摘要
目的筛选卵巢癌实验诊断比较适合的肿瘤标志,进一步探究其临界值。方法将研究对象分为卵巢癌组、卵巢良性肿瘤组和健康体检组,采用电化学发光法检测CA153、CA724、CA125、人附睾蛋白4(human epididymis protein 4,HE4)等项目的浓度,并以文献报道的方法计算卵巢恶性肿瘤风险模型(risk of ovarian malignancy algorithm,ROMA);以SPSS软件分析研究对象肿瘤标志的表达差异,并分析它们的诊断指数(灵敏度+特异性);根据受试者工作特征曲线(receiver operating characteristics,ROC),试确定各肿瘤标志的最佳临界值。结果三组的肿瘤标志表达水平存在显著差异(P<0.05);ROMA的诊断能力最佳(诊断指数1.94,灵敏度0.94、特异性1.00),其它依次是HE4(诊断指数1.86,灵敏度0.89、特异性0.97)、CA125(诊断指数1.75,灵敏度0.83、特异性0.91);依据曲线下面积(area under the curve,AUC),诊断卵巢癌能力大小的肿瘤标志(或项目)分别是:ROMA、HE4、CA125、CA153、CA724。结论诊断卵巢癌应优先考虑CA125、HE4和ROMA等项目,相应临界值可拟定为90.96 U/ml、81.38 pmol/l和37.22%。
Objective The aim of this work is to screen suitable tumor markers (TMs)and to determine their cutoff values in patients with ovarian cancer. Methods Serum samples of 179 women,including 77 ovarian cancer(group A),62 ovarian benign cysts(group B),and 40 healthy controls(group C),were tested for CA153,CA724,CA125 and Human Epididymis Protein 4(HE4)using fully automated instruments (Roche Elecsys cobas E). Risk of Ovarian Malignancy Algorithm(ROMA)values were calculated according to references. Differences of tumor marker concentration among three groups were evaluated by SPSS software. Using Receiver Operating Characteristics (ROC)curve,the cutoff values of tumor markers were determined. Results Significant differences of TMs between group A and group B,also group A and group C,were observed(P〈0. 05). ROMA was the best diagnostic item with a good sensitivity(0. 94)and specificity(1. 00),and the other recommended diagnostic items for ovarian cancer were HE4(0. 89 for sensitivity and 0. 97 for specificity)and CA125(0. 83 and 0. 91,respectively). According to area under the curve(AUC),the diagnostic ability of TMs for ovarian cancer was,successively,ROMA,HE4,CA125,CA153,CA724. Conclusions This study show that CA125,HE4 and ROMA should be recommended to diagnosis ovarian cancer,and the preliminary cutoff values for CA125,HE4 and ROMA are 90. 96 U/ml,81. 38 pmol/l and 37. 22 %,respectively.
出处
《中国肿瘤外科杂志》
CAS
2014年第5期288-291,共4页
Chinese Journal of Surgical Oncology