肝癌切除术和TACE治疗后肿瘤复发的预测模型构建与评估

Construction and evaluation of a nomogram prediction model for tumor recurrence after transcatheter arterial chemoembolization and deletion of liver cancer surgery

目的:分析肝癌切除术后行经肝动脉化疗栓塞(TACE)治疗患者肿瘤复发的影响因素,并基于影响因素构建临床预测模型和绘制列线图。方法:选取在本院进行根治性切除术后行TACE治疗的肝癌患者215例,收集患者临床特征资料及性别、年龄等人口学特征资料,并对患者进行为期1年的随访,根据肿瘤是否复发将患者分为复发组与未复发组。单因素及多因素Logistic回归分析肝癌切除术和TACE治疗后肿瘤复发的影响因素,并构建基于影响因素的临床预测模型,然后通过H-L拟合度曲线、ROC曲线及曲线下面积(AUC)对模型进行评估。最后,将预测模型可视化为列线图。结果:57例患者发生肿瘤复发,占比26.51%;对纳入患者资料进行单因素分析,结果显示,性别、血管癌栓、包膜、肿瘤直径、分化程度、卫星灶、浸透肝被膜为肝癌切除术和TACE治疗后复发的影响因素(P<0.05),但与年龄、肝硬化、乙肝、手术切缘、术中腹腔积液、病理分类、肝门总阻断时间及肝功能Child-Pugh分级无关(P>0.05);将具有显著差异的单因素分析结果纳入进行多因素Logistic回归分析,结果显示,血管癌栓、低分化、肿瘤直径≥5 cm为肝癌切除术和TACE治疗后复发的独立危险因素(P<0.05);基于多因素Logistic回归分析结果构建肝癌术后TACE治疗后复发的预测模型,对该模型以H-L拟合度曲线评估其校准度,结果显示,χ;=6.432,P=0.415。以ROC曲线下面积(AUC)评估该模型的区分度,结果显示,AUC为0.880,敏感度为87.7%,特异度为85.4%。结论:血管癌栓、低分化、肿瘤直径≥5 cm为肝癌术后TACE治疗后肿瘤复发的独立危险因素,以此构建的预测模型具有较好的预测价值及区分度、校准度,可作为临床早期干预的有效手段。

Objective:To analyze the influencing factors of tumor recurrence in patients undergoing transhepatic arterial chemoembolization(TACE)after deletion of liver cancer surgery,and construct a nomogram prediction model,and then evaluate the model.Methods:A total of 215 liver cancer patients treated with TACE after radical resection in our hospital were selected to collect clinical and demographic characteristics such as gender and age,and followed up for 1 year,and divided into relapsed and unrelapsed groups according to whether the tumor relapsed.The influencing factors of univariate and multivariate logistic regression were analyzed for liver cancer recurrence after TACE and liver cancer,and a clinical predictive model based on influencing factors was constructed,and then evaluated by H-L fitting curve,receiver operating characteristic(ROC)curve and area under the curve(AUC).Finally,the predictive model was visualized as a nomogram.Results:In this study,the included patients were followed up for 1 year,and the results showed that 57 patients had tumor recurrence,accounting after TACE treatment of liver cancer(P<0.05),but it had no connection with age,liver cirrhosis,hepatitis B,surgical margin,intraoperative ascites,case classification,total hilar block time and Child-Pugh classification of liver function(P>0.05);the results of univariate analysis with significant differences were included in the multivariate logistic regression analysis,the results showed that vascular tumor thrombus,poor differentiation,and tumor diameter≥5 cm were independent risk factors for recurrence after TACE treatment of liver cancer(P<0.05);based on the results of multivariate Logistic regression analysis,a nomogram prediction model for recurrence of liver cancer after TACE was constructed,and the model was evaluated with an H-L fit curve to evaluate its predictive effectiveness,the results showed thatχ;=6.432,P=0.415.The area under the ROC curve was used to evaluate the discrimination of the model,the results showed that the area under the ROC curve was 0.880,the sensitivity was 87.7%,and the specificity was 85.4%.Conclusions:Vascular tumor thrombus,poor differentiation,and tumor diameter≥5 cm are independent risk factors for tumor recurrence after postoperative TACE of liver cancer.The predictive model of nomogram constructed on this basis has good predictive value and distinguishability,and it can be used as an effective means of early clinical intervention.Vascular cancer embolus,low differentiation and tumor diameter of 5 cm are independent risk factors for tumor recurrence after postoperative TACE of liver cancer.The model constructed by this method has a good predictive value,differentiation and calibration,which can be used as an effective means for early clinical intervention.