Automatic segmentation of corneal deposits from corneal stromal dystrophy images via deep learning. (October 2021)

• Record Type: Journal Article
• Authors: Deshmukh, Mihir; Liu, Yu-Chi; Rim, Tyler Hyungtaek; Venkatraman, Anandalakshmi; Davidson, Matthew; Yu, Marco; Kim, Hong Seok; Lee, Geunyoung; Jun, Ikhyun; Mehta, Jodhbir S.; Kim, Eung Kweon
• Is Part Of: Computers in biology and medicine. Volume 137(2021)
• Publication Date: 2021-10
• Subjects: Ophthalmology -- Cornea -- Corneal stromal dystrophy -- Automated segmentation -- Deep learning -- Slit-lamp photographs -- Real-world setting; Medicine -- Data processing -- Periodicals; Biology -- Data processing -- Periodicals; 610.285
• DOI: 10.1016/j.compbiomed.2021.104675
• Languages: English
• ISSNs: 0010-4825

Abstract

Abstract: Background: Granular dystrophy is the most common stromal dystrophy. To perform automated segmentation of corneal stromal deposits, we trained and tested a deep learning (DL) algorithm from patients with corneal stromal dystrophy and compared its performance with human segmentation. Methods: In this retrospective cross-sectional study, we included slit-lamp photographs by sclerotic scatter from patients with corneal stromal dystrophy and real-world slit-lamp photographs via various techniques (diffuse illumination, tangential illumination, and sclerotic scatter). Our data set included 1007 slit-lamp photographs of semi-automatically generated handcraft masks on granular and linear lesions from corneal stromal dystrophy patients (806 for the training set and 201 for test set). For external test (140 photographs), we applied the DL algorithm and compared between automated and human segmentation. For performance, we estimated the intersection of union (IoU), global accuracy, and boundary F1 (BF) score for segmentation. Results: In 201 internal test set, IoU, global accuracy, and BF score with 95 % confidence Interval were 0.81 (0.79–0.82), 0.99 (0.98–0.99), and 0.93 (0.92–0.95), respectively. In 140 heterogenous external test set as a real-world data, those were 0.64 (0.61–0.67), 0.95 (0.94–0.96), and 0.70 (0.64–0.76) via DL algorithm and 0.56 (0.51–0.61), 0.95 (0.94–0.96), and 0.70 (0.65–0.74) via human rater, respectively. Conclusions: We developed an automated segmentation DL algorithm for corneal stromal deposits in patients with corneal stromal dystrophy. Segmentation on corneal deposits was accurate via the DL algorithm in the well-controlled dataset and showed reasonable performance in a real-world setting. We suggest this automatic segmentation of corneal deposits helps to monitor the disease and can evaluate possible new treatments. Highlights: Automated segmentation of corneal stromal deposits was developed for patients with corneal stromal dystrophy. The deep-learning algorithm was able to segment corneal deposits accurately in the well-controlled dataset. The algorithm gave a reasonable performance in a real-world setting, comparable to the human rater. Automatic segmentation of corneal deposits may help to monitor the disease and can evaluate possible new treatments.