Automatic non-proliferative diabetic retinopathy screening system based on color fundus image

Background

Non-proliferative diabetic retinopathy is the early stage of diabetic retinopathy. Automatic detection of non-proliferative diabetic retinopathy is significant for clinical diagnosis, early screening and course progression of patients.

Methods

This paper introduces the design and implementation of an automatic system for screening non-proliferative diabetic retinopathy based on color fundus images. Firstly, the fundus structures, including blood vessels, optic disc and macula, are extracted and located, respectively. In particular, a new optic disc localization method using parabolic fitting is proposed based on the physiological structure characteristics of optic disc and blood vessels. Then, early lesions, such as microaneurysms, hemorrhages and hard exudates, are detected based on their respective characteristics. An equivalent optical model simulating human eyes is designed based on the anatomical structure of retina. Main structures and early lesions are reconstructed in the 3D space for better visualization. Finally, the severity of each image is evaluated based on the international criteria of diabetic retinopathy.

Results

The system has been tested on public databases and images from hospitals. Experimental results demonstrate that the proposed system achieves high accuracy for main structures and early lesions detection. The results of severity classification for non-proliferative diabetic retinopathy are also accurate and suitable.

Conclusions

Our system can assist ophthalmologists for clinical diagnosis, automatic screening and course progression of patients.

     

Automated Multi-Lesion Detection for Referable Diabetic Retinopathy in Indigenous Health Care

Diabetic Retinopathy (DR) is a complication of diabetes mellitus that affects more than one-quarter of the population with diabetes, and can lead to blindness if not discovered in time. An automated screening enables the identification of patients who need further medical attention. This study aimed to classify retinal images of Aboriginal and Torres Strait Islander peoples utilizing an automated computer-based multi-lesion eye screening program for diabetic retinopathy. The multi-lesion classifier was trained on 1,014 images from the São Paulo Eye Hospital and tested on retinal images containing no DR-related lesion, single lesions, or multiple types of lesions from the Inala Aboriginal and Torres Strait Islander health care centre. The automated multi-lesion classifier has the potential to enhance the efficiency of clinical practice delivering diabetic retinopathy screening. Our program does not necessitate image samples for training from any specific ethnic group or population being assessed and is independent of image pre- or post-processing to identify retinal lesions. In this Aboriginal and Torres Strait Islander population, the program achieved 100% sensitivity and 88.9% specificity in identifying bright lesions, while detection of red lesions achieved a sensitivity of 67% and specificity of 95%. When both bright and red lesions were present, 100% sensitivity with 88.9% specificity was obtained. All results obtained with this automated screening program meet WHO standards for diabetic retinopathy screening.     

A diabetic retinopathy detection method using an improved pillar K-means algorithm

The paper presents a new approach for medical image segmentation. Exudates are a visible sign of diabetic retinopathy that is the major reason of vision loss in patients with diabetes. If the exudates extend into the macular area, blindness may occur. Automated detection of exudates will assist ophthalmologists in early diagnosis. This segmentation process includes a new mechanism for clustering the elements of high-resolution images in order to improve precision and reduce computation time. The system applies K-means clustering to the image segmentation after getting optimized by Pillar algorithm; pillars are constructed in such a way that they can withstand the pressure. Improved pillar algorithm can optimize the K-means clustering for image segmentation in aspects of precision and computation time. This evaluates the proposed approach for image segmentation by comparing with Kmeans and Fuzzy C-means in a medical image. Using this method, identification of dark spot in the retina becomes easier and the proposed algorithm is applied on diabetic retinal images of all stages to identify hard and soft exudates, where the existing pillar K-means is more appropriate for brain MRI images. This proposed system help the doctors to identify the problem in the early stage and can suggest a better drug for preventing further retinal damage.     

A Hybrid Color Space for Skin Detection Using Genetic Algorithm Heuristic Search and Principal Component Analysis Technique

Color is one of the most prominent features of an image and used in many skin and face detection applications. Color space transformation is widely used by researchers to improve face and skin detection performance. Despite the substantial research efforts in this area, choosing a proper color space in terms of skin and face classification performance which can address issues like illumination variations, various camera characteristics and diversity in skin color tones has remained an open issue. This research proposes a new three-dimensional hybrid color space termed SKN by employing the Genetic Algorithm heuristic and Principal Component Analysis to find the optimal representation of human skin color in over seventeen existing color spaces. Genetic Algorithm heuristic is used to find the optimal color component combination setup in terms of skin detection accuracy while the Principal Component Analysis projects the optimal Genetic Algorithm solution to a less complex dimension. Pixel wise skin detection was used to evaluate the performance of the proposed color space. We have employed four classifiers including Random Forest, Naïve Bayes, Support Vector Machine and Multilayer Perceptron in order to generate the human skin color predictive model. The proposed color space was compared to some existing color spaces and shows superior results in terms of pixel-wise skin detection accuracy. Experimental results show that by using Random Forest classifier, the proposed SKN color space obtained an average F-score and True Positive Rate of 0.953 and False Positive Rate of 0.0482 which outperformed the existing color spaces in terms of pixel wise skin detection accuracy. The results also indicate that among the classifiers used in this study, Random Forest is the most suitable classifier for pixel wise skin detection applications.     

Possible new method to improve detection of diabetic retinopathy: Polaroid non-mydriatic retinal photography.

Two retinal cameras (Canon CR2 45NM and CR3 45NM) have recently become available and are capable of producing an instant colour photography of a 45 degree field of retina, including the macula and optic disc, without dilatation of the pupils being necessary. The ability of each camera to detect diabetic retinopathy was compared with that of doctors in diabetic clinics using ophthalmoscopy during busy clinic hours. The CR3 was found to be considerably superior to the CR2 in terms of quality of photograph because it can use a smaller pupil. Overall, the detection rate of the camera was more than four times higher than that of ophthalmoscopy through undilated pupils and more than twice as high as that of ophthalmoscopy through dilated pupils. Lesions missed by ophthalmoscopy but detected by the camera included soft exudates and circinate rings of hard exudates, sometimes encroaching on the macula. Though various aspects of this system of screening for diabetic retinopathy, in particular its ability to detect new retinal vessels, have not yet been assessed, the system may prove beneficial in the detection and monitoring of diabetic retinopathy.     

Automatic diagnosis of mammographic abnormalities based on hybrid features with learning classifier

Breast cancer screening is currently performed by mammography, which is limited by overlying anatomy and dense breast tissue. Computer-aided detection (CAD) systems can serve as a double reader to improve radiologist performance. In this paper, we have applied a novel approach to segmentation of suspicious region by mammogram and classification based on hybrid features with learning classifier. We formulated differentiation of lesion from normal tissue as a supervised learning problem, and applied this learning method to develop the classification algorithm. The algorithm has been verified with 164 mammograms in the mini Mammographic Image Analysis Society database. The experimental results show that the detection method has a sensitivity of 94.5% at 0.26 false positives per image. The efficiency of algorithm is measured using free receiver operating characteristics curve and the results are highlighted. We conclude that CAD technology with learning classifier has the potential to help radiologists with the task of discriminating between lesion and normal tissues.     

Detection of Neovascularization Based on Fractal and Texture Analysis with Interaction Effects in Diabetic Retinopathy

Diabetic retinopathy is a major cause of blindness. Proliferative diabetic retinopathy is a result of severe vascular complication and is visible as neovascularization of the retina. Automatic detection of such new vessels would be useful for the severity grading of diabetic retinopathy, and it is an important part of screening process to identify those who may require immediate treatment for their diabetic retinopathy. We proposed a novel new vessels detection method including statistical texture analysis (STA), high order spectrum analysis (HOS), fractal analysis (FA), and most importantly we have shown that by incorporating their associated interactions the accuracy of new vessels detection can be greatly improved. To assess its performance, the sensitivity, specificity and accuracy (AUC) are obtained. They are 96.3%, 99.1% and 98.5% (99.3%), respectively. It is found that the proposed method can improve the accuracy of new vessels detection significantly over previous methods. The algorithm can be automated and is valuable to detect relatively severe cases of diabetic retinopathy among diabetes patients.     

One-Hot Vector Hybrid Associative Classifier for Medical Data Classification

Pattern recognition and classification are two of the key topics in computer science. In this paper a novel method for the task of pattern classification is presented. The proposed method combines a hybrid associative classifier (Clasificador Híbrido Asociativo con Traslación, CHAT, in Spanish), a coding technique for output patterns called one-hot vector and majority voting during the classification step. The method is termed as CHAT One-Hot Majority (CHAT-OHM). The performance of the method is validated by comparing the accuracy of CHAT-OHM with other well-known classification algorithms. During the experimental phase, the classifier was applied to four datasets related to the medical field. The results also show that the proposed method outperforms the original CHAT classification accuracy.     

A Comprehensive Texture Segmentation Framework for Segmentation of Capillary Non-Perfusion Regions in Fundus Fluorescein Angiograms

Capillary non-perfusion (CNP) in the retina is a characteristic feature used in the management of a wide range of retinal diseases. There is no well-established computation tool for assessing the extent of CNP. We propose a novel texture segmentation framework to address this problem. This framework comprises three major steps: pre-processing, unsupervised total variation texture segmentation, and supervised segmentation. It employs a state-of-the-art multiphase total variation texture segmentation model which is enhanced by new kernel based region terms. The model can be applied to texture and intensity-based multiphase problems. A supervised segmentation step allows the framework to take expert knowledge into account, an AdaBoost classifier with weighted cost coefficient is chosen to tackle imbalanced data classification problems. To demonstrate its effectiveness, we applied this framework to 48 images from malarial retinopathy and 10 images from ischemic diabetic maculopathy. The performance of segmentation is satisfactory when compared to a reference standard of manual delineations: accuracy, sensitivity and specificity are 89.0%, 73.0%, and 90.8% respectively for the malarial retinopathy dataset and 80.8%, 70.6%, and 82.1% respectively for the diabetic maculopathy dataset. In terms of region-wise analysis, this method achieved an accuracy of 76.3% (45 out of 59 regions) for the malarial retinopathy dataset and 73.9% (17 out of 26 regions) for the diabetic maculopathy dataset. This comprehensive segmentation framework can quantify capillary non-perfusion in retinopathy from two distinct etiologies, and has the potential to be adopted for wider applications.     

Treatment of renal failure from diabetic nephropathy with cadaveric homograft

We report two patients with terminal renal failure secondary to diabetic nephropathy treated with cadaveric kidney transplantation. Neither of these patients had peripheral vascular disease or peripheral neuropathy. There was a proliferative diabetic retinopathy with hemorrhages and exudates in one patient and only background diabetic changes in the ocular fundi of the other; there have been no significant changes in visual acuity or retinopathy in either patient following the transplantation. Both have good kidney function after 8 and 15 months and are completely rehabilitated.The requirement for insulin decreased in both patients during the period of renal insufficiency and increased following transplantation; this seemed to be related to the large dose of steroids given because now that a maintenance level of steroids has been established, both patients require the same dosage of insulin as they did before the onset of renal insufficiency.     

Automated quantification of superficial retinal capillaries and large vessels for diabetic retinopathy on optical coherence tomographic angiography

Optical coherence tomography angiography (OCTA) is a relatively new technique with capillary‐level resolution, which has shown great potential for the diagnosis of diabetic retinopathy (DR). A fully automatic algorithm for the quantitative measurement of microcirculatory changes in sight‐threatening DR is presented. The foveal avascular zone (FAZ) segmentation was improved with a graph‐theoretic method and the large vessels and capillaries were separately identified and analyzed. The method was evaluated in healthy and diabetic eyes with various stages of retinopathy. Results showed that, compared with the healthy group, the diabetic group showed a significantly larger large vessel density, but a significantly smaller capillary density (P < .001). Circularity of FAZ was significantly smaller while nonperfusion area was significantly larger in the diabetic group. The combined variable of all image metrics reached an area under the ROC of 0.853 (95% CI, 0.784‐0.923) for mild to moderate nonproliferative DR and 0.950 (95% CI, 0.922‐0.979) for proliferative DR. Microvascular and FAZ changes with various DR stages can be accurately delineated using the developed automatic program. Quantitative metrics on OCTA serve as potential biomarkers for the staging of DR.     

Optimization of Regularization Parameters in Compressed Sensing of Magnetic Resonance Angiography: Can Statistical Image Metrics Mimic Radiologists' Perception?

In Compressed Sensing (CS) of MRI, optimization of the regularization parameters is not a trivial task. We aimed to establish a method that could determine the optimal weights for regularization parameters in CS of time-of-flight MR angiography (TOF-MRA) by comparing various image metrics with radiologists’ visual evaluation. TOF-MRA of a healthy volunteer was scanned using a 3T-MR system. Images were reconstructed by CS from retrospectively under-sampled data by varying the weights for the L1 norm of wavelet coefficients and that of total variation. The reconstructed images were evaluated both quantitatively by statistical image metrics including structural similarity (SSIM), scale invariant feature transform (SIFT) and contrast-to-noise ratio (CNR), and qualitatively by radiologists’ scoring. The results of quantitative metrics and qualitative scorings were compared. SSIM and SIFT in conjunction with brain masks and CNR of artery-to-parenchyma correlated very well with radiologists’ visual evaluation. By carefully selecting a region to measure, we have shown that statistical image metrics can reflect radiologists’ visual evaluation, thus enabling an appropriate optimization of regularization parameters for CS.     

Predicting Peptide Structures in Native Proteins from Physical Simulations of Fragments

It has long been proposed that much of the information encoding how a protein folds is contained locally in the peptide chain. Here we present a large-scale simulation study designed to examine the extent to which conformations of peptide fragments in water predict native conformations in proteins. We perform replica exchange molecular dynamics (REMD) simulations of 872 8-mer, 12-mer, and 16-mer peptide fragments from 13 proteins using the AMBER 96 force field and the OBC implicit solvent model. To analyze the simulations, we compute various contact-based metrics, such as contact probability, and then apply Bayesian classifier methods to infer which metastable contacts are likely to be native vs. non-native. We find that a simple measure, the observed contact probability, is largely more predictive of a peptide''s native structure in the protein than combinations of metrics or multi-body components. Our best classification model is a logistic regression model that can achieve up to 63% correct classifications for 8-mers, 71% for 12-mers, and 76% for 16-mers. We validate these results on fragments of a protein outside our training set. We conclude that local structure provides information to solve some but not all of the conformational search problem. These results help improve our understanding of folding mechanisms, and have implications for improving physics-based conformational sampling and structure prediction using all-atom molecular simulations.     

Comparison of non-mydriatic retinal photography with ophthalmoscopy in 2159 patients: mobile retinal camera study.

OBJECTIVE--To determine whether non-mydriatic Polaroid retinal photography was comparable to ophthalmoscopy with mydriasis in routine clinic screening for early, treatable diabetic retinopathy. DESIGN--Prospective study of ophthalmoscopic findings according to retinal camera screening and ophthalmoscopy and outcome of referral to ophthalmologist. SETTING--Outpatient diabetic clinics of three teaching hospitals and three district general hospitals. PATIENTS--2159 Adults selected randomly from the diabetic clinics, excluding only those registered as blind or those in wheelchairs and unable to enter the screening vehicle. MAIN OUTCOME MEASURES--Numbers of patients and eyes correctly identified by each technique as requiring referral with potentially treatable retinopathy (new vessel formation and maculopathy) and congruence in numbers of microaneurysms, haemorrhages, and exudates reported. RESULTS--Camera screening missed two cases of new vessel formation and did not identify a further 12 but indicated a need for referral. Ophthalmoscopy missed five cases of new vessel formation and indicated a need for referral in another four for other reasons. Maculopathy was reported in 147 eyes with camera screening alone and 95 eyes by ophthalmoscopy only (chi 2 = 11.2; p less than 0.001), in 66 and 29 of which respectively maculopathy was subsequently confirmed. Overall, 38 eyes received laser treatment for maculopathy after detection by camera screening compared with 17 after ophthalmoscopic detection (chi 2 = 8.0; p less than 0.01). Camera screening underestimated numbers of microaneurysms (chi 2 = 12.9; p less than 0.001) and haemorrhages (chi 2 = 7.4; p less than 0.01) and ophthalmoscopy underestimated hard exudates (chi 2 = 48.2; p less than 0.001). CONCLUSIONS--Non-mydriatic Polaroid retinal photography is at least as good as ophthalmoscopy with mydriasis in routine diabetic clinics in identifying new vessel formation and absence of retinopathy and is significantly better in detecting exudative maculopathy.     

A Multi-Branch Convolutional Neural Network for Screening and Staging of Diabetic Retinopathy Based on Wide-Field Optical Coherence Tomography Angiography

BackgroundDiabetic retinopathy (DR) is one of the major causes of blindness in adults suffering from diabetes. With the development of wide-field optical coherence tomography angiography (WF-OCTA), it is to become a gold standard for diagnosing DR. The demand for automated DR diagnosis system based on OCTA images have been fostered due to large diabetic population and pervasiveness of retinopathy cases.Materials and methodsIn this study, 288 diabetic patients and 97 healthy people were imaged by the swept-source optical coherence tomography (SS-OCT) with 12 mm × 12 mm single scan centered on the fovea. A multi-branch convolutional neural network (CNN) was proposed to classify WF-OCTA images into four grades: no DR, mild non-proliferative diabetic retinopathy (NPDR), moderate to severe NPDR, and proliferative diabetic retinopathy (PDR).ResultsThe proposed model achieved a classification accuracy of 96.11%, sensitivity of 98.08% and specificity of 89.43% in detecting DR. The accuracy of the model for DR staging is 90.56%, which is higher than that of other mainstream convolution neural network models.ConclusionThis technology enables early diagnosis and objective tracking of disease progression, which may be useful for optimal treatment to reduce vision loss.     

Peripapillary Retinal Nerve Fiber Layer Changes in Preclinical Diabetic Retinopathy: A Meta-Analysis

BackgroundDiabetic retinopathy is a microvascular neurodegenerative disorder in diabetic patients. Peripapillary retinal nerve fiber layer changes have been described in patients with preclinical diabetic retinopathy, but study results have been inconsistent.ObjectiveTo assess changes in peripapillary retinal nerve fiber layer thickness in diabetic patients with preclinical diabetic retinopathy.MethodsA literature search was conducted through PubMed, EMBASE, Web of Science and Cochrane Library. Case-control studies on RNFL thickness in preclinical diabetic retinopathy patients and healthy controls were retrieved. A meta-analysis of weighted mean difference and a sensitivity analysis were performed using RevMan 5.2 software.ResultsThirteen case-control studies containing 668 diabetic patients and 556 healthy controls were selected. Peripapillary RNFL thickness was significantly reduced in patients with preclinical diabetic retinopathy compared to healthy controls in studies applying Optical Coherence Tomography (-2.88μm, 95%CI: -4.44 to -1.32, P = 0.0003) and in studies applying Scanning Laser Polarimeter (-4.21μm, 95%CI: -6.45 to -1.97, P = 0.0002). Reduction of RNFL thickness was significant in the superior quadrant (-3.79μm, 95%CI: -7.08 to -0.50, P = 0.02), the inferior quadrant (-2.99μm, 95%CI: -5.44 to -0.54, P = 0.02) and the nasal quadrant (-2.88μm, 95%CI: -4.93 to -0.82, P = 0.006), but was not significant in the temporal quadrant (-1.22μm, 95%CI: -3.21 to 0.76, P = 0.23), in diabetic patients.ConclusionPeripapillary RNFL thickness was significantly decreased in preclinical diabetic retinopathy patients compared to healthy control. Neurodegenerative changes due to preclinical diabetic retinopathy need more attention.     

Use of mobile screening unit for diabetic retinopathy in rural and urban areas.

OBJECTIVES--To compare the effectiveness of a mobile screening unit with a non-mydriatic polaroid camera in detecting diabetic retinopathy in rural and urban areas. To estimate the cost of the service. DESIGN--Prospective data collection over two years of screening for diabetic retinopathy throughout Tayside. SETTING--Tayside region, population 390,000, area 7770 km2. SUBJECTS--961 patients in rural areas and 1225 in urban areas who presented for screening. MAIN OUTCOME MEASURES--Presence of diabetic retinopathy, need for laser photocoagulation, age, duration of diabetes, and diabetic treatment. RESULTS--Compared with diabetic patients in urban areas, those in rural areas were less likely to attend a hospital based diabetic clinic (46% (442) v 86% (1054), p < 0.001); less likely to be receiving insulin (260 (27%) v 416 (34%), p < 0.001 and also after correction for differences in age distribution); more likely to have advanced (maculopathy or proliferative retinopathy) diabetic retinopathy (13% (122) v 7% (89), p < 0.001); and more likely to require urgent laser photocoagulation for previously unrecognised retinopathy (1.4% (13) v 0.5% (6), p < 0.02). The screening programme cost 10 pounds per patient screened and 1000 pounds per patient requiring laser treatment. CONCLUSION--The mobile diabetic eye screening programme detected a greater prevalence of advanced retinopathy in diabetic patients living in rural areas. Patients in rural areas were also more likely to need urgent laser photocoagulation. Present screening procedures seem to be less effective in rural areas and rural patients may benefit more from mobile screening units than urban patients.     

The English national risk-reduction programme for preservation of sight in diabetes

This paper discusses the new national guidelines for a systematic screening programme to detect sight-threatening diabetic retinopathy in the population of people with diabetes in England. A review of the literature examines the evidence base to support screening interventions and effective management and treatments in diabetic retinopathy. The current evidence supports the establishment of a digital retinal photography system using pupil dilation. A Policy Advisory Group has been formulated by the National Screening Committee to guide the meeting of this target in England. A conclusion is made that with increased effort and organisation, health care professionals can ensure that the screening programme is successfully implemented and rates of visual impairment and blindness caused by diabetic retinopathy can be reduced significantly. (Mol Cell Biochem 261: 183–185, 2004)