Deep learning assisted quantitative assessment of histopathological markers of Alzheimer’s disease and cerebral amyloid angiopathy

Abstract

Traditionally, analysis of neuropathological markers in neurodegenerative diseases has relied on visual assessments of stained sections. Resulting semiquantitative scores often vary between individual raters and research centers, limiting statistical approaches. To overcome these issues, we have developed six deep learning-based models, that identify some of the most characteristic markers of Alzheimer’s disease (AD) and cerebral amyloid angiopathy (CAA). The deep learning-based models are trained to differentially detect parenchymal amyloid β (Aβ)-plaques, vascular Aβ-deposition, iron and calcium deposition, reactive astrocytes, microglia, as well as fibrin extravasation. The models were trained on digitized histopathological slides from brains of patients with AD and CAA, using a workflow that allows neuropathology experts to train convolutional neural networks (CNNs) on a cloud-based graphical interface. Validation of all models indicated a very good to excellent performance compared to three independent expert human raters. Furthermore, the Aβ and iron models were consistent with previously acquired semiquantitative scores in the same dataset and allowed the use of more complex statistical approaches. For example, linear mixed effects models could be used to confirm the previously described relationship between leptomeningeal CAA severity and cortical iron accumulation. A similar approach enabled us to explore the association between neuroinflammation and disparate Aβ pathologies. The presented workflow is easy for researchers with pathological expertise to implement and is customizable for additional histopathological markers. The implementation of deep learning-assisted analyses of histopathological slides is likely to promote standardization of the assessment of neuropathological markers across research centers, which will allow specific pathophysiological questions in neurodegenerative disease to be addressed in a harmonized way and on a larger scale.

Perivascular space dilation is associated with vascular amyloid-β accumulation in the overlying cortex

“…The WM was manually segmented Quantitative measures of the sections stained for Aβ and Fibrin(ogen) were obtained using the online platform Aiforia  . Details of the method have been described elsewhere [46]. Briefly, sections were uploaded on the Aiforia  platform and annotations (ground truth) were manually drawn on 10% of the dataset, in order to train convolutional neural networks (CNNs).…”

“…Two AI-models and the following derived measures were adopted for this study: 1) Aβ model: ). After quality control, few sections had to be excluded from the analyses, due to suboptimal performance of the AI-models (four for the measurements derived from the Aβ model, three for the ones derived from the fibrin model) (See [46] for details). The mean integrated density of the WM was calculated using ImageJ on the LHE-stained sections after subtracting the area occupied by the EPVS as a measure of myelin rarefaction.…”

Stereology neuron counts correlate with deep learning estimates in the human hippocampal subregions

“…Subjectivity, differences in training, recognition biases, and fatigue might potentially influence results 34 , 35 . Deep learning has been applied in various previous works in humans and animal models 26 , 27 , 30 , 31 , 53 . It was identified as an effective and reliable method for cell counting by negating interrater variability and enabling reproducibility 34 , 35 .…”

“…Two stereology methods find wide application, optical fractionator 24 and the NvVref method 21 , 25 . Deep learning techniques have offered novel ways to quantify structure and have been successfully utilized in various fields, such as the diagnosis and classification of cancer 26 , 27 , the quantification of myofibers 28 , 29 , and the identification of histopathological markers in neurodegenerative diseases 30 , 31 . It has also been applied to segment fluorescently-tagged neurons in the human and rat brain 32 , 33 .…”

Learning fast and fine-grained detection of amyloid neuropathologies from coarse-grained expert labels

“…Deep learning can augment neuropathologist expertise 10 and combine multiple expert annotations into a robust and automated labeler 7 . For more localized tasks like object detection 12 and semantic segmentation 13 , deep learning has also provided accurate and automated means of quantifying Aβ and tau neuropathologies 14,15,16 . However, such studies require significant human expert labor to create high-quality training datasets in the form of manually drawn bounding boxes or segmentations and categorical labels.…”