Back to Search
Start Over
Development of Decision Support Software for Deep Learning-Based Automated Retinal Disease Screening Using Relatively Limited Fundus Photograph Data
- Source :
- Electronics, Volume 10, Issue 2, Electronics, Vol 10, Iss 163, p 163 (2021)
- Publication Year :
- 2021
- Publisher :
- MDPI AG, 2021.
-
Abstract
- Purpose&mdash<br />This study was conducted to develop an automated detection algorithm for screening fundus abnormalities, including age-related macular degeneration (AMD), diabetic retinopathy (DR), epiretinal membrane (ERM), retinal vascular occlusion (RVO), and suspected glaucoma among health screening program participants. Methods&mdash<br />The development dataset consisted of 43,221 retinal fundus photographs (from 25,564 participants, mean age 53.38 &plusmn<br />10.97 years, female 39.0%) from a health screening program and patients of the Kangbuk Samsung Hospital Ophthalmology Department from 2006 to 2017. We evaluated our screening algorithm on independent validation datasets. Five separate one-versus-rest (OVR) classification algorithms based on deep convolutional neural networks (CNNs) were trained to detect AMD, ERM, DR, RVO, and suspected glaucoma. The ground truth for both development and validation datasets was graded at least two times by three ophthalmologists. The area under the receiver operating characteristic curve (AUC), sensitivity, and specificity were calculated for each disease, as well as their macro-averages. Results&mdash<br />For the internal validation dataset, the average sensitivity was 0.9098 (95% confidence interval (CI), 0.8660&ndash<br />0.9536), the average specificity was 0.9079 (95% CI, 0.8576&ndash<br />0.9582), and the overall accuracy was 0.9092 (95% CI, 0.8769&ndash<br />0.9415). For the external validation dataset consisting of 1698 images, the average of the AUCs was 0.9025 (95% CI, 0.8671&ndash<br />0.9379). Conclusions&mdash<br />Our algorithm had high sensitivity and specificity for detecting major fundus abnormalities. Our study will facilitate expansion of the applications of deep learning-based computer-aided diagnostic decision support tools in actual clinical settings. Further research is needed to improved generalization for this algorithm.
- Subjects :
- medicine.medical_specialty
genetic structures
diagnosis
Computer Networks and Communications
lcsh:TK7800-8360
Glaucoma
Fundus (eye)
03 medical and health sciences
0302 clinical medicine
Disease Screening
Ophthalmology
medicine
030212 general & internal medicine
Electrical and Electronic Engineering
Receiver operating characteristic
business.industry
fundus
lcsh:Electronics
deep learning
Diabetic retinopathy
Macular degeneration
medicine.disease
Confidence interval
Hardware and Architecture
Control and Systems Engineering
Signal Processing
030221 ophthalmology & optometry
Epiretinal membrane
business
Subjects
Details
- ISSN :
- 20799292
- Volume :
- 10
- Database :
- OpenAIRE
- Journal :
- Electronics
- Accession number :
- edsair.doi.dedup.....3bfec109f6c2bb11dfb97b3b2c93332f