Abstract
Diagnosis of ovarian cancer is often made at an advanced stage when prognosis is poor. Diagnosis is based on invasive procedures (laparoscopy or laparotomy) and inadequate surgical exploration may adversely affect survival. A number of biomarkers have been proposed for diagnosis or screening but few have made it into clinical practice. As with many conditions where a single sensitive and specific marker is unavailable, algorithms using multiple biomarkers in combination with clinical and/or imaging data have been assessed. In this study, Montagnana et al. assessed the Risk of Ovarian Malignancy Algorithm (ROMA) in determining risk of epithelial ovarian cancer (EOC) in women presenting with a pelvic mass.
ROMA uses CA125 and human epididymis protein 4 (HE4) concentrations and menopausal status to determine a risk score. In this study, 104 women with a pelvic mass and 49 healthy female controls were enrolled. Receiver operating characteristic (ROC) analyses were used to evaluate ROMA as well as each of the biomarkers alone.
Median CA125 and HE4 concentrations were significantly higher in patients with EOC than healthy controls and those with a benign pelvic mass (both P < 0.0001). HE4 had improved the diagnostic performance over CA125, with an area under the curve (AUC) of 0.77 in premenopausal and 0.94 in postmenopausal women (compared with 0.64 and 0.84, respectively, using CA125). Using ROMA, the AUC was 0.77 and 0.92, respectively.
The limited number of subjects in this study may have affected the results. However, the study highlights an important point – use of an algorithm should be assessed against the component parts alone to determine if its use improves the performance. In this study, HE4 alone performed slightly better than ROMA. Additionally, in the differentiation of EOC and benign pelvic mass, HE4 performed better than CA125. Further investigation is required to determine the value HE4 may have in routine practice.