Machine Learning Models for Predicting In-Hospital Mortality in Acute Aortic Dissection Patients

Area of Science:

• Cardiology
• Artificial Intelligence
• Medical Informatics

Background:

• Acute aortic dissection is a life-threatening condition with high mortality.
• Existing predictive models for acute aortic dissection mortality are limited in accuracy and flexibility.
• There is a need for improved methods to identify patients at high risk of mortality.

Purpose of the Study:

• To develop and evaluate a machine learning model for predicting in-hospital mortality in acute aortic dissection patients.
• To identify key clinical variables associated with mortality risk in acute aortic dissection.

Main Methods:

• A cohort of 1,344 acute aortic dissection patients was analyzed.
• Clinical data including demographics, symptoms, lab results, and treatments were collected.
• Machine learning algorithms, including extreme gradient boost (XGBoost), were employed for prediction.
• Model performance was assessed using the area under the receiver operating characteristic curve (AUC).

Main Results:

• The extreme gradient boost model achieved the highest predictive performance with an AUC of 0.927.
• Key predictors of mortality identified by the model include treatment strategies, type of aortic dissection, and ischemia-modified albumin levels.
• SHapley Additive exPlanations confirmed that medical treatment, type A dissection, and elevated albumin levels increase mortality risk.

Conclusions:

• Machine learning, particularly XGBoost, offers a powerful tool for predicting in-hospital mortality in acute aortic dissection.
• Clinical variables such as treatment, dissection type, and specific biomarkers are crucial for risk stratification.
• These findings can aid in clinical decision-making and potentially improve patient outcomes.