Intelligent Bio-Systems, Inc. v. Illumina Cambridge Ltd. (Fed. Cir. 2016)

By Andrew Williams --

Have you ever mixed up the obviousness determinations of "motivation to combine" and "reasonable expectation of success"?  If so, you are apparently not alone -- the Federal Circuit recently faulted the Patent Trial and Appeal Board of doing just that.  In Intelligent Bio-Systems, Inc. v. Illumina Cambridge Ltd., the Court reviewed the Board's determination that Intelligent Bio-Systems, Inc. ("IBS") failed to establish the obviousness of one of Illumina's patents, the 7,566,537 patent, in view of prior art cited in an IPR petition.  Importantly, the Court determined that the Board incorrectly found there was no "reasonable expectation of success" because the "success" on which the Board focused was identified only in prior art references.  Instead, the correct analysis required a showing of an "expectation of success" in obtaining the claims at issue when combining the cited references.  In this case, the claims at issue did not contain a limitation that would have put the "expectation of success" in question.  Nevertheless, the Federal Circuit clarified that what the Board had meant to say was that there would have been no motivation to combine the cited references.  This was true, however, because one skilled in the art would not have expected the particular combination of references to have worked for the purposes as found within those references.  In other words, the Federal Circuit clarified that if the references cited in an allegation of obviousness provide reasons why a skilled artisan would not have combined them, it negates a "motivation to combine" and not a "reasonable expectation of success" of ultimately combining the references.

The technology at issue was a method of labelling nucleic acid molecules for use in, among other things, sequencing by synthesis ("SBS").  The label can be used to identify which particular base is being incorporated into the synthesized strand at each particular location.  Equally important is the use of a protecting group, which prevents additional bases from being added until the label can be removed.  However, in order for SBS to be successful, the protecting group needs to be removed ("deblocked") at near-100% efficiency.  The inventors of Illumina's patent at issue found that a protecting group with an azido group (in their case, azidomethyl) could meet this exacting standard.  The only independent claim at issue in the '537 patent was:

1. A method of labeling a nucleic acid molecule, the method comprising incorporating into the nucleic acid molecule a nucleotide or nucleoside molecule, wherein the nucleotide or nucleoside molecule has a base that is linked to a detectable label via a cleavable linker and the nucleotide or nucleoside molecule has a ribose or deoxyribose sugar moiety, wherein the ribose or deoxyribose sugar moiety comprises a protecting group attached via the 2' or 3' oxygen atom, and said protecting group can be modified or removed to expose a 3' OH group and the protecting group comprises an azido group.

The Federal Circuit emphasized the claim language underlined above.  Importantly, the Board had found that claim 1 did not require that the protecting group be removed, and neither party challenged this finding.

This case began when IBS filed a revised petition requesting inter partes review of claims 1-6 and 8 of the '537 patent.  IBS identified three references that it alleged rendered the claims at issue obvious:  (1) WO 91/06678 ("Tsien"), (2) U.S. Patent No. 6,664,079 ("Ju"), and (3) Zavgorodny et al., 32 Tetrahedron Letters 7593 (1991) ("Zavgorodny").  Ju and Tsien were cited as teaching SBS using a labeled nucleotide and using a protecting group at the 3'-OH position.  The Board found that Zavgorodny taught an azidomethyl moiety as a suitable protecting group for the 3'-OH position of nucleosides, and that the group was cleavable under specific and mild conditions.  Nevertheless, Ju taught that for SBS to be successful, the protecting group must be removed with high yield (in order to allow the incorporation and detection of subsequent nucleotides), and Tsien taught that for quantitative deblocking, efficiency needs to be close to 100%.  Moreover, as Illumina pointed out, the prior art reference Loubinoux et al., 44 Tetrahedron 6055 (1988) ("Loubinoux"), reported that azidomethyl group removal from phenols using the same conditions as Zavgorodny resulted in no more than 60-80% removal efficiency.

When looking at the "predictable result" aspect of obviousness as articulated in KSR, a reasonable expectation of success requires that the combination of references work for its intended purpose.  The Board found that, because of the low level of removal efficiency as taught by Loubinoux, IBS's "[p]etition did not provide a specific or credible explanation why an ordinary artisan would have expected Zavgorodny's azidomethyl protecting group to meet Tsien's quantitative deblocking requirement under conditions suitable for use in Tsien's sequencing methods."  Beyond this, however, the Federal Circuit could not figure out if the Board was relying on "reasonable expectation of success," "motivation to combine," or "both."  Therefore, to the extent the decision was based on "reasonable expectation of success," the Federal Circuit found it to be flawed.  The problem was that the Board should have identified a reasonable expectation of making the claimed invention, not of combining the references for the purposes found in the prior art.  As mentioned above, the Board interpreted the claims as not requiring that the azido group be removed.  As such, Loubinoux's teaching of low deblocking efficiency was irrelevant.  And, therefore, as far as the obviousness determination turned on that factor, it was in error.

Nevertheless, as the Court put it, the Federal Circuit "sit[s] to review judgments, not opinions," citing Stratoflex, Inc. v. Aeroquip Corp., 713 F.2d 1530, 1540 (Fed. Cir. 1983).  While the language cited above from the Final Written Decision of the Board (about there being no specific or credible explanation why a skilled artisan would have expected the azido group to meet the deblocking requirement of Tsien) was not relevant to an expectation of success analysis, it was relevant to the motivation to combine.  The Federal Circuit held that, based on the findings of the Board, there was no suggestion or reason to combine Tsien or Ju with Zavgorodny to achieve the claimed invention.  In other words, the low deblocking efficiency did not impact the expectation of success, but it did provide a disincentive for combining these references that was not otherwise overcome by IBS.  Thus, the finding that Illumina's patent was not obvious was correct.

The Federal Circuit also considered IBS's contention that the Board should have considered its reply brief.  IBS attempted to explain why a skilled artisan would have modified the cleavage conditions such that the alleged high-percentage cleavage requirement would have been satisfied.  The Board did not consider this argument because it was raised for the first time in the reply brief.  On appeal, the Federal Circuit noted the expedited nature of the IPRs, and that it is incumbent on petitioners to identify "with particularity" all evidence in the petition that supports the identified grounds.  The reply brief should only address issues raised in the Patent Owner's response.  "While the Board's requirements are strict ones, they are requirements of which petitioners are aware when they seek to institute an IPR."  As such, the Federal Circuit found that the Board did not abuse its discretion in excluding this evidence.  The Board's decision in this case was therefore affirmed.

Intelligent Bio-Systems, Inc. v. Illumina Cambridge Ltd. (Fed. Cir. 2016)
Panel:  Circuit Judges O'Malley, Wallach, and Hughes
Opinion by Circuit Judge O'Malley

Comments

[Atari Man]

So in Boolean notation, the CAFC was saying,

no expectation of success of the combination --> no reason to combine

which is equivalent to

reason to combine --> expectation of success of the combination

[if a --> b is true, then the contrapositive, -b --> -a, is also true]

[Paul F. Morgan]

I thought this was an excellent write-up, especially as a good argument as to why this decision is not necessarily inconsistent with Bilski or KSR. It also tips off future IPR parties as to something else to consider adding to their declarations.
-------------
Off topic, but I think many readers would really like to see a "Patent Docs" blog by a real patent damages expert on what the real impact will be of any patent claims ending up after the ongoing interference on CRISPER - widely viewed a multi-billion dollar invention.* In particular, what patent infringement damages recovery, if any, is possible for making, using or selling plants, animals or pharmaceuticals developed by CRISPER gene splicing, vis a vis what damages are recoverable just for using that tool to making the prototypes from which the products are replicated?

*P.S. Noting that in many modern PTAB interferences it is not at all unusual for one or both parties to lose interfering claims as invalid in the interference motions period, before any priority [of invention date] contest.