Deep learning (DL) algorithms are used to diagnose diabetic retinopathy (DR). However, most of these algorithms have been trained using global data or data from patients of a single region. Using different model architectures (e.g., Inception-v3, ResNet101, and DenseNet121), we assessed the necessity of modifying the algorithms for universal society screening. We used the open-source dataset from the Kaggle Diabetic Retinopathy Detection competition to develop a model for the detection of DR severity. We used a local dataset from Taipei City Hospital to verify the necessity of model localization and validated the three aforementioned models with local datasets. The experimental results revealed that Inception-v3 outperformed ResNet101 and DenseNet121 with a foreign global dataset, whereas DenseNet121 outperformed Inception-v3 and ResNet101 with the local dataset. The quadratic weighted kappa score ($\kappa$) was used to evaluate the model performance. All models had 5–8% higher κ for the local dataset than for the foreign dataset. Confusion matrix analysis revealed that, compared with the local ophthalmologists’ diagnoses, the severity predicted by the three models was overestimated. Thus, DL algorithms using artificial intelligence based on global data must be locally modified to ensure the applicability of a well-trained model to make diagnoses in clinical environments.

Background: Ambient fine particulate matter (PM2.5) has emerged as a critical environmental threat to ocular health; however, the underlying molecular mechanisms affecting the retinal pigment epithelium (RPE) remain largely uncharacterized. This study aimed to investigate transcriptomic alterations in RPE cells following PM2.5 exposure and to identify key regulatory pathways involved. Methods: Next-generation sequencing (NGS) was used to investigate differential gene expression in ARPE-19 cells upon PM2.5 exposure. Bioinformatic analyses, including pathway enrichment and gene set enrichment analysis (GSEA), were performed to identify affected signaling cascades. Functional assays—including cell viability, wound healing, and Transwell migration—were conducted to evaluate phenotypic changes. Quantitative RT-PCR (Reverse Transcription Polymerase Chain Reaction) and ELISA (Enzyme-Linked Immunosorbent Assay) validated gene expression and transforming growth factor-beta (TGF-β) secretion. TGF-β stimulation and receptor inhibition were applied to dissect pathway involvement. Results: Comprehensive analysis revealed substantial changes in gene expression profiles, with pathway enrichment highlighting the activation of cell migration-related pathways such as focal adhesion, regulation of actin cytoskeleton, extracellular matrix (ECM)–receptor interaction, tight junction, and adherens junction. Notably, the TGF-β, MAPK (Mitogen-Activated Protein Kinase), and PI3K/AKT (Phosphoinositide 3-Kinase / Protein Kinase B) pathways were significantly modulated. Functional assays showed that PM2.5 exposure enhanced ARPE-19 cell viability and migratory capacity. Among the differentially expressed genes, angiopoietin-like 4 (ANGPTL4) was markedly upregulated following PM2.5 stimulation. Pharmacological inhibition of TGF-β signaling abrogated PM2.5-induced ANGPTL4 expression, suggesting a pivotal role of the TGF-β pathway in mediating these effects. Conclusion: These findings demonstrate that PM2.5 induces transcriptomic reprogramming and activates the TGF-β signaling cascade in RPE cells, thereby enhancing cellular migration. Specifically, ANGPTL4 was identified as a key downstream effector of this pathway. This study provides novel insights into the molecular mechanisms by which air pollution contributes to retinal disease pathogenesis and suggests potential therapeutic targets for preventing PM2.5-induced retinal injury.

Portable magnetic-assisted capsule endoscopy (MACE) provides satisfactory patient experience and safety with comparable performance in diagnosis of organic lesions when compared to conventional upper gastrointestinal endoscopy. In this study, a total of 58 homecare patients were included for MACE either in the hospital (n = 42) or at home (n = 16), with mean age of 71.1 ± 12.4 years. A total of 55 patients (94.83%) had completed the MACE with diagnosis of reflux esophagitis (43.6%), gastritis (54.5%), erosions (21.8%), fundic polyps (14.5%), peptic ulcers (25.9%), etc. Most patients (n = 47, 85.5%) were satisfied with the experience, and all patients who received MACE at home (n = 15, 100%) appreciated the convenience of endoscopy at home. Less than half of the patients (n = 24, 43.6%) could afford MACE if the expense was not covered by health insurance (USD 714). Time consumption from both traffic and capsule manipulation was also challenging for the physicians, as it took an average of 24.7 min to complete MACE, but it added up to a total of 92.7 min at home, which is about 15 times that of conventional endoscopy in hospital. More efforts are needed to ease the financial burden of patients, and optimization of workflow in community practice may help lift the obstacles revealed in this study.