NEW YORK – A public-private consortium led by Norway's Stavanger University Hospital is working to develop tools for the early detection of Alzheimer's disease or dementia.
Called Predictom, the consortium comprises 30 academic and industry partners including Novo Nordisk, GE Healthcare, Siemens Healthineers, the University of Geneva, Kings College London, and Alzheimer Europe.
The effort launched in November 2023 and runs through October 2027 and is funded by €21 million ($22.7 million), which includes €8 million from the EU, €9 million from private industry, and €4 million from UK Research and Innovation.
The project, which will be run at seven research centers across Europe, consists of two stages. In the first stage, the researchers plan to enroll 4,000 individuals, who will collect samples including stool, saliva, and blood either at home or at their general practitioner's office. They will also undergo cognitive, hearing, and eye-tracking testing.
Using the data collected in this first stage, the researchers will develop risk models for Alzheimer's or dementia that they will then use to select 650 individuals from the original 4,000 — 450 deemed high risk and 200 deemed low risk — to go through a standard diagnostic work-up for Alzheimer's disease including confirmatory testing via either cerebrospinal fluid testing or PET imaging.
Dag Aarsland, professor of old age psychiatry at King's College London and research lead at Stavanger University Hospital, said the project is in part driven by the need to prepare for new Alzheimer's therapies expected to become available in coming years. He noted that the European Medicines Agency (EMA) is currently considering marketing authorization for Eisai's Alzheimer's drug Leqembi (lecanemab) and that a decision is expected before the summer. The drug received approval from the US Food and Drug Administration last year.
The availability of Alzheimer's drugs will likely lead to a "huge demand" for diagnostic testing, Aarsland said, noting that existing diagnostic pathways will be unable to meet that demand.
"Today, any diagnostic confirmation needs to be made in a hospital either with CSF or molecular PET, so there will be a huge clog in the system," he said. "So we need to empower the primary care system to do a meaningful triage of all of these people. We hope that with the home-based and [general practitioner]-based testing, we will be able to select a much more precise cohort for final confirmation at the hospital."
While plasma amyloid-beta 42/40 (Aβ42/Aβ40) and plasma phosphorylated-tau 217 (p-tau 217) are perhaps the leading markers for triaging patients at risk of Alzheimer's disease, the Predictom project looks to cast a wide net, exploring genetic and epigenetic markers as well as participants' microbiome profiles. Such markers are not specific for Alzheimer's itself, Aarsland noted, but may provide insight into processes like oxidative stress and inflammation that are linked to brain health.
The researchers plan to integrate the different data sources using machine learning and other data-analysis approaches to develop diagnostic models for Alzheimer's and dementia.
Aarsland said that beyond improving the diagnostic process, the Predictom initiative hopes to inform efforts to develop more personalized prevention programs for individuals at high risk for the disease.
Ideally, "we can say, this person should do more of this and change their lifestyle in this direction because of the data we have, while this other person should do something else," he said. "Much more differentiated guidelines are where we would like to go."
As one part of the initiative, University of Gothenburg researchers will expand on a previous pilot project looking at whether blood-based Alzheimer's markers can be measured in self-collected finger-prick samples.
Led by Nicholas Ashton, associate professor of neurochemistry at the University of Gothenburg, the original project collected dried blood spots of both venous and finger-prick blood from 77 patients of the ACE Alzheimer's Center in Barcelona. The samples were shipped to Gothenburg at ambient temperatures, and the researchers measured the Alzheimer's biomarkers neurofilament light (NfL), glial fibrillary acidic protein (GFAP), and phosphorylated tau (p-tau 181 and p-tau 217) in both sample types. They found that in the finger-prick samples NfL, GFAP, and p-tau 217 correlated well with standard samples, while p-tau 181 did not.
Since the pilot, the researchers "have been able to make some really good progress both in terms of detectability and also some clinical validation," Ashton said. "We're now collecting from four different [memory clinics] around Europe where, when someone comes in because they have an objective concern, they are asked to participate by giving a finger-prick test along with standard blood draws and CSF or PET."
The researchers have collected finger-prick data on around 400 people through this effort, Ashton said, adding that "it looks very promising in terms of being able to identify who is positive and who is negative" for Alzheimer's pathology.
He said the finger-prick data appears most useful for identifying individuals at high risk for Alzheimer's pathology. "It's more of a rule-in than a rule-out biomarker," he noted.
Ashton and his colleagues now plan to expand their sampling to the 4,000 participants in the initial stage of the Predictom initiative. The scientists have also narrowed their focus to p-tau 217, which Ashton said research in standard blood samples has established as a leading marker for the presence of Alzheimer's disease pathology.
The Gothenburg team is also using new blood collection cards from Swedish sample collection firm Capitainer that separates plasma on the card for testing. Ashton said the new cards are scheduled for commercial release this spring.
Additionally, the researchers are changing the assay they are using for their measurements. To date, they have used a p-tau 217 test from diagnostics firm Alzpath that runs on Quanterix's Simoa immunoassay platform. They are moving to a test run on Alamar Biosciences' NULISA (NUcleic acid-Linked Immuno-Sandwich Assay) system.
"We hope this new assay will give us a little bit more sensitivity," Ashton said, suggesting that this could potentially boost the finger-prick assay's performance.
He also noted that the NULISA platform allows for greater multiplexing than the Simoa technology, letting the researchers collect measurements on a number of neurology markers. Last month, Alamar launched its Argo HT proteomics system and its NULISAseq Inflammation Panel 250, an assay panel for the system that measures 250 proteins linked to immune response. The company has not announced the launch of a neurology-focused panel, but Ashton said the panel he plans to use in his work will allow him to measure 120 proteins in a single blood spot.
The Predictom project is focused on p-tau 217, but the NULISA system provides Ashton and his colleagues a chance to reinterrogate the samples they have collected, he said. "We have no data for us to say that any marker is going to be better than 217, so we won't use it for the purposes of Predictom, but it gives us an opportunity to look at our data retrospectively and say, what could we have done differently? Are there other biomarkers that could have helped us?"
Yuling Luo, Alamar's founder, chairman, and CEO, said that the company, which has been sued by Olink for patent infringement, did not initially plan to invest significantly in neurology but that in discussions with customers observed strong demand for higher sensitivity assays, particularly given the interest in detecting neurology markers in blood.
Anna Berdine, the company's senior VP of marketing, noted that Ashton and his Gothenburg colleague Henrik Zetterberg, a prominent researcher in the neurology biomarker space, were both early-access customers of the Argo HT platform.
A Quanterix spokesperson said that the data the company has seen to date regarding NULISA versus Simoa "have been ambiguous and do not demonstrate superior analytical sensitivity." The spokesperson added that Quanterix believes its technology offers advantages in terms of performance and fast turnaround time that make it an attractive option, particularly for clinical applications.