Development and international validation of custom-engineered and code-free deep-learning models for detection of plus disease in retinopathy of prematurity: a retrospective study.

BACKGROUND: Retinopathy of prematurity (ROP), a leading cause of childhood blindness, is diagnosed through interval screening by paediatric ophthalmologists. However, improved survival of premature neonates coupled with a scarcity of available experts has raised concerns about the sustainability of this approach. We aimed to develop bespoke and code-free deep learning-based classifiers for plus disease, a hallmark of ROP, in an ethnically diverse population in London, UK, and externally validate them in ethnically, geographically, and socioeconomically diverse populations in four countries and three continents. Code-free deep learning is not reliant on the availability of expertly trained data scientists, thus being of particular potential benefit for low resource health-care settings. METHODS: This retrospective cohort study used retinal images from 1370 neonates admitted to a neonatal unit at Homerton University Hospital NHS Foundation Trust, London, UK, between 2008 and 2018. Images were acquired using a Retcam Version 2 device (Natus Medical, Pleasanton, CA, USA) on all babies who were either born at less than 32 weeks gestational age or had a birthweight of less than 1501 g. Each images was graded by two junior ophthalmologists with disagreements adjudicated by a senior paediatric ophthalmologist. Bespoke and code-free deep learning models (CFDL) were developed for the discrimination of healthy, pre-plus disease, and plus disease. Performance was assessed internally on 200 images with the majority vote of three senior paediatric ophthalmologists as the reference standard. External validation was on 338 retinal images from four separate datasets from the USA, Brazil, and Egypt with images derived from Retcam and the 3nethra neo device (Forus Health, Bengaluru, India). FINDINGS: Of the 7414 retinal images in the original dataset, 6141 images were used in the final development dataset. For the discrimination of healthy versus pre-plus or plus disease, the bespoke model had an area under the curve (AUC) of 0·986 (95% CI 0·973-0·996) and the CFDL model had an AUC of 0·989 (0·979-0·997) on the internal test set. Both models generalised well to external validation test sets acquired using the Retcam for discriminating healthy from pre-plus or plus disease (bespoke range was 0·975-1·000 and CFDL range was 0·969-0·995). The CFDL model was inferior to the bespoke model on discriminating pre-plus disease from healthy or plus disease in the USA dataset (CFDL 0·808 [95% CI 0·671-0·909, bespoke 0·942 [0·892-0·982]], p=0·0070). Performance also reduced when tested on the 3nethra neo imaging device (CFDL 0·865 [0·742-0·965] and bespoke 0·891 [0·783-0·977]). INTERPRETATION: Both bespoke and CFDL models conferred similar performance to senior paediatric ophthalmologists for discriminating healthy retinal images from ones with features of pre-plus or plus disease; however, CFDL models might generalise less well when considering minority classes. Care should be taken when testing on data acquired using alternative imaging devices from that used for the development dataset. Our study justifies further validation of plus disease classifiers in ROP screening and supports a potential role for code-free approaches to help prevent blindness in vulnerable neonates. FUNDING: National Institute for Health Research Biomedical Research Centre based at Moorfields Eye Hospital NHS Foundation Trust and the University College London Institute of Ophthalmology. TRANSLATIONS: For the Portuguese and Arabic translations of the abstract see Supplementary Materials section.

BLIND-SIDED BY COSMETIC VEIN SCLEROTHERAPY: A CASE OF OPHTHALMIC ARTERIAL OCCLUSION.

PURPOSE: Cosmetic vein sclerotherapy is increasingly used to treat varicose veins because of its effectiveness and adherence with British Pharmacopoeia specifications. We present the first documented case of ophthalmic artery occlusion resulting in panocular ischemia secondary to intravascular injection of sodium tetradecyl sulfate sclerosant in a young healthy women seeking treatment for prominent facial veins in her forehead. METHODS: The patient presented with unilateral sudden loss of vision. Funduscopy demonstrated a pale retina, cherry-red spot, and sclerosant visualized directly at the macula. She underwent emergency treatment for central retinal artery occlusion followed by fundal photographs, fluorescein angiography, and optical coherence tomography. RESULTS: Despite intervention, the vision remained no perception to light. Magnetic resonance imaging, echocardiography, and Doppler ultrasound were unremarkable. The patient later developed neovascular sequelae requiring laser pan-retinal photocoagulation. CONCLUSION: Widely regarded as safe, and approved by the U.S. Food and Drug Administration, the only published ocular side effects of foam sclerotherapy are transient visual disturbances or temporary scotomas. This case demonstrates irreversible loss of vision as a previously unreported complication. While undoubtedly rare, we believe physicians and surgeons using sclerosant in the orbital adnexa, face, nose, and sinuses should be aware of this sight-threatening complication of injection and counsel potential patients accordingly.

Variations in Treatment Delivery for Patients with Neovascular AMD in the UK: Results from an Ophthalmology Trainee Clinical Research Network Study.

INTRODUCTION: The aim of this study was to determine treatment delivery patterns for patients with neovascular age-related macular degeneration (nAMD) across the UK through an ophthalmology trainee research network delivered observational study. METHODS: Data were collected via an online tool by potential research collaborators identified by the Ophthalmology Trainee Clinical Trial Network (OCTN). Collaborators were asked to comment on periprocedural practices of treatment of nAMD in their eye unit including treatment location and injectors, clinical assessment and routine observation in patients undergoing intravitreal treatment. RESULTS: Data were available from 26 units around the United Kingdom. Survey methodology refinement was approximately 3 months, and the average response time was 4.9 ± 2.4 days. The majority of responders confirmed that treatment was undertaken as a "one-stop" service (n = 15, 58%), delivered in a clean room (n = 23, 88%). In the majority of units, doctors administered injections (n = 24, 92%), but significant treatment was also given by nurse injectors (n = 21, 81%). All collaborators reported that patients underwent visual acuity testing and optical coherence tomography imaging at all visits, but other imaging including fundus fluorescein angiography (FFA) did not take place in all cases (n = 17, 65%) and only at baseline visit. CONCLUSIONS: These results demonstrate the feasibility of conducting ophthalmology trainee led and delivered observational studies. Our results show that FFA is not routinely used in the diagnosis of nAMD in the units sampled; most injections are carried out in a clean room, and ophthalmic nurses delivering injections is a highly prevalent model of care in the UK.

Recognizing and managing retinal vein occlusion.

Retinal vein occlusion is the second most common retinal vascular disease after diabetic retinopathy and represents a significant cause of irreversible sight loss and disability in persons over the age of 50 years (The Branch Vein Occlusion Study Group, 1984).

Retinal arterial hypertrophy: the new LVH?

Prevention of target organ damage represents the El Dorado for clinicians who treat hypertension. Although many of the cardiovascular sequelae of chronic hypertension are due to large artery atherosclerosis, an equal number are due to small artery dysfunction. These microvascular complications include eye disease (retinopathy), kidney failure, diastolic dysfunction of the heart and small vessel brain disease leading to stroke syndromes, dementia and even depression. Examination of the retinal vasculature represents the only way to reliably derive information regarding small arteries responsible for these diverse pathologies. This review aims to summarise the rapidly accruing evidence indicating that easily observable abnormalities of retinal arteries reflect target organ damage elsewhere in the body of hypertensive patients. In tandem, we also present putative mechanisms by which hypertension and diabetes fundamentally change small artery structure and function and how these processes may lead to target organ damage.