NEW YORK – Myriad Genetics has implemented a new fetal DNA enrichment method for its Prequel noninvasive prenatal screening test that it hopes will further increase uptake of the test among clinicians and patients.
The improvement adds to various strategies providers of NIPT have used to enhance their offerings over the years, both on the technology side and regarding the data analysis.
In a study involving almost 1,300 blood samples from pregnant women, published in Genetics in Medicine on Sunday following an earlier release of the results in a MedRxiv preprint in July, Myriad researchers demonstrated that their method, called fetal fraction amplification (FFA), could increase the percentage of fetal cell-free DNA in almost all samples, allowing the test to provide accurate results for a variety of chromosomal abnormalities.
Most circulating DNA in maternal blood stems from the pregnant woman, while the proportion of fetal-derived DNA, or fetal fraction, can vary from well under 2 percent to more than 10 percent. In general, fetal fraction goes up over the course of pregnancy but is often low in women with an elevated body mass index.
Importantly, low fetal fraction has been associated with less reliable test results. "At a low fetal fraction, many of the labs have difficulty reporting accurate results," said Jim Goldberg, chief medical officer of Myriad Women's Health. As a result, professional guidelines recommend that test providers state the fetal fraction in their reports.
While some labs are getting reliable results from samples with low fetal fraction by tweaking their analysis methods — including Myriad itself, which showed in a study published in Prenatal Diagnosis last year that it can produce accurate results even from samples from women with high BMI and low fetal fraction — others have set a strict limit below which they don't report a finding.
According to Goldberg, commercial NIPT providers often use 4 percent fetal fraction as their cutoff. Women falling below that limit might need to repeat the test later in pregnancy, which is not ideal, and those with high BMI are sometimes discouraged from having NIPT altogether because of their high risk of receiving a no-call result, he said.
The issue of low fetal fraction is also not rare. Overall, 3.7 percent of the samples the Myriad researchers tested in their study had an initial fetal fraction below 4 percent, and among those with high BMI, 16 percent had a low fetal fraction.
To improve its assay, the Myriad team developed a technique called 'Amplify' that enriches fetal DNA in maternal blood samples. In their study, they showed that the method increased fetal fraction about 2.3-fold across samples and 3.9-fold in samples with low fetal fraction.
Having a way to increase the fetal fraction, Goldberg said, will likely improve clinicians' confidence in the results of the Prequel test. "We felt that there was a general feeling that people were not comfortable with results below 4 percent sometimes," he said, and the new method can provide "confidence to physicians and their patients that they are getting a highly accurate test."
In addition, he said, a higher fetal fraction enhances the sensitivity and specificity for calling microdeletions. In their study, the researchers showed an overall sensitivity of more than 97 percent for five common microdeletions, with 99.8 percent specificity. "We think with this enhancement, we will be able to provide new data that will help move [microdeletion reporting] forward," he said.
The enrichment process, which is based on earlier observations by others that fetal DNA fragments tend to be shorter than maternal ones, involves agarose gel electrophoresis of Illumina next-gen sequencing libraries generated from cell-free plasma DNA to separate the DNA by size. Only short DNA fragments with an average length of 140 nucleotides are eluted from the gel and used for sequencing.
Goldberg declined to provide details of how Myriad has automated the gel-based size selection process, which he said is proprietary. He noted that "it took a lot of work to develop and validate it."
He acknowledged that his group is not the first to enrich fetal DNA based on size, "but the thing that we have developed is a technique that we can apply at scale," he said. "No one that I know of has been able to automate that to scale like we've done."
"I think that overall such size fractionation would be a useful addition to the tools that we have for NIPT," said Dennis Lo, a professor of medicine and chemical pathology at the Chinese University of Hong Kong, in an email. In 2004, his group was one of the first to demonstrate that fetal-derived DNA molecules tend to be shorter than maternal ones.
"One question that is not clear is whether we should use such a size fractionation strategy for all cases, or just for those with low fetal DNA fractions or for pregnant women with prior conditions that would increase the chance of a low fetal DNA fraction (e.g. increased BMI)," he said.
Lo also suggested that reporting the initial fetal fraction is still a good idea. "I understand that the Myriad authors are of the opinion that such size enrichment strategy would obviate the need to measure the fetal DNA fraction," he said. "I am generally more cautious and would still think that the measurement of fetal DNA fraction would be a good quality control."
Myriad said it still includes fetal fraction in its reports, though it only reports the enriched fetal fraction, not the original one.
Rossa Chiu, also a professor at of chemical pathology at CUHK, cautioned that DNA size selection may potentially reduce the DNA yield, especially when the starting plasma volume was small.
Her own group has also observed increased variance in chromosome copy analysis as a result of DNA size selection. "Such worsening of the variance may or may not be salvageable by increasing the sequencing count. It all depends on the sample preparation and library construction protocols used before sequencing," she said in an email. "In short, just adding a size-based selection step may not produce favorable outcomes for some laboratories. Other components of the system may need to be retuned as well."
Mark Pertile, head of the divisions of reproductive genetics and biochemical genetics at Murdoch Children's Research Institute in Australia, and head of its NIPT laboratory, said that the Myriad researchers' "logic of enriching for fetal fraction using size selection is sound."
In an email, he noted that while fetal fraction "generates significant debate amongst NIPS providers and healthcare professionals," it is not the only quality parameter that determines the accuracy of test results. For example, in shotgun sequencing-based assays, no-call results may be reduced by increasing the sequencing depth. "This allows some NIPS laboratories to use a lower minimum fetal fraction cutoff," he said.
In addition, analysis methods play an important role. For example, his own lab, which runs the Illumina VeriSeq v2 NIPS assay, and others have used a dynamic algorithm that takes into account fetal fraction, read depth, and other quality metrics to ensure high sensitivity for trisomies at fetal fractions below 4 percent. "We have recorded very few false-negative results that can be attributed to low fetal fraction," he said.
Also, the Illumina assay calculates cell-free DNA fragment size based on sequence read data, he said, and considers this information in the analysis to predict which fragments were fetal or maternal in origin. "Bioinformatics improvements to that assay have led to some significant gains in test accuracy by incorporating cfDNA fragment size information," he said.
Myriad's method also has the potential to improve sensitivity for detecting duplications and deletions, he said, "but these results are sometimes confounded by placental mosaicism." To find out more about this, he added, "prospectively collected data will be important."
He also cautioned that maternal autoimmune and inflammatory conditions may lead to unusual cfDNA sizing patterns that are associated with low fetal fractions, "and it would be interesting to see how those referrals respond to this method."
In the meantime, Myriad has implemented fetal DNA enrichment across the board to it workflow. In late July, the company added its fetal fraction amplification method routinely for the Prequel noninvasive prenatal test, which Goldberg said did not change the test's turnaround time — about a week — or price.
So far, the method has worked extremely well, he said: among the first 10,000 samples run with fetal DNA enrichment, not a single fetal fraction remained below 4 percent. Going forward, the company predicts that more than 99 percent of samples will have an increase in fetal fraction.
Overall, Myriad's new method is adding to other improvements in the field. "The NIPS cfDNA screening world is moving rapidly and it's difficult for clinical users to keep up with these technical developments," Pertile said. "The more progressive NIPS providers are continually improving their assays' capabilities. It's an exciting space to be working in."