基于不同机器学习方法构建老年脊柱手术患者术后谵妄的预测模型OA
Prediction model for postoperative delirium in elderly patients undergoing spinal surgery based on various machine learning methods
目的 基于机器学习方法构建老年脊柱手术患者术后谵妄(POD)的预测模型,并比较不同算法的预测效能.方法 本研究为回顾性研究,纳入2018年1月至2025年5月在医院接受脊柱手术的老年患者809例,并按8∶2随机分为训练集和验证集.通过最小绝对收缩与选择算子(LASSO)回归分析筛选老年脊柱手术患者POD的预测因素,采用合成少数类过采样技术(SMOTE)算法处理类别不平衡问题,并构建9种机器学习模型.绘制ROC曲线以评估不同机器学习模型在验证集中的预测效能.利用沙普利加解释法对预测效能最优的模型进行可解释分析,并对特征变量的重要性进行排序.结果 年龄、糖尿病、脑血管疾病、饮酒史、白蛋白、术中低血压、术中失血量、手术时间、麻醉时间和术中七氟醚用量在两组间差异有统计学意义(P<0.05).LASSO分析显示,年龄、糖尿病、白蛋白、术中低血压、术中七氟醚用量和手术时间是老年脊柱手术患者POD的预测因素.在9种机器学习模型中,轻量级梯度提升机模型效能最优,AUC为0.928.特征重要性排序结果显示,白蛋白对预测POD的贡献度最大.结论 基于轻量级梯度提升机建立的老年脊柱手术患者POD模型具有较好的预测效能,可为高危患者的早期识别和围手术期干预提供参考.
Objective This study aims to develop machine learning-based prediction models for postopera‑tive delirium(POD)in elderly patients undergoing spinal surgery,and to compare the predictive performance of different algorithms.Methods This retrospective study enrolled 809 elderly patients who underwent spinal surgery at our hospital between January 2018 and May 2025.All patients were randomly divided into a training set and a validation set at an 8:2 ratio.Least absolute shrinkage and selection operator(LASSO)regression analysis was applied to identify potential predictive factors for POD,the synthetic minority oversampling technique(SMOTE)algorithm was adopted to resolve the class imbalance issue,and nine machine learning models were subsequently constructed.Receiver operating characteristic(ROC)curves were generated to assess the predictive performance of the established models in the validation set.SHapley Additive exPlanations(SHAP)analysis was utilized to inter‑pret the best-performing model and rank the importance of each feature variable.Results Statistically significant differences(P<0.05)were observed between the study groups in age,prevalence of diabetes,prevalence of cere‑brovascular disease,drinking history,albumin level,intraoperative hypotension,intraoperative blood loss,dura‑tion of surgery,duration of anesthesia,and intraoperative sevoflurane dosage.LASSO regression analysis identified age,diabetes,albumin level,intraoperative hypotension,intraoperative sevoflurane dosage,and duration of sur‑gery as independent predictive factors for POD in elderly patients undergoing spinal surgery.Among the nine tested machine learning models,the Light Gradient Boosting Machine model achieved the optimal predictive perfor‑mance,with an area under the receiver operating characteristic curve(AUC)of 0.928.Feature importance ranking indicated that albumin made the greatest contribution to POD prediction.Conclusion The prediction model for postoperative delirium in elderly patients undergoing spinal surgery based on Light Gradient Boosting Machine dem‑onstrates good predictive performance and can provide a reference for the early identification of high-risk patients and the formulation of perioperative intervention strategies.
俞芸芸;郑芳;丁伟民;方佳伟;杨庆;王传光
温州医科大学附属第五医院,丽水市中心医院麻醉科(浙江 丽水 323000)温州医科大学附属第五医院,丽水市中心医院麻醉科(浙江 丽水 323000)温州医科大学附属第五医院,丽水市中心医院麻醉科(浙江 丽水 323000)温州医科大学附属第五医院,丽水市中心医院脊柱外科(浙江 丽水 323000)温州医科大学附属第五医院,丽水市中心医院放射科(浙江 丽水 323000)温州医科大学附属第五医院,丽水市中心医院麻醉科(浙江 丽水 323000)
医药卫生
脊柱手术谵妄机器学习围手术期管理
spinal surgerydeliriummachine learningperioperative management
《实用医学杂志》 2026 (17)
3119-3126,8
浙江省医药卫生科技计划项目(编号:2025KY1946)
评论