Related Experiment Video
Updated: Jul 13, 2025
Author Spotlight: An Automated Method for Assessing Visual Acuity in Infants and Toddlers Using an Eye-Tracking System
Published on: March 17, 2023
An Artificial Intelligence System for Screening and Recommending the Treatment Modalities for Retinopathy of
Yaling Liu1, Yueshanyi Du1,2, Xi Wang3,4,5
1Shenzhen Eye Hospital, Jinan University, Shenzhen Eye Institute, Shenzhen, China.
Insights
An artificial intelligence (AI) system effectively identifies retinopathy of prematurity (ROP) severity and recommends treatments. This AI tool shows potential to enhance ROP screening in clinical settings.
Area of Science:
- Ophthalmology
- Medical Imaging
- Artificial Intelligence
Background:
- Retinopathy of prematurity (ROP) is a significant cause of visual impairment in premature infants.
- Accurate and timely diagnosis of ROP is crucial for effective treatment and prevention of vision loss.
- Current diagnostic methods rely on expert ophthalmologist interpretation of retinal images.
Purpose of the Study:
- To develop and evaluate an artificial intelligence (AI) system for identifying ROP disease status.
- To enable the AI system to recommend appropriate treatment modalities for ROP.
- To assess the performance of the AI system in comparison to human experts.
Main Methods:
- A retrospective cohort study utilized 24,495 RetCam images from 651 preterm infants.
- The AI system was trained to perform three tasks: ROP identification, severe ROP identification, and treatment modality identification (laser photocoagulation or intravitreal injections).
- AI performance was evaluated against ophthalmologist performance using 200 independent RetCam images.
Main Results:
- The AI system demonstrated high performance in all tasks, with AUCs of 0.9531 for ROP identification, 0.9132 for severe ROP identification, and 0.9360 for treatment modality identification.
- For treatment modality identification, the AI achieved 86.27% accuracy, 70.59% sensitivity, and 94.12% specificity.
- External validation showed 92.0% accuracy for the AI system across all tasks, outperforming four experienced ophthalmologists (56%-76%).
Conclusions:
- The developed AI system shows promising results for automated ROP severity identification and treatment modality recommendation.
- Algorithmic tools like this AI system have the potential to improve ROP screening efficiency in clinical practice.
- Further integration of AI as an accessory tool could enhance clinical decision-making for ROP management.
Purpose:
The purpose of this study was to develop an artificial intelligence (AI) system for the identification of disease status and recommending treatment modalities for retinopathy of prematurity (ROP).
Methods:
This retrospective cohort study included a total of 24,495 RetCam images from 1075 eyes of 651 preterm infants who received RetCam examination at the Shenzhen Eye Hospital in Shenzhen, China, from January 2003 to August 2021. Three tasks included ROP identification, severe ROP identification, and treatment modalities identification (retinal laser photocoagulation or intravitreal injections). The AI system was developed to identify the 3 tasks, especially the treatment modalities of ROP. The performance between the AI system and ophthalmologists was compared using extra 200 RetCam images.
Results:
The AI system exhibited favorable performance in the 3 tasks, including ROP identification [area under the receiver operating characteristic curve (AUC), 0.9531], severe ROP identification (AUC, 0.9132), and treatment modalities identification with laser photocoagulation or intravitreal injections (AUC, 0.9360). The AI system achieved an accuracy of 0.8627, a sensitivity of 0.7059, and a specificity of 0.9412 for identifying the treatment modalities of ROP. External validation results confirmed the good performance of the AI system with an accuracy of 92.0% in all 3 tasks, which was better than 4 experienced ophthalmologists who scored 56%, 65%, 71%, and 76%, respectively.
Conclusions:
The described AI system achieved promising outcomes in the automated identification of ROP severity and treatment modalities. Using such algorithmic approaches as accessory tools in the clinic may improve ROP screening in the future.
