Tau protein discrete aggregates in Alzheimer's disease: neuritic plaques and tangles detection and segmentation using computational histopathology

Abstract

Tau proteins in the gray matter are widely known to be a part of Alzheimer's disease symptoms. They can aggregate in three different structures within the brain: neurites, tangles, and neuritic plaques. The morphology and the spatial disposition of these three aggregates are hypothesised to be correlated to the advancement of the disease. In order to establish a behavioural disease model related to the Tau proteins aggregates, it is necessary to develop algorithms to detect and segment them automatically. We present a 5-folded pipeline aiming to perform with clinically operational results. This pipeline is composed of a non-linear colour normalisation, a CNN-based image classifier, an Unet-based image segmentation stage, and a morphological analysis of the segmented objects. The tangle detection and segmentation algorithms improve state-of-the-art performances (75.8% and 91.1% F1score, respectively), and create a reference for neuritic plaques detection and segmentation (81.3% and 78.2% F1-score, respectively). These results constitute an initial baseline in an area where no prior results exist, as far as we know. The pipeline is complete and based on a promising state-of-the-art architecture. Therefore, we consider this study a handy baseline of an impactful extension to support new advances in Alzheimer's disease. Moreover, building a fully operational pipeline will be crucial to create a 3D histology map for a deeper understanding of clinico-pathological associations in Alzheimer's disease and the histology-based evidence of disease stratification among different sub-types.

Visual Deep Learning-Based Explanation for Neuritic Plaques Segmentation in Alzheimer’s Disease Using Weakly Annotated Whole Slide Histopathological Images

“…We normalized the patches from the 128×128 dataset using Macenko and Vahadane methods, and we selected the best fold (i.e., highest Dice score in testing for the first experiment) to train, validate and test the UNet under different input color properties. Opposite to the results reported in [7], the Dice score obtained was higher using the Macenko method (0.7248 in testing) than the Vahadane (0.7098 in testing), even in validation (0.72313 for Macenko and 0.6864 for Vahadane). For a full list of results, see supplementary material.…”

“…1. Use of biomarkers: the study in [7] uses the ALZ50 (used to discover compacted structures) biomarker, while our study uses the AT8 (majorly used in clinics, helps to discover all structures). We focus on AT8 in order to stay close to clinical protocols.…”

“…For the downsampling block, we used a leaky ReLU activation function and ReLU for the upsampling block. In both blocks, we used batch-normalization following the suggestions in [3] and [7]. Dropout was used in each convolutional block with a probability of 0.5.…”

“…An FCN model was also used in [12] to classify morphologies of tau protein aggregates in the gray and white matter regions from 37 WSIs representing multiple degenerative diseases. In [7], tau aggregate analysis was done on a dataset of 6 WSIs with a combined classification-segmentation framework which achieved an F1 score of 81.3% and 75.8% on detection and segmentation tasks, respectively. Several domains in DL-based histopathological analysis of AD tauopathy remain unexplored.…”

Structural Proteomic Profiling of Cerebrospinal Fluids to Reveal Novel Conformational Biomarkers for Alzheimer’s Disease

“…Tau aggregation and the formation of neurofibrillary tangles (NFTs) is another hallmark of AD, strongly associated with progressive neuronal loss and cognitive decline. − Patients at the dementia stage of AD have been shown to display protein abundance changes with elevated unmodified tau protein expression (T-tau), as well as the presence of tau phosphorylated at threonine 181 (P-tau) . However, the efficacy of tau-targeting strategies and clinical trials of tau inhibitors in AD have yet to be satisfactory. − Biochemical changes in the brain are reflected in the cerebrospinal fluids (CSF), making it an appealing sample source for the study of AD. − …”