Sequencing-based noninvasive prenatal tests that screen for aneuploidies such as Down syndrome and Edwards syndrome are rapidly becoming part of clinical care, especially for women at elevated risk for fetal aneuploidies.
While these tests in general have very high sensitivities and specificities, one factor that physicians and providers of the test are finding to be important for accuracy is the amount of fetal DNA present in the maternal plasma, with false negatives often occurring in samples with low amounts of fetal DNA.
"It goes without saying that because we're looking at fetal aneuploidy, the amount of fetal DNA is going to impact the ability to perform testing," Ken Song, CEO at Ariosa Diagnostics, told Clinical Sequencing News.
Experts agree that fetal fraction is important, however there are differing viewpoints and practices about how much fetal DNA is required for testing as well as whether fetal fraction should be measured and reported back to the ordering physician.
The four major US companies offering noninvasive prenatal tests differ in terms of whether they report fetal fraction, how fetal fraction is considered in their algorithms and test results, and whether they draw a cut-off below which they will not report results.
Sequenom measures fetal fraction for each sample, but does not report it back to the ordering physician, and if it finds that fetal fraction is below 4 percent it will not report the result.
Ariosa measures and reports back fetal fraction, and incorporates fetal fraction into its risk prediction score. It also does not report back results with lower than 4 percent fetal fraction. Natera measures and reports back fetal fraction. When paternal DNA is available, Natera will report results back with fetal fractions as low as 3.8 percent, but without paternal DNA has a cutoff of 4 percent fetal fraction. For samples with fetal fractions below 3.8 percent, Natera still reports the fetal fraction, but says that a call of aneuploidy or not cannot be reported.
Illumina-owned Verinata Health is the only US company that does not measure or report back fetal fraction. Instead, it has attempted to overcome the issues of low fetal fraction by deeper sequencing, generating between 23 million and 25 million reads per sample.
Speaking in a Natera-sponsored webinar on the issue, Brynn Levy, co-director of the division of personalized genomic medicine at Columbia University Medical Center, said that while the average fetal fraction is between 10 percent and 15 percent — high enough for NIPT to be accurate — there is a "huge standard deviation" among women, and fetal fraction can be as low as 4 percent or as high as 30 percent.
Looking through published literature of the different tests, the vast majority of false positives and false negatives have been in cases where fetal fraction is below 8 percent, he said.
For instance, he said, at 13 weeks gestation, around 20 percent of pregnancies have fetal DNA fraction less than 10 percent. And, at a fetal fraction of less than 8 percent, studies have reported a drop in sensitivity for trisomy 21 detection to 75 percent.
A study published in Prenatal Diagnosis that looked at the impact of fetal fraction on Sequenom's MaterniT21 performance for trisomy 21 detection, found that for the 212 cases with Down syndrome there were four false negatives, all of which had fetal fractions between 4 percent and 7 percent.
Similarly, researchers from Ariosa Diagnostics found that less fetal DNA tends to cause a lower z-score, the statistic used to determine whether a fetus is aneuploid, and as such developed an algorithm that takes into account fetal fraction to give an individualized risk score of having trisomy, as opposed to simply calling trisomy or not.
In a study published in the American Journal of Obstetrics and Gynecology, company researchers found that while the test calls aneuploidy with more certainty at higher fetal fractions, the algorithm was still able to give risk scores of 99 percent or greater for fetal fractions down to 5 percent.
Verinata Health recently published a study in Clinical Chemistry, reanalyzing samples from their clinical validation study to determine fetal fraction. The company researchers found that its test should have a theoretical cutoff at fetal fractions between 1.4 percent and 3.7 percent, depending on the chromosome being analyzed.
Richard Rava, Verinata's vice president of research and development, told CSN that the company has been improving its algorithms so that in combination with deeper sequencing, it can reliably detect aneuploidies down to 2 percent fetal fraction, and the firm is planning to publish a study demonstrating this, he said.
He added that Verinata does not measure and report fetal fraction because the methods for measuring fetal fraction are "imprecise." Rava said that the company has found that the standard deviation for samples with a reported fetal fraction of 4 percent is plus or minus 3 percent.
"Measurements that are made at very low fetal fractions are imprecise," he said. Instead, he said that Verinata has "worked to lower the threshold to ensure that there wouldn't be many samples that would fall below that threshold," adding that few samples contain less than 2 percent fetal DNA.
Nonetheless, many in the field think that fetal fraction should be measured and reported.
Howard Cuckle, an adjunct professor of obstetrics and gynecology at Columbia University Medical Center and a former president of the International Society of Prenatal Diagnostics, told CSN that delivering a report back to a physician with a simple positive or negative result can be misleading because "that's what you'd expect to see on a diagnostic test," and the current noninvasive prenatal tests are "just very good screening tests."
However, fetal fraction is not the only important metric, he added, because even at lower fetal fractions, the tests still have some discriminatory power, often depending on the level of coverage.
"What we need as an end user is a report that gives the fetal fraction and ideally gives the number of reads," Cuckle said.
Factors impacting fetal fraction
Lower amounts of fetal DNA can be caused by a number of factors including maternal weight and mosaicism. Researchers from Ariosa found in a study published in Prenatal Diagnosis that as maternal weight increases, fetal fraction DNA in maternal plasma decreases.
Levy said that there are two possible scenarios – a "dilution of relatively consistent amounts of fetal cell free DNA into a larger maternal volume or an increase in concentration of maternally derived cell-free DNA as maternal weight increases."
In the Ariosa study, the firm found that of more than 22,000 maternal plasma samples drawn between July and December of 2012, about 1.9 percent of the tests resulted in requesting a redraw because they did not contain sufficient amounts of fetal DNA, which for Ariosa's test is below 4 percent.
According to Levy, the average fetal fraction for a woman that weighs 45 kilograms (99 pounds) is 18 percent, but that drops to 4 percent for women weighing 150 kilograms (330 pounds).
However, Verinata's Clinical Chemistry study found "weak but significant" association between body mass index and fetal fraction. Instead, it found that the biology itself plays a greater role.
For instance, trisomy 21 cases had an average fetal fraction of 13.5 percent, or 9.3 percent higher than the proportion of fetal DNA in euploid males.
By contrast, average fetal fractions for trisomy 18 and 13 cases were 8.9 percent and 9.0 percent, respectively. These fetal fractions were 29.7 percent and 28.3 percent lower, respectively, than the average fetal fraction of euploid males.
Mosaicism may also play a role in fetal fraction, Levy said. In mosaicism, "the contribution from the abnormal cell line is minimized by the normal cell line," which "effectively lowers the fetal fraction on cells of interest."
Moving forward, Cuckle thinks that most companies will move toward measuring and reporting fetal fraction.
He would like to see professional organizations such as the American College of Obstetrics and Gynecology issue recommendations regarding fetal fraction in order to influence company practices and result in a more standardized approach.
"The pace of development has been extremely rapid. It's all just jumped ahead before regulations could be discussed or put into place," he said.
"Clinicians should be aware of the importance of fetal fraction," Cuckle added. Right now, the "level of education of the patients and the clinicians is very low in this area. They don't know how to criticize the results they are getting."