1 2 3 4 UNITED STATES DISTRICT COURT 5 NORTHERN DISTRICT OF CALIFORNIA 6

7 PLEXXIKON INC., Case No. 4:17-cv-04405-HSG

8 Plaintiff, ORDER REGARDING MOTIONS TO 9 vs. EXCLUDE DAMAGES EXPERTS’

OPINIONS

10 NOVARTIS PHARMACEUTICALS Re: Dkt. Nos. 169, 202

CORPORATION,

11 Defendant. 12

13 Pending before the Court are the parties’ motions to exclude certain expert opinions and 14 testimony related to damages. Plaintiff Plexxikon Inc. (“Plexxikon”) moves to exclude opinions 15 and testimony of defendant’s damages expert, James E. Malackowski. Dkt. No. 169 (“Pl. Mot.”). 16 Defendant Novartis Pharmaceuticals Corporation’s (“Novartis”) moves to exclude the testimony 17 of plaintiff’s damages expert, Gregory J. Leonard. Dkt. No. 202 (“Def. Mot.”). The Court heard 18 oral argument on these motions on November 1, 2019. See Dkt. No. 341. As detailed below, the 19 Court DENIES Novartis’ motion and GRANTS IN PART and DENIES IN PART Plexxikon’s 20 motion. 21 I. BACKGROUND 22 This is a patent infringement case related to Plexxikon’s patents for kinase inhibitors. 23 Plexxikon accuses Novartis’ melanoma drug Tafinlar, which Novartis acquired for $2 billion from 24 GlaxoSmithKline in 2015. Tafinlar works by inhibiting B-Raf, a type of kinase. The parties agree 25 that the asserted patents have never been licensed. Nevertheless, both parties’ damages experts 26 rely on purportedly comparable licenses under a “hypothetical negotiation” framework in order to 27 determine a reasonable royalty for the asserted patents. 1 both parties would have had “walk away points” beyond which they would not have entered a 2 license. See Dkt. No. 403-6 (“Leonard Report”) ¶ 45. For Plexxikon, the walk-away point relates 3 to the drug Zelboraf, which is sold by its licensee, Roche.1 Zelboraf is not covered by the asserted 4 patents. See id. ¶ 78. However, Zelboraf competes directly with Tafinlar as the only other B-Raf 5 inhibitor on the market. Id. ¶ 61. Dr. Leonard opines that Plexxikon would have considered the 6 “opportunity cost” of licensing to Novartis in terms of lost royalties from Zelboraf sales. Id. ¶ 60. 7 Assuming that Zelboraf would capture the projected Tafinlar sales in the absence of a Novartis 8 license, Dr. Leonard concludes that Plexxikon would not have accepted less than a 5.5% royalty 9 rate. Id. ¶ 71. 10 Separately, Dr. Leonard opines that the Roche collaboration agreement that covers 11 Zelboraf is also the most comparable license. Id. ¶ 83. Although they involve different patents, 12 Plexxikon’s technical expert opines that the technology is comparable, and Dr. Leonard finds that 13 the markets are similar because Zelboraf and Tafinlar serve the same patients through the same 14 mechanism of action and Roche and Novartis are similarly situated. Id. ¶ 103. Since the Roche 15 agreement also involves other types of collaboration (“know-how,” identification of a single lead 16 compound, etc.), Dr. Leonard apportions the value of the intellectual property by comparing 17 royalties in countries that did and did not have patent protection. Id. ¶¶ 107-09. Dr. Leonard thus 18 concludes that without these additional factors, the effectively royalty rate for the patents would be 19 6.26%-12.52%. Id. ¶ 144. 20 Novartis’ expert, Mr. Malackowski, disagrees that the Roche collaboration agreement is 21 comparable because, among other reasons, it identifies a specific compound and provides broad 22 collaboration benefits. Dkt. No. 393-17 (“Malackowski Report”) at 102-07. By contrast, in a 23 hypothetical negotiation, Plexxikon would have provided only a bare license, leaving the difficult 24 work of identifying and developing a specific compound to Novartis. Id. at 107; see also Leonard 25 Report ¶ 128 (agreeing that only bare license would be provided). During deposition, Mr. 26

27 1 Specifically, Hoffman-La Roche Inc. and F. Hoffman-La Roche Ltd. (together, “Roche”). Dr. 1 Malackowski testified that narrow patents that identify a specific compound are more valuable 2 than broad genus patents because the latter are “like a state permit to dig for gold in California,” 3 while the former provide “a treasure map with an X on the spot.” Dkt. No. 397-5 (“Malackowski 4 Depo.”) at 37:20-38:3, 146:22-147:1. Instead of the Roche license, Mr. Malackowski opines that 5 three other license agreements between Novartis and Rigel Pharmaceuticals, Inc., Curis, Inc., and 6 Harvard Corporation (the “Rigel,” “Curis,” and “Harvard” licenses, respectively) are comparable. 7 Malackowski Report at 50, 65. 8 All three agreements are “freedom to operate” licenses. Id. at 41, 54, 62. The Rigel 9 license arose from a legal settlement based on Novartis’ desire to avoid litigation. Id. at 40-41. It 10 licensed patents related to a different kinase inhibitor for a drug used to treat a different cancer 11 type. Id. at 46-49. The Curtis license also arose from a settlement of a lawsuit in which Novartis 12 further argued that the patents were invalid. Id. at 50. Again, Novartis testified that it agreed to 13 the license because doing so was cheaper than litigation. Id. at 53-54. The licensed drug treated 14 skin cancer, not melanoma. Id. at 58. Last, the Harvard license arose when Harvard approached 15 Novartis about licensing one of its products. Id. at 64. Novartis’ 30(b)(6) witness testified that 16 “Novartis determined that it would be cheaper to take the license than risk litigation for freedom to 17 operate purposes.” Dkt. No. 397-6 (“Waibel Depo.”) at 207:25-208:3. The licensed drug targets 18 multiple myeloma, not cancer, using a different type of inhibitor. Malackowski Report at 68. 19 Mr. Malackowski opines that these licenses are comparable because the parties, the license 20 terms, the patents, and the licensed product are all comparable. Namely, Rigel and Curtis are both 21 biotechnology companies that focus on drug discovery and collaboration, similar to Plexxikon; the 22 agreements provided a bare, non-exclusive license; the patents claim a broad genus of molecules 23 without identifying a specific compound; the licensed products have similar value and were on the 24 market for comparable amounts of time; and the parties were in a similar financial state at the time 25 of the license. Id. at 42-50, 54-62, 64-71. Based considerably on these licenses, Mr. Malackowski 26 concludes that a reasonable royalty would be a lump sum of $3.5 million. Id. at 96. However, he 27 also provides an alternative reasonable royalty opinion based on a greater discounting of the 1 II. LEGAL STANDARD 2 Federal Rule of Evidence 702 allows a qualified expert to testify “in the form of an opinion or 3 otherwise” where:

4 (a) the expert’s scientific, technical, or other specialized knowledge will help the trier of fact to understand the evidence or to determine a 5 fact in issue; (b) the testimony is based on sufficient facts or data; (c) the testimony is the product of reliable principles and methods; and 6 (d) the expert has reliably applied the principles and methods to the facts of the case. 7 8 Fed. R. Evid. 702. Expert testimony is admissible under Rule 702 if it is both relevant and reliable. 9 See Daubert v. Merrell Dow Pharm., Inc., 509 U.S. 579, 597 (1993). “[R]elevance means that the 10 evidence will assist the trier of fact to understand or determine a fact in issue.” Cooper v. Brown, 510 11 F.3d 870, 942 (9th Cir. 2007); see also Primiano v. Cook, 598 F.3d 558, 564 (9th Cir. 2010) (“The 12 requirement that the opinion testimony assist the trier of fact goes primarily to relevance.”) (quotation 13 omitted). Under the reliability requirement, the expert testimony must “ha[ve] a reliable basis in the 14 knowledge and experience of the relevant discipline.”2 Primiano, 598 F.3d at 565. To ensure 15 reliability, the Court “assess[es] the [expert’s] reasoning or methodology, using as appropriate such 16 criteria as testability, publication in peer reviewed literature, and general acceptance.” Id. at 564. 17 To satisfy these standards, testimony and evidence related to patent damages must be “tied to 18 the facts of the case.” Uniloc USA, Inc. v. Microsoft Corp., 632 F.3d 1292, 1315 (Fed. Cir. 2011) 19 (quoting Daubert, 509 U.S. at 591); see id. (“If the patentee fails to tie the theory to the facts of 20 the case, the testimony must be excluded.”). “Any evidence unrelated to the claimed invention 21 does not support compensation for infringement but punishes beyond the reach of the statute.” 22 ResQNet.com, Inc. v. Lansa, Inc., 594 F.3d 860, 869 (Fed. Cir. 2010). In addition, under Federal 23 Rule of Evidence 403, a court may exclude evidence “if its probative value is substantially 24 outweighed by a danger of . . . unfair prejudice.” See, e.g., Prism Techs. LLC v. Sprint Spectrum 25 L.P., 849 F.3d 1360, 1368 (Fed. Cir. 2017). That said, when evidence meets the “minimum 26

27 2 Admission of expert testimony is evaluated under regional circuit law—here, the law of the 1 threshold” of being based on case-specific data and sound methodology, disagreements as to 2 comparability go to weight, not admissibility. ActiveVideo Networks, Inc. v. Verizon Commn’s, 3 Inc., 694 F.3d 1312, 1334 (Fed. Cir. 2012); accord Primano, 598 F.3d at 565 (“Shaky but 4 admissible evidence is to be attacked by cross examination, contrary evidence, and attention to the 5 burden of proof, not exclusion.” (quoting Daubert, 509 U.S. at 596)). 6 III. DEFENDANT’S MOTION TO EXCLUDE EXPERT TESTIMONY OF GREGORY

K. LEONARD, PH.D

7 Novartis moves to exclude Dr. Leonard’s opinions in their entirety on two grounds. First, 8 Novartis argues that Dr. Leonard’s analysis lacks reliability because he focuses on valuing the 9 exclusionary value of the asserted patents and fails to apportion damages based on the “inventive 10 contribution” of the patents. Second, Novartis argues that Dr. Leonard improperly conducts a lost 11 profit analysis by setting a “minimum” royalty based on lost Zelboraf sales without satisfying the 12 rigorous test for lost profit recovery. 13 A. Apportionment 14 A long-standing rule holds that “[t]he patentee . . . must in every case give evidence 15 tending to separate or apportion the defendant’s profits and the patentee’s damages between the 16 patented feature and the unpatented features, and such evidence must be reliable and tangible, and 17 not conjectural or speculative.” Garretson v. Clark, 111 U.S. 120, 121 (1884). That is because 18 damages in a patent case are meant to capture “the value of what was taken,” which is the patented 19 invention. Ericsson, Inc. v. D-Link Sys., Inc., 773 F.3d 1201, 1226 (Fed. Cir. 2014). Thus, “where 20 multi-component products are involved, the governing rule is that the ultimate combination of 21 royalty base and royalty rate must reflect the value attributable to the infringing features of the 22 product, and no more.” Id.; see also VirnetX, Inc. v. Cisco Sys., Inc., 767 F.3d 1308 (Fed. Cir. 23 2014). An exception to this rule, called the “entire market value rule,” permits a patentee to obtain 24 damages based on the value of the accused products as a whole “where the patented feature creates 25 the basis for consumer demand” or “substantially creates the value of the component parts.” 26 Uniloc, 632 F.3d at 1318. 27 Even where the patented features are not part of a multi-component product, damages must 1 still reflect “the incremental value that the patented invention adds to the end product.” Exmark 2 Mfg. Co. Inc. v. Briggs & Stratton Power Prods. Grp., LLC, 879 F.3d 1332, 1348 (Fed. Cir. 2018). 3 For example, in Exmark, the patent covered a lawn mower as a whole, not any single component. 4 See id. However, the court still held that damages had to be apportioned between the conventional 5 and inventive aspects of the claim. Id. Similarly, in AstraZeneca AB v. Apotex Corp., 782 F.3d 6 1324, 1338 (Fed. Cir. 2015), the court found that the patents covered an entire pharmacological 7 compound, not any one component. However, because the active ingredient was covered by a 8 different, expired patent—and was therefore conventional—damages needed to be apportioned to 9 account for the “inventive” feature of the patent, which was the subcoating. Id. at 1337-38. 10 Apportionment can be addressed “in a variety of ways.” Exmark, 879 F.3d at 1348. Aside 11 from adjusting the royalty base or royalty rate, a proper analysis of the Georgia-Pacific factors 12 may account for the value of the invention. Id. at 1348-49.3 That is because the Georgia-Pacific 13 factors already account for the “utility and advantages of the patent property over any old modes 14 or devices,” “the nature of the patented invention,” and the “portion of the realizable profit that 15 should be credited to the invention.” AstraZeneca, 782 F.3d at 1338; see also Univ. of Pittsburgh 16 v. Varian Med. Sys., Inc., 561 F. Appx 934, 947-50 (Fed. Cir. 2014) (finding that distinguishing 17 inventive features “is precisely what the Georgia-Pacific factors purport to do”). Apportionment 18 analysis does not require subtraction of the conventional elements where the inventive element is 19

20 3 The Georgia-Pacific factors include (1) the royalties received by the patent owner for the licensing of the patent-in suit, proving or tending to prove an established royalty, (2) the rates paid 21 by the licensee for the use of other patents comparable to the patent-in-suit, (3) the nature and scope of the license, (4) the licensor’s established policy and marketing program to maintain its 22 patent monopoly by not licensing others or by granting licenses under conditions, (5) the commercial relationship between the licensor and licensee, (6) the effect of selling the licensed 23 inventions on other products sold by the licensee, as well as the value of the invention to the licensor as a generator of sales of non-patented items, (7) the duration of the patent and the term of 24 the license, (8) the established profitability, commercial success, and popularity of the patented product, (9) the utility and advantages of the invention over old modes or devices, (10) the nature 25 and benefits of the patented invention, (11) the extent to which the infringer had made use of the invention, (12) the portion of the profit or of the selling price for comparable businesses to use the 26 invention or analogous inventions, (13) the portion of the realizable profit that should be credited to the invention as distinguished from non-patented elements, (14) the opinion of qualified 27 experts, and (15) the amount that a licensor and licensee would have agreed upon if both had been 1 “substantially responsible for the value of the product.” AstraZeneca, 782 F.3d at 1339. 2 Here, Novartis argues that Dr. Leonard testified during deposition that his analysis values 3 Plexxikon’s right to exclude, not the incremental value of the patented inventions. See Dkt. No. 4 393-29 (“Leonard Depo.”) at 90:22-25 (describing the purpose of his analysis as “get[ting] a 5 royalty that reflects the value of the licensor’s right to exclude”). Dr. Leonard apparently did not 6 consider the inventive aspects of the patent claims and relied entirely on Plexxikon’s technical 7 expert, Dr. Michael Metzker, for this analysis. See id. at 42:23-44:1, 40:4-17. Novartis thus 8 argues that Dr. Leonard fails to account for Novartis’ own inventive contribution in identifying 9 and developing the specific molecule embodied in Tafinlar, which has been separately patented. 10 Plexxikon responds, in the first instance, that the right to exclude is equivalent to the 11 incremental value of the invention because low-value inventions would have many noninfringing 12 alternatives.4 That may be true for narrow “improvement” patents, but it would not be true for 13 broad “blocking” patents. Cf. Acorda Therapeutics, Inc. v. Roxane Labs., Inc., 903 F.3d 1310, 14 1337 (Fed. Cir. 2018) (a patent is “blocking” where “practice of a later invention would infringe 15 the earlier patent”). For example, a patent for a car with novel tires can be avoided by making a 16 car without those tires. But a patent for a car could not be avoided by a party that invents novel 17 tires, no matter how innovative the invention. Here, Novartis argues that Plexxikon’s patents 18 cover trillions of compounds that require significant additional research and development to 19 identify specific useful compounds. The right to exclude cannot account for these latter inventive 20 contributions and thus cannot fully apportion value.5 Cf. Exmark, 879 F.3d at 1350 (finding 21 “problematic” that an expert ignored defendants’ own patents for the accused products). 22 4 Plexxikon also argues that Novartis’ cited cases do not apply because they deal with multi- 23 component products. That is incorrect: the principle stated in those cases has been applied to claims that cover the entire product. See, e.g., AstraZeneca, 782 F.3d at 1338 (citing Ericsson, 24 773 F.3d at 1233). Plexxikon cites AstraZeneca in its brief and presumably knows as much.

25 5 The current case presents an opposite scenario from the one in AstraZeneca. There, the active ingredient was known, but the invention provided an improvement in subcoating that made the 26 commercial embodiment viable. 782 F.3d at 1339. Here, Plexxikon invented the ingredient, but Novartis apparently provided the improvements that allowed the drug to be brought to market. 27 Since AstraZeneca affirmed attributing the value of the entire product to the improvement in the 1 Plexxikon further argues that Dr. Leonard nevertheless apportioned damages because he 2 relies on the Roche agreement that provides for royalties reflecting Roche’s development costs 3 and for lower royalties in countries that lack patent protection. While that may be sufficient to 4 distinguish the value of the intellectual property compared to other benefits provided under the 5 collaboration agreement (know-how, etc.), it cannot value the inventions of the asserted patents 6 because the Roche patents cover different technology. Intellectual property is not fungible, and 7 the value of one invention is not indicative of the value of another. See ResQNet.com, 594 F.3d at 8 869 (requiring “district courts performing reasonable royalty calculations to exercise vigilance 9 when considering past licenses to technologies other than the patent in suit”); LaserDynamics, Inc. 10 v. Quanta Comp., Inc., 694 F.3d 51, 79 (Fed. Cir. 2012) (“alleging a loose or vague comparability 11 between different technologies or licenses does not suffice”). 12 Nevertheless, Plexxikon provides enough evidence to show that these deficiencies are best 13 addressed through cross-examination, rather than exclusion. See ActiveVideo, 694 F.3d at 1334. 14 Dr. Leonard opines, as relevant, that both the Roche patents and the asserted patents are “critical 15 to the effectiveness of” their respective products, Zelboraf and Tafinlar. Leonard Report ¶ 103. 16 Dr. Metzker supports this opinion in his technical report. See Dkt. No. 441-1 (“Metzker Report”) 17 ¶ 113. Dr. Leonard further opines that Zelboraf and Tafinlar operate in the same competitive 18 space and serve the same patient populations. Leonard Report ¶ 103. The Federal Circuit recently 19 affirmed a jury verdict based on similar evidence of damages, where the plaintiff’s expert testified 20 regarding a license for a related, earlier technology and “the components at issue, for purposes of 21 apportionment to the value of a larger product or service, were comparable.” Elbit Sys. Land & 22 C4l Ltd. v. Hughes Network Sys., LLC, 927 F.3d 1292, 1300-01 (Fed. Cir. 2019); see also Summit 23 6, LLC v. Samsung Elecs. Co., Ltd., 802 F.3d 1283, 1296 (Fed. Cir. 2015) (permitting reasonable 24 royalty value “based upon comparable features in the marketplace”). Because the license 25 agreement “already built in apportionment,” it was sufficient to analyze the asserted patents by 26 adjusting the values of the comparable license for the related patents. Elbit, 927 F.3d at 1301. 27 Here as well, Plexxikon provides enough evidence of the comparability of the asserted patents to 1 Accordingly, the Court DENIES Novartis’ motion in this respect. 2 B. Lost Profits 3 Under 35 U.S.C. § 284, damages for patent infringement must be “adequate to compensate 4 for the infringement” but “in no event less than the reasonable royalty for the use made of the 5 invention by the infringer.” To prove actual damages in the form of lost profits, the patentee must 6 show but-for causation, which is governed by the rigorous test articulated in Panduit Corp. v. 7 Stahlin Bros. Fibre Works, Inc., 575 F.2d 1152 (6th Cir. 1978). See Rite-Hite Corp. v. Kelley Co., 8 Inc., 56 F.3d 1538, 1545 (Fed. Cir. 1005). Alternatively, a reasonable royalty provides a “floor” 9 for damages under the flexible “hypothetical negotiation” framework of Georgia Pacific, 318 F. 10 Supp. at 1120-21. Lucent Techs., Inc. v. Gateway, Inc., 580 F.3d 1301, 1324 (Fed. Cir. 2009). 11 Calculating a reasonable royalty “is not an exact science” and permits multiple reliable methods. 12 Summit 6, 802 F.3d at 1296. 13 Novartis argues that Dr. Leonard improperly relied on an incomplete lost profits analysis 14 by setting a “floor” for the license based on lost sales of Zelboraf. In particular, Novartis argues 15 that Dr. Leonard did not consider the third element of the Panduit test, which concerns Roche’s 16 manufacturing ability to make Tafinlar sales in the absence of a Novartis license. However, the 17 Federal Circuit has approved of consideration of a patent owner’s lost profits in a hypothetical 18 negotiation framework, even without considering the Panduit test. See, e.g., Asetek Danmark A/S 19 v. CMI USA Inc., 852 F.3d 1352, 1362-63 (Fed. Cir. 2017) (“[A] patent owner participating in a 20 hypothetical negotiation would consider the profits on sales it might lose as a result of granting a 21 license.”); Rite-Hite, 56 F.3d at 1554 (finding it reasonable that “an unwilling patentee would only 22 license for one-half of its expected lost profits”). The issue here is not whether Roche would have 23 actually made Tafinlar’s sales, but whether Plexxikon would reasonably believe during a 24 hypothetical negotiation that this would happen. See SEB S.A. v. Montgomery Ward & Co., Inc., 25 594 F.3d 1360, 1380 (Fed. Cir. 2010). Here, Dr. Leonard properly considered Plexxikon’s ex ante 26 expectations for lost royalties as part of a hypothetical negotiation framework, and Novartis may 27 attack this evidence on cross examination.

IV. PLAINTIFF’S MOTION TO EXCLUDE THE OPINIONS AND TESTIMONY OF

1 JAMES E. MALACKOWSKI 2 Plexxikon moves to exclude the opinions of Mr. Malackowski regarding three litigation 3 settlement licenses. Plexxikon does not move to exclude Mr. Malackowski’s rebuttal of Dr. 4 Leonard’s opinions, including his alternative royalty calculation based on the Roche collaboration 5 agreement. Instead, Plexxikon argues only that the settlement licenses are unreliable under Rules 6 403 and 408 because they concern different technologies and markets and also reflect Novartis’ 7 desire to avoid litigation. Plexxikon also argues that Mr. Malackowski should not be permitted to 8 testify that broad genus patents are worth less than narrower patents. 9 A. Settlement Licenses 10 The court has discretion to admit evidence regarding settlement licenses if they are 11 sufficiently probative of the value of the asserted patents. Prism, 849 F.3d at 1368-69. As the 12 Supreme Court recognized, litigation settlements typically reflect three components: (1) the 13 predicted judgment, discounted by its probability, (2) the costs of further litigation, and (3) the 14 cost of settlement. Id. at 1369 (citing Evans v. Jeff D., 475 U.S. 717, 734 (1986)). In the patent 15 law context, the first element can be further broken down into (a) the expected judgment for 16 infringement, and (b) the expected judgment for invalidity. See id. at 1370. Of these, only the 17 first and third elements are relevant for patent valuation, and only in relation to infringement. See 18 id. at 1369. 19 Thus, the probative value of a settlement license depends on the weight of the other 20 factors. Id. at 1370. Where a license was negotiated after a court established validity and 21 infringement, the license is “very probative of reasonable royalty” because it “duplicate[s]” the 22 hypothetical negotiation framework. AstraZeneca, 782 F.3d at 1337. On the other hand, where 23 the license was negotiated early in the suit, when expected costs of litigation and uncertainty over 24 the outcome are highest, the license is less probative. See Prism, 849 F.3d at 1369. Moreover, 25 additional litigation factors, such as the risk of enhanced damages or sanctions, will further skew 26 the license amount away from proper patent valuation. Id. at 1370; LaserDynamics, 694 F.3d at 27 78. In short, settlement licenses are treated much like other licenses, requiring the proponent of 1 927 F.3d at 1299-1300; Prism, 849 F.3d at 1370. 2 Plexxikon argues that settlement licenses must be excluded unless they are “the most 3 reliable” on the record. That is simply not the law. LaserDynamics, which Plexxikon cites for 4 this proposition, merely states that settlement licenses have been excluded until recently. See 694 5 F.3d at 77. It then distinguishes ResQNet, which earlier permitted such licenses based, in part, on 6 the settlement license in that case being “the most reliable” on record.6 594 F.3d 872. Nothing in 7 LaserDynamics or ResQNet suggests that a settlement license must be the most reliable in the 8 record to be admitted. As more recent cases show, there is no such requirement. See Elbit, 927 9 F.3d at 1299-1300; Prism, 849 F.3d at 1370; Rembrandt Wireless Techs., LP v. Samsung Elecs. 10 Co., Ltd., 853 F.3d 1370, 1381 (Fed. Cir. 2017); AstraZeneca, 782 F.3d at 1337.7 11 Nevertheless, Plexxikon is correct that the settlement licenses here do not appear to be 12 particularly probative. First, they do not involve the asserted patents and Novartis has not shown 13 that the technology is comparable. See ResQNet.com, 594 F.3d at 869 (cautioning against reliance 14 on past licenses for technologies other than the patent in suit); LaserDynamics, 694 F.3d at 79 15 (finding “vague comparability between different technologies” insufficient). Second, they were 16 entered into early in the cases, and Mr. Malackowski expressly opines that Novartis agreed to the 17 licenses to avoid litigation. The second factor of the “transaction costs of further litigation” is thus 18 a dominant factor in the license values. Prism, 849 F.3d at 1369. Third, Mr. Malackowski 19 introduces no evidence that Novartis actually infringed the licensed patents or that the value of 20 potential infringement was comparable to the reasonable value here. Finally, Novartis’ 30(b)(6) 21 witness testified that the patents in the Curtis license were invalid, which means that the patentee 22 further discounted the royalty amount to account for a potential judgment of invalidity if the case 23 continued to trial. The probative value of these licenses is thus limited while the risk of prejudice 24 from skewed damages horizons before the jury is significant. See Prism, 849 F.3d at 1370 25

26 6 By contrast, the LaserDynamics court excluded a settlement license because it was “the least reliable license by a wide margin.” 694 F.3d at 77-78. 27 1 (permitting exclusion of licenses under Rule 403).8 2 Novartis’ evidence for comparability only weakly ties those licenses “to the facts of the 3 case.” Summit 6, 802 F.3d at 1296. Although Mr. Malackowski purports to opine on the 4 comparability of the license terms, products, patents, and negotiation positions, much of that 5 evidence is superficial or insufficient to show comparability. For example, Mr. Malackowski 6 opines that the product licensed under the Rigel agreement, Zykadia, is comparable to Tafinlar 7 even through it makes only a fraction of the sales. Malackowski Report at 48. The Curtis and 8 Harvard licenses also relate to different technology and cover products that made less than 5% of 9 Tafinlar’s sales. Id. at 60-61, 70. Moreover, the comparability of the patents in all three cases is 10 based on their breadth and relation to each other, rather than the value of the inventions. Id. at 49. 11 To be sure, Mr. Malackowski does provide some evidence of comparability, particularly from 12 Novartis’ perspective (where each license provides “freedom to operate” for an already-existing 13 product). Id. at 107-08. However, in light of the litigation-focused circumstances of the 14 negotiations of the licenses, as well as the differences of the licensed products and technology, the 15 probative value in those licenses is outweighed by the risk of undue prejudice. Fed. R. Evid. 403.9 16 Accordingly, the Court GRANTS Plexxikon’s motion to exclude Mr. Malackowski’s 17 testimony regarding the Rigel, Curtis, and Harvard licenses. 18 B. Genus Patents 19 Plexxikon argues that Mr. Malackowski should not be permitted to testify that a broad 20 genus patent is less valuable than a narrower one that identifies specific compounds. The issue is 21

22 8 Although the Harvard license was not made in the context of litigation, Novartis’ 30(b)(6) witness testified that the company agreed to the license because “it would be cheaper to take the 23 license than risk litigation for freedom to operate purposes.” Waibel Depo. at 207:25-208:3. The Harvard license is therefore most probative of litigation costs, not the value of the technology. 24

9 Plexxikon also argues that the settlement licenses are inadmissible under Rule 408. However, 25 the Federal Circuit has recognized that settlement licenses are “sometimes [] admissible and sometimes [] not” and “can be pertinent to the issue of reasonable royalties.” Prism, 849 F.3d at 26 1368-69; In re MSTG, Inc., 675 F.3d 1337, 1348 (Fed. Cir. 2012). Because discoverability of information related to reasonable royalties implicates that circuit’s exclusive jurisdiction, Federal 27 Circuit law is likely to apply to the admissibility question at issue here as well. MSTG, 675 F.3d 1 closely related to the one regarding apportionment discussed supra section III.A. Namely, 2 Plexxikon argues that the value of a patent is commensurate with its ability to exclude, while 3 Novartis argues that a broad genus patent provides less of an “inventive contribution” than a 4 patent identifying specific useful compounds. 5 The Court finds, again, that this issue is properly resolved by a proper application of the 6 Georgia-Pacific factors. The need for additional development rests, in large part, on the date of 7 the hypothetical negotiation. See Integra Lifesciences I, Ltd. v. Merck KGaA, 331 F.3d 860, 870- 8 71 (Fed. Cir. 2003), rev’d on other grounds 545 U.S. 193 (2005); Hitkansut LLC v. U.S., 130 Fed. 9 Cl. 353, 393-94 (Ct. Fed. Cl. 2017). In Integra, the patents concerned certain peptide sequences 10 that the accused infringer tested as part of its own, largely independent discovery. 331 F.3d at 11 862-63. The Federal Circuit analyzed the peptides as a “research tool” and concluded that their 12 value depended on the date of the negotiation: “At the point before Merck even attempted its first 13 test on [the peptide], it would have assumed all the risks of failure—either scientific failure to 14 identify a suitable therapeutic candidate or economic failure to market a successful product,” while 15 using the tool “to confirm an already recognized drug candidate’s safety or efficacy” would result 16 in a different valuation. Id. at 870-71 & n.4. Similarly, in Hitkansut, the Court of Federal Claims 17 found that a hypothetical negotiation for patents that provide only “potential” benefits depends on 18 the state of development and commercialization. 130 F. Cl. at 393-94. Thus, the value of patents 19 that require additional development to commercialize rests on the state of the development at the 20 time of the hypothetical negotiation. In all instances, evaluation of the patents must be tied to the 21 facts of the case—broad statements about the value of genus patents in the abstract are largely not 22 helpful. See Uniloc USA, 632 F.3d at 1315. 23 Here, the parties appear to agree that the hypothetical negotiation would take place after 24 Novartis already developed Tafinlar, which might increase its willingness to pay. However, Mr. 25 Malackowski provides his opinion as part of the apportionment analysis to distinguish Novartis’ 26 own contribution and to show the non-comparability of the Roche license, which identified 27 specific compounds. These are reasonable ways of analyzing the issues, even if more abstract ] the Court DENIES this part of Plexxikon’s motion. 2 || V. CONCLUSION 3 For the foregoing reasons, the Court DENIES Novartis’ motion to exclude the testimony 4 of Dr. Leonard and GRANTS IN PART and DENIES IN PART Plexxikon’s motion to exclude 5 || the testimony of Mr. Malackowski. 6 7 8 9 IT ISSO ORDERED. 10 11 Dated: 1/12/2021

HAYWOOD S. GILLIAM, JR.

14 United States District Judge

16

Oo

Z 18

19 20 21 22 23 24 25 26 27 28