Good morning, and thank you for standing by. My name is Jeannie, and I will be your conference operator today. At this time, I would like to welcome everyone to the uniQure phase I/II AMT-130 Huntington's Disease Program update. All lines have been placed on mute to prevent any background noise. After the speaker's remarks, there will be a question and answer session. If you would like to ask a question during this time, simply press star, followed by the number one on your telephone keypad. We do ask you to limit yourself to one question and rejoin the queue for follow-up. If you would like to withdraw your question, press star one again. Thank you. I would now like to turn the conference over to Chiara Russo, Senior Director of Investor Relations.
This morning, uniQure announced additional data on patients treated with our investigational gene therapy, ifezuntirgene inilparvovec, or AMT-130, in our ongoing phase I/II clinical trials in Huntington's disease, taking place in the U.S., E.U., and the U.K. This update consists of data on clinical endpoints, an exploratory biomarker, and safety and tolerability. Joining us for this investor event and webcast are Matt Kapusta, our Chief Executive Officer, Dr. Walid Abi-Saab, our Chief Medical Officer, and to provide a clinician's perspective on the experience of patients with Huntington's disease, Dr. Victor Sung, Professor of Neurology at the University of Alabama at Birmingham, UAB, Director of the UAB Huntington's Disease Clinic, and a clinical investigator in this trial of AMT-130. Kylie O'Keefe, our Chief Commercial and Strategy Officer, will be joining us for the analyst Q&A at the end of this presentation.
The slides included in this morning's webcast have been filed with the SEC and will be available on the Events and Presentations page of uniQure's website shortly after the conclusion of this webcast. Please note that we will be making forward-looking statements during this investor call. All statements other than statements of historical fact are forward-looking statements. They are based on management's beliefs and assumptions and on information available to management only as of the date of this webcast. Our actual results could differ materially from those anticipated in these forward-looking statements for many reasons, including, without limitation, the factors described in uniQure's quarterly report on Form 10-Q, filed on July 29, 2026, and other security filings. Given these risks, you should not place undue reliance on these forward-looking statements, and we assume no obligation to update these statements, even as new information becomes available in the future.
Now, I am pleased to introduce Matt Kapusta, uniQure's CEO.
Thanks, Chiara, and good morning, everyone. Thank you for joining us on today's call. This morning, we announced additional data from our ongoing studies of AMT-130 in Huntington's disease. These results continue to provide evidence of long-term, meaningful slowing of disease progression and support the primary analysis underpinning the license applications we submitted to the FDA and MHRA. To date, no other program in Huntington's disease has demonstrated long-term disease modification after a single administration. AMT-130 does. As pioneers in this space, we are blazing a new trail in HD, learning and deepening our understanding with every data point we generate. At 48 months, we continue to see meaningful treatment effects. Huntington's disease is relentless and progressive. There is nothing available today that changes its course. Our aim now is to bring forward the first disease-modifying treatment for people living with Huntington's.
The data we announced this morning include analyses at two time points. The first is a 48-month analysis based on the first 12 high-dose patients that have reached this time point. The second is a 36-month analysis reflecting all 15 patients, including three additional patients who have reached that time point since last year's update. Starting with the 48-month data, we continue to see treatment benefit across both cUHDRS and Total Functional Capacity, or TFC, with particularly significant effect on TFC demonstrating a 61% slowing of disease progression and stable treatment benefit at 48 months. TFC is the most direct measure we have of what patients and families care about. It captures whether someone can hold a job, manage their own finances, and perform the daily tasks that allow them to live independently. It is also the agreed-upon primary endpoint of our confirmatory study.
While the primary cUHDRS analysis did not reach statistical significance in 48 months, we believe the trends are positive and meaningful for HD patients. HD patients' lives are defined by moments, and a year of meaningful moments is something no patient should have to give up. There is also an important limitation in the new updated external control. 53% of the external control data is missing at 48 months. While missing this is inherent to long-term natural history datasets, the patients who left the control were declining markedly faster than those who remained. So over time, the comparator increasingly reflected a healthier population than at baseline. This bias cuts against AMT-130, not for it. It understates treatment benefit and makes statistical comparisons at the 40-month time point more difficult.
As we advance as leaders in HD, we will continue our collaboration with key stakeholders to assess the most meaningful ways to evaluate long-term treatment benefit. When comparing cUHDRS at 48 months across doses, we do see clear dose dependencies, with the high dose showing a one-point favorable effect on cUHDRS compared to the low dose. Importantly, this is an internal comparison that does not depend on the control and is what you would expect to see if there was a genuine treatment effect. Turning to the updated 36-month data, this is an important time point from a regulatory perspective, as it's consistent with the BLA primary analysis in our planned confirmatory study. Including the three additional patients, the 36-month data demonstrate an even stronger treatment effect. We now see 80% slowing of disease progression on cUHDRS, 67% slowing on TFC, and favorable trends across all other clinical sub-domains.
As a reminder, the three additional patients at 36 months are not included in the 48-month analysis. In summary, we strongly believe AMT-130 has the potential to deliver meaningful long-term benefit to people living with Huntington's disease, and we will continue to work tirelessly to deliver this treatment option to them. Patients and loved ones waited long enough for a therapy that changes what this disease does to them. We firmly believe AMT-130 is that treatment. As we move forward, we remain laser focused in the near term on three areas. First, continued engagement with the FDA and MHRA as their reviews advance. Second, timely execution of our confirmatory study targeting initiation of patient screening before the end of this year. Third, advancing our commercial preparations so that we are ready to deliver this therapy if and when approved.
With that, let me hand it over to Walid to walk you through the data in more detail. Walid?
Thank you, Matt. Good morning, everyone. Before I go over to the updated analyses, let me set the stage by reviewing the 36 months analysis from last year. This is the study design and statistical analysis plan supporting our recent regulatory submissions to the FDA and MHRA. Our primary endpoint is the Composite Unified Huntington's Disease Rating Scale, or cUHDRS. We compare change from baseline at 36 months against an external natural history cohort, ENROLL-HD, using a propensity score methodology. At the time of last year's analysis, which we presented to you in September of 2025, 12 patients on the high dose had reached 36 months of follow-up as of the data cutoff of June 30, 2025. The main takeaway from this analysis was that AMT-130 high dose showed a statistically significant slowing of disease progression at 36 months.
75% slowing on cUHDRS and 60% slowing on Total Functional Capacity, or TFC. I draw your attention to the decline of the matched natural history cohort after three years. These correspond to an average decline of half a point per year on cUHDRS and 0.3 points per year on TFC. I will come back to these numbers later in my presentation. Today, we are updating you with the results of the analyses based on the data from the June 30, 2026 study data cutoff. 12 patients have now reached 48 months on the high dose, and 12 patients have reached 48 months on the low dose. In addition, three more patients have reached 36 months on the high dose, making the 36-month time point fully representative of all enrolled patients at the high dose, factoring in the two patients who dropped out at 12 and 15 months.
There is an important comparative update that I would like to explain to assist in understanding the interpretation of both the 36-month and the 48-month data. The 48-month protocol and statistical analysis plan, which we submitted earlier this year, pre-specified the most current ENROLL-HD dataset released in September of 2025. We will refer to this as ENROLL-HD B throughout this presentation. Last year's 36-month analysis used the January 2023 released dataset, which we will call ENROLL-HD A. ENROLL-HD B contains everything in ENROLL-HD A, plus more than 2.5 years of additional follow-up on those same participants and 6,000 additional ones. It is the largest and most contemporaneous longitudinal natural history database available. It also overlaps in time with recruitment and follow-up of High Dose Cohort two in our phase I/II studies. Contemporaneous controls are an FDA preference, and that was a significant factor in the choice.
Our primary analysis follows the same plan as the 36-month analysis in 2025 that supported the BLA submission. The primary endpoint compares the change from baseline in cUHDRS at 48 months in the high dose against the propensity-score matched control group from ENROLL-HD B. TFC is a key secondary endpoint. This slide shows the high dose in orange and the low dose in purple across all patients. The panel on the left shows all available data. The panel on the right shows data from the 12 patients at the high dose and the 12 patients at the low dose for whom we have results up to 48 months. Since last year's update, these 24 patients accrued data from month 36 to month 48. There are several key observations worth noting on this slide.
First, there is a softening of approximately half a point between month 36 and month 48 on the high dose. Two components of the cUHDRS drive this. DMT, which declines somewhat after month 36, perhaps a regression to the mean following relative stability between months 34 and 36. TMS, which was noisy, with a few patients showing very large swings between month 36 and month 42. Second, we see dose dependence across four years, suggesting a treatment effect. This observation matters because it does not rely on a matched external control. The high dose, which is 10x higher than the low dose, shows evidence suggestive of disease slowing when compared to the low dose.
Third is after four years, the low dose, which can be considered a pseudo-placebo, meaning it is expected to produce a lower therapeutic effect and could serve as an internal reference, shows a decline of approximately two points on cUHDRS. That is in line with the half a point decline per year in the matched external control in the two and three-year analyses reported over the last two years. Lastly, the results at 48 months are still missing the three patients who have not yet reached the 48-month mark in the study, and at month 36 showed a slower rate of decline than the 12 patients who reached that time point a year earlier. As noted, our primary analysis uses ENROLL-HD B.
On the left panel, that analysis shows a 44% slowing of disease progression on cUHDRS with AMT-130 high dose compared to the matched controls, not statistically significant at P equals 0.144. It is noteworthy that the decline of the matched controls at month 48 is 1.61 points, which is only slightly more than the 1.52 points at month 36 that we reported last year using the matched control from ENROLL-HD A. I will address this topic later in my presentation. On the right panel, we show the results with TFC. TFC is a key functional endpoint. It is the most direct measure available of what patients and families care about. It captures whether someone can hold a job, manage their own finances, and live independently. TFC is also favored by regulators and the primary endpoint for our confirmatory study.
In this primary analysis, TFC shows 61% slowing compared to the matched external control with a nominal P value of 0.008. Similar to cUHDRS, the decline in the matched controls from ENROLL-HD B is lower at 0.94 points at 48 months on TFC compared to 0.88 points at 36 months in last year's analysis using ENROLL-HD A. These graphs show the time course for both endpoints. Again, I'll point out that the 48-month data do not reflect the complete high-dose data set, as the three additional patients included in the 36-month data have not yet reached this time point. Both cUHDRS and TFC show continued treatment effect through 48 months. These are the results for the other cUHDRS components, which all show favorable trends relative to external controls. The Stroop test remains strongly positive with significant nominal P values.
As mentioned previously, DMT shows some softening between months 36 and 48, and total motor score is relatively noisy. Let us take a closer look at the ENROLL-HD B matched controls. Participants with longer follow-up tend to decline more slowly, which enriches natural history comparators with slower progressors over time, a known phenomenon in natural history studies in many different rare diseases. The graph on the left clearly illustrates this point in the cohort of participants matched to AMT-130 high dose. Participants whose last data point was at year one declined 0.91 points per year on cUHDRS. Those who remained through year four or beyond declined by 0.41 point per year, less than half the rate. The effect compounds over time, with each additional year of follow-up selecting for slower progressors.
Missingness in ENROLL-HD B was also greater than ENROLL-HD A, exceeding 50% at year four, which we believe further accentuates the survivor bias in this data set. The graph on the right shows the consequence. ENROLL-HD B declined 20% less than ENROLL-HD A at 48 months. Notably, ENROLL-HD B begins flattening after year two, inconsistent with known disease biology, while ENROLL-HD A continues a linear decline consistent with expectations for early manifest patients. We believe these factors understate disease progression in the external control, and therefore the treatment effect of AMT-130, and that this bias compounds with longer follow-up. Given our observations on ENROLL-HD B's underestimation of disease progression, we conducted a post hoc analysis using ENROLL-HD A, the data set used in our regulatory submissions to the FDA and MHRA.
On the left, this graph shows 54% slowing at 48 months in cUHDRS, nominally significant with a P value of 0.041. I draw your attention to the fact that in this analysis, the level of missingness is lower than ENROLL-HD B, as more than half the patients matched at baseline remained in the study at 48 months. In addition, the decline in cUHDRS at month 48 in the control group is 1.94 points, which is closer to the half point per year we have reported in the prior analyses. On the right, we show the results for TFC with a 68% slowing and a nominal P value of less than 0.001. The data appears to be robust. Patients who received the AMT-130 high dose showed very little decline four years after dosing on this measure of real-world functional abilities.
For many of these patients, that means preserving a greater degree of independence, the ability to work, and to care for oneself, which is highly clinically meaningful. This slide shows the time course over the 48 months compared to the matched controls from the ENROLL-HD A and the dashed gray line from the post hoc analysis. The solid gray line shows the matched controls from the primary analysis for reference. Let's turn to the results at 36 months, the time point used for our primary BLA submission and pre-agreed with the FDA for the confirmatory trial. These are the data we shared last year, the original 12 high-dose patients through 36 months shown in the orange dotted line.
Adding the three additional patients who have since reached 36 months, shown in the solid orange line, we see they're progressing more slowly than the original 12, favorably affecting the mean change from baseline for both cUHDRS and TFC. The gray line is the matched natural history control from ENROLL-HD B as pre-specified in the primary analysis. These are the results at month 36 from the primary analysis. AMT-130 high dose showed a robust 80% reduction in cUHDRS decline and a 67% reduction in TFC decline with nominal P values of 0.005 and 0.011, respectively. Now that the analysis includes every patient who received the high dose in the study, we believe these data further strengthen the September 2025 analysis that formed the basis of our regulatory submission.
Across the other components of the cUHDRS, AMT-130 high dose showed favorable results ranging from 59% slowing on total motor score to over 120% slowing on the two cognitive endpoints, SDMT and STROOP, both with nominally significant P values. Turning to neurofilament light chain, or NfL. From month 18 onwards, both doses trend near baseline. At the 48-month cutoff, mean change from baseline in CSF NfL was 4% above baseline on high dose and 8% below baseline on low dose. For context, healthy subjects aged 40- 60, matching our population, typically show a 1%-3% annual NfL increase, while untreated early manifest HD patients show a 10%-15% annual increase. Therefore, we believe the baseline NfL levels after four years are well below what would be expected absent effective treatment. Let's look at safety next. AMT-130 was generally well tolerated.
The most common adverse events relate to the administration procedure, all of which resolved. Since September 2025, there have been no new treatment-related serious adverse events in cohort one and two at either the low or high dose. As previously disclosed, since our September 2025 update, there was one treatment-related serious adverse event in cohort four, the low cerebral volume cohort. CNS inflammation in one high-dose patient fully resolved with a short course of corticosteroids. There was also one suicide in a low-dose patient approximately five years post-treatment, assessed as unrelated to therapy. Unfortunately, people with Huntington's disease experience a 9x higher risk of suicide relative to the general population, making it one of the disease's leading causes of death. This table provides the full adverse event data supporting the summary on the prior slide. In summary, AMT-130 continues to demonstrate meaningful slowing of disease progression.
With all patients having reached 36 months, we again see a substantial treatment effect reaffirming the basis of our BLA submission. The 48-month analysis shows continued evidence of meaningful slowing of disease progression as measured by both TFC and cUHDRS. The dose dependence observed between the low and high dose across 48 months supports evidence of biological activity. We believe survivor bias and missingness in the updated external control likely understate both disease progression in this population and the treatment effects observed with AMT-130 at four years. Overall, AMT-130 continues to be generally well tolerated. To conclude, we believe that there is clear evidence that AMT-130 has the potential to slow disease progression and provide people living with Huntington's disease the opportunity to maintain a higher quality of life for longer, which is incredibly meaningful to this community who have no approved disease-modifying treatment options.
Before I turn over this presentation to Dr. Sung, I want to say that these data will be presented at a future scientific meeting.
We continue to make progress towards bringing this potential therapy to patients. We have submitted the BLA to the FDA and expect acceptance in the fourth quarter. We also submitted the MAA to the MHRA, which has since validated the submission. We are committed to a timely execution of the confirmatory study. Preparations are on target. We expect the first patient to be screened before year-end. We are also busy advancing commercial preparations and look forward to providing more color on launch strategy early next year. With that, let me turn over to Dr. Sung, who will provide his take on these results. Victor?
Thank you, Walid, and good morning, everyone. I am Dr. Victor Sung, Director of the Huntington's Disease Clinic at the University of Alabama at Birmingham, and I am pleased to be here today to provide a clinical perspective on this additional data for AMT-130. My comments are based on my own clinical expertise and experience with AMT-130 as a site investigator on this study. Since the HD gene was first sequenced over 30 years ago, we have been on a quest to find a therapy that can slow the progression of the disease. But today, there is still nothing. There are currently no FDA-approved therapies to slow the progression of Huntington's disease. The biology is one of inevitable progressive decline. With that backdrop, the data from the AMT-130 program remain encouraging to me.
On their own, they provide stronger evidence of a clinically meaningful effect over four years, a trend that continues to point in the same direction on both cUHDRS and TFC, four years after a single administration. It is a pattern that I want to keep watching closely. But my impression is that the totality of the data point toward a meaningful treatment. I think it would be helpful to underscore at least two factors in my evaluation of the data. First, I think it is worth revisiting and emphasizing that the three-year point remains the time point that is the grounding for the BLA and confirmatory study for a reason. Three years is long enough to see a measurable difference in disease progression, but short enough to align with an individual patient's research participation window.
I certainly welcome uniQure's production of longer-term data to provide additional context, but it is not necessarily the case that longer-term data are more indicative of benefits than three-year data. Second, as a clinician working with patients with this disease, I think of the data in terms of its real-world significance. When I look at an external control that appears to decline more slowly in patients with longer follow-up, that would be contrary to my impression and knowledge of how the disease typically progresses. It is also contrary to how I see my patients progress in their disease. This speaks to why TFC matters clinically, and in my assessment of the data at 36 and 48 months, it is not just another statistical score. It tracks whether a patient can still hold employment, manage finances, perform chores, and care for themselves.
In the real world, as opposed to statistics, patients can either do these things, or they cannot. A sustained treatment difference on that measure through four years suggests to me that the treatment effect may be translating into real functional capacity for these patients, despite a less than conclusive primary analysis. I am also encouraged that the confirmatory study will be grounded in measure that relies on real-world function as an assessment. A deeper look at the data and an eventual full panel at 48 months may help understand the longitudinal clinical profile a bit better. But it is relatively clear to me that if the performance presented today were reproduced in the wider patient population, that would still constitute a meaningful improvement for patients and their families.
We will slow this disease before we stop it, and on balance, I believe this data represents a meaningful step toward doing exactly that, and I am happy to reflect on it with you today. Now I will pass it back to CEO Matt Kapusta to lead the Q&A.
Thank you very much, Dr. Sung. With that, operator, we are ready to open the call to Q&A. Thanks.
At this time, in order to ask a question, press star, then the number one on your telephone keypad. I would like to remind everyone to limit yourself to one question and re-queue for any further questions. Your first question comes from the line of Moritz Reiterer with Guggenheim Securities. Please go ahead.
Hi. Thanks so much for taking my question. This is Moritz on for Debjit. Assuming AMT-130 does receive accelerated approval, perhaps for both the uniQure team and for Dr. Sung, what do you think will be the impact of this four-year data on AMT-130's ability to gain market share? Maybe just one quick follow-up. Could you just clarify which data at this point has actually been submitted to the FDA and MHRA? Is it just the N equal 12 patients, and is it compared to ENROLL-HD A or B? Thank you so much.
Yeah. Hey, Moritz, it's Matt. Excuse me, it's Matt Kapusta. Thanks for the question. Let me take the second question first. The data that's been provided in the BLA submission is the data that we presented last year. That was the data that was agreed to with the FDA that could be the primary basis of the BLA submission, and that data incorporates 12 patients at the high dose, and it's compared against ENROLL-HD A, which was the ENROLL-HD data cut that we presented last year. For the second question, I'll pass it over to Kylie.
Thank you very much for the question. I think one of the things that's important is we remain highly confident in the ability of AMT-130 to meaningfully slow disease progression, which is of profound importance to HD patients who currently have no disease-modifying treatments. As Dr. Sung said today, we think the three-year time point is incredibly important because it balances the ability to show meaningful slowing of disease and also balances the natural history components, as well as the elements of research participation for patients. We think this four-year data is additional longer data that helps demonstrate that meaningful slowing of disease.
Yeah, and I just want to add one thing. When you step back and you look at four years after a single administration, these patients haven't even lost a full point in cUHDRS. They haven't even lost four-tenths of a point in Total Functional Capacity. This is clinically meaningful for patients, and I think with the additional follow-up data, we believe that there's a significant unmet need here that needs to be filled. Dr. Sung, do you want to add any comments?
Yeah. My main comment, like you were alluding to, Matt, is just that in a setting where we have no approved therapies that can slow the progression of the disease, the 48-month data still offers that we are still slowing the clinical progression of the disease. From the perspective of my patients that have been treated and are in the study, they're still holding very stable. I think that's kind of the lens that I look at it all is we don't have anything else that can slow the progression, and this is still doing that at four years, which is still significant.
Your next question comes from the line of Joseph Schwartz with Leerink Partners. Please go ahead.
Hi. Thanks very much. It is pretty striking how TFC holds up better than cUHDRS, even against the updated control data set, which is plagued by survival bias. Why do you think that is? What led to using the updated ENROLL-HD cut? Was this an FDA request or your own choice? Do you think the FDA will rerun the 36-month analysis with the updated data set? Do you think that that could constitute a major amendment?
All right. This is Walid. I will take the questions. There were a bunch of them. TFC is the endpoint that matters most to patients, to clinicians, and also to regulators. It measures function, and any decline there actually has immediate consequences to people's quality of life, whether they are able to hold a full-time job or take care of themselves. That is why it tends to be much more relevant. It is very impressive that after four years, as Matt indicated, the 12 patients that reached that time point only lose 0.3 points. cUHDRS tends to be a little bit more sensitive. It picks up on some of the cognitive endpoints and the motor scores. We have had some noise in the motor score, and the cognitive tests sometimes show variability, and performance could change based on what is happening on that given day.
Whereas TFC gives you a much better sense of what is going on overall with the patient. I think it reflects more how they are doing in general. I think that is what we are seeing here. In terms of why did we use the new cut of ENROLL-HD, ENROLL-HD is an ongoing study. CHDI releases data updates over time, every few years, and they released a data update in September 2025, which was between the last time we did the analysis on the three years. At that time, we used the most updated ENROLL-HD data cut.
When it was time for us to evaluate the four-year data, we did what was the right thing, which is to use the most recent ENROLL-HD data cut that provides 2.5 More years of follow-up on the existing patients and 6,000 more participants, has more contemporaneity compared to our study, and that's really what regulators usually look for. It's difficult for me to speculate what will the FDA do, whether they will rerun the analysis with the new data cut or not. That is something that we will deal with if the FDA asks us to run it or if they run it themselves.
Yeah. I will add that we have rerun the 36-month analysis with ENROLL-HD B, and the results continue to be very meaningful. The divergence between ENROLL-HD A and B really become more striking at 48 months. They're starting to diverge at 36 months, but it's more modest. The impact on that analysis, even if it is rerun, is not incredibly meaningful.
I noticed there was also a question about major amendments. It's really very difficult for us to speculate. If the FDA asks for the data, we definitely will provide it, and it's at their discretion whether they believe this is a major amendment or not.
Your next question comes from the line of Paul Matteis with Stifel. Please go ahead.
Hey, thanks for taking my questions. In the past, you commented that ENROLL-HD patients at some of the earlier time points, like two and three years, performed fairly similarly on these endpoints, as compared to some of the other Huntington's disease natural history studies. Is that still the case with ENROLL-HD B, and is that still the case at four years? I guess, taking a step back, I guess the reason why I ask is I think many of us are trying to think about what are some of the sensitivity analyses FDA might do. So one logical one would be to look at other natural history cohorts that are out there and available. So maybe that's one piece.
And then really just on that other side of that, what other sensitivity analyses here do you think might be relevant stress tests of the data that the agency could do, and how many of these have you done yourself?
Well, thanks, Paul. First, on ENROLL-HD, and I think I mentioned this a little bit in my presentation, we have started sharing comparison to external control, starting with the two-year time point. If you recall, at that point, we compared to TRACK-HD and PREDICT-HD. And at that point, again, it was equivalent to the match controls were declining by half a point on cUHDRS and 0.3 point on TFC. Last year when we updated on the three-year analysis, just because over time, there's not enough patients in TRACK-HD and PREDICT-HD to be able to be matched at the three years, they suffer from significant level of missingness. We switched to ENROLL-HD, which is more robust. And again, it showed that at three years, and actually throughout the year one, two, and three, a half a point on cUHDRS and a third of a point on TFC.
This is the first time that we see with ENROLL-HD B, and as we showed in the presentation, there's a break between starting to see at year two, very evident by year four where you see a significant break from this linear progression. In the end, averaging about 0.4 points on cUHDRS, which is 20% lower. So we do think that the data overall support the previous analyses. The other piece that I want to say is that if you look at our internal control itself on the low dose, even if you assume that there's no efficacy whatsoever on that dose, they also decline about half a point on cUHDRS. So I do think that the overall analyses do support the fact that beyond three years, ENROLL-HD B is really suffering from this survivor bias and underestimating the decline in the natural history cohorts.
In terms of sensitivity, we've run a number of sensitivity analysis, more than 10 approximately, which were pre-specified and included in the SAP. Across all of those, across the majority of those, I should say, we have better percent slowing of disease progression, a higher level of decline in the control, getting closer to that 2.0 that we talked about, and we reached statistical significance on more than half of those. In fact, the primary analysis that we shared with you was the more conservative of the group. I think that kind of answers your question.
Your next question comes from the line of Joseph Thome with TD Cowen. Please go ahead.
Hi there. Good morning, and thank you for taking my question. Maybe just the first one, can you talk a little bit about the consistency of response in the high-dose patients? I think Dr. Sung mentioned that overall his patients seem to be relatively stable. Is the decline between year three and year four driven by a few patients? Or any context around that. Then maybe for Dr. Sung, if you could just talk a little bit about if this is approved, maybe what proportion of your patients you think would be good candidates for therapy, kind of like within the first year or so of launch. Thanks.
Thanks, Joe. The data overall are very consistent. There's little changes from an outlier. Maybe one exception on cUHDRS, which is driven by that TMS that I mentioned. There was some noise in the data, and there was one particular individual that probably was pulling the average down significantly compared to last year. We're following this with the side. But overall, if you look across the remaining patients, they're all showing consistent results, and particularly on TFC, they're all continuing to progress or not progress, actually, as the case it is, between years three and year four. I'll turn it over to Dr. Sung for the second question.
Yeah, I think for the first thing, I feel like the TFC is more representative of what I am seeing in my patients. I really did not see any drop-off between year three and year four. Mine are still holding steady. I think as Walid Abi-Saab said, some of this may be statistical noise by one patient, and I think TFC kind of is more truly representative there of that stability. As far as if this were to be approved, what percentage of my overall patients do I think would be good candidates? It is complicated because it is who I think medically would be good candidates combined with who would necessarily want a brain surgery for it, and who would be able to afford it and those kinds of things. But, I think it is going to be in that
Obviously, this is really better in the early mild stage, so in the kind of 20%-30% range of my patients. I would consider strongly as a candidate for it, but I think that is still a pretty sizable population overall.
Your next question comes from the line of Eliana Merle with Barclays. Please go ahead.
Hey, guys. Thanks for taking my question. I guess, just given the lower interpretability of the external control data from year four, how should we think about the meaningfulness of even longer follow-up data at five years and beyond? Has the FDA given a stance on how they view the quality of external control data at these longer time points? Also, I guess, is there any way to mitigate statistically the survivor bias in the external control? Thanks.
Yeah, thanks, Ellie. You bring a very good point. I think there is an optimal point beyond which natural history becomes very difficult to rely on because, again, they lose a lot of patients, and you have significant missingness and survivor bias that really underestimates the treatment effect. We are working with experts to see how best to analyze the data beyond year three because in order for us to be able to interpret the results, we need to have a robust methodology. Statistically, there are ways by which one can do this. However, those are never really foolproof, and they can attenuate some of that a little bit, but not to the full extent.
We're going to be working with experts in the field to really use what is accepted as a state-of-the-art way of interpreting these kinds of things, and we will look forward to sharing that with you. In terms of the regulators, they did not say anything specific about this topic, nor was there a reason for them to do so. I cannot add anything to that point. Maybe I can remind you that the primary analysis is a three-year. That's the basis of it. As was mentioned before, we believe that is sort of the sweet spot, so to speak, between having enough decline in the natural history that we can compare to and show that the drug actually is slowing disease progression, that this is a length that the patient can participate in trials and still find it acceptable to be part of trials.
This is the optimal length, potentially, for a natural history like ENROLL-HD, beyond which it becomes a bit more difficult. That's why in our confirmatory study, three-year time point is the primary analysis. In that study, of course, we're comparing to a randomized open-label internal control standard of care.
Your next question comes from Luca Issi with RBC Capital Markets. Please go ahead.
Oh, great. Hi, team. This is Shelby on for Luca, and thanks for taking the question. Maybe for Matt. We've obviously seen that Karim Mikhail will be speaking on Monday as part of the Cell and Gene Therapy Conference by the Alliance for Regenerative Medicine, and we're all eager to hear what he will say more broadly. But what's your take on how he will approach your application specifically? Do you think he'll rely simply on his staff, or do you think he'll have more of an active role in the ultimate decision, similar to what we saw in the past with Peter Marks for DMD? Overall, just curious if you've met with the guy and your thoughts on his regulatory approach. Any color there, much appreciated. Thanks.
Yeah. Hey, Shelby. Matt here. Thanks for the question. Honestly, I think those are questions that are probably better suited for Dr. Mikail directly. It's not appropriate for me to answer questions about how he'll interact with the team or behave on our review. What I will tell you is what I've said now for a year and a half, since some of the organizational changes at the FDA, that we are focused on our data. We strongly believe in our data, and we strongly believe that what we've here demonstrated at four years, quite frankly, is unprecedented, and that this data demonstrates that we can significantly slow disease progression. It really doesn't matter who's leading CBER, who's the commissioner of the FDA, who's on the review team.
What we're relying on is the strength of the data and the incredible unmet need here that exists with HD patients and their families and the lack of options that they have.
Your next question comes from the line of Salveen Richter with Goldman Sachs. Please go ahead.
Good morning. Thanks for taking my questions. Can you just remind us why there was data missing in the original ENROLL-HD dataset and why it is believed to be higher in the updated dataset? As you think of rerunning your prior analyses here, how do we think about the 36-month data cut still being stat sig using the new ENROLL-HD? Thank you.
Thanks, Salveen. Look, data missingness is a well-known phenomenon in any longitudinal natural history. This tends to be biasing against us in this case, of course, because it underestimates the degree to which these controls are progressing. When we looked at the ENROLL-HD previous data cut, what we are calling ENROLL-HD A, which was released in January 2023, and we examined the effect of missingness over time, those effects were modest and not as important as what we have seen now with ENROLL-HD B. As a result, we were not concerned about doing the four-year analysis using an updated dataset from that same study. Now, seeing it the way it is right now, of course, it is an issue, and we need to go back and evaluate what is the best way to analyze the data going forward.
Despite that, as you saw, and thanks to the results from all the participants, all 17. So 15 reached year three, but the analysis includes also the two that dropped out early. All of these data at month three with ENROLL-HD B continue to show 80% slowing on cUHDRS and more than 60% slowing on TFC. I think those are driven mostly by the fact that these patients, at that time point, our patients who received the AMT-130 high dose, show remarkable stability. So whether you analyze them with this new data cut of ENROLL-HD, which underestimates the progression at three years, or the previous data cut, you are having a very robust effect. Now the question is, what do we do going forward? That is a separate question.
I am very heartened to see those results, and I do think that they indicate that treatment with AMT-130 is effective at slowing disease progression. The exact percent, there is always error bar around these estimates. But at the end of the day, these people are gaining significantly longer periods of stability, and that is going to translate to meaningful clinical benefit for these patients.
Your next question comes from the line of Yanan Zhu with Wells Fargo Securities. Please go ahead.
Oh, great. Thanks for taking our questions. Was wondering, again, on the ENROLL-HD A and B cuts. If the missing data at outer years, is that a result of more severe patients dropping out of the study, or is it the result that more patients simply haven't reached a longer time point, and therefore, at longer time points, you will naturally have more missing data? Related to that, I was wondering whether Dr. Sung can comment on could the shift between the A cut and the B cut be due to a changing standard of care and reflecting of that? Lastly, if I may, TFC is the confirmatory endpoint or proposed confirmatory endpoint. Is there any way that it can be given more consideration for the primary endpoint? Thanks.
Okay, thanks for the question. To answer your question, whether with the new database cut for ENROLL-HD, whether participants would not have had enough time to reach four years, that is not possible because we select the patients to have the opportunity to finish four years. We call it an effective enrollment period because otherwise we end up with a situation that you mentioned, and we end up inflating this missingness. These missingness here, while we don't know exactly why, and we're working with CHDI actually to better understand why is there a difference between the missingness with the previous ENROLL-HD data cut and the new, more recent data cut. What we do know for sure is that those for whom we have data up to one year are progressing much faster than those for whom remain in the trial at four years and beyond.
The difference is about twofold between these two time points. Clearly, those who are having a more severe progression, they are dropping off early, and therefore they're biasing this. I'll jump to the TFC bit, and then I'll give the floor to Dr. Sung. The primary analysis that we agreed to support the BLA submission was cUHDRS. That is an intermediate clinical endpoint that is reasonably likely to predict the meaningful clinical endpoint, which in the eyes of the FDA is TFC. In rare diseases, when you are evaluating the data sets that you have, in general, regulators evaluate the data from all various endpoints that you have in a trial, and it's their prerogative to use the data from TFC, even if it's not pre-specified as the primary and it's secondary.
Even with that, I will remind you that TFC was specified as a key secondary, and in our hierarchical testing for the three-year analysis, it was statistically significant following cUHDRS, both in ENROLL-HD A, which was the analysis we submitted to the FDA, and in this updated analysis at three years. From that perspective, I do not think regulators will be constrained to a great extent by what we chose. Regardless, both of them are statistic. With that, I will turn it over to Dr. Sung.
Yeah, I wanted to briefly address. I think the tone of your question really is: Is there something with ENROLL-HD where this data is not there between year three and year four? ENROLL-HD has been going on for over 10 years. We have been a site for over 10 years, so there is lots of data in there, but as Walid has said, there is just this thing with all natural history databases that patients drop out more over time. As far as why there is a difference between ENROLL-HD A and B as far as missing this in particular, there are things that we speculate about and, as Walid said, we are going to discuss that with CHDI further to get some color there. I have my own thoughts on that, but I think it is all speculative at this point.
Then to address your question, has there been a change in clinical standard of care that could account for the difference between ENROLL-HD A and B? I would say no. I do not think there are any new medications that have come out or treatments that would cause, especially in a mild group like this, all of a sudden for them to deflect and stabilize between year three and year four. I do not think it is that. I think it really is the missingness of the data, as we have discussed.
Your next question comes from the line of Patrick Trucchio with H.C. Wainwright. Please go ahead.
Yeah. Thanks for taking my question. This is Cynthia for Patrick. For the same two high-dose patients and the same match control methodology, what is a month 36- 48 cUHDRS treatment effect bridge, or what do the individual high-dose cUHDRS trajectories show between months 36 and 48? Thank you.
Cynthia, can you clarify? Do you mean the 12 subjects who completed the 48 months, how did they look at 36 months versus 48 months? Is that what you're asking?
Yes. Basically, for the same 12 high-dose patients using the same match control methodology, could you maybe walk us through how the cUHDRS treatment effect changed from 36 to month 48, including changes in the treated and control groups?
We have not conducted a statistical analysis model only on the 12 completers. The matching is done on all of our AMT-130 treated subjects. All I can tell you is that the observed difference from baseline, which we show on our slide, at year three for the 12 patients on cUHDRS was 0.38, and at year four is 0.90. That is the difference that we see in those 12 patients. I'm not sure if I'm answering your question correctly, but that's as best I understood it.
Again, if you would like to ask a question, press star one on your telephone keypad. Your next question comes from the line of Lachlan Hanbury-Brown with William Blair. Please go ahead.
Hey, guys. Thanks for taking the question. Maybe just sort of following up to a degree on that last question, on the sort of discordance in progression between cUHDRS and TFC from month 36 to month 48. Walid, I think you were saying earlier that is driven mainly by sort of noise from maybe one or two patients on other parts of the cUHDRS. Can you maybe just, A, confirm that I heard that correctly, but B, for those patients, what does that actually look like clinically? You have highlighted that TFC is sort of what matters. So what does it look like for a patient that is pretty stable on TFC but declining more on cUHDRS in the real world?
Yeah. Let me be very clear. For cUHDRS, we think that the decline or what we see in the subcomponent mainly driven by two of the scales, SDMT, where a number of patients have shown a decline between year three and year four. This is a cognitive test. TMS, which tended to be noisy with a couple of patients showing really significant change versus their previous measure, which is not really consistent with also how they were doing on TFC as well. What I can tell you is that I believe that, again, TFC measures almost how a patient has been doing and is doing with a little bit of more going back over time, a few weeks or a few months. Whereas when you do the cUHDRS, which is an evaluation of cognition, a cognitive test, and TMS, those are sort of time bound.
However you were feeling that day or they sometimes could depend on the effort that you put in and so on and so forth. cUHDRS has to be more sensitive to pick up certain things, but it is also more variable because it does not have that, what I like to call area under the curve. Like how have you been doing in the past few weeks? Because if you have lost your job or you are not able to function because you are losing performance, then that will be reflected in TFC. If one day you are a bit off in the way you are performing on a cognitive test, that will be reflected directly on cUHDRS. That is how I can best respond to this question.
Ladies and gentlemen, this concludes today's conference call. Thank you all for joining. You may now disconnect.