Ovarian cancer is the leading cause of death among gynecological tumors due to its insidious onset and lack of sensitive early diagnostic tools. About 70% of patients are already in advanced stages at the time of diagnosis, and approximately 60-70% of patients relapse or die within 5 years. Therefore, early diagnosis and treatment are key to improving patient survival rates. Although 80% of ovarian cancer patients have elevated serum CA125, only 30-50% of patients with stage I epithelial ovarian cancer show elevated levels, making it insensitive for early detection. HE4 has similar sensitivity to CA125 for diagnosing ovarian cancer but has higher specificity. The positive rate of HE4 in patients with benign gynecological diseases is significantly lower than that of CA125, especially in patients with endometriosis, where the positive rate of HE4 is 3% compared to 67% for CA125. At a specificity of 95%, the sensitivity for stage I ovarian cancer is 45.9%.

The Risk of Ovarian Malignancy Algorithm (ROMA) combines menstrual status and serum levels of CA125 and HE4 to establish a logistic regression model for predicting ovarian cancer. It is considered one of the best indicators for ovarian cancer screening. The ROMA model accurately predicted 87.3% of patients with benign diseases as low-risk for ovarian malignancy and 82.6% of patients with stage I and II epithelial ovarian cancer and all ovarian cancer patients as high-risk. Using this formula, with cut-off points of PP>13.1% (pre-menopausal women) and PP>27.7% (post-menopausal women), Moore divided patients with pelvic masses into high-risk and low-risk groups for ovarian cancer. The overall sensitivity, specificity, positive predictive value, and negative predictive value of this formula are 88.7%, 74.7%, 60.1%, and 93.9%, respectively. The ROMA model has been approved by the FDA for assessing pelvic masses in women.