Establishment and Internal Validation of a Prognostic Score for Post-hepatectomy Liver Failure Based on Functional Liver Parameters Estimated via TC-99m GSA.

Abstract

Background The 99mTc-galactosyl human serum albumin (Tc-99mGSA) scintigraphy evaluates the future remnant liver function, which is an important prognostic factor for post-hepatectomy liver failure (PHLF). This study aimed to establish a new prognostic score for PHLF, including the functional liver parameters evaluated by Tc-99mGSA scintigraphy. Materials and methods This study reviewed a single-center, retrospective 368-patient database of those who underwent open and laparoscopic hepatectomy in Meiwa Hospital from January 2016 to October 2021. Moreover, 102 patients who underwent Tc-99mGSA scintigraphy following hepatectomy were analyzed. The index of blood clearance of the tracer was calculated from the uptake ratio of heart at 15 minutes to that at 3 minutes (HH15) and the index of hepatic accumulation was calculated from the uptake ratio of liver to liver plus heart at 15 minutes after the injection (LHL15) were calculated for the general functional parameters. The maximal removal rate of Tc-99mGSA (GSARmax) was also calculated, then the GSARmax of the remnant liver (GSARmax-RL) was estimated as the future remnant liver function depending on the hepatectomy. Multivariate analysis was conducted to identify the PHLF predictor, and then a risk-scoring system was established with the 1,000-times bootstrapped validation. Results PHLF (grade ≥ B) was observed in 13 of 102 patients. Multivariate analysis revealed that PHLF was independently predicted by GSARmax-RL (<0.26 mg/min) and LHL15 (<0.89). The risk score was assigned to each itemand then classified into four subgroups, with a predicted PHLF of 3.7%, 14.4%, 42.8%, and 76.8%. Receiver operating characteristic (ROC) curve analysis demonstrated good discrimination (adjusted area under the curve (AUC) after bootstrapped validation, 0.779). The ROC curve analysis compared with other prognostic scores showed that the new model had the highest AUC values for accuracy. Conclusions The new prognostic score based on Tc-99mGSA scintigraphy could recognize patients with a high risk of progressing to PHLF and be helpful in planning therapeutic strategies.