基于机器学习的慢性阻塞性肺疾病急性加重预测模型的研究

doi:10.3969/j.issn.1006-9771.2022.06.008

《中国康复理论与实践》 ›› 2022, Vol. 28 ›› Issue (6): 678-683.doi: 10.3969/j.issn.1006-9771.2022.06.008

基于机器学习的慢性阻塞性肺疾病急性加重预测模型的研究

张博超^1,²,杨朝³,郭立泉^1,²,陈静^1,²,熊大曦^1,²()

1.中国科学技术大学生物医学工程学院（苏州）生命科学与医学部,安徽合肥市 230026
2.中国科学院苏州生物医学工程技术研究所,江苏苏州市 215163
3.南京医科大学附属苏州科技城医院呼吸内科,江苏苏州市 215163

收稿日期:2022-03-14 修回日期:2022-04-15 出版日期:2022-06-25 发布日期:2022-07-05
通讯作者: 熊大曦 E-mail:xiongdx@sibet.ac.cn
作者简介:张博超(1997-),男,汉族,浙江宁波市人,硕士研究生,主要研究方向：慢性阻塞性肺疾病智能康复。|熊大曦(1970-),男,汉族,湖北武汉市人,博士,研究员,主要研究方向：应用光电子技术。
基金资助:
江苏省自然科学基金项目(BK20201183);苏州市临床重点病种诊疗技术专项(LCZX201931);江苏省"双创人才"项目(JSSCRC2021568)

Prediction model of acute exacerbation of chronic obstructive pulmonary disease based on machine learning

ZHANG Bochao^1,²,YANG Zhao³,GUO Liquan^1,²,CHEN Jing^1,²,XIONG Daxi^1,²()

1. School of Biomedical Engineering (Suzhou), Division of Life Sciences and Medicine, University of Science and Technology of China, Hefei, Anhui 230026, China
2. Suzhou Institute of Biomedical Engineering and Technology, Chinese Academy of Sciences, Suzhou, Jiangsu 215163, China
3. Respiratory Department, the Affiliated Suzhou Science and Technology Town Hospital of Nanjing Medical University, Suzhou, Jiangsu 215163, China

Received:2022-03-14 Revised:2022-04-15 Published:2022-06-25 Online:2022-07-05
Contact: XIONG Daxi E-mail:xiongdx@sibet.ac.cn
Supported by:
Natural Science Foundation of Jiangsu Province(BK20201183);Suzhou Municipal Special Project on Diagnosis and Treatment of Key Clinical Diseases(LCZX201931);"Double Creation Talents" in Jiangsu Province(JSSCRC2021568)

摘要/Abstract

摘要：

目的针对慢性阻塞性肺疾病急性加重期(AECOPD)患者肺功能检测存在误差大、准确性差的问题,开发AECOPD患者的肺功能预测模型,通过比较不同机器学习模型的预测性能,找到最优的模型。方法选取2018年1月至2020年2月南京医科大学附属苏州科技城医院不同患病程度的慢性阻塞性肺疾病(COPD)患者90例。利用6种机器学习算法(K-最近邻、逻辑回归、支持向量机、朴素贝叶斯、决策树和随机森林)建立预测分类模型,比较受试者工作特征曲线下面积(AUC-ROC)和准确性。采用10折交叉验证对数据集进行验证。结果基于随机森林的模型相较于其他模型预测性能最佳,准确率达到0.844,AUC-ROC为0.916。结论基于随机森林的预测模型能够辅助临床医生在难以给出确切诊断时提供决策支持。

关键词: 慢性阻塞性肺疾病, 急性加重期, 机器学习, 预测模型

Abstract:

Objective In view of the problems of large errors and poor accuracy in pulmonary function testing in patients with acute exacerbation of chronic obstructive pulmonary disease (AECOPD), a predictive classification model of pulmonary function in patients with AECOPD was proposed by comparing the prediction performance of different machine learning models to find the optimal model. Methods From January, 2018 to February, 2020, 90 patients with different degrees of COPD from the Affiliated Suzhou Science and Technology Town Hospital of Nanjing Medical University were collected. Six machine learning model algorithms (K-nearest neighbor, logistic regression, support vector machine, naive Bayes, decision tree and random forest) were used to establish AECOPD predictive classification models. Their area under the curve of receiver operating characteristic (AUC-ROC) and accuracy were compared. Ten-fold cross-validation method was used to validate the data set. Results The model based on random forest worked best in predicting and classifying AECOPD patients, with an accuracy rate of 0.844 and an AUC-ROC of 0.916. Conclusion Random forest-based predictive model is a powerful tool for identifying patients with AECOPD, providing decision support when it is difficult to give a definitive diagnosis.

Key words: chronic obstructive pulmonary disease, acute exacerbation period, machine learning, prediction model

中图分类号:

R562.2

张博超,杨朝,郭立泉,陈静,熊大曦. 基于机器学习的慢性阻塞性肺疾病急性加重预测模型的研究[J]. 《中国康复理论与实践》, 2022, 28(6): 678-683.

ZHANG Bochao,YANG Zhao,GUO Liquan,CHEN Jing,XIONG Daxi. Prediction model of acute exacerbation of chronic obstructive pulmonary disease based on machine learning[J]. 《Chinese Journal of Rehabilitation Theory and Practice》, 2022, 28(6): 678-683.

图/表 5

表1

图1

表2

图2

图3

参考文献 27

[1]	TAN K S, LIM R L, LIU J, et al. Respiratory viral infections in exacerbation of chronic airway inflammatory diseases: novel mechanisms and insights from the upper airway epithelium[J]. Front Cell Dev Biol, 2020, 8: 99. doi: 10.3389/fcell.2020.00099
[2]	WHO. Chronic respiratory diseases. [EB/OL]. (2022-05-19) [2022-05-31]. https://www.who.int/news-room/fact-sheets/detail/chronic-obstructive-pulmonary-disease-(copd).
[3]	2022 GOLD Reports. Global initiative for chronic obstructive lung disease[EB/OL]. (2021-11-22) [2022-04-01]. https://goldcopd.org/.
[4]	LI X, CAO X, GUO M, et al. Trends and risk factors of mortality and disability adjusted life years for chronic respiratory diseases from 1990 to 2017: systematic analysis for the Global Burden of Disease Study 2017[J]. BMJ, 2020, 368: m234.
[5]	CONNORS JR A F, DAWSON N V, THOMAS C, et al. Outcomes following acute exacerbation of severe chronic obstructive lung disease. The SUPPORT investigators (Study to Understand Prognoses and Preferences for Outcomes and Risks of Treatments)[J]. Am J Resp Crit Care, 1996, 154(4): 959-967. doi: 10.1164/ajrccm.154.4.8887592
[6]	BARNES P J. Cellular and molecular mechanisms of asthma and COPD[J]. Clin Sci (Lond), 2017, 131(13): 1541-1558. doi: 10.1042/CS20160487
[7]	ULMER W T. Lung function: clinical importance, problems, and new results[J]. J Physiol Pharmacol, 2003, 54: 11-13.
[8]	WU C T, LI G H, HUANG C T, et al. Acute exacerbation of a chronic obstructive pulmonary disease prediction system using wearable device data, machine learning, and deep learning: development and cohort study[J]. JMIR Mhealth Uhealth, 2021, 9(5): e22591. doi: 10.2196/22591
[9]	ZHOU M, CHEN C, PENG J, et al. Fast prediction of deterioration and death risk in patients with acute exacerbation of chronic obstructive pulmonary disease using vital signs and admission history: retrospective cohort study[J]. JMIR Med Inform, 2019, 7(4): e13085. doi: 10.2196/13085
[10]	SANCHEZ-MORILLO D, FERNANDEZ-GRANERO M A, JIMÉNEZ A L. Detecting COPD exacerbations early using daily telemonitoring of symptoms and k-means clustering: a pilot study[J]. Med Biol Eng Comput, 2015, 53(5): 441-451. doi: 10.1007/s11517-015-1252-4
[11]	PIKOULA M, QUINT J K, NISSEN F, et al. Identifying clinically important COPD sub-types using data-driven approaches in primary care population based electronic health records[J]. BMC Med Inform Decis Mak, 2019, 19(1): 86. doi: 10.1186/s12911-019-0805-0
[12]	WANG C, CHEN X, DU L, et al. Comparison of machine learning algorithms for the identification of acute exacerbations in chronic obstructive pulmonary disease[J]. Comput Meth Prog Bio, 2020, 188: 105267. doi: 10.1016/j.cmpb.2019.105267
[13]	CHEN J, YANG Z, YUAN Q, et al. Prediction models for pulmonary function during acute exacerbation of chronic obstructive pulmonary disease[J]. Physiol Meas, 2020, 41(12): 125010. doi: 10.1088/1361-6579/abc792
[14]	SEN I, SARACLAR M, KAHYA Y P. Differential diagnosis of asthma and COPD based on multivariate pulmonary sounds analysis[J]. IEEE Trans Biomed Eng, 2021, 68(5): 1601-1610. doi: 10.1109/TBME.2021.3049288
[15]	SHARMA H, KUMAR S. A survey on decision tree algorithms of classification in data mining[J]. Int J Sci Res, 2016, 5(4): 2094-2097.
[16]	FERNANDEZ-GRANERO M A, SANCHEZ-MORILLO D, LEON-JIMENEZ A. An artificial intelligence approach to early predict symptom-based exacerbations of COPD[J]. Biotechnol Biotec Eq, 2018, 32(3): 778-784. doi: 10.1080/13102818.2018.1437568
[17]	MOHKTAR M S, REDMOND S J, ANTONIADES N C. Predicting the risk of exacerbation in patients with chronic obstructive pulmonary disease using home telehealth measurement data[J]. Artif Intell Med, 2015, 63(1): 51-59. doi: 10.1016/j.artmed.2014.12.003
[18]	SHAH S A, VELARDO C, FARMER A, et al. Exacerbations in chronic obstructive pulmonary disease: identification and prediction using a digital health system[J]. J Med Internet Res, 2017, 19(3): e69. doi: 10.2196/jmir.7207
[19]	VERMA V K, LIN W Y. A machine learning-based predictive model for 30-day hospital readmission prediction for COPD patients[C]. IEEE International Conference on Systems, Man, and Cybernetics, 2020.
[20]	PAN W H, CHEN J Y, HAUNG S L, et al. Reference spiro metric values in healthy Chinese never smokers in two townships of Taiwan[J]. Chin J Physiol, 1997, 40(3): 165-174.
[21]	IP M S, WAI-SAN KO F, LAU A C, et al. Updated spirometric reference values for adult Chinese in Hong Kong and implications on clinical utilization[J]. Chest, 2006, 129(2): 384-392. doi: 10.1378/chest.129.2.384
[22]	DUONG M L, ISLAM S, RANGARAJAN S, et al. Global differences in lung function by region (PURE): an international, community-based prospective study[J]. Lancet Resp Med, 2013, 1(8): 599-609.
[23]	RAO W, WANG S, DULEBA M, et al. Regenerative metaplastic clones in COPD lung drive inflammation and fibrosis[J]. Cell, 2020, 181(4): 848-864. doi: 10.1016/j.cell.2020.03.047
[24]	JIA T G, ZHAO J Q, LIU J H, et al. Serum inflammatory factor and cytokines in AECOPD[J]. Asian Pac J Trop Med, 2014, 7(12): 1005-1008. doi: 10.1016/S1995-7645(14)60177-2
[25]	KARADENIZ G, POLAT G, SENOL G, et al. C-reactive protein measurements as a marker of the severity of chronic obstructive pulmonary disease exacerbations[J]. Inflammation, 2013, 36(4): 948-953. doi: 10.1007/s10753-013-9625-z
[26]	THOMSEN M, INGEBRIGTSEN T S, MAROTT J L, et al. Inflammatory biomarkers and exacerbations in chronic obstructive pulmonary disease[J]. JAMA, 2013, 309(22): 2353-2361. doi: 10.1001/jama.2013.5732
[27]	ZHANG S T, ZHANG X Q. Clinical significance and comparison of CRP, WBC and N% in hospitalized patients with acute exacerbations of chronic obstructive pulmonary disease[J]. Chin Prac Med, 2010, 5(20): 30-31.

分级	气流受限程度	肺功能特征
GOLD1	轻度	FEV₁/FVC< 0.7,FEV₁ ≥ 80%预计值
GOLD2	中度	FEV₁/FVC< 0.7,FEV₁ 50%~< 80%预计值
GOLD3	重度	FEV₁/FVC< 0.7,FEV₁ 30%~< 50%预计值
GOLD4	极重度	FEV₁/FVC< 0.7,FEV₁ < 30%预计值

特征	COPD GOLD分级				χ²/Z/F值	P值
特征	1级	2级	3级	4级	χ²/Z/F值	P值
性别(男/女)/n	8/2	31/3	34/4	8/0	2.064	0.559
年龄/岁	70.67±10.98	74.00±7.89	73.42±6.66	72.00±8.15	0.476	0.700
身高/cm	166.44±7.99	165.18±6.17	163.5±6.49	166.44±4.65	0.929	0.430
体质量/kg	70.94±5.44	64.51±10.38	60.02±10.79	63.44±7.65	8.503	0.037
BMI/kg·m^-2	25.64±1.57	23.67±3.83	22.41±3.62	22.94±2.90	1.473	0.228
CRP/mg·L^-1	4.50(7.00)	15.50(46.75)	17.50(65.25)	12.00(28.00)	5.352	0.148
WBC/10⁹·L^-1	7.51±3.64	7.80±2.75	8.74±4.06	10.13±3.53	1.634	0.187
NEU/%	64.17±13.91	72.56±8.26	80.63±8.74	79.61±6.78	17.602	< 0.001

基于机器学习的慢性阻塞性肺疾病急性加重预测模型的研究

Prediction model of acute exacerbation of chronic obstructive pulmonary disease based on machine learning

RichHTML

PDF (PC)

可视化

摘要/Abstract

引用本文

使用本文

图/表 5

参考文献 27

相关文章 15

Metrics

本文评价

推荐阅读 0

[1]	唐流泉, 彭兴云. 血流限制下运动对绝经后慢性阻塞性肺疾病患者脆性骨折部分危险因素的疗效[J]. 《中国康复理论与实践》, 2023, 29(7): 761-769.
[2]	彭娟,王洁萍,黄炜,樊必双,虞记华,曾今,黄丽衡,安丽娟,胥方元. 阈值负荷吸气肌训练对慢性阻塞性肺疾病患者呼吸功能、运动功能及生活质量影响的Meta分析[J]. 《中国康复理论与实践》, 2022, 28(9): 1022-1031.
[3]	朱旭,刘静,董泽萍,仇大伟. 基于表面肌电图手势动作意图识别的系统综述[J]. 《中国康复理论与实践》, 2022, 28(9): 1032-1038.
[4]	王博,袁永学,张庆苏. 非经口进食吞咽障碍脑卒中患者预后的相关因素及预测模型[J]. 《中国康复理论与实践》, 2022, 28(4): 453-460.
[5]	俞快,张利,叶祥明. 慢性意识障碍患者结局预测模型的系统综述[J]. 《中国康复理论与实践》, 2022, 28(2): 190-198.
[6]	张倩,卞敏洁,何琴,黄东锋. 血管性认知障碍早期预测机器学习模型的构建[J]. 《中国康复理论与实践》, 2021, 27(9): 1072-1077.
[7]	蔡倩,张溪,荆纯祥,蔡书宾,郭明凯,李际强. 弹性抗阻运动对慢性阻塞性肺疾病康复疗效的Meta分析[J]. 《中国康复理论与实践》, 2021, 27(8): 913-922.
[8]	曾斌,刘亚康,王龙平,张鸣生. 慢性阻塞性肺疾病患者通气有效性与运动后心脏功能的关系[J]. 《中国康复理论与实践》, 2021, 27(7): 812-818.
[9]	杨肇宇,李培君,李健,刘晓丹,吴卫兵. 运动对慢性阻塞性肺疾病系统性炎症和骨骼肌功能障碍干预效果的系统综述[J]. 《中国康复理论与实践》, 2021, 27(12): 1443-1449.
[10]	王洁萍,彭娟,樊必双,滕悦,胥方元. 水中肺康复对慢性阻塞性肺疾病稳定期患者肺功能的效果[J]. 《中国康复理论与实践》, 2021, 27(11): 1329-1333.
[11]	师晨曦, 王佳妮, 肖倩, 杨经玉, 贾燕瑞, 逯勇. 基于移动健康平台的自我管理干预在慢性阻塞性肺疾病患者肺康复中的应用[J]. 《中国康复理论与实践》, 2019, 25(6): 734-739.
[12]	陈雅琪, 蒋玉宇, 周春香, 许光清, 蒋为. 以同伴教育为核心的社区肺康复实施策略的质性研究[J]. 《中国康复理论与实践》, 2019, 25(4): 487-492.
[13]	蔡亚飞, 洪毅, 王方永. 脊髓损伤神经功能变化随访与预测模型[J]. 《中国康复理论与实践》, 2019, 25(10): 1120-1124.
[14]	崔尧, 丛芳, 李建军, 高峰, 杜良杰, 杨明亮. 成年脊髓损伤患者功能结局的预测模型研究[J]. 《中国康复理论与实践》, 2019, 25(10): 1125-1132.
[15]	王丹, 刘伟, 李仲铭, 钟莲梅. 慢性阻塞性肺疾病性周围神经病的神经电生理特点[J]. 《中国康复理论与实践》, 2018, 24(7): 846-849.