AI Foundation Model Improves Ophthalmologist Diagnostic Accuracy in Randomized Controlled Trial

jameslk

9mo ago· 23 min readenInsight

100/100

Golden Brown

Bagelometer↗

Master baker tier. Every paragraph earns its place on the tray.

Score100TypeanalysisSentimentpositive

Summary

Researchers developed EyeFM, a multimodal vision-language eyecare foundation model trained on 14.5 million ocular images from global datasets. In a randomized controlled trial with 668 participants and 16 ophthalmologists, the AI copilot significantly improved diagnostic accuracy (92.2% vs 75.4%), referral rates (92.2% vs 80.5%), and patient compliance with follow-up recommendations. The study demonstrated that AI assistance can enhance ophthalmologist performance and patient outcomes in retinal disease screening.

Key quotes

· 4 pulled

Ophthalmologists with EyeFM copilot achieved higher correct diagnostic rate (92.2% versus 75.4%, P < 0.001) and referral rate (92.2% versus 80.5%, P < 0.001)

The intervention group demonstrated higher compliance with self-management (70.1% versus 49.1%, P < 0.001) and referral suggestions (33.7% versus 20.2%, P < 0.001) at follow-up

Our study provided evidence that implementing EyeFM copilot can improve the performance of ophthalmologists and the outcome of patients

Trained and validated on multimodal data from 14.5 million images from multicountry datasets, a foundation model is shown to increase diagnostic and referral accuracy of clinicians

Snippet from the RSS feed

In the context of an increasing need for clinical assessments of foundation models, we developed EyeFM, a multimodal vision–language eyecare copilot, and conducted a multifaceted evaluation, including retrospective validations, multicountry efficacy valid

You might also wanna read

Feyenally offers AI-guided smartphone vision tests for refractive error detection at home

Feyenally is a smartphone-based AI-guided vision testing tool that allows users to check for refractive errors at home without needing extra

Product Hunt·1mo ago

Tempus AI presents initial multimodal foundation model results for oncology insight generation at ASCO 2026

Tempus AI announced initial results from its Multimodal Foundation Model efforts at the 2026 ASCO Annual Meeting. The company is building no

stocktitan.net·2d ago

Study Shows Doctors' Cancer Detection Skills Decline Without AI Assistance

A new study reveals that doctors who frequently use AI for cancer detection during colonoscopies performed worse when AI was not available,

The Verge·9mo ago