Premium
Neuroimaging‐derived phenotypes in the European Prevention of Alzheimer Dementia (EPAD) Cohort Study
Author(s) -
Lorenzini Luigi,
Ingala Silvia,
Wink Alle Meije,
Kuijer Joost,
Wottschel Viktor,
Sudre Carole H.,
Haller Sven,
Molinuevo Jose,
Gispert Juan Domingo,
Cash David M.,
Thomas David L.,
Vos Sjoerd,
Carrasco Ferran Prados,
Petr Jan,
Wolz Robin,
Palombit Alessandro,
Schwarz Adam J.,
Chetelat Gaël,
Payoux Pierre,
Di Perri Carol,
Pernet Cyril,
Frisoni Giovanni B.,
Fox Nick C.,
Ritchie Craig W.,
Wardlaw Joanna M.,
Waldman Adam,
Barkhof Frederik,
Mutsaerts HenkJan
Publication year - 2021
Publication title -
alzheimer's and dementia
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 6.713
H-Index - 118
eISSN - 1552-5279
pISSN - 1552-5260
DOI - 10.1002/alz.055495
Subject(s) - neuroimaging , white matter , fluid attenuated inversion recovery , diffusion mri , fractional anisotropy , magnetic resonance imaging , dementia , hyperintensity , medicine , psychology , neuroscience , radiology , pathology , disease
Background Neuroimaging biomarkers in large‐scale multimodal studies have proven effective for early diagnosis. Image‐derived phenotypes (IDPs) are summary features derived from modalities such as magnetic resonance imaging (MRI). We provide an overview of the IDPs computed from the European Prevention of Alzheimer Dementia (EPAD) cohort study, a multi‐center European study including multimodal brain MRI. Method Baseline data from the first 1500 participants of the EPAD LCS were included. The imaging protocol consists of core (3D‐T1w, 3D‐FLAIR, 2D‐T2w, 2D‐T2*), and advanced (resting‐state fMRI, SWI, diffusion MRI, and ASL) sequences. 3D‐T1w IDPs consisted of regional volumes derived from FreeSurfer v6.0 and the Learn Embeddings for Atlas Propagation (LEAP) segmentation pipelines and 3D‐FLAIR IDPs were white matter hyperintensity (WMH) volumes measured by Bayesian Model Selection (BaMoS). Mean functional connectivity of rs‐fMRI networks was extracted with FSL MELODIC and dual‐regression analyses. For diffusion‐weighted imaging, we used FSL TBSS to quantify 48 regional fractional anisotropy (FA) values according to the JHU atlas of white matter tracts. ExploreASL was used to calculate mean Cerebral blood flow (CBF) and spatial coefficient‐of‐variation (sCoV) from ASL images. To evaluate the biological relevance of IDPs, we assessed their relationship with non‐imaging phenotypes. Result A total of 358 core and 119 advanced IDPs were derived. GM volume was inversely correlated with age, with stronger effects in medio‐temporal areas (Figure 1). Regional WMH volumetrics showed mostly frontal and parietal WM lesions that were associated with aging (Figure 2). Heterogeneous changes in within‐network functional connectivity were observed with older age, with mild differences between CDR 0 and 0.5 contrasts, mostly related to the default‐mode and frontoparietal networks (Figure 3). Age was also associated inversely with skeletonised FA values. Stronger reductions appeared in amyloid positive participants (Figure 4). Conclusion We show the relevance of IDPs from the EPAD neuroimaging dataset. The observed relationships with non‐imaging phenotypes confirm their biological relevance and are in agreement with previous studies. The proposed IDP framework may constitute a valuable resource for researchers using EPAD data, promoting reproducibility of results and easily adaptable for other studies and cohorts.