Morphological heart age from CTA: Predictions, Performance and Saliency Maps

This data distribution contains regression results for age prediction from computed tomography angiography images from the SCAPIS dataset, as well as proof-of-concept experiments concerning the prediction of known volumetric features estimated through segmentation of the images. Each sub-folder represents one experiment. For each experiment (sex-stratified), there is a csv file '0.csv'with the following structure Row 1: subject id1, subject id2, subject id3, ...Row 2: reference value1, reference value2, reference value3, ...Row 3: predicted value1, predicted value2, predicted value3, ... There is also a file 'results_summary.txt' which provides a text output of the quality measures corresponding to each experiment:Mean Absolute Error (MAE)R^2 (R2)Pearson correlation (r_p)Spearman correlation (r_s)Intraclass Correlation Coefficient (ICC) The following sub-folders/experiments contain saliency maps:- main (the main experiment with PCA from the whole heart)- main_linear (the main experiment without PCA from the whole heart)- poc_lvv (proof-of-concept: left ventricle volume)- poc_rvv (proof-of-concept: right ventricle volume)- poc_lav (proof-of-concept: left atrium volume)- poc_rav (proof-of-concept: right atrium volume)- poc_myov (proof-of-concept: myocardium volume)- poc_av (proof-of-concept: aorta volume) The feature subsets use the following encoding (for the feature subset experiments, the path contains the numbers representing the included features):1: Median Density2: Median Volume3: Stddev Density4: Stddev Volume Ethics:The subjects/images are all anonymized, with a chronological age rounded to whole months. Ethics approval was obtained from the Swedish Ethical Review Authority (Dnr 2022-07308-01) to conduct this research study related to human subjects, with associated sex and age information. All subjects provided informed written consent for their collected data to be used for research and for that research to be published. The study adheres to the Declaration of Helsinki. SCAPIS has been approved as a multicentre trial by the ethics committee at Umea University and adheres to the Declaration of Helsinki.本数据集包含来自SCAPIS数据集的计算机断层血管造影（computed tomography angiography）图像的年龄预测回归结果，以及针对通过图像分割得到的已知体积特征预测的概念验证实验。 每个子文件夹对应一项实验。针对每项按性别分层的实验，均存在一个名为0.csv的CSV文件，其结构如下： 第1行：受试者ID1、受试者ID2、受试者ID3…… 第2行：参考值1、参考值2、参考值3…… 第3行：预测值1、预测值2、预测值3…… 此外还包含一个results_summary.txt文件，其中记录了各实验对应的质量指标文本结果：平均绝对误差（Mean Absolute Error, MAE）、决定系数R²（R²）、皮尔逊相关系数（Pearson correlation, r_p）、斯皮尔曼相关系数（Spearman correlation, r_s）以及组内相关系数（Intraclass Correlation Coefficient, ICC）。 包含显著性映射图的子文件夹/实验如下： - main：基于全心主成分分析（PCA）的主实验 - main_linear：不基于全心主成分分析的主实验 - poc_lvv：概念验证实验：左心室容积 - poc_rvv：概念验证实验：右心室容积 - poc_lav：概念验证实验：左心房容积 - poc_rav：概念验证实验：右心房容积 - poc_myov：概念验证实验：心肌容积 - poc_av：概念验证实验：主动脉容积 特征子集采用以下编码规则（针对特征子集实验，其路径中包含代表所纳入特征的数字）： 1: 中位密度 2: 中位容积 3: 密度标准差 4: 容积标准差 伦理说明：所有受试者及其图像均已完成匿名化处理，其实际年龄已按整月取整。本项涉及人类受试者的研究已获得瑞典伦理审查局的伦理批准（编号Dnr 2022-07308-01），所有受试者均已签署书面知情同意书，同意将其采集的数据用于科研及科研成果发表。本研究符合《赫尔辛基宣言》。SCAPIS作为多中心临床试验，已通过于默奥大学伦理委员会审批，同样遵循《赫尔辛基宣言》。