Prediction of severe posthepatectomy liver failure using albumin-bilirubin score and liver resection rate in patients with primary liver cancer

Abstract

Background: We developed a heatmap-based model to predict severe posthepatectomy liver failure (PHLF) by integrating the albumin-bilirubin (ALBI) score and the liver resection rate in patients with primary liver cancer.

Methods: This retrospective study included 524 patients with primary liver cancer who underwent hepatectomy at our institution between January 2002 and October 2024. The liver resection rate was calculated as: (resected liver volume — tumor volume)/total functional liver volume×100. Severe PHLF was defined as Grade B or C according to the International Study Group of Liver Surgery (ISGLS) criteria. Independent risk factors for severe PHLF were identified using multivariate analysis.

Results: Severe PHLF occurred in 56 patients. Multivariate analysis revealed that a higher ALBI score (odds ratio [OR]=8.91 per 1.0 increase) and greater liver resection rate (OR=1.05 per 1% increase) were independent predictors of severe PHLF. When applying a 50% threshold for predicted PHLF incidence, the heatmap model expanded the surgical indication to 67 additional patients (13%) compared to the Makuuchi criteria. While the positive predictive value was similar between the heatmap model and Makuuchi criteria (90.9% vs. 93.4%), the heatmap model showed a markedly higher negative predictive value (58.8% vs. 32.9%).

Conclusion: The heatmap model is a valuable tool for preoperative risk assessment and surgical decision-making based on hepatic functional reserve, potentially improving the prevention of severe PHLF.