Gil Mor of the Yale School of Medicine and his colleagues based the novel screening tool on four proteins: leptin, prolactin, osteopontin and insulinlike growth factor II. Individually, each had been identified as a potential cancer biomarker, but did not provide enough information alone to determine whether or not a person was suffering from the illness. When the researchers used the new approach blindly to test a group of 106 healthy women and 100 women with ovarian cancer, they found that the test correctly diagnosed cancer victims 95 percent of the time.
The team reports that the test must be further refined before it will be applied to the general population, because screening strategies for early detection should have specificities greater than 99.6 percent in order to avoid false positive diagnoses. The authors suggest that including additional proteins that have been linked to ovarian cancer could enhance their test's sensitivity. The results were published online this week by the Proceedings of the National Academy of Sciences.