GBM-Reservoir: Brain Tumor (Glioblastoma Multiforme) MRI Dataset Collection with Ground Truth Segmentation Masks

We present a brain tumor database collection of 23,049 samples, where each sample comprises four different types of an MRI brain scan: FLAIR, T1, T1ce and T2. In addition, a segmentation mask (ground truth) for each sample is provided. These samples have been acquired via registration process of 438 samples available at the moment of registration from the original dataset provided by the BraTS 2022 Challenge. Registering each pair of existing brain scans results in two additional scans that retain a similar brain shape while featuring varying tumor locations. Consequently, by registering all possible pairs, a dataset originally consisting of n samples can be expanded to n² samples. The original dataset was collected from different institutions under standard clinical conditions, but with different equipment and imaging protocols. As a result, the image quality is heterogeneous, reflecting the diversity of clinical practices across institutions. This dataset can be utilized for various tasks, such as developing fully automated segmentation algorithms for new, unseen brain tumor cases, particularly through deep learning-based approaches, since ground truth is provided for each sample.本研究提出一个包含23049个样本的脑肿瘤数据库，每个样本包含四种不同模态的磁共振成像（Magnetic Resonance Imaging, MRI）脑部扫描序列：FLAIR、T1、T1ce及T2。此外，每个样本均附带分割掩码（标注真值，ground truth）。本数据集基于BraTS 2022挑战赛提供的原始数据集构建而成：该原始数据集在配准时刻包含438个样本，通过对其中的脑部扫描图像对进行配准操作，生成了当前的样本集。对任意一对现有脑部扫描图像进行配准，均可得到两幅额外的扫描图像：二者保留相近的脑部形态，但肿瘤位置存在差异。因此，若对所有可能的图像对进行配准，原本包含n个样本的数据集可扩展至n²个样本。原始数据集采集自多家医疗机构，均遵循标准临床诊疗流程，但使用了不同的成像设备与扫描协议。因此，数据集内图像质量存在异质性，体现了不同医疗机构间临床实践的差异。该数据集可应用于多种研究任务，例如针对全新的未见脑肿瘤病例开发全自动分割算法，尤其适用于基于深度学习的方法——因每个样本均附带标注真值。