术前血清学联合影像学检查预测肝内胆管癌根治性切除术后患者生存获益的研究

Combined preoperative serological and imaging examinations in predicting survival benefits of patients after radical resection of intrahepatic cholangiocarcinoma

背景与目的:肝内胆管癌(ICC)起病隐匿、侵袭性高,患者往往确诊时已失去了最佳手术时机,接受手术者5年生存率也极低。早期判断患者根治性切除术的生存获益至关重要。本研究依据术前影像学联合血清学指标对ICC根治性切除患者生存获益实行预测,以期对临床判断是否适宜行根治性切除提供指导与参考。方法:回顾性收集2010年1月—2021年12月于中国13家三甲医院行根治性切除的821例ICC患者的影像学与血清学检测资料。影像学指标包括:发现肝脏肿块、肝内胆管扩张、门静脉侵犯、淋巴结侵犯、腹水及结石;血清学指标包括:血红蛋白、白细胞计数、淋巴细胞计数、中性粒细胞计数、甲胎蛋白(AFP)、癌胚抗原(CEA)、糖类抗原19-9(CA19-9)、CA125、丙氨酸氨基转移酶(ALT)、总胆红素(TBIL)、白蛋白(ALB)及凝血酶原时间(PT)。通过单因素与多因素Cox回归筛选目标变量,用目标变量构建CoxPH模型并绘制列线图,用Kaplan-Meier生存分析验证评分与患者预后的关系,通过受试者工作特征(ROC)曲线及校准曲线对模型预测效能进行评估。结果:影像学发现腹水、肝内胆管扩张、淋巴结侵犯与血清学指标CEA>5μg/L、CA19-9>37 U/mL、CA125>40 U/mL是独立预后因素(均P<0.05)。用该六个变量构建CoxPH模型,根据该模型所区分的高风险组患者术后1、3、5年生存率均明显低于低风险组患者(均P<0.05);所构建的列线图具有较好的区分度及有效性。ROC曲线显示,模型1、3、5年预测的曲线下面积分别为0.711、0.721、0.782;模型1、3、5年预测效能均高于独立指标的预测效能。结论:由CA125、腹水、肝内胆管扩张、淋巴结侵犯、CEA、CA19-9这六个术前指标组成的预后模型能较好地对患者进行高低风险分层,并对ICC患者根治性切除术后生存获益进行较精准的个体化预测,对临床医生判断患者是否适宜行根治性切除具有指导意义。

Background and Aims:Intrahepatic cholangiocarcinoma(ICC)is characterized by its insidious onset and high invasiveness,often resulting in patients losing the optimal timing for surgery upon diagnosis,leading to poor 5-year survival rates following surgery.Early prediction of survival benefits from radical resection is crucial.This study was performed to predict survival benefits in ICC patients undergoing radical resection based on preoperative imaging and serum markers to provide guidance and references for clinical decisions regarding the suitability of radical resection.Methods:The imaging and serological data of 821 ICC patients who underwent radical resection in 13 tertiary-grade A-class hospitals in China from January 2010 to December 2021 were retrospectively collected.Imaging data included the presence of liver mass,intrahepatic bile duct dilation,portal vein invasion,lymph node invasion,ascites,and stones.Serum markers had hemoglobin,white blood cell count,lymphocyte count,neutrophil count,alpha-fetoprotein(AFP),carcinoembryonic antigen(CEA),carbohydrate antigen 19-9(CA19-9),CA125,alanine aminotransferase(ALT),total bilirubin(TBIL),albumin(ALB),and prothrombin time(PT).Target variables were selected through univariate and multivariate Cox regression analysis,and using these variables,a CoxPH model was constructed,and a nomogram was also visualized.Survival curves were plotted using Kaplan-Meier analysis to validate the relationship between the scores and patient outcomes.The model's predictive performance was assessed through receiver operating characteristics(ROC)and calibration curves.Results:The independent prognostic factors were imaging indicators of ascites,intrahepatic bile duct dilation,lymph node invasion,and serum markers with CEA>5μg/L,CA19-9>37 U/mL,and CA125>40 U/mL(all P<0.05).The CoxPH model built using the 6 variables demonstrated that patients in the high-risk group identified by the model had significantly lower 1-,3-,and 5-year survival rates compared to the low-risk group(all P<0.05).The model exhibited good discrimination and effectiveness through the constructed nomogram.ROC curves showed that the area under the curve(AUC)for the model's predictions at 1,3,and 5 years were 0.711,0.721,and 0.782,respectively,surpassing the predictive efficacy of individual markers.Conclusion:The prognostic model composed of preoperative CA125,ascites,intrahepatic bile duct dilation,lymph node invasion,CEA,and CA19-9 effectively stratifies patients into high-and low-risk groups and provides accurate individualized predictions of survival benefits following radical resection in ICC patients.This model offers guidance for clinical decisions regarding the suitability of radical resection.