Quality assurance for diabetic retinopathy telescreening

Artificial intelligence for telemedicine diabetic retinopathy screening: a review

PURPOSE

This study aims to compare artificial intelligence (AI) systems applied in diabetic retinopathy (DR) teleophthalmology screening, currently deployed systems, fairness initiatives and the challenges for implementation.

METHODS

The review included articles retrieved from PubMed/Medline/EMBASE literature search strategy regarding telemedicine, DR and AI. The screening criteria included human articles in English, Portuguese or Spanish and related to telemedicine and AI for DR screening. The author's affiliations and the study's population income group were classified according to the World Bank Country and Lending Groups.

RESULTS

The literature search yielded a total of 132 articles, and nine were included after full-text assessment. The selected articles were published between 2004 and 2020 and were grouped as telemedicine systems, algorithms, economic analysis and image quality assessment. Four telemedicine systems that perform a quality assessment, image preprocessing and pathological screening were reviewed. A data and post-deployment bias assessment are not performed in any of the algorithms, and none of the studies evaluate the social impact implementations. There is a lack of representativeness in the reviewed articles, with most authors and target populations from high-income countries and no low-income country representation.

CONCLUSIONS

Telemedicine and AI hold great promise for augmenting decision-making in medical care, expanding patient access and enhancing cost-effectiveness. Economic studies and social science analysis are crucial to support the implementation of AI in teleophthalmology screening programs. Promoting fairness and generalizability in automated systems combined with telemedicine screening programs is not straightforward. Improving data representativeness, reducing biases and promoting equity in deployment and post-deployment studies are all critical steps in model development.

Considerations of COVID-19 in Ophthalmology

Since its emergence in early 2020, the SARS-CoV-2 infection has had a significant impact on the entire eye care system. Ophthalmologists have been categorized as a high-risk group for contracting the virus due to the belief that the eye may be a site of inoculation and transmission of the SARS-CoV-2 infection. As a result, clinical ophthalmologists, optometrists, and eyecare professionals have had to familiarize themselves with the ocular manifestations of COVID-19, as well as its treatments and vaccines. The implementation of measures to prevent the transmission of the virus, such as restrictions, lockdowns, telemedicine, and artificial intelligence (AI), have led to substantial and potentially irreversible changes in routine clinical practice, education, and research. This has resulted in the emergence of a new mode of managing patients in a routine clinical setting. This brief review aims to provide an overview of various aspects of COVID-19 in ophthalmology, including the ocular manifestations related to the disease, the modes of transmission of SARS-CoV-2 infection, precautions taken in ophthalmic practice to prevent the spread of the virus, drugs, and vaccines used in the treatment of COVID-19, the impact of the pandemic on patients, clinicians, and the eye care system as a whole, and the future of ophthalmology conditioned by this global pandemic experience.

Risk factors associated with progression to referable retinopathy: a type 2 diabetes mellitus cohort study in the Republic of Ireland

AIM

To determine factors associated with progression to referable diabetic retinopathy in people with type 2 diabetes in the Republic of Ireland.

RESEARCH DESIGN AND METHODS

The study was conducted in a dynamic cohort of 2770 people with type 2 diabetes, recruited between April 2005 and July 2013. Systemic factors (systolic and diastolic blood pressure, HbA_1c , lipid levels, BMI) and baseline diabetic retinopathy grading results were evaluated at 4-monthly and yearly intervals, respectively. Associations between risk factors (most recently recorded value, and rate of change in value between pairs of consecutive systemic evaluations) and development of referable diabetic retinopathy were estimated using Cox proportional hazards models.

RESULTS

There was a fourfold increased risk of progression to referral when retinopathy was present at baseline vs no retinopathy at baseline (hazard ratio 4.02, 95% CI 2.80-5.78; P<0.001). Higher current values of HbA_1c (hazard ratio 1.22, 95% CI 1.11-1.34; P<0.001), systolic blood pressure (hazard ratio 1.29, 95% CI 1.15-1.45; P<0.001) and triglycerides (hazard ratio 1.10, 95% CI 1.03-1.18; P=0.004) were associated with increased risk of referral. Higher current BMI (hazard ratio 0.83, 95% CI 0.73-0.95; P=0.007) and diastolic blood pressure (hazard ratio 0.91, 95% CI 0.85-0.97; P=0.006) were associated with reduced risk of referral.

CONCLUSIONS

Presence of retinopathy at baseline was strongly associated with increased risk of referral. Modest associations between systemic factors and risk of progression to referable retinopathy were detected.

Practice Guidelines for Ocular Telehealth-Diabetic Retinopathy, Third Edition

Contributors The following document and appendices represent the third edition of the Practice Guidelines for Ocular Telehealth-Diabetic Retinopathy. These guidelines were developed by the Diabetic Retinopathy Telehealth Practice Guidelines Working Group. This working group consisted of a large number of subject matter experts in clinical applications for telehealth in ophthalmology. The editorial committee consisted of Mark B. Horton, OD, MD, who served as working group chair and Christopher J. Brady, MD, MHS, and Jerry Cavallerano, OD, PhD, who served as cochairs. The writing committees were separated into seven different categories. They are as follows: 1.Clinical/operational: Jerry Cavallerano, OD, PhD (Chair), Gail Barker, PhD, MBA, Christopher J. Brady, MD, MHS, Yao Liu, MD, MS, Siddarth Rathi, MD, MBA, Veeral Sheth, MD, MBA, Paolo Silva, MD, and Ingrid Zimmer-Galler, MD. 2.Equipment: Veeral Sheth, MD (Chair), Mark B. Horton, OD, MD, Siddarth Rathi, MD, MBA, Paolo Silva, MD, and Kristen Stebbins, MSPH. 3.Quality assurance: Mark B. Horton, OD, MD (Chair), Seema Garg, MD, PhD, Yao Liu, MD, MS, and Ingrid Zimmer-Galler, MD. 4.Glaucoma: Yao Liu, MD, MS (Chair) and Siddarth Rathi, MD, MBA. 5.Retinopathy of prematurity: Christopher J. Brady, MD, MHS (Chair) and Ingrid Zimmer-Galler, MD. 6.Age-related macular degeneration: Christopher J. Brady, MD, MHS (Chair) and Ingrid Zimmer-Galler, MD. 7.Autonomous and computer assisted detection, classification and diagnosis of diabetic retinopathy: Michael Abramoff, MD, PhD (Chair), Michael F. Chiang, MD, and Paolo Silva, MD.

Telemedicine and ocular health in diabetes mellitus

Teleretinal/teleophthalmological programs that use existing health information technology infrastructure solutions for people with diabetes increase access to and adherence to appropriate eye care. Teleophthalmological studies indicate that the single act of patients viewing their own retinal images improves self-management behaviour and clinical outcomes. In some settings this can be done at lower cost and with improved visual outcomes compared with standard eye care. Cost-effective and sustainable teleretinal surveillance for detection of diabetic retinopathy requires a combination of an inexpensive portable device for taking low light-level retinal images without the use of pharmacological dilation of the pupil and a computer-assisted methodology for rapidly detecting and diagnosing diabetic retinopathy. A more holistic telehealth-care paradigm augmented with the use of health information technology, medical devices, mobile phone and mobile health applications and software applications to improve health-care co-ordination, self-care management and education can significantly impact a broad range of health outcomes, including prevention of diabetes-associated visual loss. This approach will require a collaborative, transformational, patient-centred health-care program that integrates data from medical record systems with remote monitoring of data and a longitudinal health record. This includes data associated with social media applications and personal mobile health technology and should support continuous interactions between the patient, health-care team and the patient's social environment. Taken together, this system will deliver contextually and temporally relevant decision support to patients to facilitate their well-being and to reduce the risk of diabetic complications.

External quality assurance for image grading in the Scottish Diabetic Retinopathy Screening Programme

AIMS

To develop and evaluate an image grading external quality assurance system for the Scottish Diabetic Retinopathy Screening Programme.

METHOD

A web-based image grading system was developed which closely matches the current Scottish national screening software. Two rounds of external quality assurance were run in autumn 2008 and spring 2010, each time using the same 100 images. Graders were compared with a consensus standard derived from the top-level graders' results. After the first round, the centre lead clinicians and top-level graders reviewed the results and drew up guidance notes for the second round.

RESULTS

Grader sensitivities ranged from 60.0 to 100% (median 92.5%) in 2008, and from 62.5 to 100% (median 92.5%) in 2010. Specificities ranged from 34.0 to 98.0% (median 86%) in 2008, and 54.0 to 100% (median 88%) in 2010. There was no difference in sensitivity between grader levels, but first-level graders had a significantly lower specificity than level-two and level-three graders. In 2008, one centre had a lower sensitivity but higher specificity than the majority of centres. Following the feedback from the first round, overall agreement improved in 2010 and there were no longer any significant differences between centres.

CONCLUSIONS

A useful educational tool has been developed for image grading external quality assurance.

Telehealth practice recommendations for diabetic retinopathy, second edition

Ocular telemedicine and telehealth have the potential to decrease vision loss from DR. Planning, execution, and follow-up are key factors for success. Telemedicine is complex, requiring the services of expert teams working collaboratively to provide care matching the quality of conventional clinical settings. Improving access and outcomes, however, makes telemedicine a valuable tool for our diabetic patients. Programs that focus on patient needs, consider available resources, define clear goals, promote informed expectations, appropriately train personnel, and adhere to regulatory and statutory requirements have the highest chance of achieving success.

Implementation and first-year screening results of an ocular telehealth system for diabetic retinopathy in China

BACKGROUND

To describe implementation and first-year screening results of the first Chinese telehealth system for diabetic retinopathy (DR) - the Beixinjing Community Diabetic Retinopathy Telehealth system (BCDRT).

METHODS

BCDRT implementation was based on the acquisition of adequate digital retinographs, secure digital transmission, storage and retrieval of participants' data and reader-generated medical reports. Local diabetic residents meeting inclusion criteria were enrolled into the BCDRT system beginning in 2009. Participants recommended for further in-person examination with ophthalmologists were followed, and the consistencies in diagnoses between BCDRT and ophthalmologists for DR or macular edema were calculated.

RESULTS

A total of 471 diabetic residents participated in BCDRT screening in 2009. The proportions of total DR, proliferative DR, and diabetic macular edema were 24.42% (115 patients), 2.12% (10 patients) and 6.47% (24 patients), respectively: 56 patients consulted ophthalmologists for further in-person retinal examination with funduscopy after pupil dilation. High rates of consistency between BCDRT screening and ophthalmologists were observed for macular edema (Kappa = 0.81), moderate or severe non-proliferative DR grade (Kappa = 0.92), and other DR grades (Kappa = 1). A total of 456 (96.82%) patients were willing to participate in the next BCDRT screening.

CONCLUSIONS

BCDRT was a reliable and valid system for DR screening, and offers the potential to increase DR annual screening rates in local residents.

Diabetic retinopathy screening: a systematic review of the economic evidence

This paper systematically reviews the published literature on the economic evidence of diabetic retinopathy screening. Twenty-nine electronic databases were searched for studies published between 1998 and 2008. Internet searches were carried out and reference lists of key studies were hand searched for relevant articles. The key search terms used were 'diabetic retinopathy', 'screening', 'economic' and 'cost'. The search identified 416 papers of which 21 fulfilled the inclusion criteria, comprising nine cost-effectiveness studies, one cost analysis, one cost-minimization analysis, four cost-utility analyses and six reviews. Eleven of the included studies used economic modelling techniques and/or computer simulation to assess screening strategies. To date, the economic evaluation literature on diabetic retinopathy screening has focused on four key questions: the overall cost-effectiveness of ophthalmic care; the cost-effectiveness of systematic vs. opportunistic screening; how screening should be organized and delivered; and how often people should be screened. Systematic screening for diabetic retinopathy is cost-effective in terms of sight years preserved compared with no screening. Digital photography with telemedicine links has the potential to deliver cost-effective, accessible screening to rural, remote and hard-to-reach populations. Variation in compliance rates, age of onset of diabetes, glycaemic control and screening sensitivities influence the cost-effectiveness of screening programmes and are important sources of uncertainty in relation to the issue of optimal screening intervals. There is controversy in relation to the economic evidence on optimal screening intervals. Further research is needed to address the issue of optimal screening interval, the opportunities for targeted screening to reflect relative risk and the effect of different screening intervals on attendance or compliance by patients.

Reliability of screening methods for diabetic retinopathy

AIM

To review studies of the reliability (reproducibility) of the commonly used methods (ophthalmoscopy and inspection of retinal images) of screening for diabetic retinopathy (DR).

RESEARCH DESIGN AND METHODS

Literature search.

RESULTS

We found six studies of the intra-examiner agreement after examining the same retinal images. Three of these found an almost perfect agreement (k > 0.8) after inspecting colour slides and digital images; three other studies reported 'significant differences' in microaneurysm counts and only 39-85% agreement rates between two assessments by the same examiner. The inter-examiner agreement was reported in 24 studies. Using stereoscopic photographs, one study found almost perfect agreement after examining seven fields; another study found a substantial to moderate (k = 0.4-0.8) agreement after examining five fields and a third study found a fair agreement (k = 0.2-0.4) after examining a single field. Studies using single- or two-field monoscopic photographs also have reported agreement rates that have varied between almost perfect, substantial and moderate. In four other studies using biomicroscopy, agreement levels varied between perfect and moderate.

CONCLUSIONS

Relative to the large number of studies on the validity of the various methods for screening for DR, there are only few studies of their reliability, with a marked variability in their findings. We suggest that future studies of the effectiveness of the various methods for screening for DR should also include data on their reliability.

Mobile digital fundus screening of type 2 diabetes patients in the Finnish county of South-Ostrobothnia

In 1999 we introduced a digital, mobile eye fundus camera for screening type 2 diabetes mellitus patients (adult onset diabetes) in the Finnish county of South-Ostrobothnia. Prior to this, eye fundus examinations were only available in the regional central hospital, and no systematic screening of type 2 (adult onset) diabetes patients was possible. The total number of screenings carried out with the mobile equipment from 1999 to 2006 was 17,471 cases. There were significant benefits, most important from improved coverage of type 2 diabetes patients. Because the mobile system made eye fundus examinations available to patients in their nearest health centre, approximately 85% of all patients with diabetes in the region are now screened. On average this means that each patient is screened once every 2.5 years, which is in line with national recommendations. As a result, more of the earlier stage retinopathy cases have been diagnosed. There have also been changes in the frequency and type of procedures at the eye department of the regional central hospital. The number of cases referred to the eye department for follow-up studies has decreased from 49 in 2000 to 19 in 2005. The screening system adopted in South-Ostrobothnia has had beneficial effects for both patients and health-care providers.

A sorting system for hierarchical grading of diabetic fundus images: a preliminary study

Diabetic retinopathy is a leading cause of blindness in developed countries. Diabetic patients can prevent severe visual loss by attending regular eye examinations and receiving timely treatments. In the United States, standard protocols have been developed and refined for years to provide better screening and evaluation procedures of the fundus images. Due to the emerging number of diabetic retinopathy cases, accurate and efficient evaluations of the fundus images have become a serious burden for the ophthalmologists or care providers. While diabetic retinopathy remains too complicated to call for an automatic diagnosis system, an efficient tool to facilitate the grading process with a limited number of personnel is in great demand. The current study is to develop a sorting system with a user-friendly interface, based upon the standardized early treatment diabetic retinopathy study (ETDRS) protocol, to assist the professional graders. The raw fundus images will be screened and assigned to different graders according to their skill levels and experiences. The developed hierarchical sorting process will greatly support the graders and enhance their efficiency and throughput. The proposed hybrid intelligent system with multilevel knowledge representation is used to construct this sorting system. A preliminary case study is conducted using only the features of the spot lesion group coupled with the ETDRS standard to demonstrate its feasibility and performance. The results obtained from the case study show a promising future.