Detecting diabetic retinopathy using an artificial intelligence-based software platform: a pilot study

Abstract

Purpose: To examine the potential for the detection of diabetic retinopathy (DR) using the artificial intelligence (AI)-based software platform Retina-AI CheckEye©. Material and Methods: This was an open-label, prospective, pilot observational case-control study for the detection of DR using an AI-based software platform. The study was conducted at the sites of healthcare facilities in Chernivtsi oblast. Four hundred and eight diabetics and 256 non-diabetic controls were involved in the study. All fundus images were analyzed using the artificial intelligence (AI)-based software platform Retina-AI CheckEye©. Receiver operating characteristic (ROC) curve analysis was performed to determine the sensitivity and specificity of the DR diagnosis method. Results: Using the AI-based software platform, signs of DR in at least one eye were detected in 143 diabetics (22% of total study subjects (664 individuals; 1328 eyes) or 35% of the diabetics (408 patients)). No DR signs were detected in 322 individuals (48% of total study subjects). In 199 individuals (30% of total study subjects), the results were not obtained due to the features of the optical media and presence of certain eye diseases (in most cases, unilateral cataract or corneal opacity). This trial found 93% sensitivity and 86% specificity for the Retina-AI CheckEye-assisted detection of DR. Conclusion: An AI-based software platform, Retina-AI CheckEye©, has been for the first time developed in Ukraine. The platform was demonstrated to have a high accuracy (93% sensitivity and 86% specificity) in diagnosing DR in diabetic patients and can be used for large-scale DR screening.

Keywords

• diabetes mellitus
• diabetic retinopathy
• artificial intelligence