Machine learning models effectively predict 30-day mortality after gastrectomy, enhancing surgical decision-making.

• 4.3% of gastrectomy patients experienced 30-day mortality.
• The xgboost model outperformed logistic regression and traditional risk calculators, identifying preoperative blood urea nitrogen and age as key predictors.

Incorporating these models could significantly improve patient selection and care strategies in your practice.

• The xgboost model showed the best predictive accuracy compared to other machine learning and regression models.