Thursday, November 5, 2015

"Retrosynthetic Analysis" Fails to Establish Sufficiency

Gilead Sciences, Inc v Idenix Pharmaceuticals, Inc 2015 FC 1156 Annis J
            2,490,191 2,527,657 / sofosbuvir / SOVALDI

Yesterday’s post provided an overview of the facts in Gildead v Idenix and dealt with Gilead’s successful utility attack on Idenix’ 191 patent. Today’s post deals with Gilead’s sufficiency attack, which was also successful.

Recall that as of the filing date Idenix had not made any of the 2’-C-Me/F compounds which it claimed, and the disclosure dealt primarily with 2’-C-Me/OH compounds, which had been synthesized. Consequently, Gildead argued straightforwardly that the disclosure was not sufficient to allow a skilled person to make any of the compounds falling within the claims [413]. This is a classic "how to make" attack, which was perhaps the most common reason for invalidating a patent in the early days of UK patent law, but which we don’t often see anymore. As with utility, the result turned on the facts. This case does provide a good illustration of the point that making the invention may be insufficient as requiring “undue experimentation” even if the effort exercised by the skilled person does not rise to the level of invention. Annis J’s restrained interpretation of the SCC’s Viagra 2012 SCC 60 decision is also of interest, as are his comments on the proper approach to experiments intended to simulate the efforts of a skilled person.

How to Make
It is well established that a sufficient disclosure need not set out a recipe which can be followed mindlessly. There is no entirely standard way of expressing the permissible degree of effort. Perhaps the most common is to say that the disclosure is sufficient if it allows a skilled person, aided by their common general knowledge, to make the invention with no more than “routine experimentation.” It is also said that the disclosure is sufficient so long as the invention can be made without “undue experimentation,” (or “effort,” or "burden.”) Annis J used all of these phrases more or less interchangeably: [421], [425], [540], [547].

By the time of trial there were three known methods of making the claimed 2’-C-Me/F compounds, two of which involve a 2’-C-OH/Me intermediate. Annis J concluded that the disclosure itself did not provide instructions as to how to make either the claimed compound or the intermediate [500], so the real question was whether the common general knowledge would be sufficient to allow the skilled reader to make the claimed compound “with at best snipits of useful information well buried in the ‘191 Patent” [491]. The experts all considered whether a “retrosynthetic analysis” would suffice. This just means the skilled person would work backward from the claimed compound to see how to get to some known starting points. As Annis J pointed out “A retrosynthetic plan is at best a contingent roadmap for a proposed route through a labyrinth of research and trial and error experimentation that, when it confronts a dead end, goes back to the last step, and then again maybe the last step before that, and tries another route. There is no common general knowledge to any plan, as every plan is different and depends upon context. . . . A retrosynthetic analysis is used precisely because the synthesis path is unknown” [505]. This calls to mind the statement from Pioneer Hi-Bred [1989] 1 SCR 1623, 1641 that “if the specification contains something that necessitates the working out of a problem, the patent cannot be supported” (quoted by Annis J, with emphasis [417]). Nonetheless, Annis J did not reject the retrosynthetic analysis out of hand, but carefully considered all the various possibilities advanced by Idenix. While he did not find that any particular step rose to the level of invention [433-34], [548], he did note that there was a large number of choices facing a skilled person as to which general approach to try, which specific reagents to use, and so on. He held that “routine trial and error experimentation does not occur where there is an excess of choice, not to mention that an excess of choice entirely excludes the possibility that the disclosure of the 2’-C-Me/F compound could be made by common general knowledge” [547].

Apart from this excess of choice, Idenix faced the problem that its own chemist, Dr Griffon, starting around the filing date – which is to say, with the benefit of the information in the disclosure – had tried for almost two years to synthesize the 2’-C-Me/F compound, without success [550]. This was evidently a “major hurdle” in trying to show that the disclosure was sufficient [608]. Idenix dealt with this by trying to paint Dr Griffon as incompetent and deceitful. After a detailed review of the evidence, Annis J concluded that Dr Griffon was a meticulous chemist [685] and that there was no evidence that he had misled his superiors [686]. Annis J noted in particular that Dr Griffon’s supervisors were not called to testify, and he came to the conclusion that they were not prepared to be complicit in the attack on the professional reputation of “a long-standing, apparently competent employee whom they do not blame for Idenix’s failure to synthesize the 2’-C-Me/F compound” [677].

On the other hand, Gilead’s employee Jeremy Clark, who did not even have a PhD in chemistry, had successfully synthesized the 2’-C-Me/F compounds in only a few months. Idenix argued that this showed that the disclosure of the 191 patent was sufficient. Annis J concluded that Mr Clark had not synthesized the compound all that easily [761], and in any event, Mr Clark may have been particularly lucky, or particularly talented, and his success was not sufficient on its own to establish the sufficiency of the disclosure [771], [773].

Idenix had also argued that Dr Griffon actually had synthesized the 2’-C-Me/F compound in 2003, but due to his purported incompetence had not recognized that he had done so [628]. To establish this Idenix asked Albany Molecular Research Inc [AMRI] to simulate Dr Griffon’s 2003 experiments. While AMRI successfully synthesized the 2’-C-Me/F compounds, Annis J found that the experiments did not faithfully simulate those of Dr Griffon [692]. Further, AMRI carried out two tests in secrecy before inviting Gilead to send a representative to observe the third experiment. It also emerged that Idenix’ lawyers were directing AMRI on the protocol to be used, and Annis J inferred that Dr Griffon’s protocol was not followed in order to improve Idenix’ chances of obtaining the desired result [709], [714]. Idenix also argued that Gilead should have carried out its own tests. Annis J called this argument “facetious at best” [710]. Annis J held that given that the product being tested was Idenix’, not Gilead’s, “Idenix should have offered to retain an outside testing agency mutually agreed upon by Gilead. Similarly, it could have been transparent on working up a protocol that replicated all the steps followed by Dr Griffon and by permitting Gilead to participate fully in all the testing, all at Idenix’s costs. In these circumstances, it would have had a plausible argument, even if Gilead rejected the offers, that the results of the tests should serve considerable probative value” [713]. For these reasons Annis J gave little weight to the AMRI experiments [717].

In light of all this, Annis J held the disclosure to be insufficient [775].

Even though Idenix had not relied on Claim 1, Gilead attempted to attack the patent by attacking Claim 1, the broadest claim [180]. Relying on the SCC’s decision in Viagra 2012 SCC 60, Gilead argued that there was only one “invention” [179] and if that invention is invalid then the whole patent falls. Annis J rejected this, citing paragraph 80 of Viagra itself for the proposition that s 58 of the Act means that the claims stand or fall independently [181].

This argument was raised in a variety of places, but most prominently in the context of sufficiency [437], where Annis J remarked that “I admit to difficulty in following this argument” [438]. I understand why Gilead advanced the argument it did, as the Viagra decision is extremely difficult to reconcile with s 58 of the Act. (Gilead attempted the reconciliation by arguing that whether claims stand or fall independently depends on the particular patent [436].) I also understand why the argument was difficult to follow, because Viagra itself is difficult to follow, or, to put more bluntly, incoherent and internally contradictory: see my article, The Duty to Disclose “The Invention”: The Wrong Tool for the Job, (2013) 25 IPJ 269. Annis J held that

[439] I do not think that this is important in any event. I find that Viagra has insufficient parallels to be applied to this matter. Viagra is a case about a situation where the invention is known and the inventor is quite able to describe it, but leaves the addressee up in the air as to which of two claims contains the accurate disclosure, thereby requiring a minor research project to find out.

I entirely agree with this assessment of Viagra, which should be confined to its facts. Annis J’s holding on this point adds to a line of cases in which the FC has modestly interpreted Viagra: see herehere and here. In the conclusion to my IPJ article, I stated that “If we are lucky, the federal courts will recognize the injustice that may be wrought by the requirement to disclose ‘the invention,’ and distinguish [Viagra’s] treatment of the disclosure requirement into insignificance.” Fortunately, it looks like that is exactly what is happening.

No comments:

Post a Comment