2025 Lay Abstracts
Deep Phenotyping of Optical Coherence Tomography Biomarkers of Cardiovascular Disease Using Deep Learning and Genome Wide Analysis
Josef Huemer
Moorfields Eye Hospital NHS Foundation Trust
Cardiovascular diseases are a leading cause of death globally, creating an urgent need for better ways to identify high-risk individuals early. Emerging research suggests that tiny, hidden signs of blood flow problems around the retina’s blood vessels, visible with Optical coherence tomography (OCT) scans, could be an early warning sign for systemic cardiovascular disease. However, the exact link and the general prevalence of these signs are not well understood.
This project will investigate these signs of subclinical ischemia by developing a powerful artificial intelligence (AI) program. We will use a state-of-the-art ophthalmic foundation model trained on 1.6 million images and fine-tune it to automatically and accurately detect and measure these subtle changes in eye scans.
To do this, we will utilize the vast UK Biobank, a major health study containing OCT scans, fundus photographs and in-depth health data for approximately 82,000 individuals. By applying our AI tool to this dataset, we will determine how common these eye changes are and analyze their connection to participants’ broader health, including cardiovascular conditions, lifestyle factors, and even genetic data. This large-scale analysis will allow us to assess whether people with these signs in their eyes are more likely to have or develop heart disease.
To ensure our AI tool is robust and reliable, it will be validated using data from other large studies, including the Hamburg City Health Study. The ultimate goal is to determine if adding this information from a simple, non-invasive eye scan can improve our ability to predict a person’s risk of a future heart attack or stroke, enabling earlier preventative measures and saving lives.
Assessing Physiological Risk Factors for Retinopathy of Prematurity (ROP) Progression and Effects of Anti-VEGF Therapy on Neurodevelopment: PROGRESS-ROP Study
Kanmin Xue
University of Oxford
Retinopathy of prematurity (ROP) is a leading cause of childhood blindness, affecting babies born very early or very small. Advances in retinal imaging and the introduction of anti-VEGF injections have transformed how ROP is detected and treated, greatly improving outcomes for many infants. However, important clinical questions remain. First, can we identify practical ways to prevent ROP from worsening, so that fewer babies need treatment? Second, while anti-VEGF therapy is highly effective in controlling abnormal blood vessel growth in the eye, its longer-term effects on the developing retina and brain are not yet fully understood. The EURETINA Retinal Medicine Clinical Research Grant 2025 will support a two-year research project led by A/Prof Kanmin Xue and A/Prof Caroline Hartley at the University of Oxford. The project will investigate factors that increase the risk of ROP progression and examine how anti-VEGF treatment influences retinal and brain development. By combining advanced imaging with detailed assessments of visual and neurological function, the team will build a clearer picture of how ROP and its treatment shape early visual and brain maturation. This knowledge will help guide safer, more targeted approaches to ROP care and improve long-term visual outcomes in childhood.
2024 Lay Abstracts
Enhanced Vision and Imaging Test for Enabling Treatment Trials in Early and Intermediate AMD
Maximillian Pfau
University of Bonn
Age-related macular degeneration (AMD) is a leading cause of vision loss in older adults. Detecting AMD at its earliest stages—before irreversible retinal damage occurs—is crucial for developing effective prevention and treatment strategies. This project aims to evaluate a new approach for measuring how the eye adapts to darkness (dark adaptation) as an early indicator of AMD. The study focuses on refining fundus-tracked dark adaptometry, a technique that measures how quickly rods and cones recover after bright light exposure while precisely tracking the tested retinal area.
Work Package 1 will validate this new method in healthy volunteers and in patients with early or intermediate AMD. We will assess whether it can reliably measure dark adaptation, separate normal age-related changes from early disease, and more accurately classify AMD stages compared to existing functional tests. Participants will undergo comprehensive eye examinations and dark adaptation testing. Statistical analyses will establish normal reference data, assess test–retest reliability, and determine the method’s ability to differentiate AMD severity.
Work Package 2 will explore how structural changes detected with high-resolution optical coherence tomography (OCT) relate to visual function. Using advanced image analysis and precise alignment techniques, we will quantify subtle alterations in the outer retinal layers and compare these findings with dark adaptation performance. This structure–function correlation will help identify imaging markers that signal early functional impairment in AMD.
Work Package 3 will use artificial intelligence to connect standard clinical OCT data with dark adaptation results. The aim is to develop a predictive model that estimates rod-mediated visual function directly from OCT scans, providing a rapid and non-invasive way to detect early retinal dysfunction.
Together, these studies will improve early AMD detection and deepen our understanding of how retinal structure and function interact during disease progression.
Prediction of Treatment Needs Using a Hybrid Artificial Intelligence-Based Management Assistant (ALMA) Tool in Neovascular Age-Related Macular Degeneration
Javier Zarranz-Ventura
Fundació De Recerca Clínic Barcelona – Institut D’investigacions Biomèdiques August Pi I Sunyer (FRCB-IDIBAPS)
The ALMA project is directed to create and validate a hybrid Artificial inteLligence-based Management Assistant (ALMA) tool to identify baseline patient profiles in Neovascular Age-related Macular Degeneration (nAMD) for optimizing intravitreal therapy outcomes.
The ALMA tool will be created and trained in the multicenter, national nAMD registry Fight Retinal Blindness Spain (FRB Spain, Zarranz-Ventura et al. CEO 2022) database and associated OCT images datasets from previous studies (FRB Spain IMAGE, Martin-Pinardel et al. BJO 2024) to identify eyes achieving a series of endpoints: a) long treatment intervals (defined as 12, 14 and 16-week maximum interval between injections), b) high frequency of injections (defined as 8 injections), and c) visual outcomes (defined as <5, 5, 10 and 15 letters gain/loss from baseline) at 12 months. The variables included in the model will be: a) Demographics (age, gender, ethnicity, smoking status), b) ocular (baseline visual acuity -VA-, neovascular lesion type, previous ocular surgery, comorbidities, and treatment type) and c) OCT-based (central retinal thickness, macular volume and fluid compartment quantifications -intraretinal, subretinal and sub-retinal pigment epithelium-). The tool performance will be validated in a 12-month prospective multicentric cohort.
The clinical significance of the ALMA project is based on the identification of different patient profiles which may predict the evolution of cases at individual level, with a view to inform clinicians in the selection of the most appropriate therapy to optimize the clinical outcomes of anti-VEGF drugs as a step forward to a personalized medicine. The economic implications of this optimized treatment choice are relevant, as this tailored approach may reduce the rates of non-responders and switchers between treatments, with not only clinical benefits but also potential savings that could mitigate the substantive cost associated to intravitreal therapies.
2023 Lay Abstracts
Predicting Progressive Macular Atrophy in Patients Undergoing Anti-VEGF Therapy for Neovascular Age-Related Macular Degeneration Using Multi-Modal Imaging, Genetics, and Artificial Intelligence
Pearse Keane
University College London
Neovascular (“wet”) age-related macular degeneration (nAMD) is the leading cause of blindness among older people in European countries. Typical this involves the growth of abnormal blood vessels in the retina, the nerve tissue which lines the back of the eye. These blood vessels can lead to bleeding, fluid accumulation, scarring, and ultimately blindness. Current treatment for nAMD involves injections into the eye which block the growth of these blood vessels, helping to improve vision. However, some of the patients receiving this treatment will still lose vision due to the development of progressive wearing away (atrophy) of the retina. We still don’t know which patients will develop this atrophy, making it impossible to implement patient-specific treatment strategies.
To identify patients at risk, we will use state-of-the-art artificial intelligence (AI) tools on eye images at Moorfields Eye Hospital to outline and predict atrophic changes in patients with nAMD. First, we will describe the natural course of atrophic changes in eyes receiving treatment for nAMD. This will help to understand how the condition develops and progresses. Next, we will use AI to predict progression of atrophy over different time periods. These analyses will use both population-based and clinical data alongside images, and give us more detailed information about patients that might develop atrophy. Finally, since genetics have a major impact on AMD development and progression, we will incorporate genetic data into our prediction model.
We aim to provide a model that is capable of reliably identifying patients that are at risk for atrophy development while receiving treatment for nAMD. With the recent emergence of drugs targeting atrophic changes in AMD, the results of our study may help to prevent irreversible visual loss due to atrophic changes by providing patients with combination treatments, targeting both the aberrant vessels and the atrophic lesions.
“EYS on” Gene Editing for Retinitis Pigmentosa 25
João Pedro Marques
University of Coimbra
The project “EYS on” gene editing for Retinitis Pigmentosa 25 aims to develop a new way to treat a type of inherited vision loss by fixing a particular genetic mistake in the retina. This genetic mistake is caused by a mutation in a gene called EYS. When this gene does not work properly, it can lead to a type of vision loss that gets worse over time and has no cure. We are exploring a technique called prime editing to fix a particular genetic defect, quite common in our patient cohort, in Portugal. This technique is like molecular scissors that can cut and paste the DNA in a very precise way. It is different from other gene editing techniques because it is less likely to make unintended changes to the DNA. To test whether prime editing can fix this particular genetic mistake in the EYS gene, we will use cells from patients with inherited vision loss. In the laboratory, these cells will be turned into tridimensional (3D)-mini-retinas which we expect will recapitulate the changes observed in the patient’s retinas. We will then use prime editing technology to correct the genetic mistake in these 3D-mini-retinas/cells and see if it works. We will study whether the corrected cells can show better morphology, survival or function than the cells with the genetic mistake. If the corrected cells work or survive better, this could be a promising new treatment for inherited vision loss caused by a mutation in the EYS gene. This research is important because there are currently no treatments for inherited vision loss caused by a mutation in the EYS gene. If we will be able to develop a new treatment using prime editing, it could help many people with this condition to see better and improve their quality of life.
The Role of Mitochondria in Geographic Atrophy Progression in Age-Related Macular Degeneration
Enrico Borrelli
University of Turin
Geographic atrophy (GA), the advanced, non-neovascular stage of age-related macular degeneration (AMD), is a major cause of vision loss, resulting from the degeneration of the outer retina, retinal pigment epithelium (RPE), and choriocapillaris. Oxidative stress and mitochondrial DNA (mtDNA) damage, both associated with aging, play a critical role in the development of GA. Mitochondria, essential for energy production in RPE cells, become dysfunctional due to an increase in reactive oxygen species (ROS) and compromised mtDNA repair, leading to cellular degeneration. In AMD, extensive mtDNA damage correlates with the severity of the disease, impairing mitochondrial function and ATP production. Under normal circumstances, dysfunctional mitochondria are cleared via mitophagy; however, in AMD, this process is disrupted, promoting disease progression.
Key genes involved in mitochondrial protection, such as NFE2L2 and PGC-1α, regulate oxidative stress and autophagy. In AMD, the reduced expression of these genes leads to heightened oxidative damage. Genetic research indicates that specific mitochondrial haplogroups affect AMD risk, with haplogroup J increasing susceptibility and haplogroup H offering some protection.
This study will examine the role of mitochondrial dysfunction in GA by enrolling 600 Caucasian participants (200 controls, 200 with early/intermediate AMD, and 200 with GA). Participants will undergo comprehensive eye exams, including fundus imaging, optical coherence tomography (OCT), and genetic analysis using next-generation sequencing. The study will focus on genetic risk factors, including variants of NFE2L2 and PGC-1α, and will develop genetic risk scores (GRS) to assess their impact on the development and progression of GA. The findings may assist in identifying high-risk patients, enabling precision medicine approaches for diagnosis and treatment. By combining genetic analysis with imaging, this research aims to enhance risk stratification and facilitate the development of targeted therapies for GA.
2022 Lay Abstracts
Risk of Posterior Staphyloma in Highly Myopic Europeans: From Epidemiology to Anatomy and Back
Suzanne Yzer
Radboud University Medical Center
Background: Myopia (near-sightedness) is an endemic problem and is estimated to increase dramatically. Approximately 2.5 billion people worldwide are currently myopic. 20% of these have high myopia (refraction more than -6.00 Dioptres). This is frequently accompanied by complications, putting these patients at high risk of legal blindness. An important driver of these complications is Posterior Staphyloma (PS). PS is an outpouching of the posterior wall of the eye that results in progressive and irreversible damage of different retinal layers and eventually central visual loss. To date, there is no therapeutic intervention to constraint PS or its blinding complications. Virtually all studies on PS are from Asian populations and well documented large studies on Europeans are unavailable.
Aim: To estimate the frequency of PS in European high myopic patients and to study its functional consequences through a unique European collaboration and the constitution of a well phenotyped clinical database and biological biobank.
Approach: In a cross-sectional clinical study we will use multiple devices for imaging of the posterior part of the eye in order to grade PS in high myopic Europeans from hospitals in The Netherlands, Spain and France according to a standard grading system. DNA and additional body materials (like sclera) from high myopes with or without PS will be collected. Well characterized large cohorts will assist in the interpretation of laboratory studies, like future DNA studies and aid in unravelling the underlying mechanisms of PS development. We will perform our research in a unique European consortium to provide distinctive information on PS patients from European descent.
Outcome: Creation of a European clinical and biologic database for future studies on PS, which is an increasing major visual health issue, not only in Asians but also in the European population.
Diet, Lifestyle, Systemic Medication and Genetics: Can the Risk for AMD be Modulated? (AMD_LifeGene)
Cláudia Farinha
AIBILI – Association for Innovation and Biomedical Research on Light and Image
“The knowledge of how different risk factors, genetic and non-genetic, interconnect in AMD pathophysiology influencing its development and progression is of utmost importance.
Our group based in Coimbra performed the first epidemiologic AMD study in Portugal, the Coimbra Eye Study. Two cohorts from two primary Portuguese healthcare units (Mira and Lousã) were phenotypically characterized and risk factors were assessed. The overall prevalences of early and late AMD were similar to other large-scale population-based cohorts. However, AMD prevalence in Mira (a rural coastal town) was inferior to that in Lousã (an inland, more urbanized town). We arrived at the conclusion that the interplay between lifestyle and genetic background differences could be the cause of our reported epidemiologic findings.
With this new project, we aim to explore the interactions between AMD risk factors and their impact on AMD development and progression, in these two Portuguese populations with different reported lifestyles and AMD prevalence. Our primary objective will be to explore how the interplay between genetic, environmental, and lifestyle factors (e.g. Mediterranean diet), and systemic comorbidities/ chronic medications modulates the risk for AMD onset and progression in the global population of the Coimbra Eye Study and in each cohort (Mira vs Lousã). We will also characterize the genetics of our population, including rare variants analysis, explore genotypic-phenotypic associations and determine the 10-year incidence of AMD.
2021 Lay Abstracts
Microbiome, Metabolome and Complement Activation in Age-Related Macular Degeneration
Prof. Robert P Finger
University Hospital Mannheim
Prof Zeinab Abdullah
University Hospital Bonn
Age-related macular degeneration (AMD) is a very common retinal disease associated with ageing. Its early stages affect about one third of persons 70 or older and those are at risk of progression to late AMD stages which often cause severe irreversible vision loss. Consequently, late AMD is the main cause of severe vision impairment in Europe. Today, we still do not fully understand this disease and who might be at risk of progression. We also do not have any treatments to prevent AMD, stop or delay its progression. We know that AMD is associated with inflammation and a dysregulation of the immune system including elevated levels of inflammatory markers in the blood and a reduced capacity for waste removal called phagocytosis. Recently, the bacteria living in our guts (i.e. the microbiome) have been related to inflammation and to AMD. The microbiome impacts how our guts process food and thus the metabolites we resorb as well as levels of inflammation in our blood and elsewhere in our bodies. However, this has not been assessed all in one study yet. In order to understand AMD better we will comprehensively assess AMD status, the microbiome, the metabolome (i.e. metabolic products) and levels of inflammation as part of the immune system. All this together may help us to better understand AMD and unravel novel biomarkers that are linked to increased or decreased risk of disease development.
REGENT (RETiNal GENE ThERapy immunNE REspoNsE)
Dominik Fischer
Oxford University Hospitals NHS Foundation Trust
Viral infections can damage the eye. Antiviral immune responses can lead to additional problems for the eye. The ocular immune privilege is a clever way of nature to limit such secondary damage. However, this privilege is potentially compromised by disease processes and surgical trauma. Surgical delivery of adeno-associated virus (AAV) potentially induces gene therapy (GT) associated inflammation after ocular delivery in patients with pre-existing retinal disease. Progress has been made addressing innate immune response to AAV, but not regarding adaptive immunity: Anti-AAV antibody titres do not correlate or predict the risk of inflammation or its severity; but adaptive immune cells have the potential to limit therapeutic effects of GT (50% therapeutic effect is lost within first year via immune-mediated clearance of treated cells in liver). Immune-mediated clearance of treated cells in the eye could explain an observed decline in functional benefit following ocular GT. It is therefore crucial to better understand the risks, consequences, and best management of GT related inflammation. In this project, we utilize a unique biobank of samples taken from patients and model systems exposed to therapeutic AAV. Hypothesis driven analysis of these samples will help to 1) test, whether subretinal gene therapy leads to increase in ocular and systemic inflammation markers beyond the expected postsurgical recovery phase; 2) identify biomarkers of cellular adaptive immunity that can predict risk and severity of GT associated inflammation; and 3) identify biomarkers of adaptive immunity that can be used to quantify and monitor such an inflammatory response. Closing these knowledge gaps will provide a unique opportunity to identifying strategies of targeted intervention limiting potentially detrimental effects of an immune response against therapeutic virus used for gene therapy.
2020 Lay Abstracts
Metabolomics: A Novel Tool for Investigating the Pathogenesis of Age-related Macular Degeneration (AMDMetab)
Rufino Martins da Silva
AIBILI – Association for Innovation and Biomedical Research on Light and Image
Age-related macular degeneration (AMD) is a disease that affects the macula, the central area of the eye that is responsible for vision. Some people with AMD remain very stable over time, while other progress to the blinding stages of the disease. Currently, there is very little to offer to patients with AMD to prevent them from progressing to the late blinding forms of the disease. This is because we don’t completely understand why progression happens. We know that both genetic/ inherited factors and exposures (such as smoking) are related to progression, but we don’t have enough information about the exact mechanisms that lead to progression in some patients. This study aimed to gain insights into the mechanisms behind AMD progression. To do so, we recruited patients with AMD and a control group and assessed changes in their blood profiles at 5 years. To assess their blood profiles, we used a technique named metabolomics, which is the study of the small little particles in our blood that result from gene expression but are also affected by environmental exposures. Our group had previously shown that this technique (metabolomics) was good to assess AMD, but until now we had not followed people over 5 years. Our results in this project showed that indeed certain metabolites at baseline were associated with a higher risk of developing worse AMD, and that changes in the levels of certain metabolites (i.e., comparing their baseline levels with 5-year levels) were related to progression of the disease. This is important because it gives us information on the exact mechanisms of AMD progression and with further research can help develop treatments to stop the progression of the disease.
The C-Type Lectin Receptor CD93 as a New Target in Retinal Vascular Diseases
Gian Marco Tosi
University of Siena
Anti-VEGF drugs have revolutionized the treatment of neovascular eye disease and provide significant visual benefits. Although this strategy has proven successful in some patients, the efficacy of neovascular inhibitor therapy remains to be improved. The discovery of new targets with complementary endothelial function may provide new opportunities to circumvent resistance mechanisms and improve current therapeutic treatment. The goal of this project was to characterize a new endothelial target and develop new tools for the treatment of retinal vascular diseases. Thank you to the Euretina award, we gained relevant insights into the molecular mechanisms controlling angiogenesis in neovascular retinopathies, which has important biological and translational implications. Indeed, we have developed a new monoclonal antibody that can neutralize the novel endothelial receptor CD93 and block blood vessel growth. Through genetic engineering, we have derived a recombinant antibody that has the same neutralizing effect as the original antibody, but can be produced and purified on a large scale for various applications. For example, we have produced recombinant antibodies conjugated with toxic agents and used them to kill proliferating but not quiescent endothelial cells. In addition, we are testing the recombinant antibody in a specially generated animal model of neovascular retinopathy to provide evidence of its therapeutic potential. Our team is confident that further development of this new agent can benefit patients with neovascular eye diseases.
2019 Lay Abstracts
Juvenile X-Linked Retinoschisis: Assessing the Clinical and Genetic Spectrum in Preparation for Therapy
Camiel Boon
Amsterdam University Medical Centre
X-linked retinoschisis (XLRS) is an intriguing and relatively common genetic eye disease caused by mutations in the RS1 gene. Mutations in this gene lead to splitting between different retinal layers in affected men, which often causes severe loss of vision starting at young age. Patients with XLRS also have an increased risk for vision-threatening complications such as detachment of the retina, that regularly require surgical intervention. Currently, there is no treatment available for XLRS, but gene therapy in mice has shown promising results and first steps have been taken to study gene therapy in humans. To date, strikingly little is known about the natural disease course, clinical variability, frequency of complications, outcome of surgery in XLRS.
We will perform the clinical largest study to date on XLRS, with the following aims:
- To achieve a detailed description of the natural disease course and clinical variability.
- To assess how often complications occur and what the outcome of (surgical) intervention is.
- To determine which patients may be the best candidates for future (gene) therapy and which disease characteristics are most suitable to monitor the effectiveness of gene therapy.
Our study will be coordinated at the Expertise Centre for Hereditary Eye Diseases of the Amsterdam University Medical Centers. Participants will be collected from our own large database, in collaboration with all other Dutch expertise centers from the RD5000 consortium, as well as the Ghent University Hospital (Belgium). We expect to include at least 300 participants, with confirmed genetic testing at our certified DNA diagnostics laboratory. The data collection will include detailed genetic and clinical data with elaborate multimodal imaging and ophthalmic examinations.
This study is crucial to provide an accurate prognosis of disease progression, risk of complications, and outcome of surgery for XLRS patients. Furthermore, our results will be an important milestone for the successful patient selection for future (gene) therapeutic trials through the identification of appropriate clinical endpoints.
Deep Learning, Large Datasets, Genome-Wide Analysis and Histological Correlates to Determine the Role of Reticular Pseudodrusen in the Progression of Age-Related Macular Degeneration
Roy Schwartz
Moorfields Eye Hospital
In this study, we aimed to find new genetic links to a specific eye finding known as reticular pseudodrusen (RPD) by analysing a large health database from the UK Biobank. RPD, along with drusen, are small deposits that can form in the retina and are of interest because they are connected to eye diseases that can affect vision.
To do this, we used artificial intelligence techniques (referred to as a machine learning pipeline) to sift through eye health data from thousands of people. We were looking to identify individuals with RPD, those with another type of deposit called drusen, and a group with neither, to compare them. The work was verified by eye specialists to make sure it was accurate.
We then conducted a series of genetic studies, known as genome-wide association studies, to pinpoint the genetic differences between these groups. We were specifically interested in finding any new genetic factors linked to RPD by comparing it to drusen and the control group.
We successfully identified genetic factors that seem to be associated with RPD but not with drusen, some of which are completely new discoveries. These findings are being further validated to confirm their roles.
In summary, this study has highlighted different genetic risk factors related to RPD but not to drusen, including potential new genes related to RPD, which could help in understanding more about why this condition occurs. We are continuing our work to verify these results, and plan to share our final findings later in the year in a peer-reviewed journal.