22 June 2021

Sequenom Down-Under – Appeals Court Finds Non-Invasive Foetal DNA Test Patent-Eligible in Australia

Double helixIn 18 June 2021, a Full bench of the Federal Court of Australia (Middleton, Nicholas, and Burley JJ) unanimously upheld a decision of a single judge of the court (Beach J), finding that a method of detecting cell-free foetal DNA (cffDNA) in maternal blood serum comprises patent-eligible subject matter (i.e. a ‘manner of manufacture’) under Australian law: Ariosa Diagnostics, Inc v Sequenom, Inc [2021] FCAFC 101.  The patent at issue is Australian patent no. 727,919, covering an invention originally developed by Oxford University researchers, and subsequently transferred to Sequenom Inc.  The patent expired in March 2018, however a live dispute remains because Ariosa Diagnostics licensed its ‘Harmony Test’ – which Sequenom says (and the Full Court has agreed) infringes the patent – for use in Australia since at least September 2015.

To my mind, the result in this case is neither particularly surprising nor contentious.  The patent claims are directed to a method of detecting cffDNA.  While the method is underpinned by the naturally occurring fact – not known until its discovery by the inventors prior to March 1997 – that cffDNA is present in maternal blood serum, a useful method of detecting a previously unknown natural phenomenon, having a practical application, has long been considered patentable.  Ariosa’s arguments that Sequenom’s claims were in substance directed to the ‘mere’ discovery itself, resulted only in the production of ‘information’, and therefore unpatentable, were unsuccessful.  And while the broad scope of the main claim in this case might raise other issues, such as obviousness or sufficiency of description, these matters have also been addressed at first instance and/or on appeal, and are separate from the question of subject matter eligibility. 

Nonetheless, this case will generate some interest, if only because the result in Australia is opposed to the outcome of equivalent litigation between Sequenom and Ariosa involving a corresponding patent in the United States.  In that case, a narrower claim than in Australia was found to be directed to an unpatentable natural phenomenon.  The result was controversial, not least because a number of judges on the US Court of Appeals for the Federal Circuit (CAFC) – including Judge Linn on the original panel (Ariosa Diagnostics, Inc.v. Sequenom, Inc. (Fed. Cir. 2015)) and Judges Lourie and Dyk in a decision refusing en banc rehearing – indicated that they felt bound by the Supreme Court precedents, but did not agree with the outcome.  There was therefore great disappointment when the US Supreme Court declined to hear an appeal.

The outcome of the Australian appeal is not all bad news for Ariosa, however, with the finding of the primary judge on infringement being partially reversed.  In particular, there were periods during which the Harmony Test was not carried out in Australia, but instead samples were sent to the US for testing by Ariosa, which the primary judge found also to be infringing actions.  The Full Court has disagreed, finding that ‘importing’ the information resulting from the tests into Australia did not comprise a relevant ‘exploitation’ of the claimed method, as it might have done were the product of the method a physical article.

Background

One of the many benefits of genetic testing is the ability to carry out prenatal diagnosis of genetic disorders, most notably including (but certainly not limited to) Down Syndrome.  This can be particularly valuable when one or both parents are known to be carriers of a genetic mutation such that the baby may be at significant risk of developing a serious condition or abnormality.  The conventional method of obtaining foetal DNA in order to conduct such tests is via amniocentesis, which is an invasive procedure in which a small amount of amniotic fluid is sampled from the amniotic sac surrounding the developing foetus using a needle.  Unfortunately, amniocentesis presents a risk to both foetus and mother.

A test that would enable screening of foetal DNA in a noninvasive manner would therefore be an important breakthrough.  Fortunately, just such a breakthrough was made sometime prior to 4 March 1997 (the priority date of the patent) at Oxford University, when researchers Stephen Wainscoat and Dennis Lo discovered the presence of cffDNA in maternal blood serum, and developed techniques to isolate this DNA for testing.  Their discovery led to the grant of patents in a number of countries.  Oxford University, through its commercialisation arm Isis Innovation (now renamed as Oxford Innovation, for sadly obvious reasons), licensed the technology exclusively to US company Sequenom (the Australian patent was, in fact, assigned to Sequenom in 2014).

Some 15 years after the breakthrough by Wainscoat and Lo, Ariosa developed the Harmony Test.  Many details of the Harmony test remain confidential, and were redacted from the original judgment.  However, the primary judge described it as ‘a very impressive piece of science’ that ‘relies on some technologies and approaches that were not foreshadowed by the Patent’ including ‘the digital analysis of selected regions of the genome using digital technology that was not available as at 1997, together with sequence information derived from SNP databases that were developed several years after the priority date’: Sequenom, Inc. v Ariosa Diagnostics, Inc. [2019] FCA 1011, at [1164].  Nonetheless, the Harmony Test has been found to infringe a number of valid claims of the Sequenom patent, which cover fundamental aspects of the cffDNA detection process that continue to form part of the more sophisticated Harmony Test.  (To be fair, one would assume that Sequenom’s process has been significantly refined and improved over the years, as well.)

The US and Australian Patent Claims

The Australian cffDNA patent is broader in scope than its US counterpart.  Claim 1 of US patent no. 6,258,540 defined the invention as follows:

1.  A method for detecting a paternally inherited nucleic acid of fetal origin performed on a maternal serum or plasma sample from a pregnant female, which method comprises
        amplifying a paternally inherited nucleic acid from the serum or plasma sample and
        detecting the presence of a paternally inherited nucleic acid of fetal origin in the sample.

Claim 1 of Australian patent no. 727,919, by comparison, recites:

1.  A detection method performed on a maternal serum or plasma sample from a pregnant female, which method comprises detecting the presence of a nucleic acid of foetal origin in the sample.

The differences between the claims are essentially that the Australian claim lacks the ‘paternal inheritance’ and ‘amplification’ limitations of the US claim.  The first of these is important in practice, because it is the presence of features of the father’s genetic inheritance that distinguishes foetal DNA from the mother’s own DNA.  ‘Amplification’ refers to a well-known practice of multiplying copies of the DNA of interest to facilitate detection and further analysis.

These features are, however, also present in the Australian claims – specifically in claims 9 and 2 respectively.  Thus even if claim 1 had been found to be invalid (e.g. for being overly-broad – although this would be unlikely under the law as it was in 1998 when the application was originally filed) Sequenom could still have maintained claims of the same scope as in the US.

Relevant High Court Authorities

The foundational case for the modern application of the ‘manner of manufacture’ test for subject matter eligibility remains National Research Development Corporation v Commissioner of Patents [1959] HCA 67 (NRDC), in which the High Court repudiated the application of any ‘exact verbal formula’ for distinguishing between patentable and unpatentable subject matter. 

NRDC concerned claims directed to a method of treating crops to eradicate weeds using known chemical compounds that were not previously thought to be useful for this purpose.  A much more recent High Court decision involving the patent-eligibility of a method or process is Apotex Pty Ltd v Sanofi-Aventis Australia Pty Ltd [2013] HCA 50, in which the claims at issue were directed to a method of preventing or treating psoriasis using a pharmaceutical substance which was previously known to have other therapeutic uses.  In both cases the methods were found to be patentable, despite involving the use of known substances in substantially conventional ways, and despite not resulting in a tangible product (i.e. an article of manufacture), on the basis that the methods involved the creation of an artificial effect through human intervention, and that the effect in question had some previously unrealised economic significance.

More recently, in D'Arcy v Myriad Genetics Inc [2015] HCA 35 (Myriad), the High Court again affirmed NRDC, reconfirming (at [28]) that two necessary criteria for patentability of a claimed invention are that it be for ‘a product made, or a process producing an outcome as a result of human action’, and that it have ‘economic utility’.  These criteria have compactly been expressed as requiring an ‘artificially created state of affairs of utility in the field of economic endeavour’ (CCOM Pty Ltd v Jiejing [1994] FCA 1168, at [128]), which has sometimes mistakenly been taken as a statement of the requirement for patent-eligibility in Australia, despite the rejection of any such ‘verbal formula in NRDC’.  The plurality of the High Court in Myriad thus reminded us that while these criteria may be necessary, they are not necessarily sufficient, and where a claim is not within established boundaries of what is patentable, other considerations – including matters of policy, the proper roles of the courts and the legislature, and Australia’s obligations under international treaties – may come into play.

Myriad concerned claims directed to an isolated genetic sequence, which ostensibly defined the invention as an artificial product, in the sense that isolated nucleic acids do not exist in nature and only come into being through human action.  However, the plurality of the High Court found that, in truth, the ‘substance’ of the claims was ‘information embodied in arrangements of nucleotides’ that was ‘not "made" by human action’, but rather was ‘discerned’ (Myriad at [6]).  This placed the claimed subject matter ‘at the boundaries of the concept of “manner of manufacture” (Myriad at [93]).  The plurality concluded that to include this subject matter ‘within the scope of a “manner of manufacture” involves an extension of that concept, which is not appropriate for judicial determination’ and ‘would not contribute to coherence in the law’ (at [94]).  Thus, in contrast to the methods claimed in NRDC and Apotex, Myriad’s ‘isolated gene’ claims were found not to be eligible subject matter.

‘Detecting Nucleic Acid of Foetal Origin’ is Patent-Eligible

The appellant, Ariosa, relied substantially on Myriad in support of its appeal.  As the Full Court observed (Ariosa at [108]):

Fundamental to the appeal is the contention that what is claimed is a mere discovery of a naturally occurring phenomenon, and not a method involving a practical application that goes beyond the discovery itself. Here, the appellants identify the discovery as the fact that cffDNA from maternal plasma or serum was a source of foetal DNA. They submit that there is no artificially created state of affairs arising from the invention as claimed. Furthermore, they submit that the “end result” of each claim results in information only, being the detection of cffDNA.

As I have already noted, claim 1 of the Sequenom patent is broad.  It recites a method defined in terms of a single step, comprising ‘detecting the presence of a nucleic acid of foetal origin in the [maternal serum or plasma] sample.’  At first instance the judge construed this broad definition as inherently requiring human action in the form of firstly obtaining the maternal serum or plasma, and then analysing it to detect the presence of cffDNA.  The Full Court agreed with this approach, finding that it did not involve impermissibly reading additional features into the claim (at [151]):

Rather, his Honour was demonstrating that although the method is broadly expressed, there was no means known within the common general knowledge whereby the method of the claim could be achieved without significant human intervention. The person skilled in the art would understand that to perform the detection method within what we have described as integer (1), plasma or serum needs to be separated from blood. To “detect” the nucleic acid of foetal origin within integer (2), analysis involving the use of a molecular technique such as PCR is required.

Ariosa sought to argue (at [152]) that ‘the difficulty with this analysis is that despite the human mediation, “detection” is no different from the discovery that cffDNA is detectable.’  However, the Court found that this argument was based on a misidentification of the invention, which ‘lies not in the mere observation that cffDNA is to be found in maternal plasma (or serum), but in the explanation as to how that knowledge may be unlocked for others to use it …. It is the idea coupled with a practical means of application that makes the invention’ (at [153], emphasis added).

The Full Court agreed with the primary judge (at [154]-[155]) that ‘[t]he artificially created state of affairs is the detection of cffDNA in the tested sample’ and that ‘[t]his “product” is, by definition the result of human action and is not naturally occurring.’  Citing NRDC, Apotex, and Myriad, the Full Court explained (emphasis added):

Unlike the position in Myriad, claim 1 is not, as a matter of substance, directed to genetic information, but to a method involving the practical application of a means for identifying and discriminating between maternal and foetal nucleic acid. Although foetal nucleic acid occurs in nature, the substance of the invention is not cffDNA itself, but the identification of that particular nucleic acid as a part of a method. It is impermissible to disaggregate the integers of the method to point only to the cffDNA as the “invention”. Identification of the substance of the invention does not involve disregarding material aspects of the claim language. The invention as claimed is not merely output, but the detection process which yields an output. This is the very type of subject matter considered to fall on the correct side of the line between discovery of a scientific fact or law of nature and invention….

The Full Court also rejected (at [159]) Ariosa’s argument that ‘the “output” of the method is mere information and accordingly cannot amount to a manner of manufacture’, observing that:

Here, the invention as claimed carries into effect an idea that the presence of information within the naturally occurring code of a person will be useful. By a process of detection that information is yielded up. The claim construed as a whole necessarily involves an artificially created state of affairs yielding an outcome that is of economic utility. We reject the submission that because, disaggregated from the method, the result of the detection of foetal nucleic acid is information about a naturally occurring state of affairs, the requirements of Myriad factors 1 and 2 are not satisfied. The invention as claimed cannot be separated into discrete parts.

What About Computerised Inventions that Yield Information?

Regular readers of this blog, and other observers of developments in Australia patent law, will be aware that there have been a number of decisions of the Full Court over recent years addressing the patent-eligibility of – or, more commonly, the lack thereof – inventions involving computer-implementation.  Plainly, many such inventions, including those that have been found unpatentable, yield useful information as output.  It is therefore not surprising that Ariosa argued that finding Sequenom’s claims to be patentable would be inconsistent with many of these decisions.

The court had very little difficulty in disposing of this argument, given the very different technologies and facts at issue in in cases involving computer-implemented inventions, explaining (at [160], emphasis added):

For the reasons given above, the characterisation of the outcome of the method as “information” does not disqualify the present invention. That is not to say that every method yielding information will be patentable. The consideration of patentable subject matter does not involve a one size fits all approach. In each of those cases the question for consideration was whether or not a mere scheme or plan was nonetheless a manner of manufacture because invention lay not only in the scheme but also the means by which it was realised using computerisation…. The present case does not involve a scheme or plan implemented on the medium of a computer, which would involve different considerations.

‘Importing’ Information is not an Infringement

As I have already noted, the good news for Ariosa was that the Full Court reversed the primary judge’s finding that Sequenom’s patent was infringed by transmitting reports including the results of tests performed overseas (using a method that would have infringed the claims if conducted in Australia).  As a general rule, importing the product of a patented process carried out overseas can infringe and Australian patent, because the Patents Act 1990 contains a definition of ‘exploit’, in relation to an invention, which includes ‘where the invention is a method or process – use the method or process or do [acts including selling or importing] in respect of a product resulting from such use.’

However, as the Full Court observed in Ariosa, at [259], the Act does not contain any definition of the term ‘product’, and its meaning must therefore be discerned having regard to the context in which it is used.  In the Full Court’s view, the definition of ‘exploit’ the Act contemplates the production of of physical articles via a patented process, and was included to address ambiguities and conflicting lines of authority in the UK as to how such circumstances should be approached as regards infringement (at [262]).  As the Full Court further observed (at [263-4]) cases such as NRDC and Apotex establish that ‘there is no necessity in principle for a method or process to yield a product in order for that method or process to amount to a manner of manufacture’, and that the requisite useful result may be, for example, ‘a relatively weed-free field (NRDC)’ or ‘healthier human patients’ (Apotex).

The Full Court concluded, at [269], that ‘we do not consider that the word “product” in para (b) of the definition of “exploit” should be interpreted as extending the patentee’s monopoly to information which could not itself constitute patentable subject matter since it would have the unintended and odd consequence of permitting the patentee to obtain patent protection in respect of subject matter that has long been held to be unpatentable.’

Conclusion – Ruling Strikes a Good Balance Under Current Law

The development of non-invasive foetal genetic testing was a significant breakthrough, given the considerable benefits of such tests were often foregone due to the significant risks posed to mother and child by amniocentesis.  On the face of it, incentivising and supporting commercialisation of such developments is precisely the kind of thing for which the patent system was devised.  This is why the outcome in the US is widely regarded – including, apparently, by a number of Federal Circuit judges – as ‘wrong’, and the result of overreach by the Supreme Court in the US equivalent of the Australian Myriad decision.  In my view, it is a positive thing that the law does not appear to be taking the same path in Australia.

On the other hand, I am inclined to think that claim 1 of the Australian patent is overly broad.  This is not, however, a question of subject matter, but of what we would now regard as enablement and/or support under section 40 paragraph 2(a) and subsection (3) of the current Australian law.  Sequenom’s patent was, however, governed by the former laws of sufficiency and fair basis, which really required only that the specification disclose and enable something falling within the scope of the claims, and that the invention as claimed not be inconsistent with the invention as described.  The claims survived challenges on these grounds at first instance and on appeal, which I find unsurprising.  Whether the broadest claims would be valid if the current laws had applied is a different matter, but I suspect not.

Assuming that the Ariosa decision stands, the position in Australia now is that diagnostic tests such as Sequenom’s are patentable in principle – so long as they meet other requirements, such as novelty and inventive step – but that such expansive claims are probably no longer available, following the Raising the Bar reforms which came into effect in 2013.  Taken together, I think a good balance is struck.

Of course, it is possible that the case is not yet closed.  Ariosa is free to apply for special leave to appeal the decision to the High Court (as is Sequenom, if it disagrees with the Full Court’s conclusion on non-infringement by overseas analysis).  On balance, I think it is quite probable that Ariosa will apply for special leave, but I also think it likely that the High Court would decline to hear an appeal.  It has recently addressed ‘manner of manufacture’ in Apotex and Myriad, and the decision of the Full Court in Ariosa does not appear to raise any new issues that would warrant revisiting this territory.

0 comments:

Post a Comment


Copyright © 2014
Creative Commons License
The Patentology Blog by Dr Mark A Summerfield is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 3.0 Australia License.