DOI: 10.1002/jcla.70319 ISSN: 0887-8013

Application of Serum CA125 , HE4 , ROMA , and CPH

Weiming Feng, Chunying Zhang, Lingxin Meng, Yifei Chen, Haiyun Yu, Yanhong Zhai, Zheng Cao

ABSTRACT

Objective

This study aimed to compare the diagnostic accuracy of four serum biomarkers—carbohydrate antigen 125 (CA125), human epididymis protein 4 (HE4), the Risk of Ovarian Malignancy Algorithm (ROMA), and the Copenhagen Index (CPH‐I)—for the preoperative discrimination of benign/borderline from malignant ovarian cancer, with particular attention to menopausal status.

Methods

A retrospective study enrolled 345 patients with histologically confirmed ovarian tumors (187 benign/borderline; 158 malignant). Serum CA125 and HE4 were measured and reported by the electrochemiluminescence immunoassays, along with the calculated indices of ROMA and CPH‐I. Diagnostic performance was assessed by the area under the curve (AUC), sensitivity, and specificity, respectively.

Results

Among the 158 ovarian tumors patients, CPH‐I demonstrated superior discriminative capacity with the highest AUC (0.924), followed by ROMA (0.916), HE4 (0.897), and CA125 (0.865). Menopausal stratification identified CPH‐I as optimal for premenopausal women (AUC 0.886) and ROMA for postmenopausal women (AUC 0.928). ROMA exhibited the highest overall sensitivity (84.8%), whereas CA125 achieved the highest overall specificity (94.1%). In premenopausal patients, HE4 demonstrated exceptional specificity (98.6%), while postmenopausal women showed equivalent specificity between HE4 and CPH‐I (both 97.8%). In combined panels, CA125 + HE4 + CPH‐I yielded the highest overall accuracy (AUC 0.925), and CA125 + HE4 + ROMA provided optimal postmenopausal performance (AUC 0.930).

Conclusions

CPH‐I demonstrated favorable diagnostic performance compared with conventional biomarkers. Our findings advocate for a menopausal status‐guided clinical algorithm, utilizing a CPH‐I‐based panel for premenopausal women and a ROMA‐based panel for postmenopausal women to maximize diagnostic precision.

More from our Archive