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.

Automated image curation in diabetic retinopathy screening using deep learning.

Diabetic retinopathy (DR) screening images are heterogeneous and contain undesirable non-retinal, incorrect field and ungradable samples which require curation, a laborious task to perform manually. We developed and validated single and multi-output laterality, retinal presence, retinal field and gradability classification deep learning (DL) models for automated curation. The internal dataset comprised of 7743 images from DR screening (UK) with 1479 external test images (Portugal and Paraguay). Internal vs external multi-output laterality AUROC were right (0.994 vs 0.905), left (0.994 vs 0.911) and unidentifiable (0.996 vs 0.680). Retinal presence AUROC were (1.000 vs 1.000). Retinal field AUROC were macula (0.994 vs 0.955), nasal (0.995 vs 0.962) and other retinal field (0.997 vs 0.944). Gradability AUROC were (0.985 vs 0.918). DL effectively detects laterality, retinal presence, retinal field and gradability of DR screening images with generalisation between centres and populations. DL models could be used for automated image curation within DR screening.

Multicenter Evaluation of Diagnostic Circulating Biomarkers to Detect Sight-Threatening Diabetic Retinopathy.

Importance: It is a global challenge to provide regular retinal screening for all people with diabetes to detect sight-threatening diabetic retinopathy (STDR). Objective: To determine if circulating biomarkers could be used to prioritize people with type 2 diabetes for retinal screening to detect STDR. Design, Setting, and Participants: This cross-sectional study collected data from October 22, 2018, to December 31, 2021. All laboratory staff were masked to the clinical diagnosis, assigned a study cohort, and provided with the database containing the clinical data. This was a multicenter study conducted in parallel in 3 outpatient ophthalmology clinics in the UK and 2 centers in India. Adults 40 years and older were categorized into 4 groups: (1) no history of diabetes, (2) type 2 diabetes of at least 5 years' duration with no evidence of DR, (3) nonproliferative DR with diabetic macular edema (DME), or (4) proliferative DR. STDR comprised groups 3 and 4. Exposures: Thirteen previously verified biomarkers were measured using enzyme-linked immunosorbent assay. Main Outcomes and Measures: Severity of DR and presence of DME were diagnosed using fundus photographs and optical coherence tomography. Weighted logistic regression and receiver operating characteristic curve analysis (ROC) were performed to identify biomarkers that discriminate STDR from no DR beyond the standard clinical parameters of age, disease duration, ethnicity (in the UK) and hemoglobin A1c. Results: A total of 538 participants (mean [SD] age, 60.8 [9.8] years; 319 men [59.3%]) were recruited into the study. A total of 264 participants (49.1%) were from India (group 1, 54 [20.5%]; group 2, 53 [20.1%]; group 3, 52 [19.7%]; group 4, 105 [39.8%]), and 274 participants (50.9%) were from the UK (group 1, 50 [18.2%]; group 2, 70 [25.5%]; group 3, 55 [20.1%]; group 4, 99 [36.1%]). ROC analysis (no DR vs STDR) showed that in addition to age, disease duration, ethnicity (in the UK) and hemoglobin A1c, inclusion of cystatin C had near-acceptable discrimination power in both countries (area under the receiver operating characteristic curve [AUC], 0.779; 95% CI, 0.700-0.857 in 215 patients in the UK with complete data; AUC, 0.696; 95% CI, 0.602-0.791 in 208 patients in India with complete data). Conclusions and Relevance: Results of this cross-sectional study suggest that serum cystatin C had good discrimination power in the UK and India. Circulating cystatin-C levels may be considered as a test to identify those who require prioritization for retinal screening for STDR.

Returning to cataract surgery after a hiatus: a UK survey report.

OBJECTIVES: Cataract surgeons may periodically take time away from operating which can lead to skills fade. There is a paucity of research investigating the experiences of returning cataract surgeons and how different individual circumstances impact on their return. Our aim was to investigate the subjective experiences of UK ophthalmologists simultaneously returning to surgery following the nationwide elective surgical hiatus due to the Covid-19 pandemic. METHODS: An online survey was nationally distributed between 01/09/2020 and 29/10/2020 to registered UK ophthalmologists. Participants indicating a surgical hiatus of 8 weeks or more were included. RESULTS: 232 of 264 responses were analysed. Covid-19 was the most frequent reason for a surgical hiatus (median 15 weeks). Perceived operating difficulties were found in 29.1%. Transient anxiety (51.7%), reduced confidence, and perceived increased surgical time were commonplace. Trainees and females were more likely to encounter negative experiences (p < 0.001) and barriers to resource accessibility. Eyesi® and online videos were the most available and accessed pre-return resources. Childcare was five times more likely to present as a barrier to resource access for females than males. CONCLUSIONS: Technical skills fade such as capsulorhexis difficulties were commonly perceived by trainee surgeons in addition to transient anxiety, reported in more than half of all surgeons following a hiatus as short as 8 weeks. Eyesi® simulation offers the potential to negate technical de-skilling. Few had formal return plans or awareness of RCOphth guidance. We believe there is scope for more specialised and targeted support for future returning cataract surgeons thereby optimising patient care.

Deep learning for gradability classification of handheld, non-mydriatic retinal images.

Screening effectively identifies patients at risk of sight-threatening diabetic retinopathy (STDR) when retinal images are captured through dilated pupils. Pharmacological mydriasis is not logistically feasible in non-clinical, community DR screening, where acquiring gradable retinal images using handheld devices exhibits high technical failure rates, reducing STDR detection. Deep learning (DL) based gradability predictions at acquisition could prompt device operators to recapture insufficient quality images, increasing gradable image proportions and consequently STDR detection. Non-mydriatic retinal images were captured as part of SMART India, a cross-sectional, multi-site, community-based, house-to-house DR screening study between August 2018 and December 2019 using the Zeiss Visuscout 100 handheld camera. From 18,277 patient eyes (40,126 images), 16,170 patient eyes (35,319 images) were eligible and 3261 retinal images (1490 patient eyes) were sampled then labelled by two ophthalmologists. Compact DL model area under the receiver operator characteristic curve was 0.93 (0.01) following five-fold cross-validation. Compact DL model agreement (Kappa) were 0.58, 0.69 and 0.69 for high specificity, balanced sensitivity/specificity and high sensitivity operating points compared to an inter-grader agreement of 0.59. Compact DL gradability model performance was favourable compared to ophthalmologists. Compact DL models can effectively classify non-mydriatic, handheld retinal image gradability with potential applications within community-based DR screening.

Multimodal Imaging in a Case of Peripapillary Choroidal Neovascular Membrane Associated With Idiopathic Intracranial Hypertension.

ABSTRACT: Optical coherence tomography angiography is one of the latest noninvasive imaging modalities for visualizing the vasculature of retina and choroid. We describe its application in the diagnosis, treatment, and monitoring of a patient with peripapillary choroidal neovascular membrane in the setting of idiopathic intracranial hypertension, who responded well to a course of ranibizumab intravitreal injections.

Factors affecting cataract surgery operating time among trainees and consultants.

PURPOSE: To define factors affecting cataract surgery operating time for operating room planning, optimizing throughput, enhancing patient experiences, minimizing costs, and allocating training time. SETTING: Epsom and St. Helier University National Health Service Trust, London, United Kingdom. DESIGN: Retrospective case series. METHODS: All patients who had primary manual phacoemulsification cataract surgery from January 1, 2012, to December 30, 2016, were included. Combined anterior and posterior segment procedures and surgeons with fewer than 50 cases were excluded. Anonymized data collected were demographics, anesthesia, operating time, surgeon grade, case complexity, pupil size, pupil expander or capsular tension ring (CTR) use, intraocular lens type, posterior capsule or zonular fiber rupture or dialysis, vitreous loss, and automated anterior vitrectomy. RESULTS: From 11 067 cases, 9552 (86.3%) had a recorded operating time. The mean ± SD operating times in minutes were as follows: consultants 19 ± 10, junior 30 ± 11, intermediate 27 ± 12, senior trainees 24 ± 10, and fellows 31 ± 11. Operating time was significantly shorter for topical than for sub-Tenon or general anesthesia, especially among trainees. Consultant operating time remained unchanged with increasing case complexity, except for high-complexity cases. Small pupils, pupil expander or CTR use, posterior capsule or zonular fiber rupture or dialysis with or without vitreous loss (mean 45 ± 23) were associated with increased operating times. Iris hooks were associated with greater increases in operating time than Malyugin rings (16 minutes versus 6 minutes; P < .001). There was a modest 3-minute decrease in operating time among consultants over 5 years. CONCLUSION: Cataract surgery operating time was significantly influenced by anesthesia type, surgeon grade, high case complexity, pupil size, pupil expander use/type, CTR use, and intraoperative complications.

Iris hooks versus a pupil expansion ring: Operating times, complications, and visual acuity outcomes in small pupil cases.

PURPOSE: To compare iris hooks and a pupil expansion ring (Malyugin ring) in eyes with a small pupil. SETTING: Epsom and St. Helier University National Health Service Trust, London, United Kingdom. DESIGN: Retrospective case series. METHODS: Patients who had primary phacoemulsification cataract surgery from January 1, 2012, to December 30, 2016, were included. Combined procedures and surgeons with fewer than 50 cases were excluded. Anonymized data were demographics, surgeon grade, case complexity, iris hooks or Malyugin ring use, posterior capsule rupture, zonular fiber rupture or dialysis, vitreous loss, operating time, postoperative complications, and visual acuity. RESULTS: Of the 9552 cases included, 425 (4.4%) had a small pupil, 314 of which required a pupil expansion ring and 95 iris hooks. Compared with no pupil expander, iris hooks were associated with an additional operating time of 14 minutes for consultants and 24 minutes for trainees. The pupil expansion ring was associated with an additional operating time of 4 minutes and 6 minutes, respectively. Neither pupil expander was associated with more intraoperative complications. The pupil expansion ring was associated with higher postoperative anterior uveitis and corneal edema rates. Pupil expander cases achieved equivalent visual acuity gains. CONCLUSIONS: The pupil expansion ring and iris hooks were safe and effective in minimizing intraoperative complications in eyes with a small pupil. The pupil expansion ring was faster to use than iris hooks. Monitoring for signs of postoperative anterior uveitis or cornea edema with careful insertion or removal of pupil expansion rings is advocated.

Updated cataract surgery complexity stratification score for trainee ophthalmic surgeons.

PURPOSE: To devise a comprehensive cataract surgery complexity score system for the selection of appropriate cases for trainees using evidence-based, validated risk factors for posterior capsule rupture, patient-specific factors, and complexity stratification recommendations to minimize complications, optimize outcomes, and maximize patient safety. SETTING: Epsom and St. Helier University National Health Service Trust, London, United Kingdom. DESIGN: Retrospective cohort study. METHODS: Patients having primary phacoemulsification cataract surgery from January 1, 2011 until December 31, 2016 were included; combined corneal, glaucoma, or posterior segment procedures were excluded. Anonymized data on demographics, pupil size, pupil expander use, intraoperative and postoperative complications, and postoperative distance visual acuity were extracted. Patients were stratified by complexity score and surgeon grade (consultant, junior, intermediate, and senior trainee, and fellow). RESULTS: From 11 468 included cases, 8200 (71.5%) had a complexity score. Small pupil, pupil expander use, iris damage during phacoemulsification, zonular dialysis, postoperative raised intraocular pressure, and corneal edema (odds ratio, 3.17; 95% confidence interval, 2.05-4.92) were significantly associated with increasing complexity. Appropriate case allocation by complexity and surgeon grade resulted in no association between posterior capsule rupture and complexity score. Increasing complexity scores were associated with lower postoperative distance visual acuity. CONCLUSIONS: The updated evidence-based comprehensive cataract complexity score system is a useful tool for the stratification of case complexity and guides appropriate case selection to match trainee experience. Higher complexity scores were associated with greater intraoperative and postoperative complications and lower postoperative distance visual acuity. Patients with higher scores should be made aware of the guarded prognosis when obtaining consent.

The association between individual metabolic syndrome components, primary liver cancer and cirrhosis: A study in the Swedish AMORIS cohort.

Metabolic syndrome (MetS) is associated with non-alcoholic fatty liver disease, which may progress to cirrhosis, a significant risk factor of hepatocellular carcinoma (HCC), the commonest malignant primary liver cancer (PLC). We investigated the association between the individual components of MetS (lipids, apolipoproteins, raised glucose, diabetes and obesity), PLC and cirrhosis. A total of 509,436 participants from the Swedish AMORIS cohort, recruited between January 1985 and December 1996 (end-date December 2011), aged ≥20 with baseline triglycerides (TG), total cholesterol (TC), glucose and liver enzymes were included. Those with baseline benign liver tumours, PLC or cirrhosis were excluded. Multivariate Cox regression, adjusted for age, gender, socio-economic status, liver disease (excluding cirrhosis) and MetS factors were used to estimate the association with PLC and cirrhosis. There were 766 PLC and 2,775 cirrhosis cases over 13 years. Raised TG, low TC, raised glucose, diabetes and low HDL were associated with an increased risk of developing PLC and cirrhosis. ApoB/ApoA-I ratio were also associated with PLC, whilst low LDL, raised TG/HDL, low ApoA-I and low ApoB were associated with cirrhosis. Obesity was significantly associated with PLC but not cirrhosis. Raised TG, low TC, raised glucose and diabetes showed stronger associations with PLC in participants with cirrhosis but many participants developed PLC without cirrhosis. Individual components of MetS (lipids, apolipoproteins, raised glucose, diabetes and obesity) were associated with an increased risk of developing PLC or cirrhosis. MetS components were more strongly associated with PLC with preceding cirrhosis history but many participants developed PLC without cirrhosis.

Prostate-specific antigen testing in inner London general practices: are those at higher risk most likely to get tested?

OBJECTIVES: To investigate the association between factors influencing prostate-specific antigen (PSA) testing prevalence including prostate cancer risk factors (age, ethnicity, obesity) and non-risk factors (social deprivation and comorbidity). SETTING: A cross-sectional database of 136 inner London general practices from 1 August 2009 to 31 July 2014. PARTICIPANTS: Men aged ≥40 years without prostate cancer were included (n=150 481). PRIMARY OUTCOME: Logistic regression analyses were used to estimate the association between PSA testing and age, ethnicity, social deprivation, body mass index (BMI) and comorbidity while adjusting for age, benign prostatic hypertrophy, prostatitis and tamsulosin or finasteride use. RESULTS: PSA testing prevalence was 8.2% (2013-2014), and the mean age was 54 years (SD 11). PSA testing was positively associated with age (OR 70-74 years compared to 40-44 years: 7.34 (95% CI 6.82 to 7.90)), ethnicity (black) (OR compared to white: 1.78 (95% CI 1.71 to 1.85)), increasing BMI and cardiovascular comorbidity. Testing was negatively associated with Chinese ethnicity and with increasing social deprivation. CONCLUSIONS: PSA testing among black patients was higher compared to that among white patients, which differs from lower testing rates seen in previous studies. PSA testing was positively associated with prostate cancer risk factors and non-risk factors. Association with non-risk factors may increase the risk of unnecessary invasive diagnostic procedures.