Search

Accurate segmentation of choroidal thickness (CT) and vasculature is important to better analyze and understand the choroid-related ocular diseases. In this paper, we proposed and implemented a novel and practical method based on the deep learning algorithms, residual U-Net, to segment and quantify the CT and vasculature automatically. With limited training data and validation data, the residual U-Net was capable of identifying the choroidal boundaries as precise as the manual segmentation compared with an experienced operator. Then, the trained deep learning algorithms was applied to 217 images and six choroidal relevant parameters were extracted, we found high intraclass correlation coefficients (ICC) of more than 0.964 between manual and automatic segmentation methods. The automatic method also achieved great reproducibility with ICC greater than 0.913, indicating good consistency of the automatic segmentation method. Our results suggested the deep learning algorithms can accurately and efficiently segment choroid boundaries, which will be helpful to quantify the CT and vasculature. [ABSTRACT FROM AUTHOR]

Introduction. Retinal structural and microvascular damages reflect damage to cerebral microvasculature and neurons. We aimed to investigate neovascular unit abnormalities among patients with large-artery atherosclerosis (LAA) or small-vessel occlusion (SAA) and control subjects. Methods. Twenty-eight LAA patients, forty-one SAA patients, and sixty-five age- and gender-matched controls were recruited. Based on optical coherence tomography angiography (OCTA), retinal capillary vessel density was assessed in the general and local sectors, and the thickness of individual retinal layer was extracted from retinal structural images. The differences between structural and microvascular were analyzed. Results. The superior peripapillary retinal nerve fiber layer (pRNFL) thickness was significantly different among the three groups, and the LAA group had the thinnest thickness. Compared to the control group, the deep retinal capillary vessel density in other two stroke subgroups were significantly reduced in all regions except in the inferior region (P < 0.05), and the fractal dimension in C2 and C4 regions of deep retina was significantly lower in the LAA group (P < 0.05). Discussion. Compared with superficial microvascular network, deep microvascular network is more sensitive to ischemic stroke. In addition, we have demonstrated quadrant-specific pRNFL abnormalities in LAA and SAA patients. Superior quadrant pRNFL thickness differences between stroke subgroups may suggest that changes in retinal nerve fiber layer are more sensitive to subtype identification than changes in retinal microvascular structure. All in all, the alteration in retinal structural and microvascular may further elucidate the role of the neovascular unit in ischemic stroke, suggesting that the combination of these two indicators could be used for subtype identification to guide prognosis and establish a risk prediction model. [ABSTRACT FROM AUTHOR]