The Neuropathologic Landscape of Alzheimer's Disease in Hispanic Decedents (R01 AG062517) commenced in August of 2019 with the main objective to determine if the neuropathological landscape differs based on Hispanic ethnicity. We made great strides in cycle 1 working around the constraints imposed by the pandemic. We compiled demographic and clinical data, completed standardized histology, and detailed neuroanatomic distributions of Alzheimer disease (AD) and related disorders (ADRDs) pathologies in 185 persons who identified as non-Hispanic white (NHWs) and 92 as Hispanic decedents across three cohorts from University of California Davis, Columbia University, and University of California San Diego. We formed a rich uniformly characterized dataset with over 7,500 stained and evaluated slides (19 Terabytes), generating results supporting our central hypothesis that Hispanic decedents have different neuropathologic landscapes when compared to NHWs, revealing neuroanatomical and pathological specific differences. Specifically, the frontal cortex contained statistically significant higher semi-quantitative densities of neuritic plaques and neuropil threads in persons of Hispanic descent compared to age and gender-matched NHWs decedents while these pathologies had similar densities in parietal and temporal cortices. Cycle 1 accomplished additional key goals including: review articles promoting cohort diversity and deeper phenotyping of human tissues, methodologic papers on developing, and enhancing pipelines for quantitative area-specific assessments of neuropathologies and understanding the effects of pre-analytic variables on machine learning algorithms. Moreover, we expanded our impact by engaging communities, focused on those who identify as Hispanic, as we piloted novel crowdsourcing methods producing expert-like pathological analyses to ameliorate analytic bottlenecks in computational approaches based on a type of machine learning, convolutional neural networks (CNNs). In this competitive renewal application, we seek to leverage resources from cycle 1 of the application to further expand beyond non-Hispanic White centric paradigms providing insight into three main questions: 1) How does the way we evaluate pathologies alter demographic and clinical associations? 2) Can we use citizen science to engage the communities, especially those of Hispanic/Latino descent and aid in ground truth datasets for AD pathologies as well as provide outreach and educational opportunities? 3) how does Hispanic heritage (evaluating both self- reported and genetic admixture) relate to pathologic, demographic, and clinical manifestations of AD? To provide answers to these questions we will utilize an established multi-disciplinary group of experts and an established diverse cohort of 277 individuals with 92 Hispanic/Latino decedents and expand innovative methods for scalable quantitative neuropathology assessments. These synergistic yet independent questions will fill critical gaps to provide refined objective quantitative understanding of neuropathologic differences, leading to better precision medicine approaches especially for persons within the Hispanic community.

This proposal, focused on machine learning, citizen science, and heterogeneity within the Hispanic community, within the realm of Alzheimer disease (AD) will utilize established resources from cycle 1. We will fill critical gaps providing objective quantitative understanding of neuropathologic differences and engage the community to examine the neuropathologic landscape of AD leading to better precision medicine approaches especially for persons within the Hispanic community.