Tailored risk assessment and forecasting in intermittent claudication.

Area of Science:

• Vascular Medicine
• Machine Learning in Healthcare
• Predictive Analytics

Background:

• Current intermittent claudication management relies on guidelines, but faces challenges in implementation and patient adherence.
• This variability in care leads to inconsistent patient outcomes.
• A precise risk stratification tool is needed for personalized treatment strategies.

Purpose of the Study:

• To develop and validate a machine-learning algorithm for precise risk stratification in intermittent claudication.
• To predict individual patient outcomes across different management strategies.
• To enhance personalized medicine approaches in vascular care.

Main Methods:

• Utilized the least absolute shrinkage and selection operator (LASSO) method for feature selection.
• Developed a machine-learning model on a bootstrapped sample of 255 intermittent claudication patients.
• Predicted outcomes including chronic limb-threatening ischemia, revascularization procedures, and major adverse cardiovascular/limb events.
• Evaluated performance using ROC curves, calibration curves, and decision curve analysis.

Main Results:

• The algorithm demonstrated excellent discrimination with high Area Under the ROC Curve (AUC) values (0.836–0.896) for various 2- and 5-year outcomes.
• Calibration curves showed good consistency between predicted and actual outcomes.
• Decision curve analysis confirmed the model's clinical utility, outperforming traditional logistic regression.

Conclusions:

• The developed machine-learning algorithm accurately predicts outcomes for patients with intermittent claudication.
• This tool has the potential to significantly improve risk stratification.
• Personalized predictions can lead to enhanced patient management and better overall outcomes.