Authors: Carles Javierre Petit, Ashish A. Tamhane, Arnold M. Evia, Nazanin Makkinejad, Gady Agam, David A. Bennett, Julie A. Schneider, Konstantinos Arfanakis
Published: 2020-12-07
DOI: 10.1002/alz.039938
Source: Full article
AbstractBackgroundEnlarged perivascular spaces (EPVS) have been linked to increased risk of stroke, lower cognitive function, and vascular dementia. However, neuropathologic correlates of EPVS, and their contributions to cognition, are unclear. In this work, a deep‐learning algorithm for automatic segmentation of EPVS was developed, and allowed full quantification of EPVS in the whole brain and regionally. Then, neuropathologic correlates of EPVS and contributions of EPVS to cognition were analyzed within each region in a large community‐based cohort of older adults.MethodThis work included 287 participants of two longitudinal studies, Rush Memory and Aging Project, and Religious Orders Study (Figure 1). Ex‐vivo T2‐weighted images, and detailed neuropathologic examination by a board‐certified neuropathologist (blinded to all data) were obtained for one brain hemisphere from each participant. A total of 10 neuropathologies were assessed (Figure 2).The data from 10 participants with varying EPVS severity and manually segmented EPVS were used to train a deep learning model (Figure 3) to automatically segment EPVS in the ex‐vivo MRI data of all participants.Relative EPVS volume (EPVSrel), meaning EPVS volume normalized by white matter volume, and total number of EPVS (EPVSnum) was calculated for each participant. Similar measurements were made per lobe and basal ganglia (e.g. EPVSnum‐parietal). All measures were square root transformed to reduce skewness.ResultThe deep learning model achieved a mean Dice Similarity Coefficient of 0.655 ± 0.058 (comparison in Figure 4). Whole brain relative EPVS volume and number were associated with gross infarcts (Figure 5). Regional EPVS were associated with gross infarcts in most of the brain, and with cerebral amyloid angiopathy mainly in the occipital and temporal lobes (Figure 5). Relative EPVS volume in the occipital lobe was negatively correlated with semantic memory (‐0.443, p = 0.03) and perceptual speed (‐0.311, p = 0.048) above and beyond what was explained by neuropathologies and demographics.ConclusionThis investigation provides robust evidence on the neuropathologic correlates of EPVS and the contributions of EPVS on cognition in a large community‐based cohort of older adults. Fully quantitative assessment of EPVS was facilitated by automatic EPVS segmentation using deep learning. EPVS were shown to have associations with gross infarcts and cerebral amyloid angiopathy, and independent contributions on cognition.