Logistic回归和 ROC工作曲线评价五种肿瘤标志物在乳腺癌诊断中的价值

目的 综合应用Logistic回归和 ROC工作曲线评价癌胚抗原(carcinoembryo antigen,CEA)、血清糖类抗原125(serum carbohydrate antigen 125, CA125)、血清糖类抗原153(carbohydrate antigen153,CA153)、降钙素原 (procalcitonin,PCT)、人附睾蛋白4( human epididymis protein 4,HE4)五种肿瘤标志物在乳腺癌诊断中的价值。 方法 采用罗氏全自动电化学发光仪检测42例乳腺癌患者（乳腺癌组）和42例健康者（对照组）的血清CEA、CA125、CA153、PCT、HE4水平,分析比较乳腺癌组与对照组血清CEA、CA125、CA153、PCT、HE4水平差异;运用Logistic回归分析筛选出诊断乳腺癌有效指标,并用ROC曲线综合评价各种指标在乳腺癌诊断中的价值。 结果 乳腺癌组血清CEA、CA125、CA153、PCT、HE4水平高于对照组,差异有统计学意义（P<0.05）;把CA125、PCT从自变量中筛除（P>0.05）,设自变量X₁=CEA、X₂=CA153、X₃=HE4,建立Logistic回归方程：Logit(P)=-17.806+0.715X₁+0.157X₂+0.078X₃;通过ROC曲线分析,HE4、CEA、CA153曲线下面积分别为0.843、0.775、0.716,在单独检测中HE4的曲线下面积最大,三项联合检测曲线下面积（0.906）大于其他三项。 结论 HE4、CEA、CA153可作为诊断乳腺癌的有效指标,三项肿瘤标志物联合检测可以在一定程度上弥补单项指标的不足。

Logistic regression and ROC work curve evaluated the value of five tumor markers in breast cancer diagnosis

Objective Logistic regression and ROC curve were used to evaluate the value of CEA, CA125, CA153, PCT and HE4 tumor markers in breast cancer diagnosis. Methods Using roche automatic electrochemical luminescence instrument detection of the levels of serum CEA, CA125, CA153, PCT, HE4 in 42 patients with breast cancer and 42 cases of healthy subjects, through independent sample t-test analysis between the two groups;Logistic regression analysis was used to identify the effective indicators, and evaluate the diagnostic value of various indicators using the ROC curve. Results Serum CEA, CA125, CA153, PCT and HE4 levels of breast cancer group were significantly higher than those in the control group, and the difference was statistically significant (P<0.05).The CA125 and PCT were sifted from the independent variables (P>0.05), and the independent variables X₁=CEA, X₂=CA153, X₃=HE4, and the Logistic regression equation was established: Logit(P)=-17.806+0.715X₁+0.157X₂+0.078X₃; The area of HE4, CEA and CA153 were 0.843, 0.775 and 0.716 respectively through ROC curve, and the area of HE4 was the largest under the single test.The area under three joint detection curves (0.906) was greater than the other three. Conclusion HE4, CEA, CA153 can be used as an effective indicator for diagnosing breast cancer, and the joint detection of three tumor markers can make up the deficiency of the single index to some extent.