SAN FRANCISCO (GenomeWeb) – Rare trisomies that aren't addressed by current noninvasive prenatal DNA tests may be the cause of some failed tests or abnormal and false-positive results, according to a new study. Expanding NIPT to analyze all 24 chromosomes could potentially improve the accuracy of the tests and also identify causes of miscarriage, researchers from the National Institutes of Health, Illumina, and the University of Melbourne's Victorian Clinical Genetics Services reported.
Most NIPTs look for aneuploidies in chromosomes 21, 18, 13, and the sex chromosomes. Companies often employ whole-genome shotgun sequencing and bioinformatics that compare counts of those chromosomes against a set of reference chromosomes in the same sample. But if those reference chromosomes are not in fact normal, results can be skewed, according to a study published today in Science Translational Medicine. In addition, identifying rare autosomal trisomies can also yield clinical information, with evidence suggesting that they may cause miscarriage and late fetal demise, according to Diana Bianchi, senior author of the study and director of the National Institute of Child Health and Human Development.
Bianchi has long sought to understand the causes of false-positive results in NIPT. In the new study, she and her colleagues examined samples from nearly 90,000 women who had undergone NIPT during pregnancy, either through Illumina's CLIA lab or the Victorian Clinical Genetics Services (VCGS), which licenses Illumina's technology.
All cases had received Illumina's Verifi test, which uses a whole-genome shotgun sequencing approach, and which reports aneuploidies found in chromosomes 21, 18, 13, and the sex chromosomes. It also uses a quality control metric called normalized chromosome denominator quality (NCDQ), which flags abnormal results. The researchers looked specifically at these abnormal results, which are typically reported as failed tests, and analyzed the sequence data from all 24 chromosomes.
The NCDQ was "created to identify some sort of genome-wide imbalance that's throwing off interpretation," Bianchi said. She had received a sponsored research grant from Illumina that was administered by Tufts University Medical Center, prior to her joining the NIH, under which she conducted this study. "I've been interested in why there are test failures and false positives," she said, so the research group "designed the study to essentially open up the whole-genome sequence data and look across all the chromosomes."
Of the nearly 90,000 samples, almost 73,000 were run at Illumina's CLIA lab and just under 17,000 at VCGS. Of those, 518 and 109 samples, respectively, had received NCDQ values that had been deemed borderline, warning, or abnormal. When the researchers went back and looked at the data, they found chromosomal abnormalities in 328 and 71 of those cases, respectively. Most of the remaining cases had been flagged due to technical noise.
In both cohorts, the researchers found that trisomies in chromosomes other than those tested for were the most frequent cause of a flagged test result, accounting for 49 percent of the cases. Other causes were copy number variants and other genome-wide imbalances, such as segmental duplications and deletions.
Interestingly, Bianchi said, even though each cohort was tested separately, for both, "the percentage of samples flagged and the explanations for the sample failures were very much the same." For both cohorts, not only were rare trisomies the most common cause of test failure, but also, the "chromosomes most commonly affected were the same," she said. And, looking back at the literature from invasive testing, "the proportions are the same there, so we think there's something real happening," Bianchi said.
For instance, the most frequent rare trisomy was trisomy 7, Bianchi said, followed by trisomies in chromosomes 15, 16, and 22. These trisomies also impacted the clinical test results. In the Illumina cohort, for instance, of 246 samples that had a trisomy in a non-test chromosome, 172 involved chromosomes that were used as reference chromosomes, meaning the algorithms were basing their counts of the test chromosomes on them. Because of that, 21 false-positive monosomy calls were reported, the researchers found.
Bianchi added that a key component to the study is that for many of the samples tested by the VCGS group, there was also pregnancy outcome data, which enabled the researchers to see the clinical impacts of the rare trisomies. Of the 71 abnormal samples from that cohort, 57 had outcome data.
Of 52 samples with a rare autosomal trisomy, 22 were associated with miscarriage. Notably, 13 out of 14 samples with trisomy 15 had a reported miscarriage. There were three cases of reported second-trimester intrauterine fetal demise — one with trisomy 2 and two with trisomy 16. Also, there were 17 pregnancies that appeared to have rare autosomal trisomies and went on to have amniocentesis. Of those, seven had normal results and all but one had normal births. In three cases, placental biopsies were available, and testing identified confined placental mosaicism for the rare trisomies that NIPT had identified. Other reported conditions included uniparental disomy, intrauterine growth restriction, true fetal mosaicism, and co-twin demise. In total, abnormal results were seen in 39, or 75 percent, of the 52 cases where noninvasive prenatal testing identified a rare autosomal trisomy.
"There's a lot of discussion in the NIPT world about how much information [labs] should release," Bianchi said. This study provides evidence that at least in regards to whole-chromosome information, which for NIPTs that use a whole-genome shotgun sequencing approach is already acquired, the data can help explain false-positive tests and has clinical significance, Bianchi said.
Illumina already offers a version of its Verifi test that reports aneuploidies in all 24 chromosomes, called Verifi Plus. Both tests use the same methodology described in the study, although Nargol Faravashi, Illumina's senior manager for reproductive and genetic health, said in an email that the firm has "optimized the algorithm to now screen for additional autosomal trisomies," emphasizing that it focuses on whole-chromosomal trisomies. "By unmasking the additional chromosomes, we will provide more information to physicians/patients to enhance care," Faravashi wrote.
Bianchi said that one example of how additional information could impact clinical care would be in the case of an NIPT report identifying trisomy 15, as opposed to a failed or abnormal test result. Knowing that 13 out of 14 pregnancies with that result miscarried, "that would change my counseling," she said.
Bianchi cited another case — a woman who lost her pregnancy at 40 weeks gestation. "That's a really tragic loss, to lose the pregnancy at full term and not have an explanation," Bianchi said. In the study, the researchers identified trisomy in chromosome 4 in this case. "That may have been helpful information to allow the family to heal, and also for genetic counseling," she said.
Bianchi said her next step is to continue to study NIPT and the reasons for false positives, including cases where an NIPT has revealed a previously unknown maternal malignancy. In addition, she said she would continue to follow up on the clinical impact of rare autosomal trisomies.
"This study specifically raises the question as to how many cases of fetal growth restriction are caused by rare autosomal trisomies," she said. In addition, she is interested in looking deeper at the clinical impact of mosaicism to see whether it results in learning disabilities or newborn malformations.