Causal relationships between blood metabolites and diabetic retinopathy: a two-sample Mendelian randomization study.

Background: Diabetic retinopathy (DR) is a microvascular complication of diabetes, severely affecting patients' vision and even leading to blindness. The development of DR is influenced by metabolic disturbance and genetic factors, including gene polymorphisms. The research aimed to uncover the causal relationships between blood metabolites and DR. Methods: The two-sample mendelian randomization (MR) analysis was employed to estimate the causality of blood metabolites on DR. The genetic variables for exposure were obtained from the genome-wide association study (GWAS) dataset of 486 blood metabolites, while the genetic predictors for outcomes including all-stage DR (All DR), non-proliferative DR (NPDR) and proliferative DR (PDR) were derived from the FinnGen database. The primary analysis employed inverse variance weighted (IVW) method, and supplementary analyses were performed using MR-Egger, weighted median (WM), simple mode and weighted mode methods. Additionally, MR-Egger intercept test, Cochran's Q test, and leave-one-out analysis were also conducted to guarantee the accuracy and robustness of the results. Subsequently, we replicated the MR analysis using three additional datasets from the FinnGen database and conducted a meta-analysis to determine blood metabolites associated with DR. Finally, reverse MR analysis and metabolic pathway analysis were performed. Results: The study identified 13 blood metabolites associated with All DR, 9 blood metabolites associated with NPDR and 12 blood metabolites associated with PDR. In summary, a total of 21 blood metabolites were identified as having potential causal relationships with DR. Additionally, we identified 4 metabolic pathways that are related to DR. Conclusion: The research revealed a number of blood metabolites and metabolic pathways that are causally associated with DR, which holds significant importance for screening and prevention of DR. However, it is noteworthy that these causal relationships should be validated in larger cohorts and experiments.

Microglia-endothelial cross-talk regulates diabetes-induced retinal vascular dysfunction through remodeling inflammatory microenvironment.

Inflammation-mediated crosstalk between neuroglial cells and endothelial cells (ECs) is a fundamental feature of many vascular diseases. Nevertheless, the landscape of inflammatory processes during diabetes-induced microvascular dysfunction remains elusive. Here, we applied single-cell RNA sequencing to elucidate the transcriptional landscape of diabetic retinopathy (DR). The transcriptome characteristics of microglia and ECs revealed two microglial subpopulations and three EC populations. Exploration of intercellular crosstalk between microglia and ECs showed that diabetes-induced interactions mainly participated in the inflammatory response and vessel development, with colony-stimulating factor 1 (CSF1) and CSF1 receptor (CSF1R) playing important roles in early cell differentiation. Clinically, we found that CSF1/CSF1R crosstalk dysregulation was associated with proliferative DR. Mechanistically, ECs secrete CSF1 and activate CSF1R endocytosis and the CSF1R phosphorylation-mediated MAPK signaling pathway, which elicits the differentiation of microglia and triggers the secretion of inflammatory factors, and subsequently foster angiogenesis by remodeling the inflammatory microenvironment through a positive feedback mechanism.

Hyperglycemia-regulated tRNA-derived fragment tRF-3001a propels neurovascular dysfunction in diabetic mice.

Neurovascular dysfunction is a preclinical manifestation of diabetic complications, including diabetic retinopathy (DR). Herein, we report that a transfer RNA-derived RNA fragment, tRF-3001a, is significantly upregulated under diabetic conditions. tRF-3001a downregulation inhibits Müller cell activation, suppresses endothelial angiogenic effects, and protects against high-glucose-induced retinal ganglion cell injury in vitro. Furthermore, tRF-3001a downregulation alleviates retinal vascular dysfunction, inhibits retinal reactive gliosis, facilitates retinal ganglion cell survival, and preserves visual function and visually guided behaviors in STZ-induced diabetic mice and db/db diabetic mice. Mechanistically, tRF-3001a regulates neurovascular dysfunction in a microRNA-like mechanism by targeting GSK3B. Clinically, tRF-3001a is upregulated in aqueous humor (AH) samples of DR patients. tRF-3001a downregulation inhibits DR-induced human retinal vascular endothelial cell and Müller cell dysfunction in vitro and DR-induced retinal neurovascular dysfunction in C57BL/6J mice. Thus, targeting tRF-3001a-mediated signaling is a promising strategy for the concurrent treatment of vasculopathy and neuropathy in diabetes mellitus.

A PSD-95 peptidomimetic mitigates neurological deficits in a mouse model of Angelman syndrome.

Angelman Syndrome (AS) is a severe cognitive disorder caused by loss of neuronal expression of the E3 ubiquitin ligase UBE3A. In an AS mouse model, we previously reported a deficit in brain-derived neurotrophic factor (BDNF) signaling, and set out to develop a therapeutic that would restore normal signaling. We demonstrate that CN2097, a peptidomimetic compound that binds postsynaptic density protein-95 (PSD-95), a TrkB associated scaffolding protein, mitigates deficits in PLC-CaMKII and PI3K/mTOR pathways to restore synaptic plasticity and learning. Administration of CN2097 facilitated long-term potentiation (LTP) and corrected paired-pulse ratio. As the BDNF-mTORC1 pathway is critical for inhibition of autophagy, we investigated whether autophagy was disrupted in AS mice. We found aberrantly high autophagic activity attributable to a concomitant decrease in mTORC1 signaling, resulting in decreased levels of synaptic proteins, including Synapsin-1 and Shank3. CN2097 increased mTORC1 activity to normalize autophagy and restore hippocampal synaptic protein levels. Importantly, treatment mitigated cognitive and motor dysfunction. These findings support the use of neurotrophic therapeutics as a valuable approach for treating AS pathology.

Single-cell RNA sequencing reveals a unique pericyte type associated with capillary dysfunction.

Background: Capillary dysfunction has been implicated in a series of life- threatening vascular diseases characterized by pericyte and endothelial cell (EC) degeneration. However, the molecular profiles that govern the heterogeneity of pericytes have not been fully elucidated. Methods: Single-cell RNA sequencing was conducted on oxygen-induced proliferative retinopathy (OIR) model. Bioinformatics analysis was conducted to identify specific pericytes involved in capillary dysfunction. qRT-PCRs and western blots were conducted to detect Col1a1 expression pattern during capillary dysfunction. Matrigel co-culture assays, PI staining, and JC-1 staining was conducted to determine the role of Col1a1 in pericyte biology. IB4 and NG2 staining was conducted to determine the role of Col1a1 in capillary dysfunction. Results: We constructed an atlas of > 76,000 single-cell transcriptomes from 4 mouse retinas, which could be annotated to 10 distinct retinal cell types. Using the sub-clustering analysis, we further characterized retinal pericytes into 3 different subpopulations. Notably, GO and KEGG pathway analysis demonstrated that pericyte sub-population 2 was identified to be vulnerable to retinal capillary dysfunction. Based on the single-cell sequencing results, Col1a1 was identified as a marker gene of pericyte sub-population 2 and a promising therapeutic target for capillary dysfunction. Col1a1 was abundantly expressed in pericytes and its expression was obviously upregulated in OIR retinas. Col1a1 silencing could retard the recruitment of pericytes toward endothelial cells and aggravated hypoxia-induced pericyte apoptosis in vitro. Col1a1 silencing could reduce the size of neovascular area and avascular area in OIR retinas and suppressed pericyte-myofibroblast transition and endothelial-mesenchymal transition. Moreover, Col1a1 expression was up-regulated in the aqueous humor of the patients with proliferative diabetic retinopathy (PDR) or retinopathy of prematurity (ROP) and up-regulated in the proliferative membranes of PDR patients. Conclusions: These findings enhance the understanding of the complexity and heterogeneity of retinal cells and have important implications for future treatment of capillary dysfunction.

YY1 lactylation in microglia promotes angiogenesis through transcription activation-mediated upregulation of FGF2.

BACKGROUND: Ocular neovascularization is a leading cause of blindness. Retinal microglia have been implicated in hypoxia-induced angiogenesis and vasculopathy, but the underlying mechanisms are not entirely clear. Lactylation is a novel lactate-derived posttranslational modification that plays key roles in multiple cellular processes. Since hypoxia in ischemic retinopathy is a precipitating factor for retinal neovascularization, lactylation is very likely to be involved in this process. The present study aimed to explore the role of lactylation in retinal neovascularization and identify new therapeutic targets for retinal neovascular diseases. RESULTS: Microglial depletion by the colony-stimulating factor 1 receptor (CSF1R) inhibitor PLX3397 suppresses retinal neovascularization in oxygen-induced retinopathy. Hypoxia increased lactylation in microglia and accelerates FGF2 expression, promoting retinal neovascularization. We identify 77 sites of 67 proteins with increased lactylation in the context of increased lactate under hypoxia. Our results show that the nonhistone protein Yin Yang-1 (YY1), a transcription factor, is lactylated at lysine 183 (K183), which is regulated by p300. Hyperlactylated YY1 directly enhances FGF2 transcription and promotes angiogenesis. YY1 mutation at K183 eliminates these effects. Overexpression of p300 increases YY1 lactylation and enhances angiogenesis in vitro and administration of the p300 inhibitor A485 greatly suppresses vascularization in vivo and in vitro. CONCLUSIONS: Our results suggest that YY1 lactylation in microglia plays an important role in retinal neovascularization by upregulating FGF2 expression. Targeting the lactate/p300/YY1 lactylation/FGF2 axis may provide new therapeutic targets for proliferative retinopathies.

CircRNA Uxs1/miR-335-5p/PGF axis regulates choroidal neovascularization via the mTOR/p70 S6k pathway.

Age-related macular degeneration (AMD) is one of the leading causes of irreversible blindness in the elderly population. Neovascular AMD is the late stage, characterized by choroidal neovascularization (CNV). Non-coding RNAs have been implicated in CNV; however, the role of circular RNAs (circRNAs) has not yet been elucidated. Herein, we comprehensively investigated circRNA profiles in laser-induced CNV mouse models and patient specimens. A novel circRNA, circRNA Uxs1, was identified, and its function in CNV regulation was investigated in the present study. CircRNA Uxs1 was consistently upregulated in CNV patient specimens and CNV mouse models. Knockdown of circRNA Uxs1 interrupted the tube formation, migration, and proliferation of endothelial cells in vitro. Silencing circRNA Uxs1 in vivo alleviated neovascularization formation, as shown by the decreased size of laser spots. Mechanistically, circRNA Uxs1 functioned by binding to miR-335-5p, which further upregulated the expression of placental growth factor (PGF) gene and activated the mammalian target of rapamycin/p70 S6 Kinase (mTOR/p70 S6k) pathway. By subretinal injections of adeno-associated virus (AAV), we demonstrated the anti-angiogenic function of circRNA Uxs1 knockdown in vivo. In conclusion, circRNA Uxs1 promoted CNV by sponging miR-335-5p, which stimulated PGF expression and subsequently activated the mTOR/p70 S6k pathway. Therefore, circRNA Uxs1 may serve as a promising therapeutic target for CNV.

tRNA-derived fragment tRF-1001: A novel anti-angiogenic factor in pathological ocular angiogenesis.

Transfer RNA-derived fragments (tRFs) are a novel class of non-coding RNA transcripts and play important roles in several physiological/pathological processes. However, the role of tRFs in ocular angiogenesis remains elusive. Herein, we investigate whether the intervention of tRF-1001 expression could suppress pathological ocular angiogenesis. The results show that the levels of tRF-1001 expression were reduced in the retinas of an oxygen-induced retinopathy (OIR) model, choroidal neovascularization model, and endothelial sprouting model in vitro. Increased tRF-1001 expression could suppress ocular angiogenesis and endothelial sprouting in vivo and reduce endothelial migration, specification, and sprouting in vitro. Mechanistically, tRF-1001 regulated endothelial angiogenic effects via tRF-1001/METTL3/RBPJ-MAML1 signaling. The levels of tRF-1001 expression were downregulated in the aqueous humor of age-related macular degeneration (AMD) patients. tRF-1001 upregulation could suppress AMD aqueous humor-induced endothelial sprouting and pathological angiogenesis. Collectively, tRF-1001 acts as an anti-angiogenic factor during ocular angiogenesis. Targeting tRF-1001-mediated signaling is a therapeutic option for ocular neovascular diseases.

Silencing of circular RNA­ZYG11B exerts a neuroprotective effect against retinal neurodegeneration.

Ischemic retinal diseases are the major cause of vision impairment worldwide. Currently, there are no available treatments for ischemia­induced retinal neurodegeneration. Circular RNAs (circRNAs) have emerged as important regulators of several biological processes and human diseases. The present study investigated the role of circRNA­ZYG11B (circZYG11B; hsa_circ_0003739) in retinal neurodegeneration. Reverse transcription quantitative polymerase chain reaction (RT­qPCR) demonstrated that circZYG11B expression was markedly increased during retinal neurodegeneration in vivo and in vitro. Cell Counting Kit­8, TUNEL and caspase­3 activity assays revealed that silencing of circZYG11B was able to protect against oxidative stress­ or hypoxic stress­induced retinal ganglion cell (RGC) injury. Furthermore, immunofluorescence staining and hematoxylin and eosin staining revealed that silencing of circZYG11B alleviated ischemia/reperfusion­induced retinal neurodegeneration, as indicated by reduced RGC injury and decreased retinal reactive gliosis. In addition, luciferase reporter, biotin­coupled miRNA capture and RNA immunoprecipitation assays revealed that circZYG11B could regulate RGC function through circZYG11B/microRNA­620/PTEN signaling. Clinically, RT­qPCR assays demonstrated that circZYG11B expression was markedly increased in the aqueous humor of patients with glaucoma. In conclusion, circZYG11B may be considered a promising target for the diagnosis and treatment of retinal ischemic diseases.

Targeting long noncoding RNA-AQP4-AS1 for the treatment of retinal neurovascular dysfunction in diabetes mellitus.

BACKGROUND: Diabetic retinopathy (DR) is a leading cause of blindness in the working-age population, which is characterized by retinal neurodegeneration and vascular dysfunction. Long non-coding RNAs (LncRNAs) have emerged as critical regulators in several biological processes and disease progression. Here we investigated the role of lncRNA AQP4-AS1 in retinal neurovascular dysfunction induced by diabetes. METHODS: Quantitative RT-PCR was used to detect the AQP4-AS1 expression pattern upon diabetes mellitus-related stresses. Visual electrophysiology examination, TUNEL staining, Evans blue staining, retinal trypsin digestion and immunofluorescent staining were conducted to detect the role of AQP4-AS1 in retinal neurovascular dysfunction in vivo. MTT assays, TUNEL staining, PI/Calcein-AM staining, EdU incorporation assay transwell assay and tube formation were conducted to detect the role of AQP4-AS1 in retinal cells function in vitro. qRT-PCR, western blot and in vivo studies were conducted to reveal the mechanism of AQP4-AS1-mediated retinal neurovascular dysfunction. FINDINGS: AQP4-AS1 was significantly increased in the clinical samples of diabetic retinopathy patients, high glucose-treated Müller cells, and diabetic retinas of a murine model. AQP4-AS1 silencing in vivo alleviated retinal neurodegeneration and vascular dysfunction as shown by improved retinal capillary degeneration, decreased reactive gliosis, and reduced RGC loss. AQP4-AS1 directly regulated Müller cell function and indirectly affected endothelial cell and RGC function in vitro. Mechanistically, AQP4-AS1 regulated retinal neurovascular dysfunction through affecting AQP4 levels. INTERPRETATION: This study reveals AQP4-AS1 is involved in retinal neurovascular dysfunction and expected to become a promising target for the treatment of neurovascular dysfunction in DR. FUNDING: This work was generously supported by the grants from the National Natural Science Foundation of China (Grant No. 81800858, 82070983, 81870679 and 81970823), grants from the Medical Science and Technology Development Project Fund of Nanjing (Grant No ZKX17053 and YKK19158), grants from Innovation Team Project Fund of Jiangsu Province (No. CXTDB2017010), and the Science and Technology Development Plan Project Fund of Nanjing (Grant No 201716007, 201805007 and 201803058).

A new detection model of microaneurysms based on improved FC-DenseNet.

Diabetic retinopathy (DR) is a frequent vascular complication of diabetes mellitus and remains a leading cause of vision loss worldwide. Microaneurysm (MA) is usually the first symptom of DR that leads to blood leakage in the retina. Periodic detection of MAs will facilitate early detection of DR and reduction of vision injury. In this study, we proposed a novel model for the detection of MAs in fluorescein fundus angiography (FFA) images based on the improved FC-DenseNet, MAs-FC-DenseNet. FFA images were pre-processed by the Histogram Stretching and Gaussian Filtering algorithm to improve the quality of FFA images. Then, MA regions were detected by the improved FC-DenseNet. MAs-FC-DenseNet was compared against other FC-DenseNet models (FC-DenseNet56 and FC-DenseNet67) or the end-to-end models (DeeplabV3+ and PSPNet) to evaluate the detection performance of MAs. The result suggested that MAs-FC-DenseNet had higher values of evaluation metrics than other models, including pixel accuracy (PA), mean pixel accuracy (MPA), precision (Pre), recall (Re), F1-score (F1), and mean intersection over union (MIoU). Moreover, MA detection performance for MAs-FC-DenseNet was very close to the ground truth. Taken together, MAs-FC-DenseNet is a reliable model for rapid and accurate detection of MAs, which would be used for mass screening of DR patients.

METTL3-mediated N6-methyladenosine modification governs pericyte dysfunction during diabetes-induced retinal vascular complication.

Rationale: Microvascular complication is a major cause of morbidity and mortality among the patients with diabetes. Pericyte dysfunction is the predominant pathological manifestation of microvascular complication. N6-methyladenosine (m6A) serves as the most prevalent modification in eukaryotic mRNAs. However, the role of m6A RNA modification in pericyte dysfunction is still unclear. Methods: Quantitative polymerase chain reactions and western blots were conducted to detect the change of m6A RNA modification in pericytes and mouse retinas following diabetic stress. MTT assay, transwell migration assay, caspase 3/7 activity assay, calcein-AM/propidium iodide (PI) staining, and TUNEL staining were conducted to determine the role of METTL3 in pericyte biology in vitro. Retinal trypsin digestion, vascular permeability assay, and IB4-NG2 double immunofluorescent staining were conducted to determine the role of METTL3 in retinal pericyte dysfunction and vascular complication. RNA sequencing, RNA pull-down assays and immunoblots were conducted to clarify the mechanism of METTL3-mediated pericyte dysfunction and vascular complication. Results: The levels of m6A RNA methylation were significantly up-regulated in pericytes and mouse retinas following diabetic stress, which were caused by increased expression of METTL3. METTL3 regulated the viability, proliferation, and differentiation of pericytes in vitro. Specific depletion of METTL3 in pericytes suppressed diabetes-induced pericyte dysfunction and vascular complication in vivo. METTL3 overexpression impaired pericyte function by repressing PKC-η, FAT4, and PDGFRA expression, which was mediated by YTHDF2-dependent mRNA decay. Conclusion: METTL3-mediated m6A methylation epigenetically regulates diabetes-induced pericyte dysfunction. METTL3-YTHDF2-PKC-η/FAT4/PDGFRA signaling axis could be therapeutically targeted for treating microvascular complications.

Targeting circular RNA-MET for anti-angiogenesis treatment via inhibiting endothelial tip cell specialization.

Endothelial tip cell specialization plays an essential role in angiogenesis, which is tightly regulated by the complicated gene regulatory network. Circular RNA (circRNA) is a type of covalently closed non-coding RNA that regulates gene expression in eukaryotes. Here, we report that the levels of circMET expression are significantly upregulated in the retinas of mice with oxygen-induced retinopathy, choroidal neovascularization, and diabetic retinopathy. circMET silencing significantly reduces pathological angiogenesis and inhibits tip cell specialization in vivo. circMET silencing also decreases endothelial migration and sprouting in vitro. Mechanistically, circMET regulates endothelial sprouting and pathological angiogenesis by acting as a scaffold to enhance the interaction between IGF2BP2 and NRARP/ESM1. Clinically, circMET is significantly upregulated in the clinical samples of the patients of diabetic retinopathy. circMET silencing could reduce diabetic vitreous-induced endothelial sprouting and retinal angiogenesis in vivo. Collectively, these data identify a circRNA-mediated mechanism that coordinates tip cell specialization and pathological angiogenesis. circMET silencing is an exploitable therapeutic approach for the treatment of neovascular diseases.

Automated segmentation of macular edema for the diagnosis of ocular disease using deep learning method.

Macular edema is considered as a major cause of visual loss and blindness in patients with ocular fundus diseases. Optical coherence tomography (OCT) is a non-invasive imaging technique, which has been widely applied for diagnosing macular edema due to its non-invasive and high resolution properties. However, the practical applications remain challenges due to the distorted retinal morphology and blurred boundaries near macular edema. Herein, we developed a novel deep learning model for the segmentation of macular edema in OCT images based on DeepLab framework (OCT-DeepLab). In this model, we used atrous spatial pyramid pooling (ASPP) to detect macular edema at multiple features and used the fully connected conditional random field (CRF) to refine the boundary of macular edema. OCT-DeepLab model was compared against the traditional hand-crafted methods (C-V and SBG) and the end-to-end methods (FCN, PSPnet, and U-net) to estimate the segmentation performance. OCT-DeepLab showed great advantage over the hand-crafted methods (C-V and SBG) and end-to-end methods (FCN, PSPnet, and U-net) as shown by higher precision, sensitivity, specificity, and F1-score. The segmentation performance of OCT-DeepLab was comparable to that of manual label, with an average area under the curve (AUC) of 0.963, which was superior to other end-to-end methods (FCN, PSPnet, and U-net). Collectively, OCT-DeepLab model is suitable for the segmentation of macular edema and assist ophthalmologists in the management of ocular disease.

Suppression of choroidal neovascularization by silencing of long non-coding RNA IPW.

Long noncoding RNAs (lncRNAs) have emerged as the key regulators in the pathogenesis of human disorders. This study aimed to investigate the role of lncRNA-IPW in the progression of choroidal neovascularization (CNV) and the underlying molecular mechanism. IPW was significantly up-regulated in the choroidal tissues of laser-induced CNV mice and in the endothelial cells in response to hypoxic stress. IPW silencing led to reduced formation of CNV in laser-induced CNV model and ex vivo choroidal sprouting model, which could achieve similar therapeutic effects of anti-VEGF on CNV formation. Silencing or transgenic overexpression of IPW could alter endothelial cell viability, proliferation, migration, and tube formation ability in vitro. Mechanistically, IPW silencing led to increased expression of miR-370. Increased miR-370 could mimic the effects of IPW silencing on CNV formation and endothelial angiogenic phenotypes in vivo and in vitro. This study suggests that IPW silencing is a promising strategy for the treatment of neovascular ocular diseases.

CircRNA expression profile and functional analysis in retinal ischemia-reperfusion injury.

Retinal ischemia-reperfusion (I/R) is involved in the pathogenesis of many vision-threatening diseases. circRNAs act as key players in gene regulation and human diseases. However, the global circRNA expression profile in retinal I/R injury has not been fully uncovered. Herein, we established a murine model of retinal I/R injury and performed circRNA microarrays to identify I/R-related circRNAs. 1265 differentially expressed circRNAs were identified between I/R retinas and normal retinas. Notably, the detection of cWDR37 level in aqueous humor could discriminate glaucoma patients from cataract patients (AUC = 0.9367). cWdr37 silencing protected against hypoxic stress- or oxidative stress-induced retinal ganglion cell (RGC) injury. cWdr37 silencing alleviated IR-induced retinal neurodegeneration as shown by increased NeuN staining, reduced retinal reactive gliosis, and decreased retinal apoptosis. Collectively, this study provides a novel insight into the pathogenesis of retinal I/R injury. cWdr37 is a promising target for the diagnosis or treatment of I/R-related ocular diseases.

A Joint Model for Macular Edema Analysis in Optical Coherence Tomography Images Based on Image Enhancement and Segmentation.

Optical coherence tomography (OCT) provides the visualization of macular edema which can assist ophthalmologists in the diagnosis of ocular diseases. Macular edema is a major cause of vision loss in patients with retinal vein occlusion (RVO). However, manual delineation of macular edema is a laborious and time-consuming task. This study proposes a joint model for automatic delineation of macular edema in OCT images. This model consists of two steps: image enhancement using a bioinspired algorithm and macular edema segmentation using a Gaussian-filtering regularized level set (SBGFRLS) algorithm. We then evaluated the delineation efficiency using the following parameters: accuracy, precision, sensitivity, specificity, Dice's similarity coefficient, IOU, and kappa coefficient. Compared with the traditional level set algorithms, including C-V and GAC, the proposed model had higher efficiency in macular edema delineation as shown by reduced processing time and iteration times. Moreover, the accuracy, precision, sensitivity, specificity, Dice's similarity coefficient, IOU, and kappa coefficient for macular edema delineation could reach 99.7%, 97.8%, 96.0%, 99.0%, 96.9%, 94.0%, and 96.8%, respectively. More importantly, the proposed model had comparable precision but shorter processing time compared with manual delineation. Collectively, this study provides a novel model for the delineation of macular edema in OCT images, which can assist the ophthalmologists for the screening and diagnosis of retinal diseases.