Artificial Intelligence-Based ABI Dynamic Fluctuation Patterns Predict Adverse Vascular Events in PAD: A Multicenter...

Area of Science:

• Vascular Medicine
• Artificial Intelligence in Healthcare
• Predictive Analytics

Background:

• Peripheral arterial disease (PAD) poses significant risks for major adverse limb events (MALE).
• Current risk stratification methods for PAD patients often lack precision.
• Dynamic fluctuations in ankle-brachial index (ABI) may offer novel predictive insights.

Purpose of the Study:

• To develop and validate an AI-based predictive model for MALE in PAD patients.
• To utilize ankle-brachial index (ABI) dynamic fluctuation patterns for risk prediction.
• To establish a novel risk stratification tool for precision medicine in PAD.

Main Methods:

• A multicenter prospective cohort study of 412 PAD patients.
• Standardized ABI measurements at multiple time points (baseline to 24 months).
• Development of an ABI dynamic fluctuation index (ABI-DFI) and machine learning models (random forest, SVM, neural network) for MALE prediction.

Main Results:

• The AI model, particularly random forest, showed high predictive performance (td-AUC 0.847 in validation).
• ABI-DFI was a significant independent predictor of MALE (HR=3.42, P<0.001).
• The AI model outperformed traditional single-point ABI values in predicting MALE.

Conclusions:

• AI-driven prediction models using ABI dynamic fluctuations offer superior risk assessment for MALE in PAD.
• This approach enhances precision medicine by enabling individualized treatment decisions.
• The developed model provides a valuable tool for clinicians managing PAD patients.