AI reading label system enhances retinal disease diagnosis and training in ophthalmology

(Image Credit: AdobeStock/Who is Danny)

A new Chinese study reported that an artificial intelligence (AI) reading label system was found to enhance the diagnostic accuracy of retinal diseases among ophthalmologists and holds potential for integration into future medical education,¹ according to first author Meng Wang, MD, from the Department of Ophthalmology, Peking Union Medical College Hospital, and the Beijing Key Laboratory of Fundus Diseases Intelligent Diagnosis & Drug/Device Development and Translation, both in Beijing.

The authors explained the rationale for their study. “Ophthalmology is an ideal field for the application of AI, given that the diagnosis of ophthalmic diseases predominantly relies on image-based information.² To develop accurate and reliable AI models capable of diagnosing diseases, it is imperative to have a substantial volume of meticulously annotated image data. The quality of annotated data is crucial for training AI models with precise diagnostic capabilities, underscoring the significance of training annotators effectively,³” they commented.

AI study methodology

In their study, 16 ophthalmologists, including attending physicians and residents with levels of experience ranging from 1 to 9 years, were included. This approach both helps to develop AI models and provide practical training for ophthalmologists by exposing them to a wide array of specific disease cases and corresponding images. It is important to investigate the potential of the annotation process on the training of ophthalmologists, Weng and colleagues explained.

This multicenter study used 2 imaging modalities: optical coherence tomography (OCT) and color fundus photography (CFP) images to enhance understanding of retinal diseases and improve the efficiency of disease screening and diagnosis.⁴

The investigators evaluated the training effect of the AI annotation process on relatively junior ophthalmologists. “The findings may offer valuable insights into medical education in OCT and CFP learning in various retinal diseases,” they said.

The investigators loaded 7,777 pairs of OCT and CFP images centered on the macular region into the reading label system. The participants were divided into eight groups, and each group was assigned a senior ophthalmologist who checked the annotation results and provided standard diagnoses. All images were assigned to each group in five rounds.

The retinal diseases that were included were diabetic retinopathy (DR), retinal detachment (RD), retinal vein occlusion (RVO), dry age-related macular degeneration (AMD), wet AMD, epiretinal membrane (ERM), central serous retinopathy, macular schisis (MS), and macular hole (MH); images of normal fundi also were included.

In the reading label system, after the participants provided their diagnoses based on only the OCT and only the CFP images, they then provided the final case diagnosis (bimodal diagnosis) based on both the OCT and CFP images.

Wang and colleagues reported, “The average diagnostic accuracy for the nine retinal diseases and normal fundi improved significantly across the five rounds (p = 0.013) and is closely correlated to the duration of ophthalmology study (p = 0.007). Furthermore, significant improvements were observed in the diagnostic accuracy of both OCT (p = 0.028) and CFP (p = 0.021) modalities as the number of rounds increased. Notably, OCT single modal diagnosis demonstrated higher consistency with the final diagnosis in cases of RD, ERM, MS, and MH compared to CFP, while CFP single modal diagnosis has higher consistency in DR, RVO, and normal fundus.”

They concluded, “The AI reading label system contributes to improving the diagnostic accuracy of retinal diseases among ophthalmologists. It holds potential for widespread application in future medical education in ophthalmology.”

References

1. Wang M, Zhang X, Li D, et al. The potential of artificial intelligence reading label system on the training of ophthalmologists in retinal diseases, a multicenter bimodal multi-disease study. BMC Med Educ. 2025;25:503. https://doi.org/10.1186/s12909-025-07066-1

2. Li JO, Liu H, Ting DSJ, et al. Digital technology, telemedicine and artificial intelligence in ophthalmology: A global perspective. Prog Retin Eye Res. 2021;82:100900.

3. Hasei J, Nakahara R, Otsuka Y, et al. High quality expert annotations enhance artificial intelligence model accuracy for osteosarcoma X-ray diagnosis. Cancer Sci. 2024;115:3695–704. Wang et al. BMC Med Educ. 2025;25:503

4. Wang S, He X, Jian Z, et al. Advances and prospects of multi-modal ophthalmic artificial intelligence based on deep learning: a review. Eye Vis (Lond). 2024;11:38.