Fangyao Tang1, Phoomraphee Luenam2, An Ran Ran1, Ahmed Abdul Quadeer2, Rajiv Raman3, Piyali Sen4, Rehana Khan3, Anantharaman Giridhar5, Swathy Haridas5, Matias Iglicki6, Dinah Zur7, Anat Loewenstein8, Hermino P Negri9, Simon Szeto10, Bryce Ka Yau Lam11, Clement C Tham1, Sobha Sivaprasad4, Matthew Mckay2, Carol Y Cheung12. 1. Department of Ophthalmology and Visual Sciences, The Chinese University of Hong Kong, Hong Kong, China. 2. Department of Electronic and Computer Engineering, The Hong Kong University of Science and Technology, Hong Kong, China. 3. Shri Bhagwan Mahavir Vitreoretinal Services, Sankara Nethralaya, Chennai, India. 4. Moorfields Eye Hospital, London, United Kingdom. 5. Giridhar Eye Institute, Cochin, India. 6. Private Retina Practice, University of Buenos Aires, Buenos Aires, Argentina; Tel Aviv Medical Center, Tel Aviv, Israel; Sackler Faculty of Medicine, Tel Aviv University, Tel Aviv, Israel. 7. Tel Aviv Medical Center, Tel Aviv, Israel; Sackler Faculty of Medicine, Tel Aviv University, Tel Aviv, Israel. 8. Sackler Faculty of Medicine, Tel Aviv University, Tel Aviv, Israel; Ophthalmology Division, Tel Aviv Sourasky Medical Center, Tel Aviv, Israel. 9. Diagnostic Ophthalmology Centre, Buenos Aires, Argentina. 10. Department of Ophthalmology and Visual Sciences, The Chinese University of Hong Kong, Hong Kong, China; Hong Kong Eye Hospital, Hong Kong, China. 11. Princess Margaret Hospital, Hong Kong, China. 12. Department of Ophthalmology and Visual Sciences, The Chinese University of Hong Kong, Hong Kong, China. Electronic address: carolcheung@cuhk.edu.hk.
Abstract
PURPOSE: To develop a deep learning (DL) system that can detect referable diabetic retinopathy (RDR) and vision-threatening diabetic retinopathy (VTDR) from images obtained on ultra-widefield scanning laser ophthalmoscope (UWF-SLO). DESIGN: Observational, cross-sectional study. PARTICIPANTS: A total of 9392 UWF-SLO images of 1903 eyes from 1022 subjects with diabetes from Hong Kong, the United Kingdom, India, and Argentina. METHODS: All images were labeled according to the presence or absence of RDR and the presence or absence of VTDR. Labeling was performed by retina specialists from fundus examination, according to the International Clinical Diabetic Retinopathy Disease Severity Scale. Three convolutional neural networks (ResNet50) were trained with a transfer-learning procedure for assessing gradability and identifying VTDR and RDR. External validation was performed on 4 datasets spanning different geographical regions. MAIN OUTCOME MEASURES: Area under the receiver operating characteristic curve (AUROC); area under the precision-recall curve (AUPRC); sensitivity, specificity, and accuracy of the DL system in gradability assessment; and detection of RDR and VTDR. RESULTS: For gradability assessment, the system achieved an AUROC of 0.923 (95% confidence interval [CI], 0.892-0.947), sensitivity of 86.5% (95% CI, 77.6-92.8), and specificity of 82.1% (95% CI, 77.3-86.2) for the primary validation dataset, and >0.82 AUROCs, >79.6% sensitivity, and >70.4% specificity for the geographical external validation datasets. For detecting RDR and VTDR, the AUROCs were 0.981 (95% CI, 0.977-0.984) and 0.966 (95% CI, 0.961-0.971), with sensitivities of 94.9% (95% CI, 92.3-97.9) and 87.2% (95% CI, 81.5-91.6), specificities of 95.1% (95% CI, 90.6-97.9) and 95.8% (95% CI, 93.3-97.6), and positive predictive values (PPVs) of 98.0% (95% CI, 96.1-99.0) and 91.1% (95% CI, 86.3-94.3) for the primary validation dataset, respectively. The AUROCs and accuracies for detecting both RDR and VTDR were >0.9% and >80%, respectively, for the geographical external validation datasets. The AUPRCs were >0.9, and sensitivities, specificities, and PPVs were >80% for the geographical external validation datasets for RDR and VTDR detection. CONCLUSIONS: The excellent performance achieved with this DL system for image quality assessment and detection of RDR and VTDR in UWF-SLO images highlights its potential as an efficient and effective diabetic retinopathy screening tool.
PURPOSE: To develop a deep learning (DL) system that can detect referable diabetic retinopathy (RDR) and vision-threatening diabetic retinopathy (VTDR) from images obtained on ultra-widefield scanning laser ophthalmoscope (UWF-SLO). DESIGN: Observational, cross-sectional study. PARTICIPANTS: A total of 9392 UWF-SLO images of 1903 eyes from 1022 subjects with diabetes from Hong Kong, the United Kingdom, India, and Argentina. METHODS: All images were labeled according to the presence or absence of RDR and the presence or absence of VTDR. Labeling was performed by retina specialists from fundus examination, according to the International Clinical Diabetic Retinopathy Disease Severity Scale. Three convolutional neural networks (ResNet50) were trained with a transfer-learning procedure for assessing gradability and identifying VTDR and RDR. External validation was performed on 4 datasets spanning different geographical regions. MAIN OUTCOME MEASURES: Area under the receiver operating characteristic curve (AUROC); area under the precision-recall curve (AUPRC); sensitivity, specificity, and accuracy of the DL system in gradability assessment; and detection of RDR and VTDR. RESULTS: For gradability assessment, the system achieved an AUROC of 0.923 (95% confidence interval [CI], 0.892-0.947), sensitivity of 86.5% (95% CI, 77.6-92.8), and specificity of 82.1% (95% CI, 77.3-86.2) for the primary validation dataset, and >0.82 AUROCs, >79.6% sensitivity, and >70.4% specificity for the geographical external validation datasets. For detecting RDR and VTDR, the AUROCs were 0.981 (95% CI, 0.977-0.984) and 0.966 (95% CI, 0.961-0.971), with sensitivities of 94.9% (95% CI, 92.3-97.9) and 87.2% (95% CI, 81.5-91.6), specificities of 95.1% (95% CI, 90.6-97.9) and 95.8% (95% CI, 93.3-97.6), and positive predictive values (PPVs) of 98.0% (95% CI, 96.1-99.0) and 91.1% (95% CI, 86.3-94.3) for the primary validation dataset, respectively. The AUROCs and accuracies for detecting both RDR and VTDR were >0.9% and >80%, respectively, for the geographical external validation datasets. The AUPRCs were >0.9, and sensitivities, specificities, and PPVs were >80% for the geographical external validation datasets for RDR and VTDR detection. CONCLUSIONS: The excellent performance achieved with this DL system for image quality assessment and detection of RDR and VTDR in UWF-SLO images highlights its potential as an efficient and effective diabetic retinopathy screening tool.
Authors: Mohamed Ashraf; Jerry D Cavallerano; Jennifer K Sun; Paolo S Silva; Lloyd Paul Aiello Journal: J Clin Med Date: 2021-07-27 Impact factor: 4.964