Yang Yingjian, Li Wei, Kang Yan, Guo Yingwei, Yang Kai, Li Qiang, Liu Yang, Yang Chaoran, Chen Rongchang, Chen Huai, Li Xian, Cheng Lei
College of Medicine and Biological Information Engineering, Northeastern University, Shenyang 110169, China.
Medical Health and Intelligent Simulation Laboratory, Medical Device Innovation Center, Shenzhen Technology University, Shenzhen 518118, China.
Math Biosci Eng. 2022 Feb 17;19(4):4145-4165. doi: 10.3934/mbe.2022191.
The resting HR is an upward trend with the development of chronic obstructive pulmonary disease (COPD) severity. Chest computed tomography (CT) has been regarded as the most effective modality for characterizing and quantifying COPD. Therefore, CT images should provide more information to analyze the lung and heart relationship. The relationship between HR variability and PFT or/and COPD has been fully revealed, but the relationship between resting HR variability and COPD radiomics features remains unclear. 231 sets of chest high-resolution CT (HRCT) images from "COPD patients" (at risk of COPD and stage I to IV) are segmented by the trained lung region segmentation model (ResU-Net). Based on the chest HRCT images and lung segmentation images, 231 sets of the original lung parenchyma images are obtained. 1316 COPD radiomics features of each subject are calculated by the original lung parenchyma images and its derived lung parenchyma images. The 13 selected COPD radiomics features related to the resting HR are generated from the Lasso model. A COPD radiomics features combination strategy is proposed to satisfy the significant change of the lung radiomics feature among the different COPD stages. Results show no significance between COPD stage Ⅰ and COPD stage Ⅱ of the 13 selected COPD radiomics features, and the lung radiomics feature Y1-Y4 (P > 0.05). The lung radiomics feature F2 with the dominant selected COPD radiomics features based on the proposed COPD radiomics features combination significantly increases with the development of COPD stages (P < 0.05). It is concluded that the lung radiomics feature F2 with the dominant selected COPD radiomics features not only can characterize the resting HR but also can characterize the COPD stage evolution.
静息心率随慢性阻塞性肺疾病（COPD）严重程度的发展呈上升趋势。胸部计算机断层扫描（CT）被认为是表征和量化COPD最有效的方法。因此，CT图像应能提供更多信息来分析肺与心脏的关系。心率变异性与肺功能测试或/和COPD之间的关系已得到充分揭示，但静息心率变异性与COPD影像组学特征之间的关系仍不清楚。通过训练好的肺区域分割模型（ResU-Net）对来自“COPD患者”（有COPD风险及Ⅰ至Ⅳ期）的231组胸部高分辨率CT（HRCT）图像进行分割。基于胸部HRCT图像和肺分割图像，获得231组原始肺实质图像。通过原始肺实质图像及其衍生的肺实质图像计算每个受试者的1316个COPD影像组学特征。从套索模型中生成13个与静息心率相关的选定COPD影像组学特征。提出了一种COPD影像组学特征组合策略，以满足不同COPD阶段肺影像组学特征的显著变化。结果显示，在13个选定的COPD影像组学特征中，COPDⅠ期和COPDⅡ期之间无显著性差异，肺影像组学特征Y1-Y4（P>0.05）。基于所提出的COPD影像组学特征组合，以选定的COPD影像组学特征为主导的肺影像组学特征F2随着COPD阶段的发展显著增加（P<0.05）。结论是，以选定的COPD影像组学特征为主导的肺影像组学特征F2不仅可以表征静息心率，还可以表征COPD阶段的演变。
