Machine Learning Methods for Predicting Long-Term Mortality in Patients After Cardiac Surgery

Yue Yu; Chi Peng; Zhiyuan Zhang; Kejia Shen; Yufeng Zhang; Jian Xiao; Wang Xi; Pei Wang; Jin Rao; Zhichao Jin; Zhinong Wang

doi:10.3389/fcvm.2022.831390

Machine Learning Methods for Predicting Long-Term Mortality in Patients After Cardiac Surgery

Front Cardiovasc Med. 2022 May 3:9:831390. doi: 10.3389/fcvm.2022.831390. eCollection 2022.

Authors

Yue Yu¹, Chi Peng², Zhiyuan Zhang³, Kejia Shen⁴, Yufeng Zhang¹, Jian Xiao¹, Wang Xi¹, Pei Wang¹, Jin Rao¹, Zhichao Jin², Zhinong Wang¹

Affiliations

¹ Department of Cardiothoracic Surgery, Changzheng Hospital, Naval Medical University, Shanghai, China.
² Department of Health Statistics, Naval Medical University, Shanghai, China.
³ Department of Cardiothoracic Surgery, No. 988 Hospital of Joint Logistic Support Force of PLA, Zhengzhou, China.
⁴ Department of Personnel Administration, Second Affiliated Hospital of Naval Medical University, Shanghai, China.

Abstract

Objective: This study aims to construct and validate several machine learning (ML) algorithms to predict long-term mortality and identify risk factors in unselected patients post-cardiac surgery.

Methods: The Medical Information Mart for Intensive Care (MIMIC-III) database was used to perform a retrospective administrative database study. Candidate predictors consisted of the demographics, comorbidity, vital signs, laboratory test results, scoring systems, and treatment information on the first day of ICU admission. Four-year mortality was set as the study outcome. We used the ML methods of logistic regression (LR), artificial neural network (NNET), naïve bayes (NB), gradient boosting machine (GBM), adapting boosting (Ada), random forest (RF), bagged trees (BT), and eXtreme Gradient Boosting (XGB). The prognostic capacity and clinical utility of these ML models were compared using the area under the receiver operating characteristic curves (AUC), calibration curves, and decision curve analysis (DCA).

Results: Of 7,368 patients in MIMIC-III included in the final cohort, a total of 1,337 (18.15%) patients died during a 4-year follow-up. Among 65 variables extracted from the database, a total of 25 predictors were selected using recursive feature elimination and included in the subsequent analysis. The Ada model performed best among eight models in both discriminatory ability with the highest AUC of 0.801 and goodness of fit (visualized by calibration curve). Moreover, the DCA shows that the net benefit of the RF, Ada, and BT models surpassed that of other ML models for almost all threshold probability values. Additionally, through the Ada technique, we determined that red blood cell distribution width (RDW), blood urea nitrogen (BUN), SAPS II, anion gap (AG), age, urine output, chloride, creatinine, congestive heart failure, and SOFA were the Top 10 predictors in the feature importance rankings.

Conclusions: The Ada model performs best in predicting 4-year mortality after cardiac surgery among the eight ML models, which might have significant application in the development of early warning systems for patients following operations.

Keywords: MIMIC-III database; cardiac surgery; intensive care unit; long-term mortality; machine learning; prediction model.