Testing the performance of risk prediction models to determine progression to referable diabetic retinopathy in an Irish type 2 diabetes cohort.

BACKGROUND /AIMS: To evaluate the performance of existing prediction models to determine risk of progression to referable diabetic retinopathy (RDR) using data from a prospective Irish cohort of people with type 2 diabetes (T2D). METHODS: A cohort of 939 people with T2D followed prospectively was used to test the performance of risk prediction models developed in Gloucester, UK, and Iceland. Observed risk of progression to RDR in the Irish cohort was compared with that derived from each of the prediction models evaluated. Receiver operating characteristic curves assessed models' performance. RESULTS: The cohort was followed for a total of 2929 person years during which 2906 screening episodes occurred. Among 939 individuals followed, there were 40 referrals (4%) for diabetic maculopathy, pre-proliferative DR and proliferative DR. The original Gloucester model, which includes results of two consecutive retinal screenings; a model incorporating, in addition, systemic biomarkers (HbA1c and serum cholesterol); and a model including results of one retinopathy screening, HbA1c, total cholesterol and duration of diabetes, had acceptable discriminatory power (area under the curve (AUC) of 0.69, 0.76 and 0.77, respectively). The Icelandic model, which combined retinopathy grading, duration and type of diabetes, HbA1c and systolic blood pressure, performed very similarly (AUC of 0.74). CONCLUSION: In an Irish cohort of people with T2D, the prediction models tested had an acceptable performance identifying those at risk of progression to RDR. These risk models would be useful in establishing more personalised screening intervals for people with T2D.

Standardization of Optical Coherence Tomography Angiography Imaging Biomarkers in Diabetic Retinal Disease.

Optical coherence tomography Angiography (OCT-A) represents a revolution in the noninvasive evaluation of retinal and choroidal circulation especially in detecting early clinical signs of diabetic retinal disease (DRD). With appropriate use, OCT-A characteristics and measurements have the potential to become new imaging biomarkers in managing and treating DRD. Major challenges include (a) provision of standardized outputs from different OCT-A instruments providing standardized terminology to correctly interpret data; (b) the presence of artifacts; (c) the absence of standardized grading or interpretation method in the evaluation of DRD, similar to that already established in fundus photography; and (d) establishing how OCT-A might be able to provide surrogate markers to demonstrate blood retinal barrier breakdown and vascular leakage, commonly associated with DRD. In fact, OCT-A guidelines for DRD are still evolving. The outputs of quantitative OCT-A data offer a unique opportunity to develop tools based on artificial intelligence to assist the clinicians in diagnosing, monitoring, and managing patients with diabetes. In addition, OCT-A has the potential to become a useful tool for the evaluation of cardiovascular diseases and different neurological diseases including cognitive impairment. This article written by the members of Diabetic Retinopathy expert committee of the European Vision Clinical Research network will review the available evidence on the use of OCT-A as an imaging biomarker in DRD and discuss the limits and the current application as well as future developments for its use in both clinical practice and research trials of DRD.

Prospective evaluation of an artificial intelligence-enabled algorithm for automated diabetic retinopathy screening of 30 000 patients.

BACKGROUND/AIMS: Human grading of digital images from diabetic retinopathy (DR) screening programmes represents a significant challenge, due to the increasing prevalence of diabetes. We evaluate the performance of an automated artificial intelligence (AI) algorithm to triage retinal images from the English Diabetic Eye Screening Programme (DESP) into test-positive/technical failure versus test-negative, using human grading following a standard national protocol as the reference standard. METHODS: Retinal images from 30 405 consecutive screening episodes from three English DESPs were manually graded following a standard national protocol and by an automated process with machine learning enabled software, EyeArt v2.1. Screening performance (sensitivity, specificity) and diagnostic accuracy (95% CIs) were determined using human grades as the reference standard. RESULTS: Sensitivity (95% CIs) of EyeArt was 95.7% (94.8% to 96.5%) for referable retinopathy (human graded ungradable, referable maculopathy, moderate-to-severe non-proliferative or proliferative). This comprises sensitivities of 98.3% (97.3% to 98.9%) for mild-to-moderate non-proliferative retinopathy with referable maculopathy, 100% (98.7%,100%) for moderate-to-severe non-proliferative retinopathy and 100% (97.9%,100%) for proliferative disease. EyeArt agreed with the human grade of no retinopathy (specificity) in 68% (67% to 69%), with a specificity of 54.0% (53.4% to 54.5%) when combined with non-referable retinopathy. CONCLUSION: The algorithm demonstrated safe levels of sensitivity for high-risk retinopathy in a real-world screening service, with specificity that could halve the workload for human graders. AI machine learning and deep learning algorithms such as this can provide clinically equivalent, rapid detection of retinopathy, particularly in settings where a trained workforce is unavailable or where large-scale and rapid results are needed.

Teleophthalmology: a strategy for timely diagnosis of sight-threatening diabetic retinopathy in primary care, Concepción, Chile.

AIM: To estimate the prevalence of diabetic retinopathy (DR) in a diabetic population of the public health system. METHODS: This non-experimental, descriptive and cross-sectional study, with no direct control over the variables analysed, was carried out by the Telemedicine Unit of the University of Concepción (TELMED) and the Family Health Centers (CESFAM) of the Health Service Concepción, Chile, among a population of 7382 diabetic patients of the public health system, from October 2014 to June 2015. Two digital images for each patient's eyes were obtained and uploaded to the TELMED tele-ophthalmology platform. The images were remotely evaluated by a retina expert ophthalmologist for diagnosis. Treatment and a referral (if required) were recommended, and an individualised report for each patient was sent to the place of origin. RESULTS: The prevalence of DR in this study was 14.9%. Of these, 48.7% had mild non-proliferative DR (NPDR), 30.8% moderate NPDR, 15.9% severe NPDR, and 4.6% proliferative DR. Of the patients with DR, 17.8% were referred for fluorescein angiography, 4.3% for panretinal photocoagulation, 1.2% for vitrectomy, and 0.4% for cataract surgery. CONCLUSION: The use of telemedicine allowed an increased screening coverage for DR in diabetic patients. The different types of DR were detected in a timely manner, decreasing waiting times due to the lack of specialists in ophthalmology in the public health system.

The influence of age, duration of diabetes, cataract, and pupil size on image quality in digital photographic retinal screening.

OBJECTIVE: To evaluate the effect of age, duration of diabetes, cataract, and pupil size on the image quality in digital photographic screening. RESEARCH DESIGN AND METHODS: Randomized groups of 3,650 patients had one-field, non-mydriatic, 45 degrees digital retinal imaging photography before mydriatic two-field photography. A total of 1,549 patients were then examined by an experienced ophthalmologist. Outcome measures were ungradable image rates, age, duration of diabetes, detection of referable diabetic retinopathy, presence of early or obvious central cataract, pupil diameter, and iris color. RESULTS: The ungradable image rate for non-mydriatic photography was 19.7% (95% CI 18.4-21.0) and for mydriatic photography was 3.7% (3.1-4.3). The odds of having one eye ungradable increased by 2.6% (1.6-3.7) for each extra year since diagnosis for nonmydriatic, by 4.1% (2.7-5.7) for mydriatic photography irrespective of age, by 5.8% (5.0-6.7) for non-mydriatic, and by 8.4% (6.5-10.4) for mydriatic photography for every extra year of age, irrespective of years since diagnosis. Obvious central cataract was present in 57% of ungradable mydriatic photographs, early cataract in 21%, no cataract in 9%, and 13% had other pathologies. The pupil diameter in the ungradable eyes showed a significant trend (P < 0.001) in the three groups (obvious cataract 4.434, early cataract 3.379, and no cataract 2.750). CONCLUSIONS: The strongest predictor of ungradable image rates, both for non-mydriatic and mydriatic digital photography, is the age of the person with diabetes. The most common cause of ungradable images was obvious central cataract.