Establishment of an Automatic Diagnosis System for Corneal Endothelium Diseases Using Artificial Intelligence: A Retrospective, Large-Scale, Multicentre, Diagnostic Accuracy Study

Background: Corneal endothelium disease (CED) is a blinding eye disease whose incidence has increased in recent years, placing a large economic burden on society worldwide. Early detection and diagnosis are very important for the prevention and treatment of CED. In vivo confocal microscopy (IVCM) is a useful tool for diagnosing CED. However, the ability to read IVCM images is limited, and IVCM is often not used effectively. The ability to diagnose CED remains low in China.Methods: we developed an automatic system for detecting multiple common CEDs by introducing an enhanced compact convolutional transformer (ECCT). Specifically, we introduce a cross-head relative position encoding scheme into standard self-attention to capture contextual information among different regions and employ a token-attention feed-forward network to place greater focus on valuable abnormal regions.Findings: A total of 2723 CED images were used to train our system. It achieved an accuracy of 89.53%, and the area under the receiver operating characteristic curve (AUC) was 0.958 (95% CI, 0.943-0.971) with testing CED images obtained from multiple centres.Interpretation: Our system is the first artificial intelligence-based system for diagnosing CED worldwide. Images can be uploaded to a specified website, and automatic diagnoses can be obtained, which can be particularly helpful under pandemic conditions such as those seen during the recent COVID-19 pandemic.Funding: This work was supported by the Peking University Medicine Sailing Program for Young Scholars’ Scientific & Technological Innovation under Grant No. BMU2023YFJHPY018 and by the National Natural Science Foundation of China under Grant Nos. 81970768 and 81800801.Declaration of Interest: None to declare. Ethical Approval: The study was performed according to the tenets of the Declaration of Helsinki and was approved by the institutional review board of Peking University Third Hospital (IRB00006761-M2022834). All participants provided written informed consent to take part in the study.