Search

Medical artificial intelligence (AI) offers great potential for recognizing signs of health conditions in retinal images and expediting the diagnosis of eye diseases and systemic disorders 1 . However, the development of AI models requires substantial annotation and models are usually task-specific with limited generalizability to different clinical applications 2 . Here, we present RETFound, a foundation model for retinal images that learns generalizable representations from unlabelled retinal images and provides a basis for label-efficient model adaptation in several applications. Specifically, RETFound is trained on 1.6 million unlabelled retinal images by means of self-supervised learning and then adapted to disease detection tasks with explicit labels. We show that adapted RETFound consistently outperforms several comparison models in the diagnosis and prognosis of sight-threatening eye diseases, as well as incident prediction of complex systemic disorders such as heart failure and myocardial infarction with fewer labelled data. RETFound provides a generalizable solution to improve model performance and alleviate the annotation workload of experts to enable broad clinical AI applications from retinal imaging., (© 2023. The Author(s).)

Background: Intravitreal anti-vascular endothelial growth factor (anti-VEGF) injections play a key role in treating a range of macular diseases. The effectiveness of these therapies is dependent on patients' adherence (the extent to which a patient takes their medicines as per agreed recommendations from the healthcare provider) and persistence (continuation of the treatment for the prescribed duration) to their prescribed treatment regimens. The aim of this systematic review was to demonstrate the need for further investigation into the prevalence of, and factors contributing to, patient-led non-adherence and non-persistence, thus facilitating improved clinical outcomes., Methods: Systematic searches were conducted in Google Scholar, Web of Science, PubMed, MEDLINE, and the Cochrane Library. Studies in English conducted before February 2023 that reported the level of, and/or barriers to, non-adherence or non-persistence to intravitreal anti-VEGF ocular disease therapy were included. Duplicate papers, literature reviews, expert opinion articles, case studies, and case series were excluded following screening by two independent authors., Results: Data from a total of 409,215 patients across 52 studies were analysed. Treatment regimens included pro re nata, monthly and treat-and-extend protocols; study durations ranged from 4 months to 8 years. Of the 52 studies, 22 included a breakdown of reasons for patient non-adherence/non-persistence. Patient-led non-adherence varied between 17.5 and 35.0% depending on the definition used. Overall pooled prevalence of patient-led treatment non-persistence was 30.0% (P = 0.000). Reasons for non-adherence/non-persistence included dissatisfaction with treatment results (29.9%), financial burden (19%), older age/comorbidities (15.5%), difficulty booking appointments (8.5%), travel distance/social isolation (7.9%), lack of time (5.8%), satisfaction with the perceived improvement in their condition (4.4%), fear of injection (4.0%), loss of motivation (4.0%), apathy towards eyesight (2.5%), dissatisfaction with facilities 2.3%, and discomfort/pain (0.3%). Three studies found non-adherence rates between 51.6 and 68.8% during the COVID-19 pandemic, in part due to fear of exposure to COVID-19 and difficulties travelling during lockdown., Discussion: Results suggest high levels of patient-led non-adherence/non-persistence to anti-VEGF therapy, mostly due to dissatisfaction with treatment results, a combination of comorbidities, loss of motivation and the burden of travel. This study provides key information on prevalence and factors contributing to non-adherence/non-persistence in anti-VEGF treatment for macular diseases, aiding identification of at-risk individuals to improve real-world visual outcomes. Improvements in the literature can be achieved by establishing uniform definitions and standard timescales for what constitutes non-adherence/non-persistence., Systematic Review Registration: PROSPERO CRD42020216205., (© 2023. Crown.)

Health data that are publicly available are valuable resources for digital health research. Several public datasets containing ophthalmological imaging have been frequently used in machine learning research; however, the total number of datasets containing ophthalmological health information and their respective content is unclear. This Review aimed to identify all publicly available ophthalmological imaging datasets, detail their accessibility, describe which diseases and populations are represented, and report on the completeness of the associated metadata. With the use of MEDLINE, Google's search engine, and Google Dataset Search, we identified 94 open access datasets containing 507 724 images and 125 videos from 122 364 patients. Most datasets originated from Asia, North America, and Europe. Disease populations were unevenly represented, with glaucoma, diabetic retinopathy, and age-related macular degeneration disproportionately overrepresented in comparison with other eye diseases. The reporting of basic demographic characteristics such as age, sex, and ethnicity was poor, even at the aggregate level. This Review provides greater visibility for ophthalmological datasets that are publicly available as powerful resources for research. Our paper also exposes an increasing divide in the representation of different population and disease groups in health data repositories. The improved reporting of metadata would enable researchers to access the most appropriate datasets for their needs and maximise the potential of such resources., (Copyright © 2021 The Author(s). Published by Elsevier Ltd. This is an Open Access article under the CC BY 4.0 license. Published by Elsevier Ltd.. All rights reserved.)

Among the most noteworthy developments in ophthalmology over the last decade has been the emergence of quantifiable high-resolution imaging modalities, which are typically non-invasive, rapid and widely available. Such imaging is of unquestionable utility in the assessment of ocular disease however evidence is also mounting for its role in identifying ocular biomarkers of systemic disease, which we term oculomics . In this review, we highlight our current understanding of how retinal morphology evolves in two leading causes of global morbidity and mortality, cardiovascular disease and dementia. Population-based analyses have demonstrated the predictive value of retinal microvascular indices, as measured through fundus photography, in screening for heart attack and stroke. Similarly, the association between the structure of the neurosensory retina and prevalent neurodegenerative disease, in particular Alzheimer's disease, is now well-established. Given the growing size and complexity of emerging multimodal datasets, modern artificial intelligence techniques, such as deep learning, may provide the optimal opportunity to further characterize these associations, enhance our understanding of eye-body relationships and secure novel scalable approaches to the risk stratification of chronic complex disorders of ageing., Competing Interests: Disclosure: S.K. Wagner, None; D.J. Fu, None; L. Faes, None; X. Liu, None; J. Huemer, None; H. Khalid, None; D. Ferraz, None; E. Korot, Google Health (E); C. Kelly, None; K. Balaskas, Alimera (F), Allergan (F), Bayer (F), Heidelberg Engineering (F), Novartis (F), TopCon (F); A.K. Denniston, None; P.A. Keane, Heidelberg Engineering (F), Topcon (F), Carl Zeiss Meditec (F), Haag-Streit (F), Allergan (F), Novartis (F, S), Bayer (F, S), DeepMind (C), Optos (C), (Copyright 2020 The Authors.)

The association between socioeconomic position and morbidity and mortality has long been recognized. We evaluate the evidence for an association between multiple aspects of deprivation and ocular health in a global context. This is a systematic review of studies that evaluated deprivation in the adult population in the context of the major acquired causes of visual loss such as cataract, diabetic eye disease, glaucoma, age-related macular degeneration, and ocular trauma. The search strategy identified relevant studies reported between 1946 and August 2016, with randomized control trials, case-control, cohort, and cross-sectional study designs being selected for inclusion. The studies identified in this review from across the world demonstrate the extent to which the common themes such as low educational attainment and low income may be associated with increased incidence of various sight-threatening conditions and may adversely affect access to specialist assessment and delivery of treatment. Health inequality may always persist, but an increased recognition of the importance of the various impacts of deprivation may empower policy makers to target limited resources to the most vulnerable groups in order to deliver the greatest benefit., (Copyright © 2017 Elsevier Inc. All rights reserved.)

Purpose: To evaluate a proposed method for objective measurement of vitreous inflammation using a spectral-domain optical coherence tomography (SD OCT) device in a large cohort of uveitis eyes, including pseudophakic eyes and vitrectomized eyes., Design: Retrospective, observational cohort study., Methods: One hundred five uveitis eyes (105 patients) with different vitreous haze score grades according to standardized protocols and corresponding SD OCT images (Cirrus HD-OCT; Carl Zeiss Meditec, Dublin, California, USA) were included. Clinical data recorded included phakic status, previous vitreoretinal surgery, and anterior chamber (AC) cells and flare. SD OCT images were analyzed using custom software that provided absolute measurements of vitreous (VIT) and retinal pigment epithelium (RPE) signal intensities, which were compared to generate a relative optical density ratio with arbitrary units (VIT/RPE-relative intensity) and compared to VHS., Results: VIT/RPE-relative intensity showed a significant positive correlation with vitreous haze score (r = 0.535, P < .001) that remained significant after adjusting for factors governing media clarity, such as AC cells, AC flare, and phakic status (R(2)-adjusted = 0.424, P < .001). Significant differences were also observed between the different vitreous haze score groups (P < .001). Preliminary observation did not observe differences in VIT/RPE-relative intensity values between phakic and pseudophakic eyes (0.3522 vs 0.3577, P = .48) and between nonvitrectomized and vitrectomized eyes (0.3540 vs 0.3580, P = .52), overall and respectively for each vitreous haze score subgroup., Conclusions: VIT/RPE-relative intensity values provide objective measurements of vitreous inflammation employing an SD OCT device. Phakic status and previous vitrectomy surgery do not appear to influence these values, although these preliminary findings need validation in future studies., (Copyright © 2016 Elsevier Inc. All rights reserved.)

Systemic lupus erythematosus (SLE) is a life-threatening multisystem inflammatory condition that may affect almost any part of the eye. We provide an update for the practicing ophthalmologist comprising a systematic review of the recent literature presented in the context of current knowledge of the pathogenesis, diagnosis, and treatment of this condition. We review recent advances in the understanding of the influence of genetic and environmental factors on the development of SLE. Recent changes in the diagnostic criteria for SLE are considered. We assess the potential for novel molecular biomarkers to find a clinical application in disease diagnosis and stratification and in the development of therapeutic agents. We discuss limited forms of SLE and their differentiation from other collagen vascular disorders and review recent evidence underlying the use of established and novel therapeutics in this condition, including specific implications regarding monitoring for ocular toxicity associated with antimalarials., (Copyright © 2016 Elsevier Inc. All rights reserved.)

Importance: Patients increasingly use the Internet to access information related to their disease, but poor health literacy is known to impact negatively on medical outcomes. Multiple agencies have recommended that patient-oriented literature be written at a fourth- to sixth-grade (9-12 years of age) reading level to assist understanding. The readability of online patient-oriented materials related to ophthalmic diagnoses is not yet known., Objective: To assess the readability of online literature specifically for a range of ophthalmic conditions., Design and Setting: Body text of the top 10 patient-oriented websites for 16 different ophthalmic diagnoses, covering the full range of ophthalmic subspecialties, was analyzed for readability, source (United Kingdom vs non-United Kingdom, not for profit vs commercial), and appropriateness for sight-impaired readers., Main Outcomes and Measures: Four validated readability formulas were used: Flesch Reading Ease Score (FRES), Flesch-Kincaid Grade Level (FKGL), Simple Measure of Gobbledygook (SMOG), and Gunning Fog Index (GFOG). Data were compared with the Mann-Whitney test (for 2 groups) and Kruskal-Wallis test (for more than 2 groups) and correlation was assessed by the Spearman r., Results: None of the 160 webpages had readability scores within published guidelines, with 83% assessed as being of "difficult" readability. Not-for-profit webpages were of significantly greater length than commercial webpages (P = .02) and UK-based webpages had slightly superior readability scores compared with those of non-UK webpages (P = .004 to P < .001, depending on the readability formula used). Of all webpages evaluated, only 34% included facility to adjust text size to assist visually impaired readers., Conclusions and Relevance: To our knowledge, this is the first study to assess readability of patient-focused webpages specifically for a range of ophthalmic diagnoses. In keeping with previous studies in other medical conditions, we determined that readability scores were inferior to those recommended, irrespective of the measure used. Although readability is only one aspect of how well a patient-oriented webpage may be comprehended, we recommend the use of readability scoring when producing such resources in the future. Minimum readability policies and inclusion of facilities within webpages to maximize viewing potential for visually impaired readers are important to ensure that online ophthalmic patient information is accessible to the broadest audience possible.

In this review we consider the current evidence base for treatments in inflammatory eye disease, and in particular uveitis, from a historical perspective. We consider the challenges that have traditionally hindered progress in inflammatory eye disease including small target populations, heterogeneous disease groups, poorly defined phenotypes, diagnostic inconsistency, subjective outcome measures, specific issues around visual acuity as an outcome measure and low commercial interest. Strategies to address these issues are considered de novo and with reference to recent advances outside of ophthalmology and highlight the promise for ocular inflammation. Progress in these specialties has included the development of thriving clinical-trial cultures, public-private partnerships, pathogenetic- and structure-led drug design, efficient drug development pipelines, and biomarker-defined treatment protocols enabling personalization of medicine. Although there are challenges, these are exciting opportunities as we seek to develop safe and effective treatments for patients with inflammatory eye disease.

Turner's syndrome is one of the most common of all chromosomal abnormalities and is associated with significant ophthalmic morbidity. Turner's 1938 account included two patients with strabismus, and hitherto the condition has generated more interest among orthoptists than ophthalmologists. This systematic review of the literature seeks to redress the balance. Based on the pooled data of 274 patients with Turner's syndrome, it is the most complete evaluation so far of the prevalence and severity of ophthalmic problems in this population. This includes both a systematic review of the ophthalmic literature (via Medline) and the much larger body of work available in the orthoptic literature. Finally, we consider recent progress that enables the ophthalmologist to progress from the simple recognition of a phenotype to the correlation of genotypic variations with embryogenesis and consequent features of that phenotype.