|本期目录/Table of Contents|

[1]沈月秋,曹梦琳,许梅杰,等.基于可解释机器学习预测慢性阻塞性肺疾病患者急性加重风险研究[J].中华肺部疾病杂志,2025,(03):380-384.[doi:10.3877/cma.j.issn.1674-6902.2025.03.007]
　Shen Yueqiu,Cao Menglin,Xu Meijie,et al.Risk research of acute exacerbation in patients with chronic obstructive pulmonary disease based on interpretable machine learning model[J].,2025,(03):380-384.[doi:10.3877/cma.j.issn.1674-6902.2025.03.007]
点击复制

基于可解释机器学习预测慢性阻塞性肺疾病患者急性加重风险研究(PDF)

《中华肺部疾病杂志》[ISSN:1006-6977/CN:61-1281/TN]

卷:
期数:
2025年03期
页码:
380-384
栏目:
论著
出版日期:
2025-06-25

文章信息/Info

Title:
Risk research of acute exacerbation in patients with chronic obstructive pulmonary disease based on interpretable machine learning model
作者:
沈月秋; 曹梦琳; 许梅杰; 吴敏丹; 吴凯怡; 陆琳娟
215600 江苏,张家港市第一人民医院呼吸与危重症医学科
Author(s):
Shen Yueqiu;  Cao Menglin;  Xu Meijie;  Wu Mindan;  Wu Kaiyi;  Lu Linjuan.
Department of Respiratory and Critical Care Medicine, First People's Hospital, Zhangjiagang 215600, China
关键词:
急性加重期; 肺疾病; 慢性阻塞性;  机器学习;  风险分析
Keywords:
Acute exacerbation chronic obstructive pulmonary disease;  Machine learning;  Risk analysis
分类号:
R563
DOI:
10.3877/cma.j.issn.1674-6902.2025.03.007
摘要:
目的 分析基于可解释机器学习模型下慢性阻塞性肺疾病(chronic obstructive pulmonary, COPD)患者急性加重风险。方法 选取2021年1月至2024年6月我院收治的COPD稳定期患者80例。根据病情发生急性加重为观察组23例,未发生急性加重为对照组57例。使用LASSO回归分析筛选与急性加重相关的最佳特征,运用逻辑回归(logistic regression, LR)、K近邻分类(K-nearest neighbor classification, KNN)及极端梯度提升(extreme gradient boosting, xGBoost)4种机器学习算法开发预测评估; 采用可解释性(shapley additive explanation, SHAP)方法对最优预测进行分析。结果 单因素分析显示,观察组年龄(69.26±7.67)岁、体质量指数(22.16±2.15)kg/m2、COPD病程(8.66±2.87)年、吸烟史13例(56.52%)、FEV1/FVC(50.65±10.42)%、FEV1%pred(54.67±7.98)、降钙素原(0.78±0.21)ng/ml、纤维蛋白原(5.26±1.48)g/L、肝素结合蛋白(85.13±24.16)ng/ml、IL-6(34.17±4.73)ng/ml与对照组(64.98±8.03)岁,BMI(23.30±2.19)、(7.02±2.44)年、18例(31.58%)、(56.26±10.33)%、(59.37±8.66)、(0.64±0.19)ng/ml、(4.33±1.33)g/L、(70.26±24.67)ng/ml、(30.23±2.48)ng/ml比较差异具有统计学差异(P<0.05)。LASSO 回归结果显示,影响COPD患者急性加重的风险因素包括IL-6、年龄、肝素结合蛋白、体质量指数。基于4个特征构建ML中xGBoost预测AUC最大,为0.979,准确率为91.70%,精确率为98.00%。SHAP结果显示:贡献度前4的特征分别为IL-6、年龄、肝素结合蛋白、体质量指数; 随访 3个月期间,80例COPD患者中生存68例(85.00%),死亡12例(15.00%)。结论 COPD患者急性加重的机器学习,有助于临床识别COPD患者急性加重高风险患者。
Abstract:
Objective To explore the risk of acute exacerbation in patients with chronic obstructive pulmonary disease(COPD)based on interpretable machine learning model. Methods A total of 80 patients with COPD in stable phase admitted to our hospital from January 2021 to December 2024 were selected. Patients were divided into a observation group 23 cases and an control group 57 cases based on whether they experienced acute exacerbations. LASSO regression was used to screen for the best predictors of acute exacerbations. Four machine learning algorithms-logistic regression(logistic regression, LR), K-nearest neighbors classification(K-Nearest Neighbor Classification, KNN), and extreme gradient boosting(eXtreme Gradient Boosting, xGBoost)-were employed to develop predictive models and evaluate their performance. The SHAP method was used to analyze the interpretability of the optimal model and record the prognosis of patients. Results Univariate analysis showed that there were statistically significant differences between the observation group and the control group in age, body mass index, COPD course, smoking, FEV1/FVC, FEV1%pred, procalcitonin, fibrinogen, heparin binding protein, and IL-6(P<0.05). The regression results show that the risk factors affecting acute exacerbations in COPD patients include IL-6, age, heparin-binding protein, and body mass index. Among the four ML models constructed based on these four characteristics, Model AUC of xgboost model was the highest at 0.979, with an accuracy rate of 91.70% and a precision rate of 98.00%. The SHAP results indicate that the top four contributing features are IL-6, age, heparin-binding protein, and body mass index. During the follow-up period, 68(85.00%)of the 80 patients survived and 12 died(15.00%). Conclusion This study developed an efficient and interpretable machine learning model to predict acute exacerbation of COPD patients, which is helpful for clinical identification of high-risk patients with acute exacerbation of COPD.

参考文献/References:

1 任成山, 王关嵩, 钱桂生. 慢性阻塞性肺疾病的成因及其治疗的困惑与希望[J/CD]. 中华肺部疾病杂志(电子版), 2019, 12(2): 127-141.
2 Qian Y, Cai C, Sun M, et al. Analyses of factors associated with acute exacerbations of chronic obstructive pulmonary disease: a review[J]. Int J Chron Obstruct Pulmon Dis, 2023, 18: 2707-2723.
3 郭 静, 陶莲德, 闫祥云, 等. 机器学习在慢性阻塞性肺疾病中的研究进展[J]. 四川医学, 2024, 45(11): 1272-1276.
4 张博超, 杨 朝, 郭立泉, 等. 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.
5 Kor CT, Li YR, Lin PR, et al. Explainable machine learning model for predicting first-time acute exacerbation in patients with chronic obstructive pulmonary disease[J]. J Pers Med, 2022, 12(2): 228.
6 Chen Z, Wang J, Wang H, et al. Machine learning reveals sex differences in clinical features of acute exacerbation of chronic obstructive pulmonary disease: a multicenter cross-sectional study[J]. Front Med(Lausanne), 2023,10: 1105854.
7 赵 倩, 李荣梅, 刘 蕾, 等. 慢性阻塞性肺疾病急性加重患者非计划再入院风险预测模型的研究进展[J]. 沈阳医学院学报, 2023, 25(6): 629-633.
8 Czarnota P, Macleod JL, Gupta N, et al. Sex differences in chronic obstructive pulmonary disease: Implications for pathogenesis, diagnosis, and treatment[J]. Int J Mol Sci, 2025, 26(6): 2747.
9 Raghavan S, Hatipoglu U, Aboussouan LS. Goals of chronic obstructive pulmonary disease management: a focused review for clinicians[J]. Curr Opin Pulm Med, 2025, 31(2): 156-164.
10 Hao Y, Zhou Q, Sun Y, et al. Association of three single nucleotide polymorphisms in interleukin 6 gene with risk of chronic obstructive pulmonary disease[J]. Gene, 2022, 828: 146467.
11 Aslani MR, Amani M, Moghadas F, et al. Adipolin and IL-6 serum levels in chronic obstructive pulmonary disease[J]. Adv Respir Med, 2022, 90(5): 391-398.
12 张 华, 李玉梅, 张仕萍, 等. SAA,IL-6及sCD14对慢性阻塞性肺疾病合并肺部感染患者预后的预测价值[J]. 转化医学杂志, 2024, 13(9): 1359-1363.
13 Xiong XF, Zhu M, Wu H X, et al. Immunophenotype in acute exacerbation of chronic obstructive pulmonary disease: a cross-sectional study[J]. Respir Res, 2022, 23(1): 137.
14 Takizawa A, Shimada T, Chubachi S, et al. Exploring the pathophysiology of anemia in COPD: Insights from chest CT and longitudinal clinical data[J]. Respir Med, 2025, 240: 108046.
15 Adibi A, Sin DD, Safari A, et al. The acute COPD exacerbation prediction tool(ACCEPT): a modelling study[J]. Lancet Respir Med, 2020, 8(10): 1013-1021.
16 Wu YL, Yo CH, Hsu WT, et al. Accuracy of heparin-binding protein in diagnosing sepsis: a systematic review and Meta-analysis[J]. Crit Care Med, 2021, 49(1): e80-e90.
17 Liu Z, Li X, Chen M, et al. Heparin-binding protein and sepsis-induced coagulopathy: Modulation of coagulation and fibrinolysis via the TGF-beta signalling pathway[J]. Thromb Res, 2024, 244: 109176.
18 钱倩雯, 卢 芳. 血清肝素结合蛋白与肺功能指标对慢性阻塞性肺疾病患者病情严重程度及预后的评估价值[J/CD]. 现代医学与健康研究(电子版), 2022, 6(19): 46-50.
19 黄依璐, 王良友, 郏超男, 等. 65～74岁老年人体质量指数与慢性阻塞性肺疾病高危的关联性研究[J]. 现代实用医学, 2024, 36(8): 1034-1036, 封3.
20 李 论, 刘 毅, 李 佳, 等. 体质量指数和胸部CT特征在综合评估慢性阻塞性肺疾病中的价值[J]. 中国医学科学院学报, 2020, 42(1): 55-61.
21 Keogh E, Mark WE. Managing malnutrition in COPD: a review[J]. Respir Med, 2021, 176: 106248.
22 Zhang J, Moll M, Hobbs BD, et al. Genetically predicted body mass index and mortality in chronic obstructive pulmonary disease[J]. Am J Respir Crit Care Med, 2024, 210(7): 890-899.
23 Shin SH, Kwon SO, Kim V, et al. Association of body mass index and COPD exacerbation among patients with chronic bronchitis[J]. Respir Res, 2022, 23(1): 52.
24 Mjid M, Snene H, Hedhli A, et al. COPD patients'body composition and its impact on lung function[J]. Tunis Med, 2021, 99(2): 285-290.
25 Zhang X, Chen H, Gu K, et al. Association of body mass index with risk of chronic obstructive pulmonary disease: a systematic review and meta-analysis[J]. COPD, 2021, 18(1): 101-113.

备注/Memo

备注/Memo:
通信作者: 曹梦琳, Email: 838518984@qq.com
更新日期/Last Update: 2025-06-25