Alector, Inc. (ALEC) Earnings Call Transcript & Summary

Good day, and thank you for standing by. Welcome to the conference call being hosted today by Alector's management. [Operator Instructions] Please be advised that today's conference is being recorded. [Operator Instructions] I would now like to hand the conference over to your speaker today, Ms. Michelle Coral. You may begin.

Thank you, Kara. Good morning, and thank you all for joining us on today's call. I'm Michelle Corral, Alector's Vice President of Communications and Investor Relations. Today, we will be reviewing results from the INFRONT-2 Phase II clinical trial. These data were presented earlier today at the Alzheimer's Association International Conference, and that presentation is posted to the Investors section of the Alector website under Events and Presentations. A press release detailing the data was also issued this morning and can be found on our website as well. On today's call, we have CEO and Co-Founder, Dr. Arnon Rosenthal; Dr. Shehnaaz Suliman, Alector's President and Chief Operating Officer; and Dr. Robert Paul, our Chief Medical Officer. Dr. Sam Jackson, our SVP of Clinical Development, will also be joining us today for Q&A. As a reminder, the information discussed during this call will include forward-looking statements, which represents the company's view as of today, July 29, and 2021. We undertake no obligation to update or revise any forward-looking statements to reflect new information or future events, except as required by law. Please refer to today's press release as well as our filings with the SEC for information concerning risk factors that could cause actual results to differ materially from those expressed or implied by these statements. With that out of the way, I'd like to now hand the call over to Shehnaaz for opening remarks.

Thank you, Michelle, and good morning to everyone joining today's call. This morning, we are very pleased to have presented encouraging results from our INFRONT-2 Phase II study of AL001 in frontotemporal dementia, FTD patients at the AAIC conference held in Denver. This open-label Phase II study was started in September 2019 and intended to test the safety and tolerability of chronic once-monthly dosing of 001 in FTD patients with a progranulin mutation. We rolled over several patients from our Phase Ib multiple ascending dose trial, enrolled additional patients into the Phase II and included a new cohort of FTD patients with the C9orf72 mutation. The data set that we're going to walk through today is focused on 12 months data for up to 12 symptomatic FTD GRN patients and is the most robust data set from the study that we have been able to present to date. It includes safety, PK/PD, biomarker and clinical outcome assessment data using the CDR NACC FTLD sum of boxes. At a high level, we are very encouraged by the totality of our data set, The key findings, which are as follows: AL001 continues to have a favorable safety profile. With respect to target engagement, AL001 treatment, rapidly restored progranulin levels in plasma and CSF to normal in FTD granulin patients for the entire duration of treatment. In symptomatic FTD GRN patients, AL001 treatment demonstrated consistent effects on multiple disease-relevant biomarkers that reflect underlying disease activity, either stabilizing or returning these biomarkers towards normal levels. And the volumetric MRI data suggest that slowing of ventricular enlargement and brain atrophy in treated patients compared to a cohort of participants from the genetic FTD initiative or GENFI2 that served as matched controls, and we were thoroughly encouraged by our clinical efficacy observations. AL001 treated patients showed a slowing of clinical progression by a 3-point change or 47% on as measured by the CDR plus NACC FTLD sum of boxes scale compared to the GENFI-matched control cohort at 12 months. These positive trends observed across multiple end points increase our confidence that AL001 is working as designed with the real potential to positively impact disease progression over time. This is not only the first evidence of 001's potential clinical benefit. But the INFRONT-2 study is also the first study to demonstrate any signs of a potential treatment benefit in any FTD patient population. Frontotemporal dementia is an orphan indication estimated to affect about 150,000 patients in the U.S. and EU5, of which up to 10% are thought to have symptomatic FTD with a progranulin mutation. There are currently no approved treatments. As a reminder, we are currently enrolling patients in a Phase III randomized controlled trial with asymptomatic and symptomatic FTD progranulin patients with AL001. The Phase III will enroll approximately 180 patients to measure the effect of 001 on clinical decline by utilizing the CDR plus NACC FTLD sum of boxes scale. We will provide guidance and timing for a Phase III readout once enrollment is completed. With that backdrop, I'd like to now invite Robert Paul to provide an on-call of this morning's presentation and walk you through the data in detail. Robert?

Thank you, Shehnaaz. Good morning, everybody. I'm pleased to join you today to review the data we just presented this morning at the AAIC meeting in Denver, and it covers the Phase II study with AL001 INFRONT-2 and I will be referring to the slides from the AAIC presentations, which are posted at our website. On Slide 3, you see we can -- we give a background information on the disease. Frontotemporal dementia is a rare neurological disease that's characterized by a rapid progressive decline in behavior, judgment and language. FTD is the most common dementia diagnosed in patients under the age of 60. FTD is a fatal disease and patients typically only live 7 to 10 years before they die from the disease. They are no approved treatments. And for those actually who are following us the presentation, having the slides in place, they were posted, you can see on Slide 3, a map of various genetic mutations that are linked to frontotemporal dementia. It includes MAPT C9orf72, but also heterozygous loss of function mutation in the progranulin gene that accounts for up to 20% of all familiar cases of FTD and we have initially focused our clinical development efforts on frontotemporal dementia patients with the progranulin gene mutation. Going to Slide 4. We'll give you a little bit background on progranulin. It's a secreted protein that modulates inflammation, regulate lysosomal function and promote neuronal survival. Deficiencies in progranulin are cause for frontotemporal dementia. On this slide, you see on the figure on the left, illustrates that carriers of progranulin mutations have a more than 50% reduction in plasma and in the CSF levels of progranulin compared to noncarriers. AL001 is a human monoclonal antibody that was designed by Alector with the goal of increasing progranulin levels. We do this by blocking the main degradation pathway of progranulin, and that's the transmembrane receptor sortilin. And as illustrated on the figure on the right, blocking sortilin provides a therapeutic strategy for increasing progranulin levels and reversing the progranulin deficiency that causes FTD. Now going to Slide 5. INFRONT-2 is an ongoing Phase II open-label study with the main goal of assessing safety and tolerability of AL001 after chronic dosing. The study includes 3 cohorts, 2 cohorts with progranulin mutation carrier, 1 with asymptomatic carriers and 1 with symptomatic carriers. And then we also have included 1 cohort with symptomatic FTD 9orf patients as the first of several possible additional indications for our progranulin franchise program. In addition to its primary objective, this study also includes exploratory endpoints of efficacy such as fluid and imaging biomarker and clinical outcome assessment. In this presentation, we will focus on data from the 12 asymptomatic FTD granulin patients highlighted in blue. Now moving to the next Slide 6. Let's focus first on safety. So far, AL001 was generally safe and well tolerated across all cohorts. All AEs were mild and the 3 patients that experienced SAE of whom would -- on the study of whom discontinued study to an unrelated SAE of deep venous thrombosis. All the SAE was considered unrelated to AL001. Again, I mentioned the safety profile was comparable across the asymptomatic and symptomatic participants, as you can see on the table on Slide 6, and the results are also consistent with prior observation from Phase I and in Phase Ib studies of 001. Now let's move to Slide 7. So here, our efficacy data of the study on Slide 7, starting with an overview of the progranulin disease cascade in FTD. As I mentioned, progranulin plays a pivotal role in maintaining a healthy brain by regulating lysosomal function and innate immunity. The current working hypothesis of the pathophysiology of FTD GRN is that progranulin deficiency leads to lysosomal dysfunction and inflammation in the brain. This eventually leads to neuronal death and clinical symptoms. Based on this hypothesis, we measured the effect of AL003 on markers along this disease cascade starting with progranulin, followed by biomarkers of lysosomal function and complement activation. Then going down the steps in the disease cascade, we looked at neurodegeneration by measuring neurofilament light and more directly by measuring brain atrophy using volumetric MRI. And finally and most importantly, the clinical outcome was assessed using the CDR plus NACC FTLD sum of boxes, which is also the primary endpoint in our ongoing Phase III. We're now on Slide 8. Starting with the most proximal step in the disease cascade, you can see there that all patients started with progranulin concentration in plasma and CSF well below the normal levels characteristic of the disease. The mean CSF plasma program levels in the FTD progranulin patients were decreased at baseline in the CSF and also on -- in the plasma relative of those of aged matched controls. Treatment with AL001 led to a rapid and durable increase in progranulin levels in plasma and CSF and that fully restored progranulin levels back to the normal range of age match controls. So chronic AL001 treatment completely rescued the progranulin deficiency that is causing the disease. On Slide 9, you can see the primary data supporting the graphs that were just shown. Next slide, you see we are going now down the disease cascade when we assess the effect of AL001 treatment on lysosome and complement biomarkers, representing the key pathophysiological pathways that are impacted by progranulin in deficiency. Looking at Slide 10, we see cathepsin and LAMP1, both lysosomal proteins and C1QB, a component of the complement cascade. They were elevated in the CSF of FTD patients shown here in blue at baseline when compared to age matched controls shown in gray, suggesting that dysfunctional lysosome and [indiscernible] complement activation in FTD patients. Supportive of a downstream effect of AL001 treatment, we observed a time-dependent and durable reduction of these anolytes in the CSF levels to near control levels indicating that AL001 is restoring normal lysosomes and complement function in these patients. Table 2 on Slide 11 provides you the primary data for the graphs we just have shown you. Next to Slide 10 and 12. Next in the Cascade neurofilament light, a marker of axonal damage was assessed both in plasma and CSF. Longitudinal neurofilament levels in the plasma showed within patient variability over time, and the current data from plasma and CSF suggests that neurofilament levels are stable with large arrow bars over 1 year, as you can see on Slide 12. On Slide 13, we provide the mean neurofilament measures numerically showing that the baseline CSF plasma and plasma levels were 62.8 picograms per ml and 7.3 nanograms per ml, respectively. And after 12 months of treatment, the mean CSF plasma levels of the available samples were reduced to 46.3 picogram per ml and 6.5 nanogram per ml respectively, albeit with large variability. These data all suggest that neurofilaments are stable over 1 year, and we believe this is a good result. While neurofilament appears to be a useful prognostic biomarker of disease state. The field is still much learning what the utility of this biomarker may be as a means of predicting clinical response with both variability in neurofilament itself and with the behavior across different neurodegenerative diseases. Now going to Slide 14. Since INFRONT-2 is an open-label study, we needed a way to look at the results against a reasonable comparator. The progression of FTD while generally following a similar rapid and devastating trajectory can vary based on mutation, time of diagnosis and other factors, and thus, published studies would not provide an apple-to-apple comparison. We were able to access historical data from the GENFI registry to contextualize the results from INFRONT-2 Phase II study in order to estimate treatment effects of 001. It generated a matching placebo group for the Phase II cohort using a well-established propensity score method, which is intended to mimic randomization. Of the patients that were in the database that was provided by GENFI, we had identified, in a blinded fashion, 10 patients with the progranulin mutation that matched the FTD granulin patients that were enrolled in the Phase II study. And from the table on Slide 14, you can see that the baseline characteristics were quite balanced between the Phase II participants and the historical controls, especially for disease severity, age and neurofilament levels at baseline. The GENFI-matched control cohort will provide the context for the MRI data and the clinical benefit data that you will see in the following slides. Moving to Slide 15. Brain atrophy is a hallmark of FTD and therefore, we measure the effect of AL001 on brain tissue loss using volumetric MRI. On this slide, we see the measured volume changes of the ventricles, the whole brain and the frontotemporal cortex over 1 year. And consistent with published literature, a profound loss of brain tissue was observed in the matched control group shown in gray with the largest changes observed in the ventricles. Treatment with AL001 resulted in a numerical difference in favor of AL001 that was most obviously in the reduced enlargement in the ventricles as shown by the blue bars. On Slide 16, looking at the numerical results from the volumetric MRI assessments, the relative slowing between AL001 versus the matched control in the annualized changes are 10.9% in the whole brain, 16.7% in the frontotemporal cortex and 49.6% in the ventricles. And this leads us to the most relevant data set here today on Slide 17 to assess the effect of AL001 treatment on clinical disease progression, we measured the CDR plus NACC FT LD sum-of box scores in the Phase II patients over 1 year and compare them with a matched historical control data from GENFI. The graph on the left shows the individual results as well as the mean lines for the treatment and matched controls. Looking at the table to the right on Slide 17. We see that in the GENFI-matched control group, the CDR Sum of Box scores increased by 6.4 points from baseline over a year, indicating rapid disease progression. In contrast, the estimated annual change in the AL001 treated cohort was only 3.4 points. This indicates an impressive 47% delay in disease progression. These results are very encouraging. And if verified in the Phase II/III study, would represent a clinically meaningful treatment effect. We're looking at we're looking for have powered our ongoing Phase III study trial to look for a 40% difference. In summary, on Slide 18, here are the main takeaways of the presentation. First, AL001 was well tolerated in this study. Second, 1 year of treatment with AL001 showed a consistent movement in multiple biomarkers along the pathophysiological cascade of FTD, including an effect on slowing of ventricular enlargement and brain atrophy. And third, we observed a 47% reduction in clinical progression relative to matched controls that indicates that AL001 is able to slow clinical progression. AL001 appears by all the measures to be taken to be performing as intended. And while results from a randomized Phase III clinical trial will provide the ultimate confirmation, the data set presents today increases our confidence that AL001 may be -- may prove to be a safe and effective treatment for FTD. Arnon now will add a few comments on how this program fits within Alector's broader pipeline. Arnon, please?

Welcome, everyone. As you have just heard, Robert presenting a body of evidence, including biomarkers, imaging and clinical assessment data supporting efficacy of AL001 in FTD. This is just the beginning for our progranulin franchise. Human genetics and published animal model data support expanding our progranulin elevating therapy to ALS, Parkinson's disease, Alzheimer's disease and additional forms of FTD. And this is exactly what we are doing in partnership with GSK. AL001 was originated from our immuno-neurology platform. As you know, our therapeutic strategy and platform is conceptually similar to immuno-oncology. We recruit and harness the brain on immune system to counteract multiple disease pathologies. Today, our immuno-neurology pipeline includes 4 clinical stage candidates in 6 clinical trials. By next year, we will be leading up to 10 clinical trials with 7 immuno-neurology therapies targeting 6 distinct diseases. We are looking forward to keeping you updated on our progress, and we will now open the line to Q&A. As a reminder, with me on the line are Robert, Shehnaaz and Sam Jackson, our VP of Clinical Development. And Shehnaaz will now lead the Q&A. Shehnaaz?

Great. Thank you, Arnon. I think we have a queue ready to go. So let's -- shall we continue with the Q&A? Michelle, I think you have the roster.

Operator, if you could open the line to the first speaker, questioner?

And your first question comes from the line of Matthew Harrison with Morgan Stanley.

Maybe just a couple of questions for me. Two are broader and then 1 specific. I guess first is -- have you seen anything out of this data set, which would change your confidence level or the way you designed the Phase III study? The second one is, what do you think this data set tells us about the C9orf patients that we might get to start to see some data on? And then I think the third question is a lot of the questions I've been getting and people are trying to maybe put the volumetric MRI data into some more context. I mean how they should think about whole brain versus some of the larger differences that you observed, especially in the ventricles?

Yes, sure. Thanks for the question. Of course, we are very encouraged by the data. And actually, it increases the confidence that we can see beneficial effects on the Phase III study. We -- based on that, we're actually not having any reason to change the design of the Phase III. Again, we mentioned earlier that the study is powered to detect a 40% difference in disease progression over 96 weeks. We see here 12 months data and an effect of almost 50%. And so we were very pleased to see the data, and there will be no change currently in the design of the study. Second...

And to add one thing. The Phase II patients have the same CDR baseline as our Phase III asymptomatic patients. So we think that the Phase II adequately predict what will happen in Phase III.

Thank you, Arnon. Actually, if you look at the baseline table, they're actually in the Phase II a little bit milder than the GENFI data, and we actually expect that overall, we have a little bit mild population you also might remember that we also include in the Phase III patients that are at risk to getting alternative FTD. So these are mutation carriers that have symptoms, no symptoms or only very subtle symptoms. And they have -- but they have increased neurofilament plasma levels at baseline that's predictive of clinical progression over 96 weeks. So overall, we expect actually the patient population to be milder compared to the Phase II data set. Answering your question regarding C9orf. This is the cohort that is still enrolling. And once we have more data and we expect to give you an update next year on the C9orf data, I mean for us, it's important to see that we actually see an effect on lysosomal function or dysfunction correcting this and also on inflammation. And that's why this also increases the confidence that we will have potentially see effects in the C9orf FTD patients. And as you know, we will start very soon an ALS trial, which is placebo-controlled in C9orf ALS patients. And this will, of course, will also add to the data. And then for the volumetric MRI excellent question, I can give you a little bit more context. So I think it was in 2009, there was a paper came out on the atrophy rates in Alzheimer's disease across different brain regions. And this was based on EXPEDITION 3 Phase III studies from solanezumab just to set the context. And here, if you look, they looked at, I think, like more than 20 different regions. They have seen ventricular enlargement in the placebo arm annualized about 10%, which is roughly half of what we have seen here, and it's known that atrophy rate is way more pronounced in FTD. They have seen also a decline in whole brain tissue. I think it was around 1.5% or something 2%. And that's, again, very consistent across AD trials, but you also see that this accelerated grade atrophy is also seen now in FTD. And then we picked the frontotemporal cortex, which is how basically defines the disease. And we also see here atrophy rates that are comparable with with historic controls. And also to give you a context, you can we might see this is maybe quite noisy. In this paper of the solanezumab data, they actually compared solanezumab versus placebo. And they always saw some -- I mean, always going in the right direction. They saw actually, so to speak, treatment effect, but it was always in the range of 2%, 3% or 4%. And if you think that the drug didn't have any effect on brain atrophy and the 3% or 4% presents kind of a noise of the instrument, we always, in this study, we see double-digit changes. Of course, the sample size is too small to run any statistical analysis. But we see, across all 3 regions, double-digit treatment effects, and this is very promising to us.

Yes, the specific sort of question, I mean, there is no -- you expect a mirror image between the ventricle and the tissue, the less -- the more tissue you lose the bigger the ventricle becomes the bigger the open space is. And we see this qualitatively, we don't see this quantitatively and it's sort of -- it's a measurement. Variability issue means that the ventricles are the largest component that has to be -- that can be measured. So you see it's easier with this small sample size, it's easier to measure the changes. But qualitatively, we see what we expect, reduction in tissue and increase in the open space like the ventricle in the brain.

Your next question comes from the line of Graig Suvannavejh with Goldman Sachs.

Congrats on the data. Maybe a question just going back to the changes in volumetric MRI, and can you just provide perhaps a little bit more context on how we should think about the changes in the volumetric MRI? May I guess, translate into changes in brain volume? And specifically, what you're seeing in terms of 0.5, 0.3 point difference, how meaningful that changes? And then bigger picture, how you think that might correlate directly or indirectly to cognitive scores? And then I had a question just on the historical matched control group that you're putting together. Very interesting that you're doing that, and I think it's a good thing. But when do you think we'll get a sense of what the baseline characteristics of that group are just if you're going to try to use that for comparative purposes? And then I've got a follow-up question, please.

Yes. I can address it. So the -- I mean, the volumetric MRI, again, we focused on ventricular enlargement and brain atrophy because it's done across multiple diseases, neurodegenerative diseases. As also Arnon mentioned, ventricular enlargement, you see also across multiple neurodegenerative diseases. You see the effect is most pronounced. Also, why we're looking here at whole brain atrophy. This is a small sample size. And actually, of course, initially, you can also look at other smaller brain regions or left and right side, but this actually adds more noise with a small sample size. And again, as I explained last time, if you -- if -- with the last question, if you contextualize this with what do you see in AD trials, this is very consistent that you actually see most of the changes in the ventricular and also in the brain -- whole brain atrophy. We see also this in the frontotemporal cortex, which is how the disease is defined. Numerically, again, these numbers look small. However, again, if you compare this with other MRI results from other neurodegenerative diseases. Again, these are double-digit changes, which is at least what we think is, is more than noise. And all the AD trials that have been tried so far, and we haven't actually seen any effect on volumetric MRI. Most of the time, it actually went into the opposite direction, and that's why we were actually very pleased to see the results

Yes. And I might add, Graig, that [ Martino Barchetta ] showed some really nice analysis this week from [indiscernible] from the GENFI2 cohort where the granular mutation carriers showed a very clear association between brain atrophy and disease severity with progression that follow up after 1 year. So I think on the question of association of correlation, that looks to be pretty solid, particularly when you look at longitudinal data in the FTD progranulin mutation carriers specifically. So Robert, do you want to carry out the second question?

Yes. So the question correlates with the cognitive scores. We know that volumetric MRI actually well correlates over time with atrophy rate. Under treatment, that's -- so far, I don't know any drug in the dementia space that had ever seen an impact on volumetric MRI for several reasons. In some cases, you see actually pseudoatrophy, even MMS. So that's why we're actually really pleased to see this -- these numbers. And we hope that, of course, and actually, we have seen this in this presentation that translated into a trend in slowing in the cognitive scores. Then you also asked about the baseline characteristics of the GENFI-matched control. This is actually on Slide 14. And then here, you actually see the comparison with respect to the CDR plus NACC FTLD sum of box scores. It was 5.9 in INFRONT-2 and 5.7 in the GENFI-matched controls. So actually, if you think it's actually that our Phase II studies actually includes more severe patients. However, you can see that the range is very comparable. Age was also very comparable. And these are, we think, are the major driver of disease progression, and also neurofilament plasma levels at baseline pretty matched. Actually, again, it seems to be more severe in INFRONT-2. And that's why we think not only we had a very good -- very balanced group, but also a little bit is a disadvantage to our group. So that's why we were really very pleased to see the results.

Yes. And Graig, remember that these propensity scores were -- we were blinded to outcomes. That is a very important point. We did not have a clinical, longitudinal clinical data of these patients when we did the matching. We really were blinded to their clinical status, which adds a degree of face validity to the methodology. Did you have a follow-up question, Graig?

I did. It just goes back to NFL. And in light of how this program has advanced. I was curious as to if there are any updated thoughts on how we, either from a specific FTD perspective or just from a neurodegenerative perspective, should be thinking about NFL and the changes that you're seeing? Just wanted to know how important that is in the relative stacking of kind of the different outcome measures that you're looking at in terms of in the specific case 001 in FTD?

So maybe I can start and Robert can add. Means frontotemporal dementia is a version of a lysosomal storage disease. As you know, homozygous loss of function progranulin is a typical lysosomal storage disease. And just in this scientific meeting, there was a presentation on another lysosomal storage disease for which there is a very effective enzyme replacement therapy. And the data clearly shows that with enzyme replacement therapy that's clinically effective, it takes 2.5 to 3 years to reduce neurofilament. So the clinical benefit, in this case, proceeds neurofilament by years. And I think that possibly similar time line will happen with FTD, which is, again, another version of lysosomal storage disease. Recently, there was another publication on a drug, a clinically effective drug for ALS. And again, they showed clinical efficacy on the background on actually increasing NFL. So we think that NFL sort of there were a lot of hype -- there was a lot of hype and expectations about NFL because of multiple sclerosis and spine muscular atrophy, but this does not seem to be consistent with other disease and especially it seems to have a very slow response in lysosomal storage diseases, and this is our current expectation and Robert will be happy to expand on this.

No, I think this covers everything. I mean we know examples in the literature where the clinical benefit was not -- or is not associated with the neurofilament. There is the time component. And I think, Henrik Zetterberg, in his presentation, I think pointed this neurofilament is a slow biomarker. I think that's what he mentioned in his talk, and I think we agree with this, yes.

Your next question comes from the line of Goeff Meacham with Bank of America.

My question is, I just had a few. So the first one is from a regulatory perspective. It's clear that biomarkers can create a faster development path, and that's obviously the case for Alzheimer's with pretty accepted biomarkers. But in in FTD, do you guys have to spend more time validating brain volume changes or complement biomarkers or even progranulin? I'm just trying to get a sense for maybe if there is a faster path available to you based on biomarkers that we could tie to the, obviously, the recent Aduhelm approval? And then I have one follow-up.

Yes. Very interesting question. Normally, what the regulators require is a clear correlation between a biomarker and a clinical outcome, and that normally should be verified in multiple trials, and then they actually can accept a biomarker to replace basically a clinical endpoint. And these biomarkers were actually the first time that we see changes here. And that's also the reason why the Phase III study is placebo-controlled, randomized and the primary endpoint is a clinical endpoint. Of course, we collect a lot of biomarker data. And hopefully, we can then actually correlate this with a biomarker, the clinical outcome. And going forward, we can actually do faster trials using a biomarker as a surrogate endpoint. But so far, what we heard from the agencies, it's not time yet for a surrogate marker in FTD.

Yes. We have agreement with FDA on the design and endpoint in the current Phase III study, Geoff, and we're sticking to the course on what we need to do there and trying to enroll patients as fast as possible.

Okay. And then a follow-up question. Just on the CDR data on Slide 17. I guess when you guys evaluated the match controls, just curious if there were patients whose cognitive decline was much faster or much slower that you can tie to the mechanism? I'm just trying to get a sense for genotype-phenotype relationships.

Yes. You can see we -- actually on Slide 7, you see the light, smaller, thinner gray lines. These are actually actual data, individual data from the GENFI clinical participants. And you can see there is variability, of course. Just a reminder, this Phase II patients included patients that started with the global score on the CDR on 0.5 but also included 1 of 2. So actually, we included like very mild, but also more moderate patients. And that explains the variability. You can see there are some patients in GENFI that didn't move a lot, but also others that moved faster. And that's why it was so important for us actually to generate this matched control because we really wanted to have a similar sample size with respect to disease progression disease stage, neurofilament levels also age because these are all drivers of disease progression. And then you see also the individual lines in blue from our study. And you can see actually, GENFi only provides -- they do the assessments only once a year. That's where we have the baseline 0 and then the 12-month data. We actually did the CDR assessment every 3 months. So that's why you see there is, of course, some variability, as you would expect. But you can also see that most actually of our -- the blue lines actually above the mean lines of the GENFI. And again, I want to point out here that the confidence interval with 95% confidence interval of the difference actually does not include the null value. Again, we didn't do a formal statistical analysis given the small sample size, but this is also what we think is quite encouraging.

The next question comes from the line of Neena Bitritto-Garg with Citi.

Can you help frame for us, I guess, because I know there are some differences between kind of the Phase III population and the Phase II population here. I would expect the Phase III population to be up a little bit less severe. So can you kind of help frame for us how patients who maybe would have qualified for the Phase III study or who might be more reflective of the Phase III study population performed in Phase II? And then also just a follow-up is, so in terms of just your natural history cohort data here, there was a presentation on Monday by GENFI that showed the -- their full cohort of FTD GRN patients and specifically in the symptomatic cohort, they did show a slightly different kind of rate of progression on CDR and volumetric MRI than what you've reported here. It was a little bit slower than I think the rate of progression that you show here. So I guess, can you help us understand how reflective you think kind of the matched patient cohort is here versus kind of just the broader FTD GRN population?

Yes. The Phase II patients, actually, we think, are in average more severe than what we expect to have the patients in the Phase III. Even if you only look at the phase, the symptomatic cohort in Phase III, we have the at-risk patient and then also the symptomatic. So again, I think this really represents what we will see in Phase III. But it could be that the clinical decline is overall in the placebo arm less. But what's actually sort of promising even if we have patients here that are more severe in the disease we actually see this very impressive treatment effect. And if you think that earlier is better, that's why we think -- that's why we are so encouraged because we see this in a patient population that was actually pretty advanced. And just as a reminder, when we when we started the Phase II, this was really to assess safety and tolerability and PK/PD data over chronic dosing. And we were and we were not very strict basically excluding patients that are maybe too severe. So that's actually, I think, speaks for the phase -- for the results of the Phase II. The GENFI data, again, it's very important that you look at the baseline characteristics. So if you have a patient that starts with a global score of 0.5. Yes, the disease progression rate is slower. And you see that actually on the Slide 17 that there are some patients that, I mean, only dropped by 1 point or so over a year, but there are others, they actually declined by roughly 10 points. These are the ones that are -- have a global score of 1 and 2. And so the average is determined by how many patients you have for each group, and we think that this is very consistent. Again, this is what you saw are GENFI data. These are actually the same data set we're using where we actually got the historic controls from.

Absolutely. Yes, that was kind of where my question was coming from, just given that it is the same data set and you do have some maybe earlier symptomatic patients included here versus kind of a full symptomatic cohort in GENFI.

Yes.

Yes, there is variability, as you know, based on the genetic composition of the patients, different genetic causes lead to different progression rate and the CDR baseline also changes the progression rate. So we had to do a very accurate match control to really find the actual progression rate of patients, which are similar to our Phase II patients.

Your next question comes from the line of Yaron Werber with Cowen.

I got sort of a couple. The first one is on slide 12 and 13, you're not giving the baseline -- I'm sorry, the match control NFL levels, we can see them on Page 14. I was just curious as to why is there a reason they're not showing in 12, 13? There's obviously some kind of thought into that. And then secondly, when you look at the patient numbers, and I know not all patients were able to do all assessments at 12 months in some cases, you're showing 12 patients. In other cases, you're seeing -- you're showing about 9 in your cohort or fewer. Obviously, depending on about 12 months. But maybe give us a sense what's the difference there in terms of the measurements? And then finally, Robert, maybe for you, my question is that you are going to do an interim analysis at a year in patients who are symptomatic are you expecting to advance faster? Any sense you can give us what's the powering for that group?

Okay. For Slide 13. So we give the average baseline. So what -- again, what's the question about [indiscernible] because we have the graph and we have the table. Yes.

More on Page 12 and 13, you're not showing the matched historical GENFI Gentry control? And what's the reason for that?

Okay. Yes, actually -- yes, actually, one thing is Yes. Actually, that's a good point. I'm not sure if they actually do neurofilament levels so frequently. Of course, we have the baseline, and we also have, I think, for some of the patients, maybe 49, but not 25. And they also actually don't collect CSF data. So this -- that's why we actually didn't include the neurofilament analysis in the GENFI-matched controls because most of the biomarkers we're measuring also LAMP and cathepsin. They actually don't have the data. And that's why we compare this. You see the changes over time. We have the neurofilament levels actually from baseline. But also, as you can see, we have picked 10 match controls, but only in 9 were actually -- plasma neurofilament levels were available. So that explains it. And then you're correct, we have -- so the analysis includes 12 patients because these 12 patients were enrolled in the FTD progranulin symptomatic cohort. One patient actually discontinued between month 3 and 6, and this is the one that has the venous thrombosis, the SAE, unrelated to AL001. So there, you're down basically to 11 patients. And then 2, we had -- for the 12 patients, we had 2 like late comers, so they are actually enrolled and they were still enrolled at the time of the data cut, but they have not reached the 12 months treatment. And then the 9 patients basically we had at the data cut had actually reached the 12-month time point. However, only in 7 of the 12 -- of the 9 patients that had reached the 12-month time point, we have the cognitive assessments. As you remember, we had COVID. A lot of the sites actually were down. We were able actually to keep the patients or the sites we're able to keep the patients in the study, so they received drug. But they were not able actually to bring the patient and the caregiver, which is necessary for the CDR to the site. And that explains the different numbers. And how we actually did the analysis, and that's why I think important to say that we have 12. So what we did is, we used the random coefficient model for repeat measurements for this analysis. And that means we include all the data between baseline and 12 months data, not only the last time point but basically all patients, so all patients that contributed basically to the CDR assessments are included in this analysis. And I think that actually strengthens the analysis. And that's how the 12 patients are counted. Yes. On the interim analysis, you referred for Phase II or Phase III?

For the Phase III, Robert. I'm under the impression there will be an interim analysis at 12 months for patients who are symptomatic. I don't know if you can confirm that or may give details?

No. Actually, no. Initially, we thought about this. But then actually, based on feedback we got from multiple regulatory agency, they actually recommended not to do an interim analysis and for 2 reasons. One is actually, if you run the study and you know it's 96 weeks, it actually increases then the confidence of the treatment effect. You also then get more reliable data or results on the secondary end points. And you have to spend alpha, if you do an interim analysis, and they were actually the advised against this, and we follow this advice.

Your next question comes from the line of Tom Shrader with BTIG.

Congratulations on the data. I have a kind of a related question to the control group question. Just did you do any work to see how robust those 10 are, if you take the next closest 10 and the 10 after that? The things really reverse or do they just get smaller? Or just any comments you had on how robust that set was because it's the big driver of your effect?

Yes, sure. I mean we did extensive testing of robustness. So as I mentioned, the primary analysis were basically including all assessments between baseline and 12 months into account. We also then did the secondary analysis, where we only included the baseline and the last available time point. basically just excluding all other time points in between, so you basically get a line. And then we still saw an effect of 38%, still very impressive. And then we did a third sensitivity or robustness analysis is if you focus only on the 7 actually that reached 12 months where we basically had kind of a completer analysis. And there, we actually see an effect of 51%. So we did a lot of additional analysis, and it always went in the right -- very consistent in the right direction.

All right. Great. I want to thank Arnon for that ERT data. My question is this slow ability to reverse neurofilament. Is that mimicked in animals? Do you find if you left the disease in animal models go for a while, it gets very hard to reverse or is this all stuff we're learning from humans? Just any background data you can give would be great.

I can start and maybe also Arnon can jump in. So I think we have to learn from humans. The -- some of the disease model, specifically for FTD, there is kind of a model where there heterozygous knockout mice. And actually, there is no increase in neurofilament. So there, we can actually measure an impact on neurofilament in this disease model. So we actually have to rely a lot on the human data.

Yes, basically -- yes, there are no really human -- sort of there are no really good animal models for these diseases, including for Alzheimer's disease or frontotemporal dementia. Sort of the human disease is significantly more aggressive than mice that have one good and one bad copy of progranulin. So we are really learning all of this in human and I draw analogy to breaking distances in a car, basically by applying the therapeutics, we really press the brake on the disease the car does not -- is not going to stop immediately. There is a distance where the car will slow down, and eventually, it will stop. And I think that sort of the slowdown in cognitive decline, the slowdown in brain tissue loss, the normalization over time of biomarkers and the stabilization and indication of initial decline in neurofilament really suggests that we are in this braking distance period. And again, the data from other lysosomal storage diseases suggest that this braking may take 2 to 2.5 to 3 years. So what we learned from human is, again, that neurofilament is a very slow biomarker. And as a result, it may not be useful -- as useful as people thought. Basically the clinical benefit and in ALS and the mortal benefit in lysosomal storage diseases, all the symptoms of lysosomal pathologies are reversed by treatment before neurofilament is reversed.

Your next question comes from the line of Paul Matteis with Stifel.

I wanted to just ask about the Phase III powering one more time. And I guess, can you talk a little bit more about why you think 40% is conservative? I know it's a 50% relative delta here, but it is a small sample and things often kind of regress as you go to a broader population. And then I just had one other follow-up.

Yes, sure. First, actually, we don't go in a broader population in the Phase III...

Larger sample size. Sorry about that.

It's a larger sample size, yes. And the 40% basically is based on conversations we had with KOLs, a lot of people asked us what's clinically meaningful and say 40% is definitely clinically meaningful, maybe even less. And we see, of course, this 47%. We understand this is a small sample size, and that's why we actually don't change the design or the power of the sample size in our Phase III study. And again, this is based on data, again, from GENFI where we were able to model basically disease progression. This was very helpful in designing the Phase III study.

Okay. And sorry if I missed this, but can you talk about other clinical endpoints you look in the study? Did you look at the MMSE or anything else versus natural history that to kind of reaffirm your confidence in your overall conclusion here?

Yes. Actually, we didn't include the MMSE. We had the CTI and we have the R-bands, but there are not -- especially the R-bands they are not included in GENFI. Yes.

Okay. What did you see on the CGI for the drug-treated patients?

I don't think that we have analyzed the data yet. Again, we were focusing on the data where we can actually contextualize this with the GENFI data. And that's why it's -- otherwise, it's hard basically to compare it and draw any conclusion. And that's why we focused on the data where we have comparison from GENFI as a kind of a placebo control. And that was for volumetric MRI and the CDR.

Yes. Paul, we didn't have the data. That's the simple answer to the question. Any more questions in the queue, operator?

And your final question comes from the line of Carter Gould with Barclays.

You didn't show it, but can you speak to, I guess, just the longitudinal trends you saw on volumetric MRI? I'm just trying to understand kind of -- If you saw a, I guess, a bend in the curve really after 6 months or -- yes. Any update on that front? And then a lot of commentary here on the volumetric MRI data. But I guess just how critical is it really that the Phase III shows a static benefit on whole brain atrophy for -- I mean, I'm just trying to put this in context. You're telling us to focus on the ventricular data. I guess I'm hearing this because that's less receptible to noise, but we haven't really heard that, be it a focal point of your conversation for today. So I'm just trying to get some color there.

Yes. I mean the reason why we only show the 12-month data is because GENFI only does MRI assessments every 12 months. So that's why we were not able actually to compare this and contextualize the MRI data with -- from the Phase II with GENFI, and that's why we focused on the 12 months data. And we also know that you get a more robust readout if you wait 12 months rather than 6 months. And the other question about was volumetric MRI for the Phase III. Of course, we do also [indiscernible] atrophy. Again, the approvable endpoint is the CDR sum of boxes, FTLD. And that's how the study is powered. We will look at biomarkers, different biomarkers, including volumetric MRI that supports then the clinical data, but this will be a secondary actually even exploratory biomarker.

Yes. And Carter, I would say we're not asking folks to focus on the vMRI. Our message is focus on the totality of the data, and we are indeed most encouraged by the clinical signal that we think is robust that has face validity relative to GENFI2 and the methodology we use to contextualize the data. So that's really the message is, Hey, if you look at the data in totality, we're seeing some very encouraging trends here that actually add face validity and confidence to the design of the Phase III and the expectations around it. And the FDA will definitely still be focused on CDR and NACC FTLD, which is the primary driver of efficacy in the Phase III. And the signal we see here, we think, is very encouraging.

And there are no further questions at this time.

Great. So thank you, everyone, for joining us and for your continued high level of engagement. As a reminder, the data we reviewed has been posted on the Investors section of the website. That concludes our prepared remarks, and we very much look forward to our follow-up conversations one-on-one. Thanks again.

Thank you.

This concludes today's conference call. You may now disconnect.