How to catch ovarian cancer
when it’s curable

Researchers use proteomics to develop diagnostic tool from liquid biopsy

Published May 01 2019

Star-shaped tumor A section of an ovary stained with hematoxylin and eosin, showing a roughly star-shaped tumor. This tumor can be identified as aggressive because of its micropapillae, or outgrowths.Virginia Commonwealth University Fewer than half of ovarian cancer patients survive for five years after their diagnosis. According to the American Cancer Society, this is because only about one-fifth of ovarian cancer cases are detected early, when the chances of successful treatment and recovery are highest.

“If we could change this reality by detecting (ovarian cancer) at a curable stage, we could save many lives,” said Keren Levanon, a physician-researcher at Chaim Sheba Medical Center in Israel.

In the journal Molecular & Cellular Proteomics, researchers led by Levanon and Tamar Geiger of Tel Aviv University report a new test for ovarian cancer that outperforms previous tests. They hope it can be used to screen women who are genetically predisposed to the disease.

The researchers searched for signatures of cancer in uterine fluid sampled by liquid biopsy during surgery. They compared samples from women with ovarian cancer who had surgical treatment and from volunteers who had gynecological surgery for reasons unrelated to cancer, such as uterine fibroids or benign ovarian cysts.

Bodily fluids contain many proteins. Strong signals from the most common proteins can mask signals from smaller amounts of cancer-linked proteins that also might be present. To overcome that difficulty, researchers isolated microvesicles from the uterine fluid. Because microvesicles are shed from cells, they contain almost none of the signal-masking blood plasma proteins. Instead, they contain protein cargo that may vary between normal and malignant tissues.

Using proteomics, the researchers compared thousands of proteins in uterine microvesicles from 12 healthy volunteers and 12 cancer patients. Then they used machine-learning algorithms to search for patterns of protein abundance that differed among the samples.

“We developed a diagnostic set of nine proteins that distinguishes women with ovarian cancer from healthy women with greater sensitivity and specificity than reported before,” Levanon said.

The researchers then tested the set’s accuracy in 152 women, 37 of whom were known to have ovarian cancer. The test had 70 percent diagnostic sensitivity, meaning that it correctly detected cancer in 25 of the 37 study participants who truly had cancer, including all early-stage cases; it also had 76 percent specificity, meaning that it correctly identified about three out of every four healthy volunteers as healthy. It outperformed previous proteomics-based tests, which had less than 60 percent sensitivity.

Though the study used fluids collected during surgery, the proteomic test can also be run on samples collected in a minimally invasive procedure similar to intrauterine device insertion. The authors propose that it may help young women whose risk of developing ovarian cancer is known to be high in deciding whether to have their ovaries removed as a preventative measure. The method of isolating microvesicles from bodily fluids to detect faint cancer signals also may have promise for other difficult-to-detect types of cancer.

Related stories

Breaking dogma?
Learn more about microvesicles and their role in cell-to-cell signaling.

Bridging the gap between research and treatment
In this personal essay, Michelle Gumz writes about the importance of personalized medicine for her own cancer treatment and that of her relatives.

Also by this author:

Real-time proteomics may one day speed up cancer surgery

Keeping tabs on thousands of protein variants

A royal legacy

Laurel Oldach Laurel Oldach is a science writer for the ASBMB. Follow her on Twitter.