A two-pronged approach to screening for ovarian cancer -- using CA-125 levels and a mathematical stratification model -- demonstrated nearly 100 percent specificity and also showed improved predictive power, a prospective single-center study found.
Over an 11-year period, screening determined that 2.9 percent of women (117) were high risk, and thereby referred for a transvaginal sonography exam, according to Karen Lu, MD, chair of gynecologic oncology at the the University of Texas MD Anderson Cancer Center, and colleagues.
Eighty-two of these patients had a normal exam, 11 had benign ovarian findings, 14 did not have the exam, and 10 had suspicious findings, they wrote online in the journal Cancer.
All 10 with suspicious findings underwent surgery: four had early-stage high-grade invasive ovarian cancer, two had borderline disease, one had endometrial cancer, and three had benign ovarian tumors -- for a positive predictive value of 40%, which "greatly surpasses the clinical benchmark of 10%," researchers noted.
The specificity of the screening was 99.9 percent as it failed to detect two borderline ovarian cancers.
The two-step approach relies on serial CA-125 measurements combined with the Risk of Ovarian Cancer Algorithm (ROCA), a mathematical model based on the patient's age and CA-125 score.