基于机器学习算法构建心脏手术患者术后早期谵妄风险预测模型

左都坤; 吴卓熙; 龙宗泓; 李阳; 李加欣; 刘燕; 李洪

doi:10.16016/j.2097-0927.202301050

您当前的位置：

首页 >

文章列表页 >

基于机器学习算法构建心脏手术患者术后早期谵妄风险预测模型

专题报道：围术期大数据治理与AI算法研究 | 更新时间：2026-01-09

- 基于机器学习算法构建心脏手术患者术后早期谵妄风险预测模型
- Prediction model of early postoperative delirium risk based on machine learning algorithm in patients undergoing cardiac surgery
- 陆军军医大学学报 2023年45卷第8期页码：753-758
- 作者机构：
  
  400037 重庆, 陆军军医大学(第三军医大学)第二附属医院麻醉科
- 作者简介：
  
  [李洪，E-mail: lh78553@126.com LI Hong, E-mail: lh78553@126.com
- 基金信息：
  
  国家自然科学基金面上项目(82171265);国家重点研发计划(2018YFC0117200);重庆市技术创新与应用发展专项重点项目(CSTB2022TIAD-KPX0179)
- DOI：10.16016/j.2097-0927.202301050
  中图分类号： R319;R654.2;R749
- 收稿：2023-01-01，
  
  修回：2023-02-15，
  
  纸质出版：2023-04-30
- 稿件说明：
移动端阅览
左都坤, 吴卓熙, 龙宗泓, 等. 基于机器学习算法构建心脏手术患者术后早期谵妄风险预测模型[J]. 陆军军医大学学报, 2023,45(8):753-758.

Dukun ZUO, Zhuoxi WU, Zonghong LONG, et al. Prediction model of early postoperative delirium risk based on machine learning algorithm in patients undergoing cardiac surgery[J]. Journal of Army Medical University, 2023, 45(8): 753-758.
左都坤, 吴卓熙, 龙宗泓, 等. 基于机器学习算法构建心脏手术患者术后早期谵妄风险预测模型[J]. 陆军军医大学学报, 2023,45(8):753-758. DOI： 10.16016/j.2097-0927.202301050.

Dukun ZUO, Zhuoxi WU, Zonghong LONG, et al. Prediction model of early postoperative delirium risk based on machine learning algorithm in patients undergoing cardiac surgery[J]. Journal of Army Medical University, 2023, 45(8): 753-758. DOI： 10.16016/j.2097-0927.202301050.

摘要

目的

基于极限梯度提升树(extreme gradient boosting，XGBoost)机器学习算法，构建心脏手术术后谵妄(postoperative delirium

POD)风险的预测模型，评价其与传统Logistic回归(logistic regression

LR)模型的预测效果，为早期识别、早期干预治疗POD提供参考。

方法

采用病例-对照研究设计方案，筛选2022年3~7月陆军军医大学第二附属医院病历系统内因心脏病择期行全麻下心脏手术治疗的患者684例。根据术后3天的随访结果将其分为谵妄组(

=38)和非谵妄组(

=646)，将患者按7∶3比例随机分为训练集(479例)和测试集(205例)。使用LASSO回归分析筛选与POD相关的重要变量。LR和XGBoost用于构建预测模型。计算预测模型的受试者工作特征曲线下面积(under receiver operating characteristic curve，ROC-AUC)以及最佳阈值下的灵敏度和特异度，并比较不同模型的预测性能。

结果

心脏手术患者术后3天谵妄发生率为5.56%。与非谵妄组相比，谵妄组患者的年龄越大(

＜0.05)、糖尿病史比例越高(

＜0.05)，且术前收缩压、术后睡眠评分更低(

＜0.05)，其余资料差异均无统计学意义(

＞0.05)。最终纳入年龄、术前外周血氧饱和度、术前局部脑氧饱和度、术前收缩压、术后睡眠评分共5个变量进行建模。LR及XGBoost模型的AUC依次为0.732(95%

：0.430~1.000)、0.659(95%

：0.559~0.759)，其中LR的AUC更高，预测性能更好，但其灵敏度为50%，低于XGBoost(67%)，特异度为100%，高于XGBoost(98.5%)。

结论

基于XGBoost这种集成学习算法建立的心脏手术术后谵妄预测模型的效能并不优于传统的LR模型，LR模型可较好地预测心脏手术后谵妄的发生，为POD早期干预治疗提供参考。但XGBoost对术后谵妄的诊断更灵敏。

Abstract

Objective

To develop an early postoperative delirium (POD) risk prediction model in patients undergoing cardiac surgery based on Extreme Gradient Boosting (XGBoost) and compare its prediction performance with that of a traditional logistic regression (LR) model in order to provide reference for early identification and timely intervention of the condition.

Methods

A case-control trial was conducted on 684 patients who underwent elective cardiac surgery under general anesthesia due to heart disease in the Second Affiliated Hospital of Army Medical University from March to July 2022. According to the outcome of their 3-day follow-up after operation

the patients were divided into delirium group (

=38) and non-delirium group (

=646). The patients were randomly divided into a training set (479 patients) and a test set (205 patients) at a ratio of 7∶3. LASSO regression analysis was used to screen out important variables related to POD. LR and XGBoost were employed to construct the prediction models. The area under receiver operating characteristic curve (ROC-AUC) of the prediction m

odels and the sensitivity and specificity under the optimal threshold were calculated and the prediction performance of different models was compared.

Results

The 3-day postoperative delirium rate of the patients undergoing cardiac surgery was 5.56%. Compared with the non-delirium group

the patients in the delirium group were older (

0.05)

had a higher proportion of diabetes (

0.05)

and lower preoperative systolic blood pressure (

0.05) and postoperative sleep score (

0.05). But there were no statistical differences in other indicators (

0.05). Then finally

5 variables

including age

preoperative peripheral oxygen saturation

preoperative regional cerebral oxygen saturation

preoperative systolic blood pressure and postoperative sleep score

were included for modelling. The AUCs of LR and XGBoost models were 0.732 (95%

: 0.43~1.000) and 0.659 (95%

: 0.559~0.759)

respectively. LR model had a higher AUC value and better predictive performance

but

its sensitivity was 50%

lower than that of XGBoost model (67%)

and its specificity was 100%

higher than the other model (98.5%).

Conclusion

The predictive performance of the prediction model based on XGBoost

an integrated learning algorithm

is not superior to the traditional LR model for postoperative delirium after cardiac surgery. LR model can well predict the occurrence of delirium after cardiac surgery and provide reference for early intervention and treatment. But XGBoost is more sensitive to the diagnosis of postoperative delirium.

关键词

Keywords

references

ZHANG Y, WAN D H, CHEN M, et al . Automated machine learning-based model for the prediction of delirium in patients after surgery for degenerative spinal disease[J ] . CNS Neurosci Ther, 2023, 29(1): 282-295. DOI: 10.1111/cns.14002.

HU X Y, LIU H, ZHAO X, et al . Auto mated machine learning-based model predicts postoperative delirium using readily extractable perioperative collected electronic data[J ] . CNS Neurosci Ther, 2022, 28(4): 608-618. DOI: 10.1111/cns.13758.

BROWN C H. Delirium in the cardiac surgical ICU[J]. Curr Opin Anaesthesiol, 2014, 27(2): 117-122. DOI: 10.1097/ACO.0000000000000061.

JÄRVELÄ K, PORKKALA H, KARLSSON S, et al . Postoperative delirium in cardiac surgery patients[J ] . J Cardio-thorac Vasc Anesth, 2018, 32(4): 1597-1602. DOI: 10.1053/j.jvca.2017.12.030.

CHEN H Y, MO L, HU H J, et al . Risk factors of post-operative delirium after cardiac surgery: a meta-analysis[J ] . J Cardiothorac Surg, 2021, 16(1): 113. DOI: 10.1186/s13019-021-01496-w.

BULL M J, BOAZ L, JERMÉ M. Educating family caregivers for older adults about delirium: a systematic review[J]. Worldviews Evid Based Nurs, 2016, 13(3): 232-240. DOI: 10.1111/wvn.12154.

BANERJEE S, LIO P, JONES P B, et al . A class-contrastive human-interpretable machine learning approach to predict mortality in severe mental illness[J ] . NPJ Schizophr, 2021, 7(1): 60. DOI: 10.1038/s41 537-021-00191-y.

FORD E, SHEPPARD J, OLIVER S, et al . Automated detection of patients with dementia whose symptoms have been identified in primary care but have no formal diagnosis: a retrospective case-control study using electronic primary care records[J ] . BMJ Open, 2021, 11(1): e039248. DOI: 10.1136/bmjopen-2020-039248.

JAUK S, KRAMER D, GROßAUER B, et al . Risk prediction of delirium in hospitalized patients using machine learning: an implementation and prospective evaluation study[J ] . J Am Med Inform Assoc, 2020, 27(9): 1383-1392. DOI: 10.1093/jamia/ocaa113.

ELY E W, INOUYE S K, BERNARD G R, et al . Delirium in mechanically ventilated patients: validity and reliability of the confusion assessment method for the intensive care unit (CAM-ICU)[J ] . JAMA, 2001, 286(21): 2703-2710. DOI: 10.1001/jama.286.21.2703.

BROWN C H 4th, PROBERT J, HEALY R, et al . Cognitive decline after delirium in patients undergoing cardiac surgery[J ] . Anesthesiology, 2018, 129(3): 406-416. DOI: 10.1097/ALN.0000000000002253.

ERIKSON K, ALA-KOKKO T I, KOSKENKARI J, et al . Elevated serum S-100β in patients with septic shock is associated with delirium[J ] . Acta Anaesthesiol Scand, 2019, 63(1): 69-73. DOI: 10.1111/aas.13228.

GAILIUŠAS M, ANDREJAITIENE· J, ŠIRVINSKAS E, et al . Association between serum biomarkers and postoperative delirium after cardiac surgery[J ] . Acta Med Litu, 2019, 26(1): 8-10. DOI: 10.6001/actamedica.v26i1.3949.

BALLWEG T, WHITE M, PARKER M, et al . Association between plasma tau and postoperative delirium incidence and severity: a prospective observational study[J ] . Br J Anaesth, 2021, 126(2): 458-466. DOI: 10.1016/j.bja.2020.08.061.

KAZMIERSKI J, BANYS A, LATEK J, et al . Mild cognitive impairment with associated inflammatory and cortisol alterations as independent risk factor for postoperative delirium[J ] . Dement Geriatr Cogn Disord, 2014, 38(1/2): 65-78. DOI: 10.1159/000357454.

ZHAN S Z, HUANG B X, XUE F, et al . A functional outcome prediction model of acute traumatic spinal cord injury based on extreme gradient boost[J ] . J Orthop Surg Res, 2022, 17(1): 451. DOI: 10.1186/s13018-022-03343-7.

洪亮, 孙加奎, 沈骁, 等. 心脏手术后谵妄的危险因素分析及预测模型构建[J]. 临床麻醉学杂志, 2020, 36(12): 1195-1200. DOI: 10.12089/jca.2020.12.012.

HONG L, SUN J K, SHEN X, et al . Risk factors analysis and prediction model construction of postoperative delirium after cardiac sur-gery[J ] . J Clin Anesthesiol, 2020, 36(12): 1195-1200. DOI: 10.12089/jca.2020.12.012.

SHI Q K, MU X W, ZHANG C, et al . Risk factors for postoperative delirium in type A aortic dissection patients: a retrospective study[J ] . Med Sci Monit, 2019, 25: 3692-3699. DOI: 10.12659/MSM.913774.

DJAIANI G, SILVERTON N, FEDORKO L, et al . Dexmedetomidine versus propofol sedation reduces delirium after cardiac surgery: a randomized controlled trial[J ] . Anesthesiology, 2016, 124(2): 362-368. DOI: 10.1097/ALN.0000000000000951.

DUAN X, COBURN M, ROSSAINT R, et al . Efficacy of perioperative dexmedetomidine on postoperative delirium: systematic review and meta-analysis with trial sequential analysis of randomised controlled trials[J ] . Br J Anaesth, 2018, 121(2): 384-397. DOI: 10.1016/j.bja.2018.04.046.

GREAVES D, PSALTIS P J, DAVIS D H J, et al . Risk factors for delirium and cognitive decline following coronary artery bypass grafting surgery: a systematic review and meta-analysis[J ] . J Am Heart Assoc, 2020, 9(22): e017275. DOI: 10.1161/JAHA.120.017275.

王菲, 吕海涛, 唐新慧, 等. 术前血清尿酸水平与患者术后谵妄的关系[J]. 中华麻醉学杂志, 2022, 42(7): 807-812. DOI: 10.3760/cma.j.cn131073.20220314.00704.

WANG F, LYU H T, TANG X H, et al . Relationship between preoperative levels of serum uric acid and postoperative delirium[J ] . Chin J Anesthesiol, 2022, 42(7): 807-812. DOI: 10.3760/cma.j.cn131073.20220314.00704.

VAN DER PLOEG T, AUSTIN P C, STEYERBERG E W. Modern modelling techniques are data hungry: a simulation study for predicting dichotomous endpoints[J]. BMC Med Res Methodol, 2014, 14: 137. DOI: 10.1186/1471-2288-14-137.

浏览量

下载量

CSCD

文章被引用时，请邮件提醒。

提交

工具集

关联资源

心血管手术患者术后呼吸衰竭的危险因素及预测模型构建

基于增强静脉期CT放射组学的甲状腺乳头状癌中央区淋巴结转移预测模型构建与验证

基于5种机器学习算法构建的怀孕概率预测模型及其性能比较

高原脱适应症发生风险预测模型的构建与验证

基于机器学习算法构建组织学绒毛膜羊膜炎的预测模型