Diabetic retinopathy research -
Zhou, B. Learning deep features for discriminative localization. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition , — Brown, C. Coefficient of variation. In Applied Multivariate Statistics in Geohydrology and Related Sciences , — Springer, The probable error of a mean.
Biometrika , 1—25 Falavarjani, K. Ultra-wide-field imaging in diabetic retinopathy. Kiss, S. Ultra widefield fundus imaging for diabetic retinopathy. Diabetes Rep. Hirano, T. Aiello, L. Comparison of early treatment diabetic retinopathy study standard 7-field imaging with ultrawide-field imaging for determining severity of diabetic retinopathy.
Kernt, M. Assessment of diabetic retinopathy using nonmydriatic ultra-widefield scanning laser ophthalmoscopy optomap compared with etdrs 7-field stereo photography. Diabetes Care 35 , — Silva, P.
Nonmydriatic ultrawide field retinal imaging compared with dilated standard 7-field mm photography and retinal specialist examination for evaluation of diabetic retinopathy.
Rasmussen, M. Comparison between early treatment diabetic retinopathy study 7-field retinal photos and non-mydriatic, mydriatic and mydriatic steered widefield scanning laser ophthalmoscopy for assessment of diabetic retinopathy.
Diabetes Complicat. Peripheral lesions identified on ultrawide field imaging predict increased risk of diabetic retinopathy progression over 4 years. Download references.
This research was supported by Basic Science Research Program through the National Research Foundation of Korea NRF funded by the Ministry of Education Grant number: NRFR1D1A1A Research Institute of Radiological Science, College of Medicine, Yonsei University, Seoul, , Republic of Korea.
Department of Ophthalmology, College of Medicine, Catholic Kwandong University, International St. Department of Ophthalmology, College of Medicine, Yonsei University, Seoul, , Republic of Korea. Department of Medical Humanities and Social Sciences, College of Medicine, Yonsei University, Seoul, , Republic of Korea.
You can also search for this author in PubMed Google Scholar. Correspondence to Sangchul Yoon. Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Open Access This article is licensed under a Creative Commons Attribution 4. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material.
If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.
Reprints and permissions. Oh, K. Early detection of diabetic retinopathy based on deep learning and ultra-wide-field fundus images. Sci Rep 11 , Download citation. Received : 19 October Accepted : 04 January Published : 21 January Anyone you share the following link with will be able to read this content:.
Sorry, a shareable link is not currently available for this article. Provided by the Springer Nature SharedIt content-sharing initiative. Archives of Computational Methods in Engineering By submitting a comment you agree to abide by our Terms and Community Guidelines.
If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate. Sign up for the Nature Briefing newsletter — what matters in science, free to your inbox daily. Skip to main content Thank you for visiting nature.
nature scientific reports articles article. Download PDF. Subjects Biomedical engineering Retinal diseases. Abstract Visually impaired and blind people due to diabetic retinopathy were 2.
Introduction Diabetic retinopathy DR is responsible for 0. Methods The proposed DR detection system requires an automatic segmentation of the ETDRS 7SF to remove undesirable components such as eyelashes and skin. Figure 1. An overview of the proposed DR detection system.
Full size image. Figure 2. The overall flow of the optic disc and macular detection process. Figure 3. Sample images at each processing stages.
Figure 4. The ETDRS 7SF image segmentation process and sample images with noise. Figure 5. The ResNet model. Results Evaluation protocols In our experiments, automated DR detection systems using the two segmentation images are assessed.
Detection performance assessment To provide a comprehensive detection performance throughout the overall range of decision thresholds, the ROC curves are plotted in Fig. Figure 6. Figure 7. Table 1 The sensitivity, specificity, accuracy, and AUC results of the DR detection system using ETDRS 7SF and F1—F2 fundus images.
Full size table. Figure 8. Class activation maps generated from ETDRS 7SF and F1—F2 images for DR and normal class. Table 2 Results from the paired-sample t test. Data availability The ultra-wide-field fundus image dataset utilized for training, validation, and test was acquired from Catholic Kwandong University International St.
Code availability The code for preprocessing and learning includes intellectual property and cannot be released publicly. References Leasher, J. Article Google Scholar International Diabetes Federation. Article Google Scholar Thomas, R. Article Google Scholar MacLennan, P.
Article Google Scholar Chou, C. Article Google Scholar Fogel, A. Article Google Scholar Gulshan, V. Article Google Scholar Abràmoff, M. Article Google Scholar Gargeya, R. Article Google Scholar Sayres, R.
Article Google Scholar Sahlsten, J. Article CAS ADS Google Scholar Gulshan, V. Article Google Scholar Ting, D. Article Google Scholar Takahashi, H. Article Google Scholar Kaines, A. Article Google Scholar Nagasawa, T. Article Google Scholar He, K. Google Scholar Ronneberger, O. Google Scholar Pizer, S.
Article ADS Google Scholar Keys, R. Article MathSciNet ADS Google Scholar Kingma, D. Article Google Scholar Kiss, S. Article Google Scholar Hirano, T. Article Google Scholar Aiello, L. Article Google Scholar Kernt, M. Article Google Scholar Silva, P.
Article Google Scholar Rasmussen, M. Article Google Scholar Download references. Acknowledgements This research was supported by Basic Science Research Program through the National Research Foundation of Korea NRF funded by the Ministry of Education Grant number: NRFR1D1A1A Author information Author notes These authors contributed equally: Kangrok Oh and Hae Min Kang.
Authors and Affiliations Research Institute of Radiological Science, College of Medicine, Yonsei University, Seoul, , Republic of Korea Kangrok Oh Department of Ophthalmology, College of Medicine, Catholic Kwandong University, International St.
View author publications. Ethics declarations Competing interests The authors declare no competing interests. Additional information Publisher's note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Rights and permissions Open Access This article is licensed under a Creative Commons Attribution 4. About this article. Cite this article Oh, K. Copy to clipboard. This article is cited by Machine learning based study for the classification of Type 2 diabetes mellitus subtypes Nelson E.
Ordoñez-Guillen Jose Luis Gonzalez-Compean Edwin Aldana-Bobadilla BioData Mining Modified residual networks for severity stage classification of diabetic retinopathy Nitigya Sambyal Poonam Saini Varun Gupta Evolving Systems Eisoc with ifodpso and dcnn classifier for diabetic retinopathy recognition system Neetha Merin Thomas S.
Albert Jerome Multimedia Tools and Applications TrachomaNet: Detection and grading of trachoma using texture feature based deep convolutional neural network Belesti Yenegeta Yaregal Assabie Multimedia Tools and Applications A Literature Review of Early-Stage Diabetic Retinopathy Detection Using Deep Learning and Evolutionary Computing Techniques Sachin Bhandari Sunil Pathak Sonal Amit Jain Archives of Computational Methods in Engineering Comments By submitting a comment you agree to abide by our Terms and Community Guidelines.
Though focal laser and PRP reduce the risk of vision loss and DR progression, there are several limitations to photocoagulation. Laser is primarily a destructive procedure, such that PRP may impair peripheral vision, and decrease night vision [ 18 ]. In addition, focal laser seldom actually improves visual acuity [ 19 ].
Thus, more effective treatments for the early and late stages of DR are needed. The paradigm of DR as a primarily vascular disease was pursued further. Several studies confirmed that neuro-inflammation plays a prominent role in the pathogenesis of DR [ 20 , 21 , 22 ].
Steroids, delivered intravitreally, are effective in improving vision in patients with diabetic macular edema, though they cause cataract development and increased intraocular pressure leading to glaucoma [ 23 , 24 ]. Intravitreal steroids have also been suggested to reduce the rate of progression of DR to proliferative disease as well [ 25 , 26 ].
Increased levels of inflammatory mediators ultimately lead to breakdown of the blood-retinal-barrier, increased vascular permeability, and angiogenesis via the release of cytokines and growth factors, including vascular endothelial growth factor VEGF [ 27 , 28 , 29 ].
The resulting new pharmacotherapeutic targets led to significant changes in the management of patients with DR, and the development of a new standard of care. Several clinical trials showed the efficacy of intravitreal anti-VEGF medication in the treatment of diabetic macular edema [ 30 , 31 , 32 ].
These studies confirmed that administration of repeated monthly intravitreal ranibizumab injections an anti-VEGF medication plus prompt or deferred laser yielded a modestly greater improvement in visual acuity from baseline than laser alone with minimal side effects, leading to the adoption of a new standard of care in the treatment of diabetic macular edema.
These results suggest that there are factors in addition to VEGF that likely mediate DR progression. A large randomized clinical trial showed that an average of nine intravitreal injections of ranibizumab was non-inferior to PRP at 2 years, with fewer patients who received ranibizumab requiring vitrectomy surgery over the study period [ 34 ].
Research on the role of inflammation in DR, and the subsequent increase in vascular permeability led to significant changes in the way patients have been managed in the past decade.
Anti-VEGF therapy for diabetic macular edema has been shown in multiple randomized clinical trials to be more effective at improving vision than laser, and several cost-effectiveness analyses have confirmed the value of these treatments to patients and society [ 35 , 36 ].
The efficacy of anti-VEGF treatment has also been proven in the treatment of proliferative DR and use of these medications in the management of this condition has increased, often as a first line treatment for complications of proliferative DR, such as vitreous hemorrhage.
Other factors, however, have prevented wide-spread adoption of the practice as standard of care. For optimum results, anti-VEGF medications demand frequent administration—potentially indefinitely—resulting in concerns about the overall cost of treatment and the increased burden placed on physicians and patients by the need for frequent, consistent follow up to maintain treatment gains.
Intravitreal anti-VEGF therapy does offer, though, a viable adjuvant or alternative treatment in many cases of proliferative disease [ 37 ].
Thus, we now have surgical laser and pharmacologic anti-VEGF options to treat DR and patients and many physicians tailor these approaches, using them separately or together, to optimize benefits and convenience. Figure 1 depicts the timeline of major advances in diabetic retinopathy research to date.
This significant progress in management of diabetic retinopathy was coupled with, and made possible by, important developments in ocular imaging [ 38 ].
Optical coherence tomography OCT is a non-invasive diagnostic test that is performed in the office, providing detailed cross-sectional anatomic images of the retina. In its widest application, OCT allows for early detection of anatomical changes in the macula, such as the development of thickening and cystic spaces noted in diabetic macular edema.
OCT testing is routinely used to clinically diagnose and manage patients with diabetic macular edema, and data on central retinal thickness from the OCT is used as an end-point in large clinical trials [ 39 ]. Additional imaging techniques such as ultra-wide-field fundus photography and angiography allow better visualization of the peripheral retina than conventional cameras, and better identification of areas of poor vascular perfusion [ 40 ].
These imaging modalities help with clinical management of patients, and provide further insight into structural changes in every stage of DR. The progress of the late 20 th and early 21 st centuries has been significant, allowing improved ability to delay development of DR, and to administer treatment that significantly reduces the risk of vision loss.
However, the large projected increase in prevalence of DR, coupled with the need for frequent administration of intravitreal injections clearly indicates a need for alternative options in the future. In addition, current management strategies are either preventative intensive glycemic and blood pressure control , or targeted towards advanced disease diabetic macular edema, or proliferative DR.
Yet, a growing body of literature suggests functional decline and associated health-related quality of life reductions in earlier stages of DR [ 41 ]. Visual dysfunction in the form of decreased sensitivity on visual field testing and diminished photoreceptor function as measured by electroretinogram have been reported prior to the development of vascular lesions [ 42 , 43 ].
Thus, the paradigm on DR has changed. Alterations in the neurosensory retina, undetectable by ophthalmoscopy, are recognized as important early contributors to visual decline, and it is now established that neurosensory degeneration may precede visible vascular changes, or occur alongside them [ 44 ].
That is, the entire neurovascular unit, comprised of vascular, glial, microglial and neuronal cells, is compromised by diabetes [ 45 ]. Current treatments of anti-VEGF and lasers address these late stages of disease, but only fenofibrate, blood pressure and metabolic control have shown demonstrable effects in the pre-failure stages.
Laboratory research confirms that metabolic pathways triggered by hyperglycemia, insulin deficiency [ 47 ] and dyslipidemia [ 48 ] lead to abnormalities in both the neural retina as well as the retinal capillary bed.
A better understanding of all the molecular players in these pathways has produced several potential pharmacotherapeutic targets for DR, which includes both the inhibition of mediators of neural damage, and enhancement of agents that may be neuroprotective.
Hernández et al. In addition to finding new targets to treat earlier stages of DR, research is being conducted using nanoparticles to allow for sustained delivery of drug, as well as alternative topical drug delivery systems. Nanotechnology is currently being applied to anti-VEGF medications, and several other new mediators of inflammation and angiogenesis.
Nanoparticles are particularly designed to cross the blood-retinal barrier, thereby allowing for better penetration into the retina [ 50 ]. These methods remain under development and clinical application appears to remain in the future for DR. If treating DR in earlier stages is to truly become a common clinical practice, though, new diagnostic techniques are needed to identify changes before they are visible on exam and to track response to treatment.
Circulating biomarkers and new imaging modalities are being investigated to use as clinical indicators and new endpoints for clinical trials. Inflammatory cytokines are most often reported as circulating biomarkers associated with early DR.
Three inflammatory mediators in particular: interleukin 6 IL-6 , tumor necrosis factor α TNF-α , and C-reactive protein CRP , when combined in a z-score, are associated with development of retinopathy, nephropathy and cardiovascular disease in diabetics [ 51 ].
Inflammatory and vasoactive mediators produced and measured locally, in intraocular fluids, have also been identified. An increase in glial fibrillary acidic protein GFAP , for example, is noted in the aqueous humor of patients with diabetes and no signs of DR or with non-proliferative DR compared to age-matched healthy controls [ 52 ].
As such, new imaging techniques combined with a better understanding of biomarkers is promising for developing better diagnostic tools for early disease. Frimmel et al. recently reported a technique in which imaging probes were developed to target a specific endothelial surface molecule known to participate in breakdown of the blood-retinal barrier in DR ICAM This probe allowed for visualization of the expression of ICAM-1 on the endothelial surface in vivo in rats.
Increased visibility of the probe was noted on imaging from diabetic animals compared to controls [ 53 ], suggesting that synergistic development of biomarkers and imaging technology will allow for detection of early DR in the future.
A projected tripling in the prevalence of DR in the next several decades, however, cannot be effectively managed with new diagnostic and treatment options alone. Changes in health care delivery paradigms will also be required.
Similar to the recent shift towards team-based approaches for cancer, diabetes will need to be approached in a more comprehensive manner. Close collaboration between ophthalmologists, endocrinologists, nephrologists, nutritionists, social workers, and all others involved in the care of a diabetic patient will be imperative for an efficient use of community and health system resources, and to optimize outcomes for individual patients.
The immense improvements in the care of patients with diabetes and DR over the past 50 years are an example of the significant impact laboratory and clinical research can make on the management of chronic systemic illness.
Despite great advances, though, the projected increase in the number of patients with DR in the coming decades reminds us that there is still progress to be made. Current research leading to a better understanding of molecular pathways, development of novel therapeutic targets, and use of nanotechnology, coupled with constantly improving diagnostic and imaging technology and collaborative health care delivery systems, promises to further our ability to enhance and maintain vision in diabetic patients.
Wolfensberger TJ, Hamilton AM. Diabetic retinopathy—an historical review. Semin Ophthalmol. Article Google Scholar. Wetzig PC, Worlton JT. Treatment of diabetic retinopathy by light coagulation. Br J Ophthalmol. Article CAS PubMed PubMed Central Google Scholar. Photocoagulation treatment of proliferative diabetic retinopathy.
The second report of the diabetic retinopathy study findings. Photocoagulation for diabetic macular edema. Early treatment of diabetic retinopathy study report number 1. Arch Ophthalmol. The Diabetes Control and Complications Trial Research Group. The effect of intensive treatment of diabetes on the development and progression of long-term complications in insulin-dependent diabetes mellitus.
N Engl J Med. UK Prospective Diabetes Study UKPDS Group. Intensive blood-glucose control with sulphonylureas or insulin compared with conventional treatment and risk of complications in patients with type 2 diabetes UKPDS Erratum Lancet ; UK Prospective Diabetes Study Group.
Tight blood pressure control and risk of macrovascular and microvascular complications in type 2 diabetes: UKPDS Erratum BMJ : Article PubMed Central Google Scholar.
Leske MC, Wu S-Y, Hennis A, Hyman L, et al. Barbados eye study group hyperglycemia, blood pressure, and the 9-year incidence of diabetic retinopathy: the Barbados eye studies. Article PubMed Google Scholar. ACCORD Study Group, ACCORD Eye Study Group. Effects of medical therapies on retinopathy progression in type 2 diabetes.
Klein R, Klein BE. Are individuals with diabetes seeing better? A long term epidemiological perspective. Klein R, Lee KE, Gangnon RE, Klein BE. The year incidence of visual impairment in type 1 diabetes mellitus: the Wisconsin epidemiologic study of diabetic retinopathy.
The Eye Diseases Prevalence Research Group. The prevalence of diabetic retinopathy among adults in the United States. Drummond MF, Davies LM, Frederick III. FL; assessing the costs and benefits of medical research: the diabetic retinopathy study.
Soc Sci Med. Article CAS PubMed Google Scholar. FIELD Study Investigators. Effect of fenofibrate on the need for laser treatment for diabetic retinopathy FIELD study ; a randomized controlled trial. Mitka M. Aggressive lipid, hypertension targeting yields no benefit for some with diabetes. Estacio RO, McGarling E, Biggerstaff S, Jeffers BW, et al.
Overt albuminuria predicts diabetic retinopathy in Hispanics with NIDDM. Am J Kidney Dis. Diabetes Control and Complications Trial Research Group. Effect of pregnancy on microvascular complications in the diabetes control and complications trial.
Diabetes Care. Writing Committee for the Diabetic Retinopathy Clinical Research Network, Gross JG, Glassman AR, et al. Panretinal photocoagulation vs intravitreous ranibizumab for proliferative diabetic retinopathy: a randomized clinical trial.
Diabetic Retinopathy Clinical Research Network, Elman MJ, Aiello LP, Beck RW, et al. Randomized trial evaluating ranibizumab plus prompt or deferred laser for diabetic macular edema. Adamis AP. Is diabetic retinopathy an inflammatory disease? Kern TS. Contributions of inflammatory processes to the development of the early stages of diabetic retinopathy.
Exp Diabetes Res. Article PubMed PubMed Central Google Scholar. Tang J, Kern TS. Inflammation in diabetic retinopathy. Prog Retin Eye Res. Diabetic Retinopathy Clinical Research Network. Haller JA, Kuppermann BD, Blumenkranz MS, Williams GA, Weinberg DV, Chou C, Whitcup SM.
Randomized controlled trial of an Intravitreous Dexamethasone drug delivery system in patients with diabetic macular edema. Diabetic Retinopathy Study DRS Diabetic Retinopathy Study DRS STUDY QUESTION: Does PRP argon or xenon arc prevent severe vision loss in eyes with diabetic retinopathy?
Severe NPDR was defined as the presence at least 3 of the following: 1. Cotton wool spots 2. Venous beading 3. Intraretinal microvascular abnormalities IRMA in at least 2 contiguous overlapping photographic fields 4. STUDY DESIGN: This was a randomized, prospective multicenter clinical trial.
High-risk PDR was defined as any one of the following: 1. Any NVD with vitreous hemorrhage 3. Presence of vitreous hemorrhage or pre-retinal hemorrhage 2.
Presence of any active neovascularization 3.
Christina Diabetiv Feb 07, researvh Diabetic retinopathy research retinopathj Diabetic retinopathy research Control and Prevention. Diabetes can be a difficult disease to manage. Over time, people with diabetes can experience other health problems as a result, such as diabetic retinopathy, which can lead to vision loss and blindness. Diabetic retinopathy affects more than 8 million people in the U. Diabetes Diabetic retinopathy research retinopwthy a disease Belly fat burner motivation identified Dlabetic far back as BC Diabetic retinopathy research was researcb by retiinopathy sweet properties of urine. In Mering rehinopathy Minkowski discovered Diabetic retinopathy research relevance of the pancreas in this disease process after inducing a severe and fatal form of diabetes in a dog following removal of the pancreas. Since then, advancements in medicine have led to multiple new medication therapies and approaches to treat diabetes mellitus. Despite this, diabetes remains one of the top ten most prevalent and important non-infectious causes of morbidity and mortality worldwide. An estimated
Welche bemerkenswerte Phrase
Ich entschuldige mich, aber diese Variante kommt mir nicht heran. Kann, es gibt noch die Varianten?