Machine learning effectively predicts optimal outcomes in liver surgery

A machine learning model successfully predicted textbook outcomes in liver surgery using preoperative data from over 2,000 patients. The top-performing xgboost model achieved an area under the curve of 0.73. Significant predictors included minimally invasive techniques and patient characteristics such as lower comorbidity scores. Furthermore, meeting textbook outcome criteria was associated with improved overall survival, indicating the model’s potential utility in guiding surgical decision-making and enhancing patient care.