So next up, we have Claire McKay and she her background backgrounds in as an imaging neuroscience, and I think she spent the last 10 years working in ageing and degenerative diseases. And she's now speaking mostly today about her role as the imaging informatics lead in the dementia platform. Okay, thanks very much, Annie and Charlotte, for organising this event and asking me to come and talk to you.
And so, as I said, I'm going to talk to you today about a specific bit of work that's been done by many people working in dementia research over the last 18 months, two years. And I hope I'll show you that things are really changing in dementia research, as Ms. Nisqually outlined already for us. And it really started for all of us when the prime minister made dementia, its health care priority at the beginning of the last parliament.
And this sort of ended a long, long time of dementia being, as Edwards pointed out, radically drastically underfunded compared to the burden that it turned that it has for society. But it's a big problem. And part of the reason it was underfunded is because it's seen as a rather big, difficult, intractable problem.
So if we really want, as a biomedical research community, to make a difference in terms of the impact of this disease or a set of diseases that cause dementia, we need to crack to massive problems. The first is that we need to develop neuroprotective therapy. We need to develop ways of protecting the brain before the damage is done. That's a big problem in itself. And the second is that we need to identify the means to identify people for whom this intervention is going to be useful.
And we're not going to be able to solve either of those problems unless we have a very dramatic, large scale commitment to collaborative working, to open science, to cooperation. And that's quite a big change in the way that we as scientists and clinicians have traditionally worked. So and I'm going to underline something that Emma said, which is that there are roles for clever people in all aspects of translational neuroscience.
So down at the bottom end of this matrix, you can see the sort of nuts and bolts, the means by which we understand the mechanisms of diseases, the genetics, the cell biology, the the down the microscope stuff, the scary in the lab stuff that's that's beyond me. For example, in the middle you can see the bird, the slightly higher order, things like imaging and cognition. That's the sort of thing that I do.
And then at the end of the of the translational neuroscience spectrum, as you might think of it, you also have the development of trials. How are you going to implement really good clinical trials, develop and implement trials? How are we going to turn the developments that come from further down the and the spectrum into really good diagnostic tools that really can be used at the bedside in your clinics,
for example? And we need to do this across the range of different diseases that cause dementia. And a lot of it is also about understanding age, the normal ageing process and what this does in the brain.
And so the real the real challenge was posed again, I suppose, in 2013 at the first ever G8 summit for dementia, which in itself was a real achievement to have a G8 summit specifically about this set of diseases, where the declaration was that the G8 health ministers commit themselves to the ambition to identify a cure or disease modifying therapy by 2025. That's only 10 years away. That's incredibly ambitious about to do so.
They're going to dramatically increase the amount of funding available for dementia research. So a very, very ambitious goal and some commitment to more funding. And indeed, the amount of funding that was made available in the very next year was was very different from any other year in dementia research. So this was a list of the funding opportunities that were put together at the beginning of 2014 and added up to something like 130 million,
including the 50 or so million. That's an outline from the and that came from the UK government and on several more millions that came from the UK. So we're getting there. The things are going in the right direction, but there's still an awful lot of work to do and this is what the kind of basic problem is. So this is a cartoon of what happens in in all diseases that cause dementia. You start off in a state of cognitive health and then at some point you start to deteriorate.
And in that in that early deterioration stage, you might be described as being in the programme of the disease. And as things progressed further than you would, might be diagnosed with full blown dementia. And the problem is most of the trials, almost all of the trials have ever been done in dementia are done at this stage of the disease. So once you already know that, what the.
What the status is of a patient, they've already been diagnosed with Alzheimer's disease or Parkinson's disease or whatever it is, and that's the point at which the disease modifying trials have been targeted. There are millions, if not billions of pounds have been spent trying to find treatments at this stage of the disease. And clearly, when you think about it like this, that's too late. The damage is already done.
So we really need to be pushing back and doing the trials much earlier in the disease. But you can see the problem immediately there. So if you're trying to target your trials in the red zone, you've got much less of a signal to play with. You're not quite sure whether people have got the disease yet or not. And even if you even if they aren't being able to tell the difference between somebody who's deteriorating and not deteriorating,
you're dealing with a much shallower slope. This is this is something that changes over time. So we the only way to solve that problem is to get better early diagnostic markers. And they probably not single markers. They're probably combinations of markers that will give you the best chance of being able to see the signal early. And they have to be done in a large scale, has to be done in large numbers of subjects. And so that was really the starting point for the Dementia Platform UK.
So we can sort of set out the challenges in these three categories, the scientific challenges that we need to go early and we need to learn rapidly. We can't just be reinventing each other's wheels all the time. We need to come together as a community and and put our put the best brains that we have into action at this early stage of the disease. This poses a very large infrastructure challenge.
So the UK has invested millions and millions of pounds in cohorts like the Whitehall cohort that you heard about earlier today. But these cohorts on the 1946 cohort that you also heard about are not necessarily joined up. You can't necessarily take data from the two and put them together. They're not always focussed on dementia.
So there are actually lots of cohorts in the UK that have potentially useful information in them, but they're not necessarily been focussed on dementia and measuring the outcomes that we would be interested in. And we need to create an environment where instead of all competing with each other, all the institutions competing with each other, we create a sort of platform for studies to be done at multiple centres.
So attract the pharma companies to come and spend their big money in the UK using our experimental experimental medicine platform. I'll tell you more about on the point of all of this is the economic challenges are outlined in the previous talk with it kind of goes without saying that the the the the burden of these diseases means that we must put some energy into trying to solve these problems.
But we also need to reduce the research transaction costs a million to the 50 million pounds that gets spent every year by the UK government. We need to make sure that's being spent in the best possible way so that we're not, as I say, reinventing each of wheels. And we really need to be honest. It's incredibly expensive.
Face the trials that are failing at the moment. Mm. So the Dimensions Platform UK was put together really at lightning speed over the course of the last couple of years and each stage of the process, we didn't know that the next stage was coming. So the first stage was really about establishing the cohort's, what cohorts are out there that exist that can be potentially put to better use if we pull them together.
The second stage was about starting to think about creating this experimental medicine platform, and then the third stage was buying all the kit that we need in order to create this the this multicenter infrastructure for four trials for studies and trials. And it was always designed as a partnership between academia and industry so that we're creating what we really want to have, which is an open science environment, so that we're not all competing with each other.
So you can think of the DPRK as being a sort of three main sections. The first section is about the cohort's bringing together all of these cohorts. The second section is about developing the specialist networks and then the third section is the experimental medicine programme. And I'm just going to go through each one of these and tell you a little bit more about it. So if we start with the cohorts, there are there are at the moment something like 34 cohorts included in the UK.
But the number goes up all the time because there's no barrier to entry. If you have a cohort that you think might be useful and interesting and you're willing to share and be part of the platform and everybody is welcome and the cohorts are kind of separated into its conceptual into these three different types. So the the easiest one to to think about is the pink is the pink set.
These are these cohorts that have been put together specifically for the research they contain, very valuable but quite small numbers of people who perhaps carry a familial gene. So we know that they're likely to get dementia or those specific, well, phenotype disease cohorts excuse me. And the second set are the the green ones. And these are the very large case rich cohort.
So these these contain many thousands of participants and even millions of participants in some cases, but would not necessarily originally designed to be for dementia research, but maybe with a little bit of extra work, we can use some of the information in there to to good effect. And then the third factor, the other are the other are the cohorts that have been put together deliberately for this preclinical stage phase of dementia research.
So these are people, healthy, elderly people who may have some kind of evidence that are of of cognitive decline. But for the most part, this is a group of people who have been put together with with dementia research in mind. But at the moment, they're still healthy. And there's a special case within this, which is the UK Biobank.
The UK Biobank is an incredible resource that has been collected over the last five to eight years of 500000 people who are across the across the adult age range aged between 30 and 70. And at the moment, 100000 of them are being image. This is way bigger in imaging terms than we've ever conceived of before. So this is this is going to really change the game in terms of the bringing of epidemiology into the into the realm where we can look at complicated metrics from inside the brain.
And very importantly, these people have all consented for recontact, which means they're potentially amenable for further studies and trials in the future. And the Dimensions platform Yuki's is enhancing the UK biobank by rescaling 10000 of the 100000 so that we'll have to time point imaging on these people.
And it's not been fully decided yet, but it's likely that this will be all these 10000 will come from the older end of the spectrum, goes to ten point imaging is much more sensitive in terms of being able to see that decline. See that slope that I showed you earlier? And these people also have an extended cognitive background. So this creates what we hope will be a cohort of readiness, ready for the the new intervention trials when they come online.
And then the second bit of the of the platform is the specialist networks. And this has really been put together by this very large capital investment that the MRC have made, the UK government's made through the MRC. And the biggest chunk of this has gone to equipping the UK with five new pet MRI machines. So pet imaging allows you to see the the chemistry of the brain, the molecules in the brain.
MRI is the best way of seeing brain structure. If you put the two together, you have a new, very expensive machine. That means that you can get both sets of information at the same time. There are two already in the UK and so we'll now have a seven site pet network which will dramatically change our ability to to to be. A great place for pharma companies to come and do their trials if to just explain why that's so important.
Many of the the the large and expensive failures that you might have heard about in the media over the last few years, the amyloid targeting drugs that have not been successful in changing people's cognitive status. Well, it turned out that many in many of those trials, a lot of the participants didn't even have amyloid in the brain. So there was this was an amyloid reducing drug given to people who didn't have amyloid in the brain. So and, you know, that's that's a bit of a travesty, really.
But the best way to stop that from happening again is but before you give any amyloid busting agents, you image them first to make sure that the participants have amyloid in the brain. It's an expensive lesson to have learnt the hard way. So the the imaging network, as well as procuring some very new, very expensive new kit, will also be focussed on developing new radio like.
And so that's the chemist. That's how you attach radio labels to the compounds that you're interested in, like amyloid and cow, for example. We need to have a that we have a working group whose job it is to harmonise the way that we acquire data and analyse data. And then we need an IT infrastructure to support all of this work.
The second research network is the stem cell network and am very kindly done the job for me of explaining what stem cells are and and the very exciting science that has rapidly grown in the last few years. That comes from the ability to turn skin cells into neurones and then further differentiate the neurones into the specific types of neurones that are important in different diseases.
So that means that rather than having to rely completely on animal models, we now have a way of having human disease models in a dish that can that can contribute, for example, to screening programmes for new potential new therapies. And the stem cell network is looked after by wide margins and contains all of the major stem cell centres in the UK, including the Cambridge one that Emma talked about. And within this network, all of the aspects of stem cell biology are going to be looked after.
And so this is a great example of bringing the community together, these centres, because this is new science. These centres would all otherwise be competing with each other. So by bringing them together like this, this is a really good way of showing how we can hopefully show that we can achieve more by by collaborating, cooperating. And then the third research network is the Informatics Network.
So informatics is is is one of these sort of mysterious words that's come from nowhere and now seems to be everywhere. And what it really means is both the technology to bring together data and to analyse data in in this sort of big data world in which we're living now.
And so it's really about the mathematicians. It's the pure nerds of the of the of the medical research world that need to put together the infrastructure and then the and then figure out how we use it to ask the really complicated questions. And I'll show you an example of that in a few slides time. And I've just realised I've just announced myself as a pure nerd because, ah, my name is on this slide as I'm looking after the Imaging Informatics Network for Pardeep UK.
So if I just tell you what the other ones mean. So the core portal is the means that in a couple of years time there will be a portal, a place you can go to and you can say, OK, I have particular research question what cohorts are out there, what data is out there that might be able to answer my questions so that I don't necessarily have to go and collect all the data again. So the design of the portal is happening at the Far Institute in Swansea.
Clinical informatics is another really challenging problem, which is trying to extract research, useful information from electronic patient records on this. This means using clever computer algorithms to read natural language because, of course, you might design all the forms in the world. But busy clinicians don't necessarily have time to click through every box and put the information in the right places. So actually most of the precious information is written in the case notes.
And so you have to have clever, clever algorithms that can extract useful information from case notes, and that allows you to suddenly be looking at not the 100 people that you might have been able to collect careful information for, but a thousand 10000 people whose information might not be quite so clean and neat and tidy. But you've got 100000 people now, so you're dealing with a different scale. So so that's a big part of what's happening.
I'm going to tell you more about imaging in a second. Digital health is all about the clever stuff you can. Do with your mobile phones or your Fitbit or whatever it is that you can can hang on to yourself these days and all of the information that can be collected from those. There's a genomics network led by Julie Williams in Cardiff and a brain banking network for post-mortem analysis that's led by Shatalov in Bristol. And the Informatics Network is is all about bringing information together.
And that means that there are there's more than one informatics network. And part of what we have to do is make sure that they all talk to each other. So I'll just give you an example of what we're doing in the imaging informatics world. And in the same sort of thing happens in each of the other networks. So try not to bore you to tears with this. But there are basically three challenges that we have to we have to solve.
One is that we need to we need to have the kit that is capable of dealing with 100000 images, not just storing the images themselves, but being able to process them and run all of the pipelines that we've been used to developing on tens of subjects. We now need to be able to run on hundreds of thousands of subjects and they compute resources required for that are our kind of mind blowing. But but these days, actually, that's not the biggest challenge.
You know, the computer resources are relatively easy to come by now, and we need to expand the infrastructure for the additional 10000 scans. But that's actually the easiest problem to solve. The most difficult problem to solve is, is all of the existing data. So you heard about the Whitehall cohort today. You've also heard about the 1946 birth cohort and of the thirtyish cohort that are in the UK.
About 12 of them have a decent amount of imaging data. These imaging, all of these imaging data exist where the cohorts live. So they're all in the sort of silos that all separated from each other. They were acquired on different machines with different protocols. So we have to find ways of bringing them together so that they can be combined in meaningful ways. And so creating that sort of platform is the biggest problem to solve.
And we're going to be doing that by actually, I'm not going to bore you with how we're doing this. It's 2D. You can ask me later if you want to know how we're doing it. And then the third problem is the future. And actually the future is relatively if we get if we get the present right, then we can get number two sorted. Then the future is relatively straightforward because in the future we can design
our experiments with the appropriate consent and the appropriate governance. That data can be shared sort of right from the right, from the word go. And the future is actually already happening. We've got a couple of large multicenter studies that have been recently funded that are going to use our infrastructure. So this is the thing that I decided not to tell you. It's too boring, which is that we're going to do this in a federated way.
So we can't just bring all the data together in some central place because their individual cohorts wouldn't necessarily have the permission of the consent in order to to do this. So we have to create a federated network so that each each place in the network has its own node, so that data can be the governance structure for data can be maintained. Never mind that. OK, so then the point of all of this is so that we can do the experimental medicine and the platform.
UK have the kind of fledgling experimental medicine platform. This is this, this is the kind of the last bit to get going, really. But but the way that it's being thought about is in these three overlapping and sort of research areas. And they they map onto some of the things that Emma talked about and you'll have heard about earlier today as well. So immunity is one big area, vascular risk is another big area, and then synaptic health is another big area.
And those three those three things, they sound like they're kind of slightly foreign concepts for dementia. But if you think about it, what we're talking about is whether the blood supply is okay, how the immune system is coping and how the how the brain is actually working and how the three three things interact. So there are fledgling programmes in these three areas. But the main sort of experimental medicine thing that has got going so far is a study called Deep and Frequent Genotyping.
So here's our cartoon again and the red line, which is our challenge. And so there have been a whole load of studies over over the last 20 years or so that have tried to say my measurement is the best for detecting early the early stage of dementia, whichever dementia is, that happens to be your favourite.
So you'll have people who are imaging specialists doing imaging experiments, people who are cognitive specialists, doing cognitive experiments, people who do the blood based phenotype and CSF based phenotype. And actually, it's almost never the case that you get even two of these pieces of data being collected in the same sample so they can. Preprint phenotype is the other name for it is throwing the book at people.
So the idea here is that we're going to take a select group of individuals who are going to have every every. One of each of the best measures that we know about at the moment and if you completely novel was thrown in and there are going to be collected frequently, so individuals who are participants in the computer typing study will have a cognitive battery. They'll have an MRI scan, a PET scan, a Mexican magnetoencephalography.
They have their eyes scanned. They have a gait analysis. They have lumbar punctures. Blood's taken other physical tests, and they have all of this done on day one, day to day 30, day night and day 180. So it's an incredibly intensive amount of investigation for anybody to undergo. So you can imagine that when this was this is this is what the pharma companies most want with us, US academics, academic clinicians to do. They want us to tell them what the best measure for that red bit of the curve.
And so the answer is we don't know unless we do this study. So then the funders, the funders and the patient groups say, but nobody is going to be able to to do this. It is far too much to expect anybody to do. And so the first thing that we have to do is a feasibility study. We just completed that, actually. So that was a feasibility study carried out centres where only four participants, percent were put through this rigmarole.
And I can tell you that the participants had no problem whatsoever. It's the feasibility was challenged much more by the logistics of getting by the research distance. Actually, the research assistant to this nearly killed, not the participants. Participants were fine getting there, getting this all to happen on a single day, getting everything all the rooms lined up, all the people that you need to to make this happen was far more challenging than the participants had a lovely time.
And this was people with early Alzheimer's disease. So that's not the challenging bit of this process. But anyway, it has been deemed feasible. So the full application is being written now. OK, so what I've described today is is only part of the picture. So it seems like I've hopped around all over the place, but they weren't really going to take for biomedical research to come up with ways of preventing dementia.
Looks something like this. So we need all of the basic neuroscience to translate itself into both target development and the on the really good high quality biomarkers that we need for everything.
Actually, we need the informatics and big data to pull together, which is part of the development of biomarkers, but also part of the the making best use of the resources that we already have and creating from that these cohorts of readiness who are ready to be put into early phase trials, who have the running data that we already need for and to the risk the early phase trials. And in the last two years, a lot of these boxes have been ticked in various ways.
So Emma talked about the Drug Discovery Alliance, the EPA funded, and that's made a massive difference in terms of the target development for all of dementia. And what I've just spent most of today talking about the dimensions platform links, it talks to the informatics and readiness cohorts. And then there are some other large scale European initiatives, public private partnerships run by something called the Innovative Medicines Initiative. And these things are like 60 million euros big.
These are big, lots of money and lots of in kind contributions from pharma companies and other companies. And and they're involved in various bits of this as well. And most recently, there's a new adaptive trial for for dementia has been has been funded called iPad, if you like. The sound of that, you can look that up. It's got all sorts of snazzy stuff online these days. And then the deep and frequent phenotype is feeding into the discovery of of biomarkers.
So this is a this is this is kind of a very, very ambitious set of things that we're putting together. And, of course, it's not without very significant challenges. And this is a lovely slide that I've taken from a meeting that I was at recently from the OED and the Oxford Internet Institute, where they put together some of the challenges for open science. I think and I think I think the era that we're moving into is is will be characterised by being open science.
So we've had we've had decades of people making a mating discoveries in our field, but only in their little silos, in their little pyramid structures usually won by their research group, working away at their particular problem. And I think that to make a significant impact in this disease, we have to think again about that.
And I think the open science and if you're given the grant to collect data, then your imperative is to share that data as quickly as possible so that the best can be made of it by the by the whole community and their. I'm not going to go through each of these challenges, but there are some significant there are some significant funding challenges, you can't necessarily assume that the appropriate consent has been collected from from participants and when when cohort's were originally conceived.
And you can see that this is this has been conceptualised as an iceberg. So the technology challenge being the tip of the iceberg really is not it's not that easy to to and to pull data together in terms of the technological challenge. But actually that's in some ways the least of our problems. So I'm going to leave you with with something a little bit different. Who knows what this is yet for the Large Hadron Collider.
So Large Hadron Collider was it was the is the result of something that high energy physics community, the entire high energy physics community have of achieved by all working together. So this is the and they would be at the LHC is a collaboration between 10000 scientists and engineers from over 100 countries.
It came it was it was conceived. It was the this community decided that there was no other way they were going to be able to solve their problems other than to have a Large Hadron Collider in 1984. And I think one of the nicest bits about this story is that they actually dug the tunnel pretty quickly, this 27 kilometres big ring that was dug under Geneva. They actually did that in the late 80s.
And that point they still didn't know what the experiments would be or indeed what the particle accelerator, how it would be designed. So they already, you know, they as a community, they said, we've just got to do this. We're going to dig the tunnel anyway, even though we don't know what the answers are yet. And then so they designed the LHC itself and experiments in the 90s and that it took 10 years to build the kit itself.
And then, as you know, the the discovery, the Higgs, the probable discovery, the Higgs boson was announced in 2012 and it cost about six billion pounds. So this is an enormous investment. But this was an entire scientific community coming together and and together deciding that they're not going to go for personal glory. They're not going to worry about that their next research grant.
They're going to all work together because they care so much about finding the Higgs boson together that it's worth it. It's worth their collective 30, 30 years of their careers to do it. And the other Higgs boson paper, when it came out, has 2900 authors. I Lauffer. I don't think I think that's what we're stuck with. So I'm I'm telling you this because you're the next generation. So this is it's not going to be us. It's not going to be made.
All of this it's going to be you lot when you come through creating the LHC for dementia research. That's going to make the difference. Thank you.