An Ran Ran1,2, Xi Wang3,4, Poemen P Chan1,2,5, Noel C Chan1,6,7, Wilson Yip1,6,7, Alvin L Young1,6, Mandy O M Wong1,5, Hon-Wah Yung8, Robert T Chang9, Suria S Mannil9, Yih Chung Tham10,11,12, Ching-Yu Cheng10,11,12, Hao Chen13, Fei Li14, Xiulan Zhang14, Pheng-Ann Heng3, Clement C Tham1,2,5, Carol Y Cheung1,2. 1. Department of Ophthalmology and Visual Sciences, The Chinese University of Hong Kong, Hong Kong, Hong Kong SAR, China. 2. Lam Kin Chung. Jet King-Shing Ho Glaucoma Treatment and Research Centre, The Chinese University of Hong Kong, Hong Kong, Hong Kong SAR, China. 3. Department of Computer Science and Engineering, The Chinese University of Hong Kong, Hong Kong, Hong Kong SAR, China. 4. Department of Radiation Oncology, Stanford University School of Medicine, Stanford, Palo Alto, CA, United States. 5. Hong Kong Eye Hospital, Hong Kong, Hong Kong SAR, China. 6. Department of Ophthalmology, Prince of Wales Hospital, Hong Kong, Hong Kong SAR, China. 7. Department of Ophthalmology, Alice Ho Miu Ling Nethersole Hospital, Hong Kong, Hong Kong SAR, China. 8. Tuen Mun Eye Centre, Hong Kong, Hong Kong SAR, China. 9. Department of Ophthalmology, Byers Eye Institute, Stanford University, Palo Alto, CA, United States. 10. Singapore Eye Research Institute, Singapore National Eye Centre, Singapore, Singapore. 11. Duke-National University of Singapore Medical School, Singapore, Singapore. 12. Department of Ophthalmology, Yong Loo Lin School of Medicine, National University of Singapore, Singapore, Singapore. 13. Department of Computer Science and Engineering, The Hong Kong University of Science and Technology, Hong Kong, Hong Kong SAR, China. 14. State Key Laboratory of Ophthalmology, Zhongshan Ophthalmic Centre, Sun Yat-sen University, Guangzhou, China.
Abstract
Purpose: We aim to develop a multi-task three-dimensional (3D) deep learning (DL) model to detect glaucomatous optic neuropathy (GON) and myopic features (MF) simultaneously from spectral-domain optical coherence tomography (SDOCT) volumetric scans. Methods: Each volumetric scan was labelled as GON according to the criteria of retinal nerve fibre layer (RNFL) thinning, with a structural defect that correlated in position with the visual field defect (i.e., reference standard). MF were graded by the SDOCT en face images, defined as presence of peripapillary atrophy (PPA), optic disc tilting, or fundus tessellation. The multi-task DL model was developed by ResNet with output of Yes/No GON and Yes/No MF. SDOCT scans were collected in a tertiary eye hospital (Hong Kong SAR, China) for training (80%), tuning (10%), and internal validation (10%). External testing was performed on five independent datasets from eye centres in Hong Kong, the United States, and Singapore, respectively. For GON detection, we compared the model to the average RNFL thickness measurement generated from the SDOCT device. To investigate whether MF can affect the model's performance on GON detection, we conducted subgroup analyses in groups stratified by Yes/No MF. The area under the receiver operating characteristic curve (AUROC), sensitivity, specificity, and accuracy were reported. Results: A total of 8,151 SDOCT volumetric scans from 3,609 eyes were collected. For detecting GON, in the internal validation, the proposed 3D model had significantly higher AUROC (0.949 vs. 0.913, p < 0.001) than average RNFL thickness in discriminating GON from normal. In the external testing, the two approaches had comparable performance. In the subgroup analysis, the multi-task DL model performed significantly better in the group of "no MF" (0.883 vs. 0.965, p-value < 0.001) in one external testing dataset, but no significant difference in internal validation and other external testing datasets. The multi-task DL model's performance to detect MF was also generalizable in all datasets, with the AUROC values ranging from 0.855 to 0.896. Conclusion: The proposed multi-task 3D DL model demonstrated high generalizability in all the datasets and the presence of MF did not affect the accuracy of GON detection generally.
Purpose: We aim to develop a multi-task three-dimensional (3D) deep learning (DL) model to detect glaucomatous optic neuropathy (GON) and myopic features (MF) simultaneously from spectral-domain optical coherence tomography (SDOCT) volumetric scans. Methods: Each volumetric scan was labelled as GON according to the criteria of retinal nerve fibre layer (RNFL) thinning, with a structural defect that correlated in position with the visual field defect (i.e., reference standard). MF were graded by the SDOCT en face images, defined as presence of peripapillary atrophy (PPA), optic disc tilting, or fundus tessellation. The multi-task DL model was developed by ResNet with output of Yes/No GON and Yes/No MF. SDOCT scans were collected in a tertiary eye hospital (Hong Kong SAR, China) for training (80%), tuning (10%), and internal validation (10%). External testing was performed on five independent datasets from eye centres in Hong Kong, the United States, and Singapore, respectively. For GON detection, we compared the model to the average RNFL thickness measurement generated from the SDOCT device. To investigate whether MF can affect the model's performance on GON detection, we conducted subgroup analyses in groups stratified by Yes/No MF. The area under the receiver operating characteristic curve (AUROC), sensitivity, specificity, and accuracy were reported. Results: A total of 8,151 SDOCT volumetric scans from 3,609 eyes were collected. For detecting GON, in the internal validation, the proposed 3D model had significantly higher AUROC (0.949 vs. 0.913, p < 0.001) than average RNFL thickness in discriminating GON from normal. In the external testing, the two approaches had comparable performance. In the subgroup analysis, the multi-task DL model performed significantly better in the group of "no MF" (0.883 vs. 0.965, p-value < 0.001) in one external testing dataset, but no significant difference in internal validation and other external testing datasets. The multi-task DL model's performance to detect MF was also generalizable in all datasets, with the AUROC values ranging from 0.855 to 0.896. Conclusion: The proposed multi-task 3D DL model demonstrated high generalizability in all the datasets and the presence of MF did not affect the accuracy of GON detection generally.
Authors: Christopher Kai-Shun Leung; Marco Yu; Robert N Weinreb; Heather Kayew Mak; Gilda Lai; Cong Ye; Dennis Shun-Chiu Lam Journal: Invest Ophthalmol Vis Sci Date: 2012-10-17 Impact factor: 4.799
Authors: Michael W Marcus; Margriet M de Vries; Francisco G Junoy Montolio; Nomdo M Jansonius Journal: Ophthalmology Date: 2011-10 Impact factor: 12.079
Authors: Hassan Muhammad; Thomas J Fuchs; Nicole De Cuir; Carlos G De Moraes; Dana M Blumberg; Jeffrey M Liebmann; Robert Ritch; Donald C Hood Journal: J Glaucoma Date: 2017-12 Impact factor: 2.503
Authors: Yih-Chung Tham; Xiang Li; Tien Y Wong; Harry A Quigley; Tin Aung; Ching-Yu Cheng Journal: Ophthalmology Date: 2014-06-26 Impact factor: 12.079