A carregar...
Publicação
Predictive Factors of Major Adverse Cardiovascular Events, Myocardial Infraction and Stroke after Carotid Endarterectomy under logo-regional anesthesia: A Machine Learning Model
| Nome: | Descrição: | Tamanho: | Formato: | |
|---|---|---|---|---|
| 1.3 MB | Adobe PDF |
Autores
Resumo(s)
Carotid endarterectomy (CEA) is a surgical procedure aimed at reducing the risk of Stroke in patients with Carotid Artery disease. However, postoperative complications, including Major Adverse Cardiovascular Events (MACE), Myocardial Infarction (MI), and Stroke, remain concerns. This study focuses on developing a Machine Learning (ML) predictive model to identify pre- and post-surgery factors influencing the occurrence of these adverse events after CEA under loco-regional anesthesia. Data from patients who underwent CEA under locoregional anesthesia at the Centro Hospitalar de São João between January 2012 and January 2018 was used. Two predictive models were developed: one using only pre-surgery factors and another exclusively considering post-surgery factors. The classification models evaluated included K-Nearest Neighbors (kNN), Naive Bayes, Logistic Regression, Random Forest, Extreme Gradient Boosting (XGBoost), Decision Trees, Support Vector Machine (SVM), and Neural Networks and feature importance evaluation was performed using Shapley Additive Explanations (SHAP). For the preoperative model, Random Forest achieved the highest
predictive performance with an accuracy of 65%, followed by XGBoost with 60%. In the postsurgery model, XGBoost, Random Forest, and Logistic Regression demonstrated the best results, each reaching an accuracy of 62.5%. The strongest predictors factors identified were the number of days since surgery and the patient age. The results demonstrated that the more days that pass since surgery and the older the patient the higher is the risk of adverse events. In pre-surgery model body mass index and the presence of chronic kidney disease were found to be highly relevant, while in the post-surgery side, the occurrence of a stroke 30 days after the surgery emerged as a crucial factor.
Descrição
Dissertation presented as the partial requirement for obtaining a Master's degree in Information Management, specialization in Knowledge Management and Business Intelligence
Palavras-chave
Carotid Endarterectomy Carotid clamping Machine Learning Outcomes Stroke Myocardial Infarction Major Adverse Cardiovascular Events Risk factors Loco-regional anesthesia SDG 3 - Good health and well-being