NEW YORK (GenomeWeb) – As noninvasive prenatal testing has rapidly grown, some companies have expanded their offerings beyond trisomies 21, 18, and 13 to include sub-chromosomal deletions and duplications.
To date, though, only a couple of small studies have evaluated the sensitivity and specificity of such an approach. This week, researchers from Guangdong Women and Children Hospital in China, Chinese sequencing firm CapitalBio Genomics, and the University of California, San Diego published a study in the Proceedings of the National Academy of Sciences, demonstrating that although NIPT can detect smaller deletions and duplications, the technique may require a higher percentage of fetal DNA or a higher depth of coverage than conventional aneuploidy testing and has a high false positive rate.
The researchers evaluated their whole-genome shotgun sequencing method on 1,476 pregnant women with fetal abnormalities that were detected on an ultrasound who had also undergone invasive testing via array comparative genomic hybridization. The noninvasive test initially detected 56 of 78, or around 72 percent, of the abnormalities that were detected by aCGH. When the team increased sequencing depth and only called abnormalities larger than 1 megabase, they increased their detection rate to 69 out of 73.
Kang Zhang, a senior author of the study and chief of ophthalmic genetics at UCSD School of Medicine, said that although the false positive rate was "surprisingly high" — the study produced 58 false positives in 55 samples — the team is "certainly very excited about the results and the potential."
He added that the study was one of the first "large-scale studies that tried to expand NIPT to detect sub-chromosomal alterations." Guangdong Women and Children Hospital is continuing to conduct clinical trials of the test, he said, noting that the hospital would likely begin offering it as a clinical test by early next year.
Chinese firm CapitalBio Genomics, previously called iGenomics, performed the sequencing for the study using Thermo Fisher Scientific's Ion Proton. In 2013, iGenomics invested in 17 Proton systems for research in aneuploidy screening.
In their study, the researchers first wanted to establish a DNA threshold for diagnosing deletions and duplications. The team isolated DNA from 19 newborns who had been diagnosed by aCGH with a deletion or duplication and created mixtures with normal DNA. Three samples had two alterations, while the others had one each.
They then sequenced each mixture — containing 5 percent, 10 percent, 15 percent, 20 percent, and 30 percent abnormal DNA — to different depths of coverage. Increasing the fraction of abnormal DNA, as well as increasing sequencing depth, resulted in improved detection. At 5 percent abnormal DNA and with 3.5 million sequencing reads, the smallest alteration detected was 20 mb. Increasing sequencing depth to 15 million reads enabled of the researchers to detect a 3-mb alteration.
Previous research by Dennis Lo's group at the Chinese University of Hong Kong showed that size differences could be used to delineate cell-free fetal DNA from maternal DNA in plasma, and the researchers validated this method on a set of samples, so that they could estimate fetal fraction from maternal plasma.
Next, they tested their sequencing-based method for detecting fetal deletions and duplications in a cohort of 1,476 pregnant women who had undergone invasive testing. Fetal DNA was evaluated by aCGH and maternal plasma was evaluated via the NGS method, and the team also estimated the fetal fraction of DNA in the maternal plasma samples.
About 88.5 percent of the samples had an estimated fetal DNA fraction of greater than 10 percent, and fetal DNA fraction increased with gestational age.
Array CGH identified 78 sub-chromosomal deletions and duplications, ranging from 52 kilobases to 84 mb in 57 samples, many of which corresponded to known deletion orduplication syndromes. The NIPT method identified 56 of those 78 abnormalities when 3.5 million reads were generated per sample, with a great range in detection accuracy, depending on the size of the alteration.
The method only detected 14 out of 34, or 41 percent, of alterations less than 5 mb in size and no alterations smaller than 1 mb. By contrast, it detected all 35 alterations that were larger than 10 mb and 42 out of 45, or 93 percent, of alterations between 5 and 35 mb in size. When sequencing depth was increased to 10 million reads, the method could detect 69 out of 73, or about 95 percent, of alterations larger than 1 mb.
The researchers noted that aside from the size of the alteration itself, other factors played a role in the ability to detect them. For example, four samples where NIPT failed to identify an alteration had low fetal DNA fractions. In addition, when alterations were in repetitive regions of the genome, they were harder to detect.
The assay also produced 58 false positive results in 55 samples. Further examination of the maternal DNA, however, found that 35 of the false positives were in fact alterations carried by the mother. The finding has "significant implications for the clinical utility of NIPT to detect sub-chromosomal abnormalities, because a positive test using this assay may require follow-up testing to rule out deletions and duplications in the maternal DNA before the fetus can be diagnosed," the researchers wrote.
For the remaining 20 false positives, when the researchers resequenced the samples, they did not find those events, suggesting that the "false positives seem to be the random result of our sequencing technique and statistical model," the authors wrote.
Zhang said the researchers are now working on strategies to improve the performance of the test. He said one way to improve on the false positive detection would be to also sample maternal leukocytes, so that "you exclude the possibility that whatever abnormality you detect is not from the fetus."
In addition, Zhang said the team is working on methods to enrich fetal DNA in a plasma sample. This would enable women to be tested earlier in their pregnancies, when fetal DNA fraction is lower. In the study, average gestational age was 24 weeks, but NIPT is ideally performed at around 10 weeks of gestation.
"The ultimate goal," he said, is to diagnose not only sub-chromosomal alterations, but "to detect single base pair changes in fetal DNA."