RESUMO
Artificial intelligence (AI) for the purpose of this review is an umbrella term for technologies emulating a nephropathologist's ability to extract information on diagnosis, prognosis, and therapy responsiveness from native or transplant kidney biopsies. Although AI can be used to analyze a wide variety of biopsy-related data, this review focuses on whole slide images traditionally used in nephropathology. AI applications in nephropathology have recently become available through several advancing technologies, including (i) widespread introduction of glass slide scanners, (ii) data servers in pathology departments worldwide, and (iii) through greatly improved computer hardware to enable AI training. In this review, we explain how AI can enhance the reproducibility of nephropathology results for certain parameters in the context of precision medicine using advanced architectures, such as convolutional neural networks, that are currently the state of the art in machine learning software for this task. Because AI applications in nephropathology are still in their infancy, we show the power and potential of AI applications mostly in the example of oncopathology. Moreover, we discuss the technological obstacles as well as the current stakeholder and regulatory concerns about developing AI applications in nephropathology from the perspective of nephropathologists and the wider nephrology community. We expect the gradual introduction of these technologies into routine diagnostics and research for selective tasks, suggesting that this technology will enhance the performance of nephropathologists rather than making them redundant.
Assuntos
Inteligência Artificial , Aprendizado de Máquina , Redes Neurais de Computação , Reprodutibilidade dos Testes , SoftwareRESUMO
Conjugation of a therapeutic agent to a polymer for enhanced delivery into target cells followed by its intracellular triggered release has proved to be an effective drug delivery approach. This approach is applied to the delivery of the immune-stimulatory unmethylated cytosine-phosphate-guanine (CpG) oligonucleotide for an anti-tumour immune response after intratumoral administration. On average four CpG-1668 molecules were covalently linked to a 40-kDa amino-functionalised dextran polymer via either a non-reversible (CpG-dextran) or an intracellular redox-responsive disulfide linkage (CpG-SS-dextran). Dynamic light scattering analysis showed that both conjugates had a similar particle size and surface charge of 17 nm and -10 mV, respectively. Agarose gel electrophoresis analysis showed that CpG-SS-dextran was stable in the extracellular low glutathione (GSH) concentration range (i.e. 10-20 µM) and was cleaved at the higher intracellular GSH concentration (5 mM), while CpG-dextran was stable in both GSH concentrations. Uptake and activation assays on bone-marrow-derived dendritic cells showed no significant difference between free CpG, CpG-dextran and CpG-SS-dextran. In a mouse subcutaneous colorectal tumour model the CpG-SS-dextran showed a statistically significantly greater inhibition of tumour growth (p < 0.03) and prolonged survival (p < 0.001) compared to CpG-dextran or free CpG. These results demonstrate that the redox-triggered intracellular release of CpG from a dextran polymer carrier has promise for intratumoral therapeutic vaccination against cancer.
RESUMO
Reliable large-scale cell detection and segmentation is the fundamental first step to understanding biological processes in the brain. The ability to phenotype cells at scale can accelerate preclinical drug evaluation and system-level brain histology studies. The impressive advances in deep learning offer a practical solution to cell image detection and segmentation. Unfortunately, categorizing cells and delineating their boundaries for training deep networks is an expensive process that requires skilled biologists. This paper presents a novel self-supervised Dual-Loss Adaptive Masked Autoencoder (DAMA) for learning rich features from multiplexed immunofluorescence brain images. DAMA's objective function minimizes the conditional entropy in pixel-level reconstruction and feature-level regression. Unlike existing self-supervised learning methods based on a random image masking strategy, DAMA employs a novel adaptive mask sampling strategy to maximize mutual information and effectively learn brain cell data. To the best of our knowledge, this is the first effort to develop a self-supervised learning method for multiplexed immunofluorescence brain images. Our extensive experiments demonstrate that DAMA features enable superior cell detection, segmentation, and classification performance without requiring many annotations. In addition, to examine the generalizability of DAMA, we also experimented on TissueNet, a multiplexed imaging dataset comprised of two-channel fluorescence images from six distinct tissue types, captured using six different imaging platforms. Our code is publicly available at https://github.com/hula-ai/DAMA.
Assuntos
Encéfalo , Encéfalo/diagnóstico por imagem , Processamento de Imagem Assistida por Computador/métodos , Aprendizado de Máquina Supervisionado , Humanos , Aprendizado Profundo , Animais , Algoritmos , Neuroimagem/métodosRESUMO
The effect of dextran molecular weight on the in vitro physicochemical and immune properties of cytosine-phosphate-guanine (CpG) oligodeoxynucleotide-amino-dextran conjugates is investigated. CpG-1668 was conjugated at the 3'-end to amino-dextran of differing molecular weight (20, 40, 70 or 110-kDa) via a stable bis-aryl hydrazone linkage. Conjugate formation was confirmed by agarose gel electrophoresis and dynamic light scattering measured the size and surface charge of conjugates. Uptake and immune-stimulatory activity of CpG-dextran by antigen-presenting cells was evaluated by flow cytometry and confocal microscopy. Degradation by DNase I was monitored by loss of the fluorescent signal from labelled CpG and changes in size and zeta potential. Hydrazone bond formation (UV 354 nm) showed on average four CpG molecules conjugated per polymer. CpG-dextran prepared from 20 or 40-kDa dextran had a size of 17 nm while 70 or 110-kDa was 30 nm. CpG-dextran was preferentially taken up by dendritic cells, followed by macrophages and then B-cells. Only the 20-kDa dextran conjugate significantly enhanced uptake by bone-marrow derived dendritic cells (BMDCs) compared to free CpG. Confocal microscopy showed that CpG and CpG-dextran accumulates in the endo-lysosomal compartment of BMDCs at 24 h. All conjugates upregulated activation markers (CD40, CD80 or CD86) of BMDCs to a similar level as for free CpG. CpG-dextran 40-kDa produced highest levels of cytokines (TNF-α, IL-6, and IL-12p70) secreted by BMDCs. Enzymatic protection assays showed that the conjugate made from dextran 20-kDa provided no protection for CpG while the higher molecular weight conjugates reduced degradation by DNase I. The 40-kDa dextran conjugate produced the greatest in vitro immune activity, this was due to the conjugate being relatively small in size for cell uptake while sufficiently large enough to protect CpG from nuclease attack. These in vitro studies identify the need to consider the molecular weight of the carrier in bioconjugate design.
Assuntos
Células Dendríticas , Fator de Necrose Tumoral alfa , Fator de Necrose Tumoral alfa/metabolismo , Interleucina-6/metabolismo , Peso Molecular , Fosfatos/metabolismo , Dextranos/metabolismo , Citosina , Guanina , Citocinas , Oligodesoxirribonucleotídeos/farmacologia , Desoxirribonuclease I , Hidrazonas/farmacologiaRESUMO
Cytosine-phosphate-guanine (CpG) oligonucleotides are commonly-used vaccine adjuvants to promote the activation of antigen-presenting cells (APCs). To mount an effective immune response, CpG needs to be internalized and bind to its endosomal Toll-like receptor 9 (TLR-9) inside the APCs. Using flow cytometry and fluorescence microscopy, this article presents the cellular uptake data of the amino-dextran nanoparticle (aDNP) and aDNP loaded with CpG immobilized on its surface by either electrostatic adsorption or covalent conjugation. The uptake of fluorescently-labelled aDNPs by murine splenic dendritic cells and macrophages was determined by flow cytometry and uptake by murine bone-marrow-derived dendritic cells was evaluated by fluorescence microscopy. The data presented in this paper correlates with the in vitro immune-stimulatory activity observed for the two different CpG loading methods in the research article "Nanoparticle system based on amino-dextran as a drug delivery vehicle: immune-stimulatory CpG-oligonucleotide loading and delivery" (Nguyen et al., 2020) [1]. The data provide experimental evidence for a better understanding how the nanoparticle surface loading method of CpG influences the uptake of these nanoparticles by antigen-presenting cells as a step guide in the design of more effective vaccine formulations.
RESUMO
Computer-aided diagnosis (CAD) systems must constantly cope with the perpetual changes in data distribution caused by different sensing technologies, imaging protocols, and patient populations. Adapting these systems to new domains often requires significant amounts of labeled data for re-training. This process is labor-intensive and time-consuming. We propose a memory-augmented capsule network for the rapid adaptation of CAD models to new domains. It consists of a capsule network that is meant to extract feature embeddings from some high-dimensional input, and a memory-augmented task network meant to exploit its stored knowledge from the target domains. Our network is able to efficiently adapt to unseen domains using only a few annotated samples. We evaluate our method using a large-scale public lung nodule dataset (LUNA), coupled with our own collected lung nodules and incidental lung nodules datasets. When trained on the LUNA dataset, our network requires only 30 additional samples from our collected lung nodule and incidental lung nodule datasets to achieve clinically relevant performance (0.925 and 0.891 area under receiving operating characteristic curves (AUROC), respectively). This result is equivalent to using two orders of magnitude less labeled training data while achieving the same performance. We further evaluate our method by introducing heavy noise, artifacts, and adversarial attacks. Under these severe conditions, our network's AUROC remains above 0.7 while the performance of state-of-the-art approaches reduce to chance level.
Assuntos
Neoplasias Pulmonares , Nódulo Pulmonar Solitário , Diagnóstico por Computador , Humanos , Pulmão/diagnóstico por imagem , Neoplasias Pulmonares/diagnóstico por imagemRESUMO
The aim of this study is to prepare and characterize an amino-dextran nanoparticle (aDNP) platform and investigate two loading strategies for unmethylated cytosine-phosphate-guanine (CpG) oligonucleotide. aDNP was prepared by desolvation of amino-dextran followed by the chemical crosslinking of amino groups. Size, surface charge, and surface morphology of aDNP was determined by dynamic light scattering and transmission electron microscopy. CpG was either loaded onto aDNP by adsorption (CpG-adsorbed-aDNP) or conjugated to aDNP (CpG-conjugated-aDNP). In vitro cytokine production by bone marrow-derived dendritic cells (BMDCs) was measured by flow cytometry. aDNPs size and zeta potential could be controlled to produce uniform particles in the size range of 50 to 300 nm, surface charge of -16.5 to +14 mV, and were spherical in shape. Formulation control parameters investigated included the anti-solvent, water-to-anti-solvent ratio, level of amine functionality of dextran, and the molar ratio of glutaraldehyde to amine. aDNP could be lyophilized without additional cryoprotectant. Unloaded cationic aDNP (+13 mV) showed acceptable in vitro hemolysis. Unloaded and CpG-loaded aDNPs showed no cytotoxicity on BMDCs. CpG-loaded nanoparticles stimulated cytokine production by BMDCs, the level of cytokine production was higher for CpG-conjugated-aDNP compared to CpG-absorbed-aDNP. aDNP is a promising new drug delivery platform as its offers versatility in loading and tuning of particle properties.
RESUMO
Automatic and accurate classification of apoptosis, or programmed cell death, will facilitate cell biology research. The state-of-the-art approaches in apoptosis classification use deep convolutional neural networks (CNNs). However, these networks are not efficient in encoding the part-whole relationships, thus requiring a large number of training samples to achieve robust generalization. This paper proposes an efficient variant of capsule networks (CapsNets) as an alternative to CNNs. Extensive experimental results demonstrate that the proposed CapsNets achieve competitive performances in target cell apoptosis classification, while significantly outperforming CNNs when the number of training samples is small. To utilize temporal information within microscopy videos, we propose a recurrent CapsNet constructed by stacking a CapsNet and a bi-directional long short-term recurrent structure. Our experiments show that when considering temporal constraints, the recurrent CapsNet achieves 93.8% accuracy and makes significantly more consistent prediction than NNs.
Assuntos
Apoptose/fisiologia , Técnicas Citológicas/métodos , Processamento de Imagem Assistida por Computador/métodos , Microscopia de Contraste de Fase/métodos , Redes Neurais de Computação , Linhagem Celular Tumoral , Células/classificação , HumanosRESUMO
Complex visual data contain discriminative structures that are difficult to be fully captured by any single feature descriptor. While recent work on domain adaptation focuses on adapting a single hand-crafted feature, it is important to perform adaptation of a hierarchy of features to exploit the richness of visual data. We propose a novel framework for domain adaptation using a sparse and hierarchical network (DASH-N). Our method jointly learns a hierarchy of features together with transformations that rectify the mismatch between different domains. The building block of DASH-N is the latent sparse representation. It employs a dimensionality reduction step that can prevent the data dimension from increasing too fast as one traverses deeper into the hierarchy. The experimental results show that our method compares favorably with the competing state-of-the-art methods. In addition, it is shown that a multi-layer DASH-N performs better than a single-layer DASH-N.
RESUMO
INTRODUCTION AND AIMS: This study examined prevalence and predictors of alcohol consumption and alcohol problems in a sample of medical students in Vietnam. DESIGN AND METHODS: A cross-sectional survey using a multi-stage cluster sampling approach was conducted in 2007 in two universities in Vietnam. The students (n = 619, 100% response rate) completed questionnaires based on the Alcohol Use Disorder Identification Test. A score of >or=8 defined presence of alcohol problems. Data analyses adjusted for the cluster sampling approach. RESULTS: Overall 65.5% of students had drunk alcohol during the previous year while alcohol problems were detected in 12.5%. Male students, students who reported that their family members drank and students who reported that their flat mates were drinking were more likely to be current drinkers. Male students were 14.3 times more likely to have an Alcohol Use Disorder Identification Test score of >or=8 compared with female students (P = 0.005). DISCUSSION AND CONCLUSIONS: Intervention programs focusing on male students and their social environment are warranted. As Vietnamese society rapidly modernises prevention programs for female students may also be needed.