Sandoz AG v Bayer Intellectual Property GmbH

AGREED TERMINOLOGY 8 Prior to the trial before the primary judge, the parties filed an agreed glossary and technical primer. Extracts from the agreed technical primer are included in the judgment. Some of these are set out below: (1) Absorption refers to the process by which an API (defined below) moves from the site of administration into a site in the body, usually the bloodstream. Absorption may be defined in terms of its rate and extent. For an orally administered API, the administration site is the gastrointestinal (GI) tract. In the case of a solid oral dosage form, the total extent of absorption may be determined by a number of factors, including the extent of release, solubility, permeability, local and pre-systemic metabolism, stability characteristics of the API in the GI tract and GI tract transit times. In the case of a solid oral dosage form, the rate of absorption may be determined by factors including the rate of release (including disintegration), dissolution rate, solubility, permeability, local and pre-systemic metabolism, rate of degradation (if any) of the API in the GI tract and the presence of food in the GI tract: J [14]. (2) Anticoagulants are a class of antithrombotic drugs that inhibit the formation and/or propagation of coagulation, and/or the formation of fibrin, and are used in the treatment of thromboembolic disorders. By inhibiting the formation or growth of a clot, the anticoagulant allows the body time to break down the clot. Anticoagulants are also used prophylactically and chronically to decrease the risk of a new clot forming: J [18]. (3) API means active pharmaceutical ingredient. The words "API" and "drug" can be used interchangeably: J [19]. (4) Clot or thrombus consists of a mesh of fibrin that embeds activated platelets and red and white blood cells: J [25]. (5) Coagulation cascade refers to a series of activation steps of various enzymes (factors) involved in coagulation. Under normal physiological conditions, it is only upon an injury that the coagulation system should rapidly and locally respond at an injured blood vessel wall to generate a clot: J [28]. (6) Dosage form refers to the form in which the API is administered to the patient, such as tablets, capsules, solutions, suspensions, nasal sprays etc: J [34]. (7) Dosage regimen refers to the regimen by which an API is administered, and includes the dose (quantity of API administered), dosage form, route (eg oral, intravenous, subcutaneous) and frequency of administration: J [35]. (8) Factors refers to the main enzymes involved in the coagulation cascade: J [39]. (9) Fibrin is the fibrous protein that makes up a clot or thrombus: J [40]. (10) In vitro refers to analysis, tests or evaluations conducted outside a living organism (eg in a test tube): J [43]. (11) Thromboembolism is a general term describing the combined processes of thrombosis (where a vessel is occluded by clots in part, or in entirety), or embolism (when a clot breaks off from the formation site and lodges elsewhere in the body). Thrombotic or thromboembolic disorders involve thrombosis and/or thromboembolism, which may develop in the venous or arterial circulatory systems: J [66]. (12) Thrombosis is a pathological clot formation resulting in the obstruction of blood flow through a blood vessel. The primary mechanisms for thrombosis are the improper regulation of the coagulation cascade, and/or excessive platelet activation: J [67]. 9 Rivaroxaban itself had, as one of a class of factor Xa inhibitors, been disclosed and claimed in an earlier Bayer patent application, being WO 919: J [10].

226 Patent 10 The invention described in the 226 Patent is said to relate to a process for the preparation of a solid, orally administrable pharmaceutical composition, comprising 5-chloro-N-({(5S)-2-oxo-3-[4-(3-oxo-4-morpholinyl)-phenyl]-1,3-oxazolidin-5-yl}-methyl)-2-thiophenecarboxamide (compound I) in hydrophilized form, and "its" use [i.e. use of the pharmaceutical composition comprising compound I in hydrophilized form] for the prophylaxis and/or treatment of thromboembolic diseases. Compound I is now known as rivaroxaban. It is a direct factor Xa inhibitor. 11 The 226 Patent claims an earliest priority date of 27 November 2003 (First Priority Date). 12 The specification of the 226 Patent explains that rivaroxaban is a low molecular weight, orally administrable inhibitor of blood clotting factor Xa, which can be employed for the prophylaxis and/or treatment of various thromboembolic diseases. The 226 Patent does not otherwise describe rivaroxaban in detail (save for some references to its water solubility and bioavailability), instead referring the reader to WO 919 and incorporating the disclosure of WO 919 by way of reference (page 1, lines 6-9). 13 Claim 1 provides: Process for the preparation of a solid, orally administrable pharmaceutical composition comprising [compound I] in hydrophilized form, the process comprising: (a) preparing granules comprising the active compound (I) in hydrophilized form by moist granulation; and (b) converting the granules into the pharmaceutical composition, if appropriate with addition of pharmaceutically suitable additives. 14 Independent claims 8 and 20 are respectively for a solid, orally administrable pharmaceutical composition comprising [compound I] in hydrophilized form and the use of [compound I] in hydrophilized form for preparing a medicament for the prophylaxis and/or treatment of a thromboembolic disease.

613 Patent 15 The 613 Patent relates to a method of treatment of thromboembolic disorders by once daily administration of a rapid-release tablet comprising rivaroxaban. 16 The 613 Patent claims an earliest priority date of 31 January 2005 (Second Priority Date). 17 The background of the invention in the 613 Patent identifies the blood coagulation mechanism, the blood coagulation cascade, the fact that there are numerous factors and intrinsic and extrinsic pathways, as well as the fact that activated serine protease Xa cleaves prothrombin to thrombin, allowing a further step in the cascade to occur. Following a discussion of haemostasis, coagulation and thromboembolic disorders, there is then a discussion of the existing anticoagulants (heparin, low molecular weight heparin and warfarin) and their "often severe disadvantages" which make the effective treatment of thromboembolic disorders "very difficult and unsatisfactory": J [127]. 18 The specification then describes a novel therapeutic approach for the treatment of thromboembolic disorders which aims to inhibit factor Xa (page 2, lines 20-29). The specification identifies that oral application is the preferable route of administration, and that once daily oral application is preferred for convenience and compliance. However, it says this can be "difficult to achieve depending on the specific behaviour and properties of the drug substance, especially its plasma concentration half-life": J [130]. 19 The 613 Patent then states at page 3, lines 5-10: When the drug substance is applied in no more than a therapeutically effective amount, which is usually preferred in order to minimize the exposure of the patient with that drug substance in order to avoid potential side effects, the drug must be given approximately every half live [sic] (see for example: Malcolm Rowland, Thomas N. Tozer, in "Clinical Pharmacokinetics, Concepts and Applications", 3rd edition, Lea and Febiger, Philadelphia 1995, pp 83). 20 Following this, at page 3, line 17, the invention in the 613 Patent is identified: Surprisingly, it has now been found in patients at frequent medication that once daily oral administration of a direct factor Xa inhibitor with a plasma concentration half life time of 10 hours or less demonstrated efficacy when compared to standard therapy and at the same time was as effective as after twice daily (bid) administration. 21 From page 3a, line 18, a preferred embodiment of the invention is said to be: In a preferred embodiment, the present invention relates to [rivaroxaban], a low molecular weight, orally administrable direct inhibitor of blood clotting factor Xa (see WO-A 01/47919, whose disclosure is hereby included by way of reference) as the active ingredient. 22 On page 4, lines 6-9, the plasma concentration half-life of rivaroxaban is described: For [rivaroxaban] a plasma concentration half life of 4-6 hours has been demonstrated at steady state in humans in a multiple dose escalation study (D. Kubitza et al, … 23 Claim 1 provides: Use of [compound I] in the manufacture of a medicament which is a rapid-release tablet for the treatment of a thromboembolic disorder, wherein the tablet is administered once daily for at least five consecutive days, and wherein the [compound I] has a plasma concentration half life of 10 hours or less when orally administered to a human patient. 24 Independent claim 3 is a method for the treatment of a thromboembolic disorder comprising administration of a rapid-release tablet comprising compound I with the characteristics described in claim 1.

RELEVANT LEGISLATION 25 Sections 7(2) and 7(3) of the Patents Act set out the condition upon which an invention would not be taken to involve an inventive step when compared with the prior art base. At the relevant times, these provisions stated as follows: (2) For the purposes of this Act, an invention is to be taken to involve an inventive step when compared with the prior art base unless the invention would have been obvious to a person skilled in the relevant art in the light of the common general knowledge as it existed in the patent area before the priority date of the relevant claim, whether that knowledge is considered separately or together with the information mentioned in subsection (3). (3) The information for the purposes of subsection (2) is: (a) any single piece of prior art information; or (b) a combination of any 2 or more pieces of prior art information; being information that the skilled person mentioned in subsection (2) could, before the priority date of the relevant claim, be reasonably expected to have ascertained, understood, regarded as relevant and, in the case of information mentioned in paragraph (b), combined as mentioned in that paragraph.

Reasons of the primary judge 26 By this ground, Sandoz asserts an error by the primary judge in finding that a person skilled in the relevant art could not be reasonably expected to have ascertained WO 919 within the meaning of s 7(3) of the Patents Act: J [496]. As to s 7(3), the primary judge identified at J [425] that: There are three requirements for purported prior art to qualify under s 7(3) (as in effect at the relevant time): that a person skilled in the art could be reasonably expected to have: 1) ascertained; 2) understood; and 3) regarded as relevant the information. 27 As to the meaning of "ascertained" in s 7(3), which is the focus of ground 1, the primary judge stated as follows at J [426]: First, "ascertained" simply means "discovered" or "found out": Lockwood No 2 at [132] (per Gummow, Hayne, Callinan, Heydon and Crennan JJ). As Beach J explained in Sequenom Inc v Ariosa Diagnostics Inc (2019) 143 IPR 24 at [598], a "document could be ascertained if it was published in such a manner or form that it could reasonably have been expected to be found by a person skilled in the art". It is essential to both inventive step cases that the s 7(3) document would be ascertained by the hypothetical person skilled in the art. (Emphasis original.) 28 As to the meaning of "reasonably expected" in s 7(3), the primary judge stated at J [429]-[430]: Finally, with respect to the qualifying phrase "reasonably expected", Heerey, Kiefel and Bennett JJ observed in Commissioner of Patents v Emperor Sports Pty Ltd (2006) 67 IPR 488 at [31]: Section 7(3) does not assume an ascertainability by any and all skilled persons, of whatever description, of all publicly available prior art documents anywhere in the world. Nor does it assume that the skilled person has found the document in question, so that the only question is whether he or she has understood it and regarded it as relevant. Such a construction ignores the elements of expectation and reasonableness, as applied to the particular skilled person. Further, Beach J observed in Sequenom at [599] that: The expectation as to what the skilled addressee "could" ascertain is a reasonable expectation, not a fanciful one. That reasonable expectation is to be assessed in light of the characteristics of the relevant hypothetical skilled person and the particular problem faced, the overcoming of which is said to involve an inventive step. The inquiry is hypothetical. 29 Relevantly to this appeal, the primary judge described a series of tasks given to the expert retained by Sandoz (Professor Roberts) including by reference to search results obtained by another person retained by Sandoz (Mr Cruise), being evidence adduced by Sandoz at trial: 433 Professor Roberts (wearing his pharmaceutical scientist hat) was asked by Sandoz's instructors to undertake the following hypothetical task: Ashurst asked me to assume: (a) I am a member of a pharmaceutical product development team … at the Relevant Date; (b) the product development team is seeking to develop a product for the treatment of thromboembolic diseases containing a Factor Xa inhibitor as the active pharmaceutical ingredient; and (c) I am tasked with developing a formulation for an immediate release solid oral dosage form that is suitable to take forward through clinical trials ([First] Hypothetical Task). (Italics added.) 434 Bayer took issue with the italicised parts of the First Hypothetical Task. 435 Early in his first affidavit, Professor Roberts described the literature searches he would undertake to find any information about a target drug in order to supplement the information in the product profile that he had been given by a research team. Typically, Professor Roberts would conduct searches for the target drug by its chemical name, code name and/or any other reference names. He would conduct his searches on a number of databases, including SciFinder (which includes Chemical Abstracts), which he described as the primary database he searched, the Merck Index, Martindale, the USPTO patent database, and search engines such as Google and Yahoo to identify any publicly available records including press releases or annual reports. 436 Professor Roberts considered that the First Hypothetical Task was directed towards a formulator with a pharmaceutical science skill set (like himself) rather than a straight formulator, such as Professor Poli. 437 Professor Roberts' evidence was that he would begin the First Hypothetical Task by conducting literature searches of the kind just described for a factor Xa inhibitor. He would also conduct broader searches using the terms "anti-thrombotic" and "anti-clotting" in case there were publications or documents about products in development where it had not been disclosed that the relevant product worked via factor Xa inhibition. 438 No search results in relation to the First Hypothetical Task were annexed to Professor Roberts' affidavit. After his description of his likely searches, he was provided with a copy of WO 919, and asked what information, if any, WO 919 disclosed that he considered relevant to the First Hypothetical Task. Professor Roberts then spent six pages of his affidavit discussing the disclosure of WO 919 and his choice of a preferred compound. Bayer submits that the irresistible inference to any reasonable person undertaking the First Hypothetical Task would be that they were given WO 919 because it was relevant to the task. 439 Following his discussion of WO 919, Professor Roberts was provided with a 71-page "redacted" spreadsheet of search results (containing limited bibliographic information) and asked to assume that the spreadsheet reflected the results of a Chemical Abstracts search for patent literature relating to factor Xa inhibitors. Professor Roberts' evidence was that he did not ask for the search to be limited in that manner, and that he would not normally limit a search to only patent literature. Professor Roberts was never given the results for the searches that he identified that he would have carried out across all the databases he identified. Even if his search was limited to the Chemical Abstracts database alone, his typical approach would include searching peer-reviewed literature, patents, citations and general search engines, not just patent literature. Professor Roberts agreed that he did not instruct Ashurst to limit the search to a single database (Chemical Abstracts) and to further limit the search on that database to patents only, thereby excluding these other resources which, in the normal course, he would examine. The obvious inference is that it was Ashurst, not Professor Roberts, that so limited the search results in this manner. 440 Mr Cruise conducted the search of Chemical Abstracts limited to only patents using the search term "factor Xa inhibitors", not the term "factor Xa inhibitor" proposed by Professor Roberts. Mr Cruise's search for "factor Xa inhibitors" produced about 177 results and 1,200 pages of results. Mr Cruise agreed that it was important to his search whether "factor Xa inhibitor" or "factor Xa inhibitors" was used, and that if the search was not limited to patent literature, the results would have been "significantly greater" than 177. Mr Cruise's 1,200 pages of results were reduced (presumably by the instructors) to a 71-page spreadsheet which was provided to Professor Roberts. 441 Professor Roberts only had a limited time (of about two hours) to review the spreadsheet of results. Following a "rapid scan" of the results, he classified the results as "top priority", "yes", "potentially relevant" and "no", which he explained reflected his "initial view". Professor Roberts marked WO 919 as a "top priority" document. Professor Roberts' evidence was that he would have proceeded to retrieve the full document for each of the results he identified as "top priority", "yes" or "potentially relevant" (151 documents). Professor Roberts observed that his initial ranking was just his initial view and not absolute and that he may change his mind having reviewed a paper "because sometimes abstracts don't tell you the full story". Professor Roberts was not provided with copies of any of the 151 documents, other than WO 919 which he had been given earlier. Before making a final selection, Professor Roberts said that he would have carried out further searching over additional databases, including a citation search. (Emphasis original.) 30 The complaints by Bayer about the hypothetical task are described at J [443]-[444] and [447]. The submission by Bayer at trial was that Professor Roberts was given a starting point which included aspects of the invention, and that the formulation of the task amounted to leading Professor Roberts to the invention (which complaint it repeats in this appeal). However, the complaint by Bayer was not that factor Xa inhibitors were not known as at the priority dates or that the idea that a factor Xa inhibitor could be a solution was not part of the common general knowledge as at the priority dates. 31 After addressing the parties' submissions and certain authorities, the primary judge's reasons in relation to this issue appear at J [472]-[495]. 32 Although the primary judge recognised at J [472] that the absence of evidence concerning the nature of any hypothetical literature search is not fatal to the correct application of s 7(3) and that each case depends on its facts (which statements we accept as being correct), her Honour then stated at J [473]: Rather than rely on evidence as to each step of the search process, Sandoz seeks to take a short cut, relying on the absence of search evidence in [Merck Sharp & Dohme Corporation v Wyeth LLC (No 3) (2020) 155 IPR 1; [2020] FCA 1477] and [Otsuka Pharmaceutical Co Ltd v Generic Health Pty Ltd (No 4) (2015) 113 IPR 191; [2015] FCA 634], to jump straight to the assertion that the person skilled in the art would ascertain WO 919, without establishing the foundation for such an assertion via evidence of the search process the person skilled in the art would undertake 33 In forming the critical view that the skilled person could not be reasonably expected to have ascertained WO 919, the primary judge made the following findings: (1) WO 919 is a patent specification, and the patent records database (within Chemical Abstracts) would have been one of the databases searched by the person skilled in the art: J [475], [484]; (2) the hypothetical task given to Professor Roberts was narrower than the problem found in the common general knowledge and skewed to finding the answer: J [478]; and (3) while the evidence showed that factor Xa inhibitors were something that would have been focused on by a person skilled in the art, there was also another inhibitor of interest, being thrombin inhibitors, which would have been the subject of a search by a person skilled in the art: J [478]-[481]. 34 It follows from these findings, and the references in the primary judge's reasons to the submissions made by Bayer at trial, that the submission that the hypothetical task was narrower than the problem found in the common general knowledge was directed to the fact that, while it was recognised as at both priority dates that a factor Xa inhibitor could be a solution to the problem and would be the subject of searches by a skilled person, there were other candidates that could also be a solution. On appeal, Bayer accepted that factor Xa inhibitors would be amongst the searches that would be carried out. 35 The primary judge also made the following observations and findings in respect of the literature search conducted by Mr Cruise: (1) Mr Cruise's search was undertaken contrary to Professor Roberts' instructions as to his usual search practice. In short, instead of searching all the databases that Professor Roberts said he would typically search, using the terms Professor Roberts suggested, the search was run for "factor Xa inhibitors" across one database within Chemical Abstracts - the patent records: J [483]; (2) the patent records database (referred to above) would not have been the only database searched. Searches of the other databases would have produced peer-reviewed articles and review articles. In practical terms, seven other databases, and two other search terms ("anti-thrombotic" and "anti-clotting"), were not included: J [484]; (3) the limited search conducted by Mr Cruise produced 1,200 pages of results, but the raw results were not given to Professor Roberts. Instead, the results were "winnowed down" from 1,200 pages to a 71-page spreadsheet with no details provided as to the winnowing process: J [485]; (4) the omission of the other databases and sources of information from the search vastly reduced the volume of the search results that could be expected had the search contemplated by Professor Roberts been carried out. The volume of results was also likely further reduced to a significant, but unascertainable, extent by only focusing on factor Xa inhibitors: J [486]. 36 Implicit in this last finding is acceptance by the primary judge that, had the typical search practice undertaken by Professor Roberts been carried out, the search results would have included WO 919 albeit as part of a larger volume of search results. Before proceeding to consider the search evidence put before the primary judge in more detail, and her Honour's reasons given by reference to that evidence, we take a small detour to consider two issues which were raised by Bayer on this appeal. 37 First, Bayer seeks to introduce the notion that ground one on appeal requires consideration of whether the skilled person could reasonably be expected to have ascertained and regarded as relevant WO 919 on the basis that the primary judge's conclusion at J [477] (first sentence) treated the requirements of ascertainment and relevance as related, and as such her Honour found that the document was neither ascertained nor regarded as relevant within the meaning of s 7(3). On that basis, Bayer submits that there was no evidence that, separately to being armed with it and asked directly as to its relevance, a person skilled in the art would regard WO 919 as relevant. 38 At J [477], the primary judge stated: I do not consider that the evidence establishes that the person skilled in the art could reasonably be expected to have ascertained, understood and regarded WO 919 as relevant so as to satisfy the s 7(3) requirements. The evidence demonstrates that WO 919 could have been returned by a search of relevant patent databases but fails to demonstrate that it would have been reasonable to expect a person skilled in the art to have ascertained it. 39 We reject Bayer's submission as it is plain from the reasons as a whole that the primary judge decided that s 7(3) was not satisfied in relation to WO 919 because it had not been ascertained within the meaning of that section. We refer in particular to J [477] (second sentence), J [489] and J [496]. 40 Second, Bayer sought to impugn the evidence of Professor Roberts. 41 That was because, at J [487], the primary judge noted that, prior to being given the patent results spreadsheet, a copy of WO 919 had been provided to Professor Roberts, and he was asked to describe what information, if any, it disclosed that he considered relevant to the First Hypothetical Task. At J [488], the primary judge stated that: After his consideration of WO 919, Professor Roberts then reviewed the patent results spreadsheet and, perhaps unsurprisingly, he identified WO 919 as one of the nine documents which appeared to him most relevant and most likely to contain information of use for the First Hypothetical Task. (Emphasis added.) 42 The emphasised words suggest that the primary judge considered that Professor Roberts' selection of WO 919 from the list in the spreadsheet was influenced by being given WO 919 prior to being given the spreadsheet. It is difficult to understand the emphasised words any other way. However, her Honour then went on to state in J [488] that, "[y]es, Professor Roberts was able to justify his selection of WO 919 as a 'top priority' document from the results in the spreadsheet". With respect, that statement is correct. When one has regard to the spreadsheet and to the transcript of the cross-examination of Professor Roberts at trial, it is apparent that there were no indicators in the spreadsheet itself that WO 919 had been included in the search results (as it was listed under a different title than that contained on the WO 919 document itself) or that Professor Roberts selected that entry in the spreadsheet as "top priority" because he knew that it was WO 919. 43 It follows that, as Sandoz submits, and taking into account the primary judge's acceptance that Professor Roberts was able to justify his selection of WO 919 from the spreadsheet, there is no finding by the primary judge that Professor Roberts acted with the benefit of hindsight (and Bayer did not submit otherwise). 44 Notwithstanding this, Bayer submits that, as Professor Roberts was provided with WO 919 prior to looking at the search results, this Court should apply the reasoning of the Full Court in Minnesota Mining & Manufacturing Company v Tyco Electronics Pty Ltd (2002) 56 IPR 248; [2002] FCAFC 315. In reliance on that decision, Bayer submits that to give WO 919 to Professor Roberts was hardly calculated to result in objective evidence as to what the hypothetical uninventive skilled addressee would have ascertained in the factual environment of the common general knowledge alone. 45 During the hearing of the appeal, Bayer also referred to an answer given by Professor Roberts during his cross-examination concerning his selection of the entry which was WO 919 in the spreadsheet: MR SHAVIN: Right. Because I was suggesting to you that at least subconsciously, Professor, having looked very carefully at the 919 patent - - - PROF ROBERTS: Yes. MR SHAVIN: - - - and the compound identified on page 15 - - - PROF ROBERTS: Yes. MR SHAVIN: - - - you would have been familiar with the chemical structure of that compound, and you would recognise that when you saw it identified in the entry here? PROF ROBERTS: I didn't at the time, but I should have. And I recognise there could have been some bias that's there intuitively. But I did not recognise it at the time. 46 Based on that evidence, Bayer submits that the evidence of Professor Roberts should therefore be "given very little weight". However, Bayer did not identify any submission made to the primary judge of this kind, and so it is not clear what Bayer proposes that the Full Court should make of this submission in circumstances where the primary judge did not find that Professor Roberts acted with the benefit of hindsight and Bayer does not assert an error by the primary judge through any notice of contention. We therefore propose to do nothing with it. 47 After canvassing the circumstances in which Professor Roberts selected WO 919 from the spreadsheet of results, her Honour found at J [488] that the question of whether Professor Roberts would have selected WO 919 if he was provided with the full suite of results from searches undertaken across all his suggested databases and with broader search terms than just "factor Xa inhibitors" could not be answered on the evidence before her. The primary judge identified that there was no evidence as to how Professor Roberts would prioritise between different peer-reviewed journal articles, how he would prioritise as between journal articles and patents, and whether he might prioritise other antithrombotics over factor Xa inhibitors on the basis of the peer-reviewed articles located: J [488]. At J [489], the primary judge reiterated that the evidence did not establish that it would be reasonable for the person skilled in the art undertaking a search of the kind described by Professor Roberts but not restricted to factor Xa inhibitors to have found WO 919.

Consideration 48 The central issue in relation to ground 1 is whether the primary judge applied the correct standard for the purpose of determining whether a person skilled in the art could be reasonably expected to have ascertained WO 919. 49 The standard imposed by s 7(3) does not require proof that the hypothetical skilled person would ascertain the document. Rather, it requires proof sufficient to demonstrate a reasonable expectation that the skilled person would do so, and that is on the balance of probabilities. As cited by Besanko J in Aspirating IP Ltd v Vision Systems Limited (2010) 88 IPR 52; [2010] FCA 1061 at [457], the High Court said in Commissioner of Taxation v Peabody (1994) 181 CLR 359 at 385: A reasonable expectation requires more than a possibility. It involves a prediction as to events which would have taken place if the relevant scheme had not been entered into or carried out and the prediction must be sufficiently reliable for it to be regarded as reasonable. (Footnotes omitted.) 50 Further, as observed by Burley J in Pharmacia LLC v Juno Pharmaceuticals Pty Ltd (2022) 165 IPR 200; [2022] FCA 92 at [268]: Whether or not the prior art information can be reasonably expected to have been found depends on what the skilled person is likely to have done when faced with a problem similar to that which the patentee claims to have solved with the claimed invention. 51 As to what the skilled person is likely to have done on the facts of this case, the evidence established, and the primary judge accepted, that the person skilled in the art would have conducted searches of (inter alia) the patent database using search terms including factor Xa inhibitors. Contrary to the submission by Bayer, it therefore does not matter that the skilled person would have also conducted searches using additional search terms. 52 The evidence also established, and the primary judge accepted, that a search of the patent database would have returned results which included (inter alia) WO 919. Notably, the spreadsheet provided to Professor Roberts contained the full results of the search of the patent database albeit that not all of the information associated with each result was included. 53 As her Honour accepted implicitly, it is reasonable to expect that the skilled person would have reviewed the search results. As to the consequences of that review, an independent expert witness, Professor Roberts, identified WO 919 as a "top priority" result from the spreadsheet, and he was able to justify that identification with no finding that he did so because of hindsight. This was done as part of the hypothetical exercise which demonstrated what the skilled person was likely to have done when reviewing those search results. 54 Bayer complains that Professor Roberts was not provided with all of the documents which he identified from the spreadsheet as being top priority, being 190 documents of a total of 218 documents, such that it could not be said in those circumstances that WO 919 was ascertained because it was not demonstrated that, had he reviewed all of those additional documents, he would have selected WO 919. 55 However, we do not accept this submission. It is not necessary for evidence to be adduced that the skilled person would prefer, prioritise or select the information in question (here, WO 919) over all other information which they could be reasonably expected to have discovered or found out, including other information which was not adduced into evidence: see Otsuka Pharmaceutical Co Ltd v Generic Health Pty Ltd (No 4) (2015) 113 IPR 191; [2015] FCA 634 at [415]-[417] (Yates J). 56 Nor is it necessary in every case "that an artificial ex post facto literature search [be] carried out at the time of the conduct of patent litigation in order to establish the "ascertained" aspect of s 7(3)": see Hanwha Solutions Corp v REC Solar Pte Ltd [2023] FCA 1017 (Burley J) at [367]. 57 Despite accepting that such an approach was not necessary in all cases, the primary judge required, in effect, that such a search be performed, and that Professor Roberts select WO 919 from the broader array of search results. This is apparent from, at the least, her Honour's finding at J [488]. 58 This has the consequence in this case that, once the primary judge accepted that the search results which were adduced into evidence are a subset of the searches which would have been taken by the skilled person, it is not to the point that additional searches might have been performed (including by reference to additional search terms), or the skilled person could also be expected to have found additional documents or information, or the skilled person would have been required to review such additional documents or information. 59 This is especially as Bayer led no evidence to suggest that a reasonable search of a patent database would not have found WO 919, or that WO 919 would not have been identified out of a broader search involving other databases or search terms, or that WO 919 would not have been identified as being relevant upon being reviewed. 60 It follows that Sandoz established on the balance of probabilities that WO 919 could be reasonably expected to have been ascertained by the skilled person as required by s 7(3) of the Patents Act, and that the primary judge erred in finding otherwise. 61 For these reasons, we are satisfied that ground one has been established.

GROUND TWO 62 This ground asserts an error by the primary judge in finding that the inventions claimed in each of the 226 Patent and the 613 Patent involved an inventive step in the light of the common general knowledge together with WO 919.

WO 919 63 The primary judge addressed the disclosure of rivaroxaban in WO 919 at J [529]-[539]. As observed by the primary judge at J [530]: At page 3 [of the 226 Patent], the specification states that recently a novel therapeutic approach for the treatment and prophylaxis of thromboembolic disorders has been described. This novel therapeutic approach aims to inhibit factor Xa. WO 919 also discloses that it is an object of the invention in the patent to provide novel substances with a wide therapeutic window for controlling disorders. It states that the compounds which are disclosed should be suitable for more efficient prophylaxis and treatment of a range of thromboembolic disorders, and that another object is to provide novel anticoagulants which inhibit factor Xa with increased selectivity. 64 WO 919 discloses a large class of novel oxazolidinone derivatives of the general formula (I). WO 919 states that "very particular preference" is given to the compound that is rivaroxaban. The primary judge noted that it is the only individual compound "specifically called out in this way" and which is claimed individually in claim 7: J [531]-[532]. 65 Thus, WO 919 singled out rivaroxaban and disclosed that compound (among others) as a selective factor Xa inhibitor useful for the treatment of thromboembolic diseases because its greater therapeutic range, when compared to conventional preparations, means that there is a lower risk of bleeding for the patient, easier adjustment for the physician and a more rapid onset of action. As will be seen, both the 226 Patent and the 613 Patent contained admissions to that effect in relation to the disclosure of WO 919, as well as identifying rivaroxaban as low molecular weight and orally administrable. 66 WO 919 then discloses a series of examples which can be summarised as in vitro blood work in human blood samples and in vivo tests in rats. The first group of these evaluate the physiological activity of some of the compounds, including rivaroxaban, through the use of in vitro methods to determine factor Xa inhibition, selectivity and anticoagulant action, as well as in vivo testing of antithrombotic activity in three rat models. The second group provide methods for synthesis of the compounds. 67 In the first group of examples, there is a report on the results of testing in relation to a number of the compounds, including rivaroxaban (listed as Example 44). Based on oral administration to an unspecified number of anaesthetised fasted male rats using a pharyngeal tube, rivaroxaban is said to have an ED50 (a measure of the dose that provides the desired pharmacological effect in 50% of a study population) of 3 mg/kg. Rivaroxaban is one of only two compounds for which results from oral administration to rats are reported. 68 In the second group of examples, synthesis and structure information and other data is given for compounds, including rivaroxaban, such as its melting point (indicating the drug is pure and crystalline), its mass, its optical rotation and its IC50. According to Professor Roberts, the IC50 for rivaroxaban is very low and, assuming that the IC refers to specific factor Xa inhibition, this indicates the preferred compound is quite potent.

The Patents 69 In part 5.1 of the judgment, the primary judge addressed the 226 Patent. At J [108], her Honour identified that the compound described in the 226 Patent is now known as rivaroxaban and it is a direct factor Xa inhibitor. At J [109], the primary judge referred to the fact that the specification of the 226 Patent explains that rivaroxaban is a low molecular weight, orally administrable inhibitor of blood clotting factor Xa, which can be employed for the prophylaxis and/or treatment of various thromboembolic diseases. At J [110], the primary judge observed that the 226 Patent does not otherwise describe rivaroxaban in detail, instead referring the reader to WO 919 and incorporating the disclosure of WO 919 by reference (page 1 lines 6-9). 70 In part 5.2 of the judgment, the primary judge addressed the 613 Patent which includes, at J [133], a reference to a preferred embodiment of the invention on page 3a lines 18­22 as follows: In a preferred embodiment, the present invention relates to [rivaroxaban], a low molecular weight, orally administrable direct inhibitor of blood clotting factor Xa (see WO-A 01/47919, whose disclosure is hereby included by way of reference) as the active ingredient. 71 Neither the 226 Patent nor the 613 Patent assert any invention in the choice of active pharmaceutical ingredient. They provide for the preparation of a pharmaceutical composition and method of treatment comprising rivaroxaban, respectively. Rivaroxaban is an essential integer in each of the claims of the patents. The 226 Patent asserts a surprising result in a "special treatment" of rivaroxaban to improve its bioavailability, and the 613 Patent asserts an inventive step in administering rivaroxaban once a day, where its half-life was said to suggest more frequent dosing. However, the primary judge did not find that the formulation adopted in the 226 Patent or the once-daily dosage regime adopted in the 613 Patent involved any inventive step.

Common General Knowledge 72 The technical primer filed before the primary judge set out the agreed common general knowledge of those skilled in the art as at the First Priority Date and Second Priority Date. The primary judge considered the state of the common general knowledge based on the primer and expert evidence (given in the form of the joint expert reports and orally during joint session). In particular, her Honour made several findings concerning the stages of drug development which are central to her Honour's reasons for concluding that the 226 Patent and the 613 Patent are valid. Those findings included the following: (1) It was common ground that the drug development process can be complex, time consuming, expensive and iterative. Statistically, it is not uncommon for drug development projects to fail if, for example, the drug product is considered unsafe, lacking efficacy or not commercially viable: J [383]. (2) The experts agreed that the research and development process in relation to a new chemical entity (NCE) for small molecules is different from the development of new, improved or generic formulations of pharmaceutical products that have already obtained marketing authorisation. A starting point for generic drug formulation development includes knowing the active pharmaceutical ingredient (including form), dose, dosage form and route of administration as well as the safety and efficacy information. For an NCE, there is no reference product that provides confidence that the drug can be safe and efficacious: J [385]. (3) The research and development process in relation to an NCE is also different to the development process for extending or "repurposing" a known drug for a new indication or for a drug that is a member of a class of drugs where some members are already on the market. Professor Evans contrasted the risks inherent in trying an NCE for the first time in humans, with a drug which is the third or fourth generation of a chemical class. Less unusual or unpredictable toxicities are likely when another member of a known class of drugs was tested in humans for the first time: J [386]. (4) The stages of the drug development process for an NCE (and first in its class) identified by the experts include the following: J [388]: (a) discovery or identification of the NCE (including its chemical structure) and selection and optimization of a lead compound or compounds; (b) pre-clinical collection of data as to the chemistry, pharmacology, toxicology, physicochemical properties, pharmacokinetics and biostatistics of the new compound; (c) designing and conducting comprehensive in silico and in vitro studies to provide an indication as to the drug's potential pharmacokinetics and biological activity in humans; (d) Phase I clinical studies, also known as "first-in-human" studies, in which the drug is given to a small number of typically healthy participants; (e) Phase II clinical studies, whereby the dose or dosage regimen is administered to a small number of participants, typically patients who have the relevant condition being treated; (f) Phase III clinical studies, in which the drug is administered on a large scale to patients; and (g) Phase IV studies, which are ongoing and in which post-approval marketing surveillance is conducted to monitor the use of the product in the market. (5) Each step of the drug development program provides new data that can halt the program, lead to revision of the program, modify the development of the program or provide confidence to pursue a predetermined plan: J [389]. 73 The primary judge also made extensive findings concerning the likelihood of success of the drug development process, both generally and by reference to the compounds disclosed by WO 919. 74 For example, the primary judge found that: 417 It is not uncommon for drug development projects to fail, for example, if the drug product is considered unsafe, lacking efficacy or not commercially viable. It is also very difficult, if not impossible, to identify from a chemical structure (without more) whether a product will succeed through to marketing approval. 418 The majority of lead drug candidates that make it to pre-clinical studies never advance to clinical trials. Most drugs that enter clinical trials do not progress to receive regulatory approval. According to Venkatesh S and Lipper R A, "Role of the Development Scientist in Compound Lead Selection and Optimization" (2000) 89 (February) Journal of Pharmaceutical Sciences 145, 146, which was cited by the experts in JER1, this attrition rate is about 90%. That same article notes that the historical success rate for compounds going from pre-clinical trials into Phase I trials in humans is around 40%. 419 Professor Roberts agreed with the figure of 90% across the whole drug development process, observing that only about 30% of compounds make it from Phase I into Phase II clinical trials, and around 30% of compounds make it from Phase II into Phase III studies. According to Professor Roberts, when these chances of success are multiplied together, that gives a 90% chance that a compound entering Phase I trials will not proceed to market. Professor Roberts said that this is because all sorts of things could go wrong along the steps of the development process. 420 To give a drug the best chance of getting approval, it may be necessary in some cases for the medicinal chemist to change the structure of the compound along the way, change the dosage form, and/or change the formulation, such as to make changes as between immediate release, sustained release or delayed release dosage forms. Most of these potential problems cannot be anticipated at the time a compound candidate is selected as the lead compound. 421 These risks of failure are particularly acute for NCEs, for many of the reasons described above, particularly because the development team is starting with a "blank slate" compared to developing a generic drug, which means there is simply no way of anticipating what will be encountered in the development process. 422 The substantial risks and uncertainties associated with drug development are particularly problematic in the high risk field of thrombosis involving a novel, first in class drug. For example, as Professor Roberts recognised, there are "so many failures" and "no one knew what was going to work". Further, Professor Baker agreed that the refusal of the FDA panel to recommend the use of ximelagatran in 2004 confirmed that the task of developing an anticoagulant drug that was a selective inhibitor, in solid oral dosage form and that was going to be safe and effective, was full of risk. 75 Further and in relation to the evidence given by the experts, the primary judge also observed that: 569 The experts agreed in the joint session that the drug development process is not linear. There are uncertainties inherent in the drug development process and problems or issues may arise at any stage during the process requiring a change in project direction. The experts observed that whilst the steps of the drug development program may at first glance appear to be routine, as they are a particular series of tests that must be undertaken for a drug candidate to proceed to the next step, the reality is that the results of the tests are far from routine as each step of the drug development program can provide new and unexpected data that can halt the program, lead to the revision of the program, modify the development plan or provide confidence to pursue a predetermined plan. 570 Professor Evans observed that none of the compounds in WO 919 had ever been administered to a human being. He was concerned that there may be other adverse events or toxicities that may make them far worse than warfarin in terms of their effects on people. Professor Evans gave the examples of a hypersensitivity reaction or gastrointestinal erosion causing ulcers. A drug that might cause gastrointestinal erosion that also blocks coagulation of blood would cause disastrous gastrointestinal bleeds. Professor Roberts agreed that there were issues of potential toxicity which would have to be worked out. Professor Evans further stated that it was not possible to predict before testing that these risks would eventuate, and that the only way to find out was to conduct the usual process of selecting a compound and conducting the appropriate tests, a process which might take three to four years. 571 As to the rat tests in the examples in WO 919, Professor Evans considered that, at best, they showed that "plausibly" there was absorption of the drug in animals after oral administration. He noted that the drugs were administered by a tube directly into the rat stomach, and that WO 919 gave no information as to what the drug was dissolved in, how much was absorbed and whether there was any toxicity to the rats' stomachs. In discussion with Professor Roberts, both experts speculated that the drug may have been dissolved in dimethyl sulfoxide, a chemical not administered to humans. Professor Evans considered that the limited information derived from the rat tests was not sufficient to conclude that the drug was suitable for oral administration in humans. 572 In the context of discussing the stopped clinical trials of an orally active agent, DPC-906, reported in the Chest paper, Professor Baker agreed that the task of developing a product that was a solid oral dosage form, and that was a selective inhibitor that was going to be both safe and effective, was "full of risk". 573 Professor Roberts acknowledged that even when a lead compound has been selected, there is "absolutely" an enormous amount that can go wrong and, if it did, the consequence for a human with this class of compounds can be "catastrophic". These concerns were particularly acute with respect to direct factor inhibitors as the community were aware in 2005 of the FDA's refusal to recommend the use of ximelagatran due to concerns regarding liver toxicity, as explained above at [350].

Selection of rivaroxaban 76 With that background, the primary judge found that: (1) the experts agreed that WO 919 would be of interest and value to a drug development team looking at developing a new antithrombotic drug and that, on reading WO 919, the skilled reader would focus on the compounds within the general formula (I) described by the patentee as "very preferred". In particular, they would focus on Example 44 (rivaroxaban) because it was listed as a particular chemical, discussed in examples and the only compound the subject of a claim: J [541]; (2) the skilled person reading WO 919 is likely to select rivaroxaban as a lead candidate, if not the lead candidate, to take into further drug development work. The primary judge considered that this would involve the full suite of pre-clinical tests that Professors Roberts, Evans and Baker said they would carry out in order to consider whether Example 44 was suitable to take into Phase I, first-in-human trials: J [594]. 77 The second of these findings is of particular importance. That is because the selection of rivaroxaban would be made by the skilled person to take forward into further drug development work, notwithstanding the risks associated with drug development in the "high risk field of thrombosis involving a novel, first in class drug" (to use the primary judge's words).

Taking of routine steps with expectation of success 78 Earlier in the judgment, the primary judge discussed the relevant principles concerning inventive step, stating: 292 In Aktiebolaget Hässle v Alphapharm Pty Ltd (2002) 212 CLR 411 at [50] (AB Hässle), Gleeson CJ, Gaudron, Gummow and Hayne JJ referred to the formulation of the relevant test by Aickin J in Wellcome Foundation Ltd v VR Laboratories (Aust) Pty Ltd (1981) 148 CLR 262 at 286: The test is whether the hypothetical addressee faced with the same problem would have taken as a matter of routine whatever steps might have led from the prior art to the invention, whether they be the steps of the inventor or not. 293 Their Honours observed at [51]: What Aickin J had in mind as "routine" appears from an earlier passage in his judgment in which he was discussing the question whether evidence of the steps taken by the patentee was relevant and therefore admissible in a revocation action. His Honour said: "Evidence of what he did by way of experiment may be another matter. It might show that the experiments devised for the purpose were part of an inventive step. Alternatively it might show that the experiments were of a routine character which the uninventive worker in the field would try as a matter of course. The latter could be relevant though not decisive in every case. It may be that the perception of the true nature of the problem was the inventive step which, once taken, revealed that straightforward experiments will provide the solution. It will always be necessary to distinguish between experiments leading to an invention and subsequent experiments for checking and testing the product or process the subject of the invention." (Emphasis added in AB Hässle.) 294 Their Honours then sought to unravel the "distinct strands of thought" in Aickin J's reasoning, concluding at [53] that the correct approach is the reformulated "Cripps questions" by Graham J in Olin Mathieson Chemical Corporation v Biorex Laboratories Ltd [1970] RPC 157 at 187-188: whether the person skilled in the art at the relevant date with the relevant knowledge would directly be led, as a matter of course, to try the claimed invention in the expectation that it might well produce a useful alternative to, or a better drug than, the existing compound. 295 Their Honours warned against "the twin traps of hindsight and oversimplification" and stated that routine steps taken as a "matter of course" did not include "a course of action which was complex and detailed, as well as laborious, with a good deal of trial and error, with dead ends and the retracing of steps": AB Hässle at [58]. See also AstraZeneca HC at [15] (per French CJ); Hanwha Solutions Corp v REC Solar Pte Ltd [2023] FCA 1017 at [419]-[421], [429] (per Burley J). It is also apparent from the plurality's discussion in AB Hässle at [70]-[76] that "the fact that the person skilled in the art would consider a step or series of steps worthwhile to try or worth a try does not mean that the resulting invention is obvious or lacks an inventive step": see Australian Mud Company v Globaltech Corporation Pty Ltd (2018) 138 IPR 33 at [452] (per Besanko J). 296 The test for obviousness set out in AB Hässle is not "the test to be applied in every case" but is the appropriate test to be applied in this case, which is of a similar nature to AB Hässle: see Generic Health Pty Ltd v Bayer Pharma Aktiengesellschaft (2014) 106 IPR 381 at [71] (per Besanko, Middleton and Nicholas JJ). (Emphasis original.) 79 The primary judge later addressed two issues, namely whether the steps which would be taken by the person skilled in the art would be "routine steps" as submitted by Sandoz, and whether those steps would be taken with the expectation described in the reformulated Cripps question. 80 As to the first issue, the primary judge observed at J [596]: The identification of a potential lead candidate to take into pre-clinical testing, which far precedes trials in humans, is a very early step along the drug development pathway. There are many further steps along the way to achieving a safe and efficacious antithrombotic pharmaceutical composition, each of which presents an opportunity for failure or change in direction, including changes to chemical structure, formulation, release rate and dosage form. The risks of failure or a change in direction are even more likely when, as in this case, the compound is a first in class NCE, the side effects are entirely unknown, other unrelated compounds in the relevant field are known to have failed in early phase studies, and where the therapeutic window lies between uncontrolled bleeding and the formation of clots which can cause strokes or heart attacks. 81 The primary judge repeated her emphasis on the unknowns and risks associated with the drug development process to develop rivaroxaban at J [598], [600]-[604], [610]-[611], [613] and [621]. 82 At J [607], the primary judge stated that, "[a]t one level, all steps along the research and development path of a new drug are well established and follow a routine, or well known, path. The successful completion of one step, and the information obtained from that step, leads to the next step in the series". 83 However, the primary judge found at J [612]-[613] that: …the well-known standard series of steps of a drug development pathway are not the "routine steps" contemplated by Aickin J in Wellcome at 286. They are beyond tests carried out merely to confirm that a formulation works, or for regulatory purposes: Boehringer at [110] (per Perram, Nicholas and Burley JJ). The steps foreshadowed by Professors Roberts and Baker form part of a new, arduous, comprehensive, risky and unpredictable research project for a first in class NCE. Each test or step is undertaken with the purpose of trying to obtain information and to explore possibilities with the aim of ascertaining a full suite of information to enable a determination of whether to proceed with the compound into the next phase of testing… 84 The primary judge continued: 620 In AB Hässle, the plurality at [58] distinguished the tracing of a course of action which was complex and detailed, as well as laborious, "with a good deal of trial and error, with dead ends and the retracing of steps" from the taking of routine steps to which the hypothetical formulator was taken as a matter of course. 621 As Professor Evans observed, none of the compounds on WO 919 had ever been administered to a human being. There was the possibility that the compounds, including Example 44, may have adverse events or toxicities that may make them far worse than warfarin in terms of their side effects on people. Professor Evans gave as an example gastrointestinal erosion causing ulcers. A drug that might cause gastrointestinal erosion that also blocks coagulation of blood would cause disastrous gastrointestinal bleeds. Professor Roberts agreed that there were issues of potential toxicity with Example 44 which would have to be worked out. It is not possible to predict before testing that these risks of side effects and toxicities would eventuate. 622 In AB Hässle, the plurality referred at [67] to Buckley LJ's "voyage of discovery" passage from his judgment in Re Beecham Group Ltd's (Amoxycillin) Application [1980] RPC 261 at 296, describing it as the "vital passage": I am fully prepared to assume on the evidence before the court that [prior patent] 978, should be regarded as having made clear to one skilled in the field of penicillins that the epimers of the para-hydroxy and the meta-hydroxy compounds were likely to prove fruitful avenues of research, possibly the most promising avenues known to exist. I accept that the lines which that research would follow would be what [the opponent's] witnesses described as 'routine', ie well-known. I accept that anyone experienced in penicillin research who pursued research along those avenues would probably have found what Beecham found. But with great deference to the learned judge, I do not agree that this is enough to constitute the claim to Amoxycillin as a penicillin for administration to humans obvious for the purposes of section 14(1)(e) of the [1949 UK Act]. To reach the discovery of the particular characteristics of Amoxycillin and its suitability for treating humans the research worker would have had to embark upon a voyage of discovery. It is possible now to see that his voyage would have been short and perhaps uneventfully straightforward, but where each of his two, or possibly more, vessels would make landfall and what those places would be like would not have been obvious to him at the outset. The voyage might have been clearly worth trying but not as a means of reaching a specific hoped-for destination. (Emphasis original.) 85 As to the second issue, the primary judge stated at J [609] that: Not only must the person skilled in the art take the steps in question as a matter of routine, but they must also be carried out with an expectation of success. As was explored in the second joint session with the experts, there are risks inherent at each step along the drug development path. There is only around a 10% chance that a candidate drug will progress the entire way along the drug development path from pre-clinical testing to reach the approval stage. It is only from the successful endpoint looking back that it can be said in the case of a particular drug that there were no hurdles or obstacles that had to be overcome along the drug development pathway. This is the opposite of the s 7(2) test for inventive step which is a prospective test, looking forward from the prior art… 86 At J [613], the primary judge found that: …The pre-clinical tests are not undertaken with the requisite expectation of developing a safe and effective drug to treat thromboembolic disorders. Rather, the information gathered at each stage of the drug development process (cumulatively with the information gathered at the previous stages) will be used by the team to decide whether or not to proceed to test the compound in the next phase of clinical studies. As such, I do not consider that a person skilled in the art, equipped with the common general knowledge and the information in WO 919, would have the requisite expectation of success at the time of selecting Example 44 that the chosen compound would pass through all the drug development stages, from pre-clinical testing to successful completion of Phase III trials, to be approved for use in humans as a safe and effective once per day treatment for thromboembolic disorders. 87 In conclusion on both issues, the primary judge stated at J [623]: In my view, Bayer's drug development journey is more akin to the voyage of discovery described by Buckley LJ than the "working towards the invention with an expectation of success" referred to in AB Hässle. Although I am loathe to put a figure on the expectation of success, it seems to me that a 90% (or thereabouts) chance of Example 44 not making it through Phase III clinical trials is not in the realms of an "expectation of success". 88 While the above reasoning of the primary judge was in respect of the 226 Patent, her Honour decided that the 613 Patent was also not obvious for the same reasons: J [627].

Consideration 89 In the field of drug development, the need to carry out clinical trials and other tests in order to obtain relevant data can be regarded as routine work consistent with a finding of obviousness: see, generally, AstraZeneca AB v Apotex Pty Ltd (2015) 257 CLR 356; [2015] HCA 30 at [44] (French CJ), [93]-[95] (Kiefel J, as her Honour then was), [123] (Nettle J). In Merck Sharp & Dohme Corporation v Wyeth LLC (No 3) (2020) 155 IPR 1; [2020] FCA 1477, Burley J observed at [822] that: …Trial and error are normal, everyday parts of laboratory work and non-inventive laboratory experiments: "[t]hat is what the hypothetical skilled worker in a laboratory does - if the outcomes of experiments were known, there would be little point in doing them. That is the nature of everyday, non-inventive, research": Apotex Pty Ltd v Sanofi-Aventis (2009) 82 IPR 416; [2009] FCAFC 134 at [177] (Bennett and Middleton JJ), and to like effect AstraZeneca (FC) at [547] (Jessup J)… 90 Before the primary judge there was no evidence, and her Honour made no finding, that any particular problem, difficulty or issue would have been expected to have arisen or would have arisen during the pre-clinical and clinical trials which would conventionally be carried out following the selection of rivaroxaban as a lead candidate. 91 Nor was there any evidence, or finding, that the formulation of rivaroxaban in the 226 Patent or its dosage regime adopted in the 613 Patent would not have been identified during the course of these trials. 92 The primary judge emphasised the risk of failure to gain regulatory approval which is inherent in any drug development project. In this regard, her Honour characterised the identification of a "lead candidate" from WO 919 as being "a very early step along the drug development pathway", with many further steps "each of which present[ing] an opportunity for failure or change in direction, including changes to chemical structure, formulation, release rate and dosage form": J [596]. However, that risk is common to the field and establishes the baseline or context for the inventive step inquiry; it should not be determinative of the outcome. Further, there was no evidence, and no finding, that any "failure or change in direction" would have arisen in this case. 93 The evidence of the experts made it clear that they would conduct the necessary pre-clinical and clinical tests in taking such a compound forward in development. The evidence also made it clear that this was routine and conventional work according to a well-established development pathway. None of the experts suggested that they would hesitate in taking rivaroxaban forward because it was a "first in class" drug, or that this would change the nature of the tests that would be undertaken for that purpose. 94 In these circumstances, we consider that the primary judge placed too much emphasis on the risks and unknowns associated with the pre-clinical and clinical tests, which would have been undertaken as a matter of course following the selection of rivaroxaban as a lead candidate, in determining that those steps, which followed such a selection, would not be "routine" within the meaning contemplated by Wellcome Foundation Ltd v VR Laboratories (Aust) Pty Ltd (1981) 148 CLR 262. 95 This was, with respect, an error, especially as, notwithstanding those risks and uncertainties, the primary judge found that the skilled person reading WO 919 is likely to select rivaroxaban as a lead candidate to take into further drug development work, if not the lead candidate. 96 Further, it is implicit in this latter finding that the person skilled in the art would directly be led, as a matter of course, to undertake such drug development work in the expectation that it might well produce a useful alternative to, or a better drug than, existing compounds. Having regard to the risks and uncertainties associated with the drug development process on the facts of this case, one would not expect the person skilled in the art to even commence pre-clinical trials if such an expectation was absent. 97 This position is aligned with that of the Full Court in Generic Health Pty Ltd v Bayer Pharma Aktiengesellschaft (2014) 222 FCR 336; [2014] FCAFC 73 (Besanko, Middleton and Nicholas JJ). At [71] of that decision, the Court observed that: …It is difficult to think of a case where an expectation that an experiment might well succeed is not implicit in the characterisation of steps as routine and to be tried as a matter of course… 98 In Nichia Corporation v Arrow Electronics Australia Pty Ltd (2019) 175 IPR 187; [2019] FCAFC 2 at [88], Jagot J (with whom Besanko and Nicholas JJ agreed) cited that observation and stated that it indicates that care is needed in evaluating the character of the expectation separately from the steps the skilled addressee would have taken. 99 By her reasons, the primary judge used language which indicated that, to have the requisite expectation, the person skilled in the art would need to be able to make predictions as to efficacy, side effects and safety: J [598(e)], [611], [621] or adverse events and toxicities: J [610] and that, in order to have the requisite expectation, the research project could not be unpredictable: J [613]. 100 Further, the primary judge appeared to require that it be demonstrated that the person skilled in the art would have the requisite expectation at each stage of the drug development process, such that they need to have an expectation that the chosen compound would pass through all the drug development stages: J [613]. 101 With respect, this was not the correct approach. That is because the relevant expectation is to be measured against the ordinary level of expectation and risk inherent in routine work in the particular field. The relevant test is not knowing that steps will or would or even may well work, but merely expecting that the steps may well work: Nichia at [99]; see also Otsuka at [453] and Merck at [255]. 102 In Mylan Health Pty Ltd v Sun Pharma ANZ Pty Ltd (2020) 279 FCR 354; [2020] FCAFC 116 at [502] (Middleton, Jagot, Yates, Beach and Moshinsky JJ), the Full Court noted that: The reformulated Cripps question does not require certainty of outcome. It requires that the skilled addressee be directly led as a matter of course to try the claimed invention in the expectation that a particular research path "might well produce" a useful result (Hässle v Alphapharm at [53]). It does not require the skilled addressee to know that the steps will produce a useful result. 103 The authorities make it clear that it is not necessary for the skilled person to know a particular outcome at the outset of the hypothetical task. 104 As occurred in Nichia and referred to at [89] of the decision of Jagot J, the language used by the primary judge of prediction, as highlighted above, indicates "that the primary judge strayed from the test of steps taken in an expectation that they might well produce the invention or a useful result towards a test of an expectation of knowing that steps will produce a useful result based on predictive capacity". 105 As error has been shown, the consequence is that this Court must decide the question of obviousness itself. 106 Having regard to the findings of the primary judge, including the critical and unchallenged finding at J [594], we are satisfied that the person skilled in the art at the First Priority Date with the relevant knowledge (including the disclosure contained in WO 919) would have directly been led, as a matter of course, to try to develop rivaroxaban in the expectation that it might well produce a useful alternative to, or a better drug than, existing compounds for the treatment of thromboembolic disorders. In circumstances where the primary judge did not find that the formulation or dosing regimen involved an inventive step, and there was no finding that those matters would not have been identified during the course of the conventional clinical trials, this has the consequence that the inventions claimed in each of the 226 Patent and the 613 Patent was obvious.

GROUND 3 107 Having regard to our conclusions in relation to grounds 1 and 2, it is not strictly necessary to address ground 3. However, for completeness, we will give brief reasons for rejecting this ground. 108 This ground asserts an error by the primary judge in finding that the invention claimed in the 613 Patent involved an inventive step in the light of the common general knowledge together with the Blood Abstracts. The Blood Abstracts were relied upon in combination together: J [625].

Reasons of the primary judge 109 The primary judge found that the Blood Abstracts could be reasonably expected to have been ascertained, understood, and regarded as relevant by the person skilled in the art: J [515]. That finding is not challenged in this appeal. 110 At J [522], the primary judge observed that: …there is no suggestion in the specification of the 226 and 613 Patents that the inventors had invented rivaroxaban: indeed, it is explicitly said that the compound (that is now known as rivaroxaban) was disclosed in WO 919. What the inventors say they had done was to have invented a pharmaceutical composition with identified features (hydrophilized form or rapid release) that would enable rivaroxaban to be practically used in an orally administrable form suitable for the treatment or prophylaxis of thromboembolic disorders. Rivaroxaban is an essential integer in each of the claims of the 226 and 613 Patents. 111 At J [628]-[630], the primary judge gave an overview of the Blood Abstracts: Abstract 3003 reports a Phase I clinical trial in which the participants, who were healthy volunteers, were administered a 5 mg dose and a 30 mg dose of "BAY 59-7939". Abstract 3003 states that BAY 59-7939 was effective at inhibiting thrombin generation. It states that a single 30 mg dose of BAY 59-7939 "exerted a sustained effect in some assays of thrombin generation for up to 24 hours". Abstract 3004 reports that in a "parallel-group, randomized, single-blind, placebo controlled trial, 64 subjects received multiple oral doses of BAY 59-7939: 5 mg od [once daily], bid [twice daily], or tid [three times daily], or 10 mg, 20 mg, or 30 mg bid for 5 days with food." Abstract 3004 concludes that "BAY 59-7939 has predictable dose-dependent pharmacodynamics and pharmacokinetics without signs or symptoms of bleeding", and that "oral administration of BAY 59-7939 was safe and well tolerated in doses up to 30 mg bid". Abstract 3010 describes a single dose escalation Phase I study in which 103 male participants received a single dose of BAY 59-7939. Tablets containing doses from 1.25 mg to 80 mg, and 5 mg and 10 mg oral solutions, were administered. Abstract 3010 concludes that "BAY 59-7939 was safe and well tolerated across a wide range of oral doses (1.25-80 mg) with predictable dose-dependent pharmacodynamics and pharmacokinetics", and that "BAY 59-7939 offers predictable anticoagulation with an excellent safety profile". 112 The experts agreed that the Blood Abstracts in combination disclosed certain matters about the compound BAY 59-7939, which was "limited information": J [633]-[634]. The primary judge also made findings at J [636]-[637], which are not challenged: The Blood Abstracts do not disclose the identity or structure of compound BAY 59-7939, or even its chemical class. Nor was that information part of the common general knowledge. All that is disclosed is its pharmacological class - that it is a direct factor Xa inhibitor. Sandoz speculated that Bayer might disclose the structure of BAY 59-7939 to a person skilled in the art if asked. There was no evidence to support that proposition and Professor Crowther considered it unlikely. The Blood Abstracts also do not disclose the following information for BAY 59-7939: (a) the form of the compound used in the studies (whether it was the hydrate, a sulphate or a prodrug); (b) toxicity details; (c) formulation details; (d) chemical form (eg polymorph, crystalline, salt); (e) excipient details (which play an important role in absorption); or (f) the therapeutic window. 113 The primary judge made the following critical findings at J [645]-[646] and J [652]-[653]: I consider that ascertaining the structure of BAY 59-7939 is more than a matter of "checking and testing the product or process of the invention" after the invention has been made by "straightforward experiments" of the kind described by Aickin J in Wellcome. The identity and structure of BAY 59-7939 goes to the heart of the disclosure and information in the Blood Abstracts. Without further information, the person skilled in the art in possession of the Blood Abstracts would have no way of synthesizing compound BAY59-7939, and could not test that compound in pre-clinical testing and, depending on the results of those trials, in Phase I and II studies in humans. As Professor Roberts explained: "You can't do anything without a structure" and "without the precise compounds in the abstract, you can't use the data in the abstract" or "take them into a phase I or phase II clinical trial". … Without the structure of BAY 59-3959 [sic], or a sample of the compound from which its structure could be discerned, I do not consider that the skilled team would (or could) proceed any further. … If I am wrong on the fundamental issue above, I do not, in any event, consider that the invention claimed in the 613 Patent is obvious in light of the common general knowledge and the information disclosed in the combined Blood Abstracts.

Consideration 114 On this appeal, Sandoz complains that the primary judge erred in putting aside the Blood Abstracts on the basis that they referred to rivaroxaban as "BAY 59-7939". It submits that the evidence showed that the skilled person could conceive of an appropriate once-daily dosage regime for that compound without knowing its chemical name or formula. It submits that the fact that, in the real world, access to the compound would be required in order to evaluate that dosage regime and confirm its applicability does not undermine a conclusion of obviousness. Sandoz submits that the question posed by s 7(3) is a hypothetical one, not a real world one, and relies upon AstraZeneca at [46]; Pharmacia at [317]-[319] and [321]; and Pharmacia LLC v Juno Pharmaceuticals Pty Ltd (2022) 168 IPR 431; [2022] FCAFC 167 at [124]-[125] (Jagot, Yates and Downes JJ) (Pharmacia FC). 115 The submissions on this aspect of the appeal were amplified by oral submissions of counsel who appeared for Sandoz, who submitted that: … if what is conceived by the prior art is information that allows the person skilled in the art to conceive of the invention in its totality, it is not to the point that the person skilled in the art would not be able to put that invention into effect. The question is whether it could be conceived of. And if it could be conceived of, we say the invention is obvious. 116 Counsel submitted that the content of the information in the Blood Abstracts allowed the person skilled in the art to conceive of the invention of a once-a-day dosage regimen of the compound there identified as BAY 59-7939, which is, in fact, rivaroxaban. 117 However, as found by the primary judge at J [646], the identity and structure of Bay 59-7939 goes to the heart of the disclosure and information in the Blood Abstracts. The person skilled in the art, with the information disclosed by the Blood Abstracts, could not conceive of the invention, being the compound disclosed in claim 1 and the use of that compound to treat a thromboembolic disorder by administering it once a day in oral dosage form wherein the disclosed compound has a plasma concentration half-life of 10 hours or less when orally administered to a human patient. The issue is not, as was raised by Sandoz in oral submissions, that in the Blood Abstracts the compound was referred to as BAY 59-7939 rather than rivaroxaban. It is that the particular compound, its identity and structure, is not set out in the Blood Abstracts and would not otherwise be known to the skilled person. 118 Nothing would be done without the identity and structure of BAY 59-7939. This is an information deficiency which is not limited to the real world but which constrains the skilled person in the hypothetical. For this reason, such a finding does not trespass into factual contingencies of the kind addressed in Pharmacia FC at [124] and [128]. 119 This is not a situation where the skilled person is being equipped with the requisite information about the compound and other data relating to testing and then asking the Cripps question. Further, the invention is not just the dosage regime; it is the dosage regime associated with that particular chemical compound. Without the identity or structure of BAY 59-7939, or even its chemical class, the skilled person could not conceive of the invention as Sandoz submits, and would not be led to the invention as a matter of routine steps.

CONCLUSION AND DISPOSITION 120 It follows that the appeal should be allowed. Costs should follow the event. 121 We will direct the parties to prepare short minutes of order. I certify that the preceding one hundred and twenty-one (121) numbered paragraphs are a true copy of the Reasons for Judgment of the Honourable Justices Yates, Burley and Downes.