Early Diagnosis Of Diabetic Retinopathy Research Articles

MediDRNet: Tackling category imbalance in diabetic retinopathy classification with dual-branch learning and prototypical contrastive learning

Background and Objective:The classification of diabetic retinopathy (DR) aims to utilize the implicit information in images for early diagnosis, to prevent and mitigate the further worsening of the condition. However, existing methods are often limited by the need to operate within large, annotated datasets to show significant advantages. Additionally, the number of samples for different categories within the dataset needs to be evenly distributed, because the characteristic of sample imbalance distribution can lead to an excessive focus on high-frequency disease categories, while neglecting the less common but equally important disease categories. Therefore, there is an urgent need to develop a new classification method that can effectively alleviate the issue of sample distribution imbalance, thereby enhancing the accuracy of diabetic retinopathy classification. Methods:In this work, we propose MediDRNet, a dual-branch network model based on prototypical contrastive learning. This model adopts prototype contrastive learning, creating prototypes for different levels of lesions, ensuring they represent the core features of each lesion level. It classifies by comparing the similarity between data points and their category prototypes. Our dual-branch network structure effectively resolves the issue of category imbalance and improves classification accuracy by emphasizing subtle differences in retinal lesions. Moreover, our approach combines a dual-branch network with specific lesion-level prototypes for core feature representation and incorporates the convolutional block attention module for enhanced lesion feature identification. Results:Our experiments using both the Kaggle and UWF classification datasets have demonstrated that MediDRNet exhibits exceptional performance compared to other advanced models in the industry, especially on the UWF DR classification dataset where it achieved state-of-the-art performance across all metrics. On the Kaggle DR classification dataset, it achieved the highest average classification accuracy (0.6327) and Macro-F1 score (0.6361). Particularly in the classification tasks for minority categories of diabetic retinopathy on the Kaggle dataset (Grades 1, 2, 3, and 4), the model reached high classification accuracies of 58.08%, 55.32%, 69.73%, and 90.21%, respectively. In the ablation study, the MediDRNet model proved to be more effective in feature extraction from diabetic retinal fundus images compared to other feature extraction methods. Conclusions:This study employed prototype contrastive learning and bidirectional branch learning strategies, successfully constructing a grading system for diabetic retinopathy lesions within imbalanced diabetic retinopathy datasets. Through a dual-branch network, the feature learning branch effectively facilitated a smooth transition of features from the grading network to the classification learning branch, accurately identifying minority sample categories. This method not only effectively resolved the issue of sample imbalance but also provided strong support for the precise grading and early diagnosis of diabetic retinopathy in clinical applications, showcasing exceptional performance in handling complex diabetic retinopathy datasets. Moreover, this research significantly improved the efficiency of prevention and management of disease progression in diabetic retinopathy patients within medical practice. We encourage the use and modification of our code, which is publicly accessible on GitHub: https://github.com/ReinforceLove/MediDRNet.

Diabetic Retinopathy Lesion Segmentation Method Based on Multi-Scale Attention and Lesion Perception

The early diagnosis of diabetic retinopathy (DR) can effectively prevent irreversible vision loss and assist ophthalmologists in providing timely and accurate treatment plans. However, the existing methods based on deep learning have a weak perception ability of different scale information in retinal fundus images, and the segmentation capability of subtle lesions is also insufficient. This paper aims to address these issues and proposes MLNet for DR lesion segmentation, which mainly consists of the Multi-Scale Attention Block (MSAB) and the Lesion Perception Block (LPB). The MSAB is designed to capture multi-scale lesion features in fundus images, while the LPB perceives subtle lesions in depth. In addition, a novel loss function with tailored lesion weight is designed to reduce the influence of imbalanced datasets on the algorithm. The performance comparison between MLNet and other state-of-the-art methods is carried out in the DDR dataset and DIARETDB1 dataset, and MLNet achieves the best results of 51.81% mAUPR, 49.85% mDice, and 37.19% mIoU in the DDR dataset, and 67.16% mAUPR and 61.82% mDice in the DIARETDB1 dataset. The generalization experiment of MLNet in the IDRiD dataset achieves 59.54% mAUPR, which is the best among other methods. The results show that MLNet has outstanding DR lesion segmentation ability.

Diabetic Retinopathy Detection Using Image Processing

In today’s world, Diabetes is a very common disease that affects a lot of people’s health. As the number of diabetic patients is increasing significantly in India, there is a rise in numerous associated diseases that have disturbed society.Diabetic Retinopathy (DR) is considered to be one of such silent diseases that occur as a result of either Type 1 or Type 2 diabetes. Late diagnosis of this disease may lead to permanent eye blindness. Thus, for early diagnosis of Diabetic Retinopathy, a software-based algorithm is designed here. This technique can be promising for the pre-detection of DR without any involvement of an expert doctor and hence will save both time and money. Here, MATLAB-based image processing is used which exploits the knowledge of Computer Science and Biomedical Engineering to identify whitish lesions, cotton wool spots and hard exudates associated with DR. Based on the value of pixel counts, the image of the patient’s eye under examination is classified. Diabetic retinopathy (DR), a complication of diabetes, poses a significant threat to global vision health. Early detection and timely intervention are crucial to prevent irreversible vision loss.

Abnormalities of retinal function in type 2 diabetes mellitus patients without clinical diabetic retinopathy detected by multifocal electroretinogram

ObjectiveTo study the changes of retinal function in type 2 diabetes mellitus(DM) patients without apparently diabetic retinopathy via multifocal electroretinogram. Methods: Thirty-six type 2 DM patients (72 eyes) without visible diabetic retinopathy were selected as the experimental group, and thirty-five healthy subjects (70 eyes) were selected as the control group. All subjects were underwent multifocal electroretinogram (mf- ERG).ResultsCompared with the control group, the implicit time delay of the P1 wave in the first ring, third ring, fourth ring, and fifth ring of the experimental group was significant (t = -3.154, p = 0.004, t = -8.21, p = 0.000, t = -3.067, p = 0.004, t = -4.443, p = 0.000, respectively). The implicit time of the N1 wave in the fourth- and fifth-ring were also significantly delayed compared with the control group (t = -3.549, p = 0.001, t = 2.961, p = 0.005, respectively). Compared with the control group, the implicit time of the P1 wave and N1 wave in the temporal region of the experimental group were delayed (t = -2.148, p = 0.037, t = -2.834, p = 0.007, respectively). There were no significant difference between the experimental group and the control group of the temporal area in the amplitude density of P1 wave, N1 wave. There was no difference in the implicit time and amplitude density of the N1 and P1 waves in the nasal region between the experimental group and the control group. The multifocal electroretinogram complex parameters showed better specificity and sensitivity in the diagnosis of diabetic retinopathy.ConclusionThe multifocal electroretinogram can detect abnormal changes in the retina of type 2 DM patients without visible diabetic retinopathy. The multifocal electroretinogram complex parameter is a potential indicator for the early diagnosis of diabetic retinopathy.

The perception of diabetics in Saudi Arabia about complications of diabetes in the eyes 2022-2023

Objective: This study aimed to assess the knowledge and awareness of diabetic patients in Saudi Arabia regarding ocular complications, their practice of routine eye examinations, and diabetes management. Methods: A descriptive cross-sectional study was conducted among 430 diabetic patients in various regions of Saudi Arabia. An internet-based questionnaire was used to collect data. Results: The study revealed that 38.5% diabetic patients had excellent knowledge about ocular complications, 31.4% had average knowledge, 21.4% had poor knowledge, and 8.7% had no knowledge. Diabetic patients demonstrated higher levels of knowledge and better practices for eye exams and diabetes management compared to healthy individuals. Furthermore, a significant association was found between knowledge levels and the diagnosis of diabetic retinopathy. Education level was also correlated with knowledge, with individuals holding master\'s or PhD degrees showing the highest awareness. Conclusion: This study highlighted the importance of targeted education and awareness initiatives for diabetic patients, particularly those with lower education levels, to improve their knowledge of ocular complications. Routine eye examinations are crucial for early detection and treatment of ocular problems in diabetic patients. Early diagnosis of diabetic retinopathy is essential for better ocular outcomes. The findings provide valuable insights for healthcare providers and policymakers to enhance diabetes education and awareness programs in Saudi Arabia.

GlanceSeg: Real-time microaneurysm lesion segmentation with gaze-map-guided foundation model for early detection of diabetic retinopathy.

Early-stage diabetic retinopathy (DR) presents challenges in clinical diagnosis due to inconspicuous and minute microaneurysms (MAs), resulting in limited research in this area. Additionally, the potential of emerging foundation models, such as the segment anything model (SAM), in medical scenarios remains rarely explored. In this work, we propose a human-in-the-loop, label-free early DR diagnosis framework called GlanceSeg, based on SAM. GlanceSeg enables real-time segmentation of MA lesions as ophthalmologists review fundus images. Our human-in-the-loop framework integrates the ophthalmologist's gaze maps, allowing for rough localization of minute lesions in fundus images. Subsequently, a saliency map is generated based on the located region of interest, which provides prompt points to assist the foundation model in efficiently segmenting MAs. Finally, a domain knowledge filtering (DKF) module refines the segmentation of minute lesions. We conducted experiments on two newly-built public datasets, i.e., IDRiD and Retinal-Lesions, and validated the feasibility and superiority of GlanceSeg through visualized illustrations and quantitative measures. Additionally, we demonstrated that GlanceSeg improves annotation efficiency for clinicians and further enhances segmentation performance through fine-tuning using annotations. The clinician-friendly GlanceSeg is able to segment small lesions in real-time, showing potential for clinical applications.

Association of altered fibrinogen indexes levels as a potential biomarker in determining the possible onset of diabetic retinopathy

Research suggests that fibrinogen was related to diabetic retinopathy (DR). Then, the relationship between functional indices of fibrinogen and detailed staging of DR has not been explored. Type 2 diabetic and healthy control subjects (n = 960) were recruited in a cross-sectional study. Participants with type 2 diabetes mellitus were categorized into five stages according to their fundus lesions, and fibrinogen (Fib) and its functional indices (angle α and k value) were measured. The angle α levels increased in diabetic subjects with retinopathy compared with those without, and it was significantly elevated early in retinopathy. In contrast, the k value levels slightly decreased. Despite observing an increase in angle α levels and a decrease in k value levels during the later stages of retinopathy compared to the earlier stages, there was no statistically significant difference in the later stages. The association of the angle α and k value with DR was independent of the hyperglycaemic state and other potential confounders (OR = 1.672, 95% CI 1.489–1.876, P < 0.01; OR = 0.013, 95% CI 0.004–0.041, P < 0.01). The angle α levels and k value levels were closely correlated with retinopathy (r = 0.593, P < 0.00; r = − 0.646, P < 0.01). The ROC curve indicated that the diagnostic value of angle α and k value were (AUC = 0.897, P < 0.001; AUC = 0.859, P < 0.001). Fibrinogen function indexes, such as angle α and k value, may be valuable for the early diagnosis of DR but do not directly assess the severity of DR.

Explainable Artificial Intelligence Paves the Way in Precision Diagnostics and Biomarker Discovery for the Subclass of Diabetic Retinopathy in Type 2 Diabetics.

Diabetic retinopathy (DR), a common ocular microvascular complication of diabetes, contributes significantly to diabetes-related vision loss. This study addresses the imperative need for early diagnosis of DR and precise treatment strategies based on the explainable artificial intelligence (XAI) framework. The study integrated clinical, biochemical, and metabolomic biomarkers associated with the following classes: non-DR (NDR), non-proliferative diabetic retinopathy (NPDR), and proliferative diabetic retinopathy (PDR) in type 2 diabetes (T2D) patients. To create machine learning (ML) models, 10% of the data was divided into validation sets and 90% into discovery sets. The validation dataset was used for hyperparameter optimization and feature selection stages, while the discovery dataset was used to measure the performance of the models. A 10-fold cross-validation technique was used to evaluate the performance of ML models. Biomarker discovery was performed using minimum redundancy maximum relevance (mRMR), Boruta, and explainable boosting machine (EBM). The predictive proposed framework compares the results of eXtreme Gradient Boosting (XGBoost), natural gradient boosting for probabilistic prediction (NGBoost), and EBM models in determining the DR subclass. The hyperparameters of the models were optimized using Bayesian optimization. Combining EBM feature selection with XGBoost, the optimal model achieved (91.25 ± 1.88) % accuracy, (89.33 ± 1.80) % precision, (91.24 ± 1.67) % recall, (89.37 ± 1.52) % F1-Score, and (97.00 ± 0.25) % the area under the ROC curve (AUROC). According to the EBM explanation, the six most important biomarkers in determining the course of DR were tryptophan (Trp), phosphatidylcholine diacyl C42:2 (PC.aa.C42.2), butyrylcarnitine (C4), tyrosine (Tyr), hexadecanoyl carnitine (C16) and total dimethylarginine (DMA). The identified biomarkers may provide a better understanding of the progression of DR, paving the way for more precise and cost-effective diagnostic and treatment strategies.

Optimizing Microaneurysm Detection for Early Diagnosis of Diabetic Retinopathy with Enhanced Fundus Images

Diabetic Retinopathy (DR) stands as the primary cause of visual impairment among individuals in the workforce. Detecting it early through routine screenings and providing accurate treatment has proven effective in preventing vision loss. Fundus photography enables the capture of detailed images of the eye's interior. The automated identification of lesions offers a straightforward and economical approach. This technology aids doctors in decreasing their workload while enhancing accuracy and efficiency. The focus of our research revolves around identifying red lesions, characterized by minuscule dots in fundus images. These dots signify the presence of microaneurysms, which represent an initial stage in diabetic retinopathy detection. Given that microaneurysms serve as symptomatic indicators of diabetic retinopathy, their identification holds crucial importance. Our approach is built upon a highly efficient level set framework, designed to effectively detect the presence of these lesions within the affected images.

VEGFR-2 adhesive nanoprobes reveal early diabetic retinopathy in vivo

Diabetic retinopathy (DR) is a debilitating organ manifestation of diabetes. Absent of early diagnosis and intervention, vision tends to drastically and irreversibly decline. Previously, we showed higher vascular endothelial growth factor receptor 2 (VEGFR-2) expression in diabetic microvessels, and the suitability of this molecule as a biomarker for early DR diagnosis. However, a hurdle to translation remained generation of biodegradable nanoprobes that are sufficiently bright for in vivo detection. Here, an adhesive fluorescent nanoprobe with high brightness was developed using biodegradable materials. To achieve that, a fluorophore with bulky hydrophobic groups was encapsulated in the nanoparticles to minimize fluorophore π-π stacking, which diminishes brightness at higher loading contents. The nanoprobe selectively targeted the VEGFR-2 under dynamic flow conditions. Upon systemic injection, the nanoprobes adhered in the retinal microvessels of diabetic mice and were visualized as bright spots in live retinal microscopy. Histology validated the in vivo results and showed binding of the nanoprobes to the microvascular endothelium and firmly adhering leukocytes. Leukocytes were found laden with nanoprobes, indicating the potential for payload transport across the blood-retinal barrier. Our results establish the translational potential of these newly generated nanoprobes in early diagnosis of DR.

Multiple instance learning based classification of diabetic retinopathy in weakly-labeled widefield OCTA en face images

Diabetic retinopathy (DR), a pathologic change of the human retinal vasculature, is the leading cause of blindness in working-age adults with diabetes mellitus. Optical coherence tomography angiography (OCTA), a functional extension of optical coherence tomography, has shown potential as a tool for early diagnosis of DR through its ability to visualize the retinal vasculature in all spatial dimensions. Previously introduced deep learning-based classifiers were able to support the detection of DR in OCTA images, but require expert labeling at the pixel level, a labor-intensive and expensive process. We present a multiple instance learning-based network, MIL-ResNet,14 that is capable of detecting biomarkers in an OCTA dataset with high accuracy, without the need for annotations other than the information whether a scan is from a diabetic patient or not. The dataset we used for this study was acquired with a diagnostic ultra-widefield swept-source OCT device with a MHz A-scan rate. We were able to show that our proposed method outperforms previous state-of-the-art networks for this classification task, ResNet14 and VGG16. In addition, our network pays special attention to clinically relevant biomarkers and is robust against adversarial attacks. Therefore, we believe that it could serve as a powerful diagnostic decision support tool for clinical ophthalmic screening.

Lo screening nella retinopatia diabetica

Diabetic retinopathy (DR) is the main ocular complication of diabetes. Among all the causes of blindness, it is the only one whose global pre-valence has increased over the past 30 years. DR prevalence is expected to increase in the upcoming years. Therefore, it is essential to establish cost-effective screening programs for early diagnosis of DR, which using new imaging systems, telemedicine and artificial intelligence prevent its evolution to advanced sight-threatening stages.

Early diagnosis of diabetic retinopathy using unsupervised learning

Early diagnosis of diabetic retinopathy using unsupervised learning

Bayesian deep learning methods applied to diabetic retinopathy disease: a review

Diabetic retinopathy (DR) is a complication of diabetes that cause retinal damage; therefore, it is a leading cause of blindness. However, early detection of this disease can dramatically reduce the risk of vision loss. The main problem of early DR detection is that the manual diagnosis by ophthalmology is time-consuming, expensive, and prone to misdiagnosis. Deep learning (DL) models have aided in the early diagnosis of DR, and DL is now frequently utilized in DR detection and classification. The main issues with classical DL models is that they are incapable to quantify the uncertainty in the models, thus they are prone to make wrong decisions in complex cases. However, Bayesian deep learning (BDL) models have recently evolved as unified probabilistic framework to integrate DL and Bayesian models to provides an accurate framework to identify all sources of uncertainty in the model. This paper introduces BDL and most recent research that used BDL approaches to treat diabetic retinopathy are reviewed and discussed. A thorough comparison of the existing Bayesian approaches in this topic is also presented. In addition, available datasets for the fundus retina, which is often employed in DR, are provided and reviewed.

A 3-miRNA Risk Scoring Signature in Early Diabetic Retinopathy

The aim of our study was to investigate a comprehensive profile of streptozotocin (STZ)-induced early diabetic retinopathy (DR) mice to identify a risk scoring signature based on micorRNAs (miRNAs) for early DR diagnosis. RNA sequencing was performed to obtain the gene expression profile of retinal pigment epithelium (RPE) in early STZ-induced mice. Differentially expressed genes (DEGs) were determined with log2|fold change (FC)| > 1 and p value < 0.05. Functional analysis was carried out based on gene ontology (GO), Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analysis and the protein-protein interaction (PPI) network. We predicted the potential miRNAs via online tools and ROC curves were then conducted. Three potential miRNAs with AUC > 0.7 were explored via public datasets and a formula was further established to evaluate DR severity. In total, 298 DEGs (200 up-regulating and 98 down-regulating) were obtained through RNA sequencing. Hsa-miR-26a-5p, hsa-miR-129-2-3p and hsa-miR-217 were three predicted miRNAs with AUC > 0.7, suggesting their potential to distinguish healthy controls from early DR. The formula of DR severity score = 19.257 - 0.004 × hsa-miR-217 + 5.09 × 10-5 × hsa-miR-26a-5p - 0.003 × hsa-miR-129-2-3p was established based on regression analysis. In the present study, we investigated the candidate genes and molecular mechanisms based on RPE sequencing in early DR mice models. Hsa-miR-26a-5p, hsa-miR-129-2-3p and hsa-miR-217 could work as biomarkers for early DR diagnosis and DR severity prediction, which was beneficial for DR early intervention and treatment.

Diabetic Retinopathy: Soluble and Imaging Ocular Biomarkers.

Diabetic retinopathy (DR), the most common microvascular complication of diabetes mellitus, represents the leading cause of acquired blindness in the working-age population. Due to the potential absence of symptoms in the early stages of the disease, the identification of clinical biomarkers can have a crucial role in the early diagnosis of DR as well as for the detection of prognostic factors. In particular, imaging techniques are fundamental tools for screening, diagnosis, classification, monitoring, treatment planning and prognostic assessment in DR. In this context, the identification of ocular and systemic biomarkers is crucial to facilitate the risk stratification of diabetic patients; moreover, reliable biomarkers could provide prognostic information on disease progression as well as assist in predicting a patient's response to therapy. In this context, this review aimed to provide an updated and comprehensive overview of the soluble and anatomical biomarkers associated with DR.

Diagnosis of Diabetic Retinopathy

Abstract: A diabetes-related eye condition called diabetic retinopathy (DR) can cause blindness or blurred vision. The risk of vision loss can be significantly reduced with early DR diagnosis and treatment. Due to the heterogeneity of retinal alterations between disease stages and the significant proportion of misdiagnosed and untreated DR cases, such diagnosis is a timeconsuming and challenging endeavour. A transfer learning method is used to speed up the process while several preprocessing algorithms are used to increase accuracy. The fundus image collection from the online Kaggle datasets was used in our experiment.

Deep feature extraction of retinal images for diabetic retinopathy enhanced diagnosis

Most people of working age groups globally around are affected by diabetic retinopathy, which is the most common cause of preventable vision impairment. Recent studies have improved our comprehension of the need for more effective and affordable methods of identifying, managing, diagnosing, and treating retinal disease in clinical eye care practice. The development of computer-aided diagnosis tools must consider the significance of diabetic retinopathy screening programmes and the challenge of obtaining a valid early diagnosis of diabetic retinopathy at an affordable price. Retinal image analysis using computer-aided disease diagnosis may make it easier to screen the population for diabetes mellitus on a mass scale and may also assist physicians to make better use of their time. The creation of image processing models using big data analysis may run into many technical challenges, including the visualization analysis needed by image processing technologies, the semantic expression, and storage of large image samples, the complexity of the algorithms needed for feature extraction, recognition, and prediction analysis using big data, as well as the time and memory requirements.

Minimized Computations of Deep Learning Technique for Early Diagnosis of Diabetic Retinopathy Using IoT-Based Medical Devices.

Diabetes mellitus is the main cause of diabetic retinopathy, the most common cause of blindness worldwide. In order to slow down or prevent vision loss and degeneration, early detection and treatment are essential. For the purpose of detecting and classifying diabetic retinopathy on fundus retina images, numerous artificial intelligence-based algorithms have been put forth by the scientific community. Due to its real-time relevance to everyone's lives, smart healthcare is attracting a lot of interest. With the convergence of IoT, this attention has increased. The leading cause of blindness among persons in their working years is diabetic eye disease. Millions of people live in the most populous Asian nations, including China and India, and the number of diabetics among them is on the rise. To provide medical screening and diagnosis for this rising population of diabetes patients, skilled clinicians faced significant challenges. Our objective is to use deep learning techniques to automatically detect blind spots in eyes and determine how serious they may be. We suggest an enhanced convolutional neural network (ECNN) utilizing a genetic algorithm in this paper. The ECNN technique's accuracy results are compared to those of existing approaches like the K-nearest neighbor approach, convolutional neural network, and support vector machine with the genetic algorithm.

Semi-supervised classification of fundus images combined with CNN and GCN.

Diabetic retinopathy (DR) is one of the most serious complications of diabetes, which is a kind of fundus lesion with specific changes. Early diagnosis of DR can effectively reduce the visual damage caused by DR. Due to the variety and different morphology of DR lesions, automatic classification of fundus images in mass screening can greatly save clinicians' diagnosis time. To alleviate these problems, in this paper, we propose a novel framework-graph attentional convolutional neural network (GACNN). The network consists of convolutional neural network (CNN) and graph convolutional network (GCN). The global and spatial features of fundus images are extracted by using CNN and GCN, and attention mechanism is introduced to enhance the adaptability of GCN to topology map. We adopt semi-supervised method for classification, which greatly improves the generalization ability of the network. In order to verify the effectiveness of the network, we conducted comparative experiments and ablation experiments. We use confusion matrix, precision, recall, kappa score, and accuracy as evaluation indexes. With the increase of the labeling rates, the classification accuracy is higher. Particularly, when the labeling rate is set to 100%, the classification accuracy of GACNN reaches 93.35%. Compared with DenseNet121, the accuracy rate is improved by 6.24%. Semi-supervised classification based on attention mechanism can effectively improve the classification performance of the model, and attain preferable results in classification indexes such as accuracy and recall. GACNN provides a feasible classification scheme for fundus images, which effectively reduces the screening human resources.