On the forefront of the much anticipated arrival of genomic medicine, the field of noninvasive prenatal genetic testing (NIPT) has lived up to its expectations as a game changer. NIPT is estimated to be worth $500 million in 2013, with potential to grow to 2.38 billion by 2022.1 NIPT not only has revolutionized access to prenatal testing for genetic disorders through a mere blood test, but it also is unique enough to defy classification by both intellectual property law and the existing Food and Drug Administration (FDA) regulatory framework.
November 01, 2017
Prenatal Genetic Testing: Where Algorithms May Fail
By Aubrey Haddach and Jeffrey Licitra
Much attention has been given to the Federal Circuit’s 2015 decision in Ariosa Diagnostics, Inc. v. Sequenom, Inc.,2 invalidating the patent that protected Sequenom’s commercially offered Maternit21® test as non-patent-eligible subject matter under the Supreme Court’s Mayo v. Prometheus decision.3 Less discussed is the technology itself, a screening test for a genetic condition, which relies significantly on “next-gen” DNA sequencing and proprietary algorithms to translate massive amounts of data into clinical results. The NIPT technology is described below before turning to a discussion of the law.
Algorithms and NIPT
Advancements in computational genomics, which focuses on developing probability models to interpret the data generated by DNA sequencing, have allowed genetic testing to scale upwards. The next-gen sequencing technology that powers these tests is capable of generating billions of DNA base pair reads from a single sample. This vast quantity of data must be reduced to an interpretable form.
Using this sequencing technology and the accompanying algorithms, genetic testing can be done by isolating fragments of fetal and maternal DNA present in a pregnant woman’s blood and then amplifying those fragments. Through “amplification,” scientists produce millions of copies of the relatively small amount of DNA found in the maternal blood serum to create a sample size large enough for the sequencing algorithms. Of the total DNA found in the maternal blood serum, approximately 3.4–6.2 percent is fetal, with the rest belonging to the mother.4 This DNA is often referred to as “cellular free” because it is not found in the cell nuclei and originates from cells that die naturally, leaving DNA fragments in the expectant mother’s bloodstream.
The most common prenatal genetic tests are for trisomy disorders, such as Down syndrome (trisomy 21), Edwards syndrome (trisomy 18), and Patau syndrome (trisomy 13). The latter two conditions can lead to early miscarriage and significant rates of clinical morbidity or mortality following birth.5
Typically, humans have 23 pairs of chromosomes, with each parent contributing a single chromosome to the pair. Trisomy occurs when an extra chromosome or fragment thereof is associated with the typical pair, resulting in three chromosomes where there should only be two. A trisomy on the 21st chromosome is referred to as “trisomy 21.” Individual chromosomes themselves are made up of DNA “base pairs,” molecules that occur in tandem on opposing strands of DNA, and are abbreviated by the letters A (adenine), T (thymine), C (cytosine), and G (guanine), so as to form what is referred to as the “genetic code.” DNA is “sequenced” to determine the number of base pairs present and if those base pairs correlate to a specific gene, and thus a genetic condition.
Five companies (Sequenom, Illumina, LabCorp, Ariosa, and Natera) currently offer prenatal tests for trisomy conditions, each relying on different variations of the same sequencing methods. Though these methods all involve the same basic steps of isolating, amplifying, and sequencing DNA, each company applies those steps in different and proprietary ways. This matters because each test can offer an incorrect screening result due to errors inherent to individual testing methods and their underlying statistical assumptions.
Sequenom, Illumina, and LabCorp use a method called “massively parallel shotgun sequencing,” while Ariosa and Natera use “targeted sequencing.” Shotgun sequencing amplifies the entirety of the genetic material collected, regardless of chromosome number, and then sorts the sequenced genetic content according to previously known and expected results for the human genome. Targeted sequencing, as the name implies, amplifies only a region or set of genes on the desired chromosome.
Each company then uses its own proprietary algorithms of “quantitative read counting” to count the total number of base pairs sequenced for the chromosome region being examined. One sample alone undergoing shotgun sequencing may generate, on average, 10,800,000 base pair reads, as opposed to 32,000 for targeting sequencing.6 If the algorithm finds too many base pairs associated with a particular chromosome, then the statistical inference is that the extra DNA base pairs indicate a trisomy condition on that chromosome.
What distinguishes prenatal genetic screening from other types of genetic testing is its reliance on the tiny amount of fetal DNA fragments present in the maternal serum. That is to say, NIPT does not actually sequence the entire fetal genome because it only has fragments to work with. It is not so much reading the code as it is counting it. In this way, any amount of extra DNA present in the base pair count, even if that extra DNA is not associated with an extra chromosome, can lead to a false-positive result.
One such false-positive case involved a pregnant woman who had a copy number variation. A copy number variation occurs when there is more DNA in certain chromosome regions due to the variance in length of certain DNA strands. The nonstandard length of maternal DNA with a copy number variation, particularly when inherited in fetal DNA, and thus doubly represented in the sample, could yield extra DNA in the test result, and thus a false-positive reading.7 Interestingly, both Ariosa and Sequenom replied to this case separately, not so much to refute its findings but to explain that they each compensate for copy variance differently in their respective algorithms.8
The Validity of Screening Results
While copy variance is certainly not the only factor that can affect false-positive results for NIPTs, it is instructive because it shows how indispensable underlying statistical assumptions are to the test results. Maternal or confined placental mosaicism,9 vanishing twin pregnancies, and maternal malignancy may all yield false-positive results for trisomy because each of these conditions allows for extra DNA in the maternal blood serum.10
Various journalists have reported on the anguish experienced by parents who relied on false-positive results and unfortunately terminated the pregnancies as a result or who otherwise awaited (or planned) for the arrival of a child with a trisomy condition.11 Yet peer-reviewed journals (and the companies themselves) continue to conclude that NIPTs are fundamentally sound in their ability to make true-positive predictions at a greater than 99 percent rate for trisomy 21.12
A review of results for trisomy 18 and 13 studies showed a lower true-positive rate (the rate at which tests are both positive and correct) of 97–99 percent and 92–95 percent, respectively.13 A meta-analysis of all studies estimated the true-positive rates for trisomy 18 and trisomy 13 to be approximately 95 percent when performed in combination with the trisomy 21 test, but possibly lower when performed alone.14 Moreover, the vast majority of studies have been done on women at a high risk for trisomy, increasing the likelihood of a correct positive test result.15
As the FDA considers NIPTs to be laboratory diagnostic tests (LDTs), it does not regulate prenatal genetic testing beyond a policy of “enforcement discretion”—meaning the FDA chooses whether or not to oversee pre-market testing or impose post-market safety requirements.16 This is because the tests are developed and used entirely in one location, regardless of where the samples originate. LDTs were traditionally used by hospitals billing directly to Medicare. Their “home brew” labs have long been regulated for scientific accuracy and precision by the Clinical Laboratory Improvement Amendments (CLIA) statute.17
Notably, none of this oversight reaches the level of clinical validity, and, perhaps more importantly, adverse results are not required to be reported to the FDA. Contrastingly, clinical results for drugs and diagnostic tests otherwise falling under the FDA’s purview are subject to extensive clinical data review for marketing approval and then continuous adverse event reporting after going to market. This allows an analysis of all adverse events occurring to a far more robust population than could ever be constructed in a clinical study.
This lack of regulatory oversight for NIPTs leaves patients and medical professionals in a position to interpret the results of tests that often influence a woman’s decision to terminate pregnancy. Although NIPTs are screening tests, intended as an intermediate step before a more invasive diagnostic procedure, patients may not appreciate the nuance of a 10 mL blood sample yielding a statistical guess, even if an incredibly accurate one, at the presence of a genetic condition. Similarly, the nature of NIPTs as screening tests has left the courts to grapple with the importance of NIPT as a breakthrough technology and its place within the larger context of patent jurisprudence.
Ariosa, Mayo, and Section 101 Patent Eligibility
There are two central tenets of patent law that determine whether an invention such as NIPT is patent-eligible subject matter under 35 U.S.C. section 101. On one hand is the now infamous expression that Congress intended statutory subject matter to “include anything under the sun that is made by man.”18 On the other is the “natural law exception,” which holds that naturally occurring phenomena are not patentable.19 The problem arises when a patent on a man-made invention—even a ground-breaking one such as NIPT—relies so heavily on a scientific discovery that a court finds the patent invalid for only claiming a natural law.
Sequenom obtained the first patent on NIPT in 2005, when it purchased from Isis Innovation Ltd., the commercial research arm of Oxford University, an exclusive license to a patent for “amplifying” and “detecting” cellular-free fetal DNA (cffDNA).20 In 2011, Sequenom became the first company to offer prenatal genetic testing. Only months later, Sequenom entered into litigation in the Northern District of California with Ariosa, Natera, and Verinata over infringement of its patent. If the patent was valid, Sequenom would have exclusive control of the nascent NIPT market. If invalid, its competitor companies would be able to enter the market without risking infringement.
The task before the district court was to apply the Supreme Court’s two-part Mayo test to determine whether the patent was invalid for lack of eligible subject matter: first, decide whether the method claimed is to a natural law or abstract idea; if yes, proceed to step two, and determine if there is an “inventive concept” sufficient to overcome the patent’s reliance on an abstract idea.21 In October 2013, the district court held the patent invalid as ineligible subject matter, concluding “the only inventive concept contained in the patent to be the discovery of cffDNA, which is not patentable.”22
Sequenom appealed only to have the Federal Circuit affirm the patent’s invalidation and deny rehearing en banc.23 In June 2015, the Supreme Court declined Sequenom’s petition for certiorari, putting an end to its five-year quest to maintain the patent in the face of noninfringement suits brought by its competitors. Along the way, members of the legal community and biotech industry filed numerous briefs imploring the Federal Circuit to avoid a ruling that would nullify patent protection for future genetic screening technology.
The Mayo decision itself was comparatively “low-tech” and involved a method for administering an intravenous drug to a patient, where the dosage of the drug was increased or decreased to obtain an efficacy level based on measuring metabolites in the bloodstream.24 Calculations and dosage adjustments are decisively abstract compared to the multitude of tangible steps involved in noninvasive testing, from the separation of cffDNA in the blood sample to the next-gen sequencing and algorithmic determination of a clinical result. Do these steps constitute a sufficient “inventive concept” to transform the abstract science to patentable subject matter? The Mayo framework may ultimately be inadequate if it confuses patent-ineligible abstract methods with the several, complex, laboratory steps used in molecular diagnostics.
The Federal Circuit applied Mayo to conclude that (1) Sequenom was merely utilizing a natural law, i.e., the presence of fragmented fetal DNA in the maternal bloodstream; and (2) the patent lacked sufficient inventive concept because its claim involved the standard DNA sequencing steps of amplification and detection routinely practiced by scientists.25 However, the claim may have been written so broadly as to render these methods abstract regardless of the patentability of NIPT itself. Indeed, Judge Lourie acknowledged as much in his concurrence, denying rehearing en banc, positing that the patent may have been invalid for lack of specificity regardless of Mayo, and that “the finer filter of § 112 might be better suited to treating these as questions of patentability.”26
The judges, even in agreeing on the result, expressed misgivings about the breadth of the “natural laws” restriction imposed by Mayo. Judge Lourie acknowledged that the holding of Mayo had been correctly applied as binding Supreme Court precedent, but he found the rule “unsound” insofar as it “takes inventions of this nature out of the realm of patent-eligibility.”27 Similarly, Judge Dyk, in his own concurrence, observed that “the major defect is not that the claims lack inventive concept but rather that they are overbroad.”28 Judge Linn, concurring in the panel decision, explained he did so “bound by the sweeping language of the test” set out by the Supreme Court, noting that “the amplification and detection of cffDNA had never before been done.”29 Judge Newman, in her dissent to the denial of rehearing, viewed Sequenom’s patent to be patentable subject matter. In her opinion, the patent at issue involved the “discovery and development of a new diagnostic method” for cffDNA, itself a discovery, in contrast to a situation where “both the medicinal product and its metabolites were previously known, leaving sparse room for innovative advance.”30 In other words, the subject matter of the process patent was neither an abstract process nor natural law, and in failing that, there was no need to look for an “inventive concept.”
Much of the underlying next-gen sequencing technology is already patented or proprietary subject matter. If anything, it is the use of NIPT algorithms themselves to connect the next-gen sequencing technology to a meaningful screening result that may be the “inventive concept.”
Even if the Ariosa decision stands, the patent at issue may not have been representative of the potential patentable subject matter in NIPT. Presently, litigation among these companies continues in federal court and at the Patent Trial and Appeal Board to decide whether the patent is invalid on other bases.31
Moving Forward
Noninvasive prenatal genetic testing involves technology that at present resists classification under FDA regulations and confounds the intuitive notion that ground-breaking inventions deserve patent protection. This is in part because the technology itself is so much more complex than what the legal profession is accustomed to. NIPT is not just a traditional laboratory diagnostic test, but a screening exam for a genetic condition with clinical implications. Similarly, it is not just the application of a natural law, in the discovery of cffDNA, but a novel way of applying next-gen sequencing technology to that discovery.
The public has an interest both in seeing the biotech industry continue to innovate and in receiving the benefits of more robust testing and oversight. The two interests are not contradictory, and a better understanding of the technology that makes these tests possible should in turn lead to better laws and patient outcomes. u
Endnotes
1. Press Release, Transparency Mkt. Research, Non-Invasive Prenatal Testing (NIPT) Market (BambniTest, Harmony, informaSeq, MaterniT21 PLUS, NIFTY, Panorama, PrenaTest, verifi, VisibiliT and Others)—Global Industry Analysis, Size, Volume, Share, Growth, Trends and Forecast 2014–2022 (May 21, 2015), http://www.transparencymarketresearch.com/noninvasive-prenatal-diagnostics-market.html.
2. 788 F.3d 1371 (Fed. Cir.), reh’g en banc denied, 809 F.3d 1282 (Fed. Cir. 2015), cert. denied, 136 S. Ct. 2511 (2016).
3. Mayo Collaborative Servs. v. Prometheus Labs., Inc., 132 S. Ct. 1289 (2012).
4. Y.M. Dennis Lo, Quantitative Analysis of Fetal DNA in Maternal Plasma and Serum: Implications for Noninvasive Prenatal Diagnosis, 62 Am. J. Hum. Genetics 768 (1998); see also Yuk Ming Dennis Lo, Non-Invasive Prenatal Diagnosis by Massively Parallel Sequencing of Maternal Plasma DNA, 2 Open Biology, June 2012 (reporting that fetal DNA concentrations have been found as high as 10 percent using next-gen, digital polymerase chain reaction (PCR)).
5. Errol R. Norwitz et al., Noninvasive Prenatal Testing: The Future Is Now, 6 Revs. in Obstetrics & Gynecology 48, 49 (2013).
6. Andrew B. Sparks et al., Selective Analysis of Cell-Free DNA in Maternal Blood for Evaluation of Fetal Trisomy, 32 Prenatal Diagnosis 3, 4 (2012).
7. Matthew W. Snyder et al., Copy-Number Variation and False Positive Prenatal Aneuploidy Screening Results, 372 New Eng. J. Med. 1639 (2015).
8. Correspondence: Copy-Number Variation and False Positive Prenatal Aneuploidy Screening Results, 373 New Eng. J. Med. 2583 (2015).
9. Mosaicism occurs when cells carry more than one genotype, allowing for more DNA than normal.
10. Zandra C. Deans et al., Recommended Practice for Laboratory Reporting of Non-Invasive Prenatal Testing of Trisomies 13, 18, and 21: A Consensus Opinion, 37 Prenatal Diagnosis 699 (2017).
11. See, e.g., Beth Daley, Oversold and Misunderstood: Prenatal Screening Tests Prompt Abortions, New Eng. Ctr. for Investigative Reporting (Dec. 13, 2014), available at https://eye.necir.org/2014/12/13/prenatal-testing/; Prenatal Tests Have High Failure Rate, Triggering Abortions, NBC News (Dec. 14, 2014), http://www.nbcnews.com/health/womens-health/prenatal-tests-have-high-failure-rate-triggering-abortions-n267301.
12. Megan Allyse et al., Non-Invasive Prenatal Testing: A Review of International Implementation and Challenges, 7 Int’l J. Women’s Health 113, 114 (2015); Norwitz et al., supra note 5, at 59.
13. Allyse et al., supra note 12, at 114.
14. M.M. Gil et al., Analysis of Cell-Free DNA in Maternal Blood Screening for Fetal Aneuploidies: Updated Meta-Analysis, 45 Ultrasound Obstetrics & Gynecology 249, 261–62 (2015).
15. Norwitz et al., supra note 5, at 50.
16. U.S. Food & Drug Admin., Draft Guidance for Industry, Food and Drug Administration Staff, and Clinical Laboratories: Framework for Regulatory Oversight of Laboratory Developed Tests (LDTs) 21 (Oct. 3, 2014).
17. 42 U.S.C. § 263(a).
18. Diamond v. Chakrabarty, 447 U.S. 303, 309 (1980).
19. Funk Bros. Seed Co. v. Kalo Inoculant Co., 333 U.S. 127, 130 (1948).
20. U.S. Patent No. 6,258,540.
21. Alice Corp. Pty. Ltd. v. CLS Bank Int’l, 134 S. Ct. 2347, 2353 (2014) (“[A] court must first ‘identif[y] the abstract idea represented in the claim,’ and then determine ‘whether the balance of the claim adds significantly more.’” (alteration in original)); Mayo Collaborative Servs. v. Prometheus Labs., Inc., 132 S. Ct. at 1289, 1299 (2012) (“These other steps apparently added to the formula something that in terms of patent law’s objectives had significance—they transformed the process into an inventive application of the formula.”).
22. Ariosa Diagnostics, Inc. v. Sequenom, Inc., 19 F. Supp. 3d 938, 950 (N.D. Cal. 2013).
23. Ariosa Diagnostics, Inc. v. Sequenom, Inc., 788 F.3d 1371 (Fed Cir.), reh’g en banc denied, 809 F.3d 1282 (Fed. Cir. 2015), cert. denied, 136 S. Ct. 2511 (2016).
24. Mayo, 132 S. Ct. at 1294.
25. Ariosa, 788 F.3d at 1378 (“Thus, in this case, appending routine, conventional steps to a natural phenomenon, specified at a high level of generality, is not enough to supply an inventive concept.”).
26. Ariosa, 809 F.3d at 1286 (Lourie, J., concurring).
27. Id. at 1287.
28. Id. at 1293 (Dyk, J., concurring).
29. Ariosa, 788 F.3d at 1380–81 (Linn, J., dissenting).
30. Ariosa, 809 F.3d at 1293–94 (Newman, J., dissenting).
31. See, e.g., Verinata Health, Inc. v. Ariosa Diagnostics, Inc., No. 12-cv-05501-SI (N.D. Cal. Jan. 19, 2017).