A new machine learning model improves prediction of liver failure after major hepatectomy, enhancing surgical decision-making.

• The pilot model achieved AUCs as high as 0.904, significantly outperforming traditional models (AUCs 0.502-0.644).
• The model integrates 55 variables, including novel biomarkers, across preoperative, intraoperative, and postoperative datasets.

This innovation allows surgeons to identify high-risk patients within hours of surgery, enabling tailored perioperative strategies.

• Early risk stratification could reduce morbidity and improve outcomes in liver surgery patients.