Predicting mortality risk following major lower extremity amputation using machine learning.

Area of Science:

• Vascular Surgery
• Machine Learning in Medicine
• Predictive Analytics

Background:

• Major lower extremity amputation for vascular disease carries high perioperative risks.
• Existing outcome prediction tools for amputation are limited in accuracy.
• Accurate mortality prediction is crucial for clinical decision-making and patient counseling.

Purpose of the Study:

• To develop and validate machine learning (ML) algorithms for predicting 1-year mortality after major lower extremity amputation.
• To compare the performance of ML models against traditional logistic regression.
• To identify key predictors of mortality in this patient population.

Main Methods:

• Utilized the Vascular Quality Initiative (VQI) database (2012-2024) for patients undergoing major lower extremity amputation.
• Collected 75 features (preoperative, intraoperative, postoperative) for 22,828 patients.
• Trained and evaluated six ML models, including Extreme Gradient Boosting (XGBoost), using 10-fold cross-validation, with Area Under the Receiver Operating Characteristic Curve (AUROC) as the primary metric.

Main Results:

• The XGBoost model achieved an AUROC of 0.88 using preoperative data, significantly outperforming logistic regression (AUROC 0.70).
• The XGBoost model demonstrated excellent performance postoperatively (AUROC 0.94) with strong calibration and low Brier scores.
• Key predictors included amputation level, indication, comorbidities, and functional status, with robust performance across diverse subgroups.

Conclusions:

• Developed accurate ML models for predicting 1-year mortality post-major lower extremity amputation.
• These ML algorithms offer significant potential to enhance patient selection, counseling, and shared decision-making.
• The models provide a valuable tool for supporting patient-centered care in a high-risk surgical population.