Deep Learning-Based Clustering of OCT Images for Biomarker Discovery in Age-Related Macular Degeneration (PINNACLE Study Report 4).

Purpose: We introduce a deep learning-based biomarker proposal system for the purpose of accelerating biomarker discovery in age-related macular degeneration (AMD). Design: Retrospective analysis of a large data set of retinal OCT images. Participants: A total of 3456 adults aged between 51 and 102 years whose OCT images were collected under the PINNACLE project. Methods: Our system proposes candidates for novel AMD imaging biomarkers in OCT. It works by first training a neural network using self-supervised contrastive learning to discover, without any clinical annotations, features relating to both known and unknown AMD biomarkers present in 46 496 retinal OCT images. To interpret the learned biomarkers, we partition the images into 30 subsets, termed clusters, that contain similar features. We conduct 2 parallel 1.5-hour semistructured interviews with 2 independent teams of retinal specialists to assign descriptions in clinical language to each cluster. Descriptions of clusters achieving consensus can potentially inform new biomarker candidates. Main Outcome Measures: We checked if each cluster showed clear features comprehensible to retinal specialists, if they related to AMD, and how many described established biomarkers used in grading systems as opposed to recently proposed or potentially new biomarkers. We also compared their prognostic value for late-stage wet and dry AMD against an established clinical grading system and a demographic baseline model. Results: Overall, both teams independently identified clearly distinct characteristics in 27 of 30 clusters, of which 23 were related to AMD. Seven were recognized as known biomarkers used in established grading systems, and 16 depicted biomarker combinations or subtypes that are either not yet used in grading systems, were only recently proposed, or were unknown. Clusters separated incomplete from complete retinal atrophy, intraretinal from subretinal fluid, and thick from thin choroids, and, in simulation, outperformed clinically used grading systems in prognostic value. Conclusions: Using self-supervised deep learning, we were able to automatically propose AMD biomarkers going beyond the set used in clinically established grading systems. Without any clinical annotations, contrastive learning discovered subtle differences between fine-grained biomarkers. Ultimately, we envision that equipping clinicians with discovery-oriented deep learning tools can accelerate the discovery of novel prognostic biomarkers. Financial Disclosures: Proprietary or commercial disclosure may be found in the Footnotes and Disclosures at the end of this article.

Pachychoroid disease: review and update.

The pachychoroid disease spectrum is a phenotype characterized by alterations in choroidal vasculature which result in outer retinal and choriocapillaris damage and visual loss. The presence of pachyvessels is one of the key features of the pachychoroid phenotype. Recent imaging studies suggest that pachyvessels may form because of choroidal venous congestion in one or more quadrants. The formation of intervortex anastomosis may function as a compensatory mechanism to dissipate the increased venous pressure, while outflow obstruction has been hypothesized to occur at the site of the vortex vein exiting the sclera. This review aims to summarize recent imaging findings and discuss evolution in the understanding of pathogenesis of the pachychoroid disease spectrum. We have summarized notable treatment trials in central serous chorioretinopathy and polypoidal choroidal vasculopathy and included an update of the current diagnostic and management strategies of the entities that are part of the pachychoroid disease spectrum.

Assessing the appropriateness and completeness of ChatGPT-4's AI-generated responses for queries related to diabetic retinopathy.

OBJECTIVE: To evaluate the appropriateness of responses generated by an online chat-based artificial intelligence (AI) model for diabetic retinopathy (DR) related questions. DESIGN: Cross-sectional study. METHODS: A set of 20 questions framed from the patient's perspective addressing DR-related queries, such as the definition of disease, symptoms, prevention methods, treatment options, diagnostic methods, visual impact, and complications, were formulated for input into ChatGPT-4. Peer-reviewed, literature-based answers were collected from popular search engines for the selected questions and three retinal experts reviewed the responses. An inter-human agreement was analyzed for consensus expert responses and also between experts. The answers generated by the AI model were compared with those provided by the experts. The experts rated the response generated by ChatGPT-4 on a scale of 0-5 for appropriateness and completeness. RESULTS: The answers provided by ChatGPT-4 were appropriate and complete for most of the DR-related questions. The response to questions on the adverse effects of laser photocoagulation therapy and compliance to treatment was not perfectly complete. The average rating given by the three retina expert evaluators was 4.84 for appropriateness and 4.38 for completeness of answers provided by the AI model. This corresponds to an overall 96.8% agreement among the experts for appropriateness and 87.6% for completeness regarding AI-generated answers. CONCLUSION: ChatGPT-4 exhibits a high level of accuracy in generating appropriate responses for a range of questions in DR. However, there is a need to improvise the model to generate complete answers for certain DR-related topics.

Health transition and eye care policy planning for people with diabetic retinopathy in south Asia.

The prevalence of type 2 diabetes (T2D), associated systemic disorders, diabetic retinopathy (DR) and current health policies in south Asian countries were analysed to assess country-specific preparedness to meet the 2030 Sustainable Development Goals. The south Asian countries were classified by human development index, socio-demographic index, multidimensional poverty indices, and eye health resources for epidemiological resource-level analysis. In south Asia, the prevalence of diagnosed and undiagnosed T2D in adults aged 40 years or above, was higher in Pakistan (26.3%) and Afghanistan (71.4%), respectively; India has the highest absolute number of people with DR, and Afghanistan has the highest prevalence of DR (50.6%). In this region, out-of-pocket spending is high (∼77%). This Health Policy is a situational analysis of data available on the prevalence of DR and common eye diseases in people with T2D in south Asia and available resources to suggest tailored health policies to local needs. The common issues in the region are insufficient human resources for eye health, unequal distribution of available workforce, and inadequate infrastructure. Addressing these challenges of individuals with T2D and DR, a 10-point strategy is suggested to improve infrastructure, augment human resources, reduce out-of-pocket spending, employ targeted screening, and encourage public-private partnerships.

Exploring the fragility of meta-analyses in ophthalmology: a systematic review.

OBJECTIVE: The fragility index (FI) of a meta-analysis evaluates the extent that the statistical significance can be changed by modifying the event status of individuals from included trials. Understanding the FI improves the interpretation of the results of meta-analyses and can help to inform changes to clinical practice. This review determined the fragility of ophthalmology-related meta-analyses. METHODS: Meta-analyses of randomized controlled trials with binary outcomes published in a journal classified as 'Ophthalmology' according to the Journal Citation Report or an Ophthalmology-related Cochrane Review were included. An iterative process determined the FI of each meta-analysis. Multivariable linear regression modeling evaluated the relationship between the FI and potential predictive factors in statistically significant and non-significant meta-analyses. RESULTS: 175 meta-analyses were included. The median FI was 6 (Q1-Q3: 3-12). This meant that moving 6 outcomes from one group to another would reverse the study's findings. The FI was 1 for 18 (10.2%) of the included meta-analyses and was ≤5 for 75 (42.4%) of the included meta-analyses. The number of events (p < 0.001) and the p-value (p < 0.001) were the best predictors of the FI in both significant and non-significant meta-analyses. CONCLUSION: The statistical significance of meta-analyses in ophthalmology often hinges on the outcome of a few patients. The number of events and the p-value are the most important factors in determining the fragility of the evidence. The FI is an easily interpretable measure that can supplement the reader's understanding of the strength of the evidence being presented. PROSPERO REGISTRATION: CRD42022377589.

Acceptability of Home Monitoring for Neovascular Age-Related Macular Degeneration Reactivation: A Qualitative Study.

This study formed part of a diagnostic test accuracy study to quantify the ability of three index home monitoring (HM) tests (one paper-based and two digital tests) to identify reactivation in Neovascular age-related macular degeneration (nAMD). The aim of the study was to investigate views about acceptability and explore adherence to weekly HM. Semi-structured interviews were held with 98 patients, family members, and healthcare professionals. A thematic approach was used which was informed by theories of technology acceptance. Various factors influenced acceptability including a patient's understanding about the purpose of monitoring. Training and ongoing support were regarded as essential for overcoming unfamiliarity with digital technology. Findings have implications for implementation of digital HM in the care of older people with nAMD and other long-term conditions.

The effect of four loading intravitreal aflibercept injections on macular fluid in treatment-naïve neovascular age-related macular degeneration.

PURPOSE: To evaluate the effect of four versus three loading aflibercept injections on macular fluid resolution and visual acuity (VA) in exudative neovascular AMD (nAMD). METHODS: Multicentre, retrospective cohort study of treatment naïve nAMD eyes undergoing 3 versus 4 loading doses of aflibercept. Change in VA and fluid resolution on optical coherence tomography (OCT), were evaluated at 8 weeks post loading. The primary outcome was proportion of patients with no intraretinal (IRF) and/or subretinal (SRF) fluid at central 1 mm and whole macula at 8 weeks after loading. Data were summarised with mean ± SD for continuous variables, and n (%) for categorical variables. RESULTS: Data from 995 patients was analysed (355 patients - 4 loading doses and 640-3 loading doses). At 8 weeks post 4 loading doses proportion of eyes with neither IRF nor SRF, no IRF and no SRF were 62.8%, 88.7% and 79.2% at fovea versus 56.1%, 87.9% and 69.9% in the whole macula, respectively. Fluid resolution at both fovea and macula were significantly higher in eyes with 4 loading injections versus 3 (p = 0.0001). The mean VA change was +4.0 (±11.3) and +5.4(±13.3) letters for 3 and 4 loading doses groups (p = 0.09). CONCLUSION: Four loading dose injections of aflibercept results in higher proportion of eyes with total fluid resolution in the central subfield and total macular scan when compared to those receiving 3 loading dose injections at 8 weeks post loading phase. However, the better drying effect of 4th loading dose does not translate into better short-term VA outcomes.

A new population for primary prevention of retinal diseases; a step toward improved global eye health beyond diabetes and prediabetes.

Created with biorender.com.

Home-monitoring for neovascular age-related macular degeneration in older adults within the UK: the MONARCH diagnostic accuracy study.

Background: Most neovascular age-related macular degeneration treatments involve long-term follow-up of disease activity. Home monitoring would reduce the burden on patients and those they depend on for transport, and release clinic appointments for other patients. The study aimed to evaluate three home-monitoring tests for patients to use to detect active neovascular age-related macular degeneration compared with diagnosing active neovascular age-related macular degeneration by hospital follow-up. Objectives: There were five objectives: Estimate the accuracy of three home-monitoring tests to detect active neovascular age-related macular degeneration. Determine the acceptability of home monitoring to patients and carers and adherence to home monitoring. Explore whether inequalities exist in recruitment, participants' ability to self-test and their adherence to weekly testing during follow-up. Provide pilot data about the accuracy of home monitoring to detect conversion to neovascular age-related macular degeneration in fellow eyes of patients with unilateral neovascular age-related macular degeneration. Describe challenges experienced when implementing home-monitoring tests. Design: Diagnostic test accuracy cohort study, stratified by time since starting treatment. Setting: Six United Kingdom Hospital Eye Service macular clinics (Belfast, Liverpool, Moorfields, James Paget, Southampton, Gloucester). Participants: Patients with at least one study eye being monitored by hospital follow-up. Reference standard: Detection of active neovascular age-related macular degeneration by an ophthalmologist at hospital follow-up. Index tests: KeepSight Journal: paper-based near-vision tests presented as word puzzles. MyVisionTrack®: electronic test, viewed on a tablet device. MultiBit: electronic test, viewed on a tablet device. Participants provided test scores weekly. Raw scores between hospital follow-ups were summarised as averages. Results: Two hundred and ninety-seven patients (mean age 74.9 years) took part. At least one hospital follow-up was available for 317 study eyes, including 9 second eyes that became eligible during follow-up, in 261 participants (1549 complete visits). Median testing frequency was three times/month. Estimated areas under receiver operating curves were < 0.6 for all index tests, and only KeepSight Journal summary score was significantly associated with the lesion activity (odds ratio = 3.48, 95% confidence interval 1.09 to 11.13, p = 0.036). Older age and worse deprivation for home address were associated with lower participation (χ2 = 50.5 and 24.3, respectively, p < 0.001) but not ability or adherence to self-testing. Areas under receiver operating curves appeared higher for conversion of fellow eyes to neovascular age-related macular degeneration (0.85 for KeepSight Journal) but were estimated with less precision. Almost half of participants called a study helpline, most often due to inability to test electronically. Limitations: Pre-specified sample size not met; participants' difficulties using the devices; electronic tests not always available. Conclusions: No index test provided adequate test accuracy to identify lesion diagnosed as active in follow-up clinics. If used to detect conversion, patients would still need to be monitored at hospital. Associations of older age and worse deprivation with study participation highlight the potential for inequities with such interventions. Provision of reliable electronic testing was challenging. Future work: Future studies evaluating similar technologies should consider: Independent monitoring with clear stopping rules based on test performance. Deployment of apps on patients' own devices since providing devices did not reduce inequalities in participation and complicated home testing. Alternative methods to summarise multiple scores over the period preceding a follow-up. Trial registration: This trial is registered as ISRCTN79058224. Funding: This award was funded by the National Institute of Health and Care Research (NIHR) Health Technology Assessment programme (NIHR award ref: 15/97/02) and is published in full in Health Technology Assessment; Vol. 28, No. 32. See the NIHR Funding and Awards website for further award information.

Treatment for neovascular age-related macular degeneration, the most common cause of sight loss in those over 50 years, involves regular eye injections and frequent follow-up appointments. This is inconvenient for patients and causes capacity issues in the hospital eye service. Finding tests that could be undertaken at home that could detect if a further injection and hospital appointment was required or not would increase capacity to see those at highest risk of sight loss and also reduce the burden on patients and their carers. We investigated three different visual function tests, one paper-based and two applications on an iPod Touch_TM tablet (Apple, Cupertino, CA, USA). We wanted to see if they could detect an increase in disease activity that would require treatment, compared to the decision by a retinal specialist at a traditional hospital eye outpatient visit based on clinical examination and retinal imaging. To encourage those without a smartphone or home internet to participate, we provided both an iPod Touch and Mobile Wireless-Fidelity device with a mobile contract. None of the tests performed well enough to safely monitor patients at home. Those who were willing to participate tended to be younger, had previous experience of using smartphones, sending e-mail and internet access and were more well-off than those who chose not to participate. Some participants also experienced difficulties with the devices provided and successfully uploading the data which were not related to the extent of previous information technology experience. There were also significant technical challenges for the research team. The study helpline was used heavily, considerably more than we anticipated. These tests are not ready to be used in this context. Future studies involving mobile health technology need to carefully consider how to reach those unlikely to participate and provide sufficient technical support to support long-term follow-up.

Real World One-Year Outcomes of Treatment Intensive Neovascular Age Related Macular Degeneration switched to Faricimab.

PURPOSE: To report one-year anatomical and functional real-world outcomes of treatment intensive neovascular age related macular degeneration (nAMD) patients switched to faricimab DESIGN: Retrospective multi-centre cohort study SUBJECTS: Consecutive nAMD patients on 4-weekly treatment interval with either ranibizumab or aflibercept 2mg in the last three visits within a treat and extend protocol (high treatment burden) prior to switch to faricimab at Moorfields Eye Hospital between 5/9/2022- 5/12/2022. METHODS: Total number of nAMD patients switched to faricimab were identified from electronic medical records and those who met criteria of high treatment burden were included. Data collected included pre and post-switch visual acuity (VA), treatment intervals, length of follow-up, baseline macular morphology, changes in central subfield thickness (CST), macular fluid status and adverse events. MAIN OUTCOME MEASURES: VA, CST, presence of intraretinal fluid (IRF), subretinal fluid (SRF) and injection intervals over one-year following switch to faricimab. RESULTS: A total of 130/ 286 (45.5%) eyes met inclusion criteria of being switched due to high treatment burden and 117 were included in analysis. Prior to switch to faricimab, these eyes received mean total number of injections of 33.4±19.6 over mean of 51.3±34.9 months. Mean number of injections in 12 months preceding switch was 10.1±1.6 and mean interval of the preceding three injections was 4.2±0.3 weeks. Mean VA, CST and percentage of patients with dry macula prior to switch were 66.0±11.9 ETDRS letters, 259.6±76.0µm and 18.3% respectively. Following switch, there was no statistical difference in mean VA after each visit and at 12 months. Mean CST statistically significantly reduced following the 3rd faricimab injection and at 12 months by 20.0µm (p=0.035) and 22.1µm (p=0.041) respectively. Mean treatment intervals increased to 6.9±2.3 weeks (p<0.005) at 12 months with 42.9% and 11.4% of patients being on ≥8 weekly and ≥12 weekly treatment intervals respectively. CONCLUSION: At 12 months, patients with nAMD with previous record of high treatment burden when switched to faricimab maintained visual acuities and improved anatomical outcomes on extended treatment intervals. Physician bias is inherent in these types of observational studies so a prospective randomised controlled trial is recommended to validate these findings.

Impact of treat and extend criteria on proportions that can be extended after loading phase of 2 mg aflibercept therapy for neovascular age related macular degeneration: PRECISE Report 5.

OBJECTIVE: To study the impact of definitions of various treatment extension criteria on the proportion of patients who could be extended at their first visit after the loading phase of 2 mg aflibercept therapy for neovascular age related macular degeneration (nAMD). METHODS: Patients with nAMD initiated on the loading phase of three intravitreal doses of 2 mg aflibercept in routine clinical practice were recruited from December 2019 to August 2021. The response to the loading phase was assessed at approximately 8 weeks post-loading (up to 140 days from first injection) based on different definitions of response. The proportion of patients that qualify for interval extension based on different clinical trial criteria was also evaluated. RESULTS: A total of 722 patients with visual acuity (VA) and optical coherence tomography (OCT) scans done at all 4 visits were included. Of these 32.4% of eyes responded with complete macular fluid resolution after the first injection with no recurrence through the loading phase (super-responders) while 26.9% had persistent macular fluid in all 4 visits (true non-responders). The rest were considered suboptimal responders. Change in VA showed marked variations within each of these categories of fluid resolution. For extension of next treatment interval, if presence of any macular fluid at the post-loading visit is the only criteria considered, about 50% could be extended to 8 weeks. If both VA worsening by ≥5 letters and a > 25 µm increase in central sub-field thickness (CST) are considered, 90% will be eligible for interval extension. CONCLUSION: Clinical trial designs and pre-defined treatment extension/shortening criteria determine the proportion of patients requiring treatment in the post-loading visit. The short and long-term impact of interval extension immediate post-loading on visual outcome in clinical practice is unknown.

Comparing generative and retrieval-based chatbots in answering patient questions regarding age-related macular degeneration and diabetic retinopathy.

BACKGROUND/AIMS: To compare the performance of generative versus retrieval-based chatbots in answering patient inquiries regarding age-related macular degeneration (AMD) and diabetic retinopathy (DR). METHODS: We evaluated four chatbots: generative models (ChatGPT-4, ChatGPT-3.5 and Google Bard) and a retrieval-based model (OcularBERT) in a cross-sectional study. Their response accuracy to 45 questions (15 AMD, 15 DR and 15 others) was evaluated and compared. Three masked retinal specialists graded the responses using a three-point Likert scale: either 2 (good, error-free), 1 (borderline) or 0 (poor with significant inaccuracies). The scores were aggregated, ranging from 0 to 6. Based on majority consensus among the graders, the responses were also classified as 'Good', 'Borderline' or 'Poor' quality. RESULTS: Overall, ChatGPT-4 and ChatGPT-3.5 outperformed the other chatbots, both achieving median scores (IQR) of 6 (1), compared with 4.5 (2) in Google Bard, and 2 (1) in OcularBERT (all p ≤8.4×10-3). Based on the consensus approach, 83.3% of ChatGPT-4's responses and 86.7% of ChatGPT-3.5's were rated as 'Good', surpassing Google Bard (50%) and OcularBERT (10%) (all p ≤1.4×10-2). ChatGPT-4 and ChatGPT-3.5 had no 'Poor' rated responses. Google Bard produced 6.7% Poor responses, and OcularBERT produced 20%. Across question types, ChatGPT-4 outperformed Google Bard only for AMD, and ChatGPT-3.5 outperformed Google Bard for DR and others. CONCLUSION: ChatGPT-4 and ChatGPT-3.5 demonstrated superior performance, followed by Google Bard and OcularBERT. Generative chatbots are potentially capable of answering domain-specific questions outside their original training. Further validation studies are still required prior to real-world implementation.

Real-world service costs for neovascular-AMD clinics in the United Kingdom: structured literature review and scenario analysis.

OBJECTIVE: Current cost-effectiveness analyses (CEA) emphasize drug costs as the differentiator between NICE recommended anti-VEGF treatments but may neglect real-world non-drug costs of running nAMD services in the UK. To address this, this study identified real-world non-drug service cost items relevant to UK NHS nAMD clinics, including costs arising from operational strain (demand exceeding capacity). METHODS: Cost items were identified by a structured literature review of peer-reviewed and grey literature, and an expert panel of 10 UK-based ophthalmologists with relevance to real-world practice. These items underwent meta-synthesis and were then determined in a consensus exercise. RESULTS: Of 237 cost items identified, 217 (91.6%) met the consensus threshold of >0.51 and were included in the nAMD Service Non-Drug Cost Instrument (nAS). Sensitivity of cost items taken from UK Health Technology Assessment (HTA) using the nAS as the reference standard was low (HTAmin: 1.84%, 95% CI 0.50-4.65%; HTAmax: 70.51%, 95% CI 63.96-76.49%). False negative rates showed variable likelihood of misclassifying a service by cost burden depending on prevalence. Scenario analysis using cost magnitudes estimated annual per-patient clinic cost at £845 (within capacity) to £13,960 (under strain) compared to an HTAmin estimate of £210. Accounting for cost of strain under an assumed 50% increase in health resource utilization influenced cost-effectiveness in a hypothetical genericisation scenario. CONCLUSION: Findings suggested that HTA underestimates UK NHS nAMD clinic cost burden with cost of strain contributing substantial additional unmeasured expense with impact on CEA. Given potential undertreatment due to strain, durability is suggested as one of the relevant factors in CEA of nAMD anti-VEGF treatments due to robustness under limited capacity conditions affecting UK ophthalmology services.

When considering how well treatments work versus how much they cost, the focus is usually only on the price of the medicine itself. However, other real-world costs exist. In the UK, when treating certain eye problems such as neovascular age-related macular degeneration (nAMD), there are additional expenses related to running clinics and managing treatments that often go unnoticed. To get a better understanding of these hidden costs, the study examined factors like clinic workload and the extra expenses that come with it. Ten eye doctors in the UK were consulted for their expert opinions and numerous research papers were reviewed to identify these additional costs. The study grouped different costs in a tool called the nAMD Service Non-Drug Cost Instrument (nAS). When the findings of the nAS tool were compared to the usual methods of calculating costs, it was found that the conventional approach overlooked many of the actual expenses. Busy clinics face unique challenges, such as higher operational costs associated with staffing for extended hours, emergency appointments, extended waiting times and the potential to miss optimal treatment windows. This can lead to disease progression and the onset of comorbidities, which require more complex and costly treatments. Recognizing these real costs is crucial when making decisions about treatments, especially when treatments require more frequent visits to eye clinics. This study emphasizes the importance of considering all expenses, not just the obvious ones like medication and doctor visits when determining the most effective way to manage eye conditions like nAMD in the UK.

Metabolic memory and diabetic retinopathy: Legacy of glycemia and possible steps into future.

The response of retinal pathology to interventions in diabetic retinopathy (DR) is often independent of the glycated hemoglobin (HbA1c) values at the point of care. This is despite glucose control being one of the strongest risk factors for the development and progression of DR. Previous preclinical and clinical research has indicated metabolic memory, whereby past cumulative glucose exposure may continue to impact DR for a prolonged period. Preclinical studies have evaluated punitive metabolic memory through poor initial control of DM, whereas clinical studies have evaluated protective metabolic memory through good initial control of DM. In this narrative review, we evaluate the preclinical and clinical evidence regarding metabolic memory and discuss how this may form the basis of preventive care for DR by inducing "metabolic amnesia" in people with a history of uncontrolled diabetes in the past. While our review suggested mitochondrial biology may be one such target, research is still far from a possible clinical trial. We discuss the challenges in such research.

Study protocol: optical coherence tomography angiography for the detection of neovascular age-related macular degeneration: a comprehensive multicentre diagnostic accuracy study in the UK-the ATHENA study.

INTRODUCTION: The diagnosis of neovascular age-related macular degeneration (nAMD), the leading cause of visual impairment in the developed world, relies on the interpretation of various imaging tests of the retina. These include invasive angiographic methods, such as Fundus Fluorescein Angiography (FFA) and, on occasion, Indocyanine-Green Angiography (ICGA). Newer, non-invasive imaging modalities, predominately Optical Coherence Tomography (OCT) and Optical Coherence Tomography Angiography (OCTA), have drastically transformed the diagnostic approach to nAMD. The aim of this study is to undertake a comprehensive diagnostic accuracy assessment of the various imaging modalities used in clinical practice for the diagnosis of nAMD (OCT, OCTA, FFA and, when a variant of nAMD called Polypoidal Choroidal Vasculopathy is suspected, ICGA) both alone and in various combinations. METHODS AND ANALYSIS: This is a non-inferiority, prospective, randomised diagnostic accuracy study of 1067 participants. Participants are patients with clinical features consistent with nAMD who present to a National Health Service secondary care ophthalmology unit in the UK. Patients will undergo OCT as per standard practice and those with suspicious features of nAMD on OCT will be approached for participation in the study. Patients who agree to take part will also undergo both OCTA and FFA (and ICGA if indicated). Interpretation of the imaging tests will be undertaken by clinicians at recruitment sites. A randomised design was selected to avoid bias from consecutive review of all imaging tests by the same clinician. The primary outcome of the study will be the difference in sensitivity and specificity between OCT+OCTA and OCT+FFA (±ICGA) for nAMD detection as interpreted by clinicians at recruitment sites. ETHICS AND DISSEMINATION: The study has been approved by the South Central-Oxford B Research Ethics Committee with reference number 21/SC/0412.Dissemination of study results will involve peer-review publications, presentations at major national and international scientific conferences. TRIAL REGISTRATION NUMBER: ISRCTN18313457.

The UK clinical eye research strategy: refreshing research priorities for clinical eye research in the UK.

OBJECTIVES: To validate and update the 2013 James Lind Alliance (JLA) Sight Loss and Vision Priority Setting Partnership (PSP)'s research priorities for Ophthalmology, as part of the UK Clinical Eye Research Strategy. METHODS: Twelve ophthalmology research themes were identified from the JLA report. They were allocated to five Clinical Study Groups of diverse stakeholders who reviewed the top 10 research priorities for each theme. Using an online survey (April 2021-February 2023), respondents were invited to complete one or more of nine subspecialty surveys. Respondents indicated which of the research questions they considered important and subsequently ranked them. RESULTS: In total, 2240 people responded to the survey (mean age, 59.3 years), from across the UK. 68.1% were female. 68.2% were patients, 22.3% healthcare professionals or vision researchers, 7.1% carers, and 2.1% were charity support workers. Highest ranked questions by subspecialty: Cataract (prevention), Cornea (improving microbial keratitis treatment), Optometric (impact of integration of ophthalmic primary and secondary care via community optometric care pathways), Refractive (factors influencing development and/or progression of refractive error), Childhood onset (improving early detection of visual disorders), Glaucoma (effective and improved treatments), Neuro-ophthalmology (improvements in prevention, diagnosis and treatment of neurodegeneration affecting vision), Retina (improving prevention, diagnosis and treatment of dry age-related macular degeneration), Uveitis (effective treatments for ocular and orbital inflammatory diseases). CONCLUSIONS: A decade after the initial PSP, the results refocus the most important research questions for each subspecialty, and prime targeted research proposals within Ophthalmology, a chronically underfunded specialty given the substantial burden of disability caused by eye disease.

Subretinal transient hyporeflectivity in neovascular age-related macular degeneration and its response to a loading phase of aflibercept: PRECISE report 4.

PURPOSE: To describe the prevalence of subretinal transient hyporeflectivity (STHR) in exudative neovascular age-related macular degeneration (nAMD) and its response to a loading phase of aflibercept. METHODS: Optical coherence tomography (OCT) scans of treatment-naïve nAMD patients captured at baseline and after a loading phase of aflibercept were graded for presence of STHR, defined as a small, well-defined, round, subretinal, hyporeflective area, delimited between the ellipsoid zone (EZ) and the retinal pigmented epithelium/Bruch membrane complex. OCT parameters recorded were macular neovascularisation (MNV) subtypes, location of retinal fluids (subretinal fluid, SRF and intraretinal fluid, IRF), central retinal and choroidal thickness. Response was defined as absence of IRF and SRF. Factors associated with completely resolved STHR versus persistent STHR post-loading phase were compared. RESULTS: 2039 eyes of 1901 patients were analysed. STHR was observed in 79 eyes of 78 patients, with an estimated prevalence of 3.87% (95% CI 3.08-4.81%). STHR were seen in 44 type 1 MNV (56%), 27 with type 2 (34%), and 8 with type 3 (10%). At baseline, a total of 303 STHR were present, ranging between 1-22 per eye. The total number of STHR reduced significantly after the loading phase to 173 (p = 0.002). Complete disappearance of STHR was seen in 44 eyes (56%) and persistent STHR in the rest (44%). CONCLUSIONS: STHR may represent a marker of low-grade exudation in nAMD eyes with good response to a loading phase of aflibercept. However, its potential role as an independent nAMD activity biomarker is limited as most resolve after the loading phase.

Morph-SSL: Self-Supervision with Longitudinal Morphing for Forecasting AMD Progression from OCT Volumes.

The lack of reliable biomarkers makes predicting the conversion from intermediate to neovascular age-related macular degeneration (iAMD, nAMD) a challenging task. We develop a Deep Learning (DL) model to predict the future risk of conversion of an eye from iAMD to nAMD from its current OCT scan. Although eye clinics generate vast amounts of longitudinal OCT scans to monitor AMD progression, only a small subset can be manually labeled for supervised DL. To address this issue, we propose Morph-SSL, a novel Self-supervised Learning (SSL) method for longitudinal data. It uses pairs of unlabelled OCT scans from different visits and involves morphing the scan from the previous visit to the next. The Decoder predicts the transformation for morphing and ensures a smooth feature manifold that can generate intermediate scans between visits through linear interpolation. Next, the Morph-SSL trained features are input to a Classifier which is trained in a supervised manner to model the cumulative probability distribution of the time to conversion with a sigmoidal function. Morph-SSL was trained on unlabelled scans of 399 eyes (3570 visits). The Classifier was evaluated with a five-fold cross-validation on 2418 scans from 343 eyes with clinical labels of the conversion date. The Morph-SSL features achieved an AUC of 0.779 in predicting the conversion to nAMD within the next 6 months, outperforming the same network when trained end-to-end from scratch or pre-trained with popular SSL methods. Automated prediction of the future risk of nAMD onset can enable timely treatment and individualized AMD management.