Machine Learning-Based Random Forest to Predict 3-Year Survival after Endovascular Aneurysm Repair

Area of Science:

• Cardiovascular Surgery
• Medical Informatics
• Machine Learning in Healthcare

Background:

• Endovascular aneurysm repair (EVAR) is a common treatment for abdominal aortic aneurysms (AAAs).
• Mid-term survival rates after EVAR require improvement and better prediction methods.
• Accurate prediction of patient survival is crucial for optimizing treatment strategies.

Purpose of the Study:

• To develop and validate a machine learning model for predicting 3-year survival post-EVAR.
• To identify key preoperative and perioperative factors influencing survival after EVAR.
• To enhance risk stratification for patients undergoing EVAR for AAAs.

Main Methods:

• A random forest machine learning model was developed.
• Data from 169 patients who underwent EVAR were analyzed, incorporating 23 variables.
• Model performance was assessed using 5-fold cross-validation, evaluating AUC, accuracy, sensitivity, specificity, and F1 score.

Main Results:

• The random forest model demonstrated strong predictive performance with an AUC of 0.91 and accuracy of 81.1%.
• Significant predictors of 3-year mortality included poor nutritional status, compromised immunity, chronic kidney disease (CKD), octogenarian status, chronic obstructive pulmonary disease (COPD), small aneurysm size, and statin use.
• Cross-validation confirmed consistent model performance across different data subsets.

Conclusions:

• A machine learning-based random forest model effectively predicts 3-year survival following EVAR.
• Identified risk factors such as nutritional status, immune function, CKD, advanced age, COPD, aneurysm size, and statin use are critical for predicting mortality.
• This model can aid clinicians in assessing patient risk and tailoring treatment for AAA.