[Construction and validation of a nomogram model of early related factors for hepatic insufficiency after hemihepatectomy]

Abstract

Objectives: To investigate the early related factors for hepatic insufficiency after hemihepatectomy and to construct and validate a nomogram model. Methods: This was a retrospective cohort study.There were 207 patients with liver tumor who underwent hemihepatectomy in the Department of Hepatobiliary Surgery,Cancer Hospital,Chinese Academy of Medical Sciences from October 2016 to December 2022. Using the random number method,patients were randomly divided into a model group(n=166) and a validation group(n=41) according to an 4︰1 ratio. There were 118 males and 48 females in the modeling group,with an age (M(IQR)) of 59.0(13.3) years (range: 22.0 to 81.0 years),42 patients in the group with postoperative liver insufficiency and 124 patients in the group without postoperative liver insufficiency. There were 32 males and 9 females in the validation group, with an age of 54.0(19.0) years (rang: 25.0~81.0 years). The first results of the peripheral blood test of patients within 24 hours after surgery were collected,and the independent related factors for incomplete postoperative liver function were determined by multivariate Logistic regression analysis,and related factors of postoperative incomplete liver function were screened by best subset selection. A nomogram model of the risk of postoperative hepatic insufficiency after hemihepatectomy was constructed using R software,followed by internal and external validation of the model. Results: Multivariate logistic regression analysis showed that elevated D-dimer level and decreased antithrombin Ⅲ (AT-Ⅲ) activity within 24 hours after surgery were independent related factors for the development of postoperative hepatic insufficiency in hemihepatectomized patients. The results of the best subset selection showed that ALT,D-dimer, and AT-Ⅲ activity levels within 24 hours postoperatively were the most relevant factors for postoperative hepatic insufficiency. The R software was applied to build a nomogram prediction model based on the above three indicators in the model set,and the receiver operating characteristic(ROC) curve of the model showed an area under the curve of 0.803 and the calibration curve showed a U-index of -0.012 for the model(P=0.977). The results of the clinical decision analysis and the clinical impact curve indicated that the model had good clinical utility. The internal validation results of the Bootstrap method suggested that the model had reasonable consistency. The area under the ROC curve of the validation group model was 0.806,suggesting that the model had a good generalization prediction ability. Conclusions: The levels of ALT,D-dimer,and AT-Ⅲ activity within 24 hours after hemihepatectomy are valuable indicators for predicting liver insufficiency after hemihepatectomy. The nomogram model is reliable and can be used as an indicator for close postoperative monitoring.