|本期目录/Table of Contents|

[1]王慧泉,赵伟标,孟庆凯,等.基于无创生理参数的家用慢阻肺疾病分类方法*[J].生物医学工程研究,2023,01:23-29.
 WANG Huiquan,ZHAO Weibiao,MENG Qingkai,et al.Home-use chronic obstructive pulmonary disease classification method based on non-invasive physiological parameters[J].Journal of Biomedical Engineering Research,2023,01:23-29.
点击复制

基于无创生理参数的家用慢阻肺疾病分类方法*(PDF)

《生物医学工程研究》[ISSN:1006-6977/CN:61-1281/TN]

期数:
2023年01期
页码:
23-29
栏目:
出版日期:
2023-03-25

文章信息/Info

Title:
Home-use chronic obstructive pulmonary disease classification method based on non-invasive physiological parameters
文章编号:
1672-6278 (2023)01-0023-07
作者:
王慧泉1赵伟标1孟庆凯2马建新2童朝晖3曹志新3△
(1.天津工业大学 生命科学学院,天津 300387;2.北京怡和嘉业医疗科技股份有限公司,北京100041;3.首都医科大学北京朝阳医院 呼吸与危重症医学科,北京呼吸医学研究所,北京呼吸与危重症医学工程技术研究中心,北京 100020)
Author(s):
WANG Huiquan1 ZHAO Weibiao1 MENG Qingkai2 MA Jianxin2 TONG Zhaohui3 CAO Zhixin3
(1.School of Life Sciences, TianGong University, Tianjin 300387, China;2.BMC Medical Co., Ltd, Beijing 100041, China;3.Department of Respiratory and Critical Care Medicine,Beijing Chao-Yang Hospital, Capital Medical University, Beijing Institute of Respiratory Medicine, Beijing Engineering Research Center of Respiratory and Critical Care Medicine, Beijing 100020)
关键词:
慢阻肺特征提取机器学习呼吸疾病管理呼吸波SHAP分析
Keywords:
Chronic obstructive pulmonary diseaseFeature extractionMachine learningRespiratory disease managementRespiratory waveSHAP analysis
分类号:
R318;TP301
DOI:
10.19529/j.cnki.1672-6278.2023.01.04
文献标识码:
A
摘要:
为实现居家环境下慢性阻塞性肺疾病(chronic obstructive pulmonary disease, COPD)患者的病期监测,本研究设计了不同病期患者呼吸波形的采集实验,探究了基于呼吸波形提取的16个无创生理参数(non-invasive physiological parameters,NPP)及13个呼吸机统计参数(statistical parameters,STA)在COPD不同病期的分类效果,并基于支持向量机建立病期分类模型。结果表明,在NPP中,从压力波形中提取的峰值间隔序列一阶差分序列的标准差、均方根、方差和峰值间隔序列的变异系数对模型分类效果的贡献度最大。在三种输入不同特征参数的分类模型中,NPP、STA和NPP-STA的平均准确率分别为82%、67%和89%,AUC值分别为0.90、0.80和0.95。NPP-STA分类模型更适用于居家环境下COPD患者的病情分类,可用于日常疾病状况筛查,辅助医生对COPD患者进行诊断和管理。
Abstract:
In order to monitor the chronic obstructive pulmonary disease (COPD) patients in the home environment, we designed the respiratory waveform acquisition experiments of patients in different disease stages,explored the classification effects of 16 non-invasive physiological parameters (NPP) extracted from respiratory waveforms and 13 ventilator statistical parameters (STA) in different stages of COPD, and established a stage classification model based on support vector machine (SVM) . The results showed that in NPP, the standard deviation, root mean square, variance and variation coefficient of the first-order difference sequence of peak interval sequence extracted from the pressure waveform had the greatest contribution to the model classification effect. Among the three established classification models with different input feature parameters, the average accuracies of NPP, STA and NPP-STA classification model were 82%, 67% and 89% respectively, the AUC values were 0.90, 0.80 and 0.95 respectively. The NPP-STA classification model satisfies the disease classification of COPD patients in a home environment, and can be used for daily disease screening, assisting doctors in diagnosing and managing COPD patients.

参考文献/References

备注/Memo

备注/Memo:
(收稿日期:2022-07-27)国家重点研发计划项目(2019YFC0119400);北京市科学技术委员会临床诊断与治疗技术方案研究与应用项目(Z201100005520032);天津市研究生科研创新项目(2021YJSS082)。△通信作者Email:czx13911005116@163.com
更新日期/Last Update: 2023-04-28