Burden of undiagnosed and suboptimally controlled diabetes in selected regions of India: Results from the SMART India population-level diabetes screening study.

AIMS: To estimate the prevalence of undiagnosed diabetes and suboptimally controlled diabetes and the associated risk factors by community screening in India. METHODS: In this multi-centre, cross-sectional study, house-to-house screening was conducted in people aged ≥40 years in urban and rural areas across 10 states and one union territory in India between November 2018 and March 2020. Participants underwent anthropometry, clinical and biochemical assessments. Capillary random blood glucose and point-of-care glycated haemoglobin (HbA1c ) were used to diagnose diabetes. The prevalence of undiagnosed diabetes and suboptimal control (HbA1c ≥53 mmol/mol [≥7%]) among those with known diabetes was assessed. RESULTS: Among the 42,146 participants screened (22,150 urban, 19,996 rural), 5689 had known diabetes. The age-standardised prevalence of known diabetes was 13.1% (95% CI 12.8-13.4); 17.2% in urban areas and 9.4% in rural areas. The age-standardised prevalence of undiagnosed diabetes was 6.0% (95% CI 5.7-6.2); similar in both urban and rural areas with the highest proportions seen in the East (8.0%) and South (7.8%) regions. When we consider all people with diabetes in the population, 22.8% of individuals in urban areas and 36.7% in rural areas had undiagnosed diabetes. Almost 75% of the individuals with known diabetes had suboptimal glycaemic control. CONCLUSIONS: High prevalence of undiagnosed diabetes and suboptimally controlled diabetes emphasises the urgent need to identify and optimally treat people with diabetes to reduce the burden of diabetes.

Zone-wise examination of optical coherence tomography features and their correspondence to multifocal electroretinography in eyes with nonproliferative diabetic retinopathy.

PURPOSE: To examine (1) the retinal structure by optical coherence tomography (OCT) and function by means of multifocal electroretinography (mfERG) in eyes with and without nonproliferative diabetic retinopathy (NPDR) (2) for correspondence between local retinal function and OCT zones with retinal lesions. METHODS: One hundred and thirty-two eligible participants (30 with nonproliferative DR (NPDR) and 102 with diabetes with no DR) underwent comprehensive ophthalmic examination, optical coherence tomography for retinal thickness measures, mfERG, and ultra-wide field fundus photography. OCT Early Treatment Diabetic Retinopathy Study (ETDRS) grid was overlaid on to mfERG plots. RESULTS: Those with NPDR had significantly thicker full retinal measures in the nine (ETDRS) zones compared to no DR. mfERG P1 latencies in rings 1-6 were significantly delayed, while the response densities in rings 4-6 were lower in the NPDR group. Significant negative correlation was noted between OCT thickness and mfERG P1 response densities in many ETDRS zones. Significant positive correlation was noted between P1 latencies and OCT thickness in a few zones. The combination of cystic spaces, microaneurysms, and hard exudates were present in all zones and were associated with a decrease in P1 response densities compared to no lesions. Reduced P1 response densities were associated with a sporadic delay in the mfERG latencies and vice versa. The number of lesions did not show correspondence to the mfERG measures. CONCLUSIONS: In eyes with NPDR, retinal function is differentially correlated with the DR lesions on OCT and can be assessed using multimodal imaging modalities.

Effects of Metabolic Surgery on Diabetic Kidney Disease and Diabetic Retinopathy Among Obese Asian Indians with Type 2 Diabetes.

INTRODUCTION: To evaluate the effect of metabolic surgery on microvascular changes associated with diabetic retinopathy (DR) and diabetic kidney disease (DKD) in obese Asian Indians with type 2 diabetes (T2DM), one year after metabolic surgery. METHODS: This is a follow up study in 21 obese Asian Indians with T2DM who underwent metabolic surgery (MS). Diabetic microvascular complications were assessed before and one-year post surgery using urinary albumin, protein creatinine ratio, eGFR, retinal colour photography and Optical coherence tomography (OCT). RESULTS: Microalbuminuria (54±26 vs 28±16 vs 21±6 µg/mg, p<0.001) and protein creatinine ratio (0.4±0.1 vs 0.2±0.03 vs 0.1±0.02, p<0.05) reduced significantly 6 months and one year after Metabolic surgery (MS) respectively compared to baseline values. Estimated Glomerular Filtration (eGFR) rate and creatinine was stable and there was no decline in renal function one year after MS. DR was present in eight individuals at baseline. After metabolic surgery, 12 % of individuals achieved regression of DR and 12% individuals showed a one step regression from severe to moderate non proliferative DR while 12 % individuals progressed from moderate to severe non proliferative DR. Of the 14 (53.8%) individuals who had micro or macroalbuminuria at baseline, 43% individuals reverted back to normoalbuminuria. There was also a reduction in the usage of anti- hypertensive medications after MS. CONCLUSION: In obese Asian Indians with T2DM, metabolic surgery reduced urinary microalbuminuria and protein creatinine ratios at one-year post MS. MS resulted in stable D. Retionpathy status one-year post surgery. MS may help to improve in stabilisation of the microvascular complications in obese patients with T2DM.

Correlating the patterns of diabetic macular edema, optical coherence tomography biomarkers and grade of diabetic retinopathy with stage of renal disease.

PURPOSE: To correlate optical coherence tomography (OCT)-based morphological patterns of diabetic macular edema (DME), biomarkers and grade of diabetic retinopathy (DR) in patients with various stages of chronic kidney disease (CKD) secondary to diabetes. DESIGN: Multicentric retrospective cross-sectional study was conducted at seven centers across India. METHODS: Data from medical records of patients with DME and CKD were entered in a common excel sheet across all seven centers. Staging of CKD was based on estimated glomerular filtration rate (eGFR). RESULTS: The most common morphological pattern of DME was cystoid pattern (42%) followed by the mixed pattern (31%). The proportion of different morphological patterns did not significantly vary across various CKD stages (p = 0.836). The presence of external limiting membrane-ellipsoid zone (ELM-EZ) defects (p < 0.001) and foveal sub-field thickness (p = 0.024) showed a direct correlation with the stage of CKD which was statistically significant. The presence of hyperreflective dots (HRD) and disorganization of inner retinal layers (DRIL) showed no significant correlation with the stage of CKD. Sight threatening DR was found to increase from 70% in CKD stage 3 to 82% in stages 4 and 5 of CKD, and this was statistically significant (p = 0.03). CONCLUSION: Cystoid morphological pattern followed by mixed type was the most common pattern of DME on OCT found in patients suffering from stage 3 to 5 of CKD. However, the morphological patterns of DME did not significantly vary across various CKD stages. ELM-EZ defects may be considered as an important OCT biomarker for advanced stage of CKD.

Clinical profile and incidence of microvascular complications of childhood and adolescent onset type 1 and type 2 diabetes seen at a tertiary diabetes center in India.

AIM: To study the clinical characteristics and incidence of microvascular complications among childhood and adolescent onset type 1 (T1DM) and type 2 diabetes (T2DM) seen at a tertiary care diabetes center in India. METHODS: From our electronic medical records, we retrieved clinical and biochemical details of 4555 individuals with childhood and adolescent onset diabetes (diagnosed below the age of 20 years) seen between 1992 and 2017. T1DM was diagnosed if there was history of ketoacidosis or fasting C-peptide <0.3 PMol/mL and stimulated C-peptide <0.6 PMol/mL or if insulin treatment was required from the time of diagnosis. T2DM was diagnosed based on absence of ketosis, or fasting C-peptide ≥0.6 PMol/mL and stimulated >1.0 PMoL/mL, or response to oral hypoglycemic agents for more than 2 years. We calculated the incidence rates of retinopathy (presence of at least one definite microaneurysm by retinal photography), nephropathy (urinary albumin excretion ≥30 µg/mg of creatinine) and neuropathy (vibration perception threshold ≥20 V) per 1000 person-years of follow up. RESULTS: Among the 4555 individuals with childhood and adolescent-onset diabetes, 71.4% had T1DM, 19.5% T2DM and 9.1% other forms of diabetes. Age at first visit and duration of diabetes were significantly higher in T2DM when compared to T1DM. The age adjusted incidence of retinopathy was 52.9/1000 person years (Confidence Intervals [CI]: 42.9-62.8) in T1DM and 49.8/1000 person years (CI 30.8-68.8) in T2DM; nephropathy, 6.2 (CI 3.3-9.0) and 13.8 (CI 5.6-22.0); and neuropathy, 8.8(CI 3.6-14.0) and 24.0 (CI 9.8-38.2) in T1DM and T2DM, respectively. CONCLUSION: The incidence of microvascular complications is high among childhood and adolescent-onset T1DM and T2DM and these calls for more aggressive control of diabetes.

Need for Vitreous Surgeries in Proliferative Diabetic Retinopathy in 10-Year Follow-Up: India Retinal Disease Study Group Report No. 2.

INTRODUCTION: To report the 10-year rate of vitrectomies and the associated factors in people with proliferative diabetic retinopathy (PDR) from a multicentric cohort of people with diabetes mellitus. METHODS: Ten centres in India with established vitreoretinal (VR) services for over 10 years were invited to provide long-term data on PDR. People with Type 1 or 2 diabetes with a clinical diagnosis of active PDR in 1 or both eyes were included. Baseline data collected included age, sex, duration of diabetes, source of referral and best-corrected visual acuity, and diabetic retinopathy status in both eyes. Available follow-up data included the numbers of panretinal photocoagulation (PRP) sessions, cataract surgery, treatment of diabetic macular oedema, use of anti-vascular endothelial growth factor (VEGF) therapy, vitrectomy with or without retinal surgeries over 10 years. RESULTS: Over 10 years, 89% needed supplemental PRP after initial complete PRP. One-third required retinal surgery, 16% needed intravitreal injection. Men (74.5%) had significant higher risk for vitreous (VR) surgery. Of the group with low-risk PDR, 56.8% did not require VR surgery, p < 0.001. Of the patients who underwent cataract surgery and had intravitreal anti-VEGF injections, 78.5 and 28.2% needed subsequent vitreous (VR) surgery, p = 0.006 and <0.0001, respectively. Independent predictors of need for vitreoretinal surgery included those who underwent cataract surgery and those with poor baseline visual acuity (logMAR). Eyes at lower risk for VR surgery included the eyes previously treated with PRP and low-risk PDR at baseline. CONCLUSION: Despite initial "complete" PRP, one-third of our study cohort needed vitrectomies over 10 years, highlighting that these patients require regular follow-up for a long period of time.

Comparison Among Methods of Retinopathy Assessment (CAMRA) Study: Smartphone, Nonmydriatic, and Mydriatic Photography.

PURPOSE: We compared smartphone fundus photography, nonmydriatic fundus photography, and 7-field mydriatic fundus photography for their abilities to detect and grade diabetic retinopathy (DR). DESIGN: This was a prospective, comparative study of 3 photography modalities. PARTICIPANTS: Diabetic patients (n = 300) were recruited at the ophthalmology clinic of a tertiary diabetes care center in Chennai, India. METHODS: Patients underwent photography by all 3 modalities, and photographs were evaluated by 2 retina specialists. MAIN OUTCOME MEASURES: The sensitivity and specificity in the detection of DR for both smartphone and nonmydriatic photography were determined by comparison with the standard method, 7-field mydriatic fundus photography. RESULTS: The sensitivity and specificity of smartphone fundus photography, compared with 7-field mydriatic fundus photography, for the detection of any DR were 50% (95% confidence interval [CI], 43-56) and 94% (95% CI, 92-97), respectively, and of nonmydriatic fundus photography were 81% (95% CI, 75-86) and 94% (95% CI, 92-96%), respectively. The sensitivity and specificity of smartphone fundus photography for the detection of vision-threatening DR were 59% (95% CI, 46-72) and 100% (95% CI, 99-100), respectively, and of nonmydriatic fundus photography were 54% (95% CI, 40-67) and 99% (95% CI, 98-100), respectively. CONCLUSIONS: Smartphone and nonmydriatic fundus photography are each able to detect DR and sight-threatening disease. However, the nonmydriatic camera is more sensitive at detecting DR than the smartphone. At this time, the benefits of the smartphone (connectivity, portability, and reduced cost) are not offset by the lack of sufficient sensitivity for detection of DR in most clinical circumstances.

Assessing the appropriateness and completeness of ChatGPT-4's AI-generated responses for queries related to diabetic retinopathy.

OBJECTIVE: To evaluate the appropriateness of responses generated by an online chat-based artificial intelligence (AI) model for diabetic retinopathy (DR) related questions. DESIGN: Cross-sectional study. METHODS: A set of 20 questions framed from the patient's perspective addressing DR-related queries, such as the definition of disease, symptoms, prevention methods, treatment options, diagnostic methods, visual impact, and complications, were formulated for input into ChatGPT-4. Peer-reviewed, literature-based answers were collected from popular search engines for the selected questions and three retinal experts reviewed the responses. An inter-human agreement was analyzed for consensus expert responses and also between experts. The answers generated by the AI model were compared with those provided by the experts. The experts rated the response generated by ChatGPT-4 on a scale of 0-5 for appropriateness and completeness. RESULTS: The answers provided by ChatGPT-4 were appropriate and complete for most of the DR-related questions. The response to questions on the adverse effects of laser photocoagulation therapy and compliance to treatment was not perfectly complete. The average rating given by the three retina expert evaluators was 4.84 for appropriateness and 4.38 for completeness of answers provided by the AI model. This corresponds to an overall 96.8% agreement among the experts for appropriateness and 87.6% for completeness regarding AI-generated answers. CONCLUSION: ChatGPT-4 exhibits a high level of accuracy in generating appropriate responses for a range of questions in DR. However, there is a need to improvise the model to generate complete answers for certain DR-related topics.

Wide-field imaging with smartphone based fundus camera: grading of severity of diabetic retinopathy and locating peripheral lesions in diabetic retinopathy.

AIM: To assess the performance of smartphone based wide-field retinal imaging (WFI) versus ultra-wide-field imaging (UWFI) for assessment of sight-threatening diabetic retinopathy (STDR) as well as locating predominantly peripheral lesions (PPL) of DR. METHODS: Individuals with type 2 diabetes with varying grades of DR underwent nonmydriatic UWFI with Daytona Plus camera followed by mydriatic WFI with smartphone-based Vistaro camera at a tertiary care diabetes centre in South India in 2021-22. Grading of DR as well as identification of PPL (DR lesions beyond the posterior pole) in the retinal images of both cameras was performed by senior retina specialists. STDR was defined by the presence of severe non-proliferative DR, proliferative DR or diabetic macular oedema (DME). The sensitivity and specificity of smartphone based WFI for detection of PPL and STDR was assessed. Agreement between the graders for both cameras was compared. RESULTS: Retinal imaging was carried out in 318 eyes of 160 individuals (mean age 54.7 ± 9 years; mean duration of diabetes 16.6 ± 7.9 years). The sensitivity and specificity for detection of STDR by Vistaro camera was 92.7% (95% CI 80.1-98.5) and 96.6% (95% CI 91.5-99.1) respectively and 95.1% (95% CI 83.5-99.4) and 95.7% (95% CI 90.3-98.6) by Daytona Plus respectively. PPL were detected in 89 (27.9%) eyes by WFI by Vistaro camera and in 160 (50.3%) eyes by UWFI. However, this did not translate to any significant difference in the grading of STDR between the two imaging systems. In both devices, PPL were most common in supero-temporal quadrant (34%). The prevalence of PPL increased with increasing severity of DR with both cameras (p < 0.001). The kappa comparison between the 2 graders for varying grades of severity of DR was 0.802 (p < 0.001) for Vistaro and 0.753 (p < 0.001) for Daytona Plus camera. CONCLUSION: Mydriatic smartphone-based widefield imaging has high sensitivity and specificity for detecting STDR and can be used to screen for peripheral retinal lesions beyond the posterior pole in individuals with diabetes.

National prevalence of vision impairment and blindness and associated risk factors in adults aged 40 years and older with known or undiagnosed diabetes: results from the SMART-India cross-sectional study.

BACKGROUND: National estimates of the prevalence of vision impairment and blindness in people with diabetes are required to inform resource allocation. People with diabetes are more susceptible to conditions such as diabetic retinopathy that can impair vision; however, these are often missed in national studies. This study aims to determine the prevalence and risk factors of vision impairment and blindness in people with diabetes in India. METHODS: Data from the SMART-India study, a cross-sectional survey with national coverage of 42 147 Indian adults aged 40 years and older, collected using a complex sampling design, were used to obtain nationally representative estimates for the prevalence of vision impairment and blindness in people with diabetes in India. Vulnerable adults (primarily those who did not have capacity to provide consent); pregnant and breastfeeding women; anyone deemed too ill to be screened; those who did not provide consent; and people with type 1 diabetes, gestational diabetes, or secondary diabetes were excluded from the study. Vision impairment was defined as presenting visual acuity of 0·4 logMAR or higher and blindness as presenting a visual acuity of 1·0 logMAR or higher in the better-seeing eye. Demographic, anthropometric, and laboratory data along with geographic distribution were analysed in all participants with available data. Non-mydriatic retinal images were used to grade diabetic retinopathy, and risk factors were also assessed. FINDINGS: A total of 7910 people with diabetes were included in the analysis, of whom 5689 had known diabetes and 2221 were undiagnosed. 4387 (55·5%) of 7909 participants with available sex data were female and 3522 (44·5%) participants were male. The estimated national prevalence of vision impairment was 21·1% (95% CI 15·7-27·7) and blindness 2·4% (1·7-3·4). A higher prevalence of any vision impairment (29·2% vs 19·6%; p=0·016) and blindness (6·7% vs 1·6%; p<0·0001) was observed in those with ungradable images. In known diabetes, diabetic retinopathy (adjusted odds ratio [aOR] 3·06 [95% CI 1·25-7·51]), vision-threatening diabetic retinopathy (aOR 7·21 [3·52-14·75]), and diabetic macular oedema (aOR 5·41 [2·20-13·33]) were associated with blindness in adjusted analysis. Common sociodemographic risk factors for vision impairment and blindness include older age, lower educational attainment, and unemployment. INTERPRETATION: Based on the estimated 101 million people with diabetes in 2021 and the interpretation of the data from this study, approximately 21 million people with diabetes have vision impairment in India, of whom 2·4 million are blind. Higher prevalence is observed in those from lower socio-economic strata and policy makers should focus on these groups to reduce inequalities in health care. FUNDING: Global Challenge Research Fund of United Kingdom Research and Innovation through the Medical Research Council.

Deep learning-based detection of diabetic macular edema using optical coherence tomography and fundus images: A meta-analysis.

Diabetic macular edema (DME) is an important cause of visual impairment in the working-age group. Deep learning methods have been developed to detect DME from two-dimensional retinal images and also from optical coherence tomography (OCT) images. The performances of these algorithms vary and often create doubt regarding their clinical utility. In resource-constrained health-care systems, these algorithms may play an important role in determining referral and treatment. The survey provides a diversified overview of macular edema detection methods, including cutting-edge research, with the objective of providing pertinent information to research groups, health-care professionals, and diabetic patients about the applications of deep learning in retinal image detection and classification process. Electronic databases such as PubMed, IEEE Explore, BioMed, and Google Scholar were searched from inception to March 31, 2022, and the reference lists of published papers were also searched. The study followed the preferred reporting items for systematic review and meta-analysis (PRISMA) reporting guidelines. Examination of various deep learning models and their exhibition regarding precision, epochs, their capacity to detect anomalies for less training data, concepts, and challenges that go deep into the applications were analyzed. A total of 53 studies were included that evaluated the performance of deep learning models in a total of 1,414,169°CT volumes, B-scans, patients, and 472,328 fundus images. The overall area under the receiver operating characteristic curve (AUROC) was 0.9727. The overall sensitivity for detecting DME using OCT images was 96% (95% confidence interval [CI]: 0.94-0.98). The overall sensitivity for detecting DME using fundus images was 94% (95% CI: 0.90-0.96).

Compliance with follow-up in patients with diabetic macular edema: Eye care center vs. diabetes care center.

Purpose: The study was conducted to compare the compliance to intravitreal injection treatment and follow-up in patients with center-involving diabetic macular edema (CI-DME) and treatment outcomes between a tertiary eye care facility and a tertiary diabetes care center. Methods: A retrospective review was conducted on treatment naïve DME patients who had received intravitreal anti-vascular endothelial growth factor (anti-VEGF) injections in 2019. Participants were people with type 2 diabetes who were under regular care at the eye care center or the diabetes care center in Chennai. The outcome measures were noted at months 1, 2, 3, 6, and 12. Results: A review of 136 patients treated for CI-DME (72 from the eye care center and 64 from a diabetes care center) was carried out. The severity of diabetic retinopathy (DR) was similar in both centers. There was no statistically significant (P > 0.05) difference in the choice of initial intravitreal drug in the two centers. At 12-month follow-up, only 29.16% came for a follow-up in the eye center vs. 76.56% in a diabetes care center (P = 0.000). The multivariate logistic regression showed increasing age was associated with non-compliance in both the groups (eye care center: odds ratio [OR] 0.91; 95% confidence interval [CI] 0.82-1.21; P = 0.044) and diabetes care center (OR 1.15; 95% CI 1.02-1.29; P = 0.020). Conclusion: The follow-up rate between eye care and diabetic care center with DME showed a significant disparity. By providing comprehensive diabetes care for all complications under one roof, compliance with follow-up can be improved in people with DME.

Machine Learning-Based Diagnosis and Ranking of Risk Factors for Diabetic Retinopathy in Population-Based Studies from South India.

This paper discusses the importance of investigating DR using machine learning and a computational method to rank DR risk factors by importance using different machine learning models. The dataset was collected from four large population-based studies conducted in India between 2001 and 2010 on the prevalence of DR and its risk factors. We deployed different machine learning models on the dataset to rank the importance of the variables (risk factors). The study uses a t-test and Shapely additive explanations (SHAP) to rank the risk factors. Then, it uses five machine learning models (K-Nearest Neighbor, Decision Tree, Support Vector Machines, Logistic Regression, and Naive Bayes) to identify the unimportant risk factors based on the area under the curve criterion to predict DR. To determine the overall significance of risk variables, a weighted average of each classifier's importance is used. The ranking of risk variables is provided to machine learning models. To construct a model for DR prediction, the combination of risk factors with the highest AUC is chosen. The results show that the risk factors glycosylated hemoglobin and systolic blood pressure were present in the top three risk factors for DR in all five machine learning models when the t-test was used for ranking. Furthermore, the risk factors, namely, systolic blood pressure and history of hypertension, were present in the top five risk factors for DR in all the machine learning models when SHAP was used for ranking. Finally, when an ensemble of the five machine learning models was employed, independently with both the t-test and SHAP, systolic blood pressure and diabetes mellitus duration were present in the top four risk factors for diabetic retinopathy. Decision Tree and K-Nearest Neighbor resulted in the highest AUCs of 0.79 (t-test) and 0.77 (SHAP). Moreover, K-Nearest Neighbor predicted DR with 82.6% (t-test) and 78.3% (SHAP) accuracy.

Population-Based Artificial Intelligence Assessment of Relationship Between the Risk Factors for Diabetic Retinopathy in Indian Population.

PURPOSE: Risk factors (RFs), like 'body mass index (BMI),' 'age,' and 'gender' correlate with Diabetic Retinopathy (DR) diagnosis and have been widely studied. This study examines how these three secondary RFs independently affect the predictive capacity of primary RFs. METHODS: The dataset consisted of four population-based studies on the prevalence of DR and associated RFs in India between 2001 and 2010. An Autoencoder was employed to categorize RFs as primary or secondary. This study evaluated six primary RFs coupled independently with each secondary RF on five machine-learning models. RESULTS: The secondary RF 'gender' gave a maximum increase in Area under the curve (AUC) score to predict DR when combined separately with 'insulin treatment,' 'fasting plasma glucose,' 'hypertension history,' and 'glycosylated hemoglobin' with a maximum increase in AUC for the Naive Bayes model from 0.573 to 0.646, for the Support Vector Machines (SVM) model from 0.644 to 0.691, for the SVM model from 0.487 to 0.607, and for the Decision Tree model from 0.8 to 0.848, respectively. The secondary RFs 'age' and 'BMI' gave a maximum increase in AUC score to predict DR when combined separately with 'diabetes mellitus duration' and 'systolic blood pressure,' with a maximum increase in AUC for the SVM model from 0.389 to 0.621, and for the Decision Tree model from 0.617 to 0.713, respectively. CONCLUSION: The risk factor 'gender' was the best secondary RF in predicting DR compared to 'age' and 'BMI,' increasing the predictive power of four primary RFs.

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

Purpose: To evaluate the ability of an artificial intelligence (AI) model, ChatGPT, in predicting the diabetic retinopathy (DR) risk. Methods: 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. Results: 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). Conclusion: This study suggests that ChatGPT has the potential of a preliminary DR screening tool with further optimization needed for clinical use.

Using deep learning to detect diabetic retinopathy on handheld non-mydriatic retinal images acquired by field workers in community settings.

Diabetic retinopathy (DR) at risk of vision loss (referable DR) needs to be identified by retinal screening and referred to an ophthalmologist. Existing automated algorithms have mostly been developed from images acquired with high cost mydriatic retinal cameras and cannot be applied in the settings used in most low- and middle-income countries. In this prospective multicentre study, we developed a deep learning system (DLS) that detects referable DR from retinal images acquired using handheld non-mydriatic fundus camera by non-technical field workers in 20 sites across India. Macula-centred and optic-disc-centred images from 16,247 eyes (9778 participants) were used to train and cross-validate the DLS and risk factor based logistic regression models. The DLS achieved an AUROC of 0.99 (1000 times bootstrapped 95% CI 0.98-0.99) using two-field retinal images, with 93.86 (91.34-96.08) sensitivity and 96.00 (94.68-98.09) specificity at the Youden's index operational point. With single field inputs, the DLS reached AUROC of 0.98 (0.98-0.98) for the macula field and 0.96 (0.95-0.98) for the optic-disc field. Intergrader performance was 90.01 (88.95-91.01) sensitivity and 96.09 (95.72-96.42) specificity. The image based DLS outperformed all risk factor-based models. This DLS demonstrated a clinically acceptable performance for the identification of referable DR despite challenging image capture conditions.

Association of OCT and OCT angiography measures with the development and worsening of diabetic retinopathy in type 2 diabetes.

OBJECTIVE: To assess if optical coherence tomography (OCT) and OCT angiography (OCTA) measures are associated with the development and worsening of diabetic retinopathy (DR) over four years. METHODS: 280 participants with type 2 diabetes underwent ultra-wide field fundus photography, OCT and OCTA. OCT-derived macular thickness measures, retinal nerve fibre layer and ganglion cell-inner plexiform layer thickness and OCTA-derived foveal avascular zone area, perimeter, circularity, vessel density (VD) and macular perfusion (MP) were examined in relation to the development and worsening of DR over four years. RESULTS: After four years, 206 eyes of 219 participants were eligible for analysis. 27 of the 161 eyes (16.7%) with no DR at baseline developed new DR, which was associated with a higher baseline HbA1c and longer diabetes duration. Of the 45 eyes with non-proliferative DR (NPDR) at baseline, 17 (37.7%) showed DR progression. Baseline VD (12.90 vs. 14.90 mm/mm2, p = 0.032) and MP (31.79% vs. 36.96%, p = 0.043) were significantly lower in progressors compared to non-progressors. Progression of DR was inversely related to VD ((hazard ratio [HR] = 0.825) and to MP (HR = 0.936). The area under the receiver operating characteristic curves for VD was AUC = 0.643, with 77.4% sensitivity and 41.8% specificity for a cut-off of 15.85 mm/mm2 and for MP it was AUC = 0.635, with 77.4% sensitivity and 25.5% specificity for a cut-off of 40.8%. CONCLUSIONS: OCTA metrics have utility in predicting progression rather than the development of DR in individuals with type 2 diabetes.

Role of cystatin C in the detection of sight-threatening diabetic retinopathy in Asian Indians with type 2 diabetes.

AIM: To study the association between cystatin C and sight-threatening diabetic retinopathy (STDR) in Asian Indians with type 2 diabetes (T2DM). METHODS: In a cross-sectional study carried out at two tertiary centres in India in 2022, individuals with T2DM underwent clinical and ophthalmic assessments and estimation of serum cystatin C. Grading of DR was done by retina specialists. STDR was defined by the presence of severe non-proliferative DR (NPDR), proliferative DR (PDR) and/or diabetic macular edema. Receiver operating characteristic (ROC) curves were used to identify cystatin C cut-off value for detecting STDR. RESULTS: Among 420 individuals with T2DM (mean age 56 ± 9 years; mean duration of diabetes 14.5 ± 7.9 years), 121 (24.1 %) had No-DR, 119 (28.3 %) had No-STDR and 200 (49.6 %) had STDR. Mean cystatin C level was significantly higher in individuals with STDR compared to those with no-STDR and No-DR (1.34 vs 1.06 vs 0.93 mg/L, p < 0.001). Cystatin C cut-off value ≥1.11 mg/L had a C statistic of 0.944 (95 % CI: 0.909-0.968, p < 0.001), 96.8 % sensitivity and 78.2 % specificity for detection of STDR. CONCLUSION: Elevated serum cystatin C was strongly associated with STDR and could possibly be used as a biomarker for screening for sight-threatening diabetic retinopathy.

Inter-observer agreement in grading severity of diabetic retinopathy in wide-field fundus photographs.

OBJECTIVE: To examine the inter-observer agreement between two retina specialists in grading diabetic retinopathy (DR) severity in ultra-wide-field fundus photographs. METHODS: Two hundred and seventy patients with diabetes, who visited the vitreoretinal specialty at a tertiary eye care hospital, with or without DR underwent comprehensive ophthalmic examination, dilated retinal exam and Optos ultra-wide-field (UWF) retinal photography. Optos images were graded for DR severity based on the International Clinical Diabetic Retinopathy Disease Severity Scale by two retina specialists with same number of years of experience, masked to the clinical details of the participants. RESULTS: The two graders showed agreement in 229/270 images (84.8%) and disagreement in 41/270 images (15.2%). The unweighted kappa for agreement between graders was k = 0.715, SE = 0.037 and the weighted kappa was k = 0.838, SE = 0.022. No DR was identified in 170/270 (62.9%) patients, mild NPDR in 15/270 (5.6%) patients, moderate NPDR in 35/270 (12.9%) patients, severe NPDR in 4/270 (1.48%) patient and PDR in 5/270 (1.85%) patients by both graders. Disagreement was neither related to the learning curve of graders nor with the patient's age (p = 0.574), gender (p = 0.169), duration of diabetes (0.660) or the lens being phakic or pseudophakic (p = 0.171) on logistic regression. CONCLUSIONS: The impact of disagreement noted between observers in grading DR on UWF fundus photographs should be considered when utilizing UWF system in clinical studies.

Retinal structure-function correlation in type 2 diabetes.

OBJECTIVE: To examine the relationship of visual function as assessed by visual acuity, contrast sensitivity, and multifocal electroretinography (mfERG) to macular structural and microvascular measures on optical coherence tomography (OCT) and angiography (OCTA) in individuals with diabetes. METHODS: This is a prospective observational study conducted at a tertiary eye care centre in India. Right eyes of 121 adults with type 2 diabetes with no diabetic retinopathy (DR), mild or moderate nonproliferative DR (NPDR) were examined. Severe NPDR, proliferative DR and diabetic macular oedema were excluded. Participants underwent assessment of glycated haemoglobin (HbA1C), blood pressure, best corrected visual acuity (LogMAR), contrast sensitivity (CS), mfERG, ultrawide field fundus photography, OCT and OCTA. Correlations were assessed by Spearman's rank correlation (rho). RESULTS: Of the total of 121 eyes, 89 had No DR, 32 had mild to moderate NPDR. In the No DR group, the LogMAR acuity was significantly and negatively correlated to central subfoveal thickness (CST) (rho = -0.420), macular vessel density (rho = -0.270) and perfusion (rho = -0.270). (ii) Contrast sensitivity correlated to foveal avascular zone circularity (rho = 0.297); (iii) mfERG P1 response densities were better with higher macular perfusion index (rho = 0.240). In the NPDR group, the LogMAR acuity also showed a significant negative correlation to CST (rho = -0.379). Other correlations were not significant. CONCLUSION: Retinal and visual functional changes are evident in diabetic patients with No DR and are correlated to subclinical retinal structural changes detectable using multimodal imaging.