| 537 | 0 | 179 |
| 下载次数 | 被引频次 | 阅读次数 |
目的 构建并评估多种机器学习模型用于预测个体罹患慢性阻塞性肺疾病(COPD)的风险，为早期筛查和干预提供数据支持。方法 选取823例研究对象，其中COPD高风险组142例，低风险组681例。收集人口统计学特征、吸烟史、症状(如咳嗽、气短)及慢性阻塞性肺疾病筛查问卷评分等数据。采用4种机器学习算法——逻辑回归、随机森林、支持向量机和XGBoost构建风险预测模型。采用5折交叉验证评估模型性能，评价指标包括准确率、精确率、召回率、F1分数、受试者工作特征曲线下面积(AUC-ROC)和平均精度(AP)。另，进行了特征重要性分析。结果 逻辑回归模型表现出最佳性能(AUC=0.982,AP=0.939)，随机森林模型次之(AUC=0.975,AP=0.890)。特征重要性分析显示，吸烟史、呼吸急促症状和体质量是关键预测因子。所有模型在识别低风险人群方面均表现出色，但在识别高风险人群的能力上存在差异。结论 机器学习模型能有效预测COPD的高风险人群。逻辑回归模型展现出最优的综合性能，能高效识别COPD高危人群，可作为有价值的临床辅助筛查工具。不同模型因其性能特点差异而适用于不同的临床筛查场景，为构建分层、智能化的COPD筛查路径提供了具体的决策依据。
Abstract：Objective To construct and evaluate various machine learning models for predicting the risk of chronic obstructive pulmonary disease(COPD) in individuals, thereby providing data support for early screening and intervention.Methods A total of 823 subjects were selected for this study, comprising 142 individuals in the high-risk group for COPD and 681 individuals in the low-risk group. Data collected included demographic characteristics, smoking history, symptoms(such as cough and shortness of breath), and scores from the Chronic obstructive pulmonary disease screening questionnaire. Four machine learning algorithms—Logistic Regression, Random Forest, Support Vector Machine, and XGBoost—were utilized to construct risk prediction models. The performance of these models was assessed using 5-fold cross-validation, with evaluation metrics including accuracy, precision, recall, F1-score, area under the receiver operating characteristic curve(AUC), and average precision(AP). Furthermore, a feature importance analysis was performed.Results The Logistic Regression model exhibited superior performance, achieving an AUC of 0. 982 and an AP of 0. 939. This was closely followed by the Random Forest model, which recorded an AUC of 0. 975 and an AP of 0. 890. Feature importance analysis revealed that smoking history, symptoms of shortness of breath, and body weight were significant predictors. All models demonstrated robust performance in identifying low-risk populations; however, variations were observed in their efficacy in identifying high-risk populations.Conclusion Machine learning models have proven effective in identifying individuals at high risk for COPD. Among these, the logistic regression model exhibits the best overall performance, efficiently identifying high-risk populations and serving as a valuable clinical auxiliary screening tool. Various models, each with distinct performance characteristics, are suited to different clinical screening scenarios, thereby offering targeted decision-making support for the establishment of a hierarchical and intelligent COPD screening pathway.
[1]Adeloye D, Song P, Zhu Y, et al. Global, regional, and national prevalence of, and risk factors for, chronic obstructive pulmonary disease(COPD)in 2019:a systematic review and modelling analysis[J]. Lancet Respir Med, 2022, 10(5):447-58. doi:10. 1016/S2213-2600(21)00511-7.
[2]Ardila D, Kiraly A P, Bharadwaj S, et al. End-to-end lung cancer screening with three-dimensional deep learning on low-dose chest computed tomography[J]. Nat Med, 2019, 25(6):954-61. doi:10. 1038/s41591-019-0447-x.
[3]王志鹏，张晓刚，张宏伟，等.机器学习在腰椎间盘突出症患者预后预测模型中应用价值的系统评价[J].中国组织工程研究，2026, 30(3):740-8. doi:10. 12307/2026. 875.[3]Wang Z P, Zhang X G, Zhang H W, et al. A systematic review of application value of machine learning to prognostic prediction models for patients with lumbar disc herniation[J]. Chin J Tissue Eng Res, 2026, 30(3):740-8. doi:10. 12307/2026. 875.
[4]张娜，杜静，郭子泉，等.基于CT影像组学的机器学习模型可预测肺部纯磨玻璃结节的浸润性[J].分子影像学杂志，2025, 48(5):614-9. doi:10. 12122/j. issn. 1674-4500. 2025. 05. 13.[4]Zhang N, Du J, Guo Z Q, et al. Machine learning models based on CT radiomics can effectively predict invasiveness of pulmonary pure ground-glass nodules[J]. J Mol Imag, 2025, 48(5):614-9. doi:10. 12122/j. issn. 1674-4500. 2025. 05. 13.
[5]Riley R D, Ensor J, Snell K I E, et al. Calculating the sample size required for developing a clinical prediction model[J]. BMJ,2020, 368:m441. doi:10. 1136/bmj. m441.
[6]Zhou Y M, Chen S Y, Tian J,et al. Development and validation of a chronic obstructive pulmonary disease screening questionnaire in China[J]. Int J Tuberc Lung Dis,2013,17(12):1645-51. doi:10. 5588/ijtld. 12. 0995.
[7]Kong J, Ha D, Lee J, et al. Network-based machine learning approach to predict immunotherapy response in cancer patients[J].Nat Commun,2022,13(1):3703. doi:10. 1038/s41467-022-31535-6.
[8]Ghalwash M A, Abdelrazek S M, Eladawi N H, et al. Enhancing credit card fraud detection using DBSCAN-augmented disjunctive voting ensemble[J]. Sci Rep, 2025, 15:39754. doi:10. 1038/s41598-025-22960-w.
[9]Mohammadi E, Rastegar M, Jamshidnezhad A, et al. Machine learning improves detection of alpha thalassemia carriers compared to clinical features[J]. Sci Rep, 2025, 15:36717. doi:10. 1038/s41598-025-20605-6.
[10]Im C, Kim W, Kim H. Explainable machine learning for heatrelated illness prediction:an XGBoost–SHAP approach using Korean meteorological data[J]. Bioengineering, 2025, 12(11):1276. doi:10. 3390/bioengineering12111276.
[11]Dey S, Eapen M S, Chia C, et al. Pathogenesis, clinical features of asthma COPD overlap, and therapeutic modalities[J]. Am J Physiol Lung Cell Mol Physiol, 2022, 322(1):L64-83. doi:10. 1152/ajplung. 00121. 2021.
[12]Ko E J, Bae S O, Kang D. A baseline study of interpretable machine learning using GC-MS breath VOCs for classifying asthma,bronchiectasis, and COPD[J]. Sci Rep, 2025, 15:44392. doi:10. 1038/s41598-025-28143-x.
[13]常敏丽，由淑萍，陈晓蝶，等.基于机器学习的肺结核肺炎患者判别分析研究[J].安徽医科大学学报，2025, 60(3):507-14. doi:10. 19405/j. cnki. issn1000-1492. 2025. 03. 017.[13]Chang M L, You S P, Chen X D, et al. Discriminant analysis of pulmonary tuberculosis patients and pneumonia patients based on machine learning[J]. Acta Univ Med Anhui, 2025, 60(3):507-14. doi:10. 19405/j. cnki. issn1000-1492. 2025. 03. 017.
[14]Glyde H M G, Morgan C, Wilkinson T M A, et al. Remote patient monitoring and machine learning in acute exacerbations of chronic obstructive pulmonary disease:dual systematic literature review and narrative synthesis[J]. J Med Internet Res, 2024,26:e52143. doi:10. 2196/52143.
[15]Sobhanan A, Shrinath V, Deshpande S, et al. Beyond smoking:exploring etiotypes of chronic obstructive pulmonary disease(COPD)using the global initiative for chronic obstructive lung disease(GOLD)2023 classification[J]. Cureus, 2025, 17(10):e94579. doi:10. 7759/cureus. 94579.
[16]Martinez F J, Han M K, Lopez C, et al. Discriminative accuracy of the CAPTURE tool for identifying chronic obstructive pulmonary disease in US primary care settings[J]. JAMA, 2023, 329(6):490-501. doi:10. 1001/jama. 2023. 0128.
基本信息:
DOI：10.19405/j.cnki.issn1000-1492.2026.07.014
中图分类号:R563.9
引用信息:
[1]陈琳,赵璐娜,周玥,等.基于机器学习和COPD-SQ问卷的COPD风险预测模型构建[J].安徽医科大学学报,2026,61(07):1261-1268.DOI:10.19405/j.cnki.issn1000-1492.2026.07.014.
基金信息:
国家自然科学基金地区基金项目(编号：82560019); 癌症、心脑血管、呼吸和代谢性疾病防治研究国家科技重大专项(编号：2023ZD0506100); 兵团指导科技计划项目(编号：2023ZD019)~~
2026-05-20
2026-05-20
2026-05-20