Application of Artificial Intelligence and Deep Learning for Choroid Segmentation in Myopia

Abstract

Purpose To investigate the correlation between choroidal thickness and myopia progression using a deep learning method. Methods Two data sets, data set A and data set B, comprising of 123 optical coherence tomography (OCT) volumes, were collected to establish the model and verify its clinical utility. The proposed mask region-based convolutional neural network (R-CNN) model, trained with the pretrained weights from the Common Objects in Context database as well as the manually labeled OCT images from data set A, was used to automatically segment the choroid. To verify its clinical utility, the mask R-CNN model was tested with data set B, and the choroidal thickness estimated by the model was also used to explore its relationship with myopia. Results Compared with the result of manual segmentation in data set B, the error of the automatic choroidal inner and outer boundary segmentation was 6.72 ± 2.12 and 13.75 ± 7.57 µm, respectively. The mean dice coefficient between the region segmented by automatic and manual methods was 93.87% ± 2.89%. The mean difference in choroidal thickness over the Early Treatment Diabetic Retinopathy Study zone between the two methods was 10.52 µm. Additionally, the choroidal thickness estimated using the proposed model was thinner in high-myopic eyes, and axial length was the most significant predictor. Conclusions The mask R-CNN model has excellent performance in choroidal segmentation and quantification. In addition, the choroid of high myopia is significantly thinner than that of nonhigh myopia. Translational Relevance This work lays the foundations for mask R-CNN models that could aid in the evaluation of more intricate changes occurring in chorioretinal diseases.

IMI 2023 Digest

“…There has been a focus on imaging and segmenting (particularly using machine learning) choroidal thickness using A-scan (traditionally termed biometers ) 133 and B-scan (termed o ptical c oherence t omography ) 131 , 134 , 135 based techniques to better understand the mechanism(s) of myopia control and also as a potential predictor of long-term efficacy of treatments (see IMI 2023 white paper on choroid). 132 A model using principally baseline pupil area, 1-month change of the zone 3-mm (flat) and zone 5-mm (flat/steep) keratometry was able to predict between 54% and 63% of the variation in 1-year AL elongation with ortho-k. 136 In addition, models have been developed to predict cycloplegic refractive error from demographics, noncycloplegic SER, AL/corneal curvature radius ratio, uncorrected VA, and intraocular pressure, with the results explaining 92% to 93% of the variability in Chinese school children (aged 5–18 years) 137 , 138 and 96% in children in Japan (aged 2–9 years).…”

Choroidal layer segmentation in OCT images by a boundary enhancement network

“…They found that the retina of patients with PD became thinner, while choroidal thickness might increase. Similar to our work, Chen et al (2022) segmented the choroidal layer of highly myopic patients and non highly myopic people and compared the thickness, while they lacked the analysis of three-dimensional features, and the segmentation performance needs to be improved.…”

A Deep Learning–Based Fully Automated Program for Choroidal Structure Analysis Within the Region of Interest in Myopic Children

“…Compared to its precursor, U-Net, 56 MedT-Net had fewer errors when segmenting choroidal boundaries and provided a larger dice coefficient ( Table 1 ). Although previous models for choroidal segmentation have reported satisfactory results, 20 , 31 , 33 , 34 , 36 these were limited to choroidal segmentation alone, without consideration of the location and width of the examined choroidal area. Defining a ROI is essential for comparison across different studies and tracking longitudinal choroidal structure changes.…”