Meng Xue-Qin, Miao Hui, Xia Yong, Shen Hao, Qian Yi, Shen Feng, Guo Jia
Department of Ultrasound, Eastern Hepatobiliary Surgery Hospital, Second Military Medical University (Naval Medical University), Shanghai, China.
Department of Medical Genetics, College of Basic Medical Science, Second Military Medical University, Shanghai, China.
Asian J Surg. 2023 Jan;46(1):399-404. doi: 10.1016/j.asjsur.2022.05.001. Epub 2022 May 20.
Post-hepatectomy liver failure (PHLF) is one of the most serious complications after hepatectomy in patients with hepatocellular carcinoma (HCC), and has an association with high morbidity and mortality. This study aimed to explore the risk factors of PHLF and to establish and validate a nomogram to predict PHLF.
We retrospectively analyzed 971 HCC patients undergoing major liver resection at the Eastern Hepatobiliary Surgery Hospital between 2011 and 2016, and established a nomogram based on multivariate analyses for predicting PHLF. The predictive accuracy and discriminative ability of the nomogram were evaluated by concordance index (C-index) and calibration curve. The predictive ability of PHLF of this nomogram was compared with conventional models using receiver operating characteristic (ROC) curves.
The incidence of PHLF was 18.8%. Multivariate analysis identified age, BMI, preoperative ascites, preoperative prealbumin, spleen volume-to-platelet ratio, total bilirubin, and intraoperative blood loss as independent predictors of PHLF. The area under ROC curve (AUROC) of the predictive model was 0.668 and was higher than that of the albumin-bilirubin score (ALBI). The optimal cut-off value of the model was 124.
We constructed a nomogram for predicting PHLF based on risk factors. The nomogram can assist clinicians in identifying patients with high-risk PHLF, eventually facilitating earlier interventions and improving clinical outcomes.
肝切除术后肝衰竭（PHLF）是肝细胞癌（HCC）患者肝切除术后最严重的并发症之一，与高发病率和死亡率相关。本研究旨在探讨PHLF的危险因素，并建立和验证预测PHLF的列线图。
我们回顾性分析了2011年至2016年期间在东方肝胆外科医院接受大肝切除的971例HCC患者，并基于多因素分析建立了预测PHLF的列线图。通过一致性指数（C指数）和校准曲线评估列线图的预测准确性和判别能力。使用受试者工作特征（ROC）曲线将该列线图对PHLF的预测能力与传统模型进行比较。
PHLF的发生率为18.8%。多因素分析确定年龄、BMI、术前腹水、术前前白蛋白、脾体积与血小板比值、总胆红素和术中失血是PHLF的独立预测因素。预测模型的ROC曲线下面积（AUROC）为0.668，高于白蛋白-胆红素评分（ALBI）。该模型的最佳截断值为124。
我们基于危险因素构建了预测PHLF的列线图。该列线图可帮助临床医生识别PHLF高危患者，最终促进早期干预并改善临床结局。
