Proteomics as a research tool in clinical and experimental ophthalmology

It is estimated that 37 million people worldwide suffer from blindness and 124 million people have impaired vision. While the relatively recently developed therapies, antivascular endothelial growth factor inhibitors for the treatment of age-related macular degeneration, and prostaglandin analogues for the treatment of glaucoma are beneficial for some patients, there are many individuals with sight-threatening diseases for whom no effective pharmacological therapy is available. For many of these diseases, the molecular mechanisms remain to be comprehensively elucidated, thus precluding the design of successful therapies against specific pathological targets. The current review summarises recent attempts to elucidate molecular mechanisms of ocular diseases, including diabetic retinal disease, age-related macular degeneration and inherited blindness using proteomic methodologies. A novel hypothesis can be generated from global protein expression analysis of disease tissue, which can then be addressed with cellular and in vivo functional studies. For example, the identification of extracellular carbonic anhydrase from the vitreous of diabetic retinopathy patients using MS based proteomics led to the elucidation of a new pathway involved in intraretinal edema, which could be inhibited by a number of agents targeting different proteins in this pathway in relevant animal models. The potential of protein biomarkers for diagnosis and the identification of novel disease mechanisms are also discussed.     

Hard exudate detection in retinal fundus images using supervised learning

Neural Computing and Applications - The patients with diabetes have a chance to develop diabetic retinopathy (DR) which affects to the eyes. DR can cause blindness if the patients do not control...     

Automatic Diabetic Retinopathy Screening via Cascaded Framework Based on Image- and Lesion-Level Features Fusion

The early detection of diabetic retinopathy is crucial for preventing blindness. However, it is time-consuming to analyze fundus images manually, especially considering the increasing amount of medical images. In this paper, we propose an automatic diabetic retinopathy screening method using color fundus images. Our approach consists of three main components: edge-guided candidate microaneurysms detection, candidates classification using mixed features, and diabetic retinopathy prediction using fused features of image level and lesion level. We divide a screening task into two sub-classification tasks: 1) verifying candidate microaneurysms by a naive Bayes classifier; 2) predicting diabetic retinopathy using a support vector machine classifier. Our approach can effectively alleviate the imbalanced class distribution problem. We evaluate our method on two public databases: Lariboisìere and Messidor, resulting in an area under the curve of 0.908 on Lariboisìere and 0.832 on Messidor. These scores demonstrate the advantages of our approach over the existing methods.

     

Detection of microaneurysms using multi-scale correlation coefficients

This paper presents a new approach to the computer aided diagnosis (CAD) of diabetic retinopathy (DR)—a common and severe complication of long-term diabetes which damages the retina and cause blindness. Since microaneurysms are regarded as the first signs of DR, there has been extensive research on effective detection and localization of these abnormalities in retinal images. In contrast to existing algorithms, a new approach based on multi-scale correlation filtering (MSCF) and dynamic thresholding is developed. This consists of two levels, microaneurysm candidate detection (coarse level) and true microaneurysm classification (fine level). The approach was evaluated based on two public datasets—ROC (retinopathy on-line challenge, http://roc.healthcare.uiowa.edu) and DIARETDB1 (standard diabetic retinopathy database, http://www.it.lut.fi/project/imageret/diaretdb1). We conclude our method to be effective and efficient.     

Proteome profiling of vitreoretinal diseases by cluster analysis

Vitreous samples collected in retinopathic surgeries have diverse properties, making proteomics analysis difficult. We report a cluster analysis to evade this difficulty. Vitreous and subretinal fluid samples were collected from 60 patients during surgical operation of non‐proliferative diabetic retinopathy, proliferative diabetic retinopathy, proliferative vitreoretinopathy, and rhegmatogenous retinal detachment. For controls, we collected vitreous fluid from patients of idiopathic macular hole, epiretinal, and from a healthy postmortem donor. Proteins from these samples were subjected to quantitative proteomics using two‐dimensional gel electrophoresis. We selected 105 proteins robustly expressed among ca. 400 protein spots and subjected them to permutation test. By using permutation test analysis we observed unique variations in the expression of some of these proteins in vitreoretinal diseases when compared to the control and to each other: (i) the levels of inflammation‐associated proteins such as alpha1‐antitrypsin, apolipoprotein A4, albumin, and transferrin were significantly higher in all four types of vitreoretinal diseases, and (ii) each vitreoretinal disease elevated a unique set of proteins, which can be interpreted based on the pathology of retinopathy. Our protocol will be effective for the study of protein expression in other types of clinical samples of diverse properties.     

Adaptive machine learning classification for diabetic retinopathy

Multimedia Tools and Applications - Diabetic retinopathy is the main cause of the blindness worldwide. However, the diabetic retinopathy is hard to be detected in the initial stages, and the...     

Diabetic retinopathy detection and stage classification in eye fundus images using active deep learning

Multimedia Tools and Applications - Retinal fundus image analysis (RFIA) for diabetic retinopathy (DR) screening can be used to reduce the risk of blindness among diabetic patients. The RFIA...     

A computational model for signaling pathways in bounded small-world networks corresponding to brain size

A computational model, the bounded composite inverse-d architecture (BCIA), was developed to characterize signaling in small-world networks with large but bounded numbers of nodes, as in human brains. The model is based upon an N-dimensional symmetrical grid with borders, with complete local connections from each node and relatively fewer long-range connections. The length of the signaling pathway generated by a greedy algorithm between two nodes exhibited polylogarithmic behavior when the grid distance between the nodes was less than m, the maximal length of a long-range connection for that network. The simulated length of signaling pathway became linear with internode distance when the grid distance between the two nodes was greater than m. The intensity of long-range connections among nodes was found to be negatively related to the simulated length of signaling pathway. For a constant grid distance between nodes, the average length of a simulated signaling pathway increased with dimension of the BCIA graph. Most strikingly, BCIA simulations of networks with large but bounded numbers (10⁹–10¹³) of nodes, approximating the number of neurons in the human brain, found that the length of simulated signaling pathway can be substantially shorter than that predicted by the best known asymptotic theoretical bound in small-world networks of infinite size.     

Sensing the insulin signaling pathway with an antibody array

The development of insulin resistance and type 2 diabetes is determined by various factors, including defects within the insulin signaling pathway. Mediators of insulin resistance operate through activation of various protein kinase C isoforms, IκB kinase β (IKKβ), and/or c‐Jun N‐terminal kinase, and subsequent inhibition of the proximal insulin signaling pathway via the insulin receptor substrate 1 and Akt. These mechanisms are still largely unresolved because of the complexity of the molecular events. In this study, an expression and activation state profiling of multiple known key signaling biomolecules involved in insulin metabolic and mitogenic signaling pathways was evaluated using a phosphospecific antibody array platform. The results of the arrayed antibodies were verified by the multiplexed bead array assay and conventional Western blot analysis, and confirmed the well‐known inhibitory effects of phorbol esters on insulin signaling pathway activation. Of interest, the increase in protein kinase C signaling responses with phorbol esters was associated with activation of the lipid phosphatase PTEN and a 27 kDa HSP. Thus, this insulin signaling antibody array provides a powerful and effective way to investigate the mechanism of insulin resistance and likely assist the development of innovative therapeutic drugs for type 2 diabetes.     

Screening for diabetic retinopathy using computer based image analysis and statistical classification

Diabetic retinopathy is one of the most common causes of blindness in Europe. However, efficient therapies do exist. An accurate and early diagnosis and correct application of treatment can prevent blindness in more than 50% of all cases. Digital imaging is becoming available as a means of screening for diabetic retinopathy. As well as providing a high quality permanent record of the retinal appearance, which can be used for monitoring of progression or response to treatment, and which can be reviewed by an ophthalmologist, digital images have the potential to be processed by automatic analysis systems. We have described the preliminary development of a tool to provide automatic analysis of digital images taken as part of routine monitoring of diabetic retinopathy in our clinic. Various statistical classifiers, a Bayesian, a Mahalanobis, and a KNN classifier were tested. The system was tested on 134 retinal images. The Mahalanobis classifier had the best results: microaneurysms, haemorrhages, exudates, and cotton wool spots were detected with a sensitivity of 69, 83, 99, and 80%, respectively.     

Two‐dimensional gel electrophoresis of apolipoprotein C‐III and other serum glycoproteins for the combined screening of human congenital disorders of O‐ and N‐glycosylation

Congenital disorders of glycosylation (CDG) are inherited diseases that can affect not only the N‐glycan (e.g. CDG type I and II) but also the O‐glycan biosynthesis pathway. In the absence of specific clinical symptoms, there is a need for a reliable biological screening of these two groups of CDG. Using a few microlitres of human serum, 2‐DE and immunoblotting were applied to the separation and simultaneous detection of the isoforms of the O‐glycosylated protein apolipoprotein C‐III (apoC‐III) and of four N‐glycosylated proteins, namely alpha‐antitrypsin, alpha‐1 acid glycoprotein, haptoglobin and transferrin. For the study of O‐glycosylation, this technique allowed the reliable separation of the three fractions of apoC‐III and the determination of normal percentage values in an adult population. Concerning N‐glycosylation, the study of serum samples from patients with CDG type Ia revealed marked abnormalities systematically affecting the four 2‐DE separated N‐linked glycoproteins. 2‐DE coupled to immunoblotting using a mixture of specific antibodies could be easily and reliably employed for the combined screening of both N‐ and O‐glycosylation disorders in humans.     

Identification of different stages of diabetic retinopathy using retinal optical images

Diabetes is a disease which occurs when the pancreas does not secrete enough insulin or the body is unable to process it properly. This disease affects slowly the circulatory system including that of the retina. As diabetes progresses, the vision of a patient may start to deteriorate and lead to diabetic retinopathy. In this study on different stages of diabetic retinopathy, 124 retinal photographs were analyzed. As a result, four groups were identified, viz., normal retina, moderate non-proliferative diabetic retinopathy, severe non-proliferative diabetic retinopathy and proliferative diabetic retinopathy. Classification of the four eye diseases was achieved using a three-layer feedforward neural network. The features are extracted from the raw images using the image processing techniques and fed to the classifier for classification. We demonstrate a sensitivity of more than 90% for the classifier with the specificity of 100%.     

Detection and decision-support diagnosis of diabetic retinopathy using machine vision

This invited paper considers the results of the IMAGERET project. The goal of the project is to demonstrate how lesions in a retina caused by diabetic retinopathy can be detected from color fundus images by using machine vision methods. The project consists of the following results: an image annotation tool for medical expert annotation, diabetic retinopathy databases, an evaluation framework for development and comparison of methods, image-based and pixel-based methods, and new imaging solutions. The automated diagnosis can be seen in two steps: in the first step, it is decided whether an eye needs further analysis (too many lesions visible) or not (the eye is healthy enough). In the second step, fundus images selected for further analysis are automatically diagnosed. The developed system saves both resources of medical experts and costs in healthcare. It will further offer a tool for the health care providers to improve the quality of the life of diabetes patients. This is important since the number of diabetes patients is increasing, especially rapidly in the developed countries.     

A NEW EYENET MODEL FOR DIAGNOSIS OF DIABETIC RETINOPATHY

Diabetic retinopathy (DR) is an eye disease caused by complications of diabetes and it should be detected early for effective treatment. As diabetes progresses, the vision of a patient may start to deteriorate and lead to diabetic retinopathy. Two types were identified: nonproliferative diabetic retinopathy (NPDR) and proliferative diabetic retinopathy (PDR). In this study, to diagnose diabetic retinopathy, we have proposed a new EYENET model that was obtained by combining the modified probabilistic neural network (PNN) and a modified radial basis function neural network (RBFNN), and hence, it possesses the advantages of both models. The features such as blood vessels and hemorrhages of the NPDR image and exudates of the PDR image are extracted from the raw images using image-processing techniques and are fed to the classifier for classification. A total of 600 fundus images were used, out of which 400 were used for training, and 200 images were used for testing. Experimental results show that PNN has an accuracy of 96%, modified PNN has an accuracy of 97.5%, RBFNN has an accuracy of 93.5%, modified RBFNN has an accuracy of 95.5%, and the proposed EYENET model has an accuracy of 98.5%. This infers that our proposed model outperforms all other models.     

糖尿病性视网膜图像的深度学习分类方法

目的 糖尿病性视网膜病变（DR）是目前比较严重的一种致盲眼病,因此,对糖尿病性视网膜病理图像的自动分类具有重要的临床应用价值。基于人工分类视网膜图像的方法存在判别性特征提取困难、分类性能差、耗时费力且很难得到客观统一的医疗诊断等问题,为此,提出一种基于卷积神经网络和分类器的视网膜病理图像自动分类系统。方法 首先,结合现有的视网膜图像的特点,对图像进行去噪、数据扩增、归一化等预处理操作;其次,在AlexNet网络的基础上,在网络的每一个卷积层和全连接层前引入一个批归一化层,得到一个网络层次更复杂的深度卷积神经网络BNnet。BNnet网络用于视网膜图像的特征提取网络,对其训练时采用迁移学习的策略利用ILSVRC2012数据集对BNnet网络进行预训练,再将训练得到的模型迁移到视网膜图像上再学习,提取用于视网膜分类的深度特征;最后,将提取的特征输入一个由全连接层组成的深度分类器将视网膜图像分为正常的视网膜图像、轻微病变的视网膜图像、中度病变的视网膜图像等5类。结果 实验结果表明,本文方法的分类准确率可达0.93,优于传统的直接训练方法,且具有较好的鲁棒性和泛化性。结论 本文提出的视网膜病理图像分类框架有效地避免了人工特征提取和图像分类的局限性,同时也解决了样本数据不足而导致的过拟合问题。     

Long range molecular dynamics study of interactions of the eukaryotic glucosamine-6-phosphate synthase with fructose-6-phosphate and UDP-GlcNAc

Glucosamine-6-phosphate synthase (EC 2.6.1.16) is responsible for catalysis of the first and practically irreversible step in hexosamine metabolism. The final product of this pathway, uridine 5′ diphospho N-acetyl-d-glucosamine (UDP-GlcNAc), is an essential substrate for assembly of bacterial and fungal cell walls. Moreover, the enzyme is involved in phenomenon of hexosamine induced insulin resistance in type II diabetes, which makes of it a potential target for anti-fungal, anti-bacterial and anti-diabetic therapy.The crystal structure of isomerase domain from human pathogenic fungus Candida albicans has been solved recently but it doesn’t reveal the molecular mechanism details of inhibition taking place under UDP-GlcNAc influence, the unique feature of eukaryotic enzyme. The following study is a continuation of the previous research based on comparative molecular dynamics simulations of the structures with and without the enzyme's physiological inhibitor (UDP-GlcNAc) bound. The models used for this study included fructose-6-phosphate, one of the enzyme's substrates in its binding pocket.The simulation results studies demonstrated differences in mobility of the compared structures. Some amino acid residues were determined, for which flexibility is evidently different between the models. Importantly, it has been confirmed that the most fixed residues are related to the inhibitor binding process and to the catalysis reaction. The obtained results constitute an important step towards understanding of the inhibition that GlcN-6-P synthase is subjected by UDP-GlcNAc molecule.     

基于R-FCN算法的糖尿病眼底病变自动诊断

糖尿病眼底病变(Diabetic Retinopathy,DR)是糖尿病患者常见的致盲疾病,可使用深度学习算法对患者的糖尿病眼底图片进行图像识别,实现对糖尿病眼底病变的辅助诊断。针对以往普通卷积神经网络只能进行分类和输入尺寸固定的问题,提出了基于目标检测的区域全卷积网络(Region-based Fully Convolutional Networks,R-FCN)算法,实现同时对任意尺寸输入的糖尿病眼底图片的分类和病变区域检测。针对原始R-FCN算法对小目标(极小的出血点和血管瘤)检测困难的问题,对R-FCN算法做了一定的改进,加入特征金字塔网络(Feature Pyramid Networks,FPN)结构,升级主干网络,修改区域建议网络(Region Proposal Network,RPN)。实现结果表明,改进后的RFCN算法能以很高的正确率实现对糖尿病眼底图片的五级分类(健康、轻度、中度、重度、增殖)和病变区域检测(血管瘤、眼底出血、玻璃体出血)。