Prediction of Aneurysm Sac Shrinkage After Endovascular Aortic Repair Using Machine Learning-Based Decision Tree...

Area of Science:

• Vascular Surgery
• Medical Informatics
• Biomedical Engineering

Background:

• Abdominal aortic aneurysms (AAA) pose significant health risks.
• Endovascular aortic repair (EVAR) is a common treatment for AAA.
• Predicting aneurysm sac shrinkage post-EVAR is crucial for patient management.

Purpose of the Study:

• To develop a simple risk stratification model for predicting aneurysm sac shrinkage in patients undergoing EVAR for AAA.
• To utilize machine learning-based decision tree analysis for this prediction.

Main Methods:

• A cohort of 119 patients undergoing elective EVAR for AAA was analyzed.
• Predictors of aneurysm sac shrinkage were identified through univariable analysis (P < 0.05).
• These predictors were then incorporated into a decision tree analysis.

Main Results:

• Preoperative pulse wave velocity (PWV), operative type II endoleak, and current smoking were significant predictors.
• Decision tree analysis identified PWV as the most relevant predictor, followed by type II endoleak and smoking.
• The model generated 6 terminal nodes with shrinkage likelihoods ranging from 5.6% to 63.6%.

Conclusions:

• A decision tree model incorporating preoperative PWV, operative type II endoleak, and current smoking was established.
• This model effectively predicts aneurysm sac shrinkage in patients undergoing EVAR for AAA.
• The classification model can assist in identifying patients with high or low likelihoods of aneurysm sac shrinkage.