Experts shed light on false positives in prenatal tests

SEATTLE – As growing numbers of pregnant women seek noninvasive blood tests to screen for Down syndrome and other abnormalities, University of Washington researchers say they’ve documented a key reason for worrisome false positive results.

Extra copies of DNA on a mother’s chromosomes may skew some tests, indicating that a fetus has a disorder when the baby is actually healthy, according to UW genetics experts Dr. Hilary Gammill and Dr. Jay Shendure, co-authors of a study in Wednesday’s New England Journal of Medicine.

The problem may affect a “substantial” portion of the fraction of false-positive results for the hundreds of thousands of noninvasive prenatal tests now being performed each year on high-risk women, said Shendure.

The good news is that test manufacturers can account for the extra DNA – and adjust their algorithms. One firm says it already has.

“Once you’re aware of the problem, you’re obligated to address it,” said Shendure, an associate professor in UW’s department of genome sciences.

The review of four UW cases of so-called “discordant” pregnancies showed the extra DNA – called copy-number variations (CNVs) – was the reason two local mothers were told their babies had devastating trisomy 18, also known as Edwards syndrome. At birth the babies were normal.

“It just really causes a huge amount of stress on families,” said Gammill, an assistant professor in the UW obstetrics and gynecology department and an associate at Fred Hutchinson Cancer Research Center. “It’s hard to even put that into words.”

Some parents choose to end pregnancies when chromosomal abnormalities such as Down syndrome, Edwards syndrome or trisomy 13, also known as Patau syndrome, are detected.

Noninvasive prenatal tests (NIPTs) have been commercially available in the U.S. since 2011. They analyze small bits of fetal DNA found floating in a pregnant women’s blood to test for chromosomal abnormalities. But some women have extra DNA on the chromosomes in question, the researchers found, which bumps up the total count, creating a false-positive result.

Other causes of false-positive tests include so-called “vanishing twin syndrome,” in which one of a set of multiple fetuses is miscarried. Undetected tumors and mosaicism, in which cells within the mother carry a different genetic makeup, can also be responsible.

Several large studies have confirmed that these cell-free DNA, or cfDNA, tests have a detection rate of 99 percent for Down syndrome, with a false-positive rate of as low as 0.1 percent. A new study of nearly 16,000 women, also published in NEJM on Wednesday, confirmed the high rate of detection and a false-positive rate of 0.06 percent.

That compares with a detection rate of about 79 percent through standard first-trimester screening, and a false-positive rate of 5.4 percent, the study found.

So far, the noninvasive tests have been recommended only for high-risk women, who are more likely to have positive results. However, experts expect the tests will be expanded in the future to low-risk women in the general population, where it will be harder to predict how often the tests will be accurate.

The worry, Gammill and Shendure say, is that parents will take irreversible actions based on the results - including false positives – without documenting the conditions with confirmed diagnoses.

“That’s where it becomes very challenging,” Gammill said. “Very highly educated patients are using these tests based on what they think are high levels of accuracy.”

The UW paper is the first to document the effect of DNA copies on false positives, say the researchers and Dr. Susan G. Klugman, director of clinical genetics for the American College of Medical Genetics.

The UW scientists said they hope test makers take it into account as the industry continues to grow to an estimated $3.6 billion a year by 2019, according to an analysis by Transparency Market Research.

Ariosa Diagnostics, maker of the Harmony prenatal test, said more than 400,000 have been performed worldwide since 2012. Spokeswoman Margaret Theeuwes said the firm’s research showed that of 65,000 samples of maternal DNA, 35 copy-number variants, or 0.05 percent, were observed. All 35 were accurately diagnosed, the study said.

Representatives for Natera, which makes the Panorama NIPT, said their screen doesn’t count chromosomes at all, but instead uses single-nucleotide polymorphisms, or SNPs, to analyze the ratio of fetal to maternal DNA in the blood. It effectively accounts for the extra copies, said Dr. Susan Gross, Natera’s chief medical officer.

Other top test manufacturers, including Illumina and Sequenom Inc., declined to comment on the UW finding. But Gammill and Shendure said they’re optimistic the industry will work to reduce false-positive results.

“If you know about this and you take it into account, you can solve this,” Shendure said.