Related Experiment Video
Updated: Sep 1, 2025
Predicting Amputation using Local Circulating Mononuclear Progenitor Cells in Angioplasty-treated Patients with Critical Limb Ischemia
Published on: September 22, 2020
Machine learning algorithms to predict major bleeding after isolated coronary artery bypass grafting
Yuchen Gao1, Xiaojie Liu2, Lijuan Wang1
1Department of Anesthesiology, Fuwai Hospital, State Key Laboratory of Cardiovascular Disease, National Center of Cardiovascular Diseases, Chinese Academy of Medical Sciences and Peking Union Medical College, Beijing, China.
Insights
Machine learning models accurately predict major bleeding after cardiac surgery, outperforming existing risk scores. These advanced methods can help identify patients at higher risk for bleeding complications.
Area of Science:
- Cardiovascular Surgery
- Medical Informatics
- Machine Learning in Healthcare
Background:
- Postoperative major bleeding is a significant complication following cardiac surgery, associated with adverse patient outcomes.
- Accurate prediction of major bleeding is crucial for risk stratification and management in cardiac surgery patients.
Purpose of the Study:
- To evaluate the performance of various machine learning (ML) methods in predicting major bleeding after coronary artery bypass graft (CABG) surgery.
- To compare the predictive accuracy of ML models against established risk scores like TRUST and WILL-BLEED.
Main Methods:
- A cohort of 1,045 patients undergoing isolated CABG surgery was analyzed, with data randomly split into training (70%) and testing (30%) sets.
- Major bleeding was defined using the universal definition of perioperative bleeding (UDPB) classes 3-4.
- Several ML algorithms, including conditional inference random forest (CIRF), stochastic gradient boosting (SGBT), and random forest, were developed and compared to a logistic regression model and existing risk scores using Area Under the Curve (AUC).
Main Results:
- The prevalence of postoperative major bleeding was 7.1%.
- The CIRF model achieved the highest AUC (0.831), followed by SGBT (0.820) and random forest (0.810), indicating superior predictive performance.
- All evaluated ML models demonstrated significantly higher AUCs compared to the TRUST (0.629) and WILL-BLEED (0.557) risk scores.
Conclusions:
- Machine learning methods show high efficacy in predicting major bleeding after cardiac surgery.
- Modern ML models offer improved performance over traditional scoring systems for identifying patients at risk of major bleeding.
- These ML tools have the potential to enhance the identification of high-risk subpopulations for targeted interventions.
Objectives:
Postoperative major bleeding is a common problem in patients undergoing cardiac surgery and is associated with poor outcomes. We evaluated the performance of machine learning (ML) methods to predict postoperative major bleeding.
Methods:
A total of 1,045 patients who underwent isolated coronary artery bypass graft surgery (CABG) were enrolled. Their datasets were assigned randomly to training (70%) or a testing set (30%). The primary outcome was major bleeding defined as the universal definition of perioperative bleeding (UDPB) classes 3-4. We constructed a reference logistic regression (LR) model using known predictors. We also developed several modern ML algorithms. In the test set, we compared the area under the receiver operating characteristic curves (AUCs) of these ML algorithms with the reference LR model results, and the TRUST and WILL-BLEED risk score. Calibration analysis was undertaken using the calibration belt method.
Results:
The prevalence of postoperative major bleeding was 7.1% (74/1,045). For major bleeds, the conditional inference random forest (CIRF) model showed the highest AUC [0.831 (0.732-0.930)], and the stochastic gradient boosting (SGBT) and random forest models demonstrated the next best results [0.820 (0.742-0.899) and 0.810 (0.719-0.902)]. The AUCs of all ML models were higher than [0.629 (0.517-0.641) and 0.557 (0.449-0.665)], as achieved by TRUST and WILL-BLEED, respectively.
Conclusion:
ML methods successfully predicted major bleeding after cardiac surgery, with greater performance compared with previous scoring models. Modern ML models may enhance the identification of high-risk major bleeding subpopulations.
