Astrocyte-mediated neuroinflammation has recently been implicated as a key contributor to neurodegeneration following retinal ischemia–reperfusion (IR) injury. However, the role of miR‑16‑5p in this process remains unclear. This study aimed to investigate the function and mechanism of miR‑16‑5p. TargetScan was used to predict miR-16-5p targets, which were validated by RNA pull-down. miR‑16‑5p expression was assessed by RT‑qPCR in IR retinas and in astrocytes after oxygen–glucose deprivation/reoxygenation (OGD/R). Astrocyte activation, inflammatory cytokine, and Wip1/nuclear factor kappa B (NF‑κB) signaling were examined following miR-16-5p modulation with mimics or inhibitors in vitro and in vivo. Retinal ganglion cell (RGC) apoptosis, retinal function, and morphology were evaluated. miR‑16‑5p was found to potentially target wild-type p53-induced phosphatase 1 (Wip1) and decreased Wip1 expression. In IR-injured mouse retinas and OGD/R-treated astrocytes, miR‑16‑5p expression was significantly downregulated. This decrease was accompanied by astrocyte activation, increased TNF-α and IL-1β levels, and upregulation of Wip1 and phosphorylated NF-κB p65 (p-p65). These retinal changes indicated retinal injury, characterized by increased TUNEL-positive RGCs, elevated cleaved caspase-3 levels, retinal thinning, and reduced electroretinography (ERG) amplitudes. Treatment with miR-16-5p mimics ameliorated these molecular, cellular, structural, and functional alterations, whereas miR‑16‑5p inhibitors exacerbated them. Collectively, miR-16-5p may protect RGCs from IR-induced apoptosis by suppressing astrocyte-mediated inflammation via the Wip1/NF-κB signaling axis.

Retinal vessel segmentation is an essential step in the early diagnosis of ophthalmic diseases. However, due to the complexity of the vascular tree structure, the imbalance of vessels at different scales, the loss of information in low-contrast regions, and noise interference, achieving accurate segmentation remains a challenge. To this end, this paper proposes a Dual-Layer Semantic Fusion Network for Retinal Vessel Segmentation (DSFN-Net), which improves segmentation accuracy and the connectivity of small vessels through multi-level feature enhancement and a cross-stage attention mechanism. Specifically, we propose an enhanced convolution dual-layer semantic fusion network to strengthen the extraction of both local and global vessel features; optimize the encoder structure by introducing hierarchical feature fusion to alleviate the imbalance in vessel information extraction at different scales; design a cross-stage fusion attention mechanism to enhance vessel extraction in low-contrast regions; and employ a multi-level semantic fusion module to enhance the complementarity of features across different layers, thereby improving vessel connectivity and microvessel segmentation performance. Experimental results show that on the DRIVE, CHASE_DB1, and STARE datasets, DSFN-Net achieves F1-scores of 83.2%, 82.06%, and 84.78%, respectively, and ACC values of 97.11%, 97.64%, and 97.66%, while demonstrating strong generalization capability in cross-validation. Source code is available at https://github.com/Lrsm-sudo/DSFN_et.

PURPOSE:Diabetic retinopathy (DR) is a leading vision-threatening disease and a common complication of diabetes. However, the understanding of how metabolic dysfunction correlates with ocular inflammation in DR remains poorly understood. Therefore, in this study, we investigated the association between these two factors in DR. METHODS:Using untargeted metabolomics analysis and mass spectrometry detection, the plasma metabolic characteristics of 27 DR patients (15 with nonproliferative diabetic retinopathy [NPDR] and 12 with proliferative diabetic retinopathy [PDR]) and 8 healthy controls (HCs) were examined. Additionally, the concentrations of inflammatory factors and vascular endothelial growth factor (VEGF) in their aqueous humor were measured. Further analysis was conducted to evaluate the correlation between metabolic disorders and abnormalities in inflammatory factors and VEGF in the aqueous humor. RESULTS:Metabolomic analysis revealed significant metabolic dysregulation in DR, with PDR exhibiting particularly pronounced alterations compared to NPDR and HC. ROC analysis identified creatinine, glutamine, indoleacetic acid, and methionine as reliable biomarkers for distinguishing PDR. Elevated cytokines, such as IL-6, IL-8, and VEGF, were significantly correlated with specific metabolism pathways. CONCLUSION:Our study highlights potential biomarkers for DR stratification and provide novel insights into the link between metabolic dysfunction and ocular inflammation/ischemia in PDR.

Semi-supervised methods aim to alleviate the high cost of annotating medical images by incorporating unlabeled data into the training set. Recently, various consistency regularization methods based on the mean-teacher model have emerged. However, their performance is limited by the small number and poor quality of confident pixels in the pseudo-labels. Based on experimental observations, we propose a new argument: the performance gains of the model do not proportionally translate into improvements in pseudo-label quality, mainly due to constraints in pixel diversity representation and model expressiveness. Therefore, we propose a novel semi-supervised framework, DOC-MLE, which consists of two key components: a dynamic orthogonal constraint (DyOrCon) method and one multi-level election (MLElect) strategy. Specifically, DyOrCon imposes orthogonal constraints on multiple intermediate projection heads to enhance pixel diversity and fully exploit the model’s potential representation capacity. MLElect is designed considering both unsupervised pixel-level and supervised feature-level strategies, to generate reliable pseudo-labels. Moreover, to generate more robust prototype representations, this paper proposes new threshold filtering, edge erosion, and dynamic convolution strategies to address errors associated with low-confidence, high-confidence, and local morphological constraints. Extensive experiments on coronary angiography, polyp dataset, and retinal fundus images have proven the effectiveness of the proposed method.

PURPOSE. To reveal the role of receptor-interacting protein kinase 3 (RIPK3) in regulating macrophage inflammation and corneal neovascularization (CoNV) induced by alkali burn. METHODS. A corneal alkali burn (AB) model was established in C57BL/6J (wildtype) and RIPK3fl/flCx3cr1+/cre (RIPK3-/-, RIPK3 knockout [KO]) mice using sodium hydroxide. Anterior segment optical coherence tomography and hematoxylin and eosin staining were used to evaluate the impact of RIPK3 on corneal edema and morphology. CoNV was detected by slit-lamp microscopy and whole-mount immunofluorescence staining of cornea. Corneal macrophage and necroptotic cell death was analyzed through immunofluorescence staining and propidium iodide (PI) staining. Activation of the necroptosis pathway was examined after corneal AB by western blot. RESULTS. Necroptosis-related proteins RIPK1, RIPK3, and mixed lineage kinase domainlike (MLKL) were upregulated and activated following corneal AB. Among these, RIPK3 demonstrated the most pronounced increase. Notably, the elevated level of RIPK3 was prominently colocalized with the infiltrating F4/80+ macrophages. RIPK3 KO significantly alleviated corneal edema and morphology defects. Additionally, as the corneal morphological defects progressed, macrophages became activated, and CoNV and lymphangiogenesis (LyG) were enhanced. RIPK3 KO markedly reduced AB-induced macrophage accumulation, as well as CoNV and LyG. RIPK3 KO mice also showed a meaningful decrease in PI+ necroptotic cells. Mechanistically, AB-induced necroptosis stimulated the expression of MLKL and fibroblast growth factor 2 (FGF2), whereas RIPK3 deficiency decreased their expression. CONCLUSIONS. This study revealed that RIPK3-mediated necroptosis drives macrophage inflammation and CoNV. Targeting RIPK3 could effectively suppress these responses by inhibiting the MLKL/FGF2 pathway, making it a promising therapeutic strategy for corneal AB.

PURPOSE To investigate central and peripheral retinal and choroidal alterations using ultra-widefield swept-source optical coherence tomography angiography (UWF-SS-OCTA) in patients with obstructive sleep apnea syndrome (OSAS) stratified by disease severity. METHODS The study included patients with mild-moderate OSAS (n = 23 eyes), severe OSAS (n = 22 eyes), and healthy controls (n = 26 eyes). UWF- SS-OCTA scans centered on the fovea were used to divide the macula 8 concentric rings spanning from the central macula to the peripheral retina (0–21 mm). Vessel density (VD) in the superficial capillary plexus (SCP) and deep capillary plexus (DCP), and choroidal thickness (CT) were analyzed across all rings. Correlations with apnea-hypopnea index (AHI) and lowest oxygen saturation (LSaO₂) were also assessed. RESULTS SCP vessel density was significantly reduced only in the far periphery (ring 8) across all OSAS groups. DCP impairment was more extensive, with severe OSAS patients showing significantly decreased vessel density in peripheral rings 5–8 versus controls (all p < 0.05), particularly in the superior and inferior quadrants (all p < 0.05). Choroidal thinning in severe OSAS followed a similar pattern, being significantly reduced in rings 4–8 (all p < 0.05), particularly in the superior, inferior, and nasal quadrants (all p < 0.05). DCP vessel density and choroidal thickness in peripheral regions correlated negatively with AHI and positively with LSaO₂. CONCLUSIONS For patients with OSAS, damage to the deep capillary plexus and choroid predominantly affects the peripheral retina and is significantly associated with the severity of hypoxia.

BackgroundIn vivo confocal microscopy (IVCM) is a crucial imaging modality for assessing corneal diseases, yet distinguishing pathological features from normal variations remains challenging due to the complex multi-layered corneal structure. Existing anomaly detection methods often struggle to generalize across diverse disease manifestations. To address these limitations, we propose a Transformer-based unsupervised anomaly detection method for IVCM images, capable of identifying corneal abnormalities without prior knowledge of specific disease features.MethodsOur method consists of three submodules: an EfficientNet network, a Multi-Scale Feature Fusion Network, and a Transformer Network. A total of 7,063 IVCM images (95 eyes) were included for analysis. The model was trained exclusively on normal IVCM images to capture and differentiate structural variations across four distinct corneal layers: epithelium, sub-basal nerve plexus, stroma, and endothelium. During inference, anomaly scores were computed to distinguish pathological from normal images. The model’s performance was evaluated on both internal and external datasets, and comparative analyses were conducted against existing anomaly detection methods, including generative adversarial networks (AnoGAN), generate to detect anomaly model (G2D), and discriminatively trained reconstruction anomaly embedding model (DRAEM). Additionally, explainable anomaly maps were generated to enhance the interpretability of model decisions.ResultsThe proposed method achieved an the areas under the receiver operating characteristic curve of 0.933 on internal validation and 0.917 on an external test dataset, outperforming AnoGAN, G2D, and DRAEM in both accuracy and generalizability. The model effectively distinguished normal and pathological images, demonstrating statistically significant differences in anomaly scores (p < 0.001). Furthermore, visualization results indicated that the detected anomalous regions corresponded to morphological deviations, highlighting potential imaging biomarkers for corneal diseases.ConclusionThis study presents an efficient and interpretable unsupervised anomaly detection model for IVCM images, effectively identifying corneal abnormalities without requiring labeled pathological samples. The proposed method enhances screening efficiency, reduces annotation costs, and holds great potential for scalable intelligent diagnosis of corneal diseases.

The assessment of microcirculation in patients with stable angina holds significant value for monitoring disease progression and managing adverse cardiovascular events. This study systematically investigates the associations between central retinal arteriolar equivalent (CRAE) and central retinal venular equivalent (CRVE), with multidimensional systemic indicators in this population, to evaluate the potential of retinal vascular parameters as indicators of systemic microcirculatory status, thereby laying the groundwork for future noninvasive monitoring approaches. A cross-sectional study design was adopted, enrolling 528 patients diagnosed with stable angina between June 2022 and October 2023 (mean age: 62.13 years, 71.4% male). Retinal vessel calibers were quantified from fundus photographs using an artificial intelligence (AI)-based system that calculates CRAE and CRVE from the six largest vessels within standardized peripapillary zones. The system demonstrated excellent agreement with manual measurements (intraclass correlation coefficients: 0.82-0.95). Multivariable linear regression analyses, adjusted for key demographic and clinical confounders (including age, gender, body mass index [BMI], blood pressure, lipid profiles, and smoking status), were conducted to examine independent associations with systemic indicators. Significant correlations were found between CRAE and BMI (B = -0.071, p = 0.016), systolic blood pressure (B = -0.080, p = 0.026), D-dimer (B = -0.056, p = 0.047), diabetes mellitus (B =-0.055, p = 0.048), and CRVE (B = 0.761, p < 0.001). CRVE was significantly associated with age (B =-0.077, p = 0.011) and BMI (B = 0.077, p = 0.009). In conclusion, retinal vessels reflect the impact of systemic pathological stressors and support the potential of retinal analysis as a noninvasive approach to assess microcirculatory characteristics in this population.

OBJECTIVE:Our study aimed to assess the association between heavy metal exposure and Diabetic Retinopathy-Related Homeostatic Dysregulation (DRHD) values, a novel prognostic marker for diabetes and its complications, and to examine the mediating role of DRHD in the relationship between heavy metal exposure and diabetic mortality. METHOD:We used data from National Health and Nutrition Examination Survey (NHANES) 2003-2016, focusing on type 2 diabetes (T2D) patients. Urinary concentrations of 12 metals were measured, and DRHD values were calculated from 14 biomarkers of patients with T2D. Multivariate linear regression, weighted quantile sum regression, and quantile g-computation were applied. Principal component analysis (PCA) identified exposure patterns, and mediation analysis evaluated the role of DRHD in mortality risk. RESULTS:Among 1590 diabetic patients, cadmium (β: 0.104), antimony (β: 0.191), and tungsten (β: 0.136) significantly increased DRHD values, while cesium was negatively associated (β: -0.145). Mixed metal exposures (β: 0.123) had notable contributions from tungsten (0.279), cadmium (0.213), and antimony (0.137). PCA showed positive associations for molybdenum, tungsten, uranium (β: 0.063) and cadmium, lead, antimony (β: 0.098), while barium, cobalt, cesium, thallium (β: -0.050) were negatively associated. The mediation proportion of DRHD value for overall mortality and diabetes-cause mortality was 27.5 % and 10.8 % respectively. Meanwhile, DRHD value significantly mediated the mix-heavy metal exposures and diabetic-relative survival time. CONCLUSIONS:Our findings highlight the relationship between heavy metal exposures and DRHD values and demonstrate the mediating role of DRHD in mortality risk and survival time for diabetic patients.

The Omi/HtrA2 inhibitor 5-[5-(2-nitrophenyl) furfuryliodine]-1,3-diphenyl-2-thiobarbituric acid (Ucf-101) has shown neuroprotective effects in the central nervous system. However, whether Ucf-101 can protect retinal ganglion cells (RGCs) after retinal ischemia/reperfusion (IR) has not been investigated. We aimed to investigate the effects of Ucf-101 on RGCs apoptosis and inflammation after IR-induced retinal injury in mice. We injected Ucf-101 into the mouse vitreous body immediately after IR injury. After 7 days, hematoxylin and eosin staining was conducted to assess retinal tissue damage. Next, retrograde labeling with FluoroGold, counting of RGCs and TUNEL staining were conducted to evaluate apoptosis. Immunohistochemistry, immunofluorescence staining, and western blotting were conducted to analyze protein levels. IR injury-induced retinal tissue damage could be prevented by Ucf-101 treatment. The number of TUNEL-positive RGCs was reduced by Ucf-101 treatment in mice with IR injury. Ucf-101 treatment inhibited the upregulation of Bax, cleaved caspase-3 and cleaved caspase-9 and activated the JNK/ERK/P38 signaling pathway. Furthermore, Ucf-101 treatment inhibited the upregulation of glial fibrillary acidic protein (GFAP), vimentin, Iba1 and CD68 in mice with IR injury. Ucf-101 prevents retinal tissue damage, improves the survival of RGCs, and suppresses microglial overactivation after IR injury. Ucf-101 might be a potential target to prevent RGCs apoptosis and inflammation in neurodegenerative eye diseases.

Aims: Retinal ischemia/reperfusion (I/R) injury is implicated in the etiology of various ocular disorders. Prior research has demonstrated that bone marrow tyrosine kinase on chromosome X (BMX) contributes to the advancement of ischemic disease and inflammatory reactions. Consequently, the current investigation aims to evaluate BMX's impact on retinal I/R injury and clarify its implied mechanism of action. Main methods: This study utilized male and female systemic BMX knockout (BMX−/−) mice to conduct experiments. The utilization of Western blot assay and immunofluorescence labeling techniques was employed to investigate variations in the expression of protein and tissue localization. Histomorphological changes were observed through H&E staining and SD-OCT examination. Visual function changes were assessed through electrophysiological experiments. Furthermore, apoptosis in the retina was identified using the TUNEL assay, as well as the ELISA technique, which has been utilized to determine the inflammatory factors level. Key findings: Our investigation results revealed that the knockdown of BMX did not yield a significant effect on mouse retina. In mice, BMX knockdown mitigated the negative impact of I/R injury on retinal tissue structure and visual function. BMX knockdown effectively reduced apoptosis, suppressed inflammatory responses, and decreased inflammatory factors subsequent to I/R injury. The outcomes of the current investigation revealed that BMX knockdown partially protected the retina through downregulating phosphorylation of AKT/ERK/STAT3 pathway. Significance: Our investigation showed that BMX−/− reduces AKT, ERK, and STAT3 phosphorylation, reducing apoptosis and inflammation. Thus, this strategy protected the retina from structural and functional damage after I/R injury.

Fibroblast growth factor (FGF) is involved in the progression of glioma, a most common type of brain tumor, and breast tumors. In this study, we aim to evaluate the effects of the inhibitor PP2 on cell proliferation and migration in glioma and breast tumor cells, and to characterize the molecular mechanisms involved in these processes. The inhibitory effect of PP2 on the tumorigenic potential of C6 glioma and MDA-MB-231 cells was examined by proliferation, migration, and invasion assays, and apoptotic analysis. The molecular mechanism behind the anti-glioma activity of PP2 was investigated by immunoblotting, immunoprecipitation, phosphoprotein assay, cellular thermal shift assay (CETSA), and molecular docking modeling. PP2 suppressed the proliferation and migration of C6 glioma and MDA-MB-231 cells via FGF2. Moreover, PP2 directly blocked the enzyme activity of FGF receptor 1 (FGFR1) and Src, subsequently affecting the nuclear factor-κB and activator protein-1 signaling pathways. CETSA analysis and the docking model indicated that the TK1 domains (Val 492 ad Glu 486) of FGFR2 could be binding sites of PP2. Collectively, therefore, our findings suggest that PP2 mediates antitumor effects by targeting both FGFR1 and Src and may have applications as a therapeutic inhibitor for the treatment of glioma.

Aims To develop a metric termed the diabetic retinopathy-related homeostatic dysregulation (DRHD) value, and estimate its association with future risk of mortality in individuals with type 2 diabetes. Methods With the data of the NHANES, the biomarkers associated with DR were identified from 40 clinical parameters using LASSO regression. Subsequently, the DRHD value was constructed utilizing the Mahalanobis distance approach. In the retrospective cohort of 6420 type 2 diabetes patients, we estimated the associations between DRHD values and mortality related to all-cause, cardiovascular disease (CVD) and diabetes-specific causes using Cox proportional hazards regression models. Results A set of 14 biomarkers associated with DR was identified for the construction of DRHD value. During an average of 8 years of follow-up, the multivariable-adjusted HRs and corresponding 95 % CIs for the highest quartiles of DRHD values were 2.04 (1.76, 2.37), 2.32 (1.78, 3.01), and 2.29 (1.72, 3.04) for all-cause, CVD and diabetes-specific mortality, respectively. Furthermore, we developed a web-based calculator for the DRHD value to enhance its accessibility and usability (https://dzwxl-drhd.streamlit.app/). Conclusions Our study constructed the DRHD value as a measure to assess homeostatic dysregulation among individuals with type 2 diabetes. The DRHD values exhibited potential as a prognostic indicator for retinopathy and for mortality in patients affected by type 2 diabetes.

The phenomenon of population aging has brought forth the challenge of frailty. Nevertheless, the contribution of environmental exposure to frailty remains ambiguous. Our objective was to investigate the association between phenols, phthalates (PAEs), and polycyclic aromatic hydrocarbons (PAHs) with frailty. We constructed a 48-item frailty index using data from the National Health and Nutrition Examination Survey (NHANES). The exposure levels of 20 organic contaminants were obtained from the survey circle between 2005 and 2016. The association between individual organic contaminants and the frailty index was assessed using negative binomial regression models. The combined effect of organic contaminants was examined using weighted quantile sum (WQS) regression. Dose–response patterns were modeled using generalized additive models (GAMs). Additionally, an interpretable machine learning approach was employed to develop a predictive model for the frailty index. A total of 1566 participants were included in the analysis. Positive associations were observed between exposure to MIB, P02, ECP, MBP, MHH, MOH, MZP, MC1, and P01 with the frailty index. WQS regression analysis revealed a significant increase in the frailty index with higher levels of the mixture of organic contaminants (aOR, 1.12; 95

Glyphosate (GLY) is a widely used herbicide with potential adverse effects on public health. However, the current epidemiological evidence is limited. This study aimed to investigate the potential associations between exposure to GLY and multiple health outcomes. The data on urine GLY concentration and nine health outcomes, including type 2 diabetes mellitus (T2DM), hypertension, cardiovascular disease (CVD), obesity, chronic kidney disease (CKD), hepatic steatosis, cancers, chronic obstructive pulmonary disease (COPD), and neurodegenerative diseases (NGDs), were extracted from NHANES (2013–2016). The associations between GLY exposure and each health outcome were estimated using reverse-scale Cox regression and logistic regression. Furthermore, mediation analysis was conducted to identify potential mediators in the significant associations. The dose-response relationships between GLY exposure with health outcomes and potential mediators were analyzed using restricted cubic spline (RCS) regression. The findings of the study revealed that individuals with higher urinary concentrations of GLY had a higher likelihood of having T2DM, hypertension, CVD and obesity (p < 0.001, p = 0.005, p < 0.001 and p = 0.005, respectively). In the reverse-scale Cox regression, a notable association was solely discerned between exposure to GLY and the risk of T2DM (adjusted HR = 1.22, 95% CI: 1.10, 1.36). Consistent outcomes were also obtained via logistic regression analysis, wherein the adjusted OR and 95% CI for T2DM were determined to be 1.30 (1.12, 1.52). Moreover, the present investigation identified serum high-density lipoprotein cholesterol (HDL) as a mediator in this association, with a mediating effect of 7.14% (p = 0.040). This mediating effect was further substantiated by RCS regression, wherein significant dose-response associations were observed between GLY exposure and an increased risk of T2DM (p = 0.002) and reduced levels of HDL (p = 0.001). Collectively, these findings imply an association between GLY exposure and an increased risk of T2DM in the general adult population.

Heavy metal exposure is a common risk factor for hypertension. To develop an interpretable predictive machine learning (ML) model for hypertension based on levels of heavy metal exposure, data from the NHANES (2003–2016) were employed. Random forest (RF), support vector machine (SVM), decision tree (DT), multilayer perceptron (MLP), ridge regression (RR), AdaBoost (AB), gradient boosting decision tree (GBDT), voting classifier (VC), and K-nearest neighbour (KNN) algorithms were utilized to generate an optimal predictive model for hypertension. Three interpretable methods, the permutation feature importance analysis, partial dependence plot (PDP), and Shapley additive explanations (SHAP) methods, were integrated into a pipeline and embedded in ML for model interpretation. A total of 9005 eligible individuals were randomly allocated into two distinct sets for predictive model training and validation. The results showed that among the predictive models, the RF model demonstrated the highest performance, achieving an accuracy rate of 77.40% in the validation set. The AUC and F1 score for the model were 0.84 and 0.76, respectively. Blood Pb, urinary Cd, urinary Tl, and urinary Co levels were identified as the main influencers of hypertension, and their contribution weights were 0.0504 ± 0.0482, 0.0389 ± 0.0256, 0.0307 ± 0.0179, and 0.0296 ± 0.0162, respectively. Blood Pb (0.55–2.93 μg/dL) and urinary Cd (0.06–0.15 μg/L) levels exhibited the most pronounced upwards trend with the risk of hypertension within a specific value range, while urinary Tl (0.06–0.26 μg/L) and urinary Co (0.02–0.32 μg/L) levels demonstrated a declining trend with hypertension. The findings on the synergistic effects indicated that Pb and Cd were the primary determinants of hypertension. Our findings underscore the predictive value of heavy metals for hypertension. By utilizing interpretable methods, we discerned that Pb, Cd, Tl, and Co emerged as noteworthy contributors within the predictive model.

Purpose:Accurate identification of corneal layers with in vivo confocal microscopy (IVCM) is essential for the correct assessment of corneal lesions. This project aims to obtain a reliable automated identification of corneal layers from IVCM images.Methods:A total of 7957 IVCM images were included for model training and testing. Scanning depth information and pixel information of IVCM images were used to build the classification system. Firstly, two base classifiers based on convolutional neural networks and K-nearest neighbors were constructed. Second, two hybrid strategies, namely weighted voting method and light gradient boosting machine (LightGBM) algorithm were used to fuse the results from the two base classifiers and obtain the final classification. Finally, the confidence of prediction results was stratified to help find out model errors.Results:Both two hybrid systems outperformed the two base classifiers. The weighted area under the curve, weighted precision, weighted recall, and weighted F1 score were 0.9841, 0.9096, 0.9145, and 0.9111 for weighted voting hybrid system, and were 0.9794, 0.9039, 0.9055, and 0.9034 for the light gradient boosting machine stacking hybrid system, respectively. More than one-half of the misclassified samples were found using the confidence stratification method.Conclusions:The proposed hybrid approach could effectively integrate the scanning depth and pixel information of IVCM images, allowing for the accurate identification of corneal layers for grossly normal IVCM images. The confidence stratification approach was useful to find out misclassification of the system.Translational Relevance:The proposed hybrid approach lays important groundwork for the automatic identification of the corneal layer for IVCM images.

The precise impact of dietary components on vascular health remains incompletely understood. To identify the dietary components and their associations with abdominal aortic calcification (AAC), the data from NHANES was employed in this cross-sectional study. The LASSO method and logistic regression were utilized to identify dietary components that exhibited the strongest association with AAC. Grouped WQS regression analysis was employed to evaluate the combined effects of dietary components on AAC. Furthermore, principal component analysis was employed to identify the primary dietary patterns in the study population. The present analysis included 1862 participants, from whom information on 35 dietary macro- and micronutrient components was obtained through 24-hour dietary recall interviews. The assessment of AAC was performed utilizing dual-energy X-ray absorptiometry. The LASSO method identified 10 dietary components that were associated with AAC. Total protein, total fiber, vitamin A, and β-cryptoxanthin exhibited a negative association with AAC. Compared to the first quartile, the adjusted odds ratios (95% CIs) for the highest quartile were 0.59 (0.38, 0.93), 0.63 (0.42, 0.93), 0.59 (0.41, 0.84), and 0.68 (0.48, 0.94), respectively. Grouped WQS regression demonstrated a positive association between the lipid group and AAC (aOR: 1.29; 95% CI: 1.12, 1.50), while the proteins and phytochemical group exhibited a negative association with AAC (aOR: 0.69; 95% CI: 0.58, 0.82). For the dietary pattern analysis, high adherence to the plant-based pattern (aOR: 0.62; 95% CI: 0.44, 0.88) was associated with a lower risk of AAC, whereas the caffeine and theobromine pattern (aOR: 1.73; 95% CI: 1.25, 2.41) was associated with a higher risk of AAC. The findings of this study indicate that adopting a dietary pattern characterized by high levels of protein and plant-based foods, as well as reduced levels of fat, may offers potential advantages for the prevention of AAC.

PURPOSE:Fungal keratitis is a common cause of blindness worldwide. Timely identification of the causative fungal genera is essential for clinical management. In vivo confocal microscopy (IVCM) provides useful information on pathogenic genera. This study attempted to apply deep learning (DL) to establish an automated method to identify pathogenic fungal genera using IVCM images.METHODS:Deep learning networks were trained, validated, and tested using a data set of 3364 IVCM images that collected from 100 eyes of 100 patients with culture-proven filamentous fungal keratitis. Two transfer learning approaches were investigated: one was a combined framework that extracted features by a DL network and adopted decision tree (DT) as a classifier; another was a complete supervised DL model which used DL-based fully connected layers to implement the classification.RESULTS:The DL classifier model revealed better performance compared with the DT classifier model in an independent testing set. The DL classifier model showed an area under the receiver operating characteristic curves (AUC) of 0.887 with an accuracy of 0.817, sensitivity of 0.791, specificity of 0.831, G-mean of 0.811, and F1 score of 0.749 in identifying Fusarium, and achieved an AUC of 0.827 with an accuracy of 0.757, sensitivity of 0.756, specificity of 0.759, G-mean of 0.757, and F1 score of 0.716 in identifying Aspergillus.CONCLUSION:The DL model can classify Fusarium and Aspergillus by learning effective features in IVCM images automatically. The automated IVCM image analysis suggests a noninvasive identification of Fusarium and Aspergillus with clear potential application in early diagnosis and management of fungal keratitis.

Reactive oxygen species (ROS) overproduction plays an essential role in the etiology of ischemic/hypoxic retinopathy caused by acute glaucoma. NADPH oxidase (NOX) 4 was discovered as one of the main sources of ROS in glaucoma. However, the role and potential mechanisms of NOX4 in acute glaucoma have not been fully elucidated. Therefore, the current study aims to investigate the NOX4 inhibitor GLX351322 that targets NOX4 inhibition in acute ocular hypertension (AOH)-induced retinal ischemia/hypoxia injury in mice. Herein, NOX4 was highly expressed in AOH retinas, particularly the retinal ganglion cell layer (GCL). Importantly, the NOX4 inhibitor GLX351322 reduced ROS overproduction, inhibited inflammatory factor release, suppressed glial cell activation and hyperplasia, inhibited leukocyte infiltration, reduced retinal cell senescence and apoptosis in damaged areas, reduced retinal degeneration and improved retinal function. This neuroprotective effect is at least partially associated with mediated redox-sensitive factor (HIF-1α, NF-κB, and MAPKs) pathways by NOX4-derived ROS overproduction. These results suggest that inhibition of NOX4 with GLX351322 attenuated AOH-induced retinal inflammation, cellular senescence, and apoptosis by inhibiting the activation of the redox-sensitive factor pathway mediated by ROS overproduction, thereby protecting retinal structure and function. Targeted inhibition of NOX4 is expected to be a new idea in the treatment of acute glaucoma.