27, and then interim efficacy data in the back half of 27. That will be 20 to 40 patients on CDK46 plus and or for Vestrant as well. And so we're very excited by that. we think that's the data set that is really going to allow us to benchmark our data to the other agents in the class as well. And so with that, I'll thank you very much for your time.

Gil Blum, Analyst — Needham & Company

And Gil, looking forward to the conversation. Thank you, Anand. The floor is open for questions from the audience, but maybe I'll start with a couple. For full disclosure, I do cover Selquity. So I'm pretty familiar with the space. Maybe one point to start. This is regarding oral dosing and PK. Now, I know this is apples to oranges, so bear that in mind. If you remember, cell acuity had differences in activity between two different dosing regimens, one which involved very continuous dosing and another one that was three-on-one-off. With an oral dosing, I'm assuming that's more close to the continuous version again apples to oranges but i'd wonder like what your thoughts are

Anand Parikh, COO

as it relates to that piece of information yeah the with the three on one off it still is very similar to an on off regimen because we have those three days when we're in range and then four days where we aren't so it is similar to a pulsatile dosing that's been seen with um getter or even with CAPF assertive, which is four on three off as well. So we do see it as similar. We're just spending longer above IC90 and other critical thresholds. So we do believe that continuous dosing in this pathway can be a challenge, but we have a very similar pulsatile approach.

Gil Blum, Analyst — Needham & Company

right my other question is you see quite a lot of investment in quote-unquote next-gen selective pick three a's relay comes to mind um what prevents them from you know taking

Anand Parikh, COO

everolimus and adding it to their regimen it's a great question and one that i get a lot i think is fundamentally quite misunderstood everolimus is a wrap log and derived from a natural molecule It actually doesn't interfere with the TORQ1 complex directly, but binds to a protein, FKBP12, that interferes with the TORQ1 complex, not TORQ2. Sapinacertib, our TORQ12 inhibitor, is a next-gen molecule that is ATP-competitive and inhibits both TORQ1 and TORQ2. If you remember my slides earlier, I showed that Everolimus does not actually shut down one of the critical pathway readouts for EBP1. It only shuts down S6. So pairing a next-gen molecule with Everolimus would not give you full pathway shutdown. You would get reactivation through Torque2 and then AKT. So we think that there is no product on the market that has our unique target profile, and nor can you piece together the target profile through other approved agents or investigational agents that we're aware of.

Gil Blum, Analyst — Needham & Company

So maybe a different way of asking this question. Why do you think pharma and others are investing so much in having the most selective BIC-3A? Because we know it's a safety story. Everyone understands that. But at what point does that stop being differentiated?

Anand Parikh, COO

Yeah, I think when you look at the mutant specifics, the initial raison d'etre was that we're going to be safer on hyperglycemia, right? That's why they were developed. I just showed you hyperglycemia data that is in the most inclusive patient population, i.e. higher HbA1c cutoffs, higher fasting blood glucose, with much lower hyperglycemia, a quarter of relay, a third of scorpion, approximately, there or thereabouts. So I'm not really sure what you're buying. Of course, you can drive hyperglycemia very, very low with some of these next-gen agents potentially, but I already think that we are at a lower bound in terms of hyperglycemia. And what we hear from physicians is that what they're looking for is greater efficacy and greater survival with, you know, convenience in dosing regimen, not really continually pushing hyperglycemia to zero. We're already at 14% with an HbA1c of 8. If you drive that to zero, I'm not sure how much

Gil Blum, Analyst — Needham & Company

additional benefit or uptake you're going to get from the physicians. I think if we had a trial

Anand Parikh, COO

with hemoglobin A1c criteria of 6.4, like the others, we would similarly see decreases in hyperglycemia, where that would be TBD, but I would guess some 14%. Yeah.

Gil Blum, Analyst — Needham & Company

So just to put that in context, because it's important, right? You have quite a few of these patients are diabetic. How much leeway does this give you as it relates to dosing of patients?

Anand Parikh, COO

Yeah. I mean, in our endometrial trial, many of these patients are medication-dependent diabetics. And so we are still seeing that these patients are able to tolerate the full dose. And so one of the things that I think is important, and one of the reasons why I think only hitting PI3 kinase alpha is a mistake, is that hyperglycemia is largely driven through PI3 kinase alpha. And so what we are able to do is to spread the hit across PI3 kinase alpha, TORC1 and TORC2. So we're not leaning so heavily on pietrokinase alpha, as the mutant specifics and others, they must lean on that single node. And so not only do you have a smaller market, but that's why you're seeing the elevated hyperglycemia that even the mutant specifics are showing as compared to the multinode inhibitors. I do have one question from the audience.

Gil Blum, Analyst — Needham & Company

how do you think that the endometrial data could potentially de-risk safety and tolerability as it

Anand Parikh, COO

relates to breast? Yeah, I think it's a great question. The way I would say it is PI3 kinase inhibitors and the pathway inhibitors have only been approved in breast. Many of them have tried but failed in endometrial. In many ways, we believe that endometrial might be the tougher disease. And so if we are able to show a data set that looks promising in endometrial, we think that the read-through to breast is significant, particularly on the efficacy side and even on the tolerability side. Again, many of these patients are medication-dependent diabetics, 90% have metabolic syndrome. That is not a phenotype that is as common in breast. So if you are efficacious and tolerable in endometrial, I believe that you will be efficacious and tolerable, potentially to a greater degree even, in breast.

Gil Blum, Analyst — Needham & Company

that's very helpful and uh again um you know we think that get out illicit is a great advancement for the space especially in wild i really haven't seen much work but i do want maybe a little your perspective as it relates to oral versus iv in front line just given the fact that we're seeing more and more of these oral agents right yeah yeah coming on i mean it's one of the reasons why

Anand Parikh, COO

we really feel like we want to get this to patients quickly, is that if you consider that we've seen all of this amazing patient benefit derived in the last few months in breast cancer, be it oral surge, be it the Victoria One study, you know, the bar is increasing in breast cancer. I think that's a great thing for patients. As that happens, the bar is going to increase not not only in second line, but in first line as well. And for patients to be tied to an IV, I think is increasingly challenging, particularly as we see a greater emphasis on oral regimens in first line. So we really want to get this out there so that the women who comprise, you know, the young working women who comprise the vast majority of the breast cancer population have the ability to go about their lives without being tied to an infusion chair.

Gil Blum, Analyst — Needham & Company

I think this is all we got as it relates to Q&A today. As always, I appreciate your time, Anand.

Anand Parikh, COO

Well, thank you very much, Gil. It's fun chatting and look forward to talking more in the future.