Sensitivity and specificity of automated analysis of single-field non-mydriatic fundus photographs by Bosch DR Algorithm-Comparison with mydriatic fundus photography (ETDRS) for screening in undiagnosed diabetic retinopathy

The application of artificial intelligence in diabetic retinopathy: progress and prospects

In recent years, artificial intelligence (AI), especially deep learning models, has increasingly been integrated into diagnosing and treating diabetic retinopathy (DR). From delving into the singular realm of ocular fundus photography to the gradual development of proteomics and other molecular approaches, from machine learning (ML) to deep learning (DL), the journey has seen a transition from a binary diagnosis of "presence or absence" to the capability of discerning the progression and severity of DR based on images from various stages of the disease course. Since the FDA approval of IDx-DR in 2018, a plethora of AI models has mushroomed, gradually gaining recognition through a myriad of clinical trials and validations. AI has greatly improved early DR detection, and we're nearing the use of AI in telemedicine to tackle medical resource shortages and health inequities in various areas. This comprehensive review meticulously analyzes the literature and clinical trials of recent years, highlighting key AI models for DR diagnosis and treatment, including their theoretical bases, features, applicability, and addressing current challenges like bias, transparency, and ethics. It also presents a prospective outlook on the future development in this domain.

Diagnostic Accuracy of Artificial Intelligence-Based Automated Diabetic Retinopathy Screening in Real-World Settings: A Systematic Review and Meta-Analysis

PURPOSE

To evaluate the diagnostic accuracy of artificial intelligence (AI)-based automated diabetic retinopathy (DR) screening in real-world settings.

DESIGN

Systematic review and meta-analysis METHODS: We conducted a systematic review of relevant literature from January 2012 to August 2022 using databases including PubMed, Scopus and Web of Science. The quality of studies was evaluated using Quality Assessment for Diagnostic Accuracy Studies 2 (QUADAS-2) checklist. We calculated pooled accuracy, sensitivity, specificity, and diagnostic odds ratio (DOR) as summary measures. The study protocol was registered with the International Prospective Register of Systematic Reviews (PROSPERO - CRD42022367034).

RESULTS

We included 34 studies which utilized AI algorithms for diagnosing DR based on real-world fundus images. Quality assessment of these studies indicated a low risk of bias and low applicability concern. Among gradable images, the overall pooled accuracy, sensitivity, specificity, and DOR were 81%, 94% (95% CI: 92.0-96.0), 89% (95% CI: 85.0-92.0) and 128 (95% CI: 80-204) respectively. Sub-group analysis showed that, when acceptable quality imaging could be obtained, non-mydriatic fundus images had a better DOR of 143 (95% CI: 82-251) and studies using 2 field images had a better DOR of 161 (95% CI 74-347). Our meta-regression analysis revealed a statistically significant association between DOR and variables such as the income status, and the type of fundus camera.

CONCLUSION

Our findings indicate that AI algorithms have acceptable performance in screening for DR using fundus images compared to human graders. Implementing a fundus camera with AI-based software has the potential to assist ophthalmologists in reducing their workload and improving the accuracy of DR diagnosis.

Transparency in Artificial Intelligence Reporting in Ophthalmology-A Scoping Review

Topic

This scoping review summarizes artificial intelligence (AI) reporting in ophthalmology literature in respect to model development and validation. We characterize the state of transparency in reporting of studies prospectively validating models for disease classification.

Clinical Relevance

Understanding what elements authors currently describe regarding their AI models may aid in the future standardization of reporting. This review highlights the need for transparency to facilitate the critical appraisal of models prior to clinical implementation, to minimize bias and inappropriate use. Transparent reporting can improve effective and equitable use in clinical settings.

Methods

Eligible articles (as of January 2022) from PubMed, Embase, Web of Science, and CINAHL were independently screened by 2 reviewers. All observational and clinical trial studies evaluating the performance of an AI model for disease classification of ophthalmic conditions were included. Studies were evaluated for reporting of parameters derived from reporting guidelines (CONSORT-AI, MI-CLAIM) and our previously published editorial on model cards. The reporting of these factors, which included basic model and dataset details (source, demographics), and prospective validation outcomes, were summarized.

Results

Thirty-seven prospective validation studies were included in the scoping review. Eleven additional associated training and/or retrospective validation studies were included if this information could not be determined from the primary articles. These 37 studies validated 27 unique AI models; multiple studies evaluated the same algorithms (EyeArt, IDx-DR, and Medios AI). Details of model development were variably reported; 18 of 27 models described training dataset annotation and 10 of 27 studies reported training data distribution. Demographic information of training data was rarely reported; 7 of the 27 unique models reported age and gender and only 2 reported race and/or ethnicity. At the level of prospective clinical validation, age and gender of populations was more consistently reported (29 and 28 of 37 studies, respectively), but only 9 studies reported race and/or ethnicity data. Scope of use was difficult to discern for the majority of models. Fifteen studies did not state or imply primary users.

Conclusion

Our scoping review demonstrates variable reporting of information related to both model development and validation. The intention of our study was not to assess the quality of the factors we examined, but to characterize what information is, and is not, regularly reported. Our results suggest the need for greater transparency in the reporting of information necessary to determine the appropriateness and fairness of these tools prior to clinical use.

Financial Disclosures

Proprietary or commercial disclosure may be found in the Footnotes and Disclosures at the end of this article.

Application of non-mydriatic fundus photography-assisted telemedicine in diabetic retinopathy screening

BACKGROUND

Early screening and accurate staging of diabetic retinopathy (DR) can reduce blindness risk in type 2 diabetes patients. DR's complex pathogenesis involves many factors, making ophthalmologist screening alone insufficient for prevention and treatment. Often, endocrinologists are the first to see diabetic patients and thus should screen for retinopathy for early intervention.

AIM

To explore the efficacy of non-mydriatic fundus photography (NMFP)-enhanced telemedicine in assessing DR and its various stages.

METHODS

This retrospective study incorporated findings from an analysis of 93 diabetic patients, examining both NMFP-assisted telemedicine and fundus fluorescein angiography (FFA). It focused on assessing the concordance in DR detection between these two methodologies. Additionally, receiver operating characteristic (ROC) curves were generated to determine the optimal sensitivity and specificity of NMFP-assisted telemedicine, using FFA outcomes as the standard benchmark.

RESULTS

In the context of DR diagnosis and staging, the kappa coefficients for NMFP-assisted telemedicine and FFA were recorded at 0.775 and 0.689 respectively, indicating substantial intermethod agreement. Moreover, the NMFP-assisted telemedicine's predictive accuracy for positive FFA outcomes, as denoted by the area under the ROC curve, was remarkably high at 0.955, within a confidence interval of 0.914 to 0.995 and a statistically significant P-value of less than 0.001. This predictive model exhibited a specificity of 100%, a sensitivity of 90.9%, and a Youden index of 0.909.

CONCLUSION

NMFP-assisted telemedicine represents a pragmatic, objective, and precise modality for fundus examination, particularly applicable in the context of endocrinology inpatient care and primary healthcare settings for diabetic patients. Its implementation in these scenarios is of paramount significance, enhancing the clinical accuracy in the diagnosis and therapeutic management of DR. This methodology not only streamlines patient evaluation but also contributes substantially to the optimization of clinical outcomes in DR management.

Reply to Siyu Tan: Associations of cardiovascular health assessed by life's essential 8 with diabetic retinopathy and mortality in type 2 diabetes. The author's reply

The present study is a reply of authors regarding the commentary from Siyu Tan. In this study, we paid specific attention to (1) highlight the inclusion criterion and diagnosis of Type 2 diabetes mellitus; (2) explain the assessments of cardiovascular health and diabetic retinopathy.

Machine learning based study for the classification of Type 2 diabetes mellitus subtypes

PURPOSE

Data-driven diabetes research has increased its interest in exploring the heterogeneity of the disease, aiming to support in the development of more specific prognoses and treatments within the so-called precision medicine. Recently, one of these studies found five diabetes subgroups with varying risks of complications and treatment responses. Here, we tackle the development and assessment of different models for classifying Type 2 Diabetes (T2DM) subtypes through machine learning approaches, with the aim of providing a performance comparison and new insights on the matter.

METHODS

We developed a three-stage methodology starting with the preprocessing of public databases NHANES (USA) and ENSANUT (Mexico) to construct a dataset with N = 10,077 adult diabetes patient records. We used N = 2,768 records for training/validation of models and left the remaining (N = 7,309) for testing. In the second stage, groups of observations -each one representing a T2DM subtype- were identified. We tested different clustering techniques and strategies and validated them by using internal and external clustering indices; obtaining two annotated datasets Dset A and Dset B. In the third stage, we developed different classification models assaying four algorithms, seven input-data schemes, and two validation settings on each annotated dataset. We also tested the obtained models using a majority-vote approach for classifying unseen patient records in the hold-out dataset.

RESULTS

From the independently obtained bootstrap validation for Dset A and Dset B, mean accuracies across all seven data schemes were [Formula: see text] ([Formula: see text]) and [Formula: see text] ([Formula: see text]), respectively. Best accuracies were [Formula: see text] and [Formula: see text]. Both validation setting results were consistent. For the hold-out dataset, results were consonant with most of those obtained in the literature in terms of class proportions.

CONCLUSION

The development of machine learning systems for the classification of diabetes subtypes constitutes an important task to support physicians for fast and timely decision-making. We expect to deploy this methodology in a data analysis platform to conduct studies for identifying T2DM subtypes in patient records from hospitals.

Artificial intelligence for diabetic retinopathy in low-income and middle-income countries: a scoping review

Diabetic retinopathy (DR) is a leading cause of blindness globally. There is growing evidence to support the use of artificial intelligence (AI) in diabetic eye care, particularly for screening populations at risk of sight loss from DR in low-income and middle-income countries (LMICs) where resources are most stretched. However, implementation into clinical practice remains limited. We conducted a scoping review to identify what AI tools have been used for DR in LMICs and to report their performance and relevant characteristics. 81 articles were included. The reported sensitivities and specificities were generally high providing evidence to support use in clinical practice. However, the majority of studies focused on sensitivity and specificity only and there was limited information on cost, regulatory approvals and whether the use of AI improved health outcomes. Further research that goes beyond reporting sensitivities and specificities is needed prior to wider implementation.

Artificial Intelligence for Diabetic Retinopathy Screening Using Color Retinal Photographs: From Development to Deployment

Diabetic retinopathy (DR), a leading cause of preventable blindness, is expected to remain a growing health burden worldwide. Screening to detect early sight-threatening lesions of DR can reduce the burden of vision loss; nevertheless, the process requires intensive manual labor and extensive resources to accommodate the increasing number of patients with diabetes. Artificial intelligence (AI) has been shown to be an effective tool which can potentially lower the burden of screening DR and vision loss. In this article, we review the use of AI for DR screening on color retinal photographs in different phases of application, ranging from development to deployment. Early studies of machine learning (ML)-based algorithms using feature extraction to detect DR achieved a high sensitivity but relatively lower specificity. Robust sensitivity and specificity were achieved with the application of deep learning (DL), although ML is still used in some tasks. Public datasets were utilized in retrospective validations of the developmental phases in most algorithms, which require a large number of photographs. Large prospective clinical validation studies led to the approval of DL for autonomous screening of DR although the semi-autonomous approach may be preferable in some real-world settings. There have been few reports on real-world implementations of DL for DR screening. It is possible that AI may improve some real-world indicators for eye care in DR, such as increased screening uptake and referral adherence, but this has not been proven. The challenges in deployment may include workflow issues, such as mydriasis to lower ungradable cases; technical issues, such as integration into electronic health record systems and integration into existing camera systems; ethical issues, such as data privacy and security; acceptance of personnel and patients; and health-economic issues, such as the need to conduct health economic evaluations of using AI in the context of the country. The deployment of AI for DR screening should follow the governance model for AI in healthcare which outlines four main components: fairness, transparency, trustworthiness, and accountability.

Deep learning system to predict the 5-year risk of high myopia using fundus imaging in children

Our study aims to identify children at risk of developing high myopia for timely assessment and intervention, preventing myopia progression and complications in adulthood through the development of a deep learning system (DLS). Using a school-based cohort in Singapore comprising of 998 children (aged 6-12 years old), we train and perform primary validation of the DLS using 7456 baseline fundus images of 1878 eyes; with external validation using an independent test dataset of 821 baseline fundus images of 189 eyes together with clinical data (age, gender, race, parental myopia, and baseline spherical equivalent (SE)). We derive three distinct algorithms - image, clinical and mix (image + clinical) models to predict high myopia development (SE ≤ -6.00 diopter) during teenage years (5 years later, age 11-17). Model performance is evaluated using area under the receiver operating curve (AUC). Our image models (Primary dataset AUC 0.93-0.95; Test dataset 0.91-0.93), clinical models (Primary dataset AUC 0.90-0.97; Test dataset 0.93-0.94) and mixed (image + clinical) models (Primary dataset AUC 0.97; Test dataset 0.97-0.98) achieve clinically acceptable performance. The addition of 1 year SE progression variable has minimal impact on the DLS performance (clinical model AUC 0.98 versus 0.97 in primary dataset, 0.97 versus 0.94 in test dataset; mixed model AUC 0.99 versus 0.97 in primary dataset, 0.95 versus 0.98 in test dataset). Thus, our DLS allows prediction of the development of high myopia by teenage years amongst school-going children. This has potential utility as a clinical-decision support tool to identify "at-risk" children for early intervention.

[Retinal alterations detected by non-mydriatic retinal camera screening and referral to ophthalmology in a population with high cardiovascular risk]

OBJECTIVE

To describe the main retinal alterations detected by non-mydriatic retinal camera screening and to evaluate factors related to referral to ophthalmology in a population at high cardiovascular risk in Palmira, Colombia.

MATERIALS AND METHODS

Cross-sectional observational study, which included 11,983 photographic imaging records of patients with hypertension and diabetes mellitus from Gesencro's S.A.S. comprehensive chronic disease care program between 2018 and 2020. Risk factors associated to referral to ophthalmology were evaluated with logistic regression, and crude and adjusted ORs (odds ratios) were obtained.

RESULTS

A total of 11,880 records were analyzed; 67.7±12years old, and 69.5% were women. Among the retinal alterations were patients with diabetic retinopathy classified as more than mild in 10% and gradeI hypertensive retinopathy in 54.9% right eye, 51.9% left eye. Macular edema was also identified. Only 2069 patients (17.4%) required referral to ophthalmology, and for imaging control 82.6%. In the multivariate analysis, the risk factors associated with the probability of being referred were male gender, age 60years and older, glycosylated hemoglobin out-of-target, advanced chronic kidney disease and the microalbumin-to-creatinine ratio moderate to severely elevated.

CONCLUSION

This study makes it possible to determine the importance of screening with a non-mydriatic retinal camera in patients at high cardiovascular risk to detect retinal abnormalities and assess risk factors associated with referral to ophthalmology. Early documentation of ocular compromise in these patients could prevent and avoid visual impairment and blindness.

Non-mydriatic fundus photography as an alternative to indirect ophthalmoscopy for screening of diabetic retinopathy in community settings: a comparative pilot study in rural and tribal India

OBJECTIVES

The impending and increasing prevalence of diabetic retinopathy (DR) in India has necessitated a need for affordable and valid community outreach screening programme for DR, especially in rural and far to reach indigenous local communities. The present study is a pilot study aimed to compare non-mydriatic fundus photography with indirect ophthalmoscopy for its utilisation as a feasible and logistically convenient screening modality for DR in an older age, rural, tribal population in Western India.

DESIGN AND SETTING

This community-based, cross-sectional, prospective population study was a part of a module using Rapid Assessment of Avoidable Blindness and DR methodology in 8340 sampled participants with ≥50 years age. In this study, the diabetics identified were screened for DR using two methods: non-mydriatic fundus photography on the field by trained professionals, that were then graded by a retina specialist at the base hospital and indirect ophthalmoscopy by expert ophthalmologists in the field with masking of each other's findings for its utility and comparison.

RESULTS

The prevalence of DR, sight threatening DR and maculopathy using indirect ophthalmoscopy was found to be 12.1%, 2.1% and 6.6%, respectively. A fair agreement (κ=0.48 for DR and 0.59 for maculopathy) was observed between both the detection methods. The sensitivity and specificity of fundus photographic evaluation compared with indirect ophthalmoscopy were found to be 54.8% and 92.1% (for DR), 60.7% and 90.8% (for any DR) and 84.2% and 94.8% (for only maculopathy), respectively.

CONCLUSION

Non-mydriatic fundus photography has the potential to identify DR (any retinopathy or maculopathy) in community settings in Indian population. Its utility as an affordable and logistically convenient cum practical modality is demonstrable. The sensitivity of this screening modality can be further increased by investing in better resolution cameras, capturing quality images and training and validation of imagers.

TRIAL REGISTRATION NUMBER

CTRI/2020/01/023025; Clinical Trial Registry, India (CTRI).

Efficacy of artificial intelligence-based screening for diabetic retinopathy in type 2 diabetes mellitus patients

AIM

To explore the efficacy of artificial intelligence (AI)-based screening for diabetic retinopathy (DR) in type 2 diabetes mellitus (T2DM) patients.

METHODS

Data were obtained from 549 T2DM patients who visited the Fundus Disease Center at Henan Provincial People's Hospital from 2018/10-2020/09. DR identification and grading were conducted by two retina specialists, EyeWisdom®DSS and EyeWisdom®MCS, with ophthalmologist grading as reference standard, efficacy of EyeWisdom was evaluated according to sensitivity, specificity, positive predictive value, and negative predictive value.

RESULTS

Ophthalmologists detected 324 DR cases. Among them, there were 43 of mild non-proliferative DR (NPDR), 79 of moderate NPDR, 61 of severe NPDR, and 141 of proliferative DR (PDR). EyeWisdom®DSS detected 337 DR and EyeWisdom®MCS detected 264 DR. Sensitivity and specificity of EyeWisdom®DSS were 91.0%(95 %CI: 87.3%-93.8%) and 81.3% (95 %CI: 75.5%-86.1%), while EyeWisdom®MCS correctly identified 76.2%(95 %CI: 71.1%-80.7%) of patients with DR and 92.4%(95 %CI: 87.9%-95.4%) of patients without DR. EyeWisdom®DSS showed 76.5%(95 %CI: 69.6%-82.3%) sensitivity and 78.4%(95 %CI: 73.7%-82.5%) specificity for detecting NPDR and 64.5%(95 %CI: 56.0%-72.3%) sensitivity and 93.1%(95 %CI: 90.1%-95.3%) specificity for diagnosing PDR.

CONCLUSION

EyeWisdom®DSS is effective in screening for DR, and the accuracy of EyeWisdom®MCS was higher for identifying patients without DR. It is valuable to carry out AI-based DR screening in poorer regions.

Current status and future possibilities of retinal imaging in diabetic retinopathy care applicable to low- and medium-income countries

Diabetic retinopathy (DR) is a leading cause of blindness among adults and the numbers are projected to rise. There have been dramatic advances in the field of retinal imaging since the first fundus image was captured by Jackman and Webster in 1886. The currently available imaging modalities in the management of DR include fundus photography, fluorescein angiography, autofluorescence imaging, optical coherence tomography, optical coherence tomography angiography, and near-infrared reflectance imaging. These images are obtained using traditional fundus cameras, widefield fundus cameras, handheld fundus cameras, or smartphone-based fundus cameras. Fluorescence lifetime ophthalmoscopy, adaptive optics, multispectral and hyperspectral imaging, and multicolor imaging are the evolving technologies which are being researched for their potential applications in DR. Telemedicine has gained popularity in recent years as remote screening of DR has been made possible. Retinal imaging technologies integrated with artificial intelligence/deep-learning algorithms will likely be the way forward in the screening and grading of DR. We provide an overview of the current and upcoming imaging modalities which are relevant to the management of DR.

Situational analysis of diabetic retinopathy screening in India: How has it changed in the last three years?

Of all the eye conditions in the contemporary Indian context, diabetic retinopathy (DR) attracts the maximum attention not just of the eye care fraternity but the entire medical fraternity. Countries are at different stages of evolution in structured DR screening services. In most low and middle income countries, screening is opportunistic, while in most of the high income countries structured population-based DR screening is the established norm. To reduce inequities in access, it is important that all persons with diabetes are provided equal access to DR screening and management services. Such programs have been proven to reverse the magnitude of vision-threatening diabetic retinopathy in countries like England and Scotland. DR screening should not be considered an endpoint in itself but the starting point in a continuum of services for effective management of DR services so that the risk of vision loss can be mitigated. Till recently all DR screening programs in India were opportunistic models where persons with diabetes visiting an eye care facility were screened. Since 2016, with support from International funders, demonstration models integrating DR screening services in the public health system were initiated. These pilots showed that a systematic integrated structured DR screening program is possible in India and need to be scaled up across the country. Many DR screening and referral initiatives have been adversely impacted by the COVID-19 pandemic and advocacy with the government is critical to facilitate continuous sustainable services.

The Impact of COVID-19 on Diabetic Retinopathy Monitoring and Treatment

PURPOSE OF REVIEW

Diabetic retinopathy (DR) is one of the leading causes of vision loss worldwide. Although screening and early treatment guidelines for DR have significantly reduced the disease burden, restrictions related to the COVID-19 pandemic have changed real-world practice patterns in the management of DR. This review summarizes evolving guidelines and outcomes of the treatment of DR in the setting of the pandemic.

RECENT FINDINGS

Intravitreal injections for DR have decreased significantly globally during the pandemic, ranging from approximately 30 to nearly 100% reduction, compared to corresponding timepoints in 2019. Most studies on functional outcomes show a decrease in visual acuity on delayed follow-up. Changing practice patterns in the management of DR has led to fewer intravitreal injections and overall reduction in visual acuity on follow-up. As COVID variants emerge, it will be necessary to continue evaluating practice guidelines.

Diagnostic performance of deep-learning-based screening methods for diabetic retinopathy in primary care-A meta-analysis

Background

Diabetic retinopathy (DR) affects 10–24% of patients with diabetes mellitus type 1 or 2 in the primary care (PC) sector. As early detection is crucial for treatment, deep learning screening methods in PC setting could potentially aid in an accurate and timely diagnosis.

Purpose

The purpose of this meta-analysis was to determine the current state of knowledge regarding deep learning (DL) screening methods for DR in PC.

Data sources

A systematic literature search was conducted using Medline, Web of Science, and Scopus to identify suitable studies.

Study selection

Suitable studies were selected by two researchers independently. Studies assessing DL methods and the suitability of these screening systems (diagnostic parameters such as sensitivity and specificity, information on datasets and setting) in PC were selected. Excluded were studies focusing on lesions, applying conventional diagnostic imaging tools, conducted in secondary or tertiary care, and all publication types other than original research studies on human subjects.

Data extraction

The following data was extracted from included studies: authors, title, year of publication, objectives, participants, setting, type of intervention/method, reference standard, grading scale, outcome measures, dataset, risk of bias, and performance measures.

Data synthesis and conclusion

The summed sensitivity of all included studies was 87% and specificity was 90%. Given a prevalence of DR of 10% in patients with DM Type 2 in PC, the negative predictive value is 98% while the positive predictive value is 49%.

Limitations

Selected studies showed a high variation in sample size and quality and quantity of available data.

Progress of artificial intelligence in diabetic retinopathy screening

Diabetic retinopathy (DR) is one of the leading causes of blindness worldwide, and the limited availability of qualified ophthalmologists restricts its early diagnosis. For the past few years, artificial intelligence technology has developed rapidly and has been applied in DR screening. The upcoming technology provides support on DR screening and improves the identification of DR lesions with a high sensitivity and specificity. This review aims to summarize the progress on automatic detection and classification models for the diagnosis of DR.

Remote screening of diabetic retinopathy using ultra-widefield retinal imaging

AIMS

To study the possibility of constructing a remote interpretation system for retinal images.

METHODS

An ultra-widefield (UWF) retinal imaging device was installed in the internal medicine department specializing in diabetes to obtain fundus images of patients with diabetes. Remote interpretation was conducted at Nagoya City University using a cloud server. The medical data, severity of retinopathy, and frequency of ophthalmologic visits were analyzed.

RESULTS

Four hundred ninety-nine patients (mean age, 62.5 ± 13.4 years) were included. The duration of diabetes in 240 (48.1%) patients was less than 10 years and 433 (86.7%) patients had a hemoglobin (Hb) A1c below 8%. Regarding the retinopathy severity, 360 (72.1%) patients had no diabetic retinopathy (NDR), 63 (12.6%) mild nonproliferative retinopathy (NPDR), 38 (7.64%) moderate NPDR, 13 (2.6%) severe NPDR, and 25 (5.0%) PDR. Two hundred forty-one (48.3%) patients had an ophthalmologic consultation within 1 year, 104 (20.8%) had no history of an ophthalmologic consultation. DR was not present in 86 (82.7%) patients who had never had an ophthalmologic examination, 30 (78.9%) patients with severe NPDR or PDR had had an ophthalmologic visit within 1 year. The frequency of ophthalmic visits was correlated negatively with age, diabetes duration, HbA1c, and severity of retinopathy.

CONCLUSION

Remote interpretation of DR using UWF retinal imaging was useful for retinopathy screening. During the COVID-19 pandemic, a remote screening system that can ensure compulsory social distancing and reduce the number of ophthalmic visits can be a safe system for patients and clinicians.

The Role of Telemedicine, In-Home Testing and Artificial Intelligence to Alleviate an Increasingly Burdened Healthcare System: Diabetic Retinopathy

In the presence of the ever-increasing incidence of diabetes mellitus (DM), the prevalence of diabetic eye disease (DED) is also growing. Despite many improvements in diabetic care, DM remains a leading cause of visual impairment in working-age patients. So far, prevention has been the best way to protect vision. The sooner we diagnose DED, the more effective the treatment is. Thus, diabetic retinopathy (DR) screening, especially with imaging techniques, is a method of choice for vision protection. To alleviate the burden of diabetic patients who need ophthalmic care, telemedicine and in-home testing are used, supported by artificial intelligence (AI) algorithms. This is why we decided to evaluate current image teleophthalmology methods used for DR screening. We searched the PubMed platform for papers published over the last 5 years (2015–2020) using the following key words: telemedicine in diabetic retinopathy screening, diabetic retinopathy screening, automated diabetic retinopathy screening, artificial intelligence in diabetic retinopathy screening, smartphone diabetic retinopathy testing. We have included 118 original articles meeting the above criteria, discussing imaging diabetic retinopathy screening methods. We have found that fundus cameras, stable or mobile, are most commonly used for retinal photography, with portable fundus cameras also relatively common. Other possibilities involve the use of ultra-wide-field (UWF) imaging and even optical coherence tomography (OCT) devices for DR screening. Also, the role of smartphones is increasingly recognized in the field. Retinal fundus images are assessed by humans instantly or remotely, while AI algorithms seem to be useful tools facilitating retinal image assessment. The common use of smartphones and availability of relatively cheap, easy-to-use adapters for retinal photographs augmented by AI algorithms make it possible for eye fundus photographs to be taken by non-specialists and in non-medical setting. This opens the way for in-home testing conducted on a much larger scale in the future. In conclusion, based on current DR screening techniques, we can suggest that the future practice of eye care specialists will be widely supported by AI algorithms, and this way will be more effective.

Magnitude and Determinants of Diabetic Retinopathy Among Indian Diabetic Patients Undergoing Telescreening in India

Purpose: To determine the magnitude, determinants, and public health issues related to diabetic retinopathy (DR) in India using 2019 data from a for-profit telescreening program. Methods: Digital retinal images were captured using a nonmydriatic fundus camera and transferred via the telescreening program to a reading center. Ophthalmologists trained in DR image reading created the DR status reports. Age/sex-adjusted rates of DR, sight-threatening DR (STDR), and diabetic macular edema (DME) were calculated and correlated with known risk factors. Results: Images of 51,760 Indian diabetic patients (103,520 eyes) were reviewed. The prevalence of DR, STDR, and DME was 19.1% (95% confidence interval [CI]: 18.9-19.5), 5.1% (95% CI: 4.9-5.3), and 3.9% (95% CI: 3.7-4.1), respectively. Based on these data, we projected 14.7 million cases of DR, 3.9 million with STDR, and 3.0 million DME cases in India. Statistically significant risk factors for DR were male gender (odds ratio [OR] = 1.19, p < 0.001), older age (χ² = 270, df = 3, p < 0.001), history of cataract surgery (OR = 2.0, p < 0.001), longer duration of diabetes (χ² = 1084, p < 0.001), and type 1 diabetes (OR = 3.9, p = 0.01). There was a statistically significant variation of DR by geographic zones (χ² = 310, p < 0.001). Laser treatment coverage for STDR was 22%. Duration of diabetes (p < 0.001), cataract surgery in the past (p = 0.02), and females (p = 0.001) were predictors of STDR. Conclusion: This model of telescreening for DR provides an additional pathway for screening and preventing diabetes-related visual morbidity in India. The data from this study can be used for epidemiologic and ophthalmic health policies related to diabetes.

Effectiveness of health education and monetary incentive on uptake of diabetic retinopathy screening at a community health center in South Gujarat, India

Purpose:

The effectiveness of Accredited Social Health Activists (ASHAs) with and without monetary incentive in uptake of diabetic retinopathy (DR) screening at community health center (CHC) was compared in South Gujarat, India.

Methods:

In this non-randomized controlled trial, ASHAs were incentivized to refer people with diabetes mellitus (PwDM) from their respective villages for DR screening after people were sensitized to DM and DR. The minimum sample size was 63 people in each arm.

Results:

Of 162, 50.6% were females, 80.2% were literate, 56.2% were >50 years, 54.3% had increased random blood sugar (RBS), and 59.9% had diabetes for 5 years. The percentage of screening was significantly higher [relative risk (RR) = 4.37, 95% confidence interval (CI) 2.79, 6.84] in ASHA incentive group and health education (HE) group (RR = 3.67, 95% CI 2.35, 5.75) compared with baseline. Providing incentive to ASHAs was not found to be of extra advantage (RR = 1.19, 95% CI 0.89, 1.57). The likelihood of uptake of screening was higher among uncontrolled PwDM, poor literacy, and higher duration of diabetes in incentive phase (P < 0.001) compared with HE. The results show that age (P = 0.017), education (P = 0.015) and level of RBS (P = 0.001) of those referred were significantly associated with incentives to ASHAs.

Conclusion:

ASHAs can be used effectively to refer known PwDM for DR screening especially when DR screening program is introduced in population with low awareness and poor accessibility. When incentives are planned, additional burden on resources should be kept in mind before adapting this model of care.

Validation of Deep Convolutional Neural Network-based algorithm for detection of diabetic retinopathy - Artificial intelligence versus clinician for screening

Purpose:

Deep learning is a newer and advanced subfield in artificial intelligence (AI). The aim of our study is to validate a machine-based algorithm developed based on deep convolutional neural networks as a tool for screening to detect referable diabetic retinopathy (DR).

Methods:

An AI algorithm to detect DR was validated at our hospital using an internal dataset consisting of 1,533 macula-centered fundus images collected retrospectively and an external validation set using Methods to Evaluate Segmentation and Indexing Techniques in the field of Retinal Ophthalmology (MESSIDOR) dataset. Images were graded by two retina specialists as any DR, prompt referral (moderate nonproliferative diabetic retinopathy (NPDR) or above or presence of macular edema) and sight-threatening DR/STDR (severe NPDR or above) and compared with AI results. Sensitivity, specificity, and area under curve (AUC) for both internal and external validation sets for any DR detection, prompt referral, and STDR were calculated. Interobserver agreement using kappa value was calculated for both the sets and two out of three agreements for DR grading was considered as ground truth to compare with AI results.

Results:

In the internal validation set, the overall sensitivity and specificity was 99.7% and 98.5% for Any DR detection and 98.9% and 94.84%for Prompt referral respectively. The AUC was 0.991 and 0.969 for any DR detection and prompt referral respectively. The agreement between two observers was 99.5% and 99.2% for any DR detection and prompt referral with a kappa value of 0.94 and 0.96, respectively. In the external validation set (MESSIDOR 1), the overall sensitivity and specificity was 90.4% and 91.0% for any DR detection and 94.7% and 97.4% for prompt referral, respectively. The AUC was. 907 and. 960 for any DR detection and prompt referral, respectively. The agreement between two observers was 98.5% and 97.8% for any DR detection and prompt referral with a kappa value of 0.971 and 0.980, respectively.

Conclusion:

With increasing diabetic population and growing demand supply gap in trained resources, AI is the future for early identification of DR and reducing blindness. This can revolutionize telescreening in ophthalmology, especially where people do not have access to specialized health care.

Cost-effectiveness and diagnostic accuracy of telemedicine in macular disease and diabetic retinopathy: A systematic review and meta-analysis

OBJECTIVE

To determine cost-effectiveness and the diagnostic accuracy of teleophthalmology (TO) in the detection of macular edema (ME) and various grades of diabetic retinopathy (DR).

METHODS

MEDLINE, EMBASE, and Cochrane databases were searched for TO, ME, and DR on May 25, 2016. The search was updated on April 2, 2019. Pooled sensitivity and specificity for ME and various grades of DR were determined using Meta-Disc software. A systematic review of the articles discussing the cost-effectiveness of TO screening was also performed.

RESULTS

Thirty-three articles on the diagnostic accuracy and 28 articles on the cost-effectiveness were selected.

CONCLUSIONS

Telescreening is moderately sensitive but very specific for the diagnosis of diabetic retinopathy. Non-mydriatic Teleretinal screening services are cost-effective, decrease clinics workload, and increase patient compliance if provided free of cost in remote low socioeconomic regions.

Cost Analysis of Teleophthalmology Screening for Diabetic Retinopathy Using Teleophthalmology Billing Codes

BACKGROUND AND OBJECTIVE

To evaluate the financial sustainability of teleophthalmology screening for diabetic retinopathy (DR) using telehealth billing codes.

PATIENTS AND METHODS

The authors performed an Institutional Review Board-approved retrospective review of medical records, billing data, and quality metrics at the University of California Davis Health System from patients screened for DR through an internal medicine-based telemedicine program using CPT codes 92227 or 92228.

RESULTS

A total of 290 patients received teleophthalmology screening over a 12-month period, resulting in an increase in the DR screening rate from 49% to 63% (P < .0001). The average payment per patient was $19.86, with an estimated cost of $41.02 per patient. The projected per-patient incentive bonus was $43.06 with a downstream referral revenue of $39.38 per patient. One hundred seventy-eight clinic visits were eliminated, providing an estimated cost savings of $42.53 per patient.

CONCLUSION

Sustainable teleophthalmology screening may be achieved by billing telehealth codes but only with health care incentive bonuses, patient referrals, and by accounting for the projected cost-savings of eliminating office visits. [Ophthalmic Surg Lasers Imaging Retina. 2020;51:S26-S34.].

Detection of Diabetic Macular Edema in Optical Coherence Tomography Image Using an Improved Level Set Algorithm

Diabetic macular edema (DME) is a major cause of visual loss in the patients with diabetic retinopathy. DME detection in Optical Coherence Tomography (OCT) image contributes to the early diagnosis of diabetic retinopathy and blindness prevention. Currently, DME detection in the OCT image mainly relies on the handwork by the experienced clinician. It is a laborious, time-consuming, and challenging work to organize a comprehensive DME screening for diabetic patients. In this study, we proposed a novel algorithm for the detection and segmentation of DME region in OCT image based on the K-means clustering algorithm and improved Selective Binary and Gaussian Filtering regularized level set (SBGFRLS) algorithm named as SBGFRLS-OCT algorithm. SBGFRLS-OCT algorithm was compared with the current level set algorithms, including C-V (Chan-Vese), GAC (geodesic active contour), and SBGFRLS, to estimate the performance of DME detection. SBGFRLS-OCT algorithm was also compared with the clinician to estimate the precision, sensitivity, and specificity of DME segmentation. Compared with C-V, GAC, and SBGFRLS algorithm, the SBGFRLS-OCT algorithm enhanced the accuracy and reduces the processing time of DME detection. Compared with manual DME segmentation, the SBGFRLS-OCT algorithm achieved a comparable precision (97.7%), sensitivity (91.8%), and specificity (99.2%). Collectively, this study presents a novel algorithm for DME detection in the OCT image, which can be used for mass diabetic retinopathy screening.

Technical and imaging factors influencing performance of deep learning systems for diabetic retinopathy

Deep learning (DL) has been shown to be effective in developing diabetic retinopathy (DR) algorithms, possibly tackling financial and manpower challenges hindering implementation of DR screening. However, our systematic review of the literature reveals few studies studied the impact of different factors on these DL algorithms, that are important for clinical deployment in real-world settings. Using 455,491 retinal images, we evaluated two technical and three image-related factors in detection of referable DR. For technical factors, the performances of four DL models (VGGNet, ResNet, DenseNet, Ensemble) and two computational frameworks (Caffe, TensorFlow) were evaluated while for image-related factors, we evaluated image compression levels (reducing image size, 350, 300, 250, 200, 150 KB), number of fields (7-field, 2-field, 1-field) and media clarity (pseudophakic vs phakic). In detection of referable DR, four DL models showed comparable diagnostic performance (AUC 0.936-0.944). To develop the VGGNet model, two computational frameworks had similar AUC (0.936). The DL performance dropped when image size decreased below 250 KB (AUC 0.936, 0.900, p < 0.001). The DL performance performed better when there were increased number of fields (dataset 1: 2-field vs 1-field-AUC 0.936 vs 0.908, p < 0.001; dataset 2: 7-field vs 2-field vs 1-field, AUC 0.949 vs 0.911 vs 0.895). DL performed better in the pseudophakic than phakic eyes (AUC 0.918 vs 0.833, p < 0.001). Various image-related factors play more significant roles than technical factors in determining the diagnostic performance, suggesting the importance of having robust training and testing datasets for DL training and deployment in the real-world settings.

The Importance of Telemedicine during COVID-19 Pandemic: A Focus on Diabetic Retinopathy

Recently, telemedicine has become remarkably important, due to increased deployment and development of digital technologies. National and international guidelines should consider its inclusion in their updates. During the COVID-19 pandemic, mandatory social distancing and the lack of effective treatments has made telemedicine the safest interactive system between patients, both infected and uninfected, and clinicians. A few potential evidence-based scenarios for the application of telemedicine have been hypothesized. In particular, its use in diabetes and complication monitoring has been remarkably increasing, due to the high risk of poor prognosis. New evidence and technological improvements in telemedicine application in diabetic retinopathy (DR) have demonstrated efficacy and usefulness in screening. Moreover, despite an initial increase for devices and training costs, teleophthalmology demonstrated a good cost-to-efficacy ratio; however, no national screening program has yet focused on DR prevention and diagnosis. Lack of data during the COVID-19 pandemic strongly limits the possibility of tracing the real management of the disease, which is only conceivable from past evidence in normal conditions. The pandemic further stressed the importance of remote monitoring. However, the deployment of device and digital application used to increase screening of individuals and monitor progression of retinal disease needs to be easily accessible to general practitioners.

Artificial intelligence for diabetic retinopathy screening: a review

Diabetes is a global eye health issue. Given the rising in diabetes prevalence and ageing population, this poses significant challenge to perform diabetic retinopathy (DR) screening for these patients. Artificial intelligence (AI) using machine learning and deep learning have been adopted by various groups to develop automated DR detection algorithms. This article aims to describe the state-of-art AI DR screening technologies that have been described in the literature, some of which are already commercially available. All these technologies were designed using different training datasets and technical methodologies. Although many groups have published robust diagnostic performance of the AI algorithms for DR screening, future research is required to address several challenges, for examples medicolegal implications, ethics, and clinical deployment model in order to expedite the translation of these novel technologies into the healthcare setting.