To establish a nomogram model for predicting posthepatectomy liver failure (PHLF) in patients with HBV-related hepatocellular carcinoma (HCC).

Clinical data of 628 patients with HBV-related HCC who underwent radical hepatectomy in Eastern Hepatobiliary Surgery Hospital, the Second Military Medical University between January 2010 and December 2010 were retrospectively analyzed. According to the operation time, all patients were divided into the model establishment group (n=471) and validation group (n=157). In the model establishment group, 409 cases were males and 62 females, aged (52±21) years old on average. In the validation group, 135 cases were males and 22 females, aged (52±11) years old on average. The informed consents of all patients were obtained and the local ethical committee approval was received. The independent risk factors of PHLF in the model establishment group were identified by Logistic regression analysis. Based upon the independent risk factors, the nomogram model for predicting PHLF of HCC was established. The accuracy of nomogram model for predicting PHLF was respectively detected in two groups by computer consistency coefficient (C-index) and calibration graph method.

Multivariate Logistic regression analysis for the model establishment group revealed that, PT>13 s (OR=2.522, 95%CI:1.384-4.596; P<0.05), TB>17.1 μmol/L (OR=2.088, 95%CI:1.342-3.251; P<0.05), ALT>44 U/L (OR=1.710, 95%CI:1.141-2.562; P<0.05), positive HBeAg (OR=1.658, 95%CI: 1.058-2.597; P<0.05), intraoperative blood transfusion (OR=3.407, 95%CI:1.945-5.967; P<0.05) and liver cirrhosis (OR=1.835, 95%CI:1.200-2.805; P<0.05) were the independent risk factors for PHLF. After the establishment of nomogram model, the C-index was respectively 0.727 and 0.719 in the model establishment group and validation group. In the calibration graph, the standard curve was properly fit with the predicting calibration curve, suggesting that the model consistency was fine.

The nomogram model for predicting PHLF in patients with HBV-related HCC is successfully established. And the model offers certain guiding significance for clinical treatment of HBV-related HCC.