Decoding and reconstructing disease relations between dry eye and depression: a multimodal investigation comprising meta-analysis, genetic pathways and Mendelian randomization.

INTRODUCTION: The clinical presentations of dry eye disease (DED) and depression (DEP) often comanifest. However, the robustness and the mechanisms underlying this association were undetermined. OBJECTIVES: To this end, we set up a three-segment study that employed multimodality results (meta-analysis, genome-wide association study [GWAS] and Mendelian randomization [MR]) to elucidate the association, common pathways and causality between DED and DEP. METHODS: A meta-analysis comprising 26 case-control studies was first conducted to confirm the DED-DEP association. Next, we performed a linkage disequilibrium (LD)-adjusted GWAS and targeted phenotype association study (PheWAS) in East Asian TW Biobank (TWB) and European UK Biobank (UKB) populations. Single-nucleotide polymorphisms (SNPs) were further screened for molecular interactions and common pathways at the functional gene level. To further elucidate the activated pathways in DED and DEP, a systemic transcriptome review was conducted on RNA sequencing samples from the Gene Expression Omnibus. Finally, 48 MR experiments were implemented to examine the bidirectional causation between DED and DEP. RESULTS: Our meta-analysis showed that DED patients are associated with an increased DEP prevalence (OR = 1.83), while DEP patients have a concurrent higher risk of DED (OR = 2.34). Notably, cross-disease GWAS analysis revealed that similar genetic architecture (rG = 0.19) and pleiotropic functional genes contributed to phenotypes in both diseases. Through protein-protein interaction and ontology convergence, we summarized the pleiotropic functional genes under the ontology of immune activation, which was further validated by a transcriptome systemic review. Importantly, the inverse variance-weighted (IVW)-MR experiments in both TWB and UKB populations (p value <0.001) supported the bidirectional exposure-outcome causation for DED-to-DEP and DEP-to-DED. Despite stringent LD-corrected instrumental variable re-selection, the bidirectional causation between DED and DEP remained. CONCLUSION: With the multi-modal evidence combined, we consolidated the association and causation between DED and DEP.

YTHDF2-mediated regulations bifurcate BHPF-induced programmed cell deaths.

N6-methyladenosine (m6A) is a critical regulator in the fate of RNA, but whether and how m6A executes its functions in different tissues remains largely obscure. Here we report downregulation of a crucial m6A reader, YTHDF2, leading to tissue-specific programmed cell deaths (PCDs) upon fluorene-9-bisphenol (BHPF) exposure. Currently, Bisphenol A (BPA) substitutes are widely used in plastic manufacturing. Interrogating eight common BPA substitutes, we detected BHPF in 14% serum samples of pregnant participants. In a zebrafish model, BHPF caused tissue-specific PCDs triggering cardiac and vascular defects. Mechanistically, BHPF-mediated downregulation of YTHDF2 reduced YTHDF2-facilitated translation of m6A-gch1 for cardiomyocyte ferroptosis, and decreased YTHDF2-mediated m6A-sting1 decay for caudal vein plexus (CVP) apoptosis. The two distinct YTHDF2-mediated m6A regulations and context-dependent co-expression patterns of gch1/ythdf2 and tnfrsf1a/ythdf2 contributed to YTHDF2-mediated tissue-specific PCDs, uncovering a new layer of PCD regulation. Since BHPF/YTHDF2-medaited PCD defects were also observed in mammals, BHPF exposure represents a potential health threat.

Pluripotent Stem Cells in Clinical Cell Transplantation: Focusing on Induced Pluripotent Stem Cell-Derived RPE Cell Therapy in Age-Related Macular Degeneration.

Human pluripotent stem cells (PSCs), including both embryonic stem cells (ESCs) and induced pluripotent stem cells (iPSCs), represent valuable cell sources to replace diseased or injured tissues in regenerative medicine. iPSCs exhibit the potential for indefinite self-renewal and differentiation into various cell types and can be reprogrammed from somatic tissue that can be easily obtained, paving the way for cell therapy, regenerative medicine, and personalized medicine. Cell therapies using various iPSC-derived cell types are now evolving rapidly for the treatment of clinical diseases, including Parkinson's disease, hematological diseases, cardiomyopathy, osteoarthritis, and retinal diseases. Since the first interventional clinical trial with autologous iPSC-derived retinal pigment epithelial cells (RPEs) for the treatment of age-related macular degeneration (AMD) was accomplished in Japan, several preclinical trials using iPSC suspensions or monolayers have been launched, or are ongoing or completed. The evolution and generation of human leukocyte antigen (HLA)-universal iPSCs may facilitate the clinical application of iPSC-based therapies. Thus, iPSCs hold great promise in the treatment of multiple retinal diseases. The efficacy and adverse effects of iPSC-based retinal therapies should be carefully assessed in ongoing and further clinical trials.

Deep learning to infer visual acuity from optical coherence tomography in diabetic macular edema.

Purpose: Diabetic macular edema (DME) is one of the leading causes of visual impairment in diabetic retinopathy (DR). Physicians rely on optical coherence tomography (OCT) and baseline visual acuity (VA) to tailor therapeutic regimen. However, best-corrected visual acuity (BCVA) from chart-based examinations may not wholly reflect DME status. Chart-based examinations are subjected findings dependent on the patient's recognition functions and are often confounded by concurrent corneal, lens, retinal, optic nerve, or extraocular disorders. The ability to infer VA from objective optical coherence tomography (OCT) images provides the predicted VA from objective macular structures directly and a better understanding of diabetic macular health. Deviations from chart-based and artificial intelligence (AI) image-based VA will prompt physicians to assess other ocular abnormalities affecting the patients VA and whether pursuing anti-VEGF treatment will likely yield increment in VA. Materials and methods: We enrolled a retrospective cohort of 251 DME patients from Big Data Center (BDC) of Taipei Veteran General Hospital (TVGH) from February 2011 and August 2019. A total of 3,920 OCT images, labeled as "visually impaired" or "adequate" according to baseline VA, were grouped into training (2,826), validation (779), and testing cohort (315). We applied confusion matrix and receiver operating characteristic (ROC) curve to evaluate the performance. Results: We developed an OCT-based convolutional neuronal network (CNN) model that could classify two VA classes by the threshold of 0.50 (decimal notation) with an accuracy of 75.9%, a sensitivity of 78.9%, and an area under the ROC curve of 80.1% on the testing cohort. Conclusion: This study demonstrated the feasibility of inferring VA from routine objective retinal images. Translational relevance: Serves as a pilot study to encourage further use of deep learning in deriving functional outcomes and secondary surrogate endpoints for retinal diseases.

m6A modification confers thermal vulnerability to HPV E7 oncotranscripts via reverse regulation of its reader protein IGF2BP1 upon heat stress.

Human papillomavirus (HPV)-induced carcinogenesis critically depends on the viral early protein 7 (E7), making E7 an attractive therapeutic target. Here, we report that the E7 messenger RNA (mRNA)-containing oncotranscript complex can be selectively targeted by heat treatment. In HPV-infected cells, viral E7 mRNA is modified by N6-methyladenosine (m6A) and stabilized by IGF2BP1, a cellular m6A reader. Heat treatment downregulates E7 mRNA and protein by destabilizing IGF2BP1 without the involvement of canonical heat-shock proteins and reverses HPV-associated carcinogenesis in vitro and in vivo. Mechanistically, heat treatment promotes IGF2BP1 aggregation only in the presence of m6A-modified E7 mRNA to form distinct heat-induced m6A E7 mRNA-IGF2BP1 granules, which are resolved by the ubiquitin-proteasome system. Collectively, our results not only show a mutual regulation between m6A RNA and its reader but also provide a heat-treatment-based therapeutic strategy for HPV-associated malignancies by specifically downregulating E7 mRNA-IGF2BP1 oncogenic complex.

Genetics behind Cerebral Disease with Ocular Comorbidity: Finding Parallels between the Brain and Eye Molecular Pathology.

Cerebral visual impairments (CVIs) is an umbrella term that categorizes miscellaneous visual defects with parallel genetic brain disorders. While the manifestations of CVIs are diverse and ambiguous, molecular diagnostics stand out as a powerful approach for understanding pathomechanisms in CVIs. Nevertheless, the characterization of CVI disease cohorts has been fragmented and lacks integration. By revisiting the genome-wide and phenome-wide association studies (GWAS and PheWAS), we clustered a handful of renowned CVIs into five ontology groups, namely ciliopathies (Joubert syndrome, Bardet-Biedl syndrome, Alstrom syndrome), demyelination diseases (multiple sclerosis, Alexander disease, Pelizaeus-Merzbacher disease), transcriptional deregulation diseases (Mowat-Wilson disease, Pitt-Hopkins disease, Rett syndrome, Cockayne syndrome, X-linked alpha-thalassaemia mental retardation), compromised peroxisome disorders (Zellweger spectrum disorder, Refsum disease), and channelopathies (neuromyelitis optica spectrum disorder), and reviewed several mutation hotspots currently found to be associated with the CVIs. Moreover, we discussed the common manifestations in the brain and the eye, and collated animal study findings to discuss plausible gene editing strategies for future CVI correction.

The Post-Translational Modification Networking in WNK-Centric Hypertension Regulation and Electrolyte Homeostasis.

The with-no-lysine (WNK) kinase family, comprising four serine-threonine protein kinases (WNK1-4), were first linked to hypertension due to their mutations in association with pseudohypoaldosteronism type II (PHAII). WNK kinases regulate crucial blood pressure regulators, SPAK/OSR1, to mediate the post-translational modifications (PTMs) of their downstream ion channel substrates, such as sodium chloride co-transporter (NCC), epithelial sodium chloride (ENaC), renal outer medullary potassium channel (ROMK), and Na/K/2Cl co-transporters (NKCCs). In this review, we summarize the molecular pathways dysregulating the WNKs and their downstream target renal ion transporters. We summarize each of the genetic variants of WNK kinases and the small molecule inhibitors that have been discovered to regulate blood pressure via WNK-triggered PTM cascades.

Efficacy of a Novel Injection Lipolysis to Induce Targeted Adipocyte Apoptosis: A Randomized, Phase IIa Study of CBL-514 Injection on Abdominal Subcutaneous Fat Reduction.

BACKGROUND: CBL-514 is a novel injectable drug that may be safe and efficacious for localized abdominal subcutaneous fat reduction. OBJECTIVES: The aim of this study was to assess the safety and efficacy of CBL-514 in reducing abdominal subcutaneous fat volume and thickness. METHODS: This Phase IIa, open-label, random allocation study consisted of a 6-week treatment period and follow-up at 4 and 8 weeks following the last treatment. Participants were randomly allocated to receive 1.2 mg/cm2 (180 mg), 1.6 mg/cm2 (240 mg), or 2.0 mg/cm2 (300 mg) of CBL-514 with up to 4 treatments, each comprising 60 injections into the abdominal adipose layer. Changes in abdominal subcutaneous fat were assessed by ultrasound at follow-up visits. Treatment-emergent adverse events were recorded. RESULTS: Higher doses of CBL-514 (unit dose, 2.0 and 1.6 mg/cm2) significantly improved the absolute and percentage reduction in abdominal fat volume (P < 0.00001) and thickness (P < 0.0001) compared with baseline. Although the COVID-19 pandemic halted some participant recruitment and follow-ups, analysis was unaffected, even after sample size limitations. CONCLUSIONS: CBL-514 injection at multiple doses up to 300 mg with a unit dose of 2.0 mg/cm2 is safe, well-tolerated, and reduced abdominal fat volume and thickness by inducing adipocyte apoptosis. Although other procedures exist to treat abdominal fat, they have limitations and may cause complications. At a dose of 2.0 mg/cm2, CBL-514 safely and significantly reduced abdominal fat volume by 24.96%, making it a promising new treatment for routine, nonsurgical abdominal fat reduction in dermatologic clinics.

Hyperthermia Selectively Destabilizes Oncogenic Fusion Proteins.

The PML/RARα fusion protein is the oncogenic driver in acute promyelocytic leukemia (APL). Although most APL cases are cured by PML/RARα-targeting therapy, relapse and resistance can occur due to drug-resistant mutations. Here we report that thermal stress destabilizes the PML/RARα protein, including clinically identified drug-resistant mutants. AML1/ETO and TEL/AML1 oncofusions show similar heat shock susceptibility. Mechanistically, mild hyperthermia stimulates aggregation of PML/RARα in complex with nuclear receptor corepressors leading to ubiquitin-mediated degradation via the SIAH2 E3 ligase. Hyperthermia and arsenic therapy destabilize PML/RARα via distinct mechanisms and are synergistic in primary patient samples and in vivo, including three refractory APL cases. Collectively, our results suggest that by taking advantage of a biophysical vulnerability of PML/RARα, thermal therapy may improve prognosis in drug-resistant or otherwise refractory APL. These findings serve as a paradigm for therapeutic targeting of fusion oncoprotein-associated cancers by hyperthermia. SIGNIFICANCE: Hyperthermia destabilizes oncofusion proteins including PML/RARα and acts synergistically with standard arsenic therapy in relapsed and refractory APL. The results open up the possibility that heat shock sensitivity may be an easily targetable vulnerability of oncofusion-driven cancers.See related commentary by Wu et al., p. 300.

Application of artificial intelligence-driven endoscopic screening and diagnosis of gastric cancer.

The landscape of gastrointestinal endoscopy continues to evolve as new technologies and techniques become available. The advent of image-enhanced and magnifying endoscopies has highlighted the step toward perfecting endoscopic screening and diagnosis of gastric lesions. Simultaneously, with the development of convolutional neural network, artificial intelligence (AI) has made unprecedented breakthroughs in medical imaging, including the ongoing trials of computer-aided detection of colorectal polyps and gastrointestinal bleeding. In the past demi-decade, applications of AI systems in gastric cancer have also emerged. With AI's efficient computational power and learning capacities, endoscopists can improve their diagnostic accuracies and avoid the missing or mischaracterization of gastric neoplastic changes. So far, several AI systems that incorporated both traditional and novel endoscopy technologies have been developed for various purposes, with most systems achieving an accuracy of more than 80%. However, their feasibility, effectiveness, and safety in clinical practice remain to be seen as there have been no clinical trials yet. Nonetheless, AI-assisted endoscopies shed light on more accurate and sensitive ways for early detection, treatment guidance and prognosis prediction of gastric lesions. This review summarizes the current status of various AI applications in gastric cancer and pinpoints directions for future research and clinical practice implementation from a clinical perspective.

An Update on Gene Therapy for Inherited Retinal Dystrophy: Experience in Leber Congenital Amaurosis Clinical Trials.

Inherited retinal dystrophies (IRDs) are a group of rare eye diseases caused by gene mutations that result in the degradation of cone and rod photoreceptors or the retinal pigment epithelium. Retinal degradation progress is often irreversible, with clinical manifestations including color or night blindness, peripheral visual defects and subsequent vision loss. Thus, gene therapies that restore functional retinal proteins by either replenishing unmutated genes or truncating mutated genes are needed. Coincidentally, the eye's accessibility and immune-privileged status along with major advances in gene identification and gene delivery systems heralded gene therapies for IRDs. Among these clinical trials, voretigene neparvovec-rzyl (Luxturna), an adeno-associated virus vector-based gene therapy drug, was approved by the FDA for treating patients with confirmed biallelic RPE65 mutation-associated Leber Congenital Amaurosis (LCA) in 2017. This review includes current IRD gene therapy clinical trials and further summarizes preclinical studies and therapeutic strategies for LCA, including adeno-associated virus-based gene augmentation therapy, 11-cis-retinal replacement, RNA-based antisense oligonucleotide therapy and CRISPR-Cas9 gene-editing therapy. Understanding the gene therapy development for LCA may accelerate and predict the potential hurdles of future therapeutics translation. It may also serve as the template for the research and development of treatment for other IRDs.

Systematic review and meta-analysis of the effectiveness and safety of hydroxychloroquine in treating COVID-19 patients.

BACKGROUND: Since COVID-19 outbreak, hydroxychloroquine (HCQ) has been tested for effective therapies, and the relevant researches have shown controversial results. METHODS: Systematic review and meta-analysis were conducted after a thorough search of relevant studies from databases. Trials that have evaluated HCQ for COVID-19 treatment were recruited for statistical analysis with fixed- and random-effect models. RESULTS: Nine trials involving 4112 patients were included in present meta-analysis. It was seen that HCQ-azithromycin (HCQ-AZI) combination regimen increased the mortality rate in COVID-19 (odds ratio [OR], 2.34; 95% confidence interval [CI], 1.63-3.36) patients; however, it also showed benefits associated with the viral clearance in patients (OR, 27.18; 95% CI, 1.29-574.32). HCQ-alone when used as a therapy in COVID-19 did not reveal significant changes in mortality rate, clinical progression, viral clearance, and cardiac QT prolongation. Subsequent subgroup analysis showed that HCQ treatment could decrease mortality rate and progression to severe illness in severely infected COVID-19 patients (OR, 0.27; 95% CI, 0.13-0.58). A lower risk of mortality rate was also noted in the stratified group of >14 days follow-up period (OR, 0.27; 95% CI, 0.13-0.58) compared to ≤14 days follow-up period group that conversely showed an increased mortality rate (OR, 2.09; 95% CI, 1.41-3.10). CONCLUSION: Our results indicated that HCQ-AZI combination treatment increased mortality rate in patients with COVID-19, but it also showed benefits associated with viral clearance in patients. HCQ-alone used for treatment has revealed benefits in decreasing the mortality rate among severely infected COVID-19 group and showed potential to be used for COVID-19 treatment in long-term follow-up period group. Accordingly, more rigorous, large-scale, and long follow-up period studies in patients with COVID-19 are needed.

Concordance analysis of intrapartum cardiotocography between physicians and artificial intelligence-based technique using modified one-dimensional fully convolutional networks.

BACKGROUND: Cardiotocography is a common method of electronic fetal monitoring (EFM) for fetal well-being. Data-driven analyses have shown potential for automated EFM assessment. For this preliminary study, we used a novel artificial intelligence method based on fully convolutional networks (FCNs), with deep learning for EFM evaluation and correct recognition, and its possible role in evaluation of nonreassuring fetal status. METHODS: We retrospectively collected 3239 EFM labor records from 292 deliveries and neonatal Apgar scores between December 2018 and July 2019 at a single medical center. We analyzed these data using an FCN model and compared the results with clinical practice. RESULTS: The FCN model recognized EFM traces like physicians, with an average Cohen's kappa coefficient of agreement of 0.525 and average area under the receiver operating characteristic curve of 0.892 for six fetal heart rate (FHR) categories. The FCN model showed higher sensitivity for predicting fetal compromise (0.528 vs 0.132) but a higher false-positive rate (0.632 vs 0.012) compared with clinical practice. CONCLUSION: FCN is a modern technique that may be useful for EFM trace recognition based on its multiconvolutional layered analysis. Our model showed a competitive ability to identify FHR patterns and the potential for evaluation of nonreassuring fetal status.

Clinical manifestation and disease progression in COVID-19 infection.

Coronavirus disease 2019 (COVID-19) is mainly an infectious disease of the respiratory system transmitted through air droplets, and pulmonary symptoms constitute main presentations of this disease. However, COVID-19 demonstrates a clinically diverse manifestation ranging from asymptomatic presentation to critically illness with severe pneumonia, acute respiratory distress syndrome, respiratory failure, or multiple organ failure. Accumulating evidences demonstrated that COVID-19 has extrapulmonary involvement, including neurological, smelling sensation, cardiovascular, digestive, hepatobiliary, renal, endocrinologic, dermatologic system, and others. Over a third of COVID-19 patients manifest a wide range of neurological symptoms involving the central/peripheral nervous system. Underlying cardiovascular comorbidities were associated with detrimental outcomes, meanwhile the occurrence of cardiovascular complications correlate to poor survival. Gastrointestinal symptoms frequently occur and have been associated with a longer period of illness. Impaired hepatic functions were associated with the severity of the disease. Higher rate of acute kidney injury was reported in critically ill patients with COVID-19. Endocrinologic presentations of COVID-19 include exacerbating hyperglycemia, euglycemic ketosis, and diabetic ketoacidosis. The most common cutaneous manifestation was acro-cutaneous (pernio or chilblain-like) lesions, and other skin lesions consist of maculopapular rash, vesicular lesions, livedoid/necrotic lesions, exanthematous rashes, and petechiae. This review article summarized the general clinical signs and symptoms, radiologic features, and disease manifestation with progression in patients with COVID-19.

Current Genetic Survey and Potential Gene-Targeting Therapeutics for Neuromuscular Diseases.

Neuromuscular diseases (NMDs) belong to a class of functional impairments that cause dysfunctions of the motor neuron-muscle functional axis components. Inherited monogenic neuromuscular disorders encompass both muscular dystrophies and motor neuron diseases. Understanding of their causative genetic defects and pathological genetic mechanisms has led to the unprecedented clinical translation of genetic therapies. Challenged by a broad range of gene defect types, researchers have developed different approaches to tackle mutations by hijacking the cellular gene expression machinery to minimize the mutational damage and produce the functional target proteins. Such manipulations may be directed to any point of the gene expression axis, such as classical gene augmentation, modulating premature termination codon ribosomal bypass, splicing modification of pre-mRNA, etc. With the soar of the CRISPR-based gene editing systems, researchers now gravitate toward genome surgery in tackling NMDs by directly correcting the mutational defects at the genome level and expanding the scope of targetable NMDs. In this article, we will review the current development of gene therapy and focus on NMDs that are available in published reports, including Duchenne Muscular Dystrophy (DMD), Becker muscular dystrophy (BMD), X-linked myotubular myopathy (XLMTM), Spinal Muscular Atrophy (SMA), and Limb-girdle muscular dystrophy Type 2C (LGMD2C).

The human mitochondrial 12S rRNA m4C methyltransferase METTL15 is required for mitochondrial function.

Mitochondrial DNA gene expression is coordinately regulated both pre- and post-transcriptionally, and its perturbation can lead to human pathologies. Mitochondrial rRNAs (mt-rRNAs) undergo a series of nucleotide modifications after release from polycistronic mitochondrial RNA precursors, which is essential for mitochondrial ribosomal biogenesis. Cytosine N4-methylation (m4C) at position 839 (m4C839) of the 12S small subunit mt-rRNA was identified decades ago; however, its biogenesis and function have not been elucidated in detail. Here, using several approaches, including immunofluorescence, RNA immunoprecipitation and methylation assays, and bisulfite mapping, we demonstrate that human methyltransferase-like 15 (METTL15), encoded by a nuclear gene, is responsible for 12S mt-rRNA methylation at m4C839 both in vivo and in vitro We tracked the evolutionary history of RNA m4C methyltransferases and identified a difference in substrate preference between METTL15 and its bacterial ortholog rsmH. Additionally, unlike the very modest impact of a loss of m4C methylation in bacterial small subunit rRNA on the ribosome, we found that METTL15 depletion results in impaired translation of mitochondrial protein-coding mRNAs and decreases mitochondrial respiration capacity. Our findings reveal that human METTL15 is required for mitochondrial function, delineate the evolution of methyltransferase substrate specificities and modification patterns in rRNA, and highlight a differential impact of m4C methylation on prokaryotic ribosomes and eukaryotic mitochondrial ribosomes.

Precaution and prevention of coronavirus disease 2019 infection in the eye.

Although current studies suggested that conjunctivitis is not a common presentation of coronavirus disease 2019 (COVID-19), several studies have reported the presence of severe acute respiratory syndrome coronavirus-2 (SARS-CoV-2) in ocular secretions. Coronavirus had not yet been successfully cultured from tears or conjunctival swabs in humans, neither SARS-CoV-2 nor SARS-CoV. However, live feline coronavirus has been isolated from conjunctival swabs. In addition, infection of COVID-19 through unprotected eye exposure had been suspected in several articles. Reports of ophthalmologists and otolaryngologists died of COVID-19 also raised concern on ocular transmission. As a result, we strongly suggest that personal protective equipment (PPE) should cover the mouth, nose, and eyes of ophthalmologists, especially when conjunctivitis caused by SARS-CoV-2 is clinically indistinguishable from other viral follicular conjunctivitis.

The roles of m6A RNA modifiers in human cancer.

Like DNA and proteins, RNA is subject to numerous (over 160) covalent modifications which play critical roles to regulate RNA metabolism. Among these modifications, N-methyladenosine (mA) is the most prevalent RNA methylation on mRNA which occurs on around 25% of transcripts. The recent studies demonstrated that mA participates in many aspects of RNA processing, including splicing, nuclear exporting, translation, stabilization, etc. Therefore, it revealed a new layer of regulatory mechanism for gene expression and has been termed "RNA Epigenetics" or "Epitranscriptomics". RNA mA is regulated and exerts its functions by three groups of "mA RNA modifiers" including mA methyltransferases (writers), mA demethylases (erasers), and mA binding proteins (readers). In this review, we would summarize and discuss the current understandings of the roles of the conventional mA RNA modifiers in human cancers.

Artificial intelligence-based decision-making for age-related macular degeneration.

Artificial intelligence (AI) based on convolutional neural networks (CNNs) has a great potential to enhance medical workflow and improve health care quality. Of particular interest is practical implementation of such AI-based software as a cloud-based tool aimed for telemedicine, the practice of providing medical care from a distance using electronic interfaces. Methods: In this study, we used a dataset of labeled 35,900 optical coherence tomography (OCT) images obtained from age-related macular degeneration (AMD) patients and used them to train three types of CNNs to perform AMD diagnosis. Results: Here, we present an AI- and cloud-based telemedicine interaction tool for diagnosis and proposed treatment of AMD. Through deep learning process based on the analysis of preprocessed optical coherence tomography (OCT) imaging data, our AI-based system achieved the same image discrimination rate as that of retinal specialists in our hospital. The AI platform's detection accuracy was generally higher than 90% and was significantly superior (p < 0.001) to that of medical students (69.4% and 68.9%) and equal (p = 0.99) to that of retinal specialists (92.73% and 91.90%). Furthermore, it provided appropriate treatment recommendations comparable to those of retinal specialists. Conclusions: We therefore developed a website for realistic cloud computing based on this AI platform, available at https://www.ym.edu.tw/~AI-OCT/. Patients can upload their OCT images to the website to verify whether they have AMD and require treatment. Using an AI-based cloud service represents a real solution for medical imaging diagnostics and telemedicine.