Screening Tool For Diabetic Retinopathy Research Articles

Predicting diabetic retinopathy stage using Siamese Convolutional Neural Network

ABSTRACT Diabetic retinopathy (DR) is considered to be a leading cause of blindness in people aged 16 to 64 years, affecting around 40% of the population diagnosed with diabetes mellitus (DM). DR is usually identified through retinal image analysis, and in some countries, such as Brazil, such diagnosis is hindered by limited access to specialized care, leading to lengthy waits for ophthalmological evaluations. This scenario makes the Brazilian Diabetic Society’s annual DR check recommendation impractical for many. To address this gap, our study introduces a novel Siamese Convolutional Neural Network (SCNN) for DR prediction, usable by primary care professionals. Our SCNN analyzes pairs of eye fundus images and employs shared weights in its layers to extract essential features, facilitating similarity measurement between neural network outputs. Despite challenges with limited and imbalanced datasets, our SCNN showed effectiveness. We tested it on four datasets (IDRiD, APTOS, Messidor-1, DIARETDB0), with accuracy ranging from 67.23% (APTOS) to 96.85% (DIARETDB0). Compared to other methods, our approach consistently excelled, especially in recall analysis. These results suggest that deep learning via Siamese networks is likely to be a viable and potential DR screening tool.

The Utility of ChatGPT in Diabetic Retinopathy Risk Assessment: A Comparative Study with Clinical Diagnosis.

To evaluate the ability of an artificial intelligence (AI) model, ChatGPT, in predicting the diabetic retinopathy (DR) risk. This retrospective observational study utilized an anonymized dataset of 111 patients with diabetes who underwent a comprehensive eye examination along with clinical and biochemical assessments. Clinical and biochemical data along with and without central subfield thickness (CST) values of the macula from OCT were uploaded to ChatGPT-4, and the response from the ChatGPT was compared to the clinical DR diagnosis made by an ophthalmologist. The study assessed the consistency of responses provided by ChatGPT, yielding an Intraclass Correlation Coefficient (ICC) value of 0.936 (95% CI, 0.913-0.954, p < 0.001) (with CST) and 0.915 (95% CI, 0.706-0.846, p < 0.001) (without CST), both situations indicated excellent reliability. The sensitivity and specificity of ChatGPT in predicting the DR cases were evaluated. The results revealed a sensitivity of 67% with CST and 73% without CST. The specificity was 68% with CST and 54% without CST. However, Cohen's kappa revealed only a fair agreement between ChatGPT predictions and clinical DR status in both situations, with CST (kappa = 0.263, p = 0.005) and without CST (kappa = 0.351, p < 0.001). This study suggests that ChatGPT has the potential of a preliminary DR screening tool with further optimization needed for clinical use.

Development of LuxIA, a Cloud-Based AI Diabetic Retinopathy Screening Tool Using a Single Color Fundus Image.

Diabetic retinopathy (DR) is the leading cause of vision impairment in working-age adults. Automated screening can increase DR detection at early stages at relatively low costs. We developed and evaluated a cloud-based screening tool that uses artificial intelligence (AI), the LuxIA algorithm, to detect DR from a single fundus image. Color fundus images that were previously graded by expert readers were collected from the Canarian Health Service (Retisalud) and used to train LuxIA, a deep-learning-based algorithm for the detection of more than mild DR. The algorithm was deployed in the Discovery cloud platform to evaluate each test set. Sensitivity, specificity, accuracy, and area under the receiver operating characteristic curve were computed using a bootstrapping method to evaluate the algorithm performance and compared through different publicly available datasets. A usability test was performed to assess the integration into a clinical tool. Three separate datasets, Messidor-2, APTOS, and a holdout set from Retisalud were evaluated. Mean sensitivity and specificity with 95% confidence intervals (CIs) reached for these three datasets were 0.901 (0.901-0.902) and 0.955 (0.955-0.956), 0.995 (0.995-0.995) and 0.821 (0.821-0.823), and 0.911 (0.907-0.912) and 0.880 (0.879-0.880), respectively. The usability test confirmed the successful integration of LuxIA into Discovery. Clinical data were used to train the deep-learning-based algorithm LuxIA to an expert-level performance. The whole process (image uploading and analysis) was integrated into the cloud-based platform Discovery, allowing more patients to have access to expert-level screening tools. Using the cloud-based LuxIA tool as part of a screening program may give diabetic patients greater access to specialist-level decisions, without the need for consultation.

Diabetic Retinopathy Classification Using Deep Learning

One of the main causes of adult blindness and a frequent consequence of diabetes is diabetic retinopathy (DR). To avoid visual loss, DR must be promptly identified and classified. In this article, we suggest an automated DR detection and classification method based on deep learning applied to fundus pictures. The suggested technique uses transfer learning for classification. On a dataset of 3,662 fundus images with real-world DR severity labels, we trained and validated our model. According to our findings, the suggested technique successfully detected and classified DR with an overall accuracy of 78.14%. Our model fared better than other recent cutting-edge techniques, illuminating the promise of deep learning-based strategies for DR detection and management. Our research indicates that the suggested technique may be employed as a screening tool for DR in a clinical environment, enabling early illness diagnosis and prompt treatment.

Google Lens: A potential cost-effective screening tool for diabetic retinopathy

Background: Diabetic retinopathy (DR) is a major, sight-threatening complication of diabetes mellitus. Blindness from DR can be prevented by successful and proactive screening. However, DR is screened in less than half of the patients because of barriers in availability, affordability, accessibility, and awareness. Although artificial intelligence (AI)-based algorithms are being evaluated for DR screening, they have limitations of infrastructure, accessibility, training, and manpower cost. Therefore, simpler and more practical DR screening tools should be explored.&#x0D; Hypothesis: Google Lens, an easily available, vision- and AI-based application in most smartphones, is a potential tool for cost-effective DR screening. It recognises images through a visual analysis based on neural networking. Thus, it can recognize retinal disorders, such as DR, in images. The development and adoption of Google Lens-based DR screening would have several advantages over the conventional hospital/specialist/healthcare facility-based approach, including widespread accessibility, acceptable accuracy, reduction in the direct cost of healthcare for patients with diabetes mellitus, and active patient participation in self-care.&#x0D; Conclusions: DR screening, detection, and grading using Google Lens is a feasible and effective option. Despite current limitations, it could transform DR screening from a costly, hospital- and expert-based method to a cost-effective, self-applicable, and home-based one. However, diagnostic accuracy studies comparing the index test with Google Lens-based screening are required to determine the usability and validity of this proposed screening tool for DR.

An online diabetic retinopathy screening tool for patients with type 2 diabetes.

To develop a useful diabetic retinopathy (DR) screening tool for patients with type 2 diabetes mellitus (T2DM). A DR prediction model based on the Logistic regression algorithm was established on the development dataset containing 778 samples (randomly assigned to the training dataset and the internal validation dataset at a ratio of 7:3). The generalization capability of the model was assessed using an external validation dataset containing 128 samples. The DR risk calculator was developed through WeChat Developer Tools using JavaScript, which was embedded in the WeChat Mini Program. The model revealed risk factors (duration of diabetes, diabetic nephropathy, and creatinine level) and protective factors (annual DR screening and hyperlipidemia) for DR. In the internal and external validation, the recall ratios of the model were 0.92 and 0.89, respectively, and the area under the curve values were 0.82 and 0.70, respectively. The DR screening tool integrates education, risk prediction, and medical advice function, which could help clinicians in conducting DR risk assessments and providing recommendations for ophthalmic referral to increase the DR screening rate among patients with T2DM.

Detection of Diabetic Retinopathy from Ultra-Widefield Scanning Laser Ophthalmoscope Images: A Multicenter Deep Learning Analysis

To develop a deep learning (DL) system that can detect referable diabetic retinopathy (RDR) and vision-threatening diabetic retinopathy (VTDR) from images obtained on ultra-widefield scanning laser ophthalmoscope (UWF-SLO). Observational, cross-sectional study. A total of 9392 UWF-SLO images of 1903 eyes from 1022 subjects with diabetes from Hong Kong, the United Kingdom, India, and Argentina. All images were labeled according to the presence or absence of RDR and the presence or absence of VTDR. Labeling was performed by retina specialists from fundus examination, according to the International Clinical Diabetic Retinopathy Disease Severity Scale. Three convolutional neural networks (ResNet50) were trained with a transfer-learning procedure for assessing gradability and identifying VTDR and RDR. External validation was performed on 4 datasets spanning different geographical regions. Area under the receiver operating characteristic curve (AUROC); area under the precision-recall curve (AUPRC); sensitivity, specificity, and accuracy of the DL system in gradability assessment; and detection of RDR and VTDR. For gradability assessment, the system achieved an AUROC of 0.923 (95% confidence interval [CI], 0.892-0.947), sensitivity of 86.5% (95% CI, 77.6-92.8), and specificity of 82.1% (95% CI, 77.3-86.2) for the primary validation dataset, and >0.82 AUROCs, >79.6% sensitivity, and >70.4% specificity for the geographical external validation datasets. For detecting RDR and VTDR, the AUROCs were 0.981 (95% CI, 0.977-0.984) and 0.966 (95% CI, 0.961-0.971), with sensitivities of 94.9% (95% CI, 92.3-97.9) and 87.2% (95% CI, 81.5-91.6), specificities of 95.1% (95% CI, 90.6-97.9) and 95.8% (95% CI, 93.3-97.6), and positive predictive values (PPVs) of 98.0% (95% CI, 96.1-99.0) and 91.1% (95% CI, 86.3-94.3) for the primary validation dataset, respectively. The AUROCs and accuracies for detecting both RDR and VTDR were >0.9% and >80%, respectively, for the geographical external validation datasets. The AUPRCs were >0.9, and sensitivities, specificities, and PPVs were >80% for the geographical external validation datasets for RDR and VTDR detection. The excellent performance achieved with this DL system for image quality assessment and detection of RDR and VTDR in UWF-SLO images highlights its potential as an efficient and effective diabetic retinopathy screening tool.

Clinical validation of an artificial intelligence-based diabetic retinopathy screening tool for a national health system

ObjectiveTo evaluate the accuracy and validity of an automated diabetic retinopathy (DR) screening tool (DART, TeleDx, Santiago, Chile) that uses artificial intelligence to analyze ocular fundus photographs for potential implementation in the national Chilean DR screening programme.MethodThis was an observational study of 1123 diabetic eye exams using a validation protocol designed by the commission of the Chilean Ministry of Health personnel and retina specialists.ResultsReceiver operating characteristic (ROC) analysis indicated a sensitivity of 94.6% (95% CI: 90.9–96.9%), specificity of 74.3% (95% CI: 73.3–75%), and negative predictive value of 98.1% (95% CI: 96.8–98.9%) for the automated tool at the optimal operating point for DR screening. The area under the ROC curve was 0.915.ConclusionsThe results of this study suggest that DART is a valid tool that could be implemented in a heterogeneous health network such as the Chilean system.

Automated diabetic retinopathy screening for primary care settings using deep learning

Diabetic Retinopathy (DR) is one of the leading causes of blindness in the United States and other high-income countries. Early detection is key to prevention, which could be achieved effectively with a fully automated screening tool performing well on clinically relevant measures in primary care settings. We have built an artificial intelligence-based tool on a cloud-based platform for large-scale screening of DR as referable or non-referable. In this paper, we aim to validate this tool built using deep learning based techniques. The cloud-based screening model was developed and tested using deep learning techniques with 88702 images from the Kaggle dataset and externally validated using 1748 high-resolution images of the retina (or fundus images) from the Messidor-2 dataset. For validation in the primary care settings, 264 images were taken prospectively from two diabetes clinics in Queens, New York. The images were uploaded to the cloud-based software for testing the automated system as compared to expert ophthalmologists' evaluations of referable DR. Measures used were area under the curve (AUC), sensitivity, and specificity of the screening model with respect to professional graders. The screening system achieved a high sensitivity of 99.21% and a specificity of 97.59% on the Kaggle test dataset with an AUC of 0.9992. The system was also externally validated in Messidor-2, where it achieved a sensitivity of 97.63% and a specificity of 99.49% (AUC, 0.9985). On primary care data, the sensitivity was 92.3% overall (12/13 referable images are correctly identified), and overall specificity was 94.8% (233/251 non-referable images). The proposed DR screening tool achieves state-of-the-art performance among the publicly available datasets: Kaggle and Messidor-2 to the best of our knowledge. The performance on various clinically relevant measures demonstrates that the tool is suitable for screening and early diagnosis of DR in primary care settings.

Dynamic Pupillometry in Type 2 Diabetes: Pupillary Autonomic Dysfunction and the Severity of Diabetic Retinopathy.

PurposeAutonomic dysfunction may precede the microvascular changes that characterise diabetic retinopathy. The aim of this pilot study was to measure and compare pupillometry indices in type 2 diabetes (T2DM) patients – with and without diabetic retinopathy – and in healthy, age-matched controls.MethodsTwo hundred and eleven participants with T2DM aged 45–80 years were recruited from Dunedin Hospital Eye Department, Dunedin, New Zealand. They were categorised into three groups – no, mild/moderate, or severe diabetic retinopathy. Seventy age-matched, diabetes screen negative control participants were recruited from the Dunedin city community. Dynamic pupillometry was performed using an infrared pupillometer. The main outcome measures were maximum constriction velocity, average constriction velocity, absolute constriction amplitude, relative reflex amplitude, average dilation velocity and 75% re-dilation (recovery) time. Outcome measures were compared between study groups using the Kruskal–Wallis nonparametric test (with Dunn’s multiple comparison post-test).ResultsPupillary parasympathetic function differed between groups. Maximum constriction velocity (p <0.001) and average constriction velocity (p <0.001) were slower, and absolute constriction amplitude (p <0.001) and relative reflex amplitude (p <0.05) were lower in the three diabetes groups compared with controls. There were no significant differences in pupillary sympathetic function between the four groups for re-dilation time, but the diabetes groups had significantly slower average dilation velocity times.ConclusionPupillary light reflex is significantly impaired with diabetic retinal neuropathy, before clinically observable signs of diabetic retinopathy. Dynamic pupillometry may be a cheap, clinically relevant test, but sensitivity and specificity need to be determined before utilising as a screening tool for diabetic retinopathy.

Comparison of Telemedicine Screening of Diabetic Retinopathy by Mydriatic Smartphone-Based vs Nonmydriatic Tabletop Camera-Based Fundus Imaging.

To compare dilated smartphone-based imaging with a nonmydriatic, tabletop fundus camera as a teleophthalmology screening tool for diabetic retinopathy (DR). This was a single-institutional, cross-sectional, comparative-instrument study. Fifty-six patients at a safety-net hospital underwent teleophthalmology screening for DR using standard, nonmydriatic fundus photography with a tabletop camera (Nidek NM-1000) and dilated fundus photography using a smartphone camera with lens adapter (Paxos Scope, Verana Health). Masked graders performed standardized photo grading. Quantitative comparisons were performed employing descriptive, κ, Bland-Altman, and receiver operating characteristic analyses. Posterior segment photography was of sufficient quality to grade in 89% of mydriatic smartphone-imaged eyes and in 86% of nonmydriatic tabletop camera-imaged eyes (P = .03). Using the tabletop camera as the reference to detect moderate nonproliferative DR or worse (referral-warranted DR), mydriatic smartphone-acquired photographs were found to be 82% sensitive and 96% specific. Dilated smartphone imaging detected referral-warranted DR in 3 eyes whose tabletop camera imaging did not demonstrate referral-warranted DR. Secondary masked review of medical records for the discordances in referral-warranted status from the two imaging modalities was performed, and it revealed revised sensitivity and specificity values of 95% and 98%, respectively. Overall, there was good agreement between tabletop camera and smartphone-acquired photo grades (κ = 0.91 ± 0.1, P < .001; area under the receiver operating characteristic curve = 0.99, 95% CI, 0.98-1.00). Mydriatic smartphone-based imaging resulted in fewer ungradable photos compared to nonmydriatic table-top camera imaging and detected more patients with referral-warranted DR. Our study supports the use of mydriatic smartphone teleophthalmology as an alternative method to screen for DR.

Vessel Density around Foveal Avascular Zone as a Potential Imaging Biomarker for Detecting Preclinical Diabetic Retinopathy: An Optical Coherence Tomography Angiography Study

ABSTRACT Purpose: The purpose of this study was to investigate the changes of optical coherence tomography angiography (OCTA) parameters in diabetic retinopathy (DR) using an updated software with 3D projection artifact removal. Methods: In this cross-sectional observational study, 192 eyes of 111 patients with diabetes mellitus (DM) and 55 eyes of 34 age-matched healthy subjects were included. Diabetic patients were divided into three subgroups: without DR, with mild non-proliferative DR, and with moderate-to-severe non-proliferative DR. All eyes underwent dilated fundoscopy along with 3x3mm and 6x6mm OCTA image acquisition. Vessel density (VD), retinal thickness and foveal avascular zone (FAZ) parameters were analyzed. Correlation analyses between OCTA parameters and DR severity were also performed. Results: There was a statistically significant difference in all OCTA parameters among groups, except for superficial foveal VD in 6x6mm scan and whole image retinal thickness in both 3x3mm and 6x6mm scans, while 3x3mm scan parameters were found to be diagnostically superior to the corresponding ones of 6x6mm scan. As the DR stage progressed, the mean VD values decreased. FD-300, which is the VD of a 300-μm width annulus surrounding FAZ, demonstrated the strongest inverse correlation with DR severity (r = −0.590/rs = −0.562, p < .001) and showed the highest area under the ROC curve (AUROC = 0.833 ± 0.030, p < .001) in scan 3 × 3. Conclusion: OCTA shows progressive decrease of VD parameters with increasing DR severity. Foveal VD, FAZ area, and perimeter are not very useful indexes due to the high interindividual variability of FAZ size. OCTA and specifically FD-300 may serve as a promising DR screening tool for detecting preclinical microvascular alterations.

Relationship between corneal sensitivity and diabetic retinopathy among diabetics attending a Nigerian Teaching Hospital.

To determine the relationship between corneal sensitivity and diabetic retinopathy among diabetics attending Lagos University Teaching Hospital, compared to their age-sex-matched non-diabetics, with a view to proposing its use as a screening tool for diabetic retinopathy. The study was a hospital-based comparative cross-sectional study among diabetics aged 25-65 years and their age-sex-matched non-diabetics at Lagos University Teaching Hospital, Lagos, Nigeria. Interviewer-based questionnaire was administered. Corneal sensitivity was assessed with Cochet-Bonnet aesthesiometer. Dilated fundoscopy examination was carried out using slit-lamp bio-microscope with 78D lens and binocular indirect ophthalmoscope. A total of 120 diabetics and 120 age- and sex-matched non-diabetics were enrolled into the study. Among the diabetics, the mean corneal sensitivity was 52.6 ± 6.9 mm, which was statistically lower than among non-diabetics - 58.2 ± 3.0 mm (p-value < 0.05). Twenty-one percent of the diabetic respondents had diabetic retinopathy. The mean cornea sensitivity was statistically lower among diabetics who had developed diabetic retinopathy compared to those with no diabetic retinopathy (48.9 ± 7.3 vs. 55.3 ± 5.4, p < 0.05). The sensitivity and specificity of the aesthesiometer for proliferative diabetic retinopathy at 55 mm cut-off were 100% and 58%, respectively. Cornea sensitivity was significantly lower among diabetics with retinopathy compared to those without retinopathy. Also, the degree of cornea sensation loss among diabetics reduced significantly with severity of diabetic retinopathy. Larger validity studies need to be conducted to further assess the use of corneal aesthesiometer as a screening tool for diabetic retinopathy among non-ophthalmologists and other cadres of health workers.

Comparative evaluation of a novel solar powered low-cost ophthalmoscope (Arclight) by eye healthcare workers in Malawi

This study compared a novel low-cost solar powered direct ophthalmoscope called the Arclight with a traditional direct ophthalmoscope (TDO). After appropriate training, 25 Malawian eye healthcare workers were asked to...

A Diabetic Retinopathy Screening Tool for Low-Income Adults in Mexico.

IntroductionA national diabetic retinopathy screening program does not exist in Mexico as of 2017. Our objective was to develop a screening tool based on a predictive model for early detection of diabetic retinopathy in a low-income population.MethodsWe analyzed biochemical, clinical, anthropometric, and sociodemographic information from 1,000 adults with diabetes in low-income communities in Mexico (from 11,468 adults recruited in 2014–2016). A comprehensive ophthalmologic evaluation was performed. We developed the screening tool through the following stages: 1) development of a theoretical predictive model, 2) performance assessment and validation of the model using cross-validation and the area under the receiver operating characteristic curve (AUC ROC), and 3) optimization of cut points for the classification of diabetic retinopathy. We identified points along the AUC ROC that minimized the misclassification cost function and considered various scenarios of misclassification costs and diabetic retinopathy prevalence.ResultsTime since diabetes diagnosis, high blood glucose levels, systolic hypertension, and physical inactivity were considered risk factors in our screening tool. The mean AUC ROC of our model was 0.780 (validation data set). The optimized cut point that best represented our study population (z = −0.640) had a sensitivity of 82.9% and a specificity of 61.9%.ConclusionWe developed a low-cost and easy-to-apply screening tool to detect people at high risk of diabetic retinopathy in Mexico. Although classification performance of our tool was acceptable (AUC ROC > 0.75), error rates (precision) depend on false-negative and false-positive rates. Therefore, confirmatory assessment of all cases is mandatory.

Role of C- reactive protein in diabetic retinopathy

Introduction: Diabetes mellitus is a major medical problem throughout the world. Diabetes causes an array of long-term systemic complications that have considerable impact on the patient as well as society, as the disease typically affects individuals in their most productive years. C Reactive Protein is an acute phase reactant and an inflammatory mediator involved in the pathogenesis of Diabetic Retinopathy. Aim: The aim of this study is to determine the correlation of C-reactive protein with diabetic retinopathy and different grades of retinopathy. Methods: The study was done on 200 diabetic patients attending the ophthalmology outpatient department, who were selected by purposive sampling. The patients were subjected to a complete ocular examination. Based on the fundus examination patients were categorized into the respective study groups: - 1) Controls 2) Diabetics without Diabetic Retinopathy 3) NPDR 4) PDR 5) CSME. The patients were then subjected to C- reactive protein estimation via laboratory analysis (Turbidometry Technique). Results: The study showed that the mean CRP levels in diabetics were found to be 3mg/dl which is well within normal limits. The CRP levels were relatively higher in the PDR and CSME groups with it being elevated in 30% and 20% patients in the groups respectively. Conclusion: CRP levels in our study did not show significant correlation with diabetes mellitus. Although higher grades of retinopathy had higher CRP levels, the correlation was not significant and consistent. Hence C- reactive protein cannot be used as a reliable screening tool for diabetic retinopathy.

Application of Novel Software Algorithms to Spectral-Domain Optical Coherence Tomography for Automated Detection of Diabetic Retinopathy.

To present novel software algorithms applied to spectral-domain optical coherence tomography (SD-OCT) for automated detection of diabetic retinopathy (DR). Thirty-one diabetic patients (44 eyes) and 18 healthy, nondiabetic controls (20 eyes) who underwent volumetric SD-OCT imaging and fundus photography were retrospectively identified. A retina specialist independently graded DR stage. Trained automated software generated a retinal thickness score signifying macular edema and a cluster score signifying microaneurysms and/or hard exudates for each volumetric SD-OCT. Of 44 diabetic eyes, 38 had DR and six eyes did not have DR. Leave-one-out cross-validation using a linear discriminant at missed detection/false alarm ratio of 3.00 computed software sensitivity and specificity of 92% and 69%, respectively, for DR detection when compared to clinical assessment. Novel software algorithms applied to commercially available SD-OCT can successfully detect DR and may have potential as a viable screening tool for DR in future. [Ophthalmic Surg Lasers Imaging Retina. 2016;47:410-417.].

Automated Diabetic Retinopathy Screening and Monitoring Using Retinal Fundus Image Analysis.

Diabetic retinopathy (DR)-a common complication of diabetes-is the leading cause of vision loss among the working-age population in the western world. DR is largely asymptomatic, but if detected at early stages the progression to vision loss can be significantly slowed. With the increasing diabetic population there is an urgent need for automated DR screening and monitoring. To address this growing need, in this article we discuss an automated DR screening tool and extend it for automated estimation of microaneurysm (MA) turnover, a potential biomarker for DR risk. The DR screening tool automatically analyzes color retinal fundus images from a patient encounter for the various DR pathologies and collates the information from all the images belonging to a patient encounter to generate a patient-level screening recommendation. The MA turnover estimation tool aligns retinal images from multiple encounters of a patient, localizes MAs, and performs MA dynamics analysis to evaluate new, persistent, and disappeared lesion maps and estimate MA turnover rates. The DR screening tool achieves 90% sensitivity at 63.2% specificity on a data set of 40 542 images from 5084 patient encounters obtained from the EyePACS telescreening system. On a subset of 7 longitudinal pairs the MA turnover estimation tool identifies new and disappeared MAs with 100% sensitivity and average false positives of 0.43 and 1.6 respectively. The presented automated tools have the potential to address the growing need for DR screening and monitoring, thereby saving vision of millions of diabetic patients worldwide.

Validation of Smartphone Based Retinal Photography for Diabetic Retinopathy Screening.

AimTo evaluate the sensitivity and specificity of “fundus on phone’ (FOP) camera, a smartphone based retinal imaging system, as a screening tool for diabetic retinopathy (DR) detection and DR severity in comparison with 7-standard field digital retinal photography.DesignSingle-site, prospective, comparative, instrument validation study.Methods301 patients (602 eyes) with type 2 diabetes underwent standard seven-field digital fundus photography with both Carl Zeiss fundus camera and indigenous FOP at a tertiary care diabetes centre in South India. Grading of DR was performed by two independent retina specialists using modified Early Treatment of Diabetic Retinopathy Study grading system. Sight threatening DR (STDR) was defined by the presence of proliferative DR(PDR) or diabetic macular edema. The sensitivity, specificity and image quality were assessed.ResultsThe mean age of the participants was 53.5 ±9.6 years and mean duration of diabetes 12.5±7.3 years. The Zeiss camera showed that 43.9% had non-proliferative DR(NPDR) and 15.3% had PDR while the FOP camera showed that 40.2% had NPDR and 15.3% had PDR. The sensitivity and specificity for detecting any DR by FOP was 92.7% (95%CI 87.8–96.1) and 98.4% (95%CI 94.3–99.8) respectively and the kappa (ĸ) agreement was 0.90 (95%CI-0.85–0.95 p<0.001) while for STDR, the sensitivity was 87.9% (95%CI 83.2–92.9), specificity 94.9% (95%CI 89.7–98.2) and ĸ agreement was 0.80 (95%CI 0.71–0.89 p<0.001), compared to conventional photography.ConclusionRetinal photography using FOP camera is effective for screening and diagnosis of DR and STDR with high sensitivity and specificity and has substantial agreement with conventional retinal photography.

The reliability of single‐field fundus photography in screening for diabetic retinopathy: the Central Australian Ocular Health Study

To assess the accuracy of grading diabetic retinopathy (DR) using single-field digital fundus photography compared with clinical grading from a dilated slit-lamp fundus examination in Indigenous Australians living in Central Australia. Cross-sectional study comparing DR grades in participants with diabetes mellitus presenting for examination at remote community clinics from 1 July 2005 to 30 June 2008. Sensitivity and specificity of grading using digital photography compared with the clinical gold standard of slit-lamp fundus examination. Of the 1884 participants recruited for the study, 1040 had self-reported diabetes mellitus and, of those, 360 had fundus photographs available (706 eyes) that were able to be graded. On clinical grading, 163 eyes had any DR and 51 eyes had vision-threatening DR (VTDR). The sensitivity and specificity for detecting any DR were 74% (95% CI, 67%-80%) and 92% (95% CI, 90%-94%), respectively. The sensitivity and specificity for detecting VTDR were 86% (95% CI, 77%-96%) and 95% (95% CI, 93%-97%), respectively. Single-field digital fundus photography is a valid screening tool for DR in remote communities of central Australia and may be used to provide eye care services to this region with acceptable accuracy.