In Conversation: Tufts Geneticist Diana Bianchi on Noninvasive Prenatal Testing

November 5, 2012 | It is one year since the commercial debut of the first noninvasive prenatal test for aneuploidy by Sequenom, one of a handful of US companies developing methods to analyze cell-free fetal DNA circulating in the maternal blood. The first laboratory developed tests use massively parallel sequencing to screen for trisomies 21, 18, and 13, with more applications swiftly following suit, including monosomy X  from Verinata Health and targeted sequencing from Ariosa. The next-gen sequencing test performs a digital dosage—millions of reads are aligned and analyzed for the ratio of chromosomes 21 (in the case of Down syndrome) to the other autosomes.   

Diana Bianchi, a renowned expert in noninvasive prenatal diagnosis, is Executive Director of the Mother Infant Research Institute at Tufts Medical Center, a Professor of Pediatrics and Obstetrics and Gynecology at Tufts University School of Medicine in Boston, and also chair of the clinical advisory board at Verinata Health. She spoke to Bio-IT World editor Kevin Davies about the rapid advances in this field and where it goes from here. 

Bio-IT World: Diana, where do we stand currently in noninvasive prenatal diagnosis? 

BIANCHI: Actually, the preferred term is now noninvasive prenatal testing (NIPT), because it is not yet diagnostic. NIPT using massively parallel sequencing (MPS) of maternal plasma DNA is available clinically in the United States and represents the largest clinical application of sequencing currently being done. Three US companies are offering it: Sequenom, Verinata Health and Ariosa. A fourth company, Natera, is scheduled to go live in the near future.

The reason that NIPT is important is that it is currently standard of care for all obstetricians to offer pregnant women screening for Down syndrome (DS). That was a decision made by the American College of Obstetrics and Gynecology (ACOG) in 2007. OB/GYN is one of these specialties in which there are “edicts” from their professional organization. You pretty much have to follow their recommendations, otherwise you’re liable to be sued for malpractice. So, all obstetricians offer their patients the current version of screening for aneuploidy, which includes the measurement of certain proteins in the blood and an ultrasound examination of a fluid-filled space at the back of the fetal neck (the nuchal translucency, pioneered by Dr. Kypros Nicolaides at King’s College in London).

For as long as I’ve been practicing medicine, there’s been great interest in developing a noninvasive prenatal diagnostic test in which a maternal blood sample is used to obtain genetic information about the fetus. My medical school mentor at Stanford, Leonard Herzenberg, one of the people who first developed the flow cytometer, had a son with Down syndrome named Michael (he’s now in his 50s.) When I arrived as a medical student at Stanford and joined their lab, Len said to me, ‘I want you to develop a noninvasive prenatal test for Down syndrome using the flow cytometer’—a tiny little task! It’s amazing to me that it’s finally now a clinically available test, although the methodology and the actual target are different from what we first envisioned.

How did Verinata Health get started? 

Verinata Health started as a company dedicated to microfluidic technology based on intellectual property from MIT and Harvard. The first name of the company was Living Microsystems and it was based in Boston. It later changed its name to Artemis Health. They moved to the Bay Area and in 2011 and subsequently changed their name to Verinata Health. Verinata has the exclusive license to [Stanford professor] Steve Quake’s patents in this space, which specifically discuss the use of massively parallel sequencing to detect fetal trisomy. Verinata is suing Sequenom and Ariosa, claiming that Sequenom infringes upon the Quake patent. (Sequenom has the exclusive license to the Dennis Lo patent.) Meanwhile, Sequenom is suing Ariosa and Natera, and Ariosa and Natera are suing Sequenom. It’s a tangled web!

Has the method been studied in clinical trials? 

The first large scale clinical trial to use MPS to detect trisomy 21 came out of Dennis Lo’s laboratory in Hong Kong, using an international consortium of samples with a high prevalence of trisomy 21. They published their results in 2011, which turned out to be a watershed year in which five large-scale clinical trials that looked at the accuracy, sensitivity, and specificity of detecting aneuploidy in high risk pregnant women were published.

Until last year, laboratory scientists performing most of the trials were not blinded. The Sequenom trial that came out in October 2011 was the first to be a prospective blinded clinical trial. The second was the Verinata one, on which I’m the first author. We presented the results of this study at the annual Society for Maternal Fetal Medicine meeting in February 2012. All of the clinical trials, whether they were performed in Lo’s lab, or at one of the US companies, showed exceptional performance, with a higher detection rate than serum screening and a lower false positive rate.

What struck me immediately upon reading the relevant papers was that the false-positive rate in all of these trials was either zero or less than 0.5 per cent. The current standard of care—serum screening—is calculated to give a 5 percent false-positive rate. So there’s a real practical problem in the clinic. We have lots of pregnant women who are told that they are screen-positive for DS. They are then referred to a prenatal diagnosis center to have an invasive procedure, such as amniocentesis, to make a definitive diagnosis. The patients don’t understand, and sometimes even the OBs don’t understand, what a positive screening result means. They think that the baby has DS. This happens every day! The advantage of DNA sequencing is that the false-positive rate is so low that it’s going to drive down the number of amniocenteses needed to clarify the false-positive rate.

What is the window to do this test in terms of the pregnancy? 

Anytime from eight weeks until term. There’s a lot of good evidence that shows that the fetal DNA levels increase significantly after eight weeks. However, the test is offered clinically after 10 weeks.  Unlike serum screening, it can be offered in the third trimester as well.

How does the technology work? 

The technology works by sequencing cell-free total DNA in the pregnant woman’s plasma.   When you perform the sequencing, you don’t extract the fetal DNA, you’re extracting the total cell-free DNA from the mother, which consists of a mixture of mom’s and baby’s DNA. An incremental increase in the amount of chromosome 21 sequence is found when the fetus has trisomy 21. It’s surprisingly sensitive. The “fetal” DNA is actually not from the fetus but rather from the placenta. In the Verinata clinical trial, we were doing about 12 million reads per sample. Both Verinata and Sequenom are using a whole genome approach, whereas Ariosa performs targeted sequencing of SNPs on chromosomes 18 and 21.  Their number of reads is consequently much lower.

As for the analysis, Sequenom uses a z-score approach, and they modify what’s normal and abnormal pretty much with runs on a daily basis, which is very similar to what’s done with serum screening. Verinata has a set of proprietary normalized chromosome values that they’ve developed. They establish the values and they serve as a permanent reference. They don’t change it on a day-to-day basis, so it’s more standardized.

What could this method be used for besides trisomies? 

The first trimester ultrasound examination detects a common fetal lymphatic malformation known as a cystic hygroma, which has a very high association with aneuploidy (50%). One major reason for cystic hygroma is fetal Turner syndrome due to the presence of monosomy X. A decreased X chromosome signal without Y signal detected suggests the presence of monosomy X. Verinata’s clinical trial was the first to report on some of the fetal sex chromosome abnormalities. We’re most interested in monosomy X (Turner syndrome) because that sex chromosome aneuploidy is associated with many medical complications—for example, there’s a very high spontaneous miscarriage rate.

In the next iteration of NIPT, sub-chromosome deletions and duplications will likely become possible. Reports are starting to surface of being able to detect sub-whole chromosome deletions.

There is also a lot of concern within the DS community that the availability of easier, safer prenatal testing of Down syndrome will lead to more elective terminations of pregnancy for affected fetuses. It’s been very hard to get good numbers on what’s happening in terms of the termination rate after a diagnosis of DS. In a recent issue of Prenatal Diagnosis (of which I’m the editor), we had a great article from Jamie Natoli, a genetic counselor and epidemiologist who provided recent data on termination rates. The often quoted 90 percent termination rate is not an accurate number—it is more like 65 percent. In addition, something that doesn’t get a lot of attention in the DS community is the possibility that if you could noninvasively diagnose DS at ten weeks you could consider the possibility of fetal treatment if you decided to continue the pregnancy.

What sort of therapy are you envisioning? 

We’ve been studying the amniotic fluid transcriptome. It turns out that the cell-free RNA floating in the amniotic fluid is coming from multiple organs, including the fetal brain. And so, in a PNAS paper in 2009, we took a systems biology approach to the second trimester living fetus with Down syndrome, [and asked] what’s wrong compared to the euploid fetus? And there were some profound functional differences, most notably oxidative stress. We hypothesized that if you specifically addressed the transcriptome, and took a personalized medicine approach, you might improve neurocognitive outcome—that’s the endpoint for most DS researchers. There’s a Roche trial going on right now treating adults with DS. My vision would be to convincingly demonstrate that you could treat the pregnant woman with meaningful personalized treatment designed for the fetus, because there’s so much neural development going on in utero. If you wait to treat until birth, it’s not just that you’re losing six months—you’re losing huge biological windows of opportunity to influence neurogenesis.

Keeping in mind you’re an advisor to Verinata, how would you compare the Verinata and the Sequenom tests?  

The technology that both companies use is quite similar, but differs from the targeted approach used by Ariosa. A fourth company, Natera, will be offering the test soon using a SNP-based approach.

How quickly are other OB/GYN departments offering this? 

To give you an idea of the intense interest from pregnant women, at Tufts Medical Center we started offering the Sequenom test in November of 2011. Within a few months, we had pregnant women flying from Europe to Boston to get the test. They were actually getting on a plane and having their blood drawn and paying out of pocket just to be able to have the test.  In US-based academic centers, the test is most commonly being offered to women at high risk for fetal aneuploidy as an alternative to an invasive procedure such as amniocentesis or chorionic villus biopsy (CVS).

Even though you’re chairwoman of the clinical advisory board at Verinata, your medical center is offering the competitor’s platform. Isn’t that a bit curious?  

Because of the potential for financial conflict of interest, I do not directly counsel women about the test. The Tufts Medical Center genetic counselors do. Their choice of test provider is affected by many variables, including insurance coverage. I have worked in the field of prenatal diagnosis for 30 years as a clinician and a scientist. I want pregnant women to get the most accurate and safest test possible. I believe that [the Verinata test] performs better, but at the end of the day, if ordering a test from another company means that women don’t have to have an amniocentesis, that’s a great thing. The risk of a fetal miscarriage from an amniocentesis is about 0.5 percent. Of course I’m biased in favor of Verinata, because I know them well and know the science behind their test. On the other hand, as a journal editor, I read all the papers in this field from the other companies; they are all contributing to advancing knowledge in this area. I’m just really happy that the field is moving along so quickly.

You say this is the most extensive current clinical application of next-gen sequencing? 

Yes, it is. At the recent California Technology Assessment Forum (CTAF) hearing on NIPT (last month), collectively, the US companies stated that 30,000 clinical tests have been performed in the past year.

Daixing Zhou (formerly Life Tech) started a company in China called Berry Genomics. He told me recently that there are discussions underway in China as to whether this should be the primary noninvasive test for fetal aneuploidy. Apparently only 15 percent of pregnant women now get serum screening in mainland China, which surprised me, because you would think that with a one-child policy, everybody would want to get the test to make sure their only child was healthy. Why is the percentage so low? Zhou said it’s because the cytogenetics labs are maxed out in China—the country can only handle 100,000 amnios a year. In the US (prior to the availability of NIPT) we did about 200,000 invasive procedures a year. Over the past year since we’ve been offering NIPT at Tufts, amniocenteses have decreased by 44%.

Are there ethical issues surrounding the implementation of this non-invasive test?  

Yes. To some extent, pregnant women don’t get prenatal diagnosis for two reasons. Either they don’t want to know about the fetal status because they don’t want to be faced with any sort of decision whether or not to continue the pregnancy or they are afraid of “the big needle.” They don’t want to start down a path in which they might eventually be counseled to have an invasive procedure, which involves putting a needle somewhere in their uterus. The promise of all this technology is you won’t need the “big needle.” And that’s what the main ethical issue is. If it is easier to find out about the presence of whole chromosome fetal aneuploidy noninvasively, will more women choose to terminate their pregnancies?

How easy is it for a prenatal clinic to embrace such a rapidly developing new technology like this?  

There are a number of issues, including lack of provider education in genetics, disruption of well-validated current screening protocols, and the economic effects of decreasing numbers of invasive procedures. The rapidity of the technical progress is outpacing the ability of professional groups such as ACOG to make practice recommendations across the US. I think eventually there will be an ACOG recommendation. The concerns include: To whom are we going to offer this test? Are we going to offer it to anybody who calls on the phone and asks for it? Or, are we going to limit it to women who are at high-risk because they’re over age 35 or they’ve had an abnormal screen?

In November 2011, the Tufts genetics and maternal fetal medicine staff members met and developed a testing algorithm. We decided that if a woman asked us for the test and was willing to pay for it, it was not our role to prevent her from getting access to that test. But we would not mention it as an option to women who are at average risk for fetal aneuploidy because all of the clinical trials had been done on high-risk women—women who are having a confirmatory diagnostic test, either CVS or amniocentesis. When pregnant women are already at high-risk for fetal trisomies 13, 18, or 21, we offer the test as an alternative to amniocentesis or CVS. That’s the compromise right now. But the FDA and ACOG will not approve it until a study of average-risk pregnant women is done. Within the past two months, however, two studies have been published that suggest that the test does not perform differently in average risk women. The Blue Cross Blue Shield (BCBS) Technology Evaluation Center recently reviewed the evidence and said that the test is effective as an advanced screen for trisomy 21 in high and average risk pregnant women.

What is next the next step in rolling this out? 

I think the big issue is education of clinicians and insurers. It’s certainly safer and performs better than the current standards of care for screening. If you remove large number of amniocenteses that are being performed and look at it system wide, does sequencing reduce or increase costs? As the sequencing costs go down, it will be very cost effective. I am encouraged that the CTAF and BCBS groups have both favorably reviewed the technology. A number of insurers, including Tufts Health Plan, are already including this test as a covered benefit. Whenever I’m at a medical meeting now obstetricians come up to me and say, “You’re going to put us out of business.” Most pregnant women don’t need to have an invasive procedure with an associated risk. There will still, however, be a need for invasive procedures for more [severe] fetal indications, such as an ultrasound abnormality.

Could you in principle screen for single genes? 

Dennis Lo had a report in Science Translation Medicine in late 2010. It was a proof of principle study that could only be done in Hong Kong where there are significant resources. One case of a fetus at risk for beta thalassemia cost $200,000 to diagnose. The pregnant woman had a CVS and it was known that the fetus carried the paternal mutation. The fetus was not affected, it had a different mutation than the mother, but they wanted to see, in theory, could you detect the paternal mutation in the mother’s blood? They did, but they needed to know the fetal genome from the diagnostic procedure (CVS) and they needed to have both parents’ DNA samples available, so it’s not a screening test.

More recently, in June 2012, two reports appeared in Nature and Science Translation Medicine from the Quake and Jay Shendure labs, respectively. Both showed that it was possible to noninvasively infer the entire sequence of the fetal genome from a maternal blood sample. So, perhaps screening for single gene disorders will be incorporated into whole genome sequencing.  That’s a bit off into the future due to intensive requirements for genetic counseling regarding unanticipated findings in whole genome sequencing. In the immediate future, however, I see expanded applications of whole chromosome aneuploidy, including the sex chromosomes, followed by detection of sub chromosome deletions and duplications. But the field is moving incredibly fast from bench to bedside. It’s a very exciting time!