Evaluation of an AI system for the detection of diabetic retinopathy from images captured with a handheld portable fundus camera: the MAILOR AI study

ObjectivesTo evaluate the performance of an artificial intelligence (AI) system (Pegasus, Visulytix Ltd., UK*) at the detection of diabetic retinopathy (DR) from images captured by a handheld portable fundus camera.MethodsA cohort of 6404 patients (~80% with diabetes mellitus) was screened for retinal diseases using a handheld portable fundus camera (Pictor Plus, Volk Optical Inc., USA) at the Mexican Advanced Imaging Laboratory for Ocular Research. The images were graded for DR by specialists according to the Scottish DR grading scheme. The performance of the AI system was evaluated, retrospectively, in assessing referable DR (RDR) and proliferative DR (PDR) and compared with the performance on a publicly available desktop camera benchmark dataset.ResultsFor RDR detection, Pegasus performed with an 89.4% (95% CI: 88.0–90.7) area under the receiver operating characteristic (AUROC) curve for the MAILOR cohort, compared with an AUROC of 98.5% (95% CI: 97.8–99.2) on the benchmark dataset. This difference was statistically significant. Moreover, no statistically significant difference was found in performance for PDR detection with Pegasus achieving an AUROC of 94.3% (95% CI: 91.0–96.9) on the MAILOR cohort and 92.2% (95% CI: 89.4–94.8) on the benchmark dataset.ConclusionsPegasus showed good transferability for the detection of PDR from a curated desktop fundus camera dataset to real-world clinical practice with a handheld portable fundus camera. However, there was a substantial, and statistically significant, decrease in the diagnostic performance for RDR when using the handheld device.Subject terms: Retinal diseases, Physical examination     

Uso en la práctica clínica,de un método de cribado automatizado de retinopatía diabética derivable mediante un sistema de inteligencia artificial de diagnóstico

Background and objectiveTo compare the diagnostic performance of an autonomous diagnostic artificial intelligence (AI) system for the diagnosis of derivable diabetic retinopathy (RDR) with manual classification.Materials and methodsPatients with type 1 and type 2 diabetes participated in a diabetic retinopathy (DR) screening program between 2011-2012. 2 images of each eye were collected. Unidentifiable retinal images were obtained, one centered on the disc and one on the fovea. The exams were classified with the autonomous AI system and manually by anonymous ophthalmologists. The results of the AI system and manual classification were compared in terms of sensitivity and specificity for the diagnosis of both (RDR) and diabetic retinopathy with decreased vision (VTDR).Results10,257 retinal inages of 5,630 eyes of 2,680 subjects were included. According to the manual classification, the prevalence of RDR was 4.14% and that of VTDR 2.57%. The AI system recorded 100% (95% CI: 97-100%) sensitivity and 81.82% (95% CI: 80 -83%) specificity for RDR, and 100% (95% CI: 95-100%) of sensitivity and 94.64% (95% CI: 94-95%) of specificity for VTDR.ConclusionsCompared to the manual classification, the autonomous diagnostic AI system registered a high sensitivity (100%) and specificity (82%) in the diagnosis of RDR and macular edema in people with diabetes. Due to its immediate diagnosis, the autonomous diagnostic AI system can increase the accessibility of RDR screening in primary care settings.     

Macular Choroidal Thickness and the Risk of Referable Diabetic Retinopathy in Type 2 Diabetes: A 2-Year Longitudinal Study

PurposeThe purpose of this study was to evaluate the associations between choroidal thickness (CT) and the 2-year incidence of referable diabetic retinopathy (RDR).MethodsThis was a prospective cohort study. Patients with type 2 diabetes in Guangzhou, China, aged 30 to 80 years underwent comprehensive examinations, including standard 7-field fundus photography. Macular CT was measured using a commercial swept-source optical coherence tomography (SS-OCT) device (DRI OCT Triton; Topcon, Tokyo, Japan). The relative risk (RR) with 95% confidence intervals (CIs) was used to quantify the association between CT and new-onset RDR. The prognostic value of CT was assessed using the area under the receiver operating characteristic curve (AUC), net reclassification improvement (NRI), and integrated discrimination improvement (IDI).ResultsA total of 1345 patients with diabetes were included in the study, and 120 (8.92%) of them had newly developed RDR at the 2-year follow-up. After adjusting for other factors, the increased RDR risk was associated with greater HbA1c (RR = 1.35, 95% CI = 1.17–1.55, P < 0.001), higher systolic blood pressure (SBP; RR = 1.02, 95% CI = 1.01–1.03, P = 0.005), lower triglyceride (TG) level (RR = 0.81, 95% CI = 0.69–0.96, P = 0.015), presence of diabetic retinopathy (DR; RR = 8.16, 95% CI = 4.47–14.89, P < 0.001), and thinner average CT (RR = 0.903, 95% CI = 0.871–0.935, P < 0.001). The addition of average CT improved NRI (0.464 ± 0.096, P < 0.001) and IDI (0.0321 ± 0.0068, P < 0.001) for risk of RDR, and it also improved the AUC from 0.708 (95% CI = 0.659–0.757) to 0.761 (95% CI = 0.719–0.804).ConclusionsCT thinning measured by SS-OCT is an early imaging biomarker for the development of RDR, suggesting that alterations in CT play an essential role in DR occurrence.     

Validation of automated screening for referable diabetic retinopathy with the IDx‐DR device in the Hoorn Diabetes Care System

Purpose

To increase the efficiency of retinal image grading, algorithms for automated grading have been developed, such as the IDx‐DR 2.0 device. We aimed to determine the ability of this device, incorporated in clinical work flow, to detect retinopathy in persons with type 2 diabetes.

Methods

Retinal images of persons treated by the Hoorn Diabetes Care System (DCS) were graded by the IDx‐DR device and independently by three retinal specialists using the International Clinical Diabetic Retinopathy severity scale (ICDR) and EURODIAB criteria. Agreement between specialists was calculated. Results of the IDx‐DR device and experts were compared using sensitivity, specificity, positive predictive value (PPV) and negative predictive value (NPV), distinguishing between referable diabetic retinopathy (RDR) and vision‐threatening retinopathy (VTDR). Area under the receiver operating characteristic curve (AUC) was calculated.

Results

Of the included 1415 persons, 898 (63.5%) had images of sufficient quality according to the experts and the IDx‐DR device. Referable diabetic retinopathy (RDR) was diagnosed in 22 persons (2.4%) using EURODIAB and 73 persons (8.1%) using ICDR classification. Specific intergrader agreement ranged from 40% to 61%. Sensitivity, specificity, PPV and NPV of IDx‐DR to detect RDR were 91% (95% CI: 0.69–0.98), 84% (95% CI: 0.81–0.86), 12% (95% CI: 0.08–0.18) and 100% (95% CI: 0.99–1.00; EURODIAB) and 68% (95% CI: 0.56–0.79), 86% (95% CI: 0.84–0.88), 30% (95% CI: 0.24–0.38) and 97% (95% CI: 0.95–0.98; ICDR). The AUC was 0.94 (95% CI: 0.88–1.00; EURODIAB) and 0.87 (95% CI: 0.83–0.92; ICDR). For detection of VTDR, sensitivity was lower and specificity was higher compared to RDR. AUC's were comparable.

Conclusion

Automated grading using the IDx‐DR device for RDR detection is a valid method and can be used in primary care, decreasing the demand on ophthalmologists.     

Anemia and Diabetic Kidney Disease Had Joint Effect on Diabetic Retinopathy Among Patients With Type 2 Diabetes

PurposeWhether the association between diabetic kidney disease (DKD) and diabetic retinopathy (DR) in patients with type 2 diabetes mellitus (T2DM) is leveraged by anemia remains unclear. This study is to evaluate the joint effect of DKD and anemia on DR.MethodsData were collected from electronic medical records of 1389 patients with T2DM in the Yiwu Central Hospital of Zhejiang Province from 2018 to 2019. Based on retinal examination findings, patients were classified as without diabetic retinopathy (non-DR), non-proliferative diabetic retinopathy (NPDR), and proliferative diabetic retinopathy (PDR). Odds ratio (OR) from multinomial logistic regression models adjusting for potential risk factors of DR were used to evaluate associations of DKD, renal function measures, and anemia with risk of NPDR and PDR. Path analysis was performed to help understand the association of DKD and hemoglobin (Hb) with DR.ResultsThe study included 901 patients with non-DR, 367 patients with NPDR and 121 patients with PDR. Both high DKD risk and abnormal renal function measures were significantly associated with PDR. Anemia was associated with increased risk of NPDR (OR = 1.75, 95% confidence interval [CI] = 1.18–2.58) and PDR (OR = 3.71, 95% CI = 2.23–6.18). DKD severity and anemia had joint effect on NPDR (OR = 2.29, 95% CI = 1.32–3.96) and PDR (OR = 11.31, 95% CI = 5.95–21.51). These associations were supported by path analysis.ConclusionsDKD severity, abnormal estimated glomerular filtration rate (eGFR), and urinary albumin/creatinine ratio (UACR) were associated with increased risk of DR in patients with T2DM, and anemia had joint effect on these associations. Improving Hb level may decrease the risk of DR in patients with T2DM.     

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.Subject terms: Public health, Retinal diseases     

Prevalence of diabetic retinopahty in India: Results from the National Survey 2015-19

Purpose:Diabetes mellitus (DM) and diabetic retinopathy (DR) contribute to ocular morbidity and are emerging as diseases with significant public health impact. Our aim was to assess the countrywide prevalence of DR and sight-threatening DR (STDR) among persons with diabetes and to evaluate the coverage of DR examinations among them.Methods:The present survey was planned to estimate the burden of DR in the population aged ≥50 years for assisting in the planning and prioritization of diabetic eye services. For this survey, 21 districts with a high prevalence of DM were selected among the 31 districts where the national blindness and visual impairment survey was conducted. The total sample size was 63,000 people aged 50 years and above. DR was assessed by dilated fundus examination with indirect ophthalmoscope and was graded according to Scottish DR grading. STDR included severe nonproliferative DR, proliferative DR, and clinically significant macular edema.Results:The prevalence of diabetes in the surveyed population was 11.8%. Among them, one-third were newly diagnosed DM, that is, diagnosed at the time of the survey. The study revealed that the prevalence of DR among persons with diabetes was 16.9%, the prevalence of STDR was 3.6%, and the prevalence of mild retinopathy was 11.8%. Risk factors for DR in the current study were duration of diabetes (>10 years, OR 4.8, 95% CI: 3.3–6.9), poor glycemic control (≥200 mg/dL, OR: 1.5, 95% CI: 1.2–1.7) and insulin treatment (OR: 2.6, 95% CI: 1.7–4.1).Conclusion:The current study highlights the substantial burden of DM and DR in India and the critical need to adopt a coordinated and multisectoral approach to reduce their prevalence. There is a need for early identification of persons with diabetes and their routine screening for DR along with availability of treatment facilities.     

Comparison of early diabetic retinopathy staging in asymptomatic patients between autonomous AI-based screening and human-graded ultra-widefield colour fundus images

IntroductionComparison of diabetic retinopathy (DR) severity between autonomous Artificial Intelligence (AI)-based outputs from an FDA-approved screening system and human retina specialists’ gradings from ultra-widefield (UWF) colour images.MethodsAsymptomatic diabetics without a previous diagnosis of DR were included in this prospective observational pilot study. Patients were imaged with autonomous AI (IDx-DR, Digital Diagnostics). For each eye, two 45° colour fundus images were analysed by a secure server-based AI algorithm. UWF colour fundus imaging was performed using Optomap (Daytona, Optos). The International Clinical DR severity score was assessed both on a 7-field area projection (7F-mask) according to the early treatment diabetic retinopathy study (ETDRS) and on the total gradable area (UWF full-field) up to the far periphery on UWF images.ResultsOf 54 patients included (n = 107 eyes), 32 were type 2 diabetics (11 females). Mean BCVA was 0.99 ± 0.25. Autonomous AI diagnosed 16 patients as negative, 28 for moderate DR and 10 for having a vision-threatening disease (severe DR, proliferative DR, diabetic macular oedema). Based on the 7F-mask grading with the eye with the worse grading defining the DR stage 23 patients were negative for DR, 11 showed mild, 19 moderate and 1 severe DR. When UWF full-field was analysed, 20 patients were negative for DR, while the number of mild, moderate and severe DR patients were 12, 21, and 1, respectively.ConclusionsThe autonomous AI-based DR examination demonstrates sufficient accuracy in diagnosing asymptomatic non-proliferative diabetic patients with referable DR even compared to UWF imaging evaluated by human experts offering a suitable method for DR screening.Subject terms: Retinal diseases, Medical imaging     

Regional variation in diabetic retinopathy and associated factors in Spectrum of Eye Disease in Diabetes (SPEED) study in India—Report 5

Purpose:To study the zonal variations in diabetic retinopathy (DR) and associated factors in people with known type 2 diabetes mellitus (T2DM) attending large eye care facilities in different regions of India.Methods:In this cross-sectional eye-care facility-based study, India was divided into five zones; large eye care facilities with a good referral base and offering an entire range of care for patients with DR were invited. First-time T2DM attendees aged ≥18 years were recruited. All subjects received a comprehensive systemic and ophthalmic examination. DR and systemic diseases were classified as per the international/national standards. Findings were compared between the zones and with the national average.Results:Fourteen eye-care facilities (15% public) from five zones participated. In the cohort of 11,173 people, there were more males (59%); the average age was above 45 years, and in 57%, DM had been diagnosed more than 5 years earlier. Compared with the overall study population, the proportion of people with any DR, sight-threatening DR, and blind were higher in the east zone (42.5%, 95% confidence interval [CI]: 40.2–44.8; 24.3%, 95% CI 22.3–26.3, and 11.5%, respectively); diabetic macular edema was more frequent in the south zone (12.2%, 95% CI 11.2–13.2); people with moderate-to-severe visual impairment were more in the west zone (32.1%) and higher proportion of people in the south-central zone had systemic hypertension (56.8%, 95% CI 54.8–58.9).Conclusion:The zonal variation in DR and related vision loss could be related to variable health-seeking behavior, availability, and confidence in the available services.     

Dietary associations with diabetic retinopathy—A cohort study

Purpose:To assess the role of dietary factors in the development of diabetic retinopathy (DR) in diabetics.Methods:This prospective study was carried out on patients attending the outpatient department of ophthalmology for a period of 1 year. An interview-based 24-hour diet recall was used to document average daily dietary nutrient intakes. Each patient was subjected to a comprehensive ocular examination to look for DR.Results:A total of 261 patients attending the outpatient department of ophthalmology were the participants for this study. The mean (±SD) age of the participants was 57.73 ± 11.29 years, and 67% were men. One hundred and six participants had DR. Univariate analysis revealed sex, duration, fish (times/week), egg (yes/no), rice lunch (yes/no), rice dinner, rice (boiled/white), and total calorie intake to be associated with DR (P < 0.05). Logistic regression multivariable analysis revealed males (OR: 3.20, 95% CI: 1.65–6.19), longer duration of diabetes (OR:1.05,95% CI:1.01-1.11), antioxidant intake (OR: 3.42, 95% CI: 1.65–7.05), and consumption of rice (OR: 3.19, 95% CI: 1.17–8.69) to have significant association with DR (P < 0.05), with the odds of developing DR increasing three times in these patients. The odds of developing DR were lesser with more frequent (>2 times/week) fish consumption (OR: 0.42, 95% CI: 0.18–0.94) and in patients on pharmacological treatment for diabetes mellitus (OR: 0.16, 95% CI: 0.04–0.58). Binary logistic regression revealed chapathi consumption (OR: 9.37, 95% CI: 1.64–53.68) to be associated with severe forms and fish consumption (OR: 0.06, 95% CI: 0.01–1.06) (P < 0.05) to be associated with less severe forms of DR.Conclusion:Males, longer duration of diabetes, antioxidant intake, fish consumption, and consumption of rice were associated with the occurrence of DR. Participants with diabetes who consumed fish more frequently and those who were on pharmacological treatment for diabetes mellitus had a significantly lower risk of DR and frequent fish consumption could reduce the risk of DR progression.     

Development and validation of an offline deep learning algorithm to detect vitreoretinal abnormalities on ocular ultrasound

Purpose:We describe our offline deep learning algorithm (DLA) and validation of its diagnostic ability to identify vitreoretinal abnormalities (VRA) on ocular ultrasound (OUS).Methods:Enrolled participants underwent OUS. All images were classified as normal or abnormal by two masked vitreoretinal specialists (AS, AM). A data set of 4902 OUS images was collected, and 4740 images of satisfactory quality were used. Of this, 4319 were processed for further training and development of DLA, and 421 images were graded by vitreoretinal specialists (AS and AM) to obtain ground truth. The main outcome measures were sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV) and area under receiver operating characteristic (AUROC).Results:Our algorithm demonstrated high sensitivity and specificity in identifying VRA on OUS ([90.8%; 95% confidence interval (CI): 86.1–94.3%] and [97.1% (95% CI: 93.7–98.9%], respectively). PPV and NPV of the algorithm were also high ([97.0%; 95% CI: 93.7–98.9%] and [90.8%; 95% CI: 86.2–94.3%], respectively). The AUROC was high at 0.939, and the intergrader agreement was nearly perfect with Cohen’s kappa of 0.938. The model demonstrated high sensitivity in predicting vitreous hemorrhage (100%), retinal detachment (97.4%), and choroidal detachment (100%)Conclusion:Our offline DLA software demonstrated reliable performance (high sensitivity, specificity, AUROC, PPV, NPV, and intergrader agreement) for predicting VRA on OUS. This might serve as an important tool for the ophthalmic technicians who are involved in community eye screening at rural settings where trained ophthalmologists are not available.     

Prevalence of diabetic retinopathy in urban and rural India: A systematic review and meta-analysis

A systematic review and meta-analysis were conducted to estimate the prevalence of diabetic retinopathy (DR) in India’s urban and rural areas. Medline, Scopus, and ScienceDirect databases were searched for population-based studies published in English between January 1990 and April 2021, wherein the prevalence of DR among Indian residents with type 2 diabetes mellitus (DM) was reported. A random-effects model was used to estimate the overall, rural, and urban prevalence. Data from 10 eligible studies were aggregated for meta-analysis. The prevalence of DR was 17.44% (95% confidence interval [CI], 14.33–20.55) in urban and 14.00% (95% CI: 9.13–18.86) in rural population (P = 0.24). The overall DR prevalence was 16.10% (95% CI: 13.16–24.32), and the population prevalence was 1.63% [95% CI: 0.94–2.32]. Prevalence of DR in people with diabetes was lower in the age group of 40–49 years [13.57% (95% CI: 7.16–19.98)] than in the age group of 50–59 years [16.72% (95% CI: 12.80–20.64)] and the age group of 60 years and above [16.55% (95% CI: 12.09–21.00)]. Variability in studies was high: urban (I² = 88.90%); rural (I² = 92.14%). Pooled estimates indicate a narrow difference in DR prevalence among people with diabetes in rural and urban India. The fast urbanization and increasing diabetes prevalence in rural areas underscore the need for providing equitable eye care at the bottom of the health pyramid.     

Screening for obstructive sleep apnea in a diabetic retinopathy clinic in a tertiary care center

Purpose:To screen for obstructive sleep apnea (OSA) in patients presenting to diabetic retinopathy (DR) clinic and to correlate its presence with the severity of DR.Methods:A prospective, cross-sectional study of diabetes mellitus patients in retina clinic of a tertiary care referral center, North India (January 2019–March 2020). All were subjected to STOP-Bang Questionnaire and Epworth Sleepiness Scale (ESS) score. Patients at high OSA risk (STOP-Bang score ≥5 and ESS score ≥10) were referred to Department of Otorhinolaryngology (sleep clinic) for polysomnography. Based on Apnea Hypopnea Index (AHI), OSA was graded as mild (AHI = 5–14/h), moderate (AHI = 15–30/h), and severe (AHI >30/h). Statistical analysis was done using three models of outcome measures: (1) “No DR” versus “any DR,” (2) “Less severe DR” versus “More severe DR,” and (3) “No diabetic macular edema (DME)” versus “DME.”Results:Of 362 patients screened, 18 (4.97%) had OSA (11 mild, 5 moderate, and 2 severe). Though OSA did not show a significant association with various outcome measures, patients with moderate–severe OSA had higher odds in developing “any DR” (OR = 7.408; 95% CI = 0.533–102.898), “more severe DR” (OR = 1.961; 95% CI = 0.153–25.215), and “DME” (OR = 2.263; 95% CI = 0.357–14.355), on multiple logistic regression.Conclusion:Ours is the first screening study of OSA in DR patients in India, the diabetes capital of the world. We detected OSA in 4.97% of patients in a DR clinic, with an increased risk of “any DR,” “more severe DR,” and “DME” in the presence of moderate–severe OSA.     

Prevalence and Associated Factors of Diabetic Retinopathy among Type 2 Diabetes Mellitus Patients in Brunei Darussalam: A Cross-sectional Study

PurposeTo determine the prevalence of diabetic retinopathy (DR) and the factors associated with retinopathy among type 2 diabetes mellitus (DM) patients in Brunei Darussalam.MethodsCross-sectional study of all type 2 DM patients who attended diabetic eye screening over a 3-month period at one of four government hospitals. We assessed association between DR with the following variables: age, sex, glycated hemoglobin (HbA1c), duration of DM, hypertension, hyperlipidemia, and microalbuminuria.ResultsThere were 341 patients (female, 58.9%; mean age, 55.3 ± 11.9 years) with a mean duration of DM of 9.4 ± 7.4 years and mean serum HbA1c of 8.4% ± 1.9%. The overall prevalence of any DR was 22.6% (95% confidence interval, 18.8–27.1) with prevalence rates of 4.1% (95% confidence interval, 2.1–6.4) for proliferative DR and 9.7% (95% confidence interval, 6.8–13.2) for vision-threatening DR. Multivariate analysis showed that DR was significantly associated with certain age groups (reduced in older age groups), longer duration of DM (11 years or more), poor control (HbA1c >9.0%) and presence of any microalbuminuria.ConclusionsDR affects one in five patients with DM in Brunei Darussalam, comparable to rates reported for other Asian populations. It is especially worrying that one in ten patients with DM had vision-threatening DR. DR was significantly associated with longer duration of DM, poor control and presence of microalbuminuria but reduced in older age groups. It is important to advocate good control right from the time of diagnosis of DM and institute timely and effective management of retinopathy. DR was significantly associated with longer duration of DM, poor control of diabetes, and presence of microalbuminuria but reduced in older age groups.     

Diabetic retinopathy screening: can the viewing monitor influence the reading and grading outcomes

Purpose

To evaluate the accuracy of different viewing monitors for image reading and grading of diabetic retinopathy (DR).

Design

Single-centre, experimental case series—evaluation of reading devices for DR screening.

Method

A total of 100 sets of three-field (optic disc, macula, and temporal views) colour retinal still images (50 normal and 50 with DR) captured by FF 450 plus (Carl Zeiss) were interpreted on 27-inch iMac, 15-inch MacBook Pro, and 9.7-inch iPad. All images were interpreted by a retinal specialist and a medical officer. We calculated the sensitivity and specificity of 15-inch MacBook Pro and 9.7-inch iPad in detection of DR signs and grades with reference to the reading outcomes obtained using a 27-inch iMac reading monitor.

Results

In detection of any grade of DR, the 15-inch MacBook Pro had sensitivity and specificity of 96% (95% confidence interval (CI): 85.1–99.3) and 96% (95% CI: 85.1–99.3), respectively, for retinal specialist and 91.5% (95% CI: 78.7–97.2) and 94.3% (95% CI: 83.3–98.5), respectively, for medical officer, whereas for 9.7-inch iPad, they were 91.8% (95% CI: 79.5–97.4) and 94.1% (95% CI: 82.8–98.5), respectively, for retinal specialist and 91.3% (95% CI: 78.3–97.1) and 92.6% (95% CI: 81.3–97.6), respectively, for medical officer.

Conclusion

The 15-inch MacBook Pro and 9.7-inch iPad had excellent sensitivity and specificity in detecting DR and hence, both screen sizes can be utilized to effectively interpret colour retinal still images for DR remotely in a routine, mobile or tele-ophthalmology setting. Future studies could explore the use of more economical devices with smaller viewing resolutions to reduce cost implementation of DR screening services.     

Failure to initiate early insulin therapy – A risk factor for diabetic retinopathy in insulin users with Type 2 diabetes mellitus: Sankara Nethralaya-Diabetic Retinopathy Epidemiology and Molecular Genetics Study (SN-DREAMS,Report number 35)

Context:Insulin users have been reported to have a higher incidence of diabetic retinopathy (DR).Aim:The aim was to elucidate the factors associated with DR among insulin users, especially association between duration, prior to initiating insulin for Type 2 diabetes mellitus (DM) and developing DR.Results:Insulin users had more incidence of DR (52.9% vs. 16.3%, P < 0.0001) and sight threatening DR (19.1% vs. 2.4%, P < 0.0001) in comparison to insulin nonusers. Among insulin users, longer duration of DM (odds ratio [OR] 1.12, 95% confidence interval [CI] 1.00–1.25, P = 0.044) and abdominal obesity (OR 1.15, 95% CI 1.02–1.29, P = 0.021) was associated with DR. The presence of DR was significantly associated with longer duration (≥5 years) prior to initiating insulin therapy, overall (38.0% vs. 62.0%, P = 0.013), and in subjects with suboptimal glycemic control (32.5% vs. 67.5%, P = 0.022).Conclusions:The presence of DR is significantly associated with longer duration of diabetes (>5 years) and sub-optimal glycemic control (glycosylated hemoglobin <7.0%). Among insulin users, abdominal obesity was found to be a significant predictor of DR; DR is associated with longer duration prior to initiating insulin therapy in Type 2 DM subjects with suboptimal glycemic control.     

Does myopia decrease the risk of diabetic retinopathy in both type-1 and type-2 diabetes mellitus?

Purpose:To study the relationship between the severity of myopia and the severity of diabetic retinopathy (DR) in individuals with type 1 or type 2 diabetes mellitus (DM).Methods:This retrospective study was conducted using data from electronic medical records from a multicentric eyecare network located in various geographic regions of India. Individuals with type 1 or type 2 DM were classified according to their refractive status. Severe nonproliferative DR (NPDR), PDR, or presence of clinically significant macular edema (CSME) with any type of DR was considered as vision-threatening diabetic retinopathy (VTDR).Results:A total of 472 individuals with type-1 DM (mean age 41 ± 10 years) and 9341 individuals with type-2 DM (52 ± 9 years) were enrolled. Individuals with a hyperopic refractive error had a significant positive association with the diagnosis of VTDR (odds ratio (OR) 1.26; 95%CI 1.04–1.51, P = 0.01) and moderate nonproliferative DR (OR 1.27; 95%CI 1.02–1.59, P = 0.03) in type-2 DM; however, no significant association was found in type-1 DM. After adjusting for age, gender, anisometropia, and duration of diabetes, the presence of high myopia (< - 6 D) reduced the risk of VTDR in type 2 DM (OR 0.18; 95% CI 0.04–0.77, P = 0.02), but no association was found in type 1 DM. Mild and moderate myopia had no significant association with any forms of DR in both type-1 and type-2 DM.Conclusion:Hyperopic refractive error was found to increase the risk of VTDR in persons with type 2 DM. High-myopic refractive error is protective for VTDR in type 2 DM, but not in type-1 DM.     

Associations Between Peripapillary Retinal Nerve Fiber Layer and Choroidal Thickness With the Development and Progression of Diabetic Retinopathy

PurposeTo evaluate the role of the peripapillary retinal nerve fiber layer (pRNFL) and peripapillary choroidal thickness (pCT) in the development and progression of diabetic retinopathy (DR).MethodsThis is a cohort study based on the baseline and 2-year follow-up data of the Guangzhou Diabetic Eye Study. Patients with type 2 diabetes mellitus between the ages of 30 and 80 years were recruited from communities in Guangzhou. DR was graded by seven-field fundus photography after dilation of the pupil. pRNFL and pCT were measured via swept-source optical coherence tomography.ResultsA total of 895 patients were included in the study; of these, 748 did not have DR at baseline and 147 had DR at baseline. During the 2-year follow-up, 80 developed DR (10.7%), and 11 experienced DR progression (7.5%). After adjusting for confounding factors, a higher risk of incident DR was strongly associated with a lower average thickness of the pRNFL (risk ratio [RR] per 1 SD, 0.55; 95% confidence interval [CI], 0.42–0.72; P < 0.001) and average pCT (RR per 1 SD, 0.49; 95% CI, 0.34–0.70; P < 0.001). Adding both metrics to the DR prediction model significantly improved the discriminant ability of the model for incidences of DR (area under the curve increased by 15.38% from 0.673 to 0.777; P < 0.001).ConclusionsNeurodegeneration shown by the thinning of pRNFL and impaired choroidal circulation shown by the thinning of pCT are independently associated with DR onset, and assessing both metrics can improve the risk assessment for DR incidences.     

Automated feature-based grading and progression analysis of diabetic retinopathy

BackgroundIn diabetic retinopathy (DR) screening programmes feature-based grading guidelines are used by human graders. However, recent deep learning approaches have focused on end to end learning, based on labelled data at the whole image level. Most predictions from such software offer a direct grading output without information about the retinal features responsible for the grade. In this work, we demonstrate a feature based retinal image analysis system, which aims to support flexible grading and monitor progression.MethodsThe system was evaluated against images that had been graded according to two different grading systems; The International Clinical Diabetic Retinopathy and Diabetic Macular Oedema Severity Scale and the UK’s National Screening Committee guidelines.ResultsExternal evaluation on large datasets collected from three nations (Kenya, Saudi Arabia and China) was carried out. On a DR referable level, sensitivity did not vary significantly between different DR grading schemes (91.2–94.2.0%) and there were excellent specificity values above 93% in all image sets. More importantly, no cases of severe non-proliferative DR, proliferative DR or DMO were missed.ConclusionsWe demonstrate the potential of an AI feature-based DR grading system that is not constrained to any specific grading scheme.Subject terms: Diagnosis, Medical imaging     

Hypertensive Retinopathy and the Risk of Stroke Among Hypertensive Adults in China

PurposeThis study aimed to investigate the association between hypertensive retinopathy and the risk of first stroke, examine possible effect modifiers in hypertensive patients, and test the appropriateness of the Keith-Wagener-Barker (KWB) classification for predicting stroke risk.MethodsIn total, 9793 hypertensive participants (3727 males and 6066 females) without stroke history from the China Stroke Primary Prevention Trial were included in this study. The primary outcome was first stroke.ResultsOver a median follow-up of 4.4 years, 592 participants experienced their first stroke (509 ischemic, 77 hemorrhagic, and six unclassifiable strokes). In total, 5590 participants were diagnosed with grade 1 retinopathy (57.08%), 1466 with grade 2 retinopathy (14.97%), 231 with grade 3 retinopathy (2.36%), and three with grade 4 retinopathy (0.03%). Grades 1 and 2 were merged and classified as mild retinopathy, and grades 3 and 4 were merged and classified as severe retinopathy. There was a significant positive association between hypertensive retinopathy and the risk of first stroke and first ischemic stroke, and no effect modifiers were found. The hazard ratios (HRs) for first stroke were as follows: mild versus no retinopathy, 1.26 (95% confidence interval [CI], 1.01–1.58, P = 0.040), and severe versus no retinopathy, 2.40 (95% CI, 1.49–3.84, P < 0.001). The HRs for ischemic stroke were as follows: severe versus no retinopathy, 2.35 (95% CI, 1.41–3.90, P = 0.001), and nonsignificantly increased HRs for mild versus no retinopathy, 1.26 (95% CI, 0.99–1.60, P = 0.057).ConclusionsThere was a significant positive association between hypertensive retinopathy and the risk of first stroke in patients with hypertension, indicating that hypertensive retinopathy may be a predictor of the risk of stroke. A simplified two-grade classification system based on the KWB classification is recommended for predicting stroke risk.