Screening for Alzheimer's Disease in the Community Using an AI-Driven Screening Platform: Design of the PREDICTOM Study

Early detection of Alzheimer's disease has long been constrained by the limits of traditional clinical settings — expensive imaging, invasive procedures, and in-person assessments that are difficult to scale. Recent advances in physiological, imaging, and digital biomarkers, combined with the approval of new disease-modifying therapies, are creating a real opportunity to shift the first steps of Alzheimer's diagnosis into the home. PREDICTOM is a large-scale European study designed to test whether that shift is possible, using an AI-driven platform that combines digital cognitive testing, blood-based biomarkers, and other novel measures. BrainCheck Assess is one of the cognitive assessment tools embedded in the platform, selected as part of the home-based screening protocol across all study sites.

Research Objective

The aim of PREDICTOM is to develop and test the accuracy of an AI-driven screening platform for the risk assessment and early detection of Alzheimer's disease. The study is designed to evaluate whether a comprehensive set of established biomarkers, collected at home or in a community care setting, can be integrated into a well-defined clinical pathway that extends early detection beyond the walls of a clinic or practice.

Study Design and Methods

PREDICTOM is an observational, prospective, multi-center cohort study running from November 2023 to October 2027, with clinical sites in Norway, the United Kingdom, France, Germany, Switzerland, Spain, and Belgium. The study uses a cloud-based platform that stores a digitalized journey for each participant and hosts a collection of AI algorithms for risk assessment, early diagnosis, and prognosis.

The study is structured across three levels.

• Level 1 enrolls 4,000 adults aged 50 and older at increased risk of Alzheimer's disease and consists of fully home-based assessments, including BrainCheck Assess, questionnaires, online hearing screening, eye-tracking, and at-home biofluid collection via finger-stick blood and saliva.
• Based on an AI-driven risk stratification algorithm, 615 participants identified as high or low risk are selected for Level 2 in-clinic assessments, which include more advanced cognitive testing, EEG, MRI, venous blood, microbiome sampling, and additional hearing and eye-tracking measures.
• Level 3 involves confirmatory diagnostic evaluation using cerebrospinal fluid analysis or amyloid PET to confirm or rule out Alzheimer's pathology.

BrainCheck Assess is administered at Level 1 and evaluates attention, executive function, and memory through Trail Making A and B, Digit-Symbol Substitution, Stroop, and immediate and delayed word recognition. The battery can be completed in 10 to 15 minutes on any browser-enabled device, making it well-suited for the study's fully remote, home-based design.

What the Study Will Tell Us

PREDICTOM is designed to answer several important questions. Can AI-driven algorithms accurately identify individuals at elevated risk for Alzheimer's disease using biomarkers collected outside of a clinical setting? How accurate, feasible, and safe are these home-based screening approaches in a real-world population? And can this model be integrated into a scalable, cost-effective clinical pathway that health systems can realistically adopt?

The study's conclusions will inform future clinical guidelines for early Alzheimer's diagnosis and help determine which combination of home-based biomarkers offers the greatest predictive value when confirmed against gold-standard diagnostics like CSF and amyloid PET.

Why This Study Matters

For BrainCheck, inclusion in PREDICTOM is significant on multiple levels. The platform was selected alongside established tools like MRI, EEG, and blood-based biomarkers as part of a rigorously designed, internationally funded diagnostic protocol, validating BrainCheck's role in AI-driven risk stratification for early dementia detection. It also establishes BrainCheck as part of EU-based data generation and decentralized clinical research, expanding the evidence base beyond the U.S. healthcare context.

More broadly, PREDICTOM reflects a critical shift in how the field is approaching Alzheimer's detection, moving away from reactive, symptom-driven diagnosis and toward predictive, pre-symptomatic models powered by digital tools. With disease-modifying therapies now approved for early-stage Alzheimer's, identifying at-risk individuals earlier and more efficiently has never been more important. PREDICTOM's tiered model, in which accessible home-based tools like BrainCheck filter a large population before more resource-intensive diagnostics are applied, offers a blueprint for how health systems can meet that demand at scale. We look forward to sharing results as the study progresses toward its 2027 completion.

Citation

Brem AK, Khan Z, Radermacher J, Georgiadis K, Lazarou I, Grammatikopoulou M, Pickering E, Mitterreiter J, Aakre JA, Ashton NJ, Baquero M, Beser-Robles M, Braboszcz C, Brandt S, Brown J, Cacciamani F, Campill S, Collins C, Deshpande P, Diaz A, Durrleman S, Engelborghs S, Ferré-González L, Frisoni GB, Gjestsen MT, Gove D, Honigberg L, Huang B, Hudak A, Kaushik S, Letoha T, Marquardt G, Mendes AJ, Müllenborn M, Paletta L, Pedrosa de Barros N, Pszeida M, Vik-Mo AO, Rostamipour H, Perneczky R, Rauchmann BS, Russegger S, Schirmer T, Shadmaan A, Solana AB, Soria-Frisch A, Tegethoff P, Ribbens A, De Witte S, van der Giezen M, Nikolopoulos S, Corbett A, Fröhlich H, Aarsland D. Screening for Alzheimer's disease in the community using an AI-driven screening platform: design of the PREDICTOM study. The Journal of Prevention of Alzheimer's Disease. 2026;13(5):100545. doi:10.1016/j.tjpad.2026.100545.

Read the full publication here: https://www.sciencedirect.com/science/article/pii/S2274580726000695