An international research collaboration led by Chinese scientists has uncovered a vision-based model that promises to transform eye care worldwide.
This model demonstrates, through an international clinical trial, how artificial intelligence (AI) will soon be able to assist doctors in a variety of scenarios, from primary care clinics to specialty centers.Published in the journal Nature Medicine, the study, led by Tsinghua University and Shanghai Jiao Tong University, shares details about EyeFM, an AI system trained on 14.5 million paired ocular images and clinical texts from a global, multi-ethnic dataset. The system was created by the two universities along with international collaborators.
Forty-four ophthalmologists from China, India, Malaysia, Denmark, Equatorial Guinea, and the United States, working in primary care and specialty clinics, validated its efficacy, highlighting EyeFM's usefulness as a clinical companion.
The team also added a "physician feedback" feature, a rapid loop that makes the model suitable for use in both low-resource clinics and high-complexity specialty hospitals.
Previous similar AI tools tended to learn from only one type of data, unable to process the diverse information that physicians can. These tools typically only examined old records and were not pre-tested in a variety of scenarios and care settings. There are almost no randomized trials, and only a few studies have examined how doctors and AI can work together.
In a double-masked, single-center trial involving 668 high-risk patients in China, 16 ophthalmologists were randomly assigned to use EyeFM or standard care alone during retinal disease screening. In a preliminary analysis, EyeFM support improved diagnostic accuracy to 92.2 percent, compared with 75.4 percent in the control group.
This study provides high-level evidence that large medical AI models can support primary and specialty care, offering a ready-made guideline for transforming AI into an effective tool for everyday clinical
