Diabetic Retinopathy Grading Using ResNet Convolutional Neural Network

This study aimed to evaluate the risk of proliferative diabetic retinopathy (DR) in the fellow eye of an eye with existing proliferative DR. Our DR screening programme database listed 1513 diabetes patients alive at the time of the study. Seventy-six had proliferative DR in one or both eyes. In 28 of the 76 (37%) diabetes patients, proliferative DR was diagnosed in both eyes at the same examination. Another 28 patients developed proliferative DR in the second eye within 5 years of its diagnosis in the first eye, bringing the total number of diabetes patients with proliferative DR in both eyes at 5 years to 56 (74%). Almost all the diabetes patients eventually developed proliferative DR in the second eye. The median duration of diabetes before the development of proliferative retinopathy was 19 years for type 1 and 14 years for type 2 diabetes. Proliferative DR is a bilateral disease. Diabetes patients with proliferative DR in one eye are at high risk of developing neovascularization in the second eye and close follow-up is recommended.

BackgroundWith the diabetes mellitus (DM) prevalence increasing annually, the human grading of retinal images to evaluate DR has posed a substantial burden worldwide. SmartEye is a recently developed fundus image processing and analysis system with lesion quantification function for DR screening. It is sensitive to the lesion area and can automatically identify the lesion position and size. We reported the diabetic retinopathy (DR) grading results of SmartEye versus ophthalmologists in analyzing images captured with non-mydriatic fundus cameras in community healthcare centers, as well as DR lesion quantitative analysis results on different disease stages.MethodsThis is a cross-sectional study. All the fundus images were collected from the Shanghai Diabetic Eye Study in Diabetics (SDES) program from Apr 2016 to Aug 2017. 19,904 fundus images were acquired from 6013 diabetic patients. The grading results of ophthalmologists and SmartEye are compared. Lesion quantification of several images at different DR stages is also presented.ResultsThe sensitivity for diagnosing no DR, mild NPDR (non-proliferative diabetic retinopathy), moderate NPDR, severe NPDR, PDR (proliferative diabetic retinopathy) are 86.19, 83.18, 88.64, 89.59, and 85.02%. The specificity are 63.07, 70.96, 64.16, 70.38, and 74.79%, respectively. The AUC are PDR, 0.80 (0.79, 0.81); severe NPDR, 0.80 (0.79, 0.80); moderate NPDR, 0.77 (0.76, 0.77); and mild NPDR, 0.78 (0.77, 0.79). Lesion quantification results showed that the total hemorrhage area, maximum hemorrhage area, total exudation area, and maximum exudation area increase with DR severity.ConclusionsSmartEye has a high diagnostic accuracy in DR screening program using non-mydriatic fundus cameras. SmartEye quantitative analysis may be an innovative and promising method of DR diagnosis and grading.

To explore the association of choroidal thickness variations in type 2 diabetes mellitus (T2DM) patients with diabetic retinopathy (DR) at different stages. A total of 161 patients with T2DM were included in this study, from October 2012 to June 2014. According to Early Treatment Diabetic Retinopathy Study (ETDRS) criteria, the patients were divided into 5 groups: non-DR without diabetic macular edema (DME) group(DR-/DME- group, 45 eyes), nonproliferative diabetic retinopathy (NPDR) without DME group(NPDR+/DME- group, 58 eyes), proliferative diabetic retinopathy (PDR) without DME group (PDR+/DME- group, 12 eyes), NPDR with DME group (NPDR+/DME+ group, 41 eyes), PDR with DME group (PDR+/DME+ group, 5 eyes). Meanwhile, 60 normal subjects were enrolled as the control group. All study subjects received optical coherence tomography enhanced depth imaging (EDI-OCT) examination to detect and compare subfoveal choroidal thickness at different stages of DR. Mean SFCT was (271 ± 36), (270 ± 35), (262 ± 38), (244 ± 36), (229 ± 35) µm respectively in control, DR-/DME-, NPDR+/DME-, PDR+/DME-, NPDR+/DME+ groups. The SFCTs of PDR+/DME- and NPDR+/DME+ group were statistically lower than that of control group (P=0.004, P=0.001). The SFCT of PDR+/DME- group was lower than that of DR-/DME- group (P=0.003), and there was also a significant difference of SFCT between NPDR+/DME+ and NPDR+/DME- group (P=0.001). There was linear correlation between SFCT and the logMAR best-corrected visual acuity (r=0.397, P<0.01), but the SFCT was independent of diabetic duration, fasting blood glucose, HbA1c, axial length, diastolic blood pressure (DBP) and systolic blood pressure SBP (r=-0.024, 0.159, 0.089, 0.036, 0.143, 0.057, all P>0.05). There was no significant difference of SFCT among different DME types (F=0.071, P>0.05). The SFCT decreased with increasing severity of DR. To monitor the SFCT in T2DM patients may be helpful to evaluate the severity of DR and provide a new treatment conception.

PurposeTo characterize the natural course of diabetic retinopathy (DR) in contemporary clinical practice.Patients and MethodsThis was a retrospective analysis of US claims data collected between January 1, 2006, and April 30, 2017. Patients aged ≥18 years with continuous medical and prescription insurance coverage for 18 months before DR diagnosis (index date) and for a follow-up period of 5 years were included (N=14,490). The time and risk of progressing to severe nonproliferative DR (NPDR) or proliferative DR (PDR) and of developing diabetic macular edema (DME) were evaluated over 5 years in patients stratified by DR severity at initial diagnosis.ResultsThe estimated probability of progressing to severe NPDR or PDR within 5 years of diagnosis was 17.6% for patients with moderate NPDR versus 5.8% for mild NPDR. The probability of developing DME within 5 years was 62.6%, 44.6%, and 28.4% for patients diagnosed with severe NPDR, moderate NPDR, and PDR, respectively, versus 15.6% for mild NPDR. Among those observed to progress, median time to severe NPDR or PDR was approximately 2.0 years in patients with moderate NPDR, whereas median time to DME was approximately 0.5 years in patients with severe NPDR, 1.3 years in moderate NPDR, and 1.6 years in PDR. Relative to mild NPDR, adjusted hazard ratios (95% confidence interval) for progression to severe NPDR or PDR within 5 years were 3.12 (2.61–3.72) in patients with moderate NPDR, and for incident DME were 5.92 (5.13–6.82), 3.54 (3.22–3.91), and 1.96 (1.80–2.14) in patients with severe NPDR, moderate NPDR, and PDR, respectively.ConclusionThe risk of DR progression and DME over 5 years was highest among patients diagnosed with moderate and severe NPDR, respectively. Our findings reinforce the importance of close monitoring for these patients to avoid unobserved disease progression toward PDR and/or DME.

Objective To investigate the distribution patterns of diabetic macular edema (DME) based on optical coherence tomography (OCT), and explore its correlation with diabetic retinopathy (DR) stages and systemic factors. Methods A total of 135 patients (242 eyes) with type 2 diabetes were included in this retrospective study. There were 75 males (138 eyes) and 60 females (104 eyes), the ages were from 29 to 83 years, with an average age of (58.8±11.1) years. The general information such as height, weight, smoking history and blood glucose [such as glycosylated hemoglobin (HbA1c)], blood pressure, blood lipid, 24 hours urine protein and other examinations were collected. The diagnosis of DR and DME were made, and the staging of DR and typing of DME were performed based on fundus color imaging and OCT. DR were divided into mild non-proliferative DR (NPDR), moderate NPDR, severe NPDR and proliferative DR (PDR). DME were categorized into 4 types including sponge-like retinal swelling (SME), cystoid macular edema (CME), serous retinal detachment (SRD) and posterior hyaloid traction (PHT). The correlation between DME types and DR staging were analyzed by χ2 test and Fisher exact test. Multivariate logistic regression analysis was used to analyze the correlation between DME types and systemic factors. Results In 242 DR eyes the proportions of mild, moderate, severe NPDR and PDR were 30.99%, 32.64%, 23.14% and 13.23%, respectively. There were 199 eyes (82.23%) with DME. There were statistically significant differences in the proportion of DME in different stages of DR (χ2=21.077, P<0.01). In the 199 eyes with DME, There were 165 eyes (68.18%) of SME, 22 eyes (9.09%) of CME, 7 eyes (2.89%) of SRD and 5 eyes (2.07%) of PHT. The distribution of DME patterns in different stages of DR was statistically significant (χ2=156.273, P<0.01). Logistic regression analysis showed that the duration of diabetes, HbA1c and macroalbuminuria were independent risk factors for DME [odds ratio (OR)= 1.090, 1.510, 4.123; P<0.05], and were also independent for SME (OR=1.092, 1.445, 3.942; P<0.05); HbA1c was an independent risk factor for SRD (OR=2.337, P<0.05). Conclusions There are differences in the distribution of different DME types in each stage of DR. The duration of diabetes, HbA1c and macroalbuminuria were independent risk factors for DME and SME, and macroalbuminuria and HbA1c for CME and SRD. Key words: Macular edema/diagnosis; Risk factors; Diabetic retinopathy/diagnosis; Tomography, optical coherence

Diabetic Retinopathy caused by Diabetes Mellitus is a major vision threatening condition across the global population. Early detection and grading of diabetic retinopathy are pivotal to avoiding the associated vision impairment, and retinal image analysis serves as an effective method for the screening process. Analysing retinal images manually to detect diabetic retinopathy and grade them is a time-consuming pro cess that necessitates the involvement of experts to perform the classification. Computer-aided diagnosis of diabetic retinopathy through retinal image analysis is an effective tool to reduce the time involved in the screening process. This study proposed a convolutional neural network based Selective Feature Map Fusion architecture that leverages the fused feature maps of ResNet50 and EfficientNetV2L networks for the detection and grading of diabetic retinopathy. Feature set of ResNet50 and EfficientNetV2L captures different aspects of the underlying lesion distribution, resulting in a more comprehensive representation of the features. This approach helps the model generalize better to unseen data by incorporating diverse data aspects, thereby preventing overfitting to a specific feature set and developing a model that is more robust to noise. Selective feature fusion helps to reduce the computational overhead of the the feature fusion computation. Comparative analysis against the ResNet50 and EfficientNetV2L models individually revealed the superior performance of the fused model in both detection and grading of diabetic retinopathy. The proposed model attained an excellent level of accuracy of 95.2% in diabetic retinopathy detection with a sensitivity of 95.2% and a specificity of 95.2%, when tested on the IDRiD dataset. The model achieved a classification accuracy of 92% for diabetic retinopathy grading with a sensitivity of 88%, a specificity of 98% and an F1-score of 0.89. Moreover, the fused framework attained superior performance when compared against existing methodologies on the IDRiD dataset. The developed model exhibited robust performance also in the DeepDRiD dataset, and it can be employed successfully in the screening processes for the diagnosis and grading of diabetic retinopathy.

Diabetic retinopathy is the most severe ocular complication of diabetes and may lead to visual disability and blindness. Proliferative diabetic retinopathy (PDR) is characterized by ischemia-induced neovascularization with associated complications. An association was established between the presence of PDR, cardiovascular disease, and mortality among patients with type 1 diabetes mellitus and type 2 diabetes mellitus in epidemiological studies. However, the mechanism underlying increased cardiovascular risk in patients with PDR is still unknown. In recent years, a group of miRNAs has been linked to the pathology of diabetes mellitus. Besides, miRNAs in biofluids such as serum have been suggested as potential minimally invasive biomarkers of diabetes and vascular complications. This was a prospective study that recruited 40 human subjects: 10 healthy subjects, 10 with diabetes but without retinopathy (NDR), 10 with diabetic non-proliferative retinopathy (NPDR), and 10 with proliferative diabetic retinopathy (PDR). To examine whether serum miRNAs show altered levels at different stages of diabetic retinopathy, seven specific miRNA candidates (miR-126-3p, miR-130a-3p, miR-21-1, let-7f-5p, miR-122, miR-30c and miR-451a) were measured by qRT-PCR in RNA isolated from sera of all subjects. miR-122 levels increased in parallel with retinopathy severity: from healthy controls to NDR and from NDR to NPDR. However, when the disease progressed to PDR a marked decrease in miR-122 level was noted. This decrease was significant both compared to NPDR samples (p = 0.016) and to all non-PDR samples (p = 0.0002). Additionally, a positive trend was observed comparing miR-122 levels and the number of endothelial progenitor cells in the sera of all subjects. A significant increase in miR-122 was found in patients with diabetic retinopathy that may be related to its role in preventing angiogenesis and proliferation. The dramatic decline in patients with PDR may represent an inhibition or exhaustion of the anti-angiogenic anti-proliferative defense system. Further studies are needed to understand whether miRNA-122 has a role in the pathogenesis of diabetic retinopathy.

The aim of this study was to look for a relationship between the - 429 T/C and the - 374 T/A gene polymorphisms of the receptor of advanced glycation end products (RAGE) gene and the development of diabetic retinopathy in Caucasians with type 2 diabetes. One hundred and sixteen subjects with type 2 diabetes and diabetic retinopathy were compared to 70 diabetic subjects without diabetic retinopathy. Additionally, 76 subjects with proliferative diabetic retinopathy (a subgroup of diabetic retinopathy) were compared to 70 diabetic subjects without diabetic retinopathy. The - 429 T/C and the - 374 T/A RAGE gene distributions in patients with diabetic retinopathy (- 429 T/C: CC 0.9%, TC 25.9%, TT 73.2%; - 374 T/A: AA 9.5%, TA 47.4%, TT 43.1%) were not significantly different from those of diabetic subjects without retinopathy (- 429 T/C: CC 0%, TC 25.7%, TT 74.3%; - 374 T/A: AA 15.7%, TA 42.9%, TT 41.4%). Moreover, the - 429 T/C and the - 374 T/A RAGE gene distributions in patients with proliferative diabetic retinopathy were not statistically significantly different from those in diabetic subjects without retinopathy. Our study failed to demonstrate an association between either - 429 T/C or - 374 T/A gene polymorphism of the RAGE gene and diabetic retinopathy in Caucasians with type 2 diabetes. Additionally, we failed to demonstrate an association between either - 429 T/C or - 374 T/A gene polymorphism of the RAGE gene and proliferative diabetic retinopathy in Caucasians with type 2 diabetes.

Diabetic Mellitus is the most familiar disease around the globe. Long prevalence of diabetes causes several problems related to health. The most common issue is Diabetic Retinopathy (DR). Diabetic retinopathy is a situation in which the vessels inside the retina are vandalized, leaking harmful substances and fluids in the surrounding tissue resulting in hemorrhages, micro aneurysms in the eye and further into partial or complete vision loss. This disease if treated in the early stage can help to prevent vision loss, but since it takes time for diagnosis and there is a shortage of ophthalmologists' patients suffer vision loss even before diagnosis. Hence, early detection of DR may help in reducing the problem. Therefore, in this paper we investigate various approaches to understand the process of detecting Diabetic Retinopathy as accurately as possible and classifying them into different grades of treatable DR (NPDR) namely LO, L1 DR, L2 DR and Proliferate DR (PDR) using Deep Learning and Image Processing techniques also making some improvisations on the same to enhance the capability of other existing systems.

Diabetes mellitus, together with its complications, has been increasing in prevalence worldwide. Its complications include cardiovascular disease (e.g., myocardial infarction, stroke), neuropathy, nephropathy, and eye complications (e.g., glaucoma, cataracts, retinopathy, and macular edema). In patients with either type 1 or type 2 diabetes mellitus, diabetic retinopathy is the leading cause of visual impairment or blindness. It is characterized by progressive changes in the retinal microvasculature. The progression from nonproliferative diabetic retinopathy to a more advanced stage of moderate to severe nonproliferative diabetic retinopathy and proliferative diabetic retinopathy occurs very quickly after diagnosis of mild nonproliferative diabetic retinopathy. The etiology of diabetic retinopathy is unclear, and present treatments have limited effectiveness. Currently diabetic retinopathy can only be diagnosed by a trained specialist, which reduces the population that can be examined. A screening biomarker of diabetic retinopathy with high sensitivity and specificity would aid considerably in identifying those individuals in need of clinical assessment and treatment. The majority of the studies reviewed identified specific microRNAs in blood serum/plasma able to distinguish diabetic patients with retinopathy from those without retinopathy and for the progresion of the disease from nonproliferative diabetic retinopathy to proliferative diabetic retinopathy. In addition, certain microRNAs in vitreous humor were dysregulated in proliferative diabetic retinopathy compared to controls. A very high percentage of patients with diabetic retinopathy develop Alzheimer’s disease. Thus, identifying diabetic retinopathy by measurement of suitable biomarkers would also enable better screening and treatment of those individuals at risk of Alzheimer’s disease.

Introduction: Diabetic retinopathy (DR) remains a visually debilitating disease and is commonly classified according to its severity as non-proliferative DR (NPDR) or proliferative DR (PDR). Those suffering from PDR tend to have worse vascular complications and prognosis. Platelets exposed by vasculopathy caused by DR maybe activated to try to maintain haemostasis. This activity can be illustrated by the mean platelet component (MPC). Therefore, by MPC monitoring we may be able to predict the progression from NPDR into PDR.Purpose: To investigate the difference of MPC in patients with NPDR and PDR.Study design: Cross-sectional.Materials and methods: This study involved 71 DR patients. Preliminary data regarding the patients’ demographic characteristics, diabetes history, related diseases, medication history, and general eye examination were recorded. Fundus photographs were taken after dilating eyedrops and DR was graded by an ophthalmologist. The patients were grouped into NPDR and PDR. Mean platelet component was analyzed using the automatic hematology analyzer ADVIA 120.Results: Mean platelet component (MPC) was 26.69 g/dl (± 1.79) and 25.52 g/dl (± 1.20) in the NPDR and PDR group, respectively (p = 0.002), but was not clinically significant. In depth analysis into the DR grades differed significantly between mild NPDR and high-risk PDR (p = 0.015), and moderate NPDR and high-risk PDR (p = 0.024). Using our definition of mild DR (mild and moderate NPDR) and severe DR (high-risk and advanced PDR), there was a significant difference with mean MPC of 27.01 g/dl (± 1.64) and 25.31 g/dl (± 1.22), respectively (p = 0.001). The proportion of activated platelets was also higher in severe DR. Negative correlations were found between MPC with duration of DM (r = -0.333; p = 0.004) and MPC with systolic blood pressure (r = -0.241; p = 0.043).Conclusion: There was a significant difference in MPC between NPDR and PDR, but the results should be interpreted carefully. Further analysis between the mild and severe form of DR strengthened this finding.

Ranibizumab for Retinal Neovascularization

Influence of HLA-DR Phenotype and Myopia on the Risk of Nonproliferative and Proliferative Diabetic Retinopathy

Rapid, bloody, and blinding diabetic retinopathy

BackgroundAtherosclerosis cardiovascular disease (ASCVD) is the main cause of morbidity and mortality in type 2 diabetes mellitus (T2DM). As most diabetic patients with ASCVD are asymptomatic, it is most neglected in clinical practice. For this reason, identifying high-risk ASCVD population with intensified treatment is very important. In recent years, the relationship between diabetic retinopathy (DR) and ASCVD has caused much academic concern, but the results are inconsistent. Moreover, whether all grades of DR increase the risk of ASCVD remains controversial. Most importantly, very few data can be found in China.ObjectiveOur aim is to discuss whether all grades of DR increase the risk of ASCVD after adjustment for the traditional cardiovascular risk factors and to assess the independent contribution of DR to cardiovascular events in patients with T2DM, hoping to provide more evidence for early identification of ASCVD.Research design and methodsA total of 425 T2DM patients with complete physical and biochemical data were included in the study. The grade of DR was assessed with two 45 color digital retinal images. Based on the presence of history of ASCVD, 425 T2DM patients were divided into 2 groups: ASCVD group and non-ASCVD group.ResultsASCVD patients were older and had a significantly higher fasting plasma glucose (FPG) and glycated haemoglobin (HbA1c) and proportion of history of ASCVD. At the same time, they were more likely to be females, and had lower level of alcohol and calculated glomerular filtration rate (eGFR) than non-ASCVD patients. Their trend to develop DR with ASCVD was significantly higher than patients with non-ASCVD (χ2 = 5.805, P = 0.016). DR was an independent statistical indicator of the presence of ASCVD [odds ratio (OR) (95% CI): 2.321 (1.152–4.678), P = 0.018]. Furthermore, when DR was divided into non-proliferative retinopathy (NPDR) and proliferative retinopathy (PDR) according to its severity, only PDR was significantly associated with incident ASCVD [OR (95% CI): 8.333 (1.813–38.304), P = 0.006]. After adjusting for traditional ASCVD risk factors, such an association still existed [OR (95% CI): 7.466 (1.355–41.137), P = 0.021].ConclusionDR associates strongly with ASCVD in the Chinese population with T2DM. With the increasing severity of DR, the risk of ASCVD also increases. After adjustment for traditional risk factors, PDR is still an independent risk marker for ASCVD.