Automated classification of angle-closure mechanisms based on anterior segment optical coherence tomography images via deep learning
Ye Zhang,
Xiaoyue Zhang,
Qing Zhang,
Bin Lv,
Man Hu,
Chuanfeng Lv,
Yuan Ni,
Guotong Xie,
Shuning Li,
Nazlee Zebardast,
Yusrah Shweikh,
Ningli Wang
Affiliations
Ye Zhang
Beijing Tongren Eye Center, Beijing Key Laboratory of Ophthalmology & Visual Science, Beijing Tongren Hospital, Capital Medical University, Beijing, China
Xiaoyue Zhang
Ping an Healthcare Technology, Beijing, China
Qing Zhang
Beijing Institute of Ophthalmology, Beijing, China
Bin Lv
Ping an Healthcare Technology, Beijing, China
Man Hu
National Key Discipline of Pediatrics, Ministry of Education, Department of Ophthalmology, Beijing Children's Hospital, Capital Medical University, China
Chuanfeng Lv
Ping an Healthcare Technology, Beijing, China
Yuan Ni
Ping an Healthcare Technology, Beijing, China
Guotong Xie
Ping an Healthcare Technology, Beijing, China; Ping an Health Cloud Company Limited, Shenzhen, China; Ping an International Smart City Technology Company Limited, Shenzhen, China; Corresponding author. Ping An Healthcare Technology, Beijing, China, 100027; Ping An Health Cloud Company Limited, Shenzhen, China, 518000; Ping An International Smart City Technology Company Limited, Shenzhen, 518000, China.
Shuning Li
Beijing Tongren Eye Center, Beijing Key Laboratory of Ophthalmology & Visual Science, Beijing Tongren Hospital, Capital Medical University, Beijing, China; Corresponding author. Beijing Tongren Eye Center, Beijing Tongren Hospital, Capital Medical University; Beijing Key Laboratory of Ophthalmology and Visual Sciences. No. 1 Dong Jiao Min Xiang Street, Dongcheng District, Beijing, 100730, China.
Nazlee Zebardast
Massachusetts Eye and Ear Infirmary, Harvard Medical School Department of Ophthalmology, Boston, MA, USA
Yusrah Shweikh
Massachusetts Eye and Ear Infirmary, Harvard Medical School Department of Ophthalmology, Boston, MA, USA; Sussex Eye Hospital, University Hospitals Sussex NHS Foundation Trust, Sussex, UK
Ningli Wang
Beijing Tongren Eye Center, Beijing Key Laboratory of Ophthalmology & Visual Science, Beijing Tongren Hospital, Capital Medical University, Beijing, China; Beijing Institute of Ophthalmology, Beijing, China; Corresponding author. Beijing Institute of Ophthalmology, Beijing Tongren Eye Center, Beijing Tongren Hospital, Capital Medical University; Beijing Key Laboratory of Ophthalmology and Visual Sciences. No. 1 Dong Jiao Min Xiang Street, Dongcheng District, Beijing, 100730, China.
Purpose: To develop and validate deep learning algorithms that can identify and classify angle-closure (AC) mechanisms using anterior segment optical coherence tomography (AS-OCT) images. Methods: This cross-sectional study included participants of the Handan Eye Study aged ≥35 years with AC detected via gonioscopy or on the AS-OCT images. These images were classified by human experts into the following to indicate the predominant AC mechanism (ground truth): pupillary block, plateau iris configuration, or thick peripheral iris roll. A deep learning architecture, known as comprehensive mechanism decision net (CMD-Net), was developed to simulate the identification of image-level AC mechanisms by human experts. Cross-validation was performed to optimize and evaluate the model. Human-machine comparisons were conducted using a held-out and separate test sets to establish generalizability. Results: In total, 11,035 AS-OCT images of 1455 participants (2833 eyes) were included. Among these, 8828 and 2.207 images were included in the cross-validation and held-out test sets, respectively. A separate test was formed comprising 228 images of 35 consecutive patients with AC detected via gonioscopy at our eye center. In the classification of AC mechanisms, CMD-Net achieved a mean area under the receiver operating characteristic curve (AUC) of 0.980, 0.977, and 0.988 in the cross-validation, held-out, and separate test sets, respectively. The best-performing ophthalmologist achieved an AUC of 0.903 and 0.891 in the held-out and separate test sets, respectively. And CMD-Net outperformed glaucoma specialists, achieving an accuracy of 89.9 % and 93.0 % compared to 87.0 % and 86.8 % for the best-performing ophthalmologist in the held-out and separate test sets, respectively. Conclusions: Our study suggests that CMD-Net has the potential to classify AC mechanisms using AS-OCT images, though further validation is needed.
