Journal of cardiothoracic surgery · 2025
Machine learning in predicting preoperative intra-aortic balloon pump use in patients undergoing coronary artery bypass grafting.
СтатьяНазвание и аннотация: Перевод в очереди
Обработка статьи · этапы, ошибки и управление
Загрузка состояния…
Ошибки и ограничения
История · было → стало
Состояние обновляется каждые 5 секунд, пока блок открыт. Процент показывается только там, где известен объём работы. Пауза останавливает продолжение этапа; текущий запрос к источнику может завершиться.
Zhang Qian, Zheng Peng, Pan Yang, Li Luo, Yang Changqing, Wu Hengfang, Bian Zhiping, Zhao Sheng, Chen Xiangjian
Аннотация
BACKGROUND: Intra-aortic balloon pump (IABP) implantation in the perioperative period of cardiac surgery is an auxiliary treatment for cardiogenic shock. However, there is a lack of effective prediction models for preoperative IABP implantation.
OBJECTIVES: This study was designed to build machine learning algorithm-based models for early predicting risk factors of preoperative IABP implantation in patients who underwent coronary artery bypass grafting (CABG) surgery.
METHODS: Patients undergoing CABG were retrospectively enrolled from the hospital between January 2015 and March 2024 and divided into the preoperative and non-preoperative (including intraoperative and postoperative) IABP implantation groups. After feature selection by the cross-validation least absolute shrinkage and selection operator (LassoCV) algorithm, machine learning models were developed. The final model was considered according to its discrimination, including area under the receiver operating characteristic curve (AUC) and kolmogorov-smirnov (KS) plot.
RESULTS: The preoperative IABP group enrolled 95 (40.3%) patients. The Gaussian Naïve Bayes (GNB) model achieved the most excellent prediction ability based on its highest AUC of 0.76 (0.69-0.82) in the training set, 0.72 (0.49-0.94) in the validation set, and good KS plot and identified the top six features. The SHapley Additive exPlanations force analysis further illustrated visualized individualized prediction of preoperative IABP implantation.
CONCLUSION: Our study suggests that the GNB model achieved superior performance compared to others in predicting preoperative IABP implantation in patients undergoing CABG surgery. This may contribute to risk-prediction and decision-making in clinical practice.
OBJECTIVES: This study was designed to build machine learning algorithm-based models for early predicting risk factors of preoperative IABP implantation in patients who underwent coronary artery bypass grafting (CABG) surgery.
METHODS: Patients undergoing CABG were retrospectively enrolled from the hospital between January 2015 and March 2024 and divided into the preoperative and non-preoperative (including intraoperative and postoperative) IABP implantation groups. After feature selection by the cross-validation least absolute shrinkage and selection operator (LassoCV) algorithm, machine learning models were developed. The final model was considered according to its discrimination, including area under the receiver operating characteristic curve (AUC) and kolmogorov-smirnov (KS) plot.
RESULTS: The preoperative IABP group enrolled 95 (40.3%) patients. The Gaussian Naïve Bayes (GNB) model achieved the most excellent prediction ability based on its highest AUC of 0.76 (0.69-0.82) in the training set, 0.72 (0.49-0.94) in the validation set, and good KS plot and identified the top six features. The SHapley Additive exPlanations force analysis further illustrated visualized individualized prediction of preoperative IABP implantation.
CONCLUSION: Our study suggests that the GNB model achieved superior performance compared to others in predicting preoperative IABP implantation in patients undergoing CABG surgery. This may contribute to risk-prediction and decision-making in clinical practice.