Siegfried Karl Wagner

We investigate the association between multimodal retinal imaging parameters with prevalent and incident atrial fibrillation (AF). For this cross-sectional and longitudinal analysis we utilized data from two independent cohorts: AlzEye (retrospective, age ≥ 40 years, 2008–2018) and UK Biobank (UKB; prospective, age 40–69 years, recruited 2006–2010). Retinal sublayer thicknesses and retinovascular indices were extracted from macular optical coherence tomography and color retinal photography, respectively. AF was ascertained via self-report and hospital admissions, association between retinal metrics with prevalent and incident AF were estimated using mixed-effects linear and frailty models. In AlzEye (n = 50,651), prevalent AF (n = 5,375; 11%) was associated with significantly smaller retinal venular calibre (-0.05 µm, 95% CI: -0.08, -0.03) and thinner macular retinal nerve fibre layer (-0.69 µm, 95% CI: -0.91, -0.46), ganglion cell-inner plexiform layer (mGCIPL; -1.96 µm, 95% CI: -2.34, -1.58), inner nuclear layer (-0.57 µm, 95% CI: -0.75, -0.38), and photoreceptor layer (-0.93 µm, 95% CI: -1.38, -0.48). Venular fractal dimension was lower (-0.08, 95% CI: -0.11, -0.06), while arteriolar fractal dimension was higher (0.05, 95% CI: 0.02, 0.08) in individuals with AF. Evidence of thinner mGCIPL was replicated in UKB (-1.19 µm, 95% CI: -1.73, -0.65). In UKB (n = 39,013 baseline without AF), 838 developed AF over a mean of 4.0 years. Adjusting for sociodemographic, clinical and lifestyle factors, each SD decrease in mGCIPL thickness was associated with a significant 27% increased hazard of AF (95% CI: 5%, 54%). Differences in the inner retinal sublayers, corresponding to retinal ganglion cells, are associated with prevalent and incident AF across two large independent cohorts, supporting the potential role of widely-accessible retinal imaging for early detection and better management of AF.

Background: Large language model (LLM) agents capable of generating and executing statistical code from natural language may broaden access to clinical data analysis, yet which pipeline stages they perform reliably and which require expert oversight remain poorly defined. Objective: This study aimed to evaluate the performance and systematic failure modes of an LLM agent across 5 stages of a clinical data analysis workflow. Methods: The publicly available dataset and R script (R Foundation for Statistical Computing) were drawn from a previously published study of 12-year outcomes in 7802 patients with eyes with neovascular age-related macular degeneration at Moorfields Eye Hospital. Participants were evaluated using an LLM agent (Claude; Anthropic) across 3 interaction modes (Chat, Code, and Cowork). It was asked to perform 3 levels of data analysis practice: prompt A, to generate research questions from raw data only; prompt B, to develop a statistical analysis plan (SAP) from a high-level clinical objective, then execute it; and prompt C, to execute an analysis given an investigator-drafted SAP. Each was replicated 3 times (27 total runs). Qualitative evaluation of research question thematic coverage (prompt A), SAP completeness against a reference checklist (prompt B), and evaluation of execution outputs against validated reference values and of result text and narrative summaries against execution logs (prompts B and C) was conducted. Results: The agent generated 18 clinically grounded questions spanning 7 domains; Cowork mode uniquely reached 3 thematic areas requiring data-driven methods. All 9 SAPs correctly identified the statistical framework. Kaplan-Meier estimates were near-identical across 17 completed runs. Systematic execution errors emerged: SAP quality did not predict code correctness, and within-mode errors propagated identically across independent repetitions. Result text accurately reflected execution logs in nearly all runs, though unit propagation and an undisclosed postcrash rerun were identified. Of 17 narrative summaries, 8 were fully satisfactory; 2 runs produced clinically meaningful errors. Conclusions: LLM agents perform reliably for question generation and SAP drafting but require expert verification of formula composition, cohort boundary logic, and concordance computation before results are reported. Using an ophthalmology dataset as a controlled testbed, this study develops and applies an evaluation framework whose lessons are likely applicable across clinical specialties.

Background: Deep‐learning models are capable of predicting age from retinal scans and the difference between this and chronological age, retinal age gap, has been shown to be significantly associated with risk of mortality, cardiovascular diseases, and kidney failure. Objective: Our aim was to investigate the association between Parkinson's disease (PD) and retinal age gap in a large, real‐world population. Methods: The deep‐learning model was trained and validated using 19,200 fundus images from 11,052 participants in the UK Biobank study. This cross‐sectional analysis used data from the AlzEye study, a retrospective cohort of 154,830 patients aged 40 years and older. Retinal age gap was calculated as the difference between the retina‐predicted age by the model and chronological age. Adjusted linear mixed‐effects models were used to assess the association between PD and retinal age gap. Results: A total of 776 individuals with PD (mean [standard deviation] age, 69.6 [8.9] years; 42.9% female) and 93,944 without PD (mean age, 57.6 [13.7] years; 51.6% female) were included. After adjusting for confounders, retinal age gap was significantly greater in those with PD ( β , 5.92, 95% confidence interval: 5.33–6.51; P < 0.001). Stratified analyses showed a larger retinal age gap across all age deciles ≥50 years. Within the PD group, males and individuals with diabetes had greater retinal age gaps than females and individuals without diabetes. Conclusions: Individuals with PD aged ≥50 years demonstrate older retinal age relative to chronological age. Retinal age gap may serve as a biomarker of accelerated aging in PD. © 2026 The Author(s). Movement Disorders published by Wiley Periodicals LLC on behalf of International Parkinson and Movement Disorder Society.

Retinal age is a biological ageing marker estimated from retinal images. Its deviation from chronological age, termed the retinal age gap (RAG), has been associated with adverse health outcomes. However, fundamental characteristics of retinal age models that are critical for interpreting and applying RAG remain underexplored, including the consistency of these associations across age subgroups as well as model generalisability across different cohorts. In this study, we used 327,764 retinal images to develop a retinal age model and investigate associations between RAG and systemic diseases. We examined association consistency across age subgroups and evaluated model performance on three external datasets with distinct imaging devices and populations. RAG showed positive associations with systemic diseases in young and middle-aged subgroups, but negative associations in the aged subgroup, revealing substantial age-dependent inconsistency. We further showed that this inconsistency was explained by regression-to-the-mean effects which varied by health status. In external evaluations, the generalisability of RAG varied considerably by cohort and intended application. Large age estimation errors did not necessarily indicate limited clinical utility, instead suggesting that model generalisability should be defined and evaluated in an application-specific manner rather than based on age estimation accuracy alone. Overall, this study investigates key characteristics of retinal age models that affect the interpretation of RAG and its disease associations. Our findings highlight the importance of reporting age-dependent associations, and emphasise the need for multi-dimensional, application-specific external evaluation for reliable use of retinal age and RAG in future clinical applications. More broadly, our findings and practical recommendations may extend to wider biological age estimation models.

Acute retinal ischaemic events, ranging from central retinal artery occlusions (CRAO) and branch retinal artery occlusions (BRAO) to transient ischaemic monocular vision loss (TMVL), cause sudden and often severe vision loss, yet their broader implications for systemic vascular health are not well defined. To investigate this, we conducted a large, propensity-matched cohort study including a total of 566 patients at an academic ophthalmic hospital in London, United Kingdom, linking ophthalmic diagnoses to national records of cardiovascular events and mortality. Our findings reveal that patients experiencing a CRAO have a nearly threefold increased risk of a subsequent ischaemic stroke or myocardial infarction, when compared to matched controls. This elevated risk, however, was not apparent in individuals who suffered a BRAO or TMVL, suggesting a distinct risk profile for CRAO. Systemic comorbidities like hypertension and diabetes were prevalent in all groups. Atrial fibrillation was newly diagnosed in 6% of patients with CRAO within one year. Patients with BRAO were more likely to undergo carotid endarterectomy than those with CRAO. The substantial risk of serious cardiovascular morbidity following retinal arterial ischaemia highlights the need to ensure such patients benefit from early treatment and prolonged cardiac monitoring and rigorous implementation of secondary prevention in RAO and suspected embolic TMVL.