EPAD LCS
European Prevention of Alzheimer's Dementia Longitudinal Cohort Study (data release V.IMI)
European multi-site cohort of 2,096 adults aged 50 or over without dementia, many amyloid positive, with longitudinal 3D T1w, FLAIR, T2w and T2*w MRI, SWI, fMRI, diffusion and ASL at a subset of sites, CSF biomarkers and cognitive data. Access by request through the AD Data Initiative.
Overview
EPAD LCS is a European cohort that followed adults aged 50 or over without dementia to describe the earliest, preclinical and prodromal stages of Alzheimer's disease and to provide a pool of well-characterised people for prevention trials. It ran from 2016 to 2020 as an Innovative Medicines Initiative public-private partnership. The final data release, V.IMI (30 October 2020), holds all 2,096 participants who consented, including screen failures.
Composition
Of the participants available for analysis, about 57% are women and the mean age is about 66 years. Screening was enriched for amyloid: 37% were CSF amyloid positive at enrolment, and 358 meet criteria for preclinical Alzheimer's disease. Visits took place at baseline, 6 months and then yearly for up to three years, with MRI at every visit except the 6-month one. The release links images to cognition, CSF biomarkers, APOE genotype, demographics and vital signs. Image counts per sequence are published only for the earlier V1500.0 baseline release.
Acquisition
Baseline scans came from 21 sites with seven scanner models from Siemens, Philips and GE using a common protocol. Core sequences, run at all sites, are 3D T1w, 3D FLAIR, axial 2D T2w and axial 2D T2*-weighted gradient echo (0.9 x 0.9 x 3 mm, TE 20 ms) meant for microbleed reading. Thirteen sites with 3 T scanners added SWI with phase, resting-state fMRI, diffusion MRI and ASL. Images are offered as raw NIfTI, processed NIfTI and image-derived phenotype tables.
Annotations
A central reader (IXICO) reported radiological reads for eligibility and follow-up, including microbleed counts, lacunes and infarcts. White matter hyperintensity masks come from BaMoS segmentation of FLAIR. FreeSurfer volumes, automated quality metrics and visual quality grades of T1w images are also released. No voxel-level microbleed labels are described.
Known limitations
- Most ASL scans from older Philips software show fat-saturation artefacts and are left out of the processed data.
- Advanced sequences exist only for participants scanned at the 3 T advanced sites.
- Few participants reached the 2- and 3-year visits because the study closed in 2020.
- Ethnicity is not collected at French sites and only year of birth at Swiss sites.
Cohort
Aggregate numbers from the sources below. Bars are relative to the 2,096 subjects.
Sex
- Female 1,035 57%
- Male 793 43%
Covers 1,828 of 2,096 subjects.
Age
mean 65.7± 7.41No age bins reported.
Modality
subjects
- MRI 1,927 92%
Contrast / sequence
subjects, values can overlap
- T1-weighted 1,356 65%
- FLAIR 1,356 65%
- T2-weighted 1,356 65%
- T2*-weighted 1,356 65%
Condition
subjects, values can overlap
- Alzheimer's disease 358 17%
License and access
Our reading of the license, not legal advice. Before you use the data, read the original license and confirm that your use is allowed. We take no responsibility for how you use a dataset. Full disclaimer
A research proposal is reviewed and approved
EPAD Terms and Conditions for accessing the V.IMI Dataset
One-page terms accepted when requesting EPAD LCS data or images through the AD Data Initiative. The data serve Alzheimer's disease research, analyses should follow the research question submitted with the request and be published, data must not be passed to third parties, and publications must carry set acknowledgement elements.
Requests go through a workspace on the AD Data Initiative's AD Workbench, whose own terms of use also apply and are not covered here. Imaging requests name the sequences and formats needed.
What you can do
- Not stated
- Train ML models Not stated
- Create derived data Not stated
- Yes
What you can share
- No
- Not stated
- Share trained models Not stated
What you must do
- Yes
- Share alike No
- Yes
- Ethics approval No
- Not stated
- Release code No
- Return results No
- Delete after use No
Limits
- No
- No
Citation
Include "EPAD LCS is registered at www.clinicaltrials.gov Identifier: NCT02804789", "Data used in preparation of this article were obtained from the EPAD LCS data set V.IMI, doi:10.34688/epadlcs_v.imi_20.10.30" and the IMI (grant 115736) and Alzheimer's Association (SG 21-818099-EPAD) funding statement. MRI reference: Lorenzini L, et al. The Open-Access European Prevention of Alzheimer's Dementia (EPAD) MRI dataset and processing workflow. NeuroImage: Clinical 35, 103106 (2022). doi:10.1016/j.nicl.2022.103106
All numbers
Every number on this page, as stored in stats.csv, with its source.
| Measure | Breakdown | Value | Source |
|---|---|---|---|
| Subjects | total all participants who consented and were entered into the eCRF; includes screen failures | 2,096 | epad-dip-vimi p. 2 (Section 2) |
| Subjects | sex=female of the 1,828 non-screen-failed participants available for analysis (the table caption says 1,843) | 1,035 | saunders2022 Table 1 |
| Subjects | sex=male of the 1,828 non-screen-failed participants available for analysis (the table caption says 1,843) | 793 | saunders2022 Table 1 |
| Subjects | condition=alzheimers participants who meet criteria for preclinical AD (CSF amyloid positive at enrolment) | 358 | saunders2022 Results: Open access data |
| Subjects | modality=MR MRI scans at the baseline visit (one per participant) | 1,927 | saunders2022 Table 2 |
| Studies | total MRI sessions (the only imaging modality); sum of MRI scans at the disjoint visits baseline (1,927), 1 year (601), 2 years (249) and 3 years (6) | 2,783 | saunders2022 Table 2 |
| Subjects | contrast=T1w V1500.0 release only: eligible participants with a complete set of core sequences | 1,356 | lorenzini2022 Results: The EPAD LCS baseline imaging dataset |
| Subjects | contrast=FLAIR V1500.0 release only: eligible participants with a complete set of core sequences | 1,356 | lorenzini2022 Results: The EPAD LCS baseline imaging dataset |
| Subjects | contrast=T2w V1500.0 release only: eligible participants with a complete set of core sequences (3D T1w, 3D FLAIR, 2D T2w, 2D T2*w) | 1,356 | lorenzini2022 Results: The EPAD LCS baseline imaging dataset |
| Subjects | contrast=T2starw V1500.0 release only: eligible participants with a complete set of core sequences (3D T1w, 3D FLAIR, 2D T2w, 2D T2*w) | 1,356 | lorenzini2022 Results: The EPAD LCS baseline imaging dataset |
| Scans | contrast=swi V1500.0 release only, baseline; advanced sequences at 13 sites with 3 T scanners | 756 | lorenzini2022 Results: The EPAD LCS baseline imaging dataset |
| Scans | contrast=bold V1500.0 release only, baseline resting-state fMRI | 842 | lorenzini2022 Results: The EPAD LCS baseline imaging dataset |
| Scans | contrast=dwi V1500.0 release only, baseline | 831 | lorenzini2022 Results: The EPAD LCS baseline imaging dataset |
| Scans | contrast=asl V1500.0 release only, baseline; 62.3% showed fat-saturation artefacts and are excluded from the processed data | 858 | lorenzini2022 Results: The EPAD LCS baseline imaging dataset |
| Mean age | total of the 1,828 non-screen-failed participants | 65.7 | saunders2022 Table 1 |
| Age SD | total | 7.4 | saunders2022 Table 1 |
Sources
The keys used in the table above.
- lorenzini2022 Lorenzini L et al. 2022, The Open-Access EPAD MRI dataset and processing workflow, NeuroImage Clinical (V1500.0 release) paper
- saunders2022 Saunders S et al. 2022, The European Prevention of Alzheimer's Dementia Programme, Frontiers in Neurology paper
- epad-open-data EPAD website, Open-Source Data, Imaging, and Samples website
- epad-dip-vimi EPAD LCS Data Information Pack, V.IMI release of 30 October 2020 website