目的：应用多种机器学习算法构建耐碳青霉烯铜绿假单胞菌(CRPA)感染患者住院期间死亡风险的预测模型，验证其预测效能与可解释性，为临床识别高危患者提供科学依据。方法：回顾性收集2019年1月至2025年5月南京医科大学附属淮安第一医院收治的1 191例CRPA感染患者的临床资料。按7∶3比例随机分为训练集(833例)和验证集(358例)。针对训练集样本不平衡问题，采用合成少数类过采样技术(SMOTE)进行处理。应用最小绝对收缩和选择算子(LASSO)回归筛选预测特征，构建逻辑回归(LR)、支持向量机(SVM)、随机森林(RF)、极限梯度提升(XGBoost)等7种机器学习模型。通过受试者工作特征(ROC)曲线下面积(AUC)、校准曲线及决策曲线分析(DCA)评估模型性能，并采用沙普利加和解释(SHAP)法对最优模型进行可视化解释。结果：本组CRPA感染患者住院病死率为8.64%。LASSO回归筛选出6个关键预测因子：入住重症监护室(ICU)、气管插管、血管活性药物使用、腹腔内感染、凝血功能障碍及血清钙离子。在验证集中，SVM模型预测效能最佳，AUC为0.843(95%CI 0.757～0.928)，敏感度为0.704，特异度为0.861，优于其他6种模型。SHAP分析显示，血管活性药物使用及腹腔内感染对死亡风险预测的贡献度最大。结论：基于SVM算法构建的CRPA感染患者住院死亡风险预测模型具有良好的区分度、校准度及临床实用性，结合SHAP解释方法能有效辅助临床医生进行病情严重程度评估及早期干预。

[1] MENZIES N A, ALLWOOD B W, DEAN A S, et al.Global burden of disease due to rifampicin-resistant tuberculosis:a mathematical modeling analysis[J].Nat Commun, 2023, 14(1):6182.
[2] 卫菊, 朱秋丽, 孙喆, 等.碳青霉烯类耐药的铜绿假单胞菌感染对血液病患者病死率的影响[J].中华内科杂志, 2020, 59(5):353-359.
[3] HOWARD-ANDERSON J, BOWER C W, SMITH G, et al.Mortality in patients with carbapenem-resistant Pseudomonas aeruginosa with and without susceptibility to traditional antipseudomonal β-lactams[J].JAC Antimicrob Resist, 2021, 3(4):dlab187.
[4] SCHERTZ A R, LENOIR K M, BERTONI A G, et al.Sepsis prediction model for determining sepsis vs SIRS, qSOFA, and SOFA[J].JAMA Netw Open, 2023, 6(8):e2329729.
[5] QIU X, LEI Y P, ZHOU R X.SIRS, SOFA, qSOFA, and NEWS in the diagnosis of sepsis and prediction of adverse outcomes:a systematic review and meta-analysis[J].Expert Rev Anti Infect Ther, 2023, 21(8):891-900.
[6] NATHAN N, SCULIER J P, AMEYE L, et al.Sepsis and septic shock definitions in patients with cancer admitted in ICU[J].J Intensive Care Med, 2021, 36(3):255-261.
[7] KHAIRUDDIN M Z F, LU HUI P, HASIKIN K, et al.Occupational injury risk mitigation:machine learning approach and feature optimization for smart workplace surveillance[J].Int J Environ Res Public Health, 2022, 19(21):13962.
[8] FONSECA J, LIU X, OLIVEIRA H P, et al.Learning models for traumatic brain injury mortality prediction on pediatric electronic health records[J].Front Neurol, 2022, 13:859068.
[9] CHAO H Y, WU C C, SINGH A, et al.Using machine learning to develop and validate an in-hospital mortality prediction model for patients with suspected sepsis[J].Biomedicines, 2022, 10(4):802.
[10] LIU Y, ZHANG Y, ZHANG X, et al.Nomogram and machine learning models predict 1-year mortality risk in patients with sepsis-induced cardiorenal syndrome[J].Front Med, 2022, 9:792238.
[11] FAN Y, DONG J, WU Y, et al.Development of machine learning models for mortality risk prediction after cardiac surgery[J].Cardiovasc Diagn Ther, 2022, 12(1):12-23.
[12] WISE M G, KARLOWSKY J A, HACKEL M A, et al.In vitro activity of ceftazidime-avibactam against clinical isolates of Enterobacterales and Pseudomonas aeruginosa from sub-Saharan Africa:ATLAS Global Surveillance Program 2017-2021[J].J Glob Antimicrob Resist, 2023, 35:93-100.
[13] WANG J, YUN L, ZHAO H, et al.Combination therapy of polymyxin B and amikacin for community-acquired Pseudomonas aeruginosa pneumonia with MODS in a previously healthy patient:a case report[J].Infect Drug Resist, 2021, 14:2895-2900.
[14] YUE S, LI S, HUANG X, et al.Machine learning for the prediction of acute kidney injury in patients with sepsis[J].J Transl Med, 2022, 20(1):215.
[15] KENNEDY N, WIN T L, BANDYOPADHYAY A, et al.Insights from linking police domestic abuse data and health data in South Wales, UK:a linked routine data analysis using decision tree classification[J].Lancet Public Health, 2023, 8(8):e629-e638.
[16] LONG K, GUO D, DENG L, et al.Cross-combination analyses of random forest feature selection and decision tree model for predicting intraoperative hypothermia in total joint arthroplasty[J].J Arthroplasty, 2025, 40(1):61-69.e2.
[17] SU S, WANG J.Machine learning prediction of contents of oxygenated components in bio-oil using extreme gradient boosting method under different pyrolysis conditions[J].Bioresour Technol, 2023, 379:129040.
[18] LAI Z, CHEN X, ZHANG J, et al.Maximal margin support vector machine for feature representation and classification[J].IEEE Trans Cybern, 2023, 53(10):6700-6713.
[19] ZHANG Q, SHENG J, ZHANG Q, et al.Enhanced Harris Hawks optimization-based fuzzy k-nearest neighbor algorithm for diagnosis of Alzheimer's disease[J].Comput Biol Med, 2023, 165:107392.
[20] WU W, LI M, WU Y, et al.Cluster energy prediction based on multiple strategy fusion whale optimization algorithm and light gradient boosting machine[J].BMC Chem, 2024, 18(1):24.
[21] KHERABI Y, THY M, BOUZID D, et al.Machine learning to predict antimicrobial resistance:future applications in clinical practice?[J].Infect Dis Now, 2024, 54(3):104864.
[22] JIANG Y, WANG H W, TIAN F Y, et al.Predicting carbapenem-resistant Pseudomonas aeruginosa infection risk using XGBoost model and explainability[J].Sci Rep, 2025, 15(1):19737.
[23] SAINI S, PAHIL S, MOHINDRA R, et al.Diagnostic utility of sepsis screening tools, procalcitonin, and C-reactive protein in nosocomial fever of unknown origin[J].World J Crit Care Med, 2025, 14(3):106496.
[24] FLEISCHMANN-STRUZEK C, MELLHAMMAR L, ROSE N, et al.Incidence and mortality of hospital-and ICU-treated sepsis:results from an updated and expanded systematic review and meta-analysis[J].Intensive Care Med, 2020, 46(8):1552-1562.
[25] THEODOSIOU A A, READ R C.Artificial intelligence, machine learning and deep learning:potential resources for the infection clinician[J].J Infect, 2023, 87(4):287-294.
[26] PEIFFER-SMADJA N, RAWSON T M, AHMAD R, et al.Machine learning for clinical decision support in infectious diseases:a narrative review of current applications[J].Clin Microbiol Infect, 2020, 26(5):584-595.
[27] GUO Y, WANG F, MA S, et al.Relationship between atherogenic index of plasma and length of stay in critically ill patients with atherosclerotic cardiovascular disease:a retrospective cohort study and predictive modeling based on machine learning[J].Cardiovasc Diabetol, 2025, 24(1):95.
[28] XU C, ZHENG L, JIANG Y, et al.A prediction model for predicting the risk of acute respiratory distress syndrome in sepsis patients:a retrospective cohort study[J].BMC Pulm Med, 2023, 23(1):78.
[29] IBA T, LEVY J H, CONNORS J M, et al.The unique characteristics of COVID-19 coagulopathy[J].Crit Care, 2020, 24(1):360.
[30] FERNANDES C, PEREIRA L.Hypocalcemia in critical care settings, from its clinical relevance to its treatment:a narrative review[J].Anaesth Crit Care Pain Med, 2024, 43(6):101438.
[31] 叶宇翃, 李颖竹, 尹蓓.基于CiteSpace的人工智能在乳腺癌中应用现状与热点可视化分析[J].现代医学, 2025, 53(6):972-980.