AI and the Eye – Integrating deep learning andin silicosimulations to optimise diagnosis and treatment of wet macular degeneration

This protocol describes the A-EYE Study and provides information about procedures for entering participants. Every care was taken in its drafting, but corrections or amendments may be necessary. These will be circulated to investigators in the Study. Problems relating to this Study should be referred, in the first instance, to the Chief Investigator. This study will adhere to the principles outlined in the UK Policy Framework for Health and Social Care Research (v3.2 10 th October 2017). It will be conducted in compliance with the protocol, the UK General Data Protection Regulation and Data Protection Act 2018, and other regulatory requirements as appropriate. DESIGN Single centre non-interventional study of patients with age-related macular degeneration to develop computational models of disease prediction and treatment outcome involving analysis of macular OCTA scans. AIMS Primary Objective To develop a mathematical model (or in silico model) of blood flow and anti-VEGF transport in the retina that, in combination with AI-based analysis of macular OCTA scans and clinical data, can be used to predict treatment response in patients with neovascular age-related macular degeneration (nAMD). Secondary objectives To apply deep learning models in combination with in silico models of blood flow to OCTA analysis, to confirm diagnosis of nAMD and its clinical subtypes. To develop prognostic models to predict treatment outcome based on longitudinal patient follow-up. Using in silico simulations, to understand why certain patients do not respond optimally to anti-VEGF treatment. To define and simulate individualised anti-VEGF treatment for optimal response. OUTCOME MEASURES A validated in silico model of patient response to nAMD and anti-VEGF treatments tailored to individual patients using OCTA scans. Identify optimal intravitreal anti-VEGF therapy drug regime for individual patients using in silico models Improve on the classification and characterisation of neovascular AMD into its subtypes Predict risk factors for poor treatment outcomes such as retinal vascular topology POPULATION ELIGIBILITY All patients aged 55 years or more, with a new diagnosis of nAMD in at least one eye, attending the Macular Clinic at Royal Liverpool University Hospital, who have had a macular OCTA scan at baseline i.e. at the time of diagnosis. DURATION 48 months Clinical Queries Clinical queries should be directed to Dr Savita Madhusudhan who will re-direct the query to the appropriate person if necessary. Sponsor The University of Liverpool is the research Sponsor for this Study. For further information regarding the sponsorship conditions, please contact: Alex Astor Head of Research Support – Health and Life Sciences University of Liverpool Research Support Office 2nd Floor Block D Waterhouse Building 3 Brownlow Street Liverpool L69 3GL [email protected] mailto: [email protected] Funder EPSRC DTP in AI and Future Digital Health is funding the studentship associated with this study. Mr Remi Hernandez is the PhD candidate holding the studentship and Dr El-Bouri, Prof Zheng, and Dr Madhusudhan are his supervisors.