XGBoost Machine Learning Algorithm for Prediction of Outcome in Aneurysmal Subarachnoid Hemorrhage

Area of Science:

• Neurosurgery
• Medical Informatics
• Machine Learning in Healthcare

Background:

• Aneurysmal subarachnoid hemorrhage (aSAH) is associated with poor survival and functional outcomes.
• Accurate risk stratification is crucial for tailoring therapeutic strategies in aSAH patients.
• Predictive modeling can aid clinicians in identifying high-risk individuals.

Purpose of the Study:

• To develop and validate a prognostic model for aSAH patients using the XGBoost algorithm.
• To compare the predictive performance of the XGBoost model against traditional logistic regression.

Main Methods:

• A cohort of 351 aSAH patients was analyzed.
• Patients were divided into training (70%) and testing (30%) sets.
• XGBoost and logistic regression models were constructed and evaluated using AUC, sensitivity, and specificity.

Main Results:

• The XGBoost model achieved an AUC of 0.950 for mortality and 0.958 for unfavorable functional outcome.
• These AUC values significantly outperformed logistic regression (0.767 for mortality, 0.829 for unfavorable outcome).
• Factors like age, GCS, WFNS score, mFisher score, IVH, and DCI were associated with non-survival.

Conclusions:

• The XGBoost-based prognostic model demonstrates superior precision in predicting outcomes for aSAH patients compared to logistic regression.
• This model can effectively assist clinicians in identifying high-risk aSAH patients, enabling enhanced medical care and potentially improving patient outcomes.