Cell-free DNA testing: Preferred for detection of fetal aneuploidy

Dr. Lockwood, Editor in Chief, is Senior Vice President, USF Health, and Dean, Morsani College of Medicine, University of South Florida, Tampa. He can be reached at DrLockwood@advanstar.com.

I am not sure of how many genetic amniocenteses I have performed. My best guess is 5000 to 6000. I estimate that I’ve performed at least 1500 chorionic villus samplings (CVS). But like experience with mid-forceps and vaginal breech deliveries, broad experience with invasive prenatal diagnosis may soon be a thing of the past.

A recent article suggests that the number of genetic amniocenteses and CVS procedures performed in this country may soon plummet. In The New England Journal of Medicine, Bianchi and colleagues reported that assessment of maternal plasma cell-free DNA (cfDNA) in low-risk women allowed for very high detection rates (ie, 100%) of fetal trisomy 21 and trisomy 18 with very low false-positive rates (0.3% and 0.4%, respectively).¹ This study opens the door to replacement of relatively inaccurate maternal serum analyte and ultrasound-based fetal aneuploidy screening by cfDNA testing, although high-risk test results using cfDNA will still require confirmation by CVS or amniocentesis.

What is cell-free DNA and how is it detected?

Fetal cfDNA is released into the maternal circulation primarily from apoptotic placental cells, whereas maternal cfDNA is principally derived from hematopoietic cells. Fetal cfDNA tends to be found at lower levels and in shorter segments than maternal cfDNA. After 9 weeks’ gestation, the fetal fraction represents between 10% and 20% of total cfDNA in the maternal circulation and the fetal fraction increases throughout gestation, especially after 21 weeks.²

There is no practical way to separate fetal from maternal DNA but by randomly amplifying all available cfDNA in a maternal plasma sample with polymerase chain reaction (PCR) and then utilizing a “next-generation” DNA sequencing method called massive parallel sequencing (MPS), the chromosomal origin of each DNA fragment can be obtained by comparing its sequence with known chromosome-specific human genome sequences. This MPS approach allows comparisons of the relative amount of each chromosome’s DNA. Thus, for fetal trisomies 21, 18, and 13, the quantity of the tripled chromosome’s DNA will be increased compared to disomic reference chromosomes. Although the absolute increase in a given trisomic chromosome’s DNA will be proportional to the fraction of fetal versus maternal cfDNA, relatively simple mathematical modeling can be used to determine thresholds strongly associated with aneuploidy.

There are limitations to this approach. Because lower fetal fractions can affect fetal trisomy detection rates, sampling before 9 weeks’ gestation will be less accurate. The fetal fraction of cfDNA also decreases with increasing maternal weight, so obesity may also lead to higher rates of inconclusive results.² Besides the impact of the fetal fraction of cfDNA, analysis of relative trisomic chromosomal DNA content can be influenced by amplification efficiency which is, in turn, dependent upon the relative amount of guanosine-cytosine (GC) base pairs in a given chromosome.³ Because the 21st and 18th chromosomes have abundant GC content, trisomies 21 and 18 have the highest detection rates. Alternatively, chromosomes 13 and X have lower GC content and detection of trisomy 13 and monosomy X is less efficient.³ However, new bio-mathematical approaches help adjust for variation in GC content and thus improve screening efficacy.

A second cfDNA aneuploidy testing strategy utilizes sequencing of selectively amplified genomic regions of target chromosomes and then quantifies those specific sequences to determine if there is trisomy-related over-expression. This approach reduces the number of sequences to read, improving efficiency and reducing cost. However, detection rates still vary among chromosomes, with the highest sensitivity found for trisomy 21 and the lowest for trisomy 13.³ Variations of this targeted approach are utilized by several commercial providers to optimize diagnostic efficacy.

A third basic strategy for cfDNA screening again uses targeted amplification of chromosome-specific polymorphic loci, but then integrates maternal and paternal (if available) allelic data and employs Bayesian analysis to evaluate possible inheritance patterns and crossover locations in order to identify fetal chromosomal copy number.³ Because this approach does not require a reference chromosome, accuracy across all chromosomes is similar, reducing false-positive results and increasing detection rates for autosomal trisomies as well as abnormalities in sex chromosome number (eg, Turner’s and Klinefelter syndrome) and triploid pregnancies.

Eventually, these techniques will soon be used to identify common micro-deletions and eventually insertions and even single gene mutations. Indeed, the entire fetal genome may eventually be regularly sequenced and reported with these methods.

Studies of cfDNA for fetal aneuploidy screening in high-risk populations

Different companies utilize variations on these three cfDNA testing methods but all report relatively high sensitivities and low false-positive results in high-risk populations. In 2011, the companies began conducting validation studies of cfDNA testing for fetal aneuploidy detection. These studies utilized case-control formats with known abnormal samples or prospective cohort studies in high-risk populations. Meta-analysis of the results confirmed the efficacy of this general approach.⁴ Thus, for singleton, high-risk pregnancies, reported weighted pooled detection rates and false-positive rates for trisomy 21 were 99.0% (95% CI, 98.2–99.6) and 0.08% (95% CI, 0.03–0.14), respectively. For trisomy 18, these rates were 96.8% (95% CI, 94.5–98.4) and 0.15% (95% CI, 0.08–0.25), respectively, while for trisomy 13 they were 92.1% (95% CI, 85.9–96.7) and 0.20% (95% CI, 0.04–0.46).

Turner syndrome was detected with a sensitivity of 88.6% (95% CI, 83.0–93.1) and a false-positive rate of 0.12% (95% CI, 0.05–0.24) whereas other sex chromosome abnormalities had a detection rate of 93.8% (95% CI, 85.9–98.7) with a false-positive rate of 0.12% (95% CI, 0.02–0.28). These favorable findings led the American College of Obstetricians and Gynecologists (ACOG) to recommend cfDNA testing in high-risk populations in 2012.⁵ Moreover, some practices began using cfDNA for screening in low-risk populations even without available representative trials.⁶

Performance of cfDNA aneuploidy screening in low-risk populations

While studies in high-risk populations were reassuring, it was unclear how the test would perform in a general population where its positive predictive value could be assessed. Bianchi and associates have now completed such a study in the United States. The authors conducted a cohort comparison of standard aneuploidy screening versus cfDNA testing at 21 medical centers between July 2012 and January 2013.¹ All cfDNA analyses were performed on stored samples obtained from the late first to third trimesters at a single reference lab. Standard screening consisted of either: 1) first-trimester serum testing for pregnancy-associated plasma protein-A and b-human chorionic gonadotropin (b-hCG) plus sonographic nuchal translucency measurement; 2) second-trimester “quadruplet” testing with maternal serum alpha-fetoprotein, b-hCG, unconjugated estriol and inhibin-A; 3) fully integrated screening (ie, both first- and second-trimester testing, assaying for b-hCG only once, followed by an integrated risk determination); 4) serum integrated testing (ie, just first- and second-trimester analyte assays without nuchal translucency measurements); and 5) sequential testing (ie, reporting results of first- and second-trimester testing separately and sequentially).

A total of 2052 women were recruited and 1914 were available for study after loss of follow-up or lack of standard or cfDNA testing. The authors utilized meticulous follow-up of infants and fetal losses with pediatric examinations, karyotyping, or autopsy studies as appropriate. The mean maternal age of the population was 29.6 years.

For detection of trisomy 21, the authors noted 100% detection rates with both screening methods but substantially and significantly lower false-positive rates with cfDNA testing compared with standard testing (0.3% vs 3.6%). Positive predictive values were substantially and significantly higher with cfDNA testing (45.5% vs 4.2%). For trisomy 18, detection rates were again 100% for both screening methods, however, false-positive rates were again significantly lower for cfDNA compared with standard testing (0.2% vs 0.6%) and the positive predictive value was substantially and significantly higher with cfDNA testing (40.0% vs 8.3%). Testing for trisomy 13 was available in a limited number of patients and only 1 affected fetus was identified. The authors also reported that the fetal fraction of cfDNA increased from around 11.5% in the first and second trimesters to 24.6% in the third trimester.

Although the overall false-positive rate of 0.5% for trisomies 21 and 18 means that some “unnecessary” genetic amniocenteses or CVSs will still need to be performed, cfDNA would reduce the requirement for such testing by 89% compared with standard screening. Moreover, the authors report that a substantial proportion of their relatively few false-positive cfDNA samples may have resulted from confined placental mosaicism or maternal copy-number variants. Because of concern for confined placental mosaicism in “false-positive” cfDNA samples, Dr. Ronald Librizzi, Chief of Maternal Fetal Medicine for the Virtua Health System in New Jersey, recommends genetic amniocentesis rather than CVS following a positive cfDNA result (written communication, April 2014). 

The Bianchi et al study has a number of limitations including the small number of affected fetuses (ie, 5 trisomy 21, 2 trisomy 18, and 1 trisomy 13), the exclusion of two affected fetuses due to “uninformative results,” the fact that all assays were performed at the manufacturer of the MPS Illumina device reference lab, the excess of third-trimester cfDNA samples analyzed (about a third), and the lack of performance data on morbidly obese women. However, given the multitude of prior studies in high-risk patients confirming the robust detection rates accruing cfDNA testing coupled with the present study’s exceptionally low false-positive rate, Bianchi and colleagues make a convincing case that cfDNA testing should be the preferred strategy for fetal aneuploidy testing when couples are interested in such testing. Cost is a concern but the lower rate of pregnancy loss associated with the far fewer numbers of invasive prenatal tests needed with cfDNA testing should be the primary consideration.

Moreover, additional support for the use of cfDNA in low-risk populations comes from a Chinese study in 1741 women younger than age 35 that compared cfDNA testing to maternal serum second-trimester “triple” screening and found that the former achieved substantially higher sensitivity (100% vs 54.5%), specificity (99.9% vs 85.9%) and positive predictive values (91.7% vs 2.4%).⁷

Take-home message

Based on the results of the study by Bianchi et al coupled with prior reports of the efficacy of cfDNA testing in high-risk populations, we must conclude that for couples desiring fetal aneuploidy risk assessment, cfDNA testing is far superior to standard testing with nuchal translucency and/or maternal serum analyte analysis. Questions remain regarding the best technology to employ, the actual detection rate of cfDNA in everyday use, the availability of pre-test genetic counseling, cost, and the optimal invasive confirmation study. However, it is abundantly clear that our collective experience with genetic amniocentesis and CVS is about to be vastly diminished, which is a good thing, since fewer such procedures will be mean fewer unnecessary pregnancy losses. 

References

1. Bianchi DW, Parker RL, Wentworth J, et al; CARE Study Group. DNA sequencing versus standard prenatal aneuploidy screening. N Engl J Med. 2014;370(9):799–808.

2. Wang E, Batey A, Struble C, Musci T, Song K, Oliphant A. Gestational age and maternal weight effects on fetal cell-free DNA in maternal plasma. Prenat Diagn. 2013;33(7):662–666.

3. Norwitz ER, Levy B. Noninvasive prenatal testing: The future is now. Rev Obstet Gynecol. 2013;6(2):48–62.

4. Gil MM, Akolekar R, Quezada MS, Bregant B, Nicolaides KH. Analysis of cell-free DNA in maternal blood in screening for aneuploidies: Meta-analysis. Fetal Diagn Ther. 2014 Feb 8. [Epub ahead of print]

5. American College of Obstetricians and Gynecologists Committee on Genetics. Committee Opinion No. 545: Noninvasive prenatal testing for fetal aneuploidy. Obstet Gynecol. 2012;120(6):1532–1534.

6. Fairbrother G, Johnson S, Musci TJ, Song K. Clinical experience of noninvasive prenatal testing with cell-free DNA for fetal trisomies 21, 18, and 13, in a general screening population. Prenat Diagn. 2013;33(6):580–583.

7. Song Y, Liu C, Qi H, Zhang Y, Bian X, Liu J. Noninvasive prenatal testing of fetal aneuploidies by massively parallel sequencing in a prospective Chinese population. Prenat Diagn. 2013;33(7):700–706.