Dataset Information

Deep learning for Alzheimer's disease: Mapping large-scale histological tau protein for neuroimaging biomarker validation.

ABSTRACT: Abnormal tau inclusions are hallmarks of Alzheimer's disease and predictors of clinical decline. Several tau PET tracers are available for neurodegenerative disease research, opening avenues for molecular diagnosis in vivo. However, few have been approved for clinical use. Understanding the neurobiological basis of PET signal validation remains problematic because it requires a large-scale, voxel-to-voxel correlation between PET and (immuno) histological signals. Large dimensionality of whole human brains, tissue deformation impacting co-registration, and computing requirements to process terabytes of information preclude proper validation. We developed a computational pipeline to identify and segment particles of interest in billion-pixel digital pathology images to generate quantitative, 3D density maps. The proposed convolutional neural network for immunohistochemistry samples, IHCNet, is at the pipeline's core. We have successfully processed and immunostained over 500 slides from two whole human brains with three phospho-tau antibodies (AT100, AT8, and MC1), spanning several terabytes of images. Our artificial neural network estimated tau inclusion from brain images, which performs with ROC AUC of 0.87, 0.85, and 0.91 for AT100, AT8, and MC1, respectively. Introspection studies further assessed the ability of our trained model to learn tau-related features. We present an end-to-end pipeline to create terabytes-large 3D tau inclusion density maps co-registered to MRI as a means to facilitate validation of PET tracers.

SUBMITTER: Ushizima D

PROVIDER: S-EPMC8983026 | biostudies-literature | 2022 Mar

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Deep learning for Alzheimer's disease: Mapping large-scale histological tau protein for neuroimaging biomarker validation.

Ushizima Daniela D Chen Yuheng Y Alegro Maryana M Ovando Dulce D Eser Rana R Lee WingHung W Poon Kinson K Shankar Anubhav A Kantamneni Namrata N Satrawada Shruti S Junior Edson Amaro EA Heinsen Helmut H Tosun Duygu D Grinberg Lea T LT

NeuroImage 20211220

Abnormal tau inclusions are hallmarks of Alzheimer's disease and predictors of clinical decline. Several tau PET tracers are available for neurodegenerative disease research, opening avenues for molecular diagnosis in vivo. However, few have been approved for clinical use. Understanding the neurobiological basis of PET signal validation remains problematic because it requires a large-scale, voxel-to-voxel correlation between PET and (immuno) histological signals. Large dimensionality of whole hu ...[more]

PMID: 34933123

Similar Datasets

Project description:BackgroundNovel phosphorylated-tau (p-tau) blood biomarkers (e.g., p-tau181, p-tau217 or p-tau231), are highly specific for Alzheimer's disease (AD), and can track amyloid-β (Aβ) and tau pathology. However, because these biomarkers are strongly associated with the emergence of Aβ pathology, it is difficult to determine the contribution of insoluble tau aggregates to the plasma p-tau signal in blood. Therefore, there remains a need for a biomarker capable of specifically tracking insoluble tau accumulation in brain.MethodsNTA is a novel ultrasensitive assay targeting N-terminal containing tau fragments (NTA-tau) in cerebrospinal fluid (CSF) and plasma, which is elevated in AD. Using two well-characterized research cohorts (BioFINDER-2, n = 1,294, and BioFINDER-1, n = 932), we investigated the association between plasma NTA-tau levels and disease progression in AD, including tau accumulation, brain atrophy and cognitive decline.ResultsWe demonstrate that plasma NTA-tau increases across the AD continuum¸ especially during late stages, and displays a moderate-to-strong association with tau-PET (β = 0.54, p < 0.001) in Aβ-positive participants, while weak with Aβ-PET (β = 0.28, p < 0.001). Unlike plasma p-tau181, GFAP, NfL and t-tau, tau pathology determined with tau-PET is the most prominent contributor to NTA-tau variance (52.5% of total R2), while having very low contribution from Aβ pathology measured with CSF Aβ42/40 (4.3%). High baseline NTA-tau levels are predictive of tau-PET accumulation (R2 = 0.27), steeper atrophy (R2 ≥ 0.18) and steeper cognitive decline (R2 ≥ 0.27) in participants within the AD continuum. Plasma NTA-tau levels significantly increase over time in Aβ positive cognitively unimpaired (βstd = 0.16) and impaired (βstd = 0.18) at baseline compared to their Aβ negative counterparts. Finally, longitudinal increases in plasma NTA-tau levels were associated with steeper longitudinal decreases in cortical thickness (R2 = 0.21) and cognition (R2 = 0.20).ConclusionOur results indicate that plasma NTA-tau levels increase across the AD continuum, especially during mid-to-late AD stages, and it is closely associated with in vivo tau tangle deposition in AD and its downstream effects. Moreover, this novel biomarker has potential as a cost-effective and easily accessible tool for monitoring disease progression and cognitive decline in clinical settings, and as an outcome measure in clinical trials which also need to assess the downstream effects of successful Aβ removal.

Dataset Information

Deep learning for Alzheimer's disease: Mapping large-scale histological tau protein for neuroimaging biomarker validation.

Publications

Deep learning for Alzheimer's disease: Mapping large-scale histological tau protein for neuroimaging biomarker validation.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets