AI Classifies Multi-Retinal Diseases

Deep Learning-Based Automated Classification of Multi-Retinal Disease From Fundus Photography

The objective of this study is to establish deep learning (DL) algorithm to automatically classify multi-diseases from fundus photography and differentiate major vision-threatening conditions and other retinal abnormalities. The effectiveness and accuracy of the established algorithm will be evaluated in community derived dataset.

Study Overview

• Status: Unknown
• Conditions: Retinal Diseases; Deep Learning
• Intervention / Treatment: Device: Retinal multi-diseases diagnosed by DL algorithm; Other: Retinal multi-diseases diagnosed by expert panel

Detailed Description

Retinal diseases seriously threaten vision and quality of life, but they often develop insidiously. To date, deep learning (DL) algorithms have shown high prospects in biomedical science, particularly in the diagnosis of ocular diseases, such as diabetic retinopathy, age-related macular degeneration, retinopathy of prematurity, glaucoma, and papilledema. However, there is still a lack of a single algorithm that can classify multi-diseases from fundus photography.

This cross-sectional study will establish a DL algorithm to automatically classify multi-diseases from fundus photography and differentiate major vision-threatening conditions and other retinal abnormalities. We will use the receiver operating characteristic (ROC) curve to examine the ability of recognition and classification of diseases. Taken the results of the expert panel as the gold standard, we will use the evaluation indexes, such as sensitivity, specificity, accuracy, positive predictive value, negative predictive value, etc, to compare the diagnostic capacity between the AI recognition system and human ophthalmologist.

• Study Type: Observational
• Enrollment (Anticipated): 10000

Contacts and Locations

Study Locations

• China
  • Beijing
    • Beijing, Beijing, China, 100730
      • Recruiting
      • Wen-Bin Wei

Participation Criteria

Eligibility Criteria

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

• Sampling Method: Non-Probability Sample

Study Population

community derived dataset

Description

Inclusion Criteria:

• fundus photography around 45° field which covers optic disc and macula
• complete patient identification information;

Exclusion Criteria:

• incomplete patient identification information

Study Plan

How is the study designed?

Number of groups / cohorts

2

Cohorts and Interventions

Group / Cohort	Intervention / Treatment
Retinal multi-diseases diagnosed by DL algorithm	Device: Retinal multi-diseases diagnosed by DL algorithm; DL algorithm automatically classify multi-diseases from fundus photography and differentiate major vision-threatening conditions and other retinal abnormalities.
Retinal multi-diseases diagnosed by expert panel	Other: Retinal multi-diseases diagnosed by expert panel; Expert panel classifies multi-diseases from fundus photography and differentiate major vision-threatening conditions and other retinal abnormalities.

What is the study measuring?

Primary Outcome Measures

Outcome Measure	Measure Description	Time Frame
Area under curve	We will use the receiver operating characteristic (ROC) curve to examine the ability of recognition and classification of diseases. Taken the results of the expert panel as the gold standard, we will use the area under curve to compare the diagnostic capacity between the AI recognition system and human ophthalmologist.	1 week
Sensitivity and specificity	Taken the results of the expert panel as the gold standard, we will use sensitivity and specificity to compare the diagnostic capacity between the AI recognition system and human ophthalmologist.	1 week
Positive and negative predictive value	Taken the results of the expert panel as the gold standard, we will use positive and negative predictive value to compare the diagnostic capacity between the AI recognition system and human ophthalmologist.	1 week
Accuracy	Taken the results of the expert panel as the gold standard, we will use accuracy to compare the diagnostic capacity between the AI recognition system and human ophthalmologist.	1 week

Collaborators and Investigators

• Sponsor: Beijing Tongren Hospital
• Collaborators: Beijing Tulip Partner Technology Co., Ltd, China

Study record dates

Study Major Dates

• Study Start (Actual): November 1, 2020
• Primary Completion (Anticipated): November 1, 2021
• Study Completion (Anticipated): December 1, 2021

Study Registration Dates

• First Submitted: October 13, 2020
• First Submitted That Met QC Criteria: October 15, 2020
• First Posted (Actual): October 19, 2020

Study Record Updates

• Last Update Posted (Actual): December 11, 2020
• Last Update Submitted That Met QC Criteria: December 9, 2020
• Last Verified: October 1, 2020

More Information

Terms related to this study

• Additional Relevant MeSH Terms: Eye Diseases; Retinal Diseases
• Other Study ID Numbers: Retinal multi diseases

Plan for Individual participant data (IPD)

• Plan to Share Individual Participant Data (IPD)?: UNDECIDED

Drug and device information, study documents

• Studies a U.S. FDA-regulated drug product: Yes
• Studies a U.S. FDA-regulated device product: No
• product manufactured in and exported from the U.S.: No