文献速递心脑血管急症重症综合管理与镇静镇痛

可解释机器学习模型预测危重症患者新发心房颤动:一项多中心研究

Interpretable machine learning model for new-onset atrial fibrillation prediction in critically ill patients: a multi-center study.
PubMed 文献 · Crit Care · 2024-10-29
一句话结论基于MIMIC-IV与MIMIC-III数据构建的23变量XGBoost模型可预测非心脏手术危重症患者新发心房颤动,外部验证AUC为0.769,风险概率>0.6定义为高风险。

本研究利用MIMIC-IV数据训练、MIMIC-III子集外部验证,构建危重症患者新发心房颤动(NOAF)的机器学习预测模型。在16,528例MIMIC-IV患者中,1520例(9.2%)入ICU后发生心房颤动;采用LASSO回归筛选特征,比较8种机器学习算法,最终23变量XGBoost模型表现最佳,验证集AUC为0.891(0.873-0.888),外部验证AUC为0.769(0.756-0.782)。关键预测因子包括年龄、机械通气、尿量、脓毒症、血尿素氮、经皮动脉血氧饱和度、连续性肾脏替代治疗和体重,风险概率>0.6定义为高风险,并已开发便于临床医生使用的友好界面。

要点

  • 在16,528例MIMIC-IV危重症患者中,1520例(9.2%)入ICU后发生新发心房颤动。
  • 23变量XGBoost模型验证集AUC为0.891(0.873-0.888),外部验证AUC为0.769(0.756-0.782)。
  • 关键预测因子包括年龄、机械通气、尿量、脓毒症、血尿素氮、经皮动脉血氧饱和度、连续性肾脏替代治疗和体重。
  • 风险概率>0.6定义为高风险,并已开发便于临床医生使用的友好界面。
  • SHAP方法用于可视化模型特征和个体病例预测,提高模型可解释性。

适用人群与场景

  • ICU收治的危重症患者,且未接受心脏手术;训练集来自MIMIC-IV,外部验证集来自MIMIC-III子集。

诊断标准与关键阈值

  • 风险概率大于0.6定义为高风险。

研究设计与关键结果

  • 研究类型为多中心研究,使用两个非重叠的MIMIC数据集,MIMIC-IV用于训练,MIMIC-III子集用于外部验证。
  • 在16,528例MIMIC-IV患者中,1520例(9.2%)入ICU后发生心房颤动。
  • 采用LASSO回归进行特征选择,使用8种机器学习算法构建预测模型。
  • 最终构建含23个变量的模型,XGBoost表现最佳,验证集AUC为0.891(0.873-0.888),外部验证AUC为0.769(0.756-0.782)。
  • 关键预测因子包括年龄、机械通气、尿量、脓毒症、血尿素氮、经皮动脉血氧饱和度、连续性肾脏替代治疗和体重。
  • 风险概率大于0.6定义为高风险。
  • 已开发便于临床医生使用的友好界面。
  • 使用SHAP方法可视化模型特征和个体病例预测。

证据类型、等级与局限

  • 证据类型:多中心机器学习模型开发与验证研究。
  • 局限:原文未报告局限性。

对床旁实践的改变

  • 作者认为该模型可作为临床可靠工具预测非心脏手术危重症患者新发心房颤动风险;SHAP提高模型可解释性,帮助临床医生更好理解新发心房颤动的原因,有助于提前预防并改善患者结局。

提炼说明

  • 本条依据论文摘要整理;摘要未涉及「处置流程与药物剂量」「红旗征与不能漏的情况」「指南或共识出处」,故未列出这些栏目;全文请点「阅读原文」。
英文原文(摘要)· 点击展开

BACKGROUND: New-onset atrial fibrillation (NOAF) is the most common arrhythmia in critically ill patients admitted to intensive care and is associated with poor prognosis and disease burden. Identifying high-risk individuals early is crucial. This study aims to create and validate a NOAF prediction model for critically ill patients using machine learning (ML).

METHODS: The data came from two non-overlapping datasets from the Medical Information Mart for Intensive Care (MIMIC), with MIMIC-IV used for training and subset of MIMIC-III used as external validation. LASSO regression was used for feature selection. Eight ML algorithms were employed to construct the prediction model. Model performance was evaluated based on identification, calibration, and clinical application. The SHapley Additive exPlanations (SHAP) method was used for visualizing model characteristics and individual case predictions.

RESULTS: Among 16,528 MIMIC-IV patients, 1520 (9.2%) developed AF post-ICU admission. A model with 23 variables was built, with XGBoost performing best, achieving an AUC of 0.891 (0.873-0.888) in validation and 0.769 (0.756-0.782) in external validation. Key predictors included age, mechanical ventilation, urine output, sepsis, blood urea nitrogen, percutaneous arterial oxygen saturation, continuous renal replacement therapy and weight. A risk probability greater than 0.6 was defined as high risk. A friendly user interface had been developed for clinician use.

CONCLUSION: We developed a ML model to predict the risk of NOAF in critically ill patients without cardiac surgery and validated its potential as a clinically reliable tool. SHAP improves the interpretability of the model, enables clinicians to better understand the causes of NOAF, helps clinicians to prevent it in advance and improves patient outcomes.

来源:Crit Care,查看原文。版权归原作者所有。

参考来源

  1. Guan C, Gong A, Zhao Y, et al. Interpretable machine learning model for new-onset atrial fibrillation prediction in critically ill patients: a multi-center study. Crit Care. 2024. PMID: 39473013. doi:10.1186/s13054-024-05138-0

供执业医师抢救与诊疗参考,不替代临床判断;剂量请结合体重、肝肾功能和本院药品规格复核,以指南原文为准。