Purpose: Application of artificial intelligence (AI) to macular OCT scans to segment and quantify volumetric change in anatomical and pathological features during intravitreal treatment for neovascular age-related macular degeneration (AMD). Design: Retrospective analysis of OCT images from the Moorfields Eye Hospital AMD Database. Participants: A total of 2115 eyes from 1801 patients starting anti-VEGF treatment between June 1, 2012, and June 30, 2017. Methods: The Moorfields Eye Hospital neovascular AMD database was queried for first and second eyes receiving anti-VEGF treatment and had an OCT scan at baseline and 12 months. Follow-up scans were input into the AI system and volumes of OCT variables were studied at different time points and compared with baseline volume groups. Cross-sectional comparisons between time points were conducted using Mann-Whitney U test. Main Outcome Measures: Volume outputs of the following variables were studied: intraretinal fluid, sub- retinal fluid, pigment epithelial detachment (PED), subretinal hyperreflective material (SHRM), hyperreflective foci, neurosensory retina, and retinal pigment epithelium. Results: Mean volumes of analyzed features decreased significantly from baseline to both 4 and 12 months, in both first-treated and second-treated eyes. Pathological features that reflect exudation, including pure fluid components (intraretinal fluid and subretinal fluid) and those with fluid and fibrovascular tissue (PED and SHRM), displayed similar responses to treatment over 12 months. Mean PED and SHRM volumes showed less pronounced but also substantial decreases over the first 2 months, reaching a plateau postloading phase, and minimal change to 12 months. Both neurosensory retina and retinal pigment epithelium volumes showed gradual reductions over time, and were not as substantial as exudative features. Conclusions: We report the results of a quantitative analysis of change in retinal segmented features over time, enabled by an AI segmentation system. Cross-sectional analysis at multiple time points demonstrated significant associations between baseline OCT-derived segmented features and the volume of biomarkers at follow-up. Demonstrating how certain OCT biomarkers progress with treatment and the impact of pretreatment retinal morphology on different structural volumes may provide novel insights into disease mechanisms and aid the personalization of care. Data will be made public for future studies. Financial Disclosure(s): Proprietary or commercial disclosure may be found in the Footnotes and Disclosures at the end of this article. Ophthalmology Science 2024;4:100570 (c) 2024 by the American Academy of Ophthalmology. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/ 4.0/).

Purpose Neovascular age-related macular degeneration (nAMD) is a major global cause of blindness. Whilst anti-vascular endothelial growth factor (anti-VEGF) treatment is effective, response varies considerably between individuals. Thus, patients face substantial uncertainty regarding their future ability to perform daily tasks. In this study, we evaluate the performance of an automated machine learning (AutoML) model which predicts visual acuity (VA) outcomes in patients receiving treatment for nAMD, in comparison to a manually coded model built using the same dataset. Furthermore, we evaluate model performance across ethnic groups and analyse how the models reach their predictions. Methods Binary classification models were trained to predict whether patients' VA would be 'Above' or 'Below' a score of 70 one year after initiating treatment, measured using the Early Treatment Diabetic Retinopathy Study (ETDRS) chart. The AutoML model was built using the Google Cloud Platform, whilst the bespoke model was trained using an XGBoost framework. Models were compared and analysed using the What-if Tool (WIT), a novel model-agnostic interpretability tool. Results Our study included 1631 eyes from patients attending Moorfields Eye Hospital. The AutoML model (area under the curve [AUC], 0.849) achieved a highly similar performance to the XGBoost model (AUC, 0.847). Using the WIT, we found that the models over-predicted negative outcomes in Asian patients and performed worse in those with an ethnic category of Other. Baseline VA, age and ethnicity were the most important determinants of model predictions. Partial dependence plot analysis revealed a sigmoidal relationship between baseline VA and the probability of an outcome of 'Above'. Conclusion We have described and validated an AutoML-WIT pipeline which enables clinicians with minimal coding skills to match the performance of a state-of-the-art algorithm and obtain explainable predictions.

Purpose of review The purpose of this review is to describe the current status of automated deep learning in healthcare and to explore and detail the development of these models using commercially available platforms. We highlight key studies demonstrating the effectiveness of this technique and discuss current challenges and future directions of automated deep learning. Recent findings There are several commercially available automated deep learning platforms. Although specific features differ between platforms, they utilise the common approach of supervised learning. Ophthalmology is an exemplar speciality in the area, with a number of recent proof-of-concept studies exploring classification of retinal fundus photographs, optical coherence tomography images and indocyanine green angiography images. Automated deep learning has also demonstrated impressive results in other specialities such as dermatology, radiology and histopathology. Summary Automated deep learning allows users without coding expertise to develop deep learning algorithms. It is rapidly establishing itself as a valuable tool for those with limited technical experience. Despite residual challenges, it offers considerable potential in the future of patient management, clinical research and medical education. Video abstract http://links.lww.com/COOP/A44

Over the past few months, various preventative measures that aim to combat the spread of severe acute respiratory syndrome coronavirus-2 have been implemented with good results, including the use of personal protective equipment (PPE) and reinforcing general hygiene practices [[1]Vimercati L. Dell'Erba A. Migliore G. De Maria L. Caputi A. Quarato M. et al.Prevention and protection measures of healthcare workers exposed to SARS-CoV-2 in a university hospital in Bari, Apulia, Southern Italy.J Hosp Infect. 2020; 105: 454-458Abstract Full Text Full Text PDF PubMed Scopus (31) Google Scholar]. However, research by Iqbal and Chaudhuri regarding management strategies for dealing with coronavirus disease 2019 (COVID-19) in the UK concluded that current efforts are 'not translating to a sense of security' amongst the National Health Service (NHS) workforce [[2]Iqbal M.R. Chaudhuri A. COVID-19: results of a national survey of United Kingdom healthcare professionals' perceptions of current management strategy – a cross-sectional questionnaire study.Int J Surg. 2020; 79: 156-161Crossref PubMed Scopus (47) Google Scholar]. In particular, mortality statistics have highlighted a disproportionate effect on Black, Asian and Minority Ethnic (BAME) healthcare workers (HCWs). Preliminary analysis of 119 HCWs that have died in the UK with COVID-19 revealed that 64% of them were from the BAME community, despite this community representing only 21% of the workforce [[3]Health Service Journal Exclusive: deaths of NHS staff from COVID-19 analysed.2020https://www.hsj.co.uk/exclusive-deaths-of-nhs-staff-from-covid-19-analysed/7027471Google Scholar]. Furthermore, a survey of frontline doctors conducted by the British Medical Association (BMA) showed that, compared with their White colleagues, almost twice as many BAME doctors felt pressured to work in high-risk environments without adequate PPE [[4]British Medical Association BAME doctors hit worse by lack of PPE. BMA, London2020https://www.bma.org.uk/news-and-opinion/bame-doctors-hit-worse-by-lack-of-ppeGoogle Scholar]. Following guidance from NHS Employers (the employers' organization for the NHS in England) and BMA, hospital NHS trusts have responded to these concerns by implementing risk assessments for HCWs that take ethnicity into account. Outcomes of these assessments are then used to provide individualized and specific guidance to staff members, such as by suggesting modifications to their work practices. However, there are important issues to consider in the design of these risk assessment tools, which can cause considerable concern for BAME HCWs. The latest data from the UK Office of National Statistics (ONS) indicates that being from a BAME background is itself a significant factor that increases the mortality risk of COVID-19 [[5]Office for National Statistics Coronavirus (COVID-19) related deaths by ethnic group, England and Wales. ONS, London2020https://www.ons.gov.uk/peoplepopulationandcommunity/birthsdeathsandmarriages/deaths/articles/coronaviruscovid19relateddeathsbyethnicgroupenglandandwales/2march2020to15may2020Google Scholar]. However, risk assessment systems that require two or more risk factors to be present (one being that the individual is of a BAME background) will not identify BAME HCWs without additional risk factors as being at significantly higher risk. In addition, the grouping of many diverse ethnicities within the umbrella term of 'BAME' severely restricts the accuracy – and, by extension, the validity of subsequent advice – of these risk assessments. This approach fails to appreciate that 'BAME' refers to a heterogeneous group, overlooking the large variations in mortality risk between different ethnicities [[5]Office for National Statistics Coronavirus (COVID-19) related deaths by ethnic group, England and Wales. ONS, London2020https://www.ons.gov.uk/peoplepopulationandcommunity/birthsdeathsandmarriages/deaths/articles/coronaviruscovid19relateddeathsbyethnicgroupenglandandwales/2march2020to15may2020Google Scholar]. The use of this broad term can also lead to instances of confusion, seen in the way that some risk assessments clearly include mixed race under the BAME category whilst others do not specify. To avoid the consequences of using reductionist labels, we propose that risk assessments should reflect the six ethnic categories as found in the detailed reports published by ONS regarding mortality rates for COVID-19: Black, Bangladeshi/Pakistani, Indian, Chinese, Mixed and Other [[5]Office for National Statistics Coronavirus (COVID-19) related deaths by ethnic group, England and Wales. ONS, London2020https://www.ons.gov.uk/peoplepopulationandcommunity/birthsdeathsandmarriages/deaths/articles/coronaviruscovid19relateddeathsbyethnicgroupenglandandwales/2march2020to15may2020Google Scholar]. Moreover, the lack of standardization between the risk assessments issued by each of over 200 NHS trusts in the UK can result in significant variations in risk stratification between hospitals. This is not a trivial matter, as variations in risk categories will subsequently impact advice around work; an HCW at one hospital may be advised to change to a non-patient-facing role, whereas in a different hospital, they may be advised to continue to work as normal. For some HCWs, these variations in management could be the difference between life and death. Whilst it may sometimes be appropriate for individual NHS trusts to take a different approach to risk assessment, for example based on their regional circumstances, the so-far inconsistent approach has left many staff feeling worried and unsure about what they need to do to best protect themselves and their patients. These feelings have been compounded by recent reports that, to date, only 23% of NHS trusts have completed the process of risk-assessing their staff [[6]Sky News Coronavirus: NHS England apologises after investigation finds only 23% of health trusts have risk-assessed BAME staff.2020https://news.sky.com/story/coronavirus-nhs-england-apologises-after-investigation-finds-only-23-of-health-trusts-have-risk-assessed-bame-staff-12015363Google Scholar]. In conclusion, these issues highlight the need for standardized and widely implemented risk assessments that use the best-available evidence to assess the risk to BAME HCWs more accurately. Finally, although this letter has focused on HCWs, we stress that similar approaches must be taken for public-facing jobs across any other relevant industries. Moreover, steps to mitigate the short-term risk to BAME staff must continue, in parallel with ongoing investigations into contributing socio-economic or biological factors leading to the racial disparity in mortality rates of COVID-19. None declared. None.