HCFMRP Sacroiliitis (v1)

Dataset composed of magnetic resonance imaging (MRI) of the sacroiliac joints of 48 patients from the Ribeirão Preto Medical School of University of São Paulo (Brazil). The selected patients are represented in 51 studies, 99 series and 594 images in the DICOM protocol, all anonymized. For each DICOM series, 6 consecutive images (slices) of the sacroiliac joints were obtained in the coronal plane, which were evaluated according to the Spondyloarthritis Research Consortium of Canada (SPARCC) MRI diagnostic criteria. Each DICOM image is accompanied by a PSD file containing the original image and a segmentation mask in PNG format. The masks indicate the location of the sacroiliac joints. The dataset contains, therefore, 1782 image files. The files are organized into two main directories, STIR and SPAIR, representing the two fat-saturated MRI techniques we adopted for this study. Each of them is subdivided into "Positive" and "Negative", indicating whether there are findings compatible with sacroiliitis. For more information, contact us: https://mainlab.fmrp.usp.br/

本数据集包含来自巴西圣保罗大学里贝朗普雷托医学院的48名患者的骶髂关节磁共振成像（Magnetic Resonance Imaging, MRI）数据。该队列共对应51项研究、99个影像序列以及594张遵循DICOM协议的医学影像，所有数据均已完成匿名化处理。针对每一个DICOM影像序列，本数据集采集了冠状位骶髂关节的6张连续断层层面，并依照加拿大脊柱关节炎研究联合会（Spondyloarthritis Research Consortium of Canada, SPARCC）的MRI诊断标准开展评估。每张DICOM影像均附带一个PSD文件，该文件包含原始影像与PNG格式的分割掩码，掩码用于标注骶髂关节的位置。本数据集总计包含1782个图像文件。所有文件分为两个主目录：STIR与SPAIR，分别对应本研究采用的两种脂肪饱和MRI成像技术。每个主目录下又细分为"Positive"与"Negative"两个子目录，用于区分是否存在符合骶髂关节炎表现的影像学征象。如需获取更多信息，请访问：https://mainlab.fmrp.usp.br/

搜集汇总

背景与挑战

背景概述

该数据集包含48名患者的骶髂关节MRI图像，采用STIR和SPAIR两种技术分类，并标注了阳性和阴性结果。所有图像均配有DICOM文件和分割掩模，总文件量1782个，用于骶髂关节炎研究。