首页 | 本学科首页   官方微博 | 高级检索  
检索        

血清及胸腔积液中四种肿瘤标志物联合应用对良恶性肿瘤鉴别诊断价值的评估
引用本文:陈阳育,徐莉莉,伍燕兵,王臻,施焕中,夏一帆,梁宝生.血清及胸腔积液中四种肿瘤标志物联合应用对良恶性肿瘤鉴别诊断价值的评估[J].中华肿瘤防治杂志,2021(3):212-222.
作者姓名:陈阳育  徐莉莉  伍燕兵  王臻  施焕中  夏一帆  梁宝生
作者单位:首都医科大学附属北京朝阳医院呼吸与危重症医学科;北京大学医学部医学技术研究院;北京大学公共卫生学院生物统计系
基金项目:首都卫生发展科研专项项目(2020-2-1062)。
摘    要:目的探讨肿瘤标志物CEA、CA125、CA15-3及CA19-9联合应用对鉴别良性胸腔积液(BPE)和恶性胸腔积液(MPE)的价值。方法收集327例2013-01-01-2015-06-30在首都医科大学附属北京朝阳医院(174例)和华中科技大学同济医学院附属协和医院呼吸与危重症医学科(153例)住院的胸腔积液(PE)患者,其中MPE患者119例,BPE患者208例。取PE标本及配对血清标本,应用化学发光法检测CEA、CA125、CA15-3及CA19-9在血清及PE中的浓度,应用二元Logistic回归模型和L1正则化(LASSO)方法将患者基本信息与PE、血清中4种肿瘤标志物CEA、CA125、CA15-3及CA19-9进行不同方式的联合,通过受试者工作特征(ROC)曲线分析和比较不同联合诊断模型的诊断价值。结果PE中CEA+CA15-3+CA19-9的联合模型对应ROC曲线下面积(AUC)值最大(0.90),血清中此联合模型对应的AUC值也是最佳(0.863),PE中的联合模型优于血清中的联合诊断模型,P=0.0125,综合预测能力最强。PE与血清肿瘤标志物浓度差值中CEA+CA15-3+CA19-9的联合模型对应的灵敏度最佳(80.2%),特异度为79.1%。基于LASSO变量选择方法的联合模型在PE中的特异度最佳(96%),此时灵敏度为73%,阳性似然比22。以上结果均P<0.001。PE中CEA+CA15-3+CA19-9的联合模型对应的AUC值(0.90)优于PE中CEA对应的AUC值(0.824),P<0.001。结论联合应用CEA、CA15-3及CA19-9在诊断效能、灵敏度、特异度、阳性似然比等方面优于其他组合,且优于PE中的CEA对BPE/MPE的鉴别诊断价值。

关 键 词:胸腔积液  肿瘤标志物  联合  诊断  价值

Evaluation of combined application of four tumor markers in serum and pleural effusion in differential diagnosis value of benign and malignant tumors
CHEN Yang-yu,XU Li-li,WU Yan-bing,WANG Zhen,SHI Huan-zhong,XIA Yi-fan,LIANG Bao-sheng.Evaluation of combined application of four tumor markers in serum and pleural effusion in differential diagnosis value of benign and malignant tumors[J].Chinese Journal of Cancer Prevention and Treatment,2021(3):212-222.
Authors:CHEN Yang-yu  XU Li-li  WU Yan-bing  WANG Zhen  SHI Huan-zhong  XIA Yi-fan  LIANG Bao-sheng
Institution:(Department of Respiratory and Critical Care Medicine,Beijing Chaoyang Hospital Affiliated to Capital Medical University,Beijing 100020,China;Peking University,Beijing 100191,China)
Abstract:Objective To investigate the diagnostic value of tumor markers CEA,CA125,CA15-3,and CA19-9 in differentiating benign pleural effusion(BPE)and malignant pleural effusion(MPE).Methods From January 1 st,2013 to June30 th,2015,327 patients with pleural effusion(PE)admitted in Beijing Chaoyang Hospital(174 cases)and Union Hospital Affiliated to Tongji Medical College of HUST(153 cases)were collected,including 119 patients with MPE and 208 patients with BPE.Take PE specimens and matched serum specimens,apply chemiluminescence method to detect the concentration of CEA,CA125,CA15-3 and CA19-9 in serum and PE,and apply binary logistic regression model and L1 regularization(Least absolute shrinkage and selectionator operator operator,LASSO)method to combine the basic information of the patient with the four tumor markers of CEA,CA125,CA15-3 and CA19-9 in PE and serum in different ways.The diagnostic value of different combined diagnostic models are analyzed and compared through receiver operating characteristics(ROC)curve.Results The area under the ROC curve(AUC)corresponding to the joint model of CEA+CA15-3+CA19-9 in PE had the maximum value of 0.90,and the model was with strongest comprehensive predictive ability.The sensitivity corresponding to the joint model of CEA+CA15-3+CA19-9 in the concentration difference between PE and serum tumor markers was the best(80.2%),with a specificity of 79.1%.The joint model based on the LASSO variable selection had the best specificity in PE(96%),and the sensitivity was 73%.The P values of the above results were less than 0.001.Conclusions The combined application of CEA,CA15-3 and CA19-9 is superior to other combinations in terms of diagnostic efficiency,sensitivity,specificity,and positive likelihood ratio,but it is not significantly better than the differential diagnosis of CEA in PE for BPE/MPE.Conclusion The combined application of CEA,CA15-3,and CA19-9 is superior to other combinations in terms of diagnostic efficiency,sensitivity,specificity,and positive likelihood ratio,and is superior to the value of CEA in PE for differential diagnosis of BPE/MPE.
Keywords:Pleural effusion  Tumor Markers  Combinations  Diagnosis  Value
本文献已被 维普 等数据库收录!
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号