BackgroundAlthough ocular manifestations of neurofibromatosis type 1 (NF1) have been described extensively in Western cohorts, systematic data in Asian populations remain limited, and age-related phenotypic variations have not been comprehensively characterized. This study investigated the prevalence and multidimensional characteristics of NF1-related ocular manifestations in Asian populations, using age-stratified analysis of uveal and retinal abnormalities.MethodsIn this cross-sectional study, 228 Chinese patients with NF1 underwent comprehensive ophthalmic evaluations. Examinations included slit-lamp biomicroscopy, ultra-widefield (UWF) fundus photography, near-infrared reflectance (NIR) imaging, and optical coherence tomography (OCT) to assess retinal and uveal abnormalities.ResultsA total of 228 patients with NF1 (46.1% male; median age 14 years) were enrolled. Lisch nodules were detected in 82.82% of patients, with highest prevalence during puberty and greater counts in older individuals (P < 0.001). Lisch nodules showed a tendency for inferior distribution and likely progressive darkening with age. Choroidal abnormalities were identified in 89.94% of patients, exceeding the prevalence of Lisch nodules (P < 0.001). Both the number and area of choroidal abnormalities correlated positively with age (P < 0.001), with predominant distribution at the posterior pole. Retinal vascular abnormalities (RVAs), retinal astrocytic hamartomas (RAHs), and iris mammillations were observed in 9.47%, 1.83%, and 8.37% of patients, respectively.ConclusionsIn Asian patients with NF1, older individuals exhibited a greater burden of uveal abnormalities. Lisch nodules demonstrated preliminary evidence of an age-dependent darkening phenotype. Choroidal abnormalities were more prevalent than Lisch nodules and consistently present across age groups, making them valuable diagnostic markers in paediatric patients. Compared with Western cohorts, Chinese paediatric patients with NF1 exhibited earlier onset of both Lisch nodules and choroidal abnormalities. These findings refine ethnicity-specific ocular profiles of NF1 and underscore the importance of timely ophthalmic surveillance, particularly in children.Supplementary InformationThe online version contains supplementary material available at 10.1186/s40662-026-00488-y.

BackgroundMounting evidence indicates metabolic dysregulation in diabetic corneal neuropathy (DCN). This study elucidates how the chromatin remodeler Brahma-related gene 1 (BRG1) orchestrates glycolytic reprogramming to drive neurodegeneration and epithelial repair defects in DCN.MethodsType 1 diabetic mice were established via streptozotocin (STZ) injection. Glycolysis was inhibited using 2-deoxy-D-glucose (2-DG) to assess its role in DCN pathogenesis. BRG1 expression was modulated by subconjunctival plasmid delivery (overexpression/knockdown). Pathway screening identified BRG1 downstream effectors, and phosphatidylinositol 3-kinase/protein kinase B (PI3K/AKT) inhibition (LY294002) confirmed regulatory hierarchy. Glycolytic flux was evaluated via Western blotting and immunofluorescence; corneal nerve integrity and epithelial healing were assessed by βIII-tubulin staining and sodium fluorescein assay.ResultsHyperglycemia upregulated BRG1 and glycolytic enzymes in diabetic corneal nerves. BRG1 overexpression exacerbated epithelial repair delay and neurodegeneration, while knockdown partially reversed damage. BRG1 overexpression activated PI3K/AKT transcription, and PI3K/AKT inhibition did not alter BRG1 levels but rescued BRG1-induced pathologies.ConclusionsGlycolytic reprogramming is a critical driver of DCN progression. BRG1 activates PI3K/AKT signaling to enhance glycolytic flux, thereby regulating DCN pathogenesis. Targeting this axis may offer novel therapeutic strategies.Supplementary InformationThe online version contains supplementary material available at 10.1186/s40662-026-00474-4.

PurposeTo assess the functional and ocular surface anti-inflammatory outcomes of epithelium-off accelerated corneal cross-linking (ACXL) in adolescents with progressive keratoconus associated with allergic ocular surface disease and dry eye disease (DED) characterized by elevated tear matrix metalloproteinase-9 (MMP-9) concentrations.MethodsProspective interventional case series of 30 eyes from 15 patients (mean age 16.41 ± 2.36 years; Krumeich stage II) undergoing epi-off ACXL. Outcomes at baseline and 1, 3, 6, and 12 months included corrected distance visual acuity (CDVA), maximum keratometry (Kmax), minimum corneal thickness (MCT), computerized non-invasive tear break-up time (cBUT), Ocular Surface Disease Index (OSDI), and tear MMP-9 (point-of-care test). In vivo qualitative confocal microscopy (IVCM) investigation provided supportive imaging. Paired t-tests were used and results reported with 95% confidence intervals (CI).ResultsCDVA improved to 0.09 logMAR at 12 months (≈ 0.81 decimal; 95% CI: 0.10–0.08 logMAR; P < 0.001). Kmax decreased from 55.00 to 53.75 D (95% CI: 53.55–53.95 D; Δ = − 1.25 D; P < 0.001), indicating ectasia stabilization. cBUT increased from 10.11 to 14.41 s (95% CI: 14.11–14.71; P < 0.01). OSDI decreased to 12.15 (95% CI: 11.65–12.65). Tear MMP-9 levels diminished from 64.79 to 16.15 ng/mL (P < 0.0001) and the proportion < 38.6 ng/mL reached 86.7% of the study cohort at 12 months. IVCM documented disappearance of inflammatory infiltrates. No postoperative persistent adverse events occurred.ConclusionsEpi-off ACXL stabilized ectasia, improving visual and ocular surface outcomes, markedly lowering tear MMP-9 levels. Although exploratory, these findings are consistent with a potential ocular surface anti-inflammatory and neuromodulatory role of ACXL, meriting validation in studies involving inflammatory DED beyond keratoconus.Supplementary InformationThe online version contains supplementary material available at 10.1186/s40662-025-00469-7.

BackgroundWhile parental astigmatism is a known risk factor for childhood astigmatism, the molecular genetic basis remains elusive. Previous genetic studies, largely confined to adult corneal and refractive astigmatism (CA and RA), did not address internal astigmatism (IA) and astigmatism vector components. We aimed to determine whether genes previously identified to have associations with corneal curvature (CR), CA, and RA in adults similarly occur for CR, CA, RA, IA, and astigmatism vector components (J0 and J45) in children.MethodsFourteen polymorphisms in nine loci were genotyped in 2167 Chinese children. Linear and logistic regression analyses were conducted to evaluate the association of the polymorphisms with CR, CA, RA, IA, and astigmatism vector components, which were determined by keratometry, cycloplegic refraction, or Fourier transformation.ResultsFMNL2 rs1579050 was associated with CA (additive: β = 0.158, P = 0.0028; dominant: β = 0.163, P = 0.0034), J0(CA) (additive: β = 0.081, P = 0.0031), and an increased risk of dichotomous RA (additive: OR = 1.609, P = 0.0028; dominant: OR = 1.671, P = 0.0020), whereas NHSL1 rs4896367 was associated with J0(IA) (recessive: β = 0.058, P = 0.0002) and a lower risk of dichotomous IA (recessive: OR = 0.577, P = 0.0007). PDGFRA rs2228230 was also associated with J0(IA) (dominant: β = −0.034, P = 0.0012). The predisposition to CA and RA increased with the risk alleles of FMNL2 rs1579050.ConclusionsThese findings reveal genetic contributions to childhood astigmatism and demonstrate that vector-based decomposition may facilitate more precise mapping of its genetic determinants.Supplementary InformationThe online version contains supplementary material available at 10.1186/s40662-025-00464-y.

BackgroundDue to the lack of quality-controlled quantitative data on meibomian gland (MG) morphology in children and adolescents, this study aims to establish a Children and Adolescents Meibomian Gland (CAMG) dataset.MethodsA total of 1114 quality-controlled upper eyelid infrared images were collected from 730 children and adolescent subjects using the Oculus Keratograph 5 M. Images underwent preprocessing and multi-stage expert quality control screening before segmentation. Morphological parameters including gland area, gland dropout ratio, gland length and width, number of glands, and total glands ratio were extracted using an AI model. The dataset, comprising images, annotations, and demographic information, is openly accessible on Figshare, with AI model codes available on GitHub to support research reproducibility and algorithm optimization.ResultsThe dataset includes 1114 high-resolution quality-controlled upper eyelid images from 730 subjects (mean age 11.80 ± 2.39 years; 46.77% male), accompanied by AI-assisted segmentation annotations and corresponding morphological measurements. The U-Net segmentation model achieved an accuracy of 97.49%, a Dice coefficient of 89.72%, and an intersection over union (IoU) of 81.67%. Quantitative analysis revealed that MG parameters remained relatively stable in adolescents compared to children. Females exhibited significantly wider and larger MGs than males. Similar sex-related differences were also observed in the central five MGs. Males exhibited a higher MG count compared to females.ConclusionsCAMG is a publicly available MG dataset for children and adolescents to support AI-based individualized clinical assessments. The dataset's transparent quality control processes establish a foundation for epidemiological research, promoting cross-institutional collaboration and AI-driven advancements in ophthalmology.Supplementary InformationThe online version contains supplementary material available at 10.1186/s40662-025-00460-2.

BackgroundThis retrospective study aimed to identify risk factors for subretinal fibrosis (SF) and evaluate the response to anti-vascular endothelial growth factor (anti-VEGF) therapy in patients with myopic choroidal neovascularization (mCNV), with a specific focus on the role of dilated choroidal vessels (DCVs) in disease progression.MethodsIn this retrospective study, patients with high myopia (spherical equivalent < −6.0 D, pathological myopia, Asian ethnicity) and active mCNV lesions, diagnosed between 2021 to 2023, were evaluated. The location of DCVs and mCNV was assessed, and macular thickness, submacular choroid thickness, best-corrected visual acuity, CNV area, and flow density were measured at baseline and during follow-up. The presence of posterior staphyloma was evaluated at baseline. SF around the mCNV was evaluated lesions during follow-up. The time to SF detection was recorded using survival analysis. Risk factors for SF were analyzed using Kaplan–Meier and multivariable Cox regression analyses.ResultsA total of 46 eyes from 46 patients were included, with a mean age of 54.17 ± 14.37 years, and a baseline spherical equivalent of 12.36 ± 3.21 D. The logarithm of the minimum angle of resolution for the mean visual acuity was 0.70 (0.40–1.30), and the mean macular thickness was 313.11 ± 63.57 μm at baseline. DCV was detected in 29 of the 46 eyes (63.0%), and the median time to detect SF was 43.41 [95% confidence interval (CI): 37.27–49.55] months. Multivariable Cox regression analysis identified submacular DCV [hazard ratio (HR): 14.93, 95% CI: 5.72–38.91, P < 0.001) and absence of posterior staphyloma (HR: 43.48, 95% CI: 12.15–156.32, P = 0.002) as independent predictors of SF. The presence of DCV under the fovea compared to the peripheral zone achieved a poorer therapeutic response and was prone to progress to SF after anti-VEGF therapy (P = 0.041).ConclusionsSubmacular DCV is associated with poor therapeutic response to anti-VEGF therapy and an increased risk of SF in patients with mCNV.

BackgroundGlaucoma causes permanent blindness. Current treatments have limited effectiveness, necessitating novel therapeutic strategies. We aimed to identify potential drug targets for glaucoma by integrating multi-trait and multi-omic analyses.MethodsWe sourced druggable gene expression and protein abundance summary-level data from quantitative trait loci studies, and genetic associations with glaucoma from a large-scale multi-trait analysis. We employed proteome and transcriptome Mendelian randomization (MR) and colocalisation to identify potential therapeutic targets, glaucoma endophenotype MR to explore the potential mechanisms of identified associations, and phenome-wide MR to investigate possible adverse effects of candidate targets.ResultsWe identified CPXM1 and FLT4 as tier 1; INSR as tier 2; and CPZ and PXDN as tier 3 druggable genes. Genetically predicted higher levels of CPXM1 [odds ratio (OR): 0.86, 95% confidence interval (CI): 0.81–0.91, PFDR < 0.001], FLT4 (OR: 0.74, 95% CI: 0.64 − 0.87, PFDR = 0.033), INSR (OR: 0.58, 95% CI: 0.43 − 0.78, PFDR = 0.042), and CPZ (OR: 0.55, 95% CI: 0.40 − 0.74, PFDR = 0.033) were associated with decreased glaucoma risk while those of PXDN (OR: 1.33, 95% CI: 1.15 − 1.54, PFDR = 0.033) with increased risk. The associations for CPXM1 (OR: 0.53, 95% CI: 0.39 − 0.73, P < 0.001) and FLT4 (OR: 0.86, 95% CI: 0.78 − 0.95, P = 0.005) were confirmed transcriptome-wide and colocalisation was confirmed for CPXM1 [posterior probability H4 (PPH4) = 0.940], FLT4 (PPH4 = 0.701), and INSR (PPH4 = 0.706). The protective effects of CPXM1 and CPZ may be attributed to intraocular pressure-lowering activities. The risk associated with PXDN is due to its involvement in glaucomatous neuropathy. No significant adverse effects were identified.ConclusionsThis study provides novel insights into glaucoma pathophysiology and promotes pharmaceutical target innovation.

BackgroundDespite extensive research on keratoconus (KC) detection with traditional machine learning models, stacking ensemble learning approaches remain underexplored. This paper presents a stacking ensemble learning method to enhance automated KC screening.MethodsThis study utilizes a clinical dataset containing detailed corneal data from 2491 cases classified as non-KC (NKC), subclinical KC (SCKC) and clinical KC (CKC). Each cornea is represented by 79 features extracted from Pentacam imaging. Following extensive pre-processing, key corneal features that are strongly correlated with the target diagnosis are identified. These features are the keratometry of the steepest anterior point, surface variance index, vertical asymmetry index, height decentration index, and height asymmetry index. A novel stacking ensemble model is developed using the selected features to improve corneal classification into NKC, SCKC, and CKC by integrating top tree-based classifiers (random forest, gradient boosting, decision trees) with a support vector machine meta-classifier.ResultsThe pre-processing and feature selection techniques reduced the model's parameters to just 6.33% of the original dataset, improving classification performance, and cutting over 85% of the training time. The performance of the developed model was validated and tested on unseen data. Experimental results showed that the model outperforms existing studies, achieving 99.72% accuracy, precision, sensitivity, F1, and F2 scores, with a Matthews correlation coefficient of 0.995. It accurately classified all NKC and CKC cases, with just one misclassification involving an SCKC case. The model also demonstrated consistent performance on 100 additional unseen test cases, underscoring its generalizability and robustness in KC screening.ConclusionsBy combining the strengths of diverse base models and key Pentacam indices, the stacking ensemble approach ensures reliable, accurate KC screening, providing clinicians with an automated tool for early detection and better patient management.

BackgroundTransepithelial photorefractive keratectomy (transPRK) can be safely and predictably performed to correct low-to-high astigmatism. This study explored the effects of fixation stability, corneal density (CD), ocular residual astigmatism (ORA), and the surgically-induced change in the epithelial thickness (ΔET) on the efficacy of astigmatism correction by transPRK.MethodsEighty-three consecutive patients who underwent transPRK to correct myopia and myopic astigmatism were divided into two groups according to refractive astigmatism [high refractive astigmatism (RA) group: ≥ 2.0 D, n = 31; low RA group: < 2.0 D, n = 52]. Fixation stability was evaluated by measuring the lateral movement of the pupil center on the eye tracker images. The CD was measured using a Pentacam Scheimpflug imaging system, epithelial thickness mapping was performed using optical coherence tomography, and the ORA was determined using vector analysis. Multiple linear regression analyses were performed to identify factors associated with the correction index (CI) and angle of error (AOE).ResultsAt 6 months postoperatively, the RA was higher in the high RA group (− 0.66 ± 0.44 D) than in the low RA group (− 0.29 ± 0.29 D, P < 0.001), whereas no significant differences were found in CI or AOE between two groups. Multiple linear regression analyses showed that for the low RA group, preoperative anterior CD of the central 2 mm (CD0-2A, β = − 0.482, P = 0.011) and ΔET (β = 0.295, P = 0.041), were associated with CI, whereas the vector length of the pupil center shift (PCVL, β = − 0.404, P = 0.005) and ΔET (β = − 0.293, P = 0.036) were associated with AOE. For the high RA group, ΔET (β = 0.519, P = 0.038) was associated with CI, whereas static cyclotorsion (β = − 0.493, P = 0.040) was associated with AOE. No significant associations were found between ORA and CI or AOE.ConclusionsPostoperative changes in epithelial thickness were associated with the efficacy of transPRK in both the low and high RA groups, whereas the pupil center shift and anterior CD were associated with the efficacy of transPRK in the low RA group.

Vitreoretinal lymphoma (VRL) is often a diffuse large B-cell lymphoma in nature, and patients may have or eventually develop central nervous system lymphoma, which frequently leads to a poor prognosis. Currently, there are no international or domestic clinical guidelines specifically for the diagnosis and treatment of VRL, and no standardized diagnostic procedures or treatment evaluation systems for this disease. VRL is clinically characterized by prominent vitreous opacities, multiple lesions beneath the retinal pigment epithelium or subretinal, and intraretinal infiltration, making it one of the most common masquerade syndromes in ophthalmology. To promote early diagnosis and standardized treatment of VRL, the Ocular Immunology Group of the Chinese Medical Association Ophthalmology Branch has developed "Clinical Guidelines for the Diagnosis and Treatment of Vitreoretinal Lymphoma in Chinese Patients (2024)", based on extensive references to diagnosis, treatment experiences and relevant clinical recommendations. The working group systematically reviewed and comprehensively summarized the latest research evidence from both domestic and international sources. Using the Oxford Evidence Level System, we assessed the quality of evidence and the strength of recommendations. This guideline provides crucial academic references and clinical practice guidance for the diagnosis and treatment of VRL patients. This guideline, including VRL diagnostic methods, processes, and treatment recommendations is suitable for clinical practice in China and is intended to assist ophthalmologists in clinical diagnosis and treatment of VRL.