Genome-wide association identifies novel ROP risk loci in a multiethnic cohort.

We conducted a genome-wide association study (GWAS) in a multiethnic cohort of 920 at-risk infants for retinopathy of prematurity (ROP), a major cause of childhood blindness, identifying 1 locus at genome-wide significance level (p < 5×10-8) and 9 with significance of p < 5×10-6 for ROP ≥ stage 3. The most significant locus, rs2058019, reached genome-wide significance within the full multiethnic cohort (p = 4.96×10-9); Hispanic and European Ancestry infants driving the association. The lead single nucleotide polymorphism (SNP) falls in an intronic region within the Glioma-associated oncogene family zinc finger 3 (GLI3) gene. Relevance for GLI3 and other top-associated genes to human ocular disease was substantiated through in-silico extension analyses, genetic risk score analysis and expression profiling in human donor eye tissues. Thus, we identify a novel locus at GLI3 with relevance to retinal biology, supporting genetic susceptibilities for ROP risk with possible variability by race and ethnicity.

Patterns of Gene Expression, Splicing, and Allele-Specific Expression Vary among Macular Tissues and Clinical Stages of Age-Related Macular Degeneration.

Age-related macular degeneration (AMD) is a leading cause of blindness, and elucidating its underlying disease mechanisms is vital to the development of appropriate therapeutics. We identified differentially expressed genes (DEGs) and differentially spliced genes (DSGs) across the clinical stages of AMD in disease-affected tissue, the macular retina pigment epithelium (RPE)/choroid and the macular neural retina within the same eye. We utilized 27 deeply phenotyped donor eyes (recovered within a 6 h postmortem interval time) from Caucasian donors (60-94 years) using a standardized published protocol. Significant findings were then validated in an independent set of well-characterized donor eyes (n = 85). There was limited overlap between DEGs and DSGs, suggesting distinct mechanisms at play in AMD pathophysiology. A greater number of previously reported AMD loci overlapped with DSGs compared to DEGs between disease states, and no DEG overlap with previously reported loci was found in the macular retina between disease states. Additionally, we explored allele-specific expression (ASE) in coding regions of previously reported AMD risk loci, uncovering a significant imbalance in C3 rs2230199 and CFH rs1061170 in the macular RPE/choroid for normal eyes and intermediate AMD (iAMD), and for CFH rs1061147 in the macular RPE/choroid for normal eyes and iAMD, and separately neovascular AMD (NEO). Only significant DEGs/DSGs from the macular RPE/choroid were found to overlap between disease states. STAT1, validated between the iAMD vs. normal comparison, and AGTPBP1, BBS5, CERKL, FGFBP2, KIFC3, RORα, and ZNF292, validated between the NEO vs. normal comparison, revealed an intricate regulatory network with transcription factors and miRNAs identifying potential upstream and downstream regulators. Findings regarding the complement genes C3 and CFH suggest that coding variants at these loci may influence AMD development via an imbalance of gene expression in a tissue-specific manner. Our study provides crucial insights into the multifaceted genomic underpinnings of AMD (i.e., tissue-specific gene expression changes, potential splice variation, and allelic imbalance), which may open new avenues for AMD diagnostics and therapies specific to iAMD and NEO.

Integrated multi-omics single cell atlas of the human retina.

Single-cell sequencing has revolutionized the scale and resolution of molecular profiling of tissues and organs. Here, we present an integrated multimodal reference atlas of the most accessible portion of the mammalian central nervous system, the retina. We compiled around 2.4 million cells from 55 donors, including 1.4 million unpublished data points, to create a comprehensive human retina cell atlas (HRCA) of transcriptome and chromatin accessibility, unveiling over 110 types. Engaging the retina community, we annotated each cluster, refined the Cell Ontology for the retina, identified distinct marker genes, and characterized cis-regulatory elements and gene regulatory networks (GRNs) for these cell types. Our analysis uncovered intriguing differences in transcriptome, chromatin, and GRNs across cell types. In addition, we modeled changes in gene expression and chromatin openness across gender and age. This integrated atlas also enabled the fine-mapping of GWAS and eQTL variants. Accessible through interactive browsers, this multimodal cross-donor and cross-lab HRCA, can facilitate a better understanding of retinal function and pathology.

Single-cell multiomics of the human retina reveals hierarchical transcription factor collaboration in mediating cell type-specific effects of genetic variants on gene regulation.

BACKGROUND: Systematic characterization of how  genetic variation modulates gene regulation in a cell type-specific context is essential for understanding complex traits. To address this question, we profile gene expression and chromatin accessibility in cells from healthy retinae of 20 human donors through single-cell multiomics and genomic sequencing. RESULTS: We map eQTL, caQTL, allelic-specific expression, and allelic-specific chromatin accessibility in major retinal cell types. By integrating these results, we identify and characterize regulatory elements and genetic variants effective on gene regulation in individual cell types. The majority of identified sc-eQTLs and sc-caQTLs display cell type-specific effects, while the cis-elements containing genetic variants with cell type-specific effects are often accessible in multiple cell types. Furthermore, the transcription factors whose binding sites are perturbed by genetic variants tend to have higher expression levels in the cell types where the variants exert their effects, compared to the cell types where the variants have no impact. We further validate our findings with high-throughput reporter assays. Lastly, we identify the enriched cell types, candidate causal variants and genes, and cell type-specific regulatory mechanism underlying GWAS loci. CONCLUSIONS: Overall, genetic effects on gene regulation are highly context dependent. Our results suggest that cell type-dependent genetic effect is driven by precise modulation of both trans-factor expression and chromatin accessibility of cis-elements. Our findings indicate hierarchical collaboration among transcription factors plays a crucial role in mediating cell type-specific effects of genetic variants on gene regulation.

DeMixSC: a deconvolution framework that uses single-cell sequencing plus a small benchmark dataset for improved analysis of cell-type ratios in complex tissue samples.

Bulk deconvolution with single-cell/nucleus RNA-seq data is critical for understanding heterogeneity in complex biological samples, yet the technological discrepancy across sequencing platforms limits deconvolution accuracy. To address this, we introduce an experimental design to match inter-platform biological signals, hence revealing the technological discrepancy, and then develop a deconvolution framework called DeMixSC using the better-matched, i.e., benchmark, data. Built upon a novel weighted nonnegative least-squares framework, DeMixSC identifies and adjusts genes with high technological discrepancy and aligns the benchmark data with large patient cohorts of matched-tissue-type for large-scale deconvolution. Our results using a benchmark dataset of healthy retinas suggest much-improved deconvolution accuracy. Further analysis of a cohort of 453 patients with age-related macular degeneration supports the broad applicability of DeMixSC. Our findings reveal the impact of technological discrepancy on deconvolution performance and underscore the importance of a well-matched dataset to resolve this challenge. The developed DeMixSC framework is generally applicable for deconvolving large cohorts of disease tissues, and potentially cancer.

A multi-omics atlas of the human retina at single-cell resolution.

Cell classes in the human retina are highly heterogeneous with their abundance varying by several orders of magnitude. Here, we generated and integrated a multi-omics single-cell atlas of the adult human retina, including more than 250,000 nuclei for single-nuclei RNA-seq and 137,000 nuclei for single-nuclei ATAC-seq. Cross-species comparison of the retina atlas among human, monkey, mice, and chicken revealed relatively conserved and non-conserved types. Interestingly, the overall cell heterogeneity in primate retina decreases compared with that of rodent and chicken retina. Through integrative analysis, we identified 35,000 distal cis-element-gene pairs, constructed transcription factor (TF)-target regulons for more than 200 TFs, and partitioned the TFs into distinct co-active modules. We also revealed the heterogeneity of the cis-element-gene relationships in different cell types, even from the same class. Taken together, we present a comprehensive single-cell multi-omics atlas of the human retina as a resource that enables systematic molecular characterization at individual cell-type resolution.

A systems biology approach uncovers novel disease mechanisms in age-related macular degeneration.

Age-related macular degeneration (AMD) is a leading cause of blindness, affecting 200 million people worldwide. To identify genes that could be targeted for treatment, we created a molecular atlas at different stages of AMD. Our resource is comprised of RNA sequencing (RNA-seq) and DNA methylation microarrays from bulk macular retinal pigment epithelium (RPE)/choroid of clinically phenotyped normal and AMD donor eyes (n = 85), single-nucleus RNA-seq (164,399 cells), and single-nucleus assay for transposase-accessible chromatin (ATAC)-seq (125,822 cells) from the retina, RPE, and choroid of 6 AMD and 7 control donors. We identified 23 genome-wide significant loci differentially methylated in AMD, over 1,000 differentially expressed genes across different disease stages, and an AMD Müller state distinct from normal or gliosis. Chromatin accessibility peaks in genome-wide association study (GWAS) loci revealed putative causal genes for AMD, including HTRA1 and C6orf223. Our systems biology approach uncovered molecular mechanisms underlying AMD, including regulators of WNT signaling, FRZB and TLE2, as mechanistic players in disease.

Genome-wide association identifies novel ROP risk loci in a multi-ethnic cohort.

We conducted a genome-wide association study (GWAS) in a multiethnic cohort of 920 at-risk infants for retinopathy of prematurity (ROP), a major cause of childhood blindness, identifying 2 loci at genome-wide significance level (p<5×10-8) and 7 at suggestive significance (p<5×10-6) for ROP ≥ stage 3. The most significant locus, rs2058019, reached genome-wide significance within the full multiethnic cohort (p=4.96×10-9); Hispanic and Caucasian infants driving the association. The lead single nucleotide polymorphism (SNP) falls in an intronic region within the Glioma-associated oncogene family zinc finger 3 (GLI3) gene. Relevance for GLI3 and other top-associated genes to human ocular disease was substantiated through in-silico extension analyses, genetic risk score analysis and expression profiling in human donor eye tissues. Thus, we report the largest ROP GWAS to date, identifying a novel locus at GLI3 with relevance to retinal biology supporting genetic susceptibilities for ROP risk with possible variability by race and ethnicity.

Heritable Risk and Protective Genetic Components of Glaucoma Medication Non-Adherence.

Glaucoma is the leading cause of irreversible blindness, affecting 76 million globally. It is characterized by irreversible damage to the optic nerve. Pharmacotherapy manages intraocular pressure (IOP) and slows disease progression. However, non-adherence to glaucoma medications remains problematic, with 41-71% of patients being non-adherent to their prescribed medication. Despite substantial investment in research, clinical effort, and patient education protocols, non-adherence remains high. Therefore, we aimed to determine if there is a substantive genetic component behind patients' glaucoma medication non-adherence. We assessed glaucoma medication non-adherence with prescription refill data from the Marshfield Clinic Healthcare System's pharmacy dispensing database. Two standard measures were calculated: the medication possession ratio (MPR) and the proportion of days covered (PDC). Non-adherence on each metric was defined as less than 80% medication coverage over 12 months. Genotyping was done using the Illumina HumanCoreExome BeadChip in addition to exome sequencing on the 230 patients (1) to calculate the heritability of glaucoma medication non-adherence and (2) to identify SNPs and/or coding variants in genes associated with medication non-adherence. Ingenuity pathway analysis (IPA) was utilized to derive biological meaning from any significant genes in aggregate. Over 12 months, 59% of patients were found to be non-adherent as measured by the MPR80, and 67% were non-adherent as measured by the PDC80. Genome-wide complex trait analysis (GCTA) suggested that 57% (MPR80) and 48% (PDC80) of glaucoma medication non-adherence could be attributed to a genetic component. Missense mutations in TTC28, KIAA1731, ADAMTS5, OR2W3, OR10A6, SAXO2, KCTD18, CHCHD6, and UPK1A were all found to be significantly associated with glaucoma medication non-adherence by whole exome sequencing after Bonferroni correction (p < 10-3) (PDC80). While missense mutations in TINAG, CHCHD6, GSTZ1, and SEMA4G were found to be significantly associated with medication non-adherence by whole exome sequencing after Bonferroni correction (p < 10-3) (MPR80). The same coding SNP in CHCHD6 which functions in Alzheimer's disease pathophysiology was significant by both measures and increased risk for glaucoma medication non-adherence by three-fold (95% CI, 1.62-5.8). Although our study was underpowered for genome-wide significance, SNP rs6474264 within ZMAT4 (p = 5.54 × 10-6) was found to be nominally significant, with a decreased risk for glaucoma medication non-adherence (OR, 0.22; 95% CI, 0.11-0.42)). IPA demonstrated significant overlap, utilizing, both standard measures including opioid signaling, drug metabolism, and synaptogenesis signaling. CREB signaling in neurons (which is associated with enhancing the baseline firing rate for the formation of long-term potentiation in nerve fibers) was shown to have protective associations. Our results suggest a substantial heritable genetic component to glaucoma medication non-adherence (47-58%). This finding is in line with genetic studies of other conditions with a psychiatric component (e.g., post-traumatic stress disorder (PTSD) or alcohol dependence). Our findings suggest both risk and protective statistically significant genes/pathways underlying glaucoma medication non-adherence for the first time. Further studies investigating more diverse populations with larger sample sizes are needed to validate these findings.

Placental Inflammation Significantly Correlates with Reduced Risk for Retinopathy of Prematurity.

Retinopathy of prematurity (ROP), a blinding condition affecting preterm infants, is an interruption of retinal vascular maturation that is incomplete when born preterm. Although ROP demonstrates delayed onset following preterm birth, representing a window for therapeutic intervention, there are no curative or preventative measures available for this condition. The in utero environment, including placental function, is increasingly recognized for contributions to preterm infant disease risk. The current study identified a protective association between acute placental inflammation and preterm infant ROP development using logistic regression, with the most significant association found for infants without gestational exposure to maternal preeclampsia and those with earlier preterm birth. Expression analysis of proteins with described ROP risk associations demonstrated significantly decreased placental high temperature requirement A serine peptidase-1 (HTRA-1) and fatty acid binding protein 4 protein expression in infants with acute placental inflammation compared with those without. Within the postnatal peripheral circulation, HTRA-1 and vascular endothelial growth factor-A demonstrated inverse longitudinal trends for infants born in the presence of, compared with absence of, acute placental inflammation. An agnostic approach, including whole transcriptome and differential methylation placental analysis, further identify novel mediators and pathways that may underly protection. Taken together, these data build on emerging literature showing a protective association between acute placental inflammation and ROP development and identify novel mechanisms that may inform postnatal risk associations in preterm infants.

Expression and subcellular localization of USH1C/harmonin in human retina provides insights into pathomechanisms and therapy.

Usher syndrome (USH) is the most common form of hereditary deaf-blindness in humans. USH is a complex genetic disorder, assigned to three clinical subtypes differing in onset, course and severity, with USH1 being the most severe. Rodent USH1 models do not reflect the ocular phenotype observed in human patients to date; hence, little is known about the pathophysiology of USH1 in the human eye. One of the USH1 genes, USH1C, exhibits extensive alternative splicing and encodes numerous harmonin protein isoforms that function as scaffolds for organizing the USH interactome. RNA-seq analysis of human retinae uncovered harmonin_a1 as the most abundant transcript of USH1C. Bulk RNA-seq analysis and immunoblotting showed abundant expression of harmonin in Müller glia cells (MGCs) and retinal neurons. Furthermore, harmonin was localized in the terminal endfeet and apical microvilli of MGCs, presynaptic region (pedicle) of cones and outer segments (OS) of rods as well as at adhesive junctions between MGCs and photoreceptor cells (PRCs) in the outer limiting membrane (OLM). Our data provide evidence for the interaction of harmonin with OLM molecules in PRCs and MGCs and rhodopsin in PRCs. Subcellular expression and colocalization of harmonin correlate with the clinical phenotype observed in USH1C patients. We also demonstrate that primary cilia defects in USH1C patient-derived fibroblasts could be reverted by the delivery of harmonin_a1 transcript isoform. Our studies thus provide novel insights into PRC cell biology, USH1C pathophysiology and development of gene therapy treatment(s).

Widespread exposure to Crimean-Congo haemorrhagic fever in Uganda might be driven by transmission from Rhipicephalus ticks: Evidence from cross-sectional and modelling studies.

BACKGROUND: Crimean-Congo haemorrhagic fever (CCHF) is a widespread tick-borne viral infection, present across Africa and Eurasia, which might pose a cryptic public health problem in Uganda. We aimed to understand the magnitude and distribution of CCHF risk in humans, livestock and ticks across Uganda by synthesising epidemiological (cross-sectional) and ecological (modelling) studies. METHODS: We conducted a cross-sectional study at three urban abattoirs receiving cattle from across Uganda. We sampled humans (n = 478), livestock (n = 419) and ticks (n = 1065) and used commercially-available kits to detect human and livestock CCHF virus (CCHFV) antibodies and antigen in tick pools. We developed boosted regression tree models to evaluate the correlates and geographical distribution of expected tick and wildlife hosts, and of human CCHF exposures, drawing on continent-wide data. FINDINGS: The cross-sectional study found CCHFV IgG/IgM seroprevalence in humans of 10·3% (7·8-13·3), with antibody detection positively associated with reported history of tick bite (age-adjusted odds ratio = 2·09 (1·09-3·98)). Cattle had a seroprevalence of 69·7% (65·1-73·4). Only one Hyalomma tick (CCHFV-negative) was found. However, CCHFV antigen was detected in Rhipicephalus (5·9% of 304 pools) and Amblyomma (2·9% of 34 pools) species. Modelling predicted high human CCHF risk across much of Uganda, low environmental suitability for Hyalomma, and high suitability for Rhipicephalus and Amblyomma. INTERPRETATION: Our epidemiological and ecological studies provide complementary evidence that CCHF exposure risk is widespread across Uganda. We challenge the idea that Hyalomma ticks are consistently the principal reservoir and vector for CCHFV, and postulate that Rhipicephalus might be important for CCHFV transmission in Uganda, due to high frequency of infected ticks and predicted environmental suitability. FUNDING: UCL Global Challenges Research Fund (GCRF) and Pan-African Network on Emerging and Re-Emerging Infections (PANDORA-ID-NET) funded by the European and Developing Countries Clinical Trials Partnership (EDCTP) under the EU Horizon 2020 Framework Programme for Research and Innovation.

Human Herpesvirus-6 Detection in Cerebrospinal Fluid on the BioFire FilmArray Meningitis/Encephalitis Panel in a High Human Immunodeficiency Virus-Prevalence African Setting.

The prevalence and clinical relevance of human herpesvirus-6 (HHV-6) detection in cerebrospinal fluid (CSF) using multiplex polymerase chain reaction (PCR) testing in patients with suspected meningoencephalitis in high human immunodeficiency virus-prevalence African settings are not known. We describe the clinical and laboratory characteristics of 13 patients with HHV-6 CSF PCR positivity in Botswana.

AMD Genomics: Non-Coding RNAs as Biomarkers and Therapeutic Targets.

Age-related macular degeneration (AMD) is a progressive neurodegenerative disease that is the world's leading cause of blindness in the aging population. Although the clinical stages and forms of AMD have been elucidated, more specific prognostic tools are required to determine when patients with early and intermediate AMD will progress into the advanced stages of AMD. Another challenge in the field has been the appropriate development of therapies for intermediate AMD and advanced atrophic AMD. After numerous negative clinical trials, an anti-C5 agent and anti-C3 agent have recently shown promising results in phase 3 clinical trials, in terms of slowing the growth of geographic atrophy, an advanced form of AMD. Interestingly, both drugs appear to be associated with an increased incidence of wet AMD, another advanced form of the disease, and will require frequent intravitreal injections. Certainly, there remains a need for other therapeutic agents with the potential to prevent progression to advanced stages of the disease. Investigation of the role and clinical utility of non-coding RNAs (ncRNAs) is a major advancement in biology that has only been minimally applied to AMD. In the following review, we discuss the clinical relevance of ncRNAs in AMD as both biomarkers and therapeutic targets.

Artificial Intelligence, Heuristic Biases, and the Optimization of Health Outcomes: Cautionary Optimism.

The use of artificial intelligence (AI) and machine learning (ML) in clinical care offers great promise to improve patient health outcomes and reduce health inequity across patient populations. However, inherent biases in these applications, and the subsequent potential risk of harm can limit current use. Multi-modal workflows designed to minimize these limitations in the development, implementation, and evaluation of ML systems in real-world settings are needed to improve efficacy while reducing bias and the risk of potential harms. Comprehensive consideration of rapidly evolving AI technologies and the inherent risks of bias, the expanding volume and nature of data sources, and the evolving regulatory landscapes, can contribute meaningfully to the development of AI-enhanced clinical decision making and the reduction in health inequity.

Implication of N-Methyl-d-Aspartate Receptor in Homocysteine-Induced Age-Related Macular Degeneration.

Age-related macular degeneration (AMD) is a leading cause of vision loss. Elevated homocysteine (Hcy) (Hyperhomocysteinemia) (HHcy) has been reported in AMD. We previously reported that HHcy induces AMD-like features. This study suggests that N-Methyl-d-aspartate receptor (NMDAR) activation in the retinal pigment epithelium (RPE) is a mechanism for HHcy-induced AMD. Serum Hcy and cystathionine-ß-synthase (CBS) were assessed by ELISA. The involvement of NMDAR in Hcy-induced AMD features was evaluated (1) in vitro using ARPE-19 cells, primary RPE isolated from HHcy mice (CBS), and mouse choroidal endothelial cells (MCEC); (2) in vivo using wild-type mice and mice deficient in RPE NMDAR (NMDARR-/-) with/without Hcy injection. Isolectin-B4, Ki67, HIF-1α, VEGF, NMDAR1, and albumin were assessed by immunofluorescence (IF), Western blot (WB), Optical coherence tomography (OCT), and fluorescein angiography (FA) to evaluate retinal structure, fluorescein leakage, and choroidal neovascularization (CNV). A neovascular AMD patient's serum showed a significant increase in Hcy and a decrease in CBS. Hcy significantly increased HIF-1α, VEGF, and NMDAR in RPE cells, and Ki67 in MCEC. Hcy-injected WT mice showed disrupted retina and CNV. Knocking down RPE NMDAR improved retinal structure and CNV. Our findings underscore the role of RPE NMDAR in Hcy-induced AMD features; thus, NMDAR inhibition could serve as a promising therapeutic target for AMD.

Pseudoexfoliation and Cataract Syndrome Associated with Genetic and Epidemiological Factors in a Mayan Cohort of Guatemala.

The Mayan population of Guatemala is understudied within eye and vision research. Studying an observational homogenous, geographically isolated population of individuals seeking eye care may identify unique clinical, demographic, environmental and genetic risk factors for blinding eye disease that can inform targeted and effective screening strategies to achieve better and improved health care distribution. This study served to: (a) identify the ocular health needs within this population; and (b) identify any possible modifiable risk factors contributing to disease pathophysiology within this population. We conducted a cross-sectional study with 126 participants. Each participant completed a comprehensive eye examination, provided a blood sample for genetic analysis, and received a structured core baseline interview for a standardized epidemiological questionnaire at the Salama Lions Club Eye Hospital in Salama, Guatemala. Interpreters were available for translation to the patients' native dialect, to assist participants during their visit. We performed a genome-wide association study for ocular disease association on the blood samples using Illumina's HumanOmni2.5-8 chip to examine single nucleotide polymorphism SNPs in this population. After implementing quality control measures, we performed adjusted logistic regression analysis to determine which genetic and epidemiological factors were associated with eye disease. We found that the most prevalent eye conditions were cataracts (54.8%) followed by pseudoexfoliation syndrome (PXF) (24.6%). The population with both conditions was 22.2%. In our epidemiological analysis, we found that eye disease was significantly associated with advanced age. Cataracts were significantly more common among those living in the 10 districts with the least resources. Furthermore, having cataracts was associated with a greater likelihood of PXF after adjusting for both age and sex. In our genetic analysis, the SNP most nominally significantly associated with PXF lay within the gene KSR2 (p < 1 × 10-5). Several SNPs were associated with cataracts at genome-wide significance after adjusting for covariates (p < 5 × 10-8). About seventy five percent of the 33 cataract-associated SNPs lie within 13 genes, with the majority of genes having only one significant SNP (5 × 10-8). Using bioinformatic tools including PhenGenI, the Ensembl genome browser and literature review, these SNPs and genes have not previously been associated with PXF or cataracts, separately or in combination. This study can aid in understanding the prevalence of eye conditions in this population to better help inform public health planning and the delivery of quality, accessible, and relevant health and preventative care within Salama, Guatemala.

Deepfakes in Ophthalmology: Applications and Realism of Synthetic Retinal Images from Generative Adversarial Networks.

Purpose: Generative adversarial networks (GANs) are deep learning (DL) models that can create and modify realistic-appearing synthetic images, or deepfakes, from real images. The purpose of our study was to evaluate the ability of experts to discern synthesized retinal fundus images from real fundus images and to review the current uses and limitations of GANs in ophthalmology. Design: Development and expert evaluation of a GAN and an informal review of the literature. Participants: A total of 4282 image pairs of fundus images and retinal vessel maps acquired from a multicenter ROP screening program. Methods: Pix2Pix HD, a high-resolution GAN, was first trained and validated on fundus and vessel map image pairs and subsequently used to generate 880 images from a held-out test set. Fifty synthetic images from this test set and 50 different real images were presented to 4 expert ROP ophthalmologists using a custom online system for evaluation of whether the images were real or synthetic. Literature was reviewed on PubMed and Google Scholars using combinations of the terms ophthalmology, GANs, generative adversarial networks, ophthalmology, images, deepfakes, and synthetic. Ancestor search was performed to broaden results. Main Outcome Measures: Expert ability to discern real versus synthetic images was evaluated using percent accuracy. Statistical significance was evaluated using a Fisher exact test, with P values ≤ 0.05 thresholded for significance. Results: The expert majority correctly identified 59% of images as being real or synthetic (P = 0.1). Experts 1 to 4 correctly identified 54%, 58%, 49%, and 61% of images (P = 0.505, 0.158, 1.000, and 0.043, respectively). These results suggest that the majority of experts could not discern between real and synthetic images. Additionally, we identified 20 implementations of GANs in the ophthalmology literature, with applications in a variety of imaging modalities and ophthalmic diseases. Conclusions: Generative adversarial networks can create synthetic fundus images that are indiscernible from real fundus images by expert ROP ophthalmologists. Synthetic images may improve dataset augmentation for DL, may be used in trainee education, and may have implications for patient privacy.