Clinical Management of BRCA Mutation Carriers
Level of evidence: 5
Risk of ovarian cancer algorithm
A novel modification of CA 125 screening is based on the hypothesis that rising CA 125 levels over time may provide better ovarian cancer screening performance characteristics than simply classifying CA 125 as normal or abnormal, based on an arbitrary cut-off value. This has been implemented in the form of the risk of ovarian cancer algorithm (ROCA), an investigational statistical model that incorporates serial CA 125 test results and other covariates into a computation which produces an estimate of the likelihood that ovarian cancer is present in the screened subject. The first report of this strategy, based on reanalysis of 5,550 average-risk women from the Stockholm Ovarian Cancer screening trial, suggested that ovarian cancer cases and controls could be distinguished with 99.7% sensitivity, 83% specificity, and a PPV of 16%. That PPV represents an eightfold increase over the 2% PPV reported with a single measure of CA 125. This report was followed by applying the ROCA to 33,621 serial CA 125 values obtained from the 9,233 average-risk postmenopausal women in a prospective British ovarian cancer screening trial. The area under the receiver operator curve increased from 84% to 93% (P = 0.01) for ROCA compared with a fixed CA 125 cutoff. These observations represented the first evidence that preclinical detection of ovarian cancer might be improved using this screening strategy. A prospective study of 13,000 normal volunteers aged 50 years and older in England used serial CA 125 values and the ROCA to stratify participants into low, intermediate, and elevated risk subgroups. Each had its own prescribed management strategy, including TVUS and repeat CA 125 either annually (low risk) or at 3 months (intermediate risk). Using this protocol, ROCA was found to have a specificity of 99.8% and a PPV of 19%.
Two prospective trials in England utilized the ROCA. The United Kingdom Collaborative Trial of Ovarian Cancer Screening (UKCTOCS) randomly assigned normal-risk women to either (1) no screening, (2) annual ultrasound, or (3) multimodal screening (n = 202,638; accrual completed; follow-up ends in 2014), and the U.K. Familial Ovarian Cancer Screening Study (UKFOCSS) targeted high-risk women (accrual completed). There are also two high-risk cohorts using the ROCA under evaluation in the United States: the Cancer Genetics Network ROCA Study (n = 2,500; follow-up complete; analysis underway) and the Gynecologic Oncology Group Protocol 199 (GOG-0199; enrollment complete; follow-up ends in late 2011). Thus, additional data regarding the utility of this currently investigational screening strategy will become available within the next few years.
Level of evidence: 4
Miscellaneous new markers
A wide array of new candidate ovarian cancer biomarkers has been described during the past decade, including HE4; mesothelin; kallekreins 6, 10, and 11; osteopontin; prostasin; M-CSF; OVX1; lysophosphatidic acid; vascular endothelial growth factor (VEGF) B7-H4; and interleukins 6 and 8, to name just a few.[138,139,140] These have been singly studied, in combination with CA 125, or in various other permutations. Most of the study populations are relatively small and comprise highly-selected known ovarian cancer cases and healthy controls of the type evaluated in early biomarker development phases 1 and 2. Results have not been consistently replicated in multiple studies; presently, none are considered ready for widespread clinical application.