Quanterix Corporation (QTRX) Earnings Call Transcript & Summary

So first of all, I want to thank Charlotte and Henrik, 2 of the top thought leaders in the landscape of trying to get an Alzheimer drug across the goal line, collaborating with many of the largest pharmas in the world. I'm very honored to be introducing another one of our Powering Precision Health segments. So I think it was only a week or 2 ago, I had Charlotte as well as Tatiana and Kevin Knopp on, who -- we've discussed similar topics, but more about how proteomics is going to play a major role in evolving the precision health. And one of the key points is for value creation, giving back more to the world than we take we think is so important for the future of companies that are going to create a lot of value in the world. You really have to give a lot more than you take, which means you need to be mission based. So it used to be a lot of companies would invent an technology, then they would go try to find the killer app. In our case, we want to start with what we consider to be Powering Precision Health, which is the ecosystem of all of the top thought leaders, and then go into the companies that basically could fulfill the killer apps that are identified within Powering Precision Health. And so to that end, when we look at Powering Precision Health, we always start with what we call the Di5, Disruptive Innovation 5-Step, starting with a greater purpose. If we can do it for the world, it's going to represent a significant advance where we give more back than we ever take. And in this case, we believe getting innovation to patients very rapidly across cancer, brain health and COVID are 3 of the top priorities. And we're starting with trying to get therapies across the goal line, but longer term, looking at prevention. So there's really a combination of biomarkers and wearable devices that we're really honing in on. And we think proteomics is the most phenotypic area for exploring within this ecosystem. And you can see the range of participants in this ecosystem. It's fairly significant. It ranges from the payer groups, folks like the UnitedHealth Group and Aetna and others as well as the policy-making bodies and agencies like the FDA, RADx, NIH, CDC. We also have many of the universities, leading universities around the world and medical centers, the large pharmas, the large diagnostic companies as well as the advocacy groups like Susan Komen and Alzheimer's Association. And what we really are proud about is we also include bankers and investors, where many of them are sponsors to try to help these companies and these technologies evolve. And so that's our greater purpose and our greater vision is expressed in this PPH ecosystem. Now moving to the biomarkers of protein. We've shown this slide numerous times where current detection levels on the x-axis don't allow you to see many cancers or many neuro diseases until very late stage, sometimes Stage 4 or even after death through an autopsy. And it typically requires spinal taps or invasive PET imaging to be able to see many of these different diseases. And so if you want to improve the detection timing, getting it earlier when it's more treatable and do it with less invasive testing, you need these better technology for seeing a sensitivity. The protein is typically lowest when you're healthy, elevates as you get sick, and it's also lowest in the least invasive samples like blood and saliva. So moving cancer, neurology and COVID down to the right where you could see the continuum of asymptomatic all the way into sick care is a big goal that we're trying to strive for. And you can see on this slide that there are ways today to see inside the brain through PET scans, which are pretty invasive as well as the cerebral spinal tap. And Henrik, you once told me that that's not so bad to get a spinal tap, but I know that the suites might see things a little differently than the rest of us around the world. Very painful, very expensive, very invasive. And so to the extent that we can use blood biomarkers to move neuro from up top right down to the bottom left, this is key. And you can see in these images, there's this area of asymptomatic where you can see the tauopathy and the amyloid pathology. And what's been really breakthrough over the last couple of years from folks like Henrik and Charlotte is the ability to use blood plasma to see that inner circle of pre cognitive impairment neurology pathology has already started for, say, in Alzheimer's, but seeing it early. And this is we think creating correlations in blood is an area of major opportunity and focus from the 2 leaders that we're talking with today. We think proteomics offers a lot more what we'll call utility because it's more phenotypic. But yet, it's pretty much the content is fairly undiscovered at this point. And the identical twins show us that you could be born with the same DNA, but as you expand through your life and have environmental triggers, you could have different disease cohorts achieved in an identical twin cohort by just the way you live your life. And so the protein helps to reveal that. And so we've put a lot of our focus on the protein where at the tip of the iceberg, historically, there's about 1,300 proteins being measured by Luminex, ProteinSimple, MSD, Quanterix, other companies. And Abbott, Roche and Siemens are measuring about 200 proteins and are creating about $25 billion of revenue looking at those 200 proteins single-plex, believe it or not. And so when you expand underneath of that pyramid into the significant sensitivity, we think there's another 1,000 proteins that could create a lot of correlative benefit to see in disease noninvasively and early. And that's been our focus to transform the traditional telephone into the iPhone for the field of medicine is through proteomics. And all these companies now are utilizing these biomarkers to try to get drugs across the goal line sooner and also to see disease earlier in a lot of the research explorations that are occurring. And so we built this slide here to just show the overall value chain. There's discovery of proteins early on, then you downstream get the utility of moving those correlative proteins that have a lot of value into diagnostics. And that's where the TAMs really explode. And really, the TAM today, we show $6 billion in that discovery area. It's really probably only $1 billion today. But as you know, there's been many new companies entering that discovery phase, which we think is great for the field it will bring more proteins. And -- But again, Siemens, Roche and Abbott through all single-plex and a lot of that TAM today is based on measuring patients that have already got symptoms. 90% of that revenue coming from patients that already have symptoms. Moving that upstream into asymptomatic or presymptomatic allows the expansion, we believe, of the more like $20 billion TAM or $30 billion up to $100 billion because it will allow to shift the cost of a lot of today's health care into early detection and early intervention with various new therapies. So we're really excited that this blue block -- blue box on the right represents a lot of the opportunity. So Quanterix today, when you look at this pipeline of proteomics, the discoverers are on the left-hand side. And we sit towards the end in translation where you take the proteins that show the most utility, and you put a lot of sensitivity on them, so you can see them noninvasively to really fuel the drug trials. And we ultimately recently have had good success with COVID being able to see it before symptoms. We now have a couple of EUAs. And so a lot of our diagnostic focus at Quanterix will be evolving into Alzheimer's, MS and TBI with what we've been able to advance in COVID. And we have formed some relationships with Siemens for our NfL and Abbott for nonexclusive rights to our Simoa. We also think that the liquid biopsy field will continue to expand. And you can see some of the companies here. We want to partner in that area versus go there direct. But where we see a lot of the excitement on the right-hand side is in health screens from the payer groups where they can see and monitor, survey their membership quarterly to see disease long before the symptoms present when health care can be administered and outcomes created much more productively. But I would still say the left-hand side, the discovery, the TAMs are much smaller on the left than they are on the right, but it's where the action is currently beginning to pick up with a lot of new companies entering and we're really excited to be teaming up and supporting a lot of that activity. So just to go back into the neuro landscape. A biomarker gets discovered, it gets applied and adopted by the pharmas, increases the probability that they can get a drug across the goal line by a large percentage. And we know in neuro, there really hasn't been an Alzheimer drug approved. So the pharma companies are much more willing to work with us to show validity and utility. And these are some of the key biomarkers that we're using in neuro shown on the right. Then ultimately, we would like to see a drug get approved and then early diagnostics by the payer groups to enable this to continue being fueled as an overall cycle of innovation. And in the area of the specific Alzheimer cascade, we showed this slide earlier top left, these images. But we now are continuing to advance an ADplex. We just launched the 4-plex, the Simoa 4-Plex, But we also are working on a next-generation 100x increase in sensitivity, enabling a 10-plex capability with high throughput, low cost with LIMS integration. That's our vision for the end of 2022. And we've begun to allow certain of our partners to get access -- early access to this to allow it to evolve. And Henrik is one of those thought leaders that's utilizing it. And you can see that this correlation between blood and these images is an important part of the early detection. So on the left, there's a lot of biomarkers today that are measured from these invasive procedures. On the right is the dry blood spot or the -- just these venous draws that we're trying to continue to evolve all of our biomarkers across all these neuro diseases. Two drugs got approved this year -- last year from Novartis and Roche for MS using our NfL biomarker, which was an exciting advance. And when we look at some of the attention we got early in the last 2 years around seeing the Alzheimer's cascade, 16 to 20 years elevations of the actual biomarkers before dementia hits. And then watching Lilly and Biogen right now as well as Takeda and others trying to get these different aducanumab and donanemab across the goal line, we really want to try to keep advancing using the center here, this new publication that Henrik himself put out around p-Tau181, the correlation with CSF and PET and also being able to stratify out Lewy bodies dementia and frontal temporal dementia. So we've seen so much activity in liquid biopsy over the last 5 years. But we are talking more about the liquid MRI for the brain or seeing brain health noninvasively. And we do think that a tipping point for really allowing both legs of the stool of research and diagnostics to really explode and be an opportunity for the world to benefit is when that first Alzheimer drug gets across the goal line. So today, we've got these 2 incredible panelists. I'm going to go to Charlotte first and ask her to introduce herself and make some comments. But today, we're talking primarily around Alzheimer's and the payer disruption. And Charlotte, I'm going to share screen here so that you can grab the screen and take it from here, if you don't mind.

I don't. Thanks a lot, Kevin, for this nice introduction into the field, and thank you for giving me the opportunity to share some of our recent data. I will give you a glimpse into a couple of data that we just generated to give you an idea on where we are going to and heading to. So to make it very clear which Kevin already stated this in his introduction, for Alzheimer's disease, we really need biomarkers to help us make the diagnosis. And why is that? Because Alzheimer's disease knows a very long preclinical phase that starts already 20 years before we see the clinical symptoms. So during those 20 years, at a certain point in time, we do see changes in biomarkers. But we are dependent also on those biomarkers because by definition, the clinical symptoms are not visible yet. So by definition, we also need biomarkers then to indicate that the drug or another treatment has any effect on the brain pathology. And luckily for Alzheimer's disease, we do have biomarkers, Kevin already indicated that. So we do see an increase in the cerebrospinal fluid and phospho tau levels and that reflect the core pathology tangle formation. And there is another core pathology in Alzheimer's disease, the plaque formation that's reflected in reductions in CSF biomarker levels of amyloid beta. And we also see better seen accumulation of amyloid beta. But of course, we want to replace the CSF analysis by blood-based biomarkers because that allows serial analysis, for example, and more democratic inclusion of patients into clinical trials, for example. So here is one of the results that we generated with -- of the early-stage kits for amyloid analysis. And we analyzed in this very preclinical stage the amyloid levels and questions whether these -- whether there are changes, reductions already in this very early stage in blood-based amyloid levels and whether these are predictive of conversion to clinical decline, so to MCI or dementia. And here, you see one of the graphs that we generated, and it's a clear dose-response relationship for the different reductions, level of reductions. And already in this preclinical stage, we do see that reductions in amyloid levels within the blood plasma are predictive for future decline. So that was good news. But with these old kits, the amyloid isoform that was detected was not the specific first amino acid, so the interim amino acids of amyloid beta. So we decided to develop another version of this assay, which is specific for the detection of the first amino acids of both amyloid beta 1-42 and 40. And here you see the results. And just this week, accepted by scientific reports, you see the results of the specificity of this assay -- sorry, the technical specificity. So at the left part of the graph, you can see the levels of different isoforms of amyloid beta, so mainly different in the first amino acids and how they reacted in our novel amyloid beta 1-42 Simoa assay. So you see only a little bit of noise and some elevation at a really high concentration of Abeta 1-40, so some cross reactivity. And the dash line is the level of the noise. Going to this right graph, we can see that there is a lot of aspecific binding. So this assay, this -- yes, previous version of the assay can detect a beta 3-42, 1-42 and also 2-42. So different isoforms, nice dose response relationship, but that's not what we want. So we think that there is added value of an assay that's more specific for the first amino acids for amyloid beta, and it's especially relevant for target engagement. Using this assay, the beauty of Simoa is that you can multiplex. So we multiplexed the amyloid beta 1-42 and 40 with GFAP and neurofilament light. And using the combination of these 3 biomarkers, the area under the curve for discrimination of people that are PET positive was increased, and it's an area under curve of 88%, and that's comparable to mass spectrometry analysis only for Abeta 42/40. But at least with this high throughput assay on Simoa, we can reach the same sensitivity and specificity. And yes, the advantage is the broader implementation possibilities of the Simoa assay compared to mass spectrometry. And we also related the levels of the different biomarkers that we measured in plasma with cognition. This is in AD population and controls. And we do see that there is a good correlation of these different biomarkers with cognitive measures on all domains that we tested. 0So if there is no relation, it would all cross this 0 line. So here, we see a lack of relation for plasma amyloid beta. But especially the plasma GFAP in purple and in orange neurofilament light they were related to the cognitive measures. So that gives us a good indication that we can use those biomarkers also for monitoring cognition, for example, in a clinical trial. We next wanted to know and get back again to this preclinical population. We wanted to address the question whether serum GFAP and neurofilament light were also prognostic for conversion to dementia. So in the earlier study that I showed you, we looked into amyloid beta 42 to 40. But now we looked into those neurodegeneration astrocytosis biomarkers. So we did see at baseline already an increase in the serum GFAP levels of those that became demented at follow-up. So at baseline, the levels were already different. And the same accounts for neurofilament light. So we see an increase in the demented portions. And then if we look into the predictive value of those different plasma biomarkers, we observed a nice dose response relationship of GFAP levels. And so they were predictive for cognitive decline and conversion to dementia. And for neurofilament light, we do see some relationship but it was less pronounced. So another set of data that's really recent, which I want to share with you is an analysis in twin cohorts. So in the twin cohort, we have access to blood of healthy aging individuals. So even after 10 years, we do not see a cognitive decline. But at 10 years, we were able to measure their amyloid status, either by PET or by CSF analysis. And this gave us the opportunity to look 10 years back and study how the biomarkers related at that time. And here is one of the results. So having access to 2 different biomarkers, blood samples, 10 years apart, we could also study the slopes. And I think this is very remarkable and interesting. So to remind you, at baseline but also at follow up, these people are cognitively normal. So when we look at -- we divide the group into amyloid positive and negative at follow-up, you can see that the slopes for those 2 groups for amyloid beta 42 to 40, they are similar, largely overlapping. But for p-Tau, the assay that we use here also, we see already an increase at baseline. So at minus 10 -- so before we measured the amyloid positivity, we do see an increase already in plasma p-Tau. So it's really an early starter along the continuum of Alzheimer's disease. And for GFAP, we see the same picture. So -- and it's very interesting because here in the statistics table, you can see whether there is an interaction of amyloid positivity with time. And especially for plasma pTau, we do see an interaction. So that means that the slope of the plasma levels over time was different for those who were amyloid positive compared to those who were negative, not that statistics. You can also see it by eye this line has a steeper slope. The same holds true more or less for GFAP, but that was only a tendency so not significant. So this is the only pTau data that I will share with you because I'm sure that Henrik Zetterberg will show you much more of the recent data that were generated for pTau so far. So now one word about implementation. For implementation of the blood-based biomarkers, we need standard operating procedures. We knew from the CSF analysis that standard operating procedures are important. Pre-analytics can really affect your results. So we started off to analyze that as soon as possible also for the blood-based biomarkers. Here is just one of the results that we generated. You can split this graph in 2. We measured the centrifugation delays or how that affected the blood-based biomarker levels. So no delay, that's the reference. 1 hour, 3 hours and 24 hours at room temperature. And at the right side, we kept the samples for the same time period, but now at 4 degrees. Different amyloid assays were used. The reddish one are mass spect, the grayish one the Simoas. And here were some ELISAs. And you can see that the same behavior, so independent of the methodology used, we do see a reduction in amyloid levels after 24 hours kept -- being standing at room temperature before centrifugation. But luckily, we could mitigate this effect by keeping the samples on ice. Here, we measure the storage delay in the lower part of the graph. And here, that means after centrifugation, we left the samples either at 4 hours or 24 hours again at room temperature or at -- or on ice. And again, at room temperature in 24 hours, we see a reduction for all assays for the -- yes, amyloid measured by all assays, essentially similar as what we observed when we had this delay before centrifugation. Now that's remarkable isn't it because then after centrifugation, there is no cell contact with the plasma anymore, but we still see the same effects. So the cell interaction is not the cause of the decline. After 2 weeks at 4 degrees, 2-8 degrees, we also see a similar detrimental effect. So if you want to send your samples after centrifugation to some central lab, it's better not to do that on ice, but you can keep them for the time being at minus 20, or minus 80, of course, that's the ideal situation. So these experiments and more of them were the basis of a standard operating procedure that we developed together with Alzheimer's Association. And this is currently further elaborated. But already, this is a very good stop to start with because it's also generic. It's -- so it's suitable for analysis for other biomarkers that we tested so far like GFAP, neurofilament light and also for pTau. So with that, I want to make some final conclusion. It's a long road of biomarker development, going through all those different steps, the identification, development, validation and implementation. And for Alzheimer's CSF biomarkers, it has taken us 20 years. But with all those learnings and all those exciting technological developments, I think we will beat it. And for the Alzheimer's blood biomarkers, it's my sincere hope, but I have also confidence that we will reach that in 5 years, starting from identification. Thank you very much.

So Charlotte, as you take off your screen and Henrik grabs the screen for his presentation, I'd like to comment, you've been such a special part of our Powering Precision Health foundation and been one of the top thought leaders for quite a number of years. And I have to say, I don't know if you remember the one we had in Boston maybe 4 years ago, Al Sandrock, the Chief Medical Officer and Head of all of R&D at Biogen, was on the stage, and he said, "I think with Simoa technology, within 10 years, we're going to be able to see Alzheimer's 10 years before dementia." And I was on the phone with him over the weekend and he's like, it only took like a year. And so I look at your 5-year advance and prediction. And I must say, seeing you talk about storage conditions and centrifugal standard procedures and trying to hone in, it feels to me like we're moving 4 or 5 key biomarkers simultaneously that have different roles for either stratification or early detection, but you're applying a lot of principles that's allowing a lot of parallel processing. And I don't think I've seen the field move at this pace in my 5 or 6 years being very honed in working with you. Are you sensing like a different level of opportunity now that we're getting so close?

Yes, I do. Yes, but maybe it's also naive that I think that is happening in 5 years. But okay, you can pin me down on it and we will talk again in 5 years. But especially the last year has seen such exciting results for plasma pTau. And I'm sure Henrik will allude on that. And -- yes. So every day, it's like a mushroom coming up. There is another paper on it. And it started already with neurofilament light. So it seems that everything that we touch nowadays and also thanks to the Simoa technology, yes, it looks like I'm a salesperson, but I think this is the reality. So for ultrasensitive technologies, we're now able to make big steps. But it's also due to the field being organized in such a way that collaboration is super simple, and you can go at accelerated pace. And we had the learnings from the CSF fields. And everyone knows and feels the urgency to address it as soon as possible without any delay because, yes, we should not wait any longer. And of course, the -- yes, the possibility that we will have some drugs in the near future that also is an important factor for accelerating all activities now.

Excellent. All right, Charlotte, well done. And really looking forward, Henrik, to -- and by the way, Charlotte, I believe you're a Ph.D.. I don't know if you're an M.D. I know that Henrik is both a neurologist and a Ph.D., and the 2 of you have collaborated, and we've had so many collaborations over the years, and there's probably another 15 or 20 neurologists that are in this ecosystem of ours that are publishing, as you pointed out. And I think there's now 100 -- or 1,200 publications, of which over 600 of them are for neurology. And Henrik, you were one of the very first to use Simoa. So I'm really looking forward to hearing your comments today.

Yes. I could just continue a little bit what was Charlotte also -- in relation to what Charlotte talked about now because I think actually the Alzheimer research field is -- I get a sense that it's unusually collaborative. And I think that is because we still lack the treatment. When the mutations were discovered, that was before I was in this field. Then I've heard it was a little bit of a race towards sort of the Nobel Prize finding the cure it will be. So then the field has been handled over the years. And we know that we need each other and that we need to work together. And one cool thing with this field also is that I think the first company or research group presenting a disease modifying treatment, they will be saluted. I mean they will be [indiscernible] by everyone because we need this. And the first successful drug with a good clinical outcome against which we can link the biomarker data, that will also be like a key unlocking the field because then we can start to look at biomarker data and see how well they predict the clinical outcome, which will be super exciting in the next phase, I think. And we have been working with Simoa now for many, many years, actually. And we have really liked the platform and one way through which we have been working on this is, of course, to use the possibility of doing home brewing. So I am an M.D., but I'm not an neurologist, my specialty is laboratory medicine. So I -- we at a lab want to be able to make our own experiments and build assays and try different things. And then having Simoa, this type of open platform for discovery for investigator-initiated discovery work, it's amazing. And then the platform then has the possibility of also full -- running fully automated assays on instruments that are now stable and you can use it in a clinical lab. That's -- having that breadth of possibilities, biomarker discovery work, wet lab at the bench, trying out different assays, trying different things and then having well standardized and stable assays. That's a unique combination, I think, and the sensitivity, of course. So this is what we started out with when we try to make our phospho tau test. We tried out many different antibody combinations. That's what you see in the A panel here was the one that really worked the best. So Tau12 in combination with the phospho tau antibody commercially available. And then we combine them into one assay that became quite sensitive quite rapidly. It was not a long project. Trying out the antibodies took a while, but it was -- but once we had them, it was quite fast. So we got an assay with a low limit of quantification of 0.5 picogram per milliliter. We could show it detects phosphorylated but not phosphorylated Tau. And we could start using -- look at these different assay performance characteristics with the recovery that is around 100% good CVs within and between days. And this was the first pilot study we did. So basically, we took from our clinical sample routing workflow patients where we had CSF data indicating that they were biomarker positive regarding the CSF biomarkers. And then we had their paired blood samples and we could compare quite rapidly phospho tau 181 concentrations and got this nice result, which encouraged us. And this -- we could also look at CSF serum phosphor Tau correlations and found a nice positive correlation. It's not perfect, but it's -- yes, we never expect a perfect correlation either. So this was encouraging, especially thinking about what new and what other CNS drive molecules sometimes do not correlate that well in [indiscernible]. And there was additional confirmatory studies we did just to continue to see that we could reproducibly detect this. This was people who were classified according to amyloid status. And you saw that we could see this phospho tau 181 increase in Abeta positive individuals. From this, we have paired plasma and serum samples and the correlation between the 2 matrixes is good. It's -- but you can see that the slope is not balanced. So it's good -- it's important to choose 1 of the matrixes when you do studies. And here was the first clinical study -- cohort study we did. That's another reason why this field has been moving so fast forward. I think all the clinical cohorts that have been built across the globe now. I mean they are so nice and well -- deeply phenotyped clinical cohorts through which we can examine the biomarker candidates and see if they really reflect brain changes or not. So we don't have to rely on clinical phenotypes only, but also the brain changes. So the TRIAD cohort was built by Pedro Rosa-Neto of McGill University. You see young people, cognitively unimpaired elderly, MCI patients and AD patients with a clear increase in these 2 categories. This was another set of samples from their cohort. And a striking finding here that we saw early on was that the plasma phospho tau 181 concentration could completely separate patients with frontal temporal dementias from Alzheimer's disease patients. There's no overlap at all, which made me think about the clinical -- the memory clinic situation where you have people with different types of -- according to problems and the blood test that actually could help sorting out these 2 rather common neurodegenerative dementias. Young controls here, cognitively unimpaired again, a little bit higher MCI patients and AD patients clearly elevated. And then when we looked at correlations with Tau PET, it -- there is a correlation with Tau PET and we could sort of Braak stage the patients. And you can see that plasma phosphor Tau levels, they increase. And then eventually, they seem to plateau here in the higher Braak stages. So people seem to move up in plasma phosphor Tau and then become increasingly positive in Tau PET over time. And -- but there is a correlation with amyloid also. And this is where we have had to revise our view on the Tau biomarkers because they respond to amyloid. And that is why they do not work in non-Alzheimer's tauopathies, which is a category of diseases where we have to do more development work. And we are trying now -- Charlotte is trying -- and many of you on the call are also looking into this, I'm sure. Those diseases are progressive supranuclear palsy, Tau positive frontotemporal dementia and some forms of corticobasal syndrome. Actually, the chronic traumatic encephalopathy could be regarded as a secondary tauopathy where the classical tau biomarkers do not work that well. So those markers, Charlotte and I have been working on mostly in terms of tau, or Alzheimer type tau pathology markers, one could say, responsive to amyloid. So the first thing that will happen is that you get amyloid deposits. Neurons exposed to those -- or close to those deposits will react to the amyloid, phosphorylate and secrete tau and that is sort of the whole biomarker cascade in Alzheimer's. And then, of course, there are other tissue reactions that most likely also involve the astrocytes, as Charlotte pointed out, with the GFAP test. So we have the possibility now of looking in Alzheimer's on different types of responses to the amyloid pathology. This is in familial Alzheimer's disease. Noncarriers here, presymptomatic carriers and symptomatic carriers. There are a couple of high individuals with our assay. These are rather young people, they shouldn't have -- they don't have Alzheimer's. We are a little bit interested in whether there is something in regards to the sample handling that could explain this. For example, delayed centrifugation, perhaps Charlotte could comment on this later also. It looks like the phosphor tau levels -- group level data indicate that they start to increase 12, 13, 14 years before expected year of onset in familial Alzheimer's disease. So it is a really early marker as is amyloid, of course. Some people have asked a little bit about if I should -- I mean, are there amyloid tests needed? And I definitely think they are. So amyloid -- it would be good to have an amyloid pTau ratio in lab that works well to detect onset of amyloid, and then I could look at the response in terms of Tau phosphorylation also. I think that will be helpful when we have the treatment. The phosphor tau -- the plasma phosphor tau levels are -- can detect preclinical amyloid pathology. And this is from the 1946 cohort, the brain substudy of this large population-based birth cohort in the U.K. So here around -- a little bit less than 500 people were invited for cognitive examination, amyloid PET imaging and CSF sampling for some of them also and blood, of course. These are -- were then cognitively normal. And around 20% had preclinical amyloid pathology and the phosphor tau test could detect them with an AUC of 0.72. So not perfect but still something. This is in Down syndrome. So due to the triplication of the APP gene due to trisomy 21, these individuals have quite -- have a strong risk of getting Alzheimer's disease and amyloid deposition. And that is well known and that happens much like in familial Alzheimer's disease when people are in their 30s or 40s. So here are asymptomatic Down syndrome individuals with a plasma phosphor tau levels in the -- progressive in terms of cognition progressively deteriorating Down syndrome individuals and these are Down -- individuals with dementia. So you see that the biomarker can detect this in Down context as well. This is a bit of a business liability. This was a study we had the fortune to be able to do. It wasn't designed to study phosphor tau. It was a study where at King's College in London, where people enrolled in a brain donation program. So they gave plasma samples. They underwent cognitive examination. They also subscribed to donating their brains when they died. And this study was ongoing for many years, from 2001 to 2012. And there are different time points in these studies that there are blood samples collected at different time points prior to death. And then the neuropathological examination was done at King's College by the neuropathologists there. And then in 2019 and '20, we did the measurements with our new -- our phosphor tau assay. And then here are the results. So to the left, you see 8 years -- samples 8 years prior to postmortem control individuals, non-AD dementia and AD dementia clinical diagnosis. And these are the -- this is the pattern on samples collected 4 years prior to postmortem and then 2 years prior to postmortem. But you see that the pattern is really identical. So it looks like the biomarker signals Alzheimer's neuropathology early. And this is the data where there were -- in relation to Braak staging of tau pathology. And you see this type of step-wise increase from Braak stage I, II, III, IV and V, VI, step-wise increase, visible at all time points. And this is when you look at the neuropathological diagnosis at postmortem and compare across different neurodegenerative changes. So here you see people who were normal in their brain pathology, AD patients, patients with AD and CAA, the cerebral amyloid angiopathy. And then you see Lewy body pathology with Alzheimer's and Alzheimer's also with TDP-43 pathology and all these that have Alzheimer's pathology have increased phosphor tau as a common trait. The other neurodegenerative diseases were not increased. So it is an Alzheimer's-specific test. And here is just another longitudinal display of these biomarker measures, basically showing that already 8 years prior to postmortem the biomarker works to detect this pathology, to detect at postmortem examination. We also completed the ADNI study. So if you're interested in playing around with the plasma phosphor tau concentrations, you can download all these data from the LONI website, the open access data repository where the plasma phosphor tau levels have been uploaded. And this is an analysis Alexis Moscoso did. He works as a postdoc in -- PhD student in Michael Schulz team here in Gothenburg. So here you see the baseline plasma phosphor tau levels. And then you see the change over time in picogram per milliliter per year. You see that in -- when phosphor tau levels start to increase, the slope of the increase goes up also. I think this actually works well with the data Charlotte presented. Then eventually, it looks like there could be some kind of plateau here. And this is in relation to different imaging and then the CSF measures. I think actually this is the most interesting thing here. So the first thing that happens in Alzheimer's disease is that CSF Abeta 42 levels go down. And then that is followed by amyloid positivity. So the brain amyloid will become possible to detect by amyloid PET and then CSF phosphor tau levels increase. And at time point 0, you have the plasma phosphor tau positivity. So the amyloid markers are -- seem to be a little bit earlier, but plasma phosphor tau is quite early marker as well. Good -- and then, of course, we are continuing to try to develop better biomarkers in this field and one important thing is to try to get more sensitive. And then the Quanterix people have developed this amazing study. I really encourage you to read this paper. So basically, it is the Simoa technology with single molecule counting, but an optimized version where we actually reduce the number of capture beads and combine that with more efficient seasonal analysis. I think I understand that paper -- perhaps not all of it, but some of it. By this, you could actually make the -- some of the assays 100-fold more sensitive than the regular -- for the regular Simoa. And we now have an SRX instrument that has been upgraded to be -- so that we can play around with this. And we have started. I don't have data to show here because we just started. So what we will do here is to examine mildly positive phosphorylated tau forms. And we talk about hyperphosphorylated tau but what we have been measuring in all of our work is basically mono-phosphorylated tau, and we think that correlates with hyperphosphorylated. But here I believe we'll see if we can start to detect tau forms that contain more than one phosphorylation site, perhaps, by combining different phosphor -- Anti-phospho-Tau antibodies with each other and see what that looks like. Then for amyloid beta, we have pyroglutamate-modified Abeta. And this is exactly what Lilly is targeting in their recent drug trial. This is a form of beta amyloid that forms in the brain tissue. It is not secreted by neurons. At least we have not -- never been able to see it secreted by neurons. But by mass spec, we can see it in the brain tissue. When we try to measure it in fluids, we get not that good results, and we are struggling with the sensitivity of the assays. And this is an assay we will evaluate on this upgraded [indiscernible] instrument. Yet another form of amyloid, which is super sticky, that is Abeta43, that contains one more amino acid in the [indiscernible] making it even more water insoluble and more aggregation prone. And this is type of a beta form that we would be keen to be able to measure. So these are things we're working on with this new technology, and we hope that we will be able to share some data soon. Yes. I would like to thank you all for listening, and these guys were the ones who were most active in the development of this in-house plasma phospho tau 181 asset together with Kevin and me in Gothenburg. Thank you.

Henrik, really incredible. And again, the whole point of these Powering Precision Health podcast panels is to just get information out there to others that are in the field that are also trying to explore different pathways. And over the years, it's amazing to me the number of new collaborative opportunities that's emanated from sessions like this, Henrik, when you would present or Charlotte would present. I know Charlotte is looking at some major upcoming studies that we're talking with her around that could implicate many of the pharmas as well because part of what we're hearing and seeing, and before I say this, I'll say that there's probably 10 questions that we'll ensure post this meeting, we'll get answers to each of those that asked the questions. We probably won't be able to answer all of them live today, but we'll get those answers to you. But clearly, in the field, there's a belief that if we can see precognitive impairment cohorts, we could enrich the cohort and provide a better opportunity for the drug to have efficacy. And so there's a significant push right now for most of our pharma collaborators wanting to see the pathology earlier. And they're also wanting to stratify out different types of pathology that might look like Alzheimer's, but the agent won't be efficacious towards it, like Lewy bodies dementia or frontotemporal dementia. And it does seem like when I listen to your presentations on 181, there is no question, Henrik, that you are advancing the field on both of those fronts, seeing it earlier and also seeing it in a more enriched way. And by doing this, the drug companies have a low invasive way to significantly enhance their cohort to improve the probability then that the agent is going to have efficacious results. And so that's a big piece of the drive right now. One of the questions that's coming up, and I hear it from most of the pharma companies too is around pTau-231. You've done a lot of work here with 181, it has incredible impact, and we've already launched this product, I know Quanterix has. And we have the 4-plex that was another question that's already been launched. But the 231 is an area can we see even earlier? Is it -- will it enable even earlier pathology and understanding that the amyloid pathologies already occurred and seen, as you point out, that you see deposits of the tau post the amyloid. Could 231 -- coupled then with higher sensitivity of 100x, could that be -- that combination allow maybe another major advance in the discrimination at the real low early precognitive impairment cohorts?

I think 100x sensitivity will not be needed. I think actually the Simoa assay for 231, we have published a paper on 231 in collaboration with ADX using the [indiscernible] and developing an assay that works really well and seems to be a little bit earlier than 181. 217, we have struggled with a bit and we will try it on 100x to see if we can get a sensitivity that is -- that makes it possible to measure also control levels. Because our assay work -- for 217 works in CSF and it works in AD patients, but the controls are so low. And it might be that this isn't -- the hypothesis that we hear from Randall Bateman's team and also other people is that 217 is a pathological tau phosphorylation. 181, we all have in our CSF or rather a little bit. And then when it increases, that's bad. Newborns have very high levels of 181. Most likely the 181 phosphorylation together with 231 and perhaps other phosphorylation sites, that also relates to synaptic pruning plasticity processes. And I -- for one thing that the Abeta, the basic mechanism is that Abeta blocks induced pathological pruning, sort of hijacks a normal physiological process in the brain and that pruning signal that the Abeta can do in the developing brain could be something that is reflected by increased phospho tau 181, whereas 217 might be more pathological. And perhaps that explains why it would be important with even more sensitivity for 217, at least that's our experience. But then for the other tauopathies, I also think we will have good -- great benefit from increased sensitivity from 100x because some of those tau forms might be depleted from the fluids and not increased in the fluids. I mean that's one -- it looks like tau aggregation in the non-AD tauopathies happens inside neurons and perhaps we should look for lost fragments instead of released fragments. Abeta makes neurons spit out their tau in one way or another. But in non-Abeta -- nonamyloid neurodegenerative disease, it looks like the Abeta stays in the neurons. And perhaps then there are some more released forms of tau that do not get released that well in such neurons. Much like the Abeta 42/40 reduction. Perhaps there is a tau fragment that is reduced in a beta tauopathies that could give the meaningful information.

So Henrik, just listening to you answer that question, describing the pruning potential effect and how these levels of sensitivity are starting to open up whole new ways of thinking about where the pathology might be coming from and going and given the drug companies a better chance at really honing in with an agent that's got the best chance of performing, I must say it's just so exciting because I don't think you could have answered the question the way you just did probably 6 weeks ago because of the advances that have occurred in the last 6 weeks are actually needed to be able to answer the question the way you did, and that's how rapidly this field begins -- it's beginning to hone in. I look back at cholesterol and troponin for the [indiscernible] and just how they evolve didn't have this level of scientific rigor coming from so many different centers from around the world. And you're all publishing in a way that you could all collaborate opposite those publications and continue to advance. And I know we're running out of time. And Charlotte, I would like to come back to you if I could because some of what I think I've heard from Henrik today further fires me up around the whole pTauopathy and how sometimes you have non and positive amyloid pathology that the taus could act different ways. And we already know that in the area of MS, just simple NfL, neurofilament light, has been a way to track whether or not a given drug is actually having -- being effective on a specific patient. And the sensitivity is needed to look at a patient level. And you've got some upcoming trials in the area of Alzheimer's. And I look back at what we've done with MS and NfL, some of those on the call are asking about algorithms around these panels. Like is it possible to start looking at data algorithms across how NfL might act versus amyloid beta 40/42 versus now maybe 3 different pTaus, 181, 217 and 231 or 3 -- even heard 235 starting to come in. How do we get data to start to be looked at across the panel and use it in algorithms? And I'm curious if you could comment on this because it seems like the algorithmic opportunity here is significant for less than 10 analytes.

Yes. I think it's a very good point. So especially because we will analyze more and more markers, so then we will become more and more dependent on algorithms to start to interpret the results together and -- of this combined analysis. So yes, we could envision that we develop an algorithm where we first, look at neurofilament light. And if it's negative, then maybe you look at another biomarker like pTau and amyloid beta. But it also depends on the clinical question and the context of use, whether we use it in a memory clinic or whether we use it at primary care or in a trial. So this whole area of developing algorithms is really about to start. And we have a project targeting that as well. So I hope to be able to share more data in about half a year. And I think one of the next steps that we should do, so first, we should define our algorithms in respective cohorts and validate it. But then the next step is to validate also those algorithms in more noisy populations like a general memory clinic population, for example, and not just a highly selected one who also donated CSF or PET -- or underwent PET analysis because that -- there is some selection bias in there. So up till now, we looked at large retrospective cohorts, and it looks fantastic. But I think it's very important to now go to the real world's evaluation because that's the next step that we should undertake.

So Charlotte, in closing out this discussion with you and Henrik, I find it this -- the technology itself of being able to see with a lot of precision, same answer with very high levels of sensitivity that also allows some level of automation and even low cost. Having that technology is one key component. The second one then is having the data and the ability to analyze the data to try to hone in on where algorithmic opportunities exist. And then to me, the third category that seems to be really evolving now is the samples themselves. And you heard Henrik describe that that's one of the advantages, is that we've got so many well-characterized sample sets all over the world. But to your point, the utility of that given sample set could be biased based on the way those samples were collected or whatever. And so how we get the sample understanding around all the collaborators, what samples are out there that exist and how could they be potentially utilized to help us see utility coupled then with the data algorithms to say, in a given sample set, if your use cases, we already know they've got memory loss. What's the best algorithm to kind of make sure that group gets the best medication and then ultimately benefits from it versus someone maybe in a health care group where they've got members that are long before precognitive impairment, they want to know now because they want their -- an agent to stop the progression of Alzheimer's now that they have it. That cohort and that sample set might look very different, right? So there's a lot of different use cases. So there's a major field here that is beginning to evolve and the 2 of you are sitting on top of it. And I'm just so proud to be part of this journey around how to use Precision Health into the -- the whole cascade of neurology. And I just want to thank the 2 of you because don't forget dry blood spots and even less invasive samples might even require more sensitivity. And I think in the end, we really want to be able to allow this field to get where cardiology has gotten to. And it deserves it. You've worked your entire lives at it. And to me, the field has got so much trauma around the world, so much pain and suffering for neurodegeneration across all fields, and concussion was brought up in some of the questions. By the way, Henrik, is one of the experts in the world as is Ramon Diaz in the area of concussions, a lot of great data there as well. But I want to thank the 2 of you, and we'll be doing this again, I'm sure, many times. And questions, direct them to the 2 of them or to us, and we'll further collaborate you with both Charlotte and Henrik. Thank you very much.

Thanks a lot.

Thank you. You're welcome.

Bye-bye.

Take care. Bye-bye.