Xinyu Zhao, MD | Chinese Academy of Medical Sciences, Youxin Chen, MD; Erqian Wang, MD; Mingyue Luo, MD; Qing Zhao, MD; Wenfei Zhang, MD; Xingwang, MD; Shiyu Cheng, MD; Zuyi Yang, MD
Competition Sponsor: Chinese Academy of Medical Sciences
Awardee Year: 2025
To address China’s severe shortage of fundus specialists and the disparity in rural medical resources, this project develops an intelligent screening system integrating ultra-widefield (UWF) fundus imaging and a multi-disease deep learning model. By utilizing a 200° UWF view (covering 82% of the retina), the system significantly reduces peripheral lesion miss rates. The AI model, trained on 60,000 images from 26 hospitals, screens for 25 fundus conditions with an average AUC of 0.915. A streamlined rural workflow enables non-mydriatic imaging in one minute, with 5G facilitating cloud-based real-time analysis and AI-driven patient communication. This approach cuts screening costs to 1/5–1/8 of traditional methods, reduces late-stage blindness by 18–22%, and lowers annual per-patient treatment costs by ¥8,500–¥12,000. As the world’s first UWF-AI system covering 25 diseases, it provides a scalable solution for population-wide fundus disease prevention and supports the Healthy China initiative.