A Predictive Score For Post-hepatectomy Liver Failure In Patients With Hepatocellular Carcinoma After Liver Resection: The PA-TBil-PLR Score

Objective:To investigate the value of the PA-TBil-PLR score in predicting post-hepatectomy liver failure(PHLF)in patients with hepatocellular carcinoma(HCC)after liver resection.Methods:A retrospective analysis was conducted on 558 patients with HCC who underwent radical surgery at Affiliated Hospital of Guilin Medical University and The Second Affiliated Hospital of Guilin Medical University from June 2017to December 2022.The patients were randomly divided into a training group(420cases)and a validation group(138 cases)according to a 3:1 random matching principle.The training group was further divided into a PHLF group(52 cases)and a control group(368 cases)based on whether PHLF occurred.Univariate analysis was used to screen for predictive factors of PHLF.Binary multivariate logistic regression analysis was used to identify that target variables are independent predictive factors,with which we establish a scoring model to predict PHLF.The predictive ability of the model for PHLF was evaluated using the area under the receiver operating characteristic curve(AUC)and compared with classic scores such as Child-Pugh score,Model for End-Stage Liver Disease(MELD)score,Albumin-Bilirubin(ALBI)score,and Platelet-Albumin-Bilirubin(PALBI)score.The optimal cutoff value of the score was determined based on the Youden index.And patients were divided into high-risk and low-risk groups using binary classification.The correlation between the risk groups and PHLF was analyzed using the validation group.Results:A total of 67 patients developed PHLF in this study,with 52 cases in the training group and 15 cases in the validation group.There was no statistically significant difference in baseline characteristics between the two groups(P>0.05).Compared to the control group,the PHLF group had increased age,larger tumor size,high rates of liver metastasis,vascular invasion,liver cirrhosis,significant portal hypertension,ascites,resection of≥3 liver segments,intraoperative blood loss≥400 m L,and HBV-DNA≥10~3IU/m L.PLT,PLR,TBil,and AST levels were increased,while ALB and PA levels were decreased,PT was prolonged,and Child-Pugh score,MELD score,ALBI score,PALBI score,and BCLC stage were higher.These differences were statistically significant(P<0.05).Binary logistic regression analysis showed that age,extent of liver resection,HBV-DNA,PLR,TBil,and PA were independent predictors of PHLF.In the training group,the PA-TBil-PLR score(AUC=0.937)was a better predictor of PHLF than Child-Pugh score(AUC=0.865),MELD score(0.777),ALBI score(0.902),and PALBI score(0.797).Similar results were obtained in the validation group(AUC:0.952 vs 0.764-0.885).To exclude the influence of high-level HBV-DNA on PHLF,the predictive ability of each score in the validation group was calculated according to HBV-DNA<103IU/m L,with similar results obtained.To eliminate the effect of large liver resection and small residual liver volume on PHLF,the predictive ability of each score in the validation group was calculated according to resection of<3 liver segments,with similar results obtained.The optimal cutoff value was determined to be 1.388 according to the Youden index,with a sensitivity of 84.0%and a specificity of 91.8%for predicting PHLF.The incidence of PHLF was significantly higher in the high-risk group(score≥1.388)than in the low-risk group(score<1.388).Conclusion:The PA-TBil-PLR score has good predictive ability for PHLF in patients undergoing liver resection for HCC and performs better than Child-Pugh score,MELD score,ALBI score,and PALBI score.It can be used as a new,noninvasive,reliable,simple,and economical model for predicting PHLF.