文献类型：期刊文献

英文题名：Development and external validation of a machine learning model for predicting the 28-day mortality risk in patients with sepsis complicated by acute respiratory failure in the ICU

作者：Xu, Yunpeng[1];Lei, Ting[1,2,3];Yang, Zi[4];Guo, Hong[5];Zhu, Lei[6];Wang, Jianlin[4,7];Liu, Hao[4];Liang, Tianhu[8];Lin, Qinglin[6];Yao, Guang[2];Yao, Zhiqiang[2];Liu, Jian[1,5]

第一作者：Xu, Yunpeng

通信作者：Liu, J[1];Yang, Z[2];Guo, H[3];Liu, J[3]

机构：[1]Lanzhou Univ, Clin Med Coll 1, Lanzhou, Gansu, Peoples R China;[2]Lanzhou Univ, Dept Gynecol, Hosp 1, Lanzhou, Gansu, Peoples R China;[3]First Hosp Lanzhou Univ, Branch Ctr Natl Clin Res Ctr Obstet & Gynecol Dis, Dept Gynecol, Lanzhou, Gansu, Peoples R China;[4]Lanzhou Univ, Informat Ctr, Hosp 1, Lanzhou, Gansu, Peoples R China;[5]Gansu Prov Cent Hosp, Gansu Prov Matern & Child Hlth Hosp, Dept Crit Care Med, Lanzhou, Gansu, Peoples R China;[6]Lanzhou Univ, Dept Crit Care Med, Hosp 1, Lanzhou, Gansu, Peoples R China;[7]Gansu Univ Chinese Med, Sch Med Informat Engn, Lanzhou, Gansu, Peoples R China;[8]Lanzhou Univ, Res Ctr Clin Med, First Hosp 1, Lanzhou, Gansu, Peoples R China

第一机构：Lanzhou Univ, Clin Med Coll 1, Lanzhou, Gansu, Peoples R China

通信机构：[1]corresponding author), Lanzhou Univ, Clin Med Coll 1, Lanzhou, Gansu, Peoples R China;[2]corresponding author), Lanzhou Univ, Informat Ctr, Hosp 1, Lanzhou, Gansu, Peoples R China;[3]corresponding author), Gansu Prov Cent Hosp, Gansu Prov Matern & Child Hlth Hosp, Dept Crit Care Med, Lanzhou, Gansu, Peoples R China.

年份：2026

卷号：6

期号：2

起止页码：175

外文期刊名：JOURNAL OF INTENSIVE MEDICINE

收录：Scopus(收录号：2-s2.0-105027232387);WOS:【ESCI(收录号:WOS:001747141600001)】；

基金： This work was supported by the National Natural Science Foundation of China (no. 82460379 , to J. Liu) ; the Gansu Provincial Science and Technology Program (Joint Scientific Research Fund Major Project , no. 24JRRA934 to J. Liu) ; the Gansu Provincial Science and Technology Program (Key R&D Project, no. 23YFGA0037 to J. Wang) ; and the Natural Science Foundation of Gansu Province (no. 22JR5RA1064, to Q. Lin) . The Gansu Provincial Science and Technology Program (no. 26JRRA340, to Z. Yang) ; and the Basic Research Project of the First Hospital of Lanzhou University (no. ldyyyn2025-10, to Z. Yang) .

语种：英文

外文关键词：Sepsis; Respiratory failure; Prediction model; Machine learning

摘要：Background: Sepsis complicated by acute respiratory failure (ARF) is a common and severe condition among patients admitted to the intensive care unit (ICU), associated with high mortality. Accurate prediction of shortterm outcomes is crucial for optimizing clinical decisions and treatment strategies. Therefore, we aimed to develop and validate an interpretable machine learning (ML) model to predict the 28-day mortality risk in ICU patients with sepsis complicated by ARF. Methods: This retrospective study included ICU patients with sepsis complicated by ARF from the Medical Information Mart for Intensive Care IV (MIMIC-IV, v3.1) database as the training cohort, and patients from the eICU-CRD (v2.0) database as the external validation cohort. Candidate predictors were initially identified based on clinical guidelines and expert consensus, and the Boruta algorithm was applied to determine the optimal feature set. Seven ML algorithms, namely random forest, XGBoost, logistic regression, AdaBoost, gradient boosting, CatBoost, and neural network, were compared. Model performance was evaluated using the area under the receiver operating characteristic curve (AUC), accuracy, sensitivity, specificity, positive predictive value, and negative predictive value. To enhance interpretability, feature importance was assessed using SHapley Additive exPlanations (SHAP) analysis. These results were integrated to construct a practical prognostic prediction platform. Results: The training cohort (2975 deaths) comprised 12,597 ICU patients with sepsis complicated by ARF from the MIMIC-IV (v3.1) database, while 890 patients from the eICU-CRD (v2.0) database (102 deaths) were included for external validation. Among the ML models, XGBoost achieved the best performance in the training cohort (AUC: 0.812; accuracy: 0.772; sensitivity: 0.605; specificity: 0.823). In the external validation cohort, XGBoost demonstrated good generalizability (AUC: 0.714). SHAP analysis identified PaO2, alanine aminotransferase, albumin, age, Acute Physiology Score (APS) III, lactate, urine output, and respiratory rate as the most influential predictors of 28-day mortality. Accordingly, a short-term mortality prediction platform was developed. Conclusions: We successfully developed an efficient, interpretable predictive model based on the XGBoost algorithm, accurately predicting 28-day mortality risk for ICU patients with sepsis complicated by ARF. Demonstrating stable performance and strong generalizability, this model holds promise as a clinical decision-support tool for the early identification of high-risk patients and optimization of personalized treatments.