Prediction model for vasospasm via snuffbox distal radial artery approach

Area of Science:

• Cardiovascular Interventions
• Medical Imaging & Diagnostics
• Machine Learning in Medicine

Background:

• Distal transradial access (dTRA) is increasingly used for interventional procedures.
• Vascular spasm is a potential complication of dTRA, impacting procedural success and patient outcomes.
• Predictive models are needed to identify patients at high risk for dTRA-related vascular spasm.

Purpose of the Study:

• To develop and validate a predictive model for vascular spasm associated with dTRA.
• To integrate demographic, ultrasonographic (anatomical and hemodynamic), and intraoperative variables into the predictive model.
• To utilize machine learning for enhanced predictive accuracy.

Main Methods:

• Retrospective analysis of 340 patients undergoing dTRA procedures.
• Random allocation into training (n=238) and validation (n=102) cohorts.
• Development of multivariate logistic regression, random forest, and support vector machine models.
• Performance evaluation using Area Under the Curve (AUC), calibration curves, and decision curve analysis.
• SHapley Additive exPlanations (SHAP) for model interpretability.

Main Results:

• Advanced age, male sex, and larger radial artery diameter were protective factors against vasospasm.
• Higher peak systolic velocity, increased puncture attempts, and larger sheath size were risk factors.
• The random forest model demonstrated superior performance with a validation AUC of 0.773.
• SHAP analysis highlighted puncture attempts and age as key predictive variables.

Conclusions:

• A multimodal random forest model effectively predicts dTRA-related vascular spasm.
• The model integrates demographic, ultrasonographic, and procedural data for risk stratification.
• This provides an objective tool for preoperative risk assessment and personalized surgical planning.