Development and Validation of a Deep Learning System for Multiple Ocular Fundus Diseases Using Retinal Images

Development and Validation of a Deep Learning System for Multiple Ocular Fundus Diseases Using Retinal Images: a Multi-center Prospective Study

Retinal images can reflect both fundus and systemic conditions (diabetes and cardiovascular disease) and firstly to be used for medical artificial intelligence (AI) algorithm training due to its advantages of clinical significance and easy to obtain. Here, the investigators developed a single network model that can mine the characteristics among multiple fundus diseases, which was trained by plenty of fundus images with one or several disease labels (if they have) in each of them. The model performance was compared with those of both native and international ophthalmologists. The model was further tested by datasets with different camera types and validated by three external datasets prospectively collected from the clinical sites where the model would be applied.

Study Overview

• Status: Unknown
• Conditions: Ophthalmological Disorder
• Intervention / Treatment: Other: diagnostic

• Study Type: Observational
• Enrollment (Anticipated): 300000

Contacts and Locations

Study Locations

• China
  • Guangdong
    • Guangzhou, Guangdong, China, 510060
      • Recruiting
      • Zhongshan Ophthalmic Center, Sun Yat-sen Univerisity
      • Contact: Haotian Lin, Ph.D; Phone Number: +86-020-87330274; Email: gddlht@aliyun.com
      • Principal Investigator: Haotian Lin, Ph.D

Participation Criteria

Eligibility Criteria

• Ages Eligible for Study: Child; Adult; Older Adult
• Accepts Healthy Volunteers: No
• Genders Eligible for Study: All

• Sampling Method: Probability Sample

Study Population

Retinal images were collected from different health care institutes all over China and other countries around the world.

Description

Inclusion Criteria:

• The quality of fundus images should clinical acceptable. More than 80% of the fundus image area including four main regions (optic disk, macular, upper and lower retinal vessel archs) are easy to read and discriminate.

Exclusion Criteria:

• Images with light leakage (>30% of area), spots from lens flares or stains, and overexposure were excluded from further analysis.

Study Plan

How is the study designed?

Design Details

• Observational Models: Other
• Time Perspectives: Other

Number of groups / cohorts

3

Cohorts and Interventions

Group / Cohort	Intervention / Treatment
Training dataset; Retinal images collected from hospitals and multiple screening sites all over China
Validation dataset; Retinal images separated from training dataset	Other: diagnostic; Training dataset was used to train the deep learning model, which was validated and tested by other two datasets.
Testing dataset; Retinal images prospectively collected from the hospitals and ocular disease screening sites totally different from training dataset	Other: diagnostic; Training dataset was used to train the deep learning model, which was validated and tested by other two datasets.

What is the study measuring?

Primary Outcome Measures

Outcome Measure	Measure Description	Time Frame
Area under the receiver operating characteristic curve of the deep learning system	The investigators will calculate the area under the receiver operating characteristic curve of deep learning system and compare this index between deep learning system and human doctors.	baseline

Secondary Outcome Measures

Outcome Measure	Measure Description	Time Frame
Sensitivity of the deep learning system	The investigators will calculate the sensitivity of deep learning system and compare this index between deep learning system and human doctors.	baseline
Specificity of the deep learning system	The investigators will calculate the specificity of deep learning system and compare this index between deep learning system and human doctors.	baseline

Collaborators and Investigators

• Sponsor: Sun Yat-sen University

Publications and helpful links

General Publications

• Lin D, Xiong J, Liu C, Zhao L, Li Z, Yu S, Wu X, Ge Z, Hu X, Wang B, Fu M, Zhao X, Wang X, Zhu Y, Chen C, Li T, Li Y, Wei W, Zhao M, Li J, Xu F, Ding L, Tan G, Xiang Y, Hu Y, Zhang P, Han Y, Li JO, Wei L, Zhu P, Liu Y, Chen W, Ting DSW, Wong TY, Chen Y, Lin H. Application of Comprehensive Artificial intelligence Retinal Expert (CARE) system: a national real-world evidence study. Lancet Digit Health. 2021 Aug;3(8):e486-e495. doi: 10.1016/S2589-7500(21)00086-8.

Study record dates

Study Major Dates

• Study Start (Actual): January 1, 2014
• Primary Completion (Anticipated): February 1, 2020
• Study Completion (Anticipated): May 1, 2020

Study Registration Dates

• First Submitted: December 23, 2019
• First Submitted That Met QC Criteria: December 24, 2019
• First Posted (Actual): December 30, 2019

Study Record Updates

• Last Update Posted (Actual): December 30, 2019
• Last Update Submitted That Met QC Criteria: December 24, 2019
• Last Verified: December 1, 2019

More Information

Terms related to this study

• Additional Relevant MeSH Terms: Eye Diseases
• Other Study ID Numbers: CCPMOH2019- China8

Plan for Individual participant data (IPD)

• Plan to Share Individual Participant Data (IPD)?: No

Drug and device information, study documents

• Studies a U.S. FDA-regulated drug product: No
• Studies a U.S. FDA-regulated device product: No