摘要目的 探讨综合运用Logistic回归和受试者工作特征曲线(ROC曲线)分析四项肿瘤标志物对肺癌的诊断价值.方法 采用放射免疫法检测1 12例原发性肺癌和74例肺良性疾病患者血清中癌胚抗原(CEA)、糖类抗原-125(CA125)、细胞角蛋白片段19(CYFRA21-1)和神经元特异性烯醇化酶(NSE)的表达水平.通过Logistic回归建立回归模型,用ROC曲线分析4项肿瘤标志物在肺癌诊断中的意义.结果 肺癌患者血清中CEA、CA125、CYFRA21-1和NSE的表达水平[4.53(2.22-11.53) ng/ml、28.97(11.39-62.10) U/ml、4.05(2.29-8.18) ng/ml、14.11(11.35-24.12)ng/ml]明显高于肺良性疾病患者[2.08(1.45-2.52) ng/ml、12.90(9.80-19.44)U/ml、1.53(1.21-2.17) ng/ml、11.38(9.07-12.80) ng/ml],差异有统计学意义(均P＜0.01).通过Logistic回归建立回归方程Y=1/[1 +EXP(4.902-0.394X1-0.627X2-0.165X3)],经ROC曲线分析,新变量Y的ROC曲线下面积(AUC)为0.915 ±0.020,敏感度79.46％、特异度93.24％、准确度84.95％.结论 运用Logistic回归和ROC曲线综合分析可提高肺癌的诊断.

abstractsObjective To explore the diagnostic value of four tumor markers analyzed with Logistic regression and receiver operator characteristic (ROC) curve in patients with lung cancer.Methods The serum levels of carcinoembryonic antigen (CEA) 、carbohydrate antigen 125 (CA125),cytokeratin 19 fragment (CYFRA21-1) and neuron specific enolase (NSE) were determined by radioimmunoassay in 112 patients with lung cancer and 74 patients with benign pulmonary disease.Four tumor markers were analyzed by Logistic regression and ROC curve.Results The serum levels of CEA,CA125,CYFRA21-1 and NSE in lung cancer patients were [4.53(2.22-11.53)ng/ml,28.97 (11.39-62.10) U/ml,4.05(2.29-8.18)ng/ml,14.11 (11.35-24.12) ng/ml],respectively,which were significantly higher than those in health adults[2.08 (1.45-2.52) ng/ml,12.90 (9.80-19.44) U/ml,1.53 (1.21-2.17) ng/ml,11.38 (9.07-12.80) ng/ml] (all P ＜ 0.01).According to regression equation Y =1/[1 + EXP (4.902-0.394X1-0.627X2-0.165X3)],the area under the ROC curve (AUC),sensitivity,specificity,and accuracy of new variable Y were 0.915 ± 0.020,79.46％,93.24％,and 84.95％,respectively.Conclusions Application of logistic regression and ROC curve increases diagnostic accuracy in lung cancer.