4DCT XCAT phantom dataset for "Resolving Variable Respiratory Motion From Unsorted 4D Computed Tomography", MICCAI2024

The XCAT phantom dataset used in our paper "Resolving Variable Respiratory Motion From Unsorted 4D Computed Tomography", MICCAI 2024. For details on how to use the dataset, please visit https://github.com/Yuliang-Huang/4DCT-irregular-motion. If you find this dataset useful for your research, please cite our MICCAI paper.Note: We thank Prof Paul Segars from Duke University for the permission of releasing this single simulation example. If you want to use the XCAT software, please contact him.Introduction to the dataset:A digital phantom of thoracic region with motion was generated by the 4DXCAT software [1] and post-processed by the cid-X software [2]. The simulation was controlled by two respiration traces, i.e. the motion of the chest and the diaphragm, which were measured from 2D Cine MRI from a real patient. The diaphragm trace was deliberately delayed by 1 second to add hysteresis, i.e. the breathing would follow different paths during inhalation and exhalation. The chest trace could simulate the skin marker or belt signals normally used to sort 4DCT data in clinical practice and was stored in the rpm_signal.txt file.The phantom dataset consisted of images of thoracic regions at 182 timepoints in NIFTI format, each with size 355x280x115 voxels and resolution of 1x1x3 mm. Binary masks of the tumour were also provided for each timepoint. The dynamic volumes and tumor masks can be found in the compressed ground_truth.zip folder.The ref_empty_image.nii.gz file is an image with zero voxel values that define the reference image space. The ref_tumor_mask.nii.gz is the ground-truth tumor mask at the time average position.References:[1]Segars, W.P., Sturgeon, G., Mendonca, S., Grimes, J., Tsui, B.M.: 4d xcat phantom<br>for multimodality imaging research. Medical physics 37(9), 4902–4915 (2010)[2]Eiben, B., Bertholet, J., Menten, M.J., Nill, S., Oelfke, U., McClelland, J.R.: Con-<br>sistent and invertible deformation vector fields for a breathing anthropomorphic<br>phantom: a post-processing framework for the xcat phantom. Physics in Medicine<br>&amp; Biology 65(16), 165005 (2020)<br>

本研究发表于MICCAI 2024的论文《从未排序的四维计算机断层扫描中解析可变呼吸运动》所使用的XCAT体模数据集。关于该数据集的详细使用方法，请访问https://github.com/Yuliang-Huang/4DCT-irregular-motion。若该数据集对你的研究有所助益，请引用我们的MICCAI论文。
注意：我们感谢杜克大学Paul Segars教授授权发布此单个仿真示例。若您希望使用XCAT软件，请联系他。
数据集简介：
本数据集由4DXCAT软件[1]生成带呼吸运动的胸部数字体模，并通过cid-X软件[2]进行后处理。本次仿真采用两条呼吸运动轨迹控制，分别为胸部与膈肌的运动，两条轨迹均采集自真实患者的二维电影MRI（2D Cine MRI）测量结果。我们刻意将膈肌轨迹延迟1秒以引入呼吸迟滞效应，即呼吸在吸气与呼气阶段会遵循不同的运动路径。胸部轨迹可模拟临床中常用于对四维计算机断层扫描（4D Computed Tomography，以下简称4DCT）数据进行排序的皮肤标记物或呼吸感应带信号，该轨迹存储于rpm_signal.txt文件中。
本体模数据集包含182个时间点的胸部区域图像，格式为NIFTI，单张图像尺寸为355×280×115体素，空间分辨率为1×1×3 mm。每个时间点均附带对应的肿瘤二值掩码。动态影像体积与肿瘤掩码可在压缩包ground_truth.zip中获取。
ref_empty_image.nii.gz文件为体素值全零的图像，用于定义标准参考图像空间。ref_tumor_mask.nii.gz为对应时间平均位置的肿瘤真值掩码。
参考文献：
[1] Segars, W.P., Sturgeon, G., Mendonca, S., Grimes, J., Tsui, B.M.: 用于多模态成像研究的4D XCAT体模. 医学物理学, 37(9), 4902–4915 (2010)
[2] Eiben, B., Bertholet, J., Menten, M.J., Nill, S., Oelfke, U., McClelland, J.R.: 面向呼吸拟人化体模的一致可逆变形矢量场：XCAT体模的后处理框架. 医学与生物学物理学, 65(16), 165005 (2020)

搜集汇总

背景与挑战

背景概述

该数据集是为研究可变呼吸运动在未排序4D计算机断层扫描中的应用而设计的XCAT幻影数据集，包含182个时间点的胸部区域图像和肿瘤掩码，图像尺寸为355x280x115体素，分辨率为1x1x3毫米。数据集还提供了参考图像空间和肿瘤掩码的基准文件。