NIPT (non-invasive prenatal testing) continues to get lots of attention lately. Indeed, we've written about it extensively on this blog. None of this is suprising because NIPT is a new technology that is continually evolving. Two years ago, I wrote about NIPT here and provided information showing it's excellent diagnostic sensitivity and specificity. To be clear: these tests are more accurate than traditional biochemical screening for detecting fetal aneuploides but they are still screening tests, meaning that positive (or abnormal) test results must be confirmed with diagnostic testing.
As is commonplace, with time comes experience and the lens of scruitiny has recently been focused on the positive predicitive value (PPV) of NIPT. What's a PPV? It's the proportion of true positive results divided by the number of all positive results. For NIPT testing, it answers the question: "What is the probability that a positive result means that the fetus is affected?" It is very important to stress that the PPV of any test is not intrinsic to the test. The PPV is also dependent on the prevalence of the condition in the tested population. If the condition is very rare in the tested population, then the PPV will likely be low, meaning that a positive result is more likely to be a false positive. The opposite is also true (positive test results are more likely to be "true" when the condition is highly prevalent).
NIPT is done to screen for fetal aneuploidies (extra copies of specific chromosomes) such as Down syndrome (trisomy 21), Edwards syndrome (trisomy 18), and Patau syndrome (trisomy 13). The prevelance of each of these disorders is influenced by the woman's age. As examples, the prevalence of each in a 35-year-old woman with a fetus at 10 weeks’ gestational age is 1:185, 1:470, and 1:1,500, respectively. As you might expect, the less prevalent a condition is, the more likley a positive result will be falsely positive.
This has been demonstrated for NIPT. A study published earlier this year evaluated the concordance of NIPT and cytogenetic results among cases with positive or negative NIPT results. The study examined test results from 109 consecutive specimens that were either prenatally and/or postnatally studied by fluorescence in situ hybridization, karyotyping, and/or oligo–single-nucleotide polymorphism microarray (as the definitive, or diagnostic, test). NIPT testing was performed with the Panorama (Natera, San Carlos, CA), Harmony (Ariosa Diagnostics, San Jose, CA), MaterniT21 (Sequenom, San Diego, CA), or Verifi (Illumina, Redwood City, CA).
The PPV for T21 was highest at 93% followed by a 64% PPV for T18. The PPV for T13 was only 44%. Given the prevalence of each of these conditions, these data aren't all that surprising but they are still rather alarming. Why? Because several studies have claimed NIPT tests are >99% specific (e.g. ~1% false-positive rate). As the authors of the study described here state: "To an average clinician, the claim that a test is >99% specific leads him or her to expect that the false-positive rate will be <1%."
As I stated this above and in several other posts on this blog (but is worth emphasizing again): NIPT is a screening test, not a diagnostic test and it cannot be considered a replacement for diagnostic testing.