肝细胞癌切除术后早期复发和转移的危险因素分析及Nomogram模型构建

Risk factors analysis and Nomogram model construction of early recurrence and metastasis after resection of hepatocellular carcinoma

目的分析影响肝细胞癌(HCC)切除术后早期复发和转移的危险因素,并构建其Nomogram预测模型。方法回顾性分析2014年12月至2019年12月于武汉市第三医院肝胆外科行手术切除治疗的248例HCC患者的病历资料。基于本院的电子病历数据库,收集入组患者的人口学、临床病理特征和手术相关资料。所有入组患者均于入院48 h内采集外周静脉血样本,进行血常规、血液生物化学和肿瘤标志物检测。按照标准方案于患者术后开始随访,以无复发生存率(RFS)作为主要研究结局,观察HCC切除术后早期复发和转移情况。采用Cox比例风险回归模型分析HCC切除术后早期复发和转移的危险因素。基于多因素Cox比例风险回归模型分析结果构建Nomogram模型。采用ROC曲线评估Nomogram模型预测HCC切除术后早期复发和转移的效能。绘制校准曲线评估Nomogram模型预测RFS与实际值的一致性。结果248例HCC患者的中位随访时间为37.5个月(四分位距:23.0~55.0),其中106例出现肝内复发或转移。患者的1年、3年和5年RFS分别为92.34%、66.13%和57.26%。与未复发组相比,复发组中Child-Pugh分级为B级、肿瘤大小≥5 cm、有微血管侵犯、BCLC分期为B~C期、血清甲胎蛋白(AFP)≥400 ng/mL、系统免疫炎症指数(SII)≥375.3、AST/ALT≥1.03、白蛋白与胆红素比值(ALBI)≥-2.24、γ-谷氨酰转肽酶(GGT)/血小板计数(PLT)≥0.36、预后营养指数(PNI)≤47.2、血小板计数与淋巴细胞计数比值(PLR)≥113.4、中性粒细胞计数与淋巴细胞计数比值(NLR)≥2.14、C反应蛋白与白蛋白比值(CAR)≥0.15和格拉斯哥预后评分(GPS)为1~2分的患者占比均显著升高(P均<0.05)。多因素Cox比例风险回归模型分析结果显示,肿瘤大小、血清AFP、SII和CAR均是影响HCC切除术后早期复发和转移的独立危险因素(P均<0.05)。Nomogram模型预测HCC切除术后早期复发和转移的一致性指数(C-index)为0.787(95%CI:0.693~0.881)。ROC曲线分析结果显示,Nomogram模型预测HCC患者术后1年、3年和5年RFS的曲线下面积(AUC)分别为0.849、0.857和0.885。校准曲线结果显示,Nomogram模型预测HCC切除术后患者的1年、3年和5年RFS与实际值的一致性均较好。结论肿瘤大小、血清AFP、SII和CAR均是影响HCC切除术后早期复发和转移的独立危险因素,基于这4项影响因素构建的Nomogram模型对于患者早期复发和转移的预测效能较高,具有一定的临床应用价值。

Objective This paper intends to analyze the risk factors of early recurrence and metastasis of hepatocellular carcinoma(HCC)after resection,and to construct its Nomogram prediction model.Methods From December 2014 to December 2019,the medical records of 248 HCC patients who underwent surgical resection in the Department of Hepatobiliary Surgery of Wuhan Third Hospital were analyzed retrospectively.Based on the electronic medical record database of the said hospital,the demographic,clinical and pathological characteristics and operation-related data of the patients were collected.Peripheral venous blood samples were collected from all patients within 48 hours after admission,and blood routine,blood biochemistry and tumor markers were detected.Follow-up was started after operation according to the standard scheme,with the recurrence-free survival rate(RFS)as the main outcome,to observe the early recurrence and metastasis of HCC after resection.The Cox proportional hazard regression model was used to analyze the risk factors of early recurrence and metastasis after HCC resection.Based on the analysis results of multivariate Cox proportional hazard regression model,the Nomogram model was constructed.The ROC curve was used to evaluate the effectiveness of Nomogram model in predicting early recurrence and metastasis of HCC after resection.A calibration curve was drawn to evaluate the consistency between the predicted RFS of Nomogram model and the actual value.Results The median follow-up time of 248 HCC patients is 37.5 months(with an interquartile range of 23.0 to 55.0),of which 106 cases have intrahepatic recurrence or metastasis.The RFS rates of 1,3,and 5 years are 92.34%,66.13%,and 57.26%respectively.Compared with the non-recurrent group,the recurrent group had a Child-Pugh of grade B,a tumor size≥5 cm,microvascular invasion and Barcelona clinic liver cancer(BCLC)stage of B–C,which are significantly higher than those in non-recurrent group(P<0.05).In the recurrent group,the alpha-fetoprotein(AFP)≥400 ng/mL,systemic inflammatory index(SII)≥375.3,AST/ALT≥1.03,and albumin to bilirubin ratio(ALBI)≥﹣2.24,gamma-glutamyl transpeptidase(GGT)/platelet count(PLT)≥0.36,prognostic nutritional index(PNI)≤47.2.The ratio of platelet to lymphocyte ratio(PLR)≥113.4,neutrophils to lymphocyte ratio(NLR)≥2.14,C-reactive protein to albumin ratio(CAR)≥0.15,and patients with a Glasgow prognostic score(GPS)of 1–2 are significantly higher than those of the nonrecurrent group(P<0.05).The multivariate Cox proportional hazard regression model shows that the tumor size,serum AFP level,SII,and CAR are independent risk factors for early postoperative recurrence and metastasis of HCC(P<0.05).The C-index of the Nomogram model for predicting recurrence and metastasis of HCC is 0.787(95%CI:0.693–0.881).The ROC analysis shows that the area under the curve(AUC)of RFS in HCC patients predicted by the Nomogram model at 1 year,3 years,and 5 years after HCC operation are 0.849,0.857,and 0.885,respectively.The calibration curve indicates that the 1-year,3-year,and 5-year survival rates predicted by the Nomogram model are in good agreement with the actual survival rates.Conclusions Tumor size,serum AFP level,SII,and CAR are independent risk factors for early recurrence and metastasis of HCC after operation.The Nomogram prediction model based on multivariate Cox proportional hazard regression analysis has a good prediction ability for early recurrence and metastasis of HCC,and has certain clinical application value.