Multilevel predictors categorization for post-CABG atrial fibrillation prediction
Preprint
Abstract
Background: Postoperative atrial fibrillation (PoAF) is known as common coronary artery bypass grafting (CABG) complication. Despite its association with increased risk of ischemic stroke, bleeding, acute renal failure and mortality there is still no ideal predictive tool with proper clinical interpretability. Methods: A retrospective single-center cohort study enrolled 1305 electronic medical records of patients with elective isolated CABG. PoAF was identified in 280 (21.5%) patients. Prognostic models with continuous variables were developed utilizing multivariate logistic regression (MLR), random forest and eXtreme gradient boosting methods. Predictors were dichotomized via grid search for optimal cut-off points, centroid calculation, and Shapley additive explanation (SHAP). For multilevel categorization, we proposed to use threshold values combination identified during dichotomization, as well as ranking cut-off thresholds by MLR weighting coefficients (multimetric categorization method). Results: Based on multistage selection, nine PoAF predictors were identified and validated. After categorization, prognostic models with continuous and multilevel categorical variables were developed. Multilevel categorical models advantage lies in their ability to explain PoAF prediction results and provide clinical interpretation, with comparable quality (AUC: 0.802 and 0.795). Conclusions: Multilevel predictor categorization shows promise for PoAF predictions explanation, with the developed models demonstrating high accuracy and transparency in their conclusions. Highlights: Validation of new 1st diagnosed atrial fibrillation predictors were performed in patients with coronary heart disease after coronary artery bypass grafting with subsequent development of predictive models utilizing machine learning methods. A new multilevel categorization method was tested, allowing to identify threshold values of predictors with the greatest predictive value, which were classified as risk factors for postoperative atrial fibrillation. The best quality metrics (AUC - 0.802) were demonstrated by a stochastic gradient boosting prognostic model based on predictors identified by the multilevel categorization method.Keywords
This publication has 21 references indexed in Scilit:
- Interpretable machine learning for in-hospital mortality risk prediction in patients with ST-elevation myocardial infarction after percutaneous coronary interventionsComputers in Biology and Medicine, 2024
- Machine Learning for Predicting Postoperative Atrial Fibrillation After Cardiac Surgery: A Scoping Review of Current LiteratureThe American Journal of Cardiology, 2023
- Machine Learning Models of Postoperative Atrial Fibrillation Prediction After Cardiac SurgeryJournal of Cardiothoracic and Vascular Anesthesia, 2022
- Machine Learning Outperforms Existing Clinical Scoring Tools in the Prediction of Postoperative Atrial Fibrillation During Intensive Care Unit Admission After Cardiac SurgeryHeart, Lung and Circulation, 2021
- A new approach for interpretability and reliability in clinical risk prediction: Acute coronary syndrome scenarioArtificial Intelligence in Medicine, 2021
- The persistent problem of new-onset postoperative atrial fibrillation: A single-institution experience over two decadesThe Journal of Thoracic and Cardiovascular Surgery, 2011
- Left Atrial Size and Postoperative Atrial FibrillationJACC, 2006
- A Multicenter Risk Index for Atrial Fibrillation After Cardiac SurgeryJAMA, 2004
- Atrial Fibrillation Following Coronary Artery Bypass Graft SurgeryJAMA, 1996
- Hazards of postoperative atrial arrhythmiasThe Annals of Thoracic Surgery, 1993