Novel Algorithm Detects Ovarian Cancer in Patients Carrying BRCA Mutations


Researchers have developed an algorithm for differentiating between healthy women and those with ovarian cancer, as well as differentiating BRCA1 mutated-disease from wild-type disease.

Their findings were published in PLOS One (online December 15, 2017; doi:10.1371/journal.pone.0189641).

Current screening techniques for ovarian cancer in women at high risk consists of a combination of carbohydrate antigen 125 (CA125) and transvaginal ultrasound. However, these tests have low sensitivity and specificity. Researchers hypothesize that screening for ovarian cancer could be improved by using several biomarkers, which have been evaluated in average risk patients but not in patients who carry BRCA mutations.

Daphne Gschwantler-Kaulich, department of obstetrics and gynecology, Cancer Comprehensive Cancer, Medical University Vienna (Austria), and colleagues utilized a multiplex, bead-based, immuno-assay system to analyze biomarkers in 26 healthy wild-type patients, 26 healthy BRCA1 mutation carriers, 28 patients with wild-type ovarian cancer, and 26 patients with BRCA1-mutated ovarian cancer. Researchers analyzed the concentrations of leptin, prolactin, osteopontin, insulin-like growth factor II, macrophage inhibitory factor, CA125, and human epididymis antigen 4.

Results of the analysis showed a high overall sensitivity (94.3%) in differentiating healthy control patients from patients with ovarian cancer. Comparable results were observed in the wild-type subgroup (sensitivity, 92.8%; AUC, 0.988; P = 1.7e-15) as well as in BRCA1 mutation carriers (sensitivity, 95.2%; AUC, 0.978; P = 1.7e-15) at an overall specificity of 92.3%.


Related Content

USPSTF Updates Guidelines for Ovarian Cancer Screening

Ovarian Cancer Screening: Harms Still Outweigh Benefits, USPSTF Says


The algorithm researchers used also helped identify healthy BRCA1 mutation carriers when compared with healthy wild-type women (sensitivity, 88.4%; specificity, 80.7%; AUC, 0.895; P = 6e-08), with a less-pronounced association with patients with ovarian cancer (sensitivity, 66.7%; specificity, 67.8%; AUC, 0.724; P = .00065).

In their concluding remarks, researchers wrote that their algorithm successfully differentiates between healthy women and ovarian cancer patients, while also differentiating BRCA mutation carriers from wild-type patients. “Large prospective trials with mainly early-stage ovarian cancer cases are warranted” to validate this suggested benefit to the current ovarian cancer detection strategies, they wrote.—Zachary Bessette