Manual counting of mitotic tumor cells in tissue sections constitutes one of the strongest prognostic markers for breast cancer. This procedure, however, is time-consuming and error-prone. We developed a method to automatically detect mitotic figures in breast cancer tissue sections based on convolutional neural networks (CNNs). Application of CNNs to hematoxylin and eosin (H&E) stained histological tissue sections is hampered by: (1) noisy and expensive reference standards established by pathologists, (2) lack of generalization due to staining variation across laboratories, and (3) high computational requirements needed to process gigapixel whole-slide images (WSIs). In this paper, we present a method to train and evaluate CNNs to specifically solve these issues in the context of mitosis detection in breast cancer WSIs. First, by combining image analysis of mitotic activity in phosphohistone-H3 (PHH3) restained slides and registration, we built a reference standard for mitosis detection in entire H&E WSIs requiring minimal manual annotation effort. Second, we designed a data augmentation strategy that creates diverse and realistic H&E stain variations by modifying the hematoxylin and eosin color channels directly. Using it during training combined with network ensembling resulted in a stain invariant mitosis detector. Third, we applied knowledge distillation to reduce the computational requirements of the mitosis detection ensemble with a negligible loss of performance. The system was trained in a single-center cohort and evaluated in an independent multicenter cohort from The Cancer Genome Atlas on the three tasks of the Tumor Proliferation Assessment Challenge (TUPAC). We obtained a performance within the top-3 best methods for most of the tasks of the challenge.
The small heat shock protein family (sHsp) comprises molecular chaperones able to interact with incorrectly folded proteins. Alzheimer's disease (AD) is characterized by pathological lesions such as senile plaques (SPs), cerebral amyloid angiopathy (CAA) and neurofibrillary tangles (NFTs), predominantly consisting of the incorrectly folded proteins amyloid-beta (Abeta) and tau respectively. The aim of this study was to investigate the association of the chaperones Hsp20, HspB2, alphaB-crystallin and Hsp27 with the pathological lesions of AD brains. For this purpose, a panel of well-characterized antibodies directed against these sHsps was used in immunohistochemistry and immunoblotting. We observed extracellular expression of Hsp20, Hsp27 and HspB2 in classic SPs, and Hsp20 expression in diffuse SPs. In addition, extracellular expression of HspB2 was observed in CAA. Both Hsp27 and alphaB-crystallin were also observed in astrocytes associated with both SPs and CAA. Furthermore, none of the sHsps were observed in NFTs in AD brains. We conclude that specific sHsp species may be involved in the pathogenesis of either SPs or CAA in AD.
Cerebral vascular amyloid beta-protein (Abeta) deposition, also known as cerebral amyloid angiopathy, is a common pathological feature of Alzheimer's disease. Additionally, several familial forms of cerebral amyloid angiopathy exist including the Dutch (E22Q) and Iowa (D23N) mutations of Abeta. Increasing evidence has associated cerebral microvascular amyloid deposition with neuroinflammation and dementia in these disorders. We recently established a transgenic mouse model (Tg-SwDI) that expresses human vasculotropic Dutch/Iowa mutant amyloid beta-protein precursor in brain. Tg-SwDI mice were shown to develop early-onset deposition of Abeta exhibiting high association with cerebral microvessels. Here we present quantitative temporal analysis showing robust and progressive accumulation of cerebral microvascular fibrillar Abeta accompanied by decreased cerebral vascular densities, the presence of apoptotic cerebral vascular cells, and cerebral vascular cell loss in Tg-SwDI mice. Abundant neuroinflammatory reactive astrocytes and activated microglia strongly associated with the cerebral microvascular fibrillar Abeta deposits. In addition, Tg-SwDI mouse brain exhibited elevated levels of the inflammatory cytokines interleukin-1beta and -6. Together, these studies identify the Tg-SwDI mouse as a unique model to investigate selective accumulation of cerebral microvascular amyloid and the associated neuroinflammation.
We report on four families affected by a clinical presentation of complex hereditary spastic paraplegia (HSP) due to recessive mutations in DDHD2, encoding one of the three mammalian intracellular phospholipases A(1) (iPLA(1)). The core phenotype of this HSP syndrome consists of very early-onset (<2 years) spastic paraplegia, intellectual disability, and a specific pattern of brain abnormalities on cerebral imaging. An essential role for DDHD2 in the human CNS, and perhaps more specifically in synaptic functioning, is supported by a reduced number of active zones at synaptic terminals in Ddhd-knockdown Drosophila models. All identified mutations affect the protein's DDHD domain, which is vital for its phospholipase activity. In line with the function of DDHD2 in lipid metabolism and its role in the CNS, an abnormal lipid peak indicating accumulation of lipids was detected with cerebral magnetic resonance spectroscopy, which provides an applicable diagnostic biomarker that can distinguish the DDHD2 phenotype from other complex HSP phenotypes. We show that mutations in DDHD2 cause a specific complex HSP subtype (SPG54), thereby linking a member of the PLA(1) family to human neurologic disease.
Alzheimer's disease (AD) is characterized by pathological lesions, such as senile plaques (SPs) and cerebral amyloid angiopathy (CAA), both predominantly consisting of a proteolytic cleavage product of the amyloid-beta precursor protein (APP), the amyloid-beta peptide (Abeta). CAA is also the major pathological lesion in hereditary cerebral hemorrhage with amyloidosis of the Dutch type (HCHWA-D), caused by a mutation in the gene coding for the Abeta peptide. Several members of the small heat shock protein (sHsp) family, such as alphaB-crystallin, Hsp27, Hsp20 and HspB2, are associated with the pathological lesions of AD, and the direct interaction between sHsps and Abeta has been demonstrated in vitro. HspB8, also named Hsp22 of H11, is a recently discovered member of the sHsp family, which has chaperone activity and is observed in neuronal tissue. Furthermore, HspB8 affects protein aggregation, which has been shown by its ability to prevent formation of mutant huntingtin aggregates. The aim of this study was to investigate whether HspB8 is associated with the pathological lesions of AD and HCHWA-D and whether there are effects of HspB8 on Abeta aggregation and Abeta-mediated cytotoxicity. We observed the expression of HspB8 in classic SPs in AD brains. In addition, HspB8 was found in CAA in HCHWA-D brains, but not in AD brains. Direct interaction of HspB8 with Abeta(1-42), Abeta(1-40) and Abeta(1-40) with the Dutch mutation was demonstrated by surface plasmon resonance. Furthermore, co-incubation of HspB8 with D-Abeta(1-40) resulted in the complete inhibition of D-Abeta(1-40)-mediated death of cerebrovascular cells, likely mediated by a reduction in both the beta-sheet formation of D-Abeta(1-40) and its accumulation at the cell surface. In contrast, however, with Abeta(1-42), HspB8 neither affected beta-sheet formation nor Abeta-mediated cell death. We conclude that HspB8 might play an important role in regulating Abeta aggregation and, therefore, the development of classic SPs in AD and CAA in HCHWA-D.
Inefficient clearance of A beta, caused by impaired blood-brain barrier crossing into the circulation, seems to be a major cause of A beta accumulation in the brain of late-onset Alzheimer's disease patients and hereditary cerebral hemorrhage with amyloidosis Dutch type. We observed association of receptor for advanced glycation end products, CD36, and low-density lipoprotein receptor (LDLR) with cerebral amyloid angiopathy in both Alzheimer's disease and hereditary cerebral hemorrhage with amyloidosis Dutch type brains and increased low-density lipoprotein receptor-related protein-1 (LRP-1) expression by perivascular cells in cerebral amyloid angiopathy. We investigated if these A beta receptors are involved in A beta internalization and in A beta-mediated cell death of human cerebrovascular cells and astrocytes. Expression of both the LRP-1 and LDLR by human brain pericytes and leptomeningeal smooth muscle cells, but not by astrocytes, increased on incubation with A beta. Receptor-associated protein specifically inhibited A beta-mediated up-regulation of LRP-1, but not of LDLR, and receptor-associated protein also decreased A beta internalization and A beta-mediated cell death. We conclude that especially LRP-1 and, to a minor extent, LDLR are involved in A beta internalization by and A beta-mediated cell death of cerebral perivascular cells. Although perivascular cells may adapt their A beta internalization capacity to the levels of A beta present, saturated LRP-1/LDLR-mediated uptake of A beta results in degeneration of perivascular cells.
Variations in the color and intensity of hematoxylin and eosin (H&E) stained histological slides can potentially hamper the effectiveness of quantitative image analysis. This paper presents a fully automated algorithm for standardization of whole-slide histopathological images to reduce the effect of these variations. The proposed algorithm, called whole-slide image color standardizer (WSICS), utilizes color and spatial information to classify the image pixels into different stain components. The chromatic and density distributions for each of the stain components in the hue-saturation-density color model are aligned to match the corresponding distributions from a template whole-slide image (WSI). The performance of the WSICS algorithm was evaluated on two datasets. The first originated from 125 H&E stained WSIs of lymph nodes, sampled from 3 patients, and stained in 5 different laboratories on different days of the week. The second comprised 30 H&E stained WSIs of rat liver sections. The result of qualitative and quantitative evaluations using the first dataset demonstrate that the WSICS algorithm outperforms competing methods in terms of achieving color constancy. The WSICS algorithm consistently yields the smallest standard deviation and coefficient of variation of the normalized median intensity measure. Using the second dataset, we evaluated the impact of our algorithm on the performance of an already published necrosis quantification system. The performance of this system was significantly improved by utilizing the WSICS algorithm. The results of the empirical evaluations collectively demonstrate the potential contribution of the proposed standardization algorithm to improved diagnostic accuracy and consistency in computer-aided diagnosis for histopathology data.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.