Predicting Outcomes Following Lower Extremity Endovascular Revascularization Using Machine Learning.

Area of Science:

• Vascular Surgery
• Machine Learning in Medicine
• Outcome Prediction

Background:

• Peripheral artery disease (PAD) poses significant risks during lower extremity endovascular revascularization.
• Current tools for predicting perioperative outcomes in these procedures are limited.
• Developing accurate prediction models is crucial for patient management and risk stratification.

Purpose of the Study:

• To develop and validate machine learning algorithms for predicting 30-day adverse outcomes after lower extremity endovascular revascularization.
• To identify key preoperative predictors of major adverse limb events or death.

Main Methods:

• Utilized the National Surgical Quality Improvement Program targeted vascular database (2011-2021).
• Included 38 preoperative demographic and clinical variables for 21,886 patients undergoing endovascular revascularization.
• Trained six machine learning models, evaluating performance using the area under the receiver operating characteristic curve (AUC).

Main Results:

• The extreme gradient boosting model achieved the highest AUC of 0.93 for predicting 30-day major adverse limb event or death.
• This significantly outperformed logistic regression (AUC 0.72).
• Top predictors included chronic limb-threatening ischemia, tibial intervention, and congestive heart failure.

Conclusions:

• Machine learning models demonstrate high accuracy in predicting 30-day outcomes using preoperative data.
• The models exhibit good discrimination and calibration, offering a valuable tool for risk assessment.
• Prospective validation is recommended to confirm generalizability and external validity.