Functional Liver Imaging Score to Predict Clinically Significant PHLF for Hepatocellular Carcinoma After Resection.

PURPOSE

To develop a model based on Functional Liver Imaging Score (FLIS) to estimate the risk of clinically significant post-hepatectomy liver failure (PHLF) for hepatocellular carcinoma (HCC) after resection.

PATIENTS AND METHODS

This retrospective study analyzed 885 patients with HCC who undergoing liver resection at our medical center between January 2017 and December 2021. Patients were randomly (7:3) assigned to development (n=620) or internal validation (n=265) cohorts. Univariable and multivariable logistic regression analyses were performed to identify independent risk factors for clinically significant PHLF, defined as grade B or C PHLF by the International Study Group of Liver Surgery. Predictive performance was assessed by the area under receiver operator characteristic curves (AUC).

RESULTS

Clinically significant PHLF occurred in 7.7% of the development cohort and 7.2% of the internal validation cohort. Multivariate analysis identified FLIS, major resection and ALBI score as independent predictors of clinically significant PHLF, and a model combining these three variables predicted failure in the development cohort (AUC 0.746, 95% CI 0.673-0.820) and internal validation cohort (AUC 0.717, 95% CI 0.595-0.838). The same model also predicted mortality within 90 days after surgery in the development cohort (AUC 0.704, 95% CI 0.575-0.832) and internal validation cohort (AUC 0.717, 95% CI 0.586-0.848). In both cohorts, overall survival rate was significantly lower among patients whom the model placed at high risk of clinically significant PHLF than among those at low risk.

CONCLUSION

The combination of FLIS and other easily acquired clinical data may reliably predict clinically significant PHLF and mortality in hepatocellular carcinoma.