Predicting the Severity of Infrapopliteal Artery Lesions in Patients With Peripheral Artery Disease Using...

Area of Science:

• Cardiovascular Medicine
• Medical Informatics
• Machine Learning in Healthcare

Background:

• Infrapopliteal arterial disease is a complex subtype of peripheral artery disease (PAD).
• Accurate assessment of lesion severity is crucial for effective treatment planning.
• Current methods may not fully capture the nuances of disease severity.

Purpose of the Study:

• To develop an interpretable machine learning model for assessing infrapopliteal artery lesion severity.
• To identify key clinical predictors of disease severity.
• To create a tool supporting clinical decision-making in PAD.

Main Methods:

• Utilized clinical data from 1062 PAD patients (2019-2024).
• Developed and compared 10 predictive models, including Gradient Boosting Machine (GBM).
• Employed feature selection and external validation for model robustness.

Main Results:

• The GBM model achieved an AUC of 0.891 in the validation set.
• Six key predictors were identified for model construction.
• The model demonstrated good calibration and clinical utility via decision curve analysis.

Conclusions:

• The developed GBM model accurately predicts infrapopliteal artery disease severity.
• This interpretable, data-driven tool can enhance risk stratification and clinical decision-making for PAD patients.
• The model offers a non-invasive, rapid decision-support system for clinicians.