Translational Vision Science & Technology

Purpose: To evaluate the performance of secure cloud-based large language models (LLMs) in extracting glaucoma diagnosis, type, and severity from free-text clinical notes in the electronic health record (EHR). Design: Retrospective chart review analysis. Participants: 1,250 subjects from the Bascom Palmer Ophthalmic Repository. Methods: Clinical notes of glaucoma-related encounters between 2014 and 2024 were extracted from the Bascom Palmer Ophthalmic Repository. Two fellowship-trained glaucoma specialists annotated clinical notes for glaucoma presence, type, and severity at the eye level. The dataset was split into development (10%), validation (10%), and test (80%) sets. Development and validation sets were used for prompt engineering and refinement, and the held-out test set was used for evaluation. Five LLMs (Claude Opus 4.6, DeepSeek-V3.2, GPT-5.2, Grok 4.1, and Qwen3.6-35B-A3B) were accessed via Azure AI Foundry within HIPAA-compliant containers. Model performance was assessed using standard metrics. Clinician-entered ICD-10 codes were also compared with adjudicated labels. Main Outcome Measures: Gwet AC1, accuracy, sensitivity, specificity, and F1-score. Results: Inter-grader agreement was high for glaucoma detection (Gwet AC1= 0.930 (95% CI: 0.917-0.945), type classification (Gwet AC1= 0.917 (95% CI: 0.904-0.930), and severity staging (Gwet AC1= 0.901 (95% CI: 0.884-0.916). For glaucoma diagnosis, LLMs demonstrated high overall accuracy, with Claude achieving 97.5%, DeepSeek 96.0%, GPT 96.2%, Grok 94.4%, and Qwen 95.5%. F1 scores for glaucoma detection ranged from 95.4% to 98.9% across models. For glaucoma type classification, accuracies were 97.1%, 94.2%, 94.2%, 94.0%, and 94.4% for Claude, DeepSeek, GPT, Grok, and Qwen, respectively. F1 scores for the most prevalent type (POAG) ranged from 96.3% to 98.9%. For severity staging, accuracies were 95.0%, 94.8%, 94.5%, 94.0%, and 95.2%, respectively, with F1 scores ranging from 89.7% to 96.3% across severity categories and models. ICD-10 codes demonstrated substantially lower performance for type and severity staging, with overall accuracies of 89.2% and 58.5%, respectively. Conclusions: Secure cloud-based LLMs accurately extracted glaucoma diagnosis, type, and severity information from free-text ophthalmology notes, achieving performance approaching expert clinician adjudication while substantially outperforming ICD-based phenotyping approaches, particularly for disease severity classification. These findings demonstrate the potential of LLMs to transform unstructured clinical documentation into scalable, research-ready phenotypic data for large-scale glaucoma cohort development and EHR-based ophthalmic research.

PurposeTo evaluate associations between green autofluorescence (GAF) and structural and functional measures relevant to retinal and optic neuropathies, and to establish normative GAF values across the optic nerve head (ONH), macula, and papillofoveal bundle (PFB) in rhesus macaques. MethodsEighty-two macaques with normal ONH morphology by spectral-domain optical coherence tomography (SD-OCT) were included with a mean {+/-} SD age of 12.76 {+/-} 7.18 (range 0.11-29.39) years. The GAF images were acquired in the ONH, macula and PFB with the OcuMet Beacon. In a subset of macaques (n=19), pattern electroretinogram (PERG) and photopic full-field ERG including the photopic negative response (PhNR) were recorded. ResultsThe GAF significantly increased with age in the ONH, macula and PFB. After adjusting by age, there were no sex differences, but IOP showed a positive association with macular GAF. At the ONH, higher GAF correlated with thinner retinal nerve fiber layer, inner and outer segment complex, and total retinal thickness. In the macula, inner nuclear layer thickness was positively associated with GAF, whereas outer plexiform layer and inner and outer segment complex were inversely associated. The PERG amplitudes inversely tracked ONH GAF. ConclusionsGAF rises with age and IOP, couples to retinal structure, and at the ONH, aligns with inner-retinal functional indices. This study provides a regional reference for GAF in rhesus macaques. Translational RelevanceNormative GAF data in healthy rhesus macaques provide a framework for interpreting this noninvasive signal in translational studies of retinal and optic nerve disease.

Objective: To investigate the performance of anterior segment (AS) OCT quantitative imaging of anterior chamber inflammation in uveitis patients with diverse demographics. Design: Prospective cross-sectional study. Participants: 144 adult patients managed at a tertiary care service in the UK Methods: Repeated swept-source ASOCT imaging was performed pre- and post-pupil dilation (i.e. 4 scan sets). Inflammation was quantified using a validated human in the loop automated image analysis pipeline, Minuscule Cell Detection (MCD), which identified and counted putative inflammatory cells on AS-OCT. Main Outcome Measures: Test-retest variability of ASOCT and diagnostic accuracy of various ASOCT derived measurands (minimum, maximum, median counts per cross sectional image, and total counts across volume image sets per eye or MINCC, MAXCC, MEDCC and TOTCC) versus Standardization of Uveitis Nomenclature (SUN) grading system as assessed by a uveitis specialist. Results: A total of 281 eyes were included in the analysis. Median age was 48 years (IQR 36 to 64). Strong test-retest measurand reliability was demonstrated, with a 95% tolerance interval ratio 0.3 to 3.0. The best diagnostic performances for SUN activity were observed with the MINCC threshold of 3 particles (negative predictive value for clinical activity of 89.8%, 95% CI 83.0 to 94.1). Associations between ASOCT measurands and patient age (adjusted coefficient 7.5 additional particles, 95% CI 0.5 to 14.6, p<0.04 for age over 60 years versus under 44), and pigment load (52.8, 11.8 to 92.9, p<0.01 in eyes with AC pigment versus without) were noted. Conclusions: ASOCT assessment of anterior chamber inflammation in uveitis meets current recommendations for quantitative imaging biomarkers, demonstrating strong repeatability, linearity with clinical assessment scores and stability with pupil dilation and patient characteristics of ethnicity and lens status. The absence of variability in diagnostic indices across derived measurands suggests similar performance across different acquisition protocols. Further longitudinal cross-platform studies are needed to determine limitations of use.

Purpose: To assess the feasibility of quantitative fundus autofluorescence (FAF) measurement in early age-related macular degeneration (AMD) using the freely available ImageJ software, to characterize signal intensity across FAF patterns, and to evaluate interobserver reproducibility in pattern classification. Methods: Single-center, non-blinded, retrospective, consecutive-case analytical study. FAF images acquired with Spectralis OCT+HRA (Heidelberg Engineering) from patients with early dry AMD seen at a tertiary referral center between January 2010 and September 2016 were analyzed. A standardized 300x300-pixel region of interest (ROI) centered on the fovea was evaluated in ImageJ v2.0.0-rc54/1.51h (Fiji distribution). Mean, minimum, and maximum autofluorescence (AF) pixel intensity were recorded. Each image was independently classified according to the Bindewald classification system by two graders; a third senior grader adjudicated discordances. Cohen's kappa (k) was used to assess interobserver agreement. Results: Of 423 patients with available FAF studies, 107 had dry AMD; 45 met quality and diagnostic criteria for early AMD and were included in the quantitative analysis. Mean age was 73.47 +/- 8.1 years; 62.2% were female. Mean FAF intensity was 120.26 (range 74.76-160.79); mean minimum was 32.07 (range 3-63) and mean maximum was 205.80 (range 125-255). Seven of eight Bindewald patterns were identified; the stippled pattern was absent. The most frequent pattern was minimal changes (31.1%), followed by increased focal (24.4%) and patchy (15.6%). Reticular pattern showed the highest mean AF (143.8), while lace pattern showed the lowest (88.4). Interobserver agreement for Bindewald pattern classification was almost perfect (k = 0.969; 95% CI, 0.908-1.000; p < 0.001). Agreement for lesion extent was moderate (k = 0.531) and for foveal involvement was substantial (k = 0.622). Conclusions: Quantitative FAF evaluation of early AMD using ImageJ is feasible and reproducible. ImageJ represents a cost-free alternative for multimodal retinal image analysis, with potential for automated screening applications in resource-limited settings. Keywords: age-related macular degeneration; fundus autofluorescence; ImageJ; quantitative autofluorescence; image analysis; Bindewald classification; interobserver agreement

Purpose: New therapeutic strategies such as optogenetics have created a need for accurate tracking of inner retina degeneration in Retinitis pigmentosa (RP) patients. We introduce two tailored deep learning models to segment the RNFL (retinal nerve fibre layer), GCIPL (ganglion cell inner plexiform layer), INL (inner nuclear layer), CFT (central foveal thickness) and RPE (retinal pigment epithelium) in RP: The first is based on a Segment Anything Model (SAM), the second on nnU-Net. To our knowledge, SAM has not yet been applied to retinal layers in OCT data. Methods: SD-OCT images of a retrospective cohort of 37 RP patients were included. Data for four training cycles were prepared semi-automatically in MATLAB, then assessed and corrected by three expert graders. 1,700 segmented B-Scans from two open datasets were used for pretraining. For post-processing, semantic retinal boundary detection was developed. The final models, OCT-SAM and nnU-Net, were trained on 228 annotated RP scans. Detected layer thicknesses were validated against manual segmentation at 90 random points in 30 OCT B-Scans. Finally, OCT-SAM was tested on three RP cases with retrospective, longitudinal OCT data. Results: nnU-Net achieved a precision, recall and F-1 score of 0.96 while OCT-SAM performance resulted in slightly lower values of 0.93, 0.8 and 0.85, respectively. OCT-SAM measurements had low bias and good agreement with manual annotations, confirming reliability. Conclusions: OCT-SAM enabled fast data annotation and tool integration, whereas nnU-Net provided the best segmentation performance. OCT-SAM demonstrated longitudinal reproducibility and detected RP-characteristic pathologies and degenerative changes. Future work will extend OCT-SAM to 3D OCT segmentation.

Abstract Background. Quantitative optical coherence tomography (OCT) measurements are essential for retinal disease monitoring, yet leading vendors store acquisition data in undocumented proprietary formats or encode measurements exclusively in private DICOM tags inaccessible to open systems. Methods. We present Transducin, an open-source Python library that reverse-engineers the undocumented Optopol Revo FC130 and Revo 60 .OPT binary format and extracts quantitative measurements from Zeiss Cirrus HDOCT private DICOM tags, generating TID 1500 Structured Reports with SNOMEDCT coded findings for both platforms. A novel finding, that OCTPARAMS tag 23 encodes ocular laterality through the arithmetic sign of the foveal horizontal position, enables geometry based laterality inference requiring no operator data entry, validated across 18 files from two device models and four software versions with 100% accuracy. Results. The primary corpus of 452 Optopol .OPT files (73 patients, 7 acquisition types) was parsed with 100% success. Cross-version compatibility was confirmed across SOCT versions 11.5.0 through 21.5.0, spanning approximately eight years of software development. The Zeiss Cirrus pipeline generated TID 1500 SRs for all 41 applicable studies (100%), yielding CMT 203to 630um and RNFL 53 to123 um across a clinically representative range. Conclusions. Transducin provides the first publicly documented specification of the Optopol .OPT format and the first open-source multivendor pipeline generating SNOMEDCT coded DICOM Structured Reports from both Optopol Revo and Zeiss Cirrus devices, closing a gap explicitly confirmed by both manufacturers' own documentation. The code is available at https://github.com/oftalmos-org/transducin (Apache License 2.0).

Retinal and choroidal vascular diseases are major causes of vision loss that require frequent intravitreal anti-VEGF therapy. Anti-angiogenic peptide AXT107 demonstrated efficacy in preclinical studies and was advanced to the clinical stage. To provide for sustained delivery of the peptide and avoid complications with intravitreal injection, we evaluated suprachoroidal delivery of AXT107 microparticles (MP-AXT107). The original, soluble AXT107 formulation was ineffective at inhibiting laser-induced choroidal neovascularization (CNV) in our rat model and was consequently reformulated as microparticles. MP-AXT107 demonstrated high peptide incorporation efficiency, reproducible morphology, and physical and chemical stability for at least 9 months under refrigerated storage. In the rat CNV model, suprachoroidal MP-AXT107 significantly reduced neovascular area by approximately 60% relative to vehicle controls. Safety and durability were evaluated in a 9-month GLP toxicology study in Gottingen minipigs following a single suprachoroidal injection of vehicle or MP-AXT107 (0.125-1.25 mg/eye). Transient increases in IOP and mild ocular inflammatory findings were observed immediately following administration but resolved rapidly without lasting effects. No treatment-related adverse ocular findings were observed during the remainder of the study, and the highest tested dose (1.25 mg/eye) was established as the no-observed-adverse-effect level. Bioanalysis at study completion demonstrated persistent AXT107 localization primarily within choroid/RPE and scleral tissues, with no signs of systemic exposure. Collectively, these findings demonstrate that suprachoroidal delivery of MP-AXT107 enables sustained anti-angiogenic activity with favorable ocular safety and prolonged tissue retention, supporting further clinical development as a durable therapy for retinal and choroidal vascular diseases.

Purpose: To evaluate the safety, procedural performance, sample recovery, and surgeon preference of an ophthalmic needle designed specifically for anterior chamber (AC) paracentesis. Methods: In this multicenter study, AC paracentesis was performed in clinic and operating-room settings using a 32-gauge x 4-mm needle with low dead space. The procedure was evaluated using a standardized physician survey. Prespecified outcomes included procedure-related adverse events (primary outcome), needle entry and handling, aspiration and sample recovery, comparative performance versus a 30-gauge needle, and physician preference for future use. Results: A total of 110 needle uses by eight surgeons were included. No ocular complications occurred, including lens or iris injury, hyphema, AC collapse, wound leak, hypotony, infection, or retinal complication, and no procedure required needle exchange or conversion to another device. Two technical events without ocular sequelae were noted, in which needle entry was partial thickness and did not reach the AC (1.8%; exact 95% CI, 0.2%-6.4%). Physicians rated needle entry, handling and sample recovery as good or excellent. Compared with a 30-gauge needle, the study needle was rated as at least comparable across all assessed domains. All surgeons rated it better or much better for intra-procedural safety and preferred it for future AC taps. Conclusions and Relevance: This short, 32-gauge low-dead-space ophthalmic needle demonstrated a favorable safety profile and was preferred over a 30-gauge needle by all surgeons. As aqueous humor liquid biopsy expands in clinical diagnostics and trials, an ophthalmic-specific needle design may help improve the consistency and safety of aqueous humor collection for molecular analysis and broader clinical use. Keywords: Anterior chamber paracentesis; Aqueous humor; Liquid biopsy; Low dead space; Ophthalmic needle

Purpose: To validate an automated pipeline to detect and quantify focal retinal and choroidal pulsation areas that are synchronous with the cardiac cycle in video indocyanine green angiography (ICGA). Design: Retrospective, observational, hypothesis-generating validation study Subjects, Participants: Consecutive patients with a diagnosis of central serous chorioretinopathy (CSCR) in one or both eyes. Methods: Videos were acquired on Heidelberg HRA+OCT. The pipeline consisted of three steps: signal extraction, foci detection, and quantification. After registration of the constituent frames, each pixel's intensity signal was analyzed at the presumed cardiac frequency (tested from a sample of three detectable frequencies). A synchrony score combining local phase coherence with oscillation amplitude was then derived and computed using a standard deviation ({sigma}) above each video's background oscillation value. Two masked graders marked the retinal and choroidal pulsation areas twice. We compared detection of the pulsation areas against grader consensus using a receiver operating characteristic curve (using multiple grid sizes to divide the scan area) and, separately, using a signal-based area-reduction method to obtain an optimum {sigma} value. Main Outcome Measures: Agreement between the automated algorithm and human graders in detection of pulsation foci, and the optimum threshold multiplier ({sigma}). Results: We studied 20 ICGA videos from 20 eyes. At the 16-pixel grid size, the pipeline achieved a mean area under the curve (AUC) of 0.914, sensitivity of 0.86, and specificity of 0.80. Grader agreement improved with larger grid size, reaching substantial-to-strong levels for choroidal annotations. The two independent validation methods demonstrated similar {sigma} values that differed by 0.62{sigma}, supporting {sigma}=4.0 as the optimum value. Conclusions: We report the first automated method to quantify retinal and choroidal vascular pulsation on video ICGA. It measures pixels that oscillate over time with the presumed cardiac cycle and works reliably at the spatial scale (grid level) where experts agree. Pulsatile hemodynamics may add a new vascular biomarker for glaucoma, diabetes, hypertension, and pachychoroid diseases.

Purpose: To evaluate the efficacy of large language models (LLMs) in extracting medication-related information from glaucoma clinical notes in the electronic health record (EHR). Design: Cross-sectional. Subjects: 1,250 subjects in the Bascom Palmer Ophthalmic Repository. Methods: Extracted clinical notes from glaucoma-related encounters between 2014 and 2024 were labeled by two glaucoma specialists with a third serving as an adjudicator. Graders were asked to label current topical medications (CTM), proposed changes to topical medications ({Delta}TM), current oral medications (COM), and proposed changes to oral medications ({Delta}OM) in a structured fashion. The dataset was split into development (10%), validation (10%), and test (80%) sets stratified by clinician. Development and validation sets were used to engineer and refine prompts, and the held-out test set was used for model assessment. Five LLMs (Claude Opus 4.6, DeepSeek-V3.2, GPT 5.2, Grok 4.1, and Qwen3.6-35B-A3B) were accessed via Microsoft Azure AI Foundry within a HIPAA-compliant environment. Inter-grader agreement was assessed with Gwet AC1. LLM performance was initially assessed in a binary fashion with F1 scores, and the degree of text match among positive cases was evaluated using exact match accuracy and Jaccard Index (JI). Main Outcome Measures: F1 score, exact match accuracy, JI. Results: Gwet AC1 for intergrader agreement was 0.799, 0.888, 0.985, and 0.988 for CTM, {Delta}TM, COM, and {Delta}OM, respectively. F1 scores for CTM were 0.985, 0.971, 0.978, 0.968, and 0.970 for Claude, Deepseek, GPT, Grok, and Qwen, respectively; for {Delta}TM: 0.905, 0.826, 0.897, 0.842, 0.855, respectively; for COM: 0.923, 0.887, 0.899, 0.906, 0.894, respectively; for {Delta}OM: 0.958, 0.815, 0.937, 0.835, 0.940, respectively. Among positive cases, range of exact match accuracies for CTM (N=1354) was 0.730- 0.882 and range of JIs was 0.809-0.918. For {Delta}TM (N=404), exact match accuracy range was 0.619-0.780 and JI range was 0.668-0.827. For COM (N=47), exact match accuracy range was 0.766-0.872 and JI range was 0.765-0.870. For {Delta}OM (N=25), exact match accuracy range was 0.583-0.920 and JI range was 0.583-0.922. Conclusions: The GLLaucoMed pipeline demonstrated high performance in extracting and standardizing medication data from unstructured clinical notes, including both current medications and proposed changes. Claude and GPT exhibited the strongest performance.

Background/Aims: Neovascular glaucoma (NVG) is a severe, secondary glaucoma. This study aimed to identify factors associated with vision, intraocular pressure (IOP), and ocular pain outcomes. Methods: The cohort included all patients diagnosed with NVG during 2008-2024 at Helsinki University Hospital, Finland. Linear mixed-effects models used pre-specified covariates, whereas machine learning was given the full longitudinal data with biomicroscopic findings as an exploratory approach. Results: 626 patients were analysed. Worse baseline vision and a closed angle were associated with worse follow-up vision. Treatments were associated with lower IOP and less pain rather than better vision. Age, sex and comorbidity were largely not associated with the outcomes. Glaucoma drainage devices showed the greatest initial IOP reduction (-10.2 mmHg, 95% confidence interval, CI -11.9 to -8.6 mmHg), followed by transscleral cyclophotocoagulation (TSCPC, -4.7 mmHg, 95% CI -5.8 to -3.7 mmHg) and peripheral retinal cryotherapy (-2.2 mmHg, 95% CI -3.1 to -1.4 mmHg). TSCPC and cryotherapy were also associated with reduced pain (odds ratio 0.51 and 0.46). Pan-retinal photocoagulation and anti-VEGF showed smaller IOP reductions, with a pain reduction for pan-retinal photocoagulation only. Both methods agreed, and machine learning added no novel clinical findings. Conclusions: Vision in this cohort was largely set by the state of the eye at diagnosis. IOP control and pain relief therefore remain realistic goals even when sight cannot be saved. Peripheral retinal cryotherapy stood out, linked to both lower IOP and less pain, seldom reported in NVG. These associations from a large, unselected cohort identify treatments worth comparing prospectively.

Objectives: Amblyopia is a pediatric visual disorder traditionally treated by patching the fellow eye, though many patients retain residual amblyopia post-treatment. Increasing evidence suggests that visual plasticity allows treat-ment beyond the classical therapeutic window. AMBER evaluated the efficacy of binocular serious games in virtual reality (VR) in residual amblyopia. Methods and Analysis: The monocentric, prospective, randomized, crossover trial (reported as case series) includ-ed 14 anisometropic, strabismic, or mixed residual amblyopia patients (6-35 years; 5 children, 9 adults). Participants underwent two 2-month intervention phases: optical correction (standard care) and standard care plus VR games (2.5 h/week), each with a 2-month follow-up. Best-corrected visual acuity (BCVA), stereoacuity, and reading speed were assessed (5 timepoints) using the Sloan and Landolt charts, the Titmus, TNO, Lang II, Asteroid, and Mnread tests. Compliance and adverse events (AE) were recorded. Results: VR training improved BCVA in 10 amblyopic eyes (Landolt and Sloan), with more pronounced effects in anisometropic patients. Six patients showed improved stereoacuity (Titmus; 4x mixed, 1x anisometropic, 1x stra-bismic amblyopia), persistent only in children (1x strabismic, 1x mixed amblyopia). Four improvements were ob-served with TNO (1x), Lang II (1x), Asteroid (0x), and MNread (1x). Despite positive trends, when comparing re-sults of individual patients, between both eyes, and with standard treatment, consistency of improvements cannot be conclusively demonstrated. One non-severe AE (dizziness) was reported. Conclusions: Following individual cases, VR training improved BCVA and stereoacuity, particularly in children and patients with high compliance. However, considering the cohort as a whole, consistency of effects has to be confirmed in larger groups. Thus, the methodologically sophisticated AMBER study revealed differences in VR treatment efficacy between amblyopia types, children/adults, endpoints and tests, offering precious data for the design of meaningful future studies. It shows that neurovisual plasticity gauged by VR-games offers safe, engaging treatment options for residual amblyopia.

PurposeOculocutaneous albinism type 1 (OCA1) is an inherited disorder caused by tyrosinase (TYR) gene mutations. Affected individuals experience visual impairment and severe photosensitivity from ocular hypomelanosis, with no current treatments. We evaluated the safety and efficacy of a TYR-encoding adeno-associated virus (AAV) vector in healthy rhesus macaques as a potential OCA1 treatment. MethodsA novel AAV2-based capsid (ATX002) was packaged with the human VMD2 promoter and TYR (hTYR) fused with mGreenLantern (mGL). Two adult rhesus macaques were injected with ATX002-hVMD2-hTYR-mGL subretinally (OD) and intravitreally (OS). Safety and efficacy were assessed via comprehensive ophthalmic examination, fundus photography, spectral-domain optical coherence tomography (SD-OCT), and full-field electroretinography at baseline and defined timepoints up to 12 weeks post-injection, followed by post-mortem immunohistochemistry (IHC). ResultsBoth subretinal doses induced localized hypermelanosis by 3 weeks post-injection, which persisted through the study endpoint and was accompanied by measurable thickening of the retinal pigment epithelium (RPE) on SD-OCT. Histological IHC confirmed successful RPE transduction via robust mGL fluorescence, corroborating in vivo findings by revealing localized RPE hyperplasia and transgene-expressing cells adjacent to regions of de novo hypermelanosis. Intravitreal delivery did not induce any changes to the RPE. Transient uveitis was observed but successfully managed with anti-inflammatory treatment. ConclusionsSubretinal AAV-TYR delivery is a safe and effective approach with the potential to induce RPE pigmentation. These findings support the use of AAV-TYR gene therapy for OCA1, demonstrating efficacy and a manageable safety profile in a large-animal model, and provide a critical bridge toward human clinical translation.

Light reaching the retina is a primary regulator of human circadian physiology, acting largely through melanopsin-expressing retinal ganglion cells with peak short-wavelength sensitivity. Delivering known, repeatable retinal doses outside the laboratory is difficult because conventional light sources leave viewing geometry, gaze, and ambient conditions uncontrolled. Consumer extended-reality (XR) glasses fix a bright binocular display in constant geometry relative to the eye, but their suitability as calibrated photic stimulators has not been established. Here we validate a commercial micro-OLED XR display (VITURE Luma Ultra) for controlled retinal photostimulation. A purpose-built host application renders exact 8-bit RGB stimuli while independently controlling hardware brightness and logging all intensity-determining state; spectral radiance was measured at the retinal position of a 3D-printed phantom head with an open-source miniature spectroradiometer, anchored to absolute units by a luminance transfer calibration. The blue primary peaks at 461 nm (FWHM 43 nm), is spectrally invariant across a >10-fold intensity range, and at maximum output delivers an estimated 299 lx melanopic equivalent daylight illuminance, above consensus daytime recommendations, while remaining roughly two orders of magnitude below photobiological safety limits. The red primary is visually effective with minimal melanopic drive (melanopic DER 0.10), enabling spectrally shifted evening stimulation. Unlike the immersive virtual-reality headsets previously used for calibrated light delivery, the see-through form factor preserves the wearer's view of the surroundings--relevant for clinical monitoring in supervised settings such as the intensive care unit. These results show that consumer XR glasses can serve as a dose-calibrated platform for wearable photostimulation using an open-source measurement chain, and provide groundwork for application-layer dose-response studies.

Glaucoma is a leading cause of irreversible blindness worldwide. Ophthalmologists diagnose glaucoma through a structured reasoning process by sequentially evaluating optic nerve head characteristics before reaching a final diagnosis, whereas existing AI systems typically perform direct image classification without providing clinically meaningful reasoning. We present the first clinically annotated fundus reasoning dataset, comprising 1,077 fundus photographs paired with expert-authored six-step diagnostic reports. Building on this dataset, we develop a reasoning-driven vision-language framework that explicitly models the ophthalmologist's diagnostic workflow by generating structured clinical reasoning prior to diagnosis. The generated reports are clinically validated, achieving the best performance across all evaluated clinical findings, including a cup-to-disc ratio mean absolute error of 0.070, an ISNT Kendall distance of 1.73, and the highest semantic agreement with expert reports (BERTScore-F1 = 0.874). The resulting framework also improves glaucoma diagnosis, achieving a balanced accuracy of $94.7\%$ and precision of $94.8\%$, demonstrating that explicitly modeling expert clinical reasoning simultaneously improves interpretability and diagnostic performance. Code and data are available at \url{https://glaucoma-cot.github.io/}.

Purpose The direct effect of iStent inject on intraocular pressure (IOP) in patients with glaucoma is difficult to quantify in pragmatic trials where rates of post-surgical IOP-lowering therapy differ between intervention groups. We aimed to quantify the causal effect of iStent inject on unmedicated IOP at 12- and 24-months post-surgery. Methods Adults with mild-to-moderate glaucoma were 1:1 randomised to receive cataract surgery with iStent inject or cataract surgery alone at an Australian hospital (2017-2020, NCT03106181). IOP-lowering medications were prescribed as per clinician discretion. An exploratory analysis was used to estimate the controlled direct effect of iStent inject on IOP, analogous to the effect expected if all participants had undergone medication washout prior to assessment. Results Ninety-five eyes from 80 people were included (67.4% male, mean age 73.0 years, mean baseline IOP 17.1 mmHg). IOP-lowering medication was required for 53% of eyes in each group at 12 months (n=76); at 24 months (n=86) it was required for 43% and 64% in the active and control groups, respectively. Mean IOP was similar between intervention groups at each outcome visit. The controlled direct effect favoured the iStent inject group at 12 months (-2.1-mmHg difference, 95% CI -4.0,-0.3) but was attenuated at 24 months (-0.5 mmHg-difference, 95% CI -2.6,1.6). Conclusion Although the iStent inject was estimated to have an effect on lowering unmedicated IOP at 12 months, this effect had largely disappeared by 24 months. Medication washout is recommended when safe and practical in future trials to estimate these direct effects with more certainty.

Purpose: To investigate corneal subbasal nerve plexus alterations using in vivo corneal confocal microscopy (IVCM) in patients with Autosomal Recessive Spastic Ataxia of Charlevoix-Saguenay (ARSACS) and Spastic Paraplegia Type 7 (SPG7). Methods: This cross-sectional pilot study included eight ARSACS patients, five SPG7 patients, and twenty age- and sex-matched healthy controls. All participants underwent neurological and ophthalmological examination followed by central corneal imaging using IVCM. Quantitative corneal nerve parameters were analyzed with automated software, and correlations with clinical severity scales were assessed. Results: The mean age was 34.2 +/- 3.4 years in controls, 34.5 +/- 0.7 years in the ARSACS group, and 38.2 +/- 3.5 years in the SPG7 group. Corneal nerve branch density (CNBD) and corneal nerve total branch density (CTBD) were significantly lower in ARSACS and SPG7 patients compared with healthy controls. CNFD, CNFL, CNFA, CNFW, and CNFrD were lower in ARSACS and SPG7 patients compared with healthy controls; however, these differences did not reach statistical significance. No statistically significant differences in IVCM parameters were detected between ARSACS and SPG7 patients. Spearman correlation analysis did not show significant correlations between corneal nerve parameters and FARS, SARA, ADL scores, or disease duration. Conclusion: IVCM revealed reduced corneal nerve branching parameters in patients with ARSACS and SPG7. These findings indicate involvement of the corneal subbasal nerve plexus and support the potential role of corneal confocal microscopy as a non-invasive ocular imaging modality for evaluating peripheral neural alterations in hereditary spastic ataxias.

ObjectiveTo systematically review automated nerve morphometry tools and independently benchmark their performance on independent optic nerve datasets. DesignSystematic review and comparative benchmarking study. ControlsBenchmarking was performed using paraphenylenediamine-stained mouse (n = 85) and rat (n = 44) optic nerve images with manually annotated axon counts as ground truth. MethodsPublished studies describing automated or semi-automated neural tissue morphometry tools were identified through systematic searches of PubMed, Embase, and Scopus through January 2026 following PRISMA guidelines. Data extraction covered 70 fields across tool capabilities, imaging modality, species, automation level, and validation approach. Eighteen eligible tools (8 deep learning [DL], 10 classical computer vision [CV]) were benchmarked on both mouse and rat independent datasets. Main Outcome MeasuresPerformance was assessed by mean absolute percentage error (MAPE), Pearson correlation, and median predicted-to-ground-truth ratio. Tools were ranked per image and compared using Friedman tests with Nemenyi post-hoc analysis. ResultsSeventy-one studies met inclusion criteria, spanning from 1999 to 2026. Deep learning methods represented 38% (27/71) of studies, increasing from 0% before 2017 to over 55% of publications after 2020. Axon counting was the most common output (73%, 52/71), while only 35% (25/71) reported g-ratio. Among benchmarked tools, Marina (CV, 2010) achieved the lowest average MAPE (32.9%). The top five tools (MAPE ranging from 32.9 to 44.8%) included both CV and DL methods and were statistically indistinguishable by Friedman-Nemenyi analysis (p > 0.05). Performance varied substantially across datasets: AxonJ (CV) achieved the second best MAPE on rat images (27.7%) but the worst on mouse images (438.6%). ConclusionsNo single tool demonstrated consistently superior performance across both datasets. Classical and deep learning approaches achieved comparable accuracy for axon counting. Tool selection should be guided by target species, tissue preparation protocol, and desired morphometric outputs. This systematic review and independent benchmarking study provide an evidence base for tool selection in optic nerve research.

Background/Objective: Diabetic macular oedema (DMO) is a leading cause for visual impairment primarily managed with intravitreal anti-VEGF agents such as ranibizumab (RBZ). Levosulpiride (LSP), a prokinetic medication, was recently repositioned as a safe oral treatment for naive DMO. Here, we investigated the adjuvant effect of oral LSP in combination with intravitreal RBZ injections for treating persistent DMO. Subjects/Methods: Double-blinded, dual-centre, phase 2 trial in patients with centre-involving DMO randomly assigned to be orally treated with placebo (15 patients, 18 eyes) or LSP (18 patients, 19 eyes) along with 3 successive (4 weeks apart) RBZ intravitreal injections and a 24-week follow-up. Results: Baseline best-corrected visual acuity (BCVA) improved (p[&le;]0.04) at week 12 in both RBZ+placebo and RBZ+LSP, but improvement was maintained (p=0.009) at week 24 only in RBZ+LSP. In agreement, longitudinal changes from baseline in BCVA from weeks 12 to 24 defined superior (p=0.02) visual gains measured by the Area Under the Curve (AUC) in RBZ+LSP vs. RBZ+placebo. The baseline value of mean central foveal thickness (CFT) decreased (p[&le;]0.002) in both groups at week 12 and CFT reduction was significant (p=0.006) at week 24 only in RBZ+LSP. Also, longitudinal changes from baseline in CFT resulted in a higher AUC reduction (p[&le;]0.04) at weeks 4 to12 in RBZ+LSP vs. RBZ+placebo. No significant adverse side effects were detected. Conclusions: Adjunctive LSP showed functional and anatomical benefits over the first-line therapy with RBZ. Adjuvant properties may involve the LSP-induced intraocular upregulation and downregulation of vasoinhibin and VEGF, respectively. Larger clinical trials are warranted.

Purpose: To investigate how artificial intelligence (AI) systems detect referrable diabetic retinopathy (DR) from retinal photographs by analysing heatmap patterns and determining their overlap with DR features. Methods: Fifty-four AI systems were developed using 27 backbone architectures, with each implemented as both binary-referable and multi-class grading models based on the International Clinical Diabetic Retinopathy (ICDR) grading scale. Models were trained on images from DDR, BRSET and Kaggle datasets. After training, each model analysed 749 images with DR feature annotations, with Grad-CAM heatmaps generated and compared to pixel-level annotations of microaneurysms, haemorrhages, exudates, cotton wool spots, venous beading, intraretinal microvascular abnormalities and neovascularisation. Results: All models achieved acceptable predictive performance (AUROC >0.8 for most architectures). Heatmap analysis revealed consistent attention to the macular region with relative neglect of the optic disc. Exudates and cotton wool spots were highlighted most frequently by the heatmaps, with venous beading and neovascularisation at the disc showing poor overall coverage for binary referable classifiers. Models grading per the ICDR scale demonstrated high coverage for all features. Substantial variability was observed between architectures, suggesting different feature detection capabilities. Interestingly, the heatmap analysis indicated that the models were using different logic to the ICDR grading scale definitions. Conclusion: AI models do not uniformly rely on all DR features when detecting referable DR, limiting their predictive performance in unusual presentations. Heatmap aggregation analysis provides a scalable method for analysing model behaviour, allowing strengths and weaknesses to be identified. These findings may help improve clinician's trust and acceptance of AI.