Related Experiment Video
Updated: Jul 8, 2025
07:25
Predicting Amputation using Local Circulating Mononuclear Progenitor Cells in Angioplasty-treated Patients with Critical Limb Ischemia
Published on: September 22, 2020
3.4K
Using Machine Learning (XGBoost) to Predict Outcomes After Infrainguinal Bypass for Peripheral Artery Disease
Ben Li1,2,3,4, Naomi Eisenberg5, Derek Beaton6
1Department of Surgery, University of Toronto, Toronto, ON, Canada.
Annals of Surgery
|December 20, 2023
Summary
Machine learning models accurately predict outcomes after infrainguinal bypass surgery. These advanced algorithms significantly outperform traditional logistic regression for predicting major adverse limb events or death.
Area of Science:
- Vascular Surgery
- Medical Informatics
- Machine Learning in Healthcare
Background:
- Infrainguinal bypass is crucial for peripheral artery disease but involves high surgical risks.
- Existing tools for predicting outcomes after infrainguinal bypass are limited.
Purpose of the Study:
- To develop and evaluate machine learning algorithms for predicting patient outcomes following infrainguinal bypass.
- To compare the performance of machine learning models against traditional logistic regression.
Main Methods:
- Utilized the Vascular Quality Initiative database (2003-2023) with 59,784 infrainguinal bypass cases.
- Developed and trained six machine learning models using preoperative, intraoperative, and postoperative variables.
- Evaluated models using Area Under the Receiver Operating Characteristic Curve (AUROC) and Brier scores.
Main Results:
- The XGBoost machine learning model achieved superior prediction accuracy, with an AUROC of 0.94 (preoperative) to 0.96 (postoperative).
- Logistic regression showed significantly lower predictive performance (AUROC 0.61).
- Models demonstrated good calibration and robustness across different stages of patient care.
Conclusions:
- Machine learning models offer a highly accurate method for predicting outcomes after infrainguinal bypass.
- These ML models surpass the predictive capabilities of logistic regression, potentially improving patient management.
