OCTA-500: A retinal dataset for optical coherence tomography angiography study.

Optical coherence tomography angiography (OCTA) is a novel imaging modality that has been widely utilized in ophthalmology and neuroscience studies to observe retinal vessels and microvascular systems. However, publicly available OCTA datasets remain scarce. In this paper, we introduce the largest and most comprehensive OCTA dataset dubbed OCTA-500, which contains OCTA imaging under two fields of view (FOVs) from 500 subjects. The dataset provides rich images and annotations including two modalities (OCT/OCTA volumes), six types of projections, four types of text labels (age/gender/eye/disease) and seven types of segmentation labels (large vessel/capillary/artery/vein/2D FAZ/3D FAZ/retinal layers). Then, we propose a multi-object segmentation task called CAVF, which integrates capillary segmentation, artery segmentation, vein segmentation, and FAZ segmentation under a unified framework. In addition, we optimize the 3D-to-2D image projection network (IPN) to IPN-V2 to serve as one of the segmentation baselines. Experimental results demonstrate that IPN-V2 achieves an about 10% mIoU improvement over IPN on CAVF task. Finally, we further study the impact of several dataset characteristics: the training set size, the model input (OCT/OCTA, 3D volume/2D projection), the baseline networks, and the diseases. The dataset and code are publicly available at: https://ieee-dataport.org/open-access/octa-500.

Corrections to "Image Projection Network: 3D to 2D Image Segmentation in OCTA Images".

In the above article [1], there is an error in (3). Instead of [Formula: see text] It should be [Formula: see text].

Choroidal neovascularization fusion visualization for spectral-domain optical coherence tomography and optical coherence tomography angiography images.

PURPOSE: The spectral-domain optical coherence tomography (SD-OCT) images and OCT angiography (OCTA) images can provide complementary information for choroidal neovascularization (CNV) visualization. We expected to simultaneously display multifaceted characteristics of CNV in a single projection image. METHODS: We proposed a fusion method for CNV visualization by combining structural and angiographic images, which mainly involves four steps: (a) Generate SD-OCT and OCTA projection images from original volumetric data with retinal layer restriction. (b) For SD-OCT projection images, enhance retinal vessels and CNV. (c) For OCTA images, detect CNV region based on multimodal data and display the neovascularization in false color. (d) A maximum fusion strategy was adopted to generate the fused images. RESULTS: Experimental results with 30 cases from 30 patients demonstrate that the fused images are more effective in displaying CNV than single-modality projection images. The average information entropy and the mean gradient in the CNV regions for SD-OCT projection images, OCTA images, and the fusion images are 4.66/0.21, 5.45/0.45, and 6.8/0.58, respectively. CONCLUSIONS: The proposed method is more effective for CNV visualization than the conventional single-modality image-based method. The proposed method can combine complementary information from multimodal images and provide a satisfying visual effect.

Case report: the first case of unilateral retinal pigment epithelium dysgenesis in China.

BACKGROUND: Unilateral retinal pigment epithelium dysgenesis (URPED) is a rare condition and is characterized by a unilateral and solitary lesion in the peripapillary region. The lesion presents with central atrophy, peripheral fibrosis, and hyperplastic changes in the retinal pigment epithelium (RPE). Herein, we report the first case of URPED in a Chinese individual using multimodal imaging techniques such as en-face optical coherence tomography (OCT) and optical coherence tomography angiography (OCTA). CASE PRESENTATION: A 10-year-old girl presented with 20/40 vision in her left eye. Presented as a solitary, unilateral, large and yellowish-white lesion, with fringe-like margins was observed in the posterior pole and lower middle periphery of the left eye continuous with the optic nerve, indicated URPED. Infrared fundal (IR) images showed that the fringe-like contour of the lesion was visible, with diffuse hyperreflective signals specifically in the fovea, while with peripheral dark spots, having a typical leopard-spot like appearance. Fundus autofluorescence (FAF) revealed a markedly scalloped lesion containing a hypoautofluorescence area mixed with an isoautofluorescence area. Spectral-domain optical coherence tomography (SD-OCT) revealed the outer segments of photoreceptors presented with an inhomogeneous signal in the fovea, with a weak local signal. The ellipsoid and interdigitation zones were thinner than normal, while the RPE/Bruch's complex was not flat, with locally visible protrusions. En-face OCT image at the level of the RPE zone showed a mottled hyperreflective signal with peripheral hyporeflective spots, fringe-like margin lesions. OCTA of the avascular area of the fovea in the superficial, deep, and outer retinal layers appeared to be oval in shape, the choroid capillary layer revealed an increase in the density of the choroidal vasculature in the fovea. CONCLUSIONS: This is the first report on URPED in China. Both en-face OCT and OCTA were essential in observing and studying the disease. Further investigation is required to better define the en-face OCT and OCTA features of URPED and clarify the disease characteristics and prognosis.

Image Projection Network: 3D to 2D Image Segmentation in OCTA Images.

We present an image projection network (IPN), which is a novel end-to-end architecture and can achieve 3D-to-2D image segmentation in optical coherence tomography angiography (OCTA) images. Our key insight is to build a projection learning module (PLM) which uses a unidirectional pooling layer to conduct effective features selection and dimension reduction concurrently. By combining multiple PLMs, the proposed network can input 3D OCTA data, and output 2D segmentation results such as retinal vessel segmentation. It provides a new idea for the quantification of retinal indicators: without retinal layer segmentation and without projection maps. We tested the performance of our network for two crucial retinal image segmentation issues: retinal vessel (RV) segmentation and foveal avascular zone (FAZ) segmentation. The experimental results on 316 OCTA volumes demonstrate that the IPN is an effective implementation of 3D-to-2D segmentation networks, and the uses of multi-modality information and volumetric information make IPN perform better than the baseline methods.