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

Abstract

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

Assessing Diabetic Retinopathy Staging With AI: A Comparative Analysis Between Pseudocolor and LED Imaging

“… 12 The deployment of AI algorithms has been associated with high sensitivity and specificity in detecting RDR based on conventional color fundus photographs. 13 – 15 However, the integration of AI models into other existing camera systems, especially UWF imaging, poses significant technical challenges that demand careful considerations, 16 and the performance of these algorithms across different imaging modalities remains an area of limited understanding. Our study aimed to bridge this gap by comparing the performance of the AI algorithm across LED imaging and pseudocolor modalities, both with and without SWL.…”

Recommendations for initial diabetic retinopathy screening of diabetic patients using large language model-based artificial intelligence in real-life case scenarios

“…According to the ADA, AI can be used as an alternative to traditional screening methods in DR [ 29 ]. AI’s current role in DR is to screen retinal images for the presence or absence of DR or sight-threatening DR [ 30 ]. However, AI should not be used in patients who have known DR, have received prior DR treatment, or have symptoms of vision impairment.…”

Variability in Grading Diabetic Retinopathy Using Retinal Photography and Its Comparison with an Automated Deep Learning Diabetic Retinopathy Screening Software

“…Multiple studies have consistently reported the favourable attributes of automated deep learning diabetic retinopathy screening software, including its high accuracy, time efficiency, and cost-effectiveness, in comparison to human graders [ 26 , 30 , 31 , 32 ]. However, despite the enhancement of precision in diabetic retinopathy screening, several real-world challenges remain, including practical obstacles such as the use of mydriasis to enhance image quality, which has the potential to cause an angle closure attack [ 17 ]. Additionally, integrating this technology into healthcare systems raises concerns about the accuracy of diagnosis, and the governance of artificial intelligence in healthcare must adhere to established guidelines for fairness, transparency, trustworthiness, and accountability to protect the interests of all stakeholders [ 17 ].…”

“…There are numerous commercially available diabetic retinopathy screening algorithms worldwide, including IDx-DR, RetmarkerDR, EyeArt, Singapore SERI-NUS, Google Inc, Bosch DR algorithm, Retinalyze, and Messidor-2 [ 12 , 17 ]. These screening algorithms, including the DiaRetDB algorithm utilised in this study, achieve a desirable equilibrium between high sensitivity and specificity.…”

“…This technology can assist ophthalmologists in diagnosing and monitoring the progression of diabetic retinopathy, enabling earlier and more effective treatment [ 14 , 15 ]. However, the effectiveness of these diabetic retinopathy screening methods remains debatable [ 11 , 12 , 16 , 17 ], with retinal photography still being the preferred screening test for early diabetic retinopathy detection [ 18 ]. Typically, the grading of these photographs is performed by non-physicians [ 19 ], healthcare workers of different specialties and levels of expertise, including optometrist [ 16 , 20 ], ophthalmologists, endocrinologists [ 16 ] and general practitioners [ 16 ].…”