基于肝癌三联检的原发性肝细胞癌微血管侵犯的术前预测研究

AFP,AFP-L3and PIVAK triple biomarkers based-preoperative prediction for microvascular invasion of primary hepatocellular carcinoma

为探讨基于肝癌三联检的原发性肝细胞癌(hepatocellular carcinoma,HCC)微血管侵犯(microvascular invasion,MVI)的术前预测价值,回顾性收集上海东方肝胆外科医院2015年1月至2016年6月所有符合纳入、排除标准的570例患者构建MVI预测模型,采用单因素Logistic回归分析确定影响MVI发生的独立危险因素,这些独立危险因素随后被用来建立相应的预测模型。根据多因素Logistic回归分析结果,预测MVI术前因素有:年龄、血小板(platelet,PLT)、白蛋白(albumin,Alb)、log(AFP)、AFP-L3、log(PIVKA),其受试者工作特征(receiver operating characteristic,ROC)曲线AUC达到0.71(0.67~0.75)。由此可见,基于实验室数据建立的Logistic回归模型能较准确地在术前预测MVI的发生。利用该模型,可评估术前患者发生MVI的风险,从而指导合理的外科治疗。

To investigate the preoperative predictive value of microvascular invasion(MVI)based on triple screening of primary hepatocellular carcinoma(HCC),the data of 570consecutive patients who underwent liver resection for HCC at the Shanghai Eastern Hepatobiliary Surgery Hospital from January 2015to June 2016were collected.Univariate Logistic regression was used to determine the independent risk factors affecting the occurrence of MVI.These independent risk factors were then used to establish the corresponding prediction model.According to multivariate Logistic regression,the preoperative predictors of MVI were age,platelet(PLT),albumin(Alb),log(AFP),AFP-L3and log(PIVKA).The AUC of the receiver operating characteristic(ROC)curve was 0.71(0.66-0.75).The Logistic regression model achieved an optimal preoperative prediction of MVI in HCC patients.Using the model,the risk for an individual patient to harbor MVI could be determined,which could lead to a rational therapeutic choice.