RVO-Lesion: A Dual-Task OCT Dataset for Joint Segmentation and Detection of Macular Lesions in Retinal Vein Occlusion

Retinal vein occlusion (RVO) is one of the most common vision-threatening retinal diseases, primarily due to complications such as macular edema (ME). Optical coherence tomography (OCT), a non-invasive imaging modality, has become an essential tool for the diagnosis, treatment monitoring, and clinical evaluation of RVO, owing to its ability to clearly visualize fine retinal structures and macular fluid distribution. OCT facilitates precise segmentation, abnormality detection, and quantitative analysis, playing a critical role in clinical decision-making. However, the development of automated algorithms for RVO-ME analysis has been hindered by the lack of high-quality, manually segmented datasets (e.g., RVO-SD).To address this limitation, we have constructed a manually annotated RVO-ME segmentation dataset. The dataset comprises 3,012 OCT B-scan images collected from 130 patients, encompassing a total of 146 eyes.Each image is annotated with segmentation labels for five critical retinal structures—subretinal fluid (SRF), intraretinal fluid (IRF), the ellipsoid zone (EZ), the external limiting membrane (ELM) and highly reflective foci (HF) . This dataset serves as a valuable resource for evaluating the accuracy and robustness of various segmentation algorithms related to RVO, and significantly facilitates the development of artificial intelligence models for RVO-related disease analysis.

视网膜静脉阻塞（Retinal vein occlusion, RVO）是最常见的威胁视力的视网膜疾病之一，其临床危害主要源于黄斑水肿（macular edema, ME）等并发症。光学相干断层扫描（Optical coherence tomography, OCT）作为一种非侵入性成像技术，可清晰可视化视网膜细微结构与黄斑积液分布，因此已成为RVO诊断、治疗监测与临床评估的核心工具。OCT可支持精准分割、异常检测与定量分析，在临床决策制定中发挥关键作用。然而，由于缺乏高质量的手动分割数据集（如RVO-SD），针对RVO-ME分析的自动化算法开发一直受到掣肘。为解决这一局限，我们构建了一款经手动标注的RVO-ME分割数据集。该数据集包含从130名患者中采集的3012幅OCT B扫描图像，共计覆盖146只患眼。每幅图像均针对五种关键视网膜结构标注了分割标签：视网膜下液（subretinal fluid, SRF）、视网膜内液（intraretinal fluid, IRF）、椭圆体区（ellipsoid zone, EZ）、外界膜（external limiting membrane, ELM）以及高反射灶（highly reflective foci, HF）。本数据集可作为评估各类与RVO相关的分割算法准确性与鲁棒性的宝贵资源，同时将显著推动用于RVO相关疾病分析的人工智能模型开发。