Original data of manuscript entitled "Spatial quantification of clinical biomarker pharmacokinetics through deep learning-based segmentation and signal-oriented analysis of MSOT data" by Hoffmann and Gerst et al.

This dataset contains all original data used in the manuscript entitled "Spatial quantification of clinical biomarker pharmacokinetics through deep learning-based segmentation and signal-oriented analysis of MSOT data" by Hoffmann and Gerst et al.

Contained data:

1. Anatomical_RoiSets.zip
This folder contains anatomical ROI sets of liver, spine and aorta regions used in the manuscript.
2. Annotation-comparison_ROIs.zipThis folder contains all ROI sets used for the comparison of manual and automated annotation.
3. Deep-learning_manual-annotation_data.zipThis folder contains the manual annotations used for the deep learning-based segmentation of animals.
4. Deep-learning_training.zipThis folder contains the the code and data used for training of the deep learning-based segmentation of MSOT images.
5. Signal-oriented_analysis_data.zip
This folder contains all data used for the signal-oriented analysis.
The data consists of two different treatments: Sham (control animals) and PCI (diseased animals). Each animal folder is identified by a unique name and contains an MSOT image in .tif file format and a region of interest (ROI) in .roi file format. The MSOT image is a time-resolved 4d image with 4 channels (water, indocyanine green, deoxygenated haemoglobin and oxygenated haemoglobin). The ROI derived by deep learning-based segmentation comprises the whole animal and is referred to as animal-ROI in the manuscript.

6. Tissue-oriented_analysis_data.zip
This folder contains all data used for the tissue-oriented analysis.
The data consists of two different treatments: Sham (control animals) and PCI (diseased animals). For each animal there exist two files prefixed with the animal identifier: a pre-processed MSOT image in .tif file format and a region of interest (ROI) in .roi file format. The MSOT image is a pre-processed, time-resolved 3d image of the indocyanine green channel (see manuscript for details on image pre-processing). The ROI comprises the liver tissue and is The third folder contains the MSOT image and the ROIs that were used to demonstrate the sensitivity of the tissue-oriented analysis towards slight changes in ROI positioning.

本数据集包含Hoffmann与Gerst等人发表的题为《基于深度学习分割与多光谱光声层析成像（MSOT）数据信号导向分析实现临床生物标志物药代动力学空间定量》的手稿中所用的全部原始数据。

包含的数据如下：
1. 解剖学感兴趣区域集.zip（Anatomical_RoiSets.zip）
该文件夹包含手稿中使用的肝脏、脊柱与主动脉区域的解剖学感兴趣区域（Region of Interest, ROI）集。
2. 标注对比感兴趣区域集.zip（Annotation-comparison_ROIs.zip）
该文件夹包含所有用于对比人工标注与自动标注的ROI集。
3. 深度学习人工标注数据.zip（Deep-learning_manual-annotation_data.zip）
该文件夹包含用于动物深度学习分割的人工标注数据。
4. 深度学习训练数据.zip（Deep-learning_training.zip）
该文件夹包含用于训练MSOT图像深度学习分割模型的代码与数据。
5. 信号导向分析数据.zip（Signal-oriented_analysis_data.zip）
该文件夹包含所有用于信号导向分析的数据。
本数据集包含两类不同处理方式的样本：假手术（Sham，对照组动物）与PCI（患病动物）。每个动物文件夹以唯一名称标识，内含格式为.tif的MSOT图像，以及格式为.roi的ROI文件。该MSOT图像为含4个通道的时间分辨四维图像，分别对应水、吲哚菁绿、脱氧血红蛋白与氧合血红蛋白。经深度学习分割得到的ROI覆盖整个动物个体，在手稿中被称为动物ROI（animal-ROI）。
6. 组织导向分析数据.zip（Tissue-oriented_analysis_data.zip）
该文件夹包含所有用于组织导向分析的数据。
本数据集包含两类不同处理方式的样本：假手术（Sham，对照组动物）与PCI（患病动物）。针对每只动物，存在两个以动物标识符为前缀的文件：格式为.tif的预处理MSOT图像，以及格式为.roi的ROI文件。该MSOT图像为针对吲哚菁绿通道的预处理时间分辨三维图像（图像预处理细节详见手稿）。该ROI覆盖肝脏组织。第三个文件夹包含用于展示组织导向分析对ROI定位微小变化敏感性的MSOT图像与ROI。