Deep Learning Models for Segmenting Non-perfusion Area of Color Fundus Photographs in Patients With Branch Retinal Vein Occlusion

Jinxin Miao; Jiale Yu; Wenjun Zou; Na Su; Zongyi Peng; Xinjing Wu; Junlong Huang; Yuan Fang; Songtao Yuan; Ping Xie; Kun Huang; Qiang Chen; Zizhong Hu; Qinghuai Liu

doi:10.3389/fmed.2022.794045

Deep Learning Models for Segmenting Non-perfusion Area of Color Fundus Photographs in Patients With Branch Retinal Vein Occlusion

Front Med (Lausanne). 2022 Jun 30:9:794045. doi: 10.3389/fmed.2022.794045. eCollection 2022.

Authors

Jinxin Miao¹, Jiale Yu², Wenjun Zou^{1

3}, Na Su¹, Zongyi Peng⁴, Xinjing Wu¹, Junlong Huang¹, Yuan Fang¹, Songtao Yuan¹, Ping Xie¹, Kun Huang², Qiang Chen², Zizhong Hu¹, Qinghuai Liu¹

Affiliations

¹ Department of Ophthalmology, The First Affiliated Hospital of Nanjing Medical University, Nanjing, China.
² School of Computer Science and Engineering, Nanjing University of Science & Technology, Nanjing, China.
³ Department of Ophthalmology, The Affiliated Wuxi No.2 People's Hospital of Nanjing Medical University, Wuxi, China.
⁴ The First School of Clinical Medicine, Nanjing Medical University, Nanjing, China.

Abstract

Purpose: To develop artificial intelligence (AI)-based deep learning (DL) models for automatically detecting the ischemia type and the non-perfusion area (NPA) from color fundus photographs (CFPs) of patients with branch retinal vein occlusion (BRVO).

Methods: This was a retrospective analysis of 274 CFPs from patients diagnosed with BRVO. All DL models were trained using a deep convolutional neural network (CNN) based on 45 degree CFPs covering the fovea and the optic disk. We first trained a DL algorithm to identify BRVO patients with or without the necessity of retinal photocoagulation from 219 CFPs and validated the algorithm on 55 CFPs. Next, we trained another DL algorithm to segment NPA from 104 CFPs and validated it on 29 CFPs, in which the NPA was manually delineated by 3 experienced ophthalmologists according to fundus fluorescein angiography. Both DL models have been cross-validated 5-fold. The recall, precision, accuracy, and area under the curve (AUC) were used to evaluate the DL models in comparison with three types of independent ophthalmologists of different seniority.

Results: In the first DL model, the recall, precision, accuracy, and area under the curve (AUC) were 0.75 ± 0.08, 0.80 ± 0.07, 0.79 ± 0.02, and 0.82 ± 0.03, respectively, for predicting the necessity of laser photocoagulation for BRVO CFPs. The second DL model was able to segment NPA in CFPs of BRVO with an AUC of 0.96 ± 0.02. The recall, precision, and accuracy for segmenting NPA was 0.74 ± 0.05, 0.87 ± 0.02, and 0.89 ± 0.02, respectively. The performance of the second DL model was nearly comparable with the senior doctors and significantly better than the residents.

Conclusion: These results indicate that the DL models can directly identify and segment retinal NPA from the CFPs of patients with BRVO, which can further guide laser photocoagulation. Further research is needed to identify NPA of the peripheral retina in BRVO, or other diseases, such as diabetic retinopathy.

Keywords: artificial intelligence; automatic segmentation; branch retinal vein occlusion; color fundus photograph; deep learning; non-perfusion area.