Machine Learning analysis of functional outcome and hemodynamic parameters during stroke thrombectomy under general...

Area of Science:

• Neurology
• Anesthesiology
• Data Science

Background:

• Intraoperative hemodynamics impact cerebral perfusion and neurological recovery in acute ischemic stroke (AIS) patients undergoing mechanical thrombectomy (MT).
• General anesthesia is frequently used for MT, necessitating careful hemodynamic management.

Purpose of the Study:

• To identify intraoperative blood pressure and heart rate patterns linked to 3-month functional outcomes in AIS patients treated with MT under general anesthesia.
• To leverage machine learning for analyzing complex hemodynamic relationships and interactions.

Main Methods:

• Post hoc analysis of a prospectively maintained cohort of AIS patients who underwent MT under general anesthesia.
• Utilized eXtreme Gradient Boosting (XGBoost) and SHapley Additive exPlanations (SHAP) to model and identify key hemodynamic predictors.
• Analyzed intraoperative hemodynamic variables during pre- and post-reperfusion phases.

Main Results:

• A total of 229 patients were analyzed, with 44.1% achieving functional independence (mRS 0-2) at 3 months.
• Increased time with mean arterial pressure <80 mmHg and systolic blood pressure <140 mmHg before reperfusion correlated with unfavorable outcomes.
• Elevated intraoperative heart rate prior to reperfusion was a significant marker of poor prognosis. The predictive model showed strong performance (AUC ROC = 0.85).

Conclusions:

• Distinct intraoperative hemodynamic patterns are associated with outcomes following MT under general anesthesia.
• Heart rate warrants further investigation as a critical physiological indicator during anesthetic management for stroke patients.
• These findings can inform optimized anesthetic strategies to improve functional recovery after MT.