The American College of Obstetricians and Gynecologist recommends that pregnant women, regardless of age, be offered prenatal screening or invasive testing for chromosomal abnormalities at <20 weeks’ gestation. The use of noninvasive prenatal screening (NIPS) for fetal aneuploidy using cell-free DNA (cfDNA) has increased since 2011. These commercially available screening tests have demonstrated high sensitivity and specificity to detect the common aneuploidies in clinical trials. Current guidelines from the American College of Obstetricians and Gynecologist recommend that cfDNA screening with concurrent genetic counseling be offered to women who are at increased risk of aneuploidy. Increased risk typically is defined as age >35 years at the time of delivery, ultrasound findings that suggest an increased risk of aneuploidy, a history of a previous pregnancy with trisomy, other positive screening for aneuploidy, or a parental balanced Robertsonian translocation that increases the risk of trisomy 21 or 13. In low-risk populations, emerging evidence suggests that cfDNA screening may outperform traditional first- and second-trimester screening for trisomy 21. Numerous trials in low-risk populations are on-going.

The commercial laboratories that offer these tests market directly to providers and to patients. Some of this marketing is misleading and implies that cfDNA screening currently is recommended for all women. It has been our clinical experience that this marketing is effective and leads patients to request cfDNA screening regardless of their risk status. Laboratories and investigators recommend confirmation of abnormal results with invasive testing, but both patients and providers may misinterpret marketing materials and assume that cfDNA screening can replace diagnostic testing. These tests remain screening tools, and confirmation of positive results through invasive testing is recommended. A recent study found that 6.2% of women who received a positive cfDNA screening test result proceeded with pregnancy termination without confirmatory testing. This study generated media attention and highlighted the risk of false-positive results. We believe that educating providers and patients about positive predictive value (PPV) will further reinforce the limitations of positive results and may prevent situations in which patients choose not to pursue diagnostic testing to confirm positive results. To that end, tools to educate patients about cfDNA screening and its implications are necessary.

Laboratories often report sensitivity and specificity when describing their test’s performance, using these numbers to describe their test’s accuracy. Despite validation studies that have demonstrated high sensitivity and specificity, false-positive and false-negative results have been reported. We believe that PPV is an important and underreported characteristic of a screening test that should be disclosed to patients and providers. The PPV (or, the likelihood that a positive result is a true positive) is affected by the prevalence of the disorder in the population. Therefore, in low-risk populations, a larger proportion of positive results will be false-positive results.

There are several biologic mechanisms that may produce discordant results between cfDNA screening and cytogenetic testing. These include placental mosaicism, maternal mosaicism, an unrecognized lost cotwin, or maternal malignancy. Although numerous cases of false-positive cfDNA screening have been reported, there is not a standardized mechanism for providers to report these to laboratories or to researchers. Some researchers have called for the establishment of a registry of false-positive and false-negative cases to better understand the mechanisms by which these occur.

To assist with counseling patients and referring providers about these concepts, members of the Maternal Fetal Medicine Division at the University of North Carolina–Chapel Hill developed a web-based calculator to estimate the PPV of the 4 cfDNA testing platforms that currently are available commercially in the United States (MaterniT21 Plus, Sequenom, Inc., San Diego, CA; Harmony, Ariosa Diagnostics, San Jose, CA; verifi, Illumina, Inc., San Diego, CA; Panorama Prenatal Screen, Natera, Inc., San Carlos, CA) as a function of a patient’s age- and gestational age–related risk of trisomies 21, 18 and 13. The calculator uses previously published age- and gestational age–related risk tables for trisomies 21, 18, and 13, as well as the sensitivity and specificity of each test, as published by the commercial laboratories that perform the testing on their websites or identified in validation studies for the specific testing platform. To calculate the PPV, we calculated the number of true positives (number of affected × sensitivity) and false positives (number of unaffected × [1-specificity]). The PPV is calculated by dividing the true positives by the total positives. Users of the calculator input their patient’s age, gestational age, and the cfDNA screening test that they intend to use. The calculator generates the patient’s age-related risk estimate for trisomy 21, 18, and 13 and the PPV. The application also allows users to estimate the PPV based on an a priori risk for patients whose risk at the time of screening may be increased above their age-related risk (ie, by abnormal serum screening or markers of aneuploidy on ultrasound scanning). Before the development of this tool, we manually calculated PPV for each patient with a positive screen result based on her age-related risk. Our providers and genetic counselors have been using this calculator in clinical practice for 6 months to help patients to understand their options for prenatal screening for aneuploidy before undergoing testing. When a positive result returns, we also use the tool to counsel the patient about her specific chance of a true positive. It has increased satisfaction among our genetic counselors; studies that are investigating its use and patient satisfaction are planned. The calculator does not estimate the negative predictive value of these tests because the negative predictive value is expected to be high when using a highly sensitive screening test in a low-prevalence population. Discussions of the risks of false negative test results, however, should be included in counseling.

NIPS with the use of cfDNA is an important technology that is proving to be more sensitive and specific for the detection of common fetal aneuploidies than previous screening strategies. Additionally, the implementation of screening with cfDNA is decreasing the need for invasive procedures. False-positive results have been reported, despite laboratory claims of near 100% accuracy. Although these claims are based on a test’s sensitivity and specificity, they can be misleading, especially in situations in which aneuploidy prevalence is lower. We believe that commercial laboratories should provide estimates of a patient-specific PPV for all reported results. Until then, we suggest that practitioners use our calculator to estimate patient-specific PPVs, which can then be used to counsel patients.

We would like to emphasize that this calculator was developed as an adjunct to, not a replacement of, in-person counseling. There are limitations to this calculator, namely that it does not account for errors in estimation of the age-/gestational age–related risk of aneuploidy or the confidence intervals around each tests’ sensitivity and specificity. The PPV therefore may be higher or lower than the calculator indicates. Thus, the calculator is not meant to be an exact predictor of any 1 test’s performance. As of submission, several additional commercial laboratories are running testing with the verifi platform (informaSeq Prenatal Test [LabCorp, Burlington, NC]), or passing collected samples to Illumina, Inc., for analysis (Informed Pregnancy Screen [Counsyl, Inc., San Francisco, CA], and verifi Prenatal Test by Progenity [Progenity, Inc., San Diego, CA]). Because test performance data for these laboratories currently is referencing the verifi data, we have not included these platforms separately in the calculator. In addition, the calculator is meant only for singleton pregnancies and does not include sex chromosomal abnormalities, rare trisomies that are tested in some laboratories, or microdeletions, which are also included on some panels. Given the limited data available on sensitivity and specificity of these aneuploidies, we did not include them in this calculator. However, because of the very low prevalence of microdeletion syndromes in particular, it is important for providers who order these tests to understand that the PPV will be low. For example, 22q11.2 deletion syndrome (Digeorge syndrome) is the most common microdeletion syndrome with incidence of 1:2000. A positive cfDNA result for this microdeletion syndrome has only a 5.3% chance of being a true positive.

In summary, NIPS with the use of cfDNA can improve screening and decrease the need for potentially unnecessary invasive testing for our patients. It is not without risk, however; practitioners and those involved in the marketing of these tests should be upfront with patients regarding the probability of a false-positive result.

The calculator is available at the University of North Carolina’s Center for Maternal & Infant Health website ( www.mombaby.org/NIPS ). Our goal is to keep the calculator up-to-date as more data accumulates. Our overarching objective is to illustrate how disease prevalence affects the performance and PPV of a screening test and call for each laboratory to report a PPV with every test result.