Histopathologic brain age estimation via context‐aware attention‐based deep multiple instance learning

Abstract

AbstractBackgroundWhile a large list of age‐related changes have been described in the human brain, the extent to which they represent the result of common pathological reactions versus signatures of underlying brain aging remains unclear. One approach to sorting out the relative importance of the myriad of changes is deep learning, which has emerged as a powerful tool for image analysis and computer vision. However, large digitized whole slide image datasets from human aging brains have not been previously available.MethodIn this study, we leveraged a large novel collection of uniformly processed digitized human post‐mortem brain tissue sections to create a histological brain age estimation model. We further investigated the effect of cognitive impairment and exogenous stress on the model. This was accomplished by developing a context‐aware attention‐based deep multiple instance learning model on 702 human brain tissues sections (age range 50‐110 yr) from the hippocampus stained with Luxol Fast Blue counterstained with hematoxylin and eosin (LH&E) on a brain age estimation task.ResultOur model estimated brain age within a mean absolute error of 6.2 years. Learned attention weights corresponded to neuroanatomical regions known to be vulnerable to age‐related change. We found that deviations from our estimated histopathologic brain age significantly correlated with the clinical marker of cognitive status (p = 0.042). In addition, we found evidence of significantly accelerated age (p = 1.12×10‐23) in a cohort of subjects with a neuropathological diagnosis of chronic traumatic encephalopathy (CTE), a neurodegenerative disease caused by mild yet repetitive traumatic brain injury, that displays features that overlap with aging.ConclusionThese data indicate that estimated histopathologic age can be used as a reliable pathologic correlate to identify factors that contribute to accelerated or decelerated brain aging.