血清HE4和CA125联合检测预测盆腔包块患者卵巢癌的风险

Prediction of ovarian cancer in patients with pelvic mass using combined assays of HE4 and CA125 markers

目的:探讨新型肿瘤标记物人附睾分泌蛋白4(HE4)和CA125联合检测预测盆腔包块患者上皮性卵巢癌(EOC)风险的价值。方法:将诊断为盆腔包块拟行手术的患者纳入本研究。测定患者术前血清HE4和CA125水平,分别用绝经后和绝经前预测模型(ROMA)将患者划分至高危组和低危组,评估预测模型的应用价值。结果:评估了191例患者,其中138例盆腔良性疾病,53例盆腔恶性肿瘤,包括36例EOC。绝经后组盆腔良性疾病12例,8例划分至低危组,特异性66.7%(95%CI=40.0～93.3),恶性肿瘤37例,35例划分至高危组,敏感性94.6%(95%CI=87.0～101.9),EOC28例(含9例早期),全划分至高危组,敏感性100.0%。绝经前组良性疾病126例,103例划分至低危组,特异性81.7%(95%CI=75.0～88.5),16例恶性肿瘤,12例划分至高危组,敏感性75.0%(95%CI=54.0～96.2),EOC8例(含2例早期),全划分至高危组,敏感性100.0%。结论:ROMA成功地将盆腔恶性肿瘤患者划分至高危组,其中EOC患者全被正确地划分至高危组,可用于将恶性肿瘤患者尤其是EOC患者分流至有治疗经验的肿瘤医师及治疗中心。

Objective：To explore the value with novel tumor maker human epididymis secretory protein 4（HE4）combined with CA125 to validate the predicative model（risk of ovarian malignancy algorithm,ROMA）for discriminating benign from malignant pelvic masses.Methods：Women with pelvic masses planned surgical exploration were enrolled.Preoperative serum levels of HE4 and CA125 were measured.Using the serum CA125 and HE4 values,a predictive index（PI）and predicted probability（PP）were calculated for each patient by the separate logistic regression algorithm based on their menopausal status.The patients were classified into low and high risk groups of malignancy.Results：One hundred and ninety one patients were evaluable,with 138 pelvic benign disease and 53 pelvic malignancy including 36 EOC.In the postmenopausal group,eight of 12 benign cases were classified as the low risk with a specificity of 66.7%（95% CI=40.0~93.3）,and 35 of 37 malignant tumors were classified as the high risk with sensitivity of 94.6%（95%CI=87.0~101.9）.All of 28 EOC cases（including 9 cases with early stage）were classified as the high risk with a sensitivity of 100%.In the premenopausal group,103 of 126 benign cases were classified as the low risk with a specificity of 81.7%（95%CI=75.0~88.5）,and 12 of 16 malignancy were classified as the high risk with a sensitivity of 75.0%（95%CI=54.0~96.2）.All of 8 EOC cases（including 2 cases with early stage）were classified as the high risk with a sensitivity of 100%.Conclusion：A predictive model ROMA can successfully classify patients into high and low risk groups with all of EOC correctly classified as the high risk.ROMA can help to triage patients with possible ovarian cancer to physicians and centers that are experienced in the care and management of patients with ovarian cancer.