基于血清标志物机器学习模型对慢性阻塞性肺疾病急性加重期机械通气风险的预测分析

doi:10.3877/cma.j.issn.1674-6902.2025.04.021

目的

基于血清标志物机械学习模型预测慢性阻塞性肺疾病急性加重期(acute exacerbation of chronic obstructive pulmonary disease, AECOPD)患者机械通气风险的机器学习模型。

方法

选取2023年1月至12月我院治疗的62例AECOPD患者为对象，未机械通气33例为对照组，机械通气29例为观察组，采用XGBoost预测AECOPD患者发生机械通气风险，通过受试者工作特征曲线(receiver operating characteristic, ROC)判断预测，与单项血清标志物对比验证预测。

结果

观察组超敏C反应蛋白(high-sensitivity C-reactive protein, hs-CRP) 46.30(16.55，104.20)mg/L、白蛋白(albumin, ALB) 36.90(31.05，41.65)g/L、直接胆红素(direct bilirubin, DBIL)4.00(3.05，5.85) μmol/L、总胆固醇(total cholesterol, TC)3.83(3.12，4.30)mmol/L、载脂蛋白A1(apolipoprotein A1, APOA1)(1.17±0.32)g/L等较对照组hs-CRP7.30(1.05，18.25)mg/L、ALB 40.90(36.85，44.20)g/L、DBIL 2.90(2.45，3.85)μmol/L、TC 4.33(3.72，4.90)mmol/L、APOA1(1.46±0.29)g/L差异有统计学意义(P<0.05)。XGBoost预测AECOPD患者发生机械通气风险ROC曲线下面积(area under curve, AUC)为0.831，敏感度为71.29%，特异度为90.00%，准确度为75.81%，F1分数为0.933，预测价值较hs-CRP、ALB、DBIL、TC、HDL、APOA1提升。校准曲线显示，XGBoost模型拟合优度高于其他每项评分，校准曲线得分0.075。

结论

基于血清标志物构建的机器学习模型能准确预测AECOPD患者发生机械通气的风险，有助于帮助临床制定有效的治疗方案。

关键词: 急性加重期，肺疾病，慢性阻塞性, 机械通气, 血清标志物, 机器学习模型

Objective

To construct a machine learning model based on serum biomarkers for predicting the risk of mechanical ventilation in patients with acute exacerbation of chronic obstructive pulmonary disease (AECOPD) and analyze its predictive performance.

Methods

A total of 62 AECOPD patients treated in our hospital from January to December 2023 were enrolled. Among them, 33 patients without mechanical ventilation were assigned to the control group, and 29 patients requiring mechanical ventilation were assigned to the observation group. An XGBoost model was developed to predict the risk of mechanical ventilation in AECOPD patients based on serum biomarkers. The predictive performance was evaluated using the receiver operating characteristic (ROC) curve, and the model′s predictive value was validated by comparing it with individual serum biomarkers.

Results

High-sensitivity C-reactive protein(hs-CRP)46.30 (16.55, 104.20) mg/L, albumin(ALB)36.90 (31.05, 41.65) g/L, direct bilirubin(DBIL) 4.00 (3.05, 5.85) μmol/L, total cholesterol(TC) 3.83 (3.12, 4.30) mmol/L, andapolipoprotein A1(APOA1) (1.17±0.32) g/L in the observation group showed statistically significant differences compared to the control group in hs-CRP 7.30 (1.05, 18.25) mg/L), ALB 40.90 (36.85, 44.20) g/L, DBIL 2.90(2.45, 3.85) μmol/L), TC 4.33 (3.72, 4.90) mmol/L and APOA1 (1.46±0.29) g/L (P<0.05). The XGBoost model achieved an area under the ROC area under curve (AUC) of 0.831, sensitivity of 71.29%, specificity of 90.00%, accuracy of 75.81%, and an F1-score of 0.933, outperforming individual biomarkers such as hs-CRP, ALB, DBIL, TC, HDL, and APOA1. The calibration curve demonstrated good model fit with a calibration score of 0.075.

Conclusion

The machine learning model based on serum biomarkers provides a more accurate prediction of mechanical ventilation risk in AECOPD patients, aiding clinicians in developing more effective treatment strategies.

Key words: Acute exacerbation of chronic obstructive pulmonary disease, Mechanical ventilation, Serum markers, Machine learning model

表1 两组AECOPD患者临床特征比较[M(M₂₅，M₇₅)]

临床资料	观察组(n＝29)	对照组(n＝33)	χ²/t值	P值
hs-CRP(mg/L)	46.30(16.55，104.20)	7.30(1.05，18.25)	0.811	0.000
AST/ALT	1.36(1.07，1.92)	1.27(1.08，1.46)	0.581	0.157
TP(g/L)	65.30(61.25，67.60)	68.20(64.90，71.60)	0.342	0.052
ALB(g/L)	36.90(31.05，41.65)	40.90(36.85，44.20)	0.304	0.013
TBIL(μmol/L)	11.00(7.65，14.35)	10.60(6.90，13.40)	0.496	0.833
DBIL(μmol/L)	4.00(3.05，5.85)	2.90(2.45，3.85)	0.632	0.033
IBIL(μmol/L)	5.70(4.00，9.60)	6.30(4.50，10.65)	0.397	0.245
TC(mmol/L)	3.83(3.12，4.30)	4.33(3.72，4.90)	0.319	0.024
TG(mmol/L)	0.98(0.73，1.19)	0.98(0.72，1.25)	0.463	0.811
HDL(mmol/L)	1.11(0.91，1.30)	1.43(1.16，1.72)	0.265	0.002
LDL(mmol/L)	2.14(1.50，2.42)	2.40(1.90，2.69)	0.356	0.082
APOA1(g/L)	1.17±0.32	1.46±0.29	3.719	0.000
APOB(g/L)	0.71(0.56，0.79)	0.74(0.60，0.81)	0.430	0.486
LP(a)(mg/L)	233.00(100.50，305.00)	154.00(78.00，314.00)	0.531	0.477

表1 两组AECOPD患者临床特征比较[M(M₂₅，M₇₅)]

临床资料	观察组(n＝29)	对照组(n＝33)	χ²/t值	P值
hs-CRP(mg/L)	46.30(16.55，104.20)	7.30(1.05，18.25)	0.811	0.000
AST/ALT	1.36(1.07，1.92)	1.27(1.08，1.46)	0.581	0.157
TP(g/L)	65.30(61.25，67.60)	68.20(64.90，71.60)	0.342	0.052
ALB(g/L)	36.90(31.05，41.65)	40.90(36.85，44.20)	0.304	0.013
TBIL(μmol/L)	11.00(7.65，14.35)	10.60(6.90，13.40)	0.496	0.833
DBIL(μmol/L)	4.00(3.05，5.85)	2.90(2.45，3.85)	0.632	0.033
IBIL(μmol/L)	5.70(4.00，9.60)	6.30(4.50，10.65)	0.397	0.245
TC(mmol/L)	3.83(3.12，4.30)	4.33(3.72，4.90)	0.319	0.024
TG(mmol/L)	0.98(0.73，1.19)	0.98(0.72，1.25)	0.463	0.811
HDL(mmol/L)	1.11(0.91，1.30)	1.43(1.16，1.72)	0.265	0.002
LDL(mmol/L)	2.14(1.50，2.42)	2.40(1.90，2.69)	0.356	0.082
APOA1(g/L)	1.17±0.32	1.46±0.29	3.719	0.000
APOB(g/L)	0.71(0.56，0.79)	0.74(0.60，0.81)	0.430	0.486
LP(a)(mg/L)	233.00(100.50，305.00)	154.00(78.00，314.00)	0.531	0.477

图1 XGBoost预测效果评价。图A为XGBoost预测AECOPD机械通气风险的ROC曲线；图B为XGBoost预测AECOPD机械通气风险的校准曲线注：hs-CRP为超敏C反应蛋白；ALB为白蛋白；DBIL为直接胆红素；TC为总胆固醇；HDL为高密度脂蛋白；APOA1为载脂蛋白

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	Pépin JL, Degano B, Tamisier R, et al. Remote monitoring for prediction and management of acute exacerbations in chronic obstructive pulmonary disease (AECOPD)[J]. Life (Basel), 2022, 12(4): 499.
3	李晓东，李甜，邸兴伟，等. 以呼吸频率与膈肌增厚分数比值为导向的风险预测模型对AECOPD患者无创-有创机械通气治疗切换时机的预测价值[J]. 中华危重病急救医学，2024, 36(5): 503-507.
4	刘明宗，罗仁杰，陈雪梅. 超声心输出量监测在慢性阻塞性肺疾病急性加重期机械通气患者预后评估中的价值[J]. 第三军医大学学报，2022, 44(2): 182-189.
5	Huang Y, Wang J, Chen S, et al. Red cell distribution width related to the severity of the disease in patients with AECOPD[J]. Biomark Med, 2025, 19(5): 177-186.
6	Deniel G, Cour M, Argaud L, et al. Early antibiotic therapy is associated with a lower probability of successful liberation from mechanical ventilation in patients with severe acute exacerbation of chronic obstructive pulmonary disease[J]. Ann Intensive Care, 2022, 12(1): 86.
7	Alzaabi O, Guerot E, Planquette B, et al. Predicting outcomes in patients with exacerbation of COPD requiring mechanical ventilation[J]. Ann Intensive Care, 2024, 14(1): 159.
8	Liao QQ, Mo YJ, Zhu KW, et al. Platelet-to-lymphocyte ratio (PLR), neutrophil-to-lymphocyte ratio (NLR), monocyte-to-lymphocyte ratio (MLR), and eosinophil-to-lymphocyte ratio (ELR) as biomarkers in patients with acute exacerbation of chronic obstructive pulmonary disease (AECOPD)[J]. Int J Chron Obstruct Pulmon Dis, 2024, 19: 501-518.
9	龙嗣博，陈燕，张鑫桐，等. 新型冠状病毒感染者入院时血清降钙素原、白细胞介素-6和白细胞介素-8水平及其在患者预后中的意义[J]. 实用医学杂志，2024, 40(4): 471-475.
10	解添淞，翁微微，刘伟，等. 基于CT图像影像组学的机器学习模型预测胰腺癌免疫细胞浸润及预后的初步研究[J]. 中华放射学杂志，2022, 56(4): 425-430.
11	Fu Y, Liu Y, Zhong C, et al. An enhanced machine learning-based prognostic prediction model for patients with AECOPD on invasive mechanical ventilation[J]. iScience, 2024, 27(12): 111230.
12	Bai JS, Liu JX, Zhang Q, et al. A predictive model based on blood indicators for admission to the ICU with AECOPD[J]. Clin Lab, 2022, 68(10): 10.
13	Xiao C, Cheng S, Li R, et al. Isoforskolin alleviates AECOPD by improving pulmonary function and attenuating inflammation which involves downregulation of Th17/IL-17A and NF-κB/NLRP3[J]. Front Pharmacol, 2021, 12: 721273.
14	商焕霞，郭立娟，张罗元，等. AECOPD患者肺功能指标与血清白介素相关因子水平的关系[J/CD]. 中华肺部疾病杂志(电子版), 2024, 17(1): 123-126.
15	程炜炜，张青，张诚实，等. 全身免疫炎症指数与慢性阻塞性肺疾病急性加重期病情严重程度相关性分析[J/CD]. 中华肺部疾病杂志(电子版), 2024, 17(4): 580-584.
16	张志辉，刘学松，陈思蓓，等. 人巨细胞病毒再活化在非免疫抑制机械通气重症肺炎患者中的发生及预后分析[J]. 中华危重病急救医学，2021, 33(3): 286-292.
17	Zeng Z, Ke X, Gong S, et al. Blood urea nitrogen to serum albumin ratio: a good predictor of in-hospital and 90-day all-cause mortality in patients with acute exacerbations of chronic obstructive pulmonary disease[J]. BMC Pulm Med, 2022, 22(1): 476.
18	Wu Q, Cui N, Wang B. Clinical analysis of the correlation between nutritional risk screening and prognosis of patients with chronic obstructive pulmonary disease[J]. Pak J Med Sci, 2025, 41(2): 608-613.
19	Zhang Y, Zheng SP, Hou YF, et al. A predictive model for frequent exacerbator phenotype of acute exacerbations of chronic obstructive pulmonary disease[J]. J Thorac Dis, 2023, 15(12): 6502-6514.
20	涂诗琴，高凌云. Non-HDL-C水平和ApoB/ApoA1比值与初发心肌梗死的相关性研究[J]. 中国循证心血管医学杂志，2024, 16(4): 436-440.
21	Markelic I, Hlapcic I, Rogic D, et al. Lipid profile and atherogenic indices in patients with stable chronic obstructive pulmonary disease[J]. Nutr Metab Cardiovasc Dis, 2021, 31(1)：153-161.
22	Meng J, Yang W, Chen Z, et al. ApoA1, ApoB, ApoA1/B for Pathogenic prediction of chronic obstructive pulmonary disease complicated by acute lower respiratory tract infection: A cross-sectional study[J]. Int J Chron Obstruct Pulmon Dis, 2024, 19: 309-317.
23	胡晓侠，黄颖，伍梦佐，等. APOA1和Genisini评分与冠状动脉粥样硬化快速进展密切相关[J]. 临床心血管病杂志，2024, 40(4): 286-290.
24	Zhang X, Wang Y, Xu M, et al. The malnutrition in AECOPD and its association with unfavorable outcomes by comparing PNI, GNRI with the GLIM criteria: a retrospective cohort study[J]. Front Nutr, 2024, 11: 1365462.
25	Giampá S, Viana LG, Cardozo K, et al. Does obstructive sleep apnea treatment influence lipoprotein (a) concentrations? Data from the TREATOSA-MS clinical trial[J]. Ann Am Thorac Soc, 2023, 20(12): 1831-1833.
26	魏薇，吴月，程中乐，等. IL-6与经典炎症指标预测川崎病患儿丙种球蛋白敏感性及冠状动脉损害的对比研究[J]. 安徽医科大学学报，2021, 56(2): 306-310.
27	Saad E, Maamoun B, Nimer A. Increased red blood cell distribution predicts severity of chronic obstructive pulmonary disease exacerbation[J]. J Pers Med, 2023, 13(5): 843.
28	王迪，郑元，王曼曼，等. 基于XGBoost算法的肝移植术后患者服药依从性预测模型研究[J]. 中华实验外科杂志，2024, 41(10): 2219-2222.
29	Li J, Liu S, Hu Y, et al. Predicting mortality in intensive care unit patients with heart failure using an interpretable machine learning model: retrospective cohort study[J]. J Med Internet Res, 2022, 24(8): e38082.
30	张洋，赵化荣，刘攀，等. 多因素Logistic回归模型和XGBoost模型对舌癌患者放疗期间发生口腔感染的预测价值[J]. 中国现代医学杂志，2023, 33(19): 66-73.

[1]	沈月秋, 曹梦琳, 许梅杰, 吴敏丹, 吴凯怡, 陆琳娟. 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/OL]. 中华肺部疾病杂志(电子版), 2025, 18(03): 380-384.
[2]	宋春雪, 王亮亮, 袁玺, 徐庚辰, 刘柳, 马银娟, 丁琦, 魏霞. 中性粒细胞/淋巴细胞比值在不同血嗜酸性粒细胞水平慢性阻塞性肺疾病急性加重患者中的临床意义[J/OL]. 中华肺部疾病杂志(电子版), 2025, 18(03): 416-422.
[3]	许丽妹, 吴海燕, 林海丽, 吉顺妮, 陈春汝, 符娇娇, 陈忠仁, 吴小妹. 机械通气患者呼吸机相关性肺炎的风险分析[J/OL]. 中华肺部疾病杂志(电子版), 2025, 18(03): 447-451.
[4]	曾忠平, 张任玲, 刘静, 张天莎, 艾美梅, 张朋勃. 恶性肿瘤伴急性呼吸衰竭行有创机械通气危险因素分析[J/OL]. 中华肺部疾病杂志(电子版), 2024, 17(06): 991-994.
[5]	张璇, 高杨, 房雅君, 姚艳玲. 保护性机械通气在肺癌胸腔镜肺段切除术中的临床应用[J/OL]. 中华肺部疾病杂志(电子版), 2024, 17(04): 563-567.
[6]	程炜炜, 张青, 张诚实, 冯契靓, 陈荣荣, 赵云峰. 全身免疫炎症指数与慢性阻塞性肺疾病急性加重期病情严重程度相关性分析[J/OL]. 中华肺部疾病杂志(电子版), 2024, 17(04): 580-584.
[7]	杨万荣, 任治坤, 时新颍. 沙丁胺醇雾化吸入脾多肽治疗AECOPD的疗效分析[J/OL]. 中华肺部疾病杂志(电子版), 2024, 17(04): 609-612.
[8]	方晓玉, 王婷, 赵珊, 陈锋. HALP指数对AECOPD并发呼吸衰竭患者ICU结局的预测意义[J/OL]. 中华肺部疾病杂志(电子版), 2024, 17(04): 639-641.
[9]	倪韫晖, 袁雪燕, 黄丽丽, 杨毅, 邱海波. 呼吸驱动在保护性通气策略中的研究进展[J/OL]. 中华重症医学电子杂志, 2025, 11(01): 91-94.
[10]	王谷宜, 黎家琦, 钟燕军, 余波, 吴晨方, 董海云, 徐敏, 王花芹, 唐莉, 朱艳艳, 李金秀, 吕奔. 基于智慧ICU云平台的ARDS集束化管理对ARDS患者临床结局的影响[J/OL]. 中华重症医学电子杂志, 2025, 11(01): 72-77.
[11]	苗明月, 周建新. 肺保护性镇静：应重视呼吸驱动和吸气努力的床旁评估[J/OL]. 中华重症医学电子杂志, 2024, 10(04): 325-328.
[12]	刘春峰, 徐朝晖, 施红伟, 陈瑢, 马腾飞, 李鹏飞, 袁蓉, 陈建荣, 徐爱明. 机械通气患者肌肉减少症的诊断及其对预后的影响[J/OL]. 中华临床医师杂志(电子版), 2024, 18(09): 820-825.
[13]	王子阳, 王宏宾, 刘晓旌. 血清标志物对甲胎蛋白阴性肝细胞癌诊断的研究进展[J/OL]. 中华临床医师杂志(电子版), 2024, 18(07): 677-681.
[14]	刘晴雯, 韩勇, 陈丽丹, 邓哲. 早期机械通气对成人院内心脏骤停病死率的影响：一项回顾性队列研究[J/OL]. 中华卫生应急电子杂志, 2024, 10(04): 203-206.
[15]	刘晓鹏, 柳聪艳, 杨宁, 蔡琛, 李晓兵, 王红宇, 张思森. 三穴五针联合腹部提压法在机械通气患者肺康复中的疗效[J/OL]. 中华卫生应急电子杂志, 2024, 10(04): 193-198.

阅读次数

全文

摘要

选择文件类型/文献管理软件名称

选择包含的内容

Predictive analysis of mechanical ventilation risk in acute exacerbation of chronic obstructive pulmonary disease based on a machine learning model of serum markers

模态框（Modal）标题

选择文件类型/文献管理软件名称

选择包含的内容

Predictive analysis of mechanical ventilation risk in acute exacerbation of chronic obstructive pulmonary disease based on a machine learning model of serum markers