Superiority of the ROMA index over HE4 and CA125 for diagnosing ovarian cancer in a Bangladeshi cohort

• ROMA showed higher accuracy (AUC: 0.93–0.95) than HE4 and CA-125 for ovarian cancer. • ROMA, HE4 discriminated cancer vs benign (AUC ≥ 0.93), unlike CA-125 (0.60). • ROMA validated as a specific triage tool for ovarian cancer in low-resource settings. • Study provides key biomarker performance data from a Bangladeshi cohort. Ovarian cancer is a leading cause of mortality among women globally. This study aimed to evaluate the diagnostic value of CA-125, HE4, and the ROMA index in identifying ovarian cancer in patients in Bangladesh. This case-control study was conducted at the National Institute of Cancer Research and Hospital (NICRH) in 50 patients with histologically confirmed epithelial ovarian cancer, 25 patients with benign ovarian diseases, and 25 healthy controls. Women with prior chemotherapy or hormonal therapy, non-epithelial ovarian cancer, bilateral oophorectomy, pregnancy, or serious systemic illness were excluded. Serum CA-125 and HE4 levels were measured using Chemiluminescence Immunoassay (CLIA), and the ROMA index was calculated. Receiver Operating Characteristic (ROC) analysis was performed to assess diagnostic performance. The mean ages of participants were 52.5-years for ovarian cancer patients, 38.2-years for those with benign diseases, and 45.0-years for healthy controls. The median serum levels of HE4, CA-125, and ROMA index were significantly higher in ovarian cancer patients compared to controls. ROC analysis revealed AUCs of 0.95 for ROMA, 0.92 for HE4, and 0.88 for CA-125 in differentiating cancer from healthy controls, and 0.93 for ROMA, 0.94 for HE4, and 0.60 for CA-125 in differentiating cancer from benign ovarian diseases. ROMA and HE4 significantly distinguished ovarian cancer from benign conditions ( p < 0.001), while CA-125 did not ( p = 0.165). In this study, the ROMA index appeared to perform better than HE4 and CA-125, though findings should be interpreted cautiously, considering the limited statistical power.