Case: 20-2243 Document: 28 Page: 1 Filed: 11/04/2021

United States Court of Appeals for the Federal Circuit ______________________

UNIVERSITY OF STRATHCLYDE, Appellant

v.

CLEAR-VU LIGHTING LLC, Appellee ______________________

2020-2243 ______________________

Appeal from the United States Patent and Trademark Office, Patent Trial and Appeal Board in No. IPR2019- 00431. ______________________

Decided: November 4, 2021 ______________________

CHRISTOPHER BRANTLEY KELLY, Alston & Bird LLP, At- lanta, GA, argued for appellant. Also represented by JASON P. COOPER; KIRK T. BRADLEY, Charlotte, NC; NATALIE CHRISTINE CLAYTON, New York, NY.

BRIAN J. EMFINGER, Banner & Witcoff, Ltd., Chicago, IL, argued for appellee. Also represented by MATTHEW PAUL BECKER, BINAL J. PATEL. ______________________

Before REYNA, CLEVENGER, and STOLL, Circuit Judges. Case: 20-2243 Document: 28 Page: 2 Filed: 11/04/2021

2 UNIVERSITY OF STRATHCLYDE v. CLEAR-VU LIGHTING LLC

STOLL, Circuit Judge. University of Strathclyde appeals from a final written decision of the Patent Trial and Appeal Board holding claims 1–4 of U.S. Patent No. 9,839,706 unpatentable as obvious. Because neither the Board’s finding that the prior art disclosed all claim limitations nor its finding of a rea- sonable expectation of success is supported by substantial evidence, we reverse the Board’s obviousness determina- tion. BACKGROUND I Methicillin-resistant Staphylococcus aureus (MRSA), and other Gram-positive bacteria 1 that have developed re- sistance to antibiotics, are “known to cause health prob- lems, particularly in the hospital environment.” ’706 patent col. 1 ll. 38–58. The specification of the ’706 pa- tent explains that effective methods of controlling trans- mission of and rising infection rates from antibiotic- resistant bacteria is “becoming one of the most significant problems within the healthcare industry.” Id. at col. 1 ll. 30–33. This is due, in part, to the “availability of few effective sterili[z]ation methods for environmental decon- tamination” of air and surfaces. Id. at col. 1 ll. 26–30. The specification discloses that photoinactivation is a method that has emerged for killing harmful bacteria like MRSA. Previous photoinactivation methods involved treating an infection by applying a photosensitizing agent and activating the photosensitizing agent using light

1 Gram-positive refers to the results of a “Gram” stain, which is a test for determining what type of cell wall structure a given bacterium has. Whether a bacterium is Gram-positive or -negative determines which antibiotics can be used in a treatment plan. Case: 20-2243 Document: 28 Page: 3 Filed: 11/04/2021

UNIVERSITY OF STRATHCLYDE v. CLEAR-VU LIGHTING LLC 3

energy, e.g., visible light having a wavelength in the region of 450–850 nm. Id. at col. 1 l. 59–col. 2 l. 9. While this technique has been shown to damage or kill certain bacte- ria such that their ability to grow is reduced or eliminated altogether, it suffers from “the significant practical disad- vantage that photosensiti[z]ing agents must be applied to the bacteria that are to be inactivated.” Id. at col. 2 ll. 11–15. Scientists at the University of Strathclyde, recognizing this practical disadvantage, developed a method for pho- toinactivating antibiotic-resistant bacteria like MRSA without using a photosensitizing agent. Through numer- ous experiments, the inventors found that “[e]xposing [cer- tain] bacteria to blue light, or white light containing blue light” having a wavelength in the region of 400–500 nm, “stimulate[s] an inactivation process.” Id. at col. 2 ll. 50–52, 63–64. Specifically, the inventors experimented with different filters that allowed certain wavelengths of light to reach the bacteria, finding that wavelengths in the 400–500 nm region “provide[d] a high rate of [MRSA] inac- tivation.” Id. at col. 5 ll. 11–13. These experiments led the inventors to conclude that “visible-light exposure over the wavelength range 400–450 nm is the major inducing factor for Staphylococcal [e.g., MRSA] inactivation, with in- creased inactivation occurring over the range 400–420 nm and optimum inactivation occurring at 405 nm.” Id. at col. 5 ll. 36–40. The inventors also discovered that ex- posing bacteria such as MRSA to 405 nm blue light re- quired a lower dose of light energy for inactivation compared to exposing bacteria to a broader wavelength range. For example, the inventors reported that a light dose of 945 J/cm2 was required to inactivate MRSA when it was exposed to a broad spectrum of visible light greater than 400 nm, whereas a light dose of only 45 J/cm2 was re- quired for the cultures exposed to only 405 nm blue light. See id. at col. 6 ll. 37–49. Case: 20-2243 Document: 28 Page: 4 Filed: 11/04/2021

4 UNIVERSITY OF STRATHCLYDE v. CLEAR-VU LIGHTING LLC

Based on their discoveries, the inventors developed a method of disinfection, which they claimed in the ’706 pa- tent. Claim 1 is illustrative of the claims on appeal and recites: 1. A method for disinfecting air, contact surfaces or materials by inactivating one or more patho- genic Gram-positive bacteria in the air, on the con- tact surfaces or on the materials, said method comprising exposing the one or more pathogenic Gram-positive bacteria to visible light without us- ing a photosensitizer, wherein the one or more pathogenic Gram-positive bacteria are selected from the group consisting of Methicillin-resistant Staphylococcus aureus (MRSA), Coagulase-Nega- tive Staphylococcus (CONS), Streptococcus, Enter- ococcus, and Clostridium species, and wherein a portion of the visible light that inactivates the one or more pathogenic Gram-positive bacteria consists of wavelengths in the range 400-420 nm, and wherein the method is performed outside of the hu- man body and the contact surfaces or the materials are non-living. Id. at col. 7 l. 17–col. 8 l. 5. II The Board determined that claims 1 and 3 of the ’706 patent would have been obvious over Ashkenazi 2 in view of Nitzan, 3 and that claims 2 and 4 would have been

2 Helena Ashkenazi et al., Eradication of Propioni- bacterium acnes by its endogenic porphyrins after illumina- tion with high intensity blue light, 35 J. FEMS Immunology & Med. Microbiology 17, 17–24 (2003). 3 Yeshayahu Nitzan et al., ALA induced photody- namic effects on Gram positive and negative bacteria, Case: 20-2243 Document: 28 Page: 5 Filed: 11/04/2021

UNIVERSITY OF STRATHCLYDE v. CLEAR-VU LIGHTING LLC 5

obvious in further view of Jones. 4 Because Strathclyde’s appeal is focused on the Board’s findings regarding Ashke- nazi and Nitzan, we discuss each of those references below. A Ashkenazi is an article that discusses photoeradication of Propionibacterium acnes (P. acnes), a Gram-positive bac- terium that is the leading cause of acne. Ashkenazi at 17. “In the case of P. acnes or other bacterial cells that produce porphyrins,” Ashkenazi hypothesized, “blue light may pho- toinactivate the intact bacterial cells.” Id. at 21; see also id. at 18 (“It has also been shown that when illuminated with blue light, porphyrins damage the cells very effi- ciently.”). Ashkenazi suggested that light-activated por- phyrin molecules contribute to bacterial cell death through release of free radicals. Ashkenazi provides a method for photosensitizing P. acnes, which naturally produces high amounts of por- phyrins, using δ-aminolevulinic acid (ALA), a photosensi- tizer 5 that enhances porphyrin production inside cells (i.e., endogenous porphyrins). To study the effects of ALA on inactivation, P. acnes was grown on a reinforced clostridial agar media supplemented with ALA and an unsupple- mented media (i.e., without ALA) was used as a control. It is undisputed that clostridial agar would have contained

3 Photochemical & Photobiological Scis. 430, 430–35 (2004). 4 U.S. Pat. App. Pub. No. 2005/00550070. 5 The Board construed the term “photosensitizer” to mean “a substance that, when applied to a target sub- stance, makes the target substance more sensitive to light.” Clear-Vu Lighting LLC v. Univ. of Strathclyde, No. IPR2019-00431, 2020 WL 3848045, at *7 (P.T.A.B. July 8, 2020) (Final Decision). Neither party challenges the Board’s construction on appeal. Case: 20-2243 Document: 28 Page: 6 Filed: 11/04/2021

6 UNIVERSITY OF STRATHCLYDE v. CLEAR-VU LIGHTING LLC

the vitamin riboflavin which, like ALA, is a photosensi- tizer. Thus, both of Ashkenazi’s methods involved growing P. acnes in the presence of a photosensitizer—either ribo- flavin alone or together with ALA. The cultures grew for up to 96 hours and were illumi- nated with 407–420 nm blue light at various light doses ei- ther once after 24 hours of growth or at consecutive 24-hour intervals. Both the ALA and non-ALA P. acnes cultures that were exposed to multiple, consecutive illuminations showed a decrease in viability of several orders of magni- tude compared to cultures illuminated only once, and “[t]he decrease in viability of the cultures grown with ALA and illuminated with blue light was far more significant than that of cultures grown without ALA.” Id. at 20. Ashkenazi also observed that increasing the light dose, e.g., from 75 J/cm2 to 100 J/cm2, likewise resulted in decreased via- bility for both the ALA and non-ALA P. acnes. Noting that “[t]he increase in the amounts of porphyrins in P. acnes as a result of ALA induction was significantly above the nat- ural production of this bacterium,” Ashkenazi concluded from these experiments that “the greater the intracellular amount of the porphyrin the better are the eradication re- sults.” Id. at 22. B Nitzan’s authors, two of whom co-authored the Ashke- nazi article, continued studying the effects of ALA on Gram-positive bacteria, including MRSA; the results were published just over a year after Ashkenazi’s. Part of Nitzan’s study focused on how exogenous porphyrins—i.e., a culture media supplemented with porphyrins—would af- fect inactivation. To do so, four different categories of MRSA cultures were prepared for the various bacteria, two of which were exposed to ALA, and two of which were not. Nitzan at 433 (Tbl. 5). For all of the non-ALA MRSA cul- tures, only some of which were grown on a media supple- mented with exogenous porphyrins, Nitzan reported a Case: 20-2243 Document: 28 Page: 7 Filed: 11/04/2021

UNIVERSITY OF STRATHCLYDE v. CLEAR-VU LIGHTING LLC 7

1.0 survival fraction, meaning there was “no decrease in vi- ability . . . after illumination” with 50 J/cm2 of 407–420 nm blue light. Id. at 433. III Clear-Vu Lighting LLC petitioned for inter partes re- view of claims 1–4 of the ’706 patent on four separate grounds: (1) claims 1 and 3 are anticipated by Nitzan; (2) claims 2 and 4 would have been obvious over Nitzan in view of Jones; (3) claims 1 and 3 would have been obvious over Ashkenazi in view of Nitzan; and (4) claims 2 and 4 would have been obvious over Ashkenazi in view of Nitzan and Jones. Final Decision, 2020 WL 3848045, at *1–2. The Board instituted review. See id. In its final written decision, the Board determined that Clear-Vu failed to prove by a preponderance of the evidence that claims 1 and 3 were anticipated by Nitzan, and that claims 2 and 4 would have been obvious over Nitzan in view of Jones. Id. at *22. Specifically, the Board found that Clear-Vu failed to demonstrate “that a 1.0 survival fraction measured for Nitzan’s non-ALA induced MRSA demon- strates ‘inactivation,’” which the parties agreed means, in the context of the ’706 patent, that the “bacteria are killed, or damaged so as to reduce or inhibit bacterial replication.” Id. at *10, 16; see also ’706 patent col. 2 ll. 44–46. The Board, however, agreed with Clear-Vu that claims 1 and 3 would have been obvious over Ashkenazi in view of Nitzan, and that claims 2 and 4 would likewise have been obvious over Ashkenazi in view of Nitzan and Jones. Final Decision, 2020 WL 3848045, at *22. Specifically, it found “Ashkenazi and Nitzan teach or suggest all the limi- tations of claims 1 and 3,” and that a person of ordinary skill in the art would have been motivated to combine these two references and “would have had a reasonable expecta- tion of successfully doing so.” Id. at *20. Case: 20-2243 Document: 28 Page: 8 Filed: 11/04/2021

8 UNIVERSITY OF STRATHCLYDE v. CLEAR-VU LIGHTING LLC

Regarding the scope and content of the prior art, the Board concluded that the only dispute was whether the prior art taught exposing bacteria to light without using a photosensitizer, finding that the “combined teachings of Nitzan and Ashkenazi” disclosed this limitation. Id. at *12–13 (emphasis added). In doing so, the Board sum- marily dismissed Strathclyde’s arguments that neither ref- erence taught this limitation, reasoning that Strathclyde “focus[ed] on Ashkenazi individually,” not the combined teachings of Ashkenazi and Nitzan. Id. at *13. The Board then found that a person of ordinary skill would have had a reasonable expectation of “inactivating MRSA using 407–420 nm light without applying a photo- sensitizer based on the combined teachings of Ashkenazi and Nitzan.” Id. at *16. Although neither Ashkenazi nor Nitzan achieved inactivation of any bacteria without using a photosensitizer and, as the Board found, Nitzan failed to achieve any inactivation when it exposed MRSA to 407–420 nm light without applying a photosensitizer, id. at *10, the Board nonetheless found that a skilled artisan would have reasonably expected “some” amount of inacti- vation because the claims “do not require any specific amount of inactivation,” id. at *16, 18. In doing so, the Board relied largely on Ashkenazi’s teachings that increas- ing the light doses, the number of illuminations, and the length of time the bacteria are cultured resulted in greater inactivation for both the ALA and non-ALA P. acnes to sup- port its finding that a skilled artisan applying Ashkenazi’s teachings would expect at least “some” inactivation for non- ALA MRSA. Id. at *18. The Board ultimately concluded that claims 1 and 3 were unpatentable as obvious over Ash- kenazi and Nitzan. Id. at *20. Because Strathclyde did not substantively address Clear-Vu’s arguments with respect to claims 2 and 4, instead relying on the same arguments it made with respect to claims 1 and 3, the Board deter- mined that claims 2 and 4 were unpatentable as obvious Case: 20-2243 Document: 28 Page: 9 Filed: 11/04/2021

UNIVERSITY OF STRATHCLYDE v. CLEAR-VU LIGHTING LLC 9

over Ashkenazi, Nitzan, and Jones for the same reasons as claims 1 and 3. Id. at *21. Strathclyde appeals, and we have jurisdiction under 28 U.S.C. § 1295(a)(4)(A). DISCUSSION We review the Board’s legal conclusions de novo and its factual findings for substantial evidence. Pers. Web Techs. v. Apple, Inc., 848 F.3d 987, 991 (Fed. Cir. 2017). “The sub- stantial evidence standard asks ‘whether a reasonable fact finder could have arrived at the agency’s decision,’ and ‘in- volves examination of the record as a whole, taking into ac- count evidence that both justifies and detracts from an agency’s decision.’” OSI Pharms., LLC v. Apotex Inc., 939 F.3d 1375, 1381–82 (Fed. Cir. 2019) (quoting In re Gartside, 203 F.3d 1305, 1312 (Fed. Cir. 2000)). “Obviousness is a question of law based on underlying findings of fact.” OSI Pharms., 939 F.3d at 1382 (quoting In re Kubin, 561 F.3d 1351, 1355 (Fed. Cir. 2009)). An ob- viousness determination generally requires a finding that “all claimed limitations are disclosed in the prior art,” PAR Pharm., Inc. v. TWI Pharms., Inc., 773 F.3d 1186, 1194 (Fed. Cir. 2014); cf. Koninklijke Philips N.V. v. Google LLC, 948 F.3d 1330, 1337–38 (Fed. Cir. 2020) (explaining that the common knowledge of a skilled artisan can be used to supply a missing limitation in some circumstances), and “that a person of ordinary skill in the art would have been motivated to combine or modify the teachings in the prior art and would have had a reasonable expectation of success in doing so,” OSI Pharms., 939 F.3d at 1382 (quoting Re- gents of Univ. of Cal. v. Broad Inst., Inc., 903 F.3d 1286, 1291 (Fed. Cir. 2018)). Whether the prior art discloses a claim limitation, whether a skilled artisan would have been motivated to modify or combine teachings in the prior art, and whether she would have had a reasonable expec- tation of success in doing so are questions of fact. Tech. Case: 20-2243 Document: 28 Page: 10 Filed: 11/04/2021

10 UNIVERSITY OF STRATHCLYDE v. CLEAR-VU LIGHTING LLC

Consumer Prods., Inc. v. Lighting Sci. Grp. Corp., 955 F.3d 16, 22 (Fed. Cir. 2020); OSI Pharm., 939 F.3d at 1382. On appeal, Strathclyde challenges the Board’s obvious- ness determination, arguing that the Board erred in find- ing that the combination of Ashkenazi and Nitzan teaches inactivating one or more Gram-positive bacteria without using a photosensitizer. It also asserts that the Board’s finding of a reasonable expectation of success is not sup- ported by substantial evidence. We address each issue in turn. I We begin by addressing the Board’s erroneous finding that the prior art disclosed all claim limitations. Both par- ties appear to agree that most of the limitations found in claims 1 and 3 are disclosed by Ashkenazi or Nitzan; the only dispute is whether these references teach inactivating one of the claimed Gram-positive bacteria without using a photosensitizer. The Board’s finding that this was taught by the combination of Ashkenazi and Nitzan is not sup- ported by substantial evidence. Claims 1 and 3 require both exposing the claimed Gram-positive bacteria to 400–420 nm blue light without using a photosensitizer and that the bacteria are inacti- vated as a result. While Ashkenazi discloses culturing P. acnes both with and without ALA such that it achieves inactivation after exposure to 407–420 nm blue light for (1) ALA P. acnes and (2) non-ALA P. acnes, the parties agree that the media Ashkenazi used to culture P. acnes contained the vitamin riboflavin, a photosensitizer. See Fi- nal Decision, 2020 WL 3848045, at *12; see also J.A. 2139–40 (Goodrich Decl. ¶ 121). It follows, then, that all of Ashkenazi’s P. acnes cultures were grown in the pres- ence of a photosensitizer—either riboflavin alone, or ribo- flavin together with ALA. We see nothing in Ashkenazi that discloses or suggests inactivating P. acnes, or any other bacteria, without using a photosensitizer. Case: 20-2243 Document: 28 Page: 11 Filed: 11/04/2021

UNIVERSITY OF STRATHCLYDE v. CLEAR-VU LIGHTING LLC 11

In addition, although Nitzan provides an example in which MRSA and other bacteria were exposed to 407–420 nm blue light without ALA or any other photosen- sitizer, there is no evidence that Nitzan successfully achieved inactivation under this condition. Indeed, in find- ing that Nitzan did not anticipate claims 1 and 3, the Board found Clear-Vu failed to establish that Nitzan’s non-ALA MRSA demonstrated “inactivation” as required by the claims. See Final Decision, 2020 WL 3848045, at *10. In making its contrary finding that the combination of references disclosed this limitation, the Board credited Clear-Vu’s argument that a skilled artisan would have pre- pared a MRSA culture according to the method described in Nitzan—which, unlike the media used to culture P. ac- nes in Ashkenazi’s experiments, would not have contained riboflavin—and applied Ashkenazi’s teaching that increas- ing the light energy, number of illuminations, and length of time the bacteria are cultured may result in greater in- activation for both ALA and non-ALA bacteria, as was shown for P. acnes. Id. at *13. The Board found that this combination, therefore, “disclose[d] exposing bacteria to light without using a photosensitizer.” Id. Given neither Ashkenazi nor Nitzan teaches or suggests inactivation of any bacteria without using a photosensitizer, we fail to see why a skilled artisan would opt to entirely omit a photo- sensitizer when combining these references. Indeed, the Board articulated no rational basis6—and we discern

6 Although the Board relied on and found persuasive the testimony of Clear-Vu’s expert, Dr. Sulzinski, discuss- ing this modification to Nitzan’s non-ALA MRSA, see id. at *13 (citing J.A. 2336 (Sulzinski Decl. ¶ 77)), the Board discredited testimony from this same paragraph in Dr. Sul- zinski’s declaration in its anticipation analysis because he “merely state[d] . . . his opinion, and offer[ed] no evidence” to support his statements, id. at *10 (citing J.A. 2336 Case: 20-2243 Document: 28 Page: 12 Filed: 11/04/2021

12 UNIVERSITY OF STRATHCLYDE v. CLEAR-VU LIGHTING LLC

none—for combining Ashkenazi’s P. acnes experiments, which at all times used a photosensitizer, with Nitzan’s non-ALA MRSA experiment, which did not achieve inacti- vation, to arrive at an embodiment in which MRSA is inac- tivated by exposing it to 407–420 nm blue light without using a photosensitizer. We find it particularly relevant that Nitzan itself disclosed such a photosensitizer-free em- bodiment and was wholly unsuccessful in achieving inacti- vation. Nor are we persuaded that Ashkenazi and Nitzan, ei- ther individually or in combination, “disclose[] the particu- lar ‘inactivating’ and ‘inactivates’ language” found in claims 1 and 3 as Clear-Vu suggests. Appellee’s Br. 57. Importantly, the claims require that the inactivation is a result of exposing bacteria to 400–420 nm light without us- ing a photosensitizer, which is neither taught nor suggested by the prior art of record. We decline Clear-Vu’s invitation to read the inactivation limitation in isolation, divorced from the claim as a whole. Cf. ATD Corp. v. Lydall, Inc., 159 F.3d 534, 546 (Fed. Cir. 1998) (“Obviousness cannot be based on the hindsight combination of components selec- tively culled from the prior art to fit the parameters of the patented invention.” (cleaned up)). On this record, we conclude that no reasonable fact finder could have found that the combination of Ashkenazi and Nitzan discloses inactivating one or more Gram-posi- tive bacteria without using a photosensitizer. The Board’s finding to the contrary is not supported by substantial evi- dence. II We turn next to the Board’s findings on reasonable ex- pectation of success. The Board found that a skilled artisan

(Sulzinski Decl. ¶ 77)). This inconsistency in the Board’s credibility findings further undermines its findings here. Case: 20-2243 Document: 28 Page: 13 Filed: 11/04/2021

UNIVERSITY OF STRATHCLYDE v. CLEAR-VU LIGHTING LLC 13

would have expected that MRSA could be inactivated by blue light without using a photosensitizer due to the pres- ence of at least some amount of naturally produced porphy- rin in the bacteria. We disagree. The only support for such a finding is pure conjecture coupled with hindsight reliance on the teachings in the ’706 patent. Starting with the references themselves, neither Ash- kenazi nor Nitzan provides a skilled artisan with any evi- dence or data or other promising information showing successful inactivation of P. acnes, MRSA, or any other bac- teria without using a photosensitizer. These references thus contain no suggestion that a skilled artisan would rea- sonably expect that MRSA or one of the other claimed Gram-positive bacteria could be inactivated upon exposure to 407–420 nm blue light without using a photosensitizer. The Board nevertheless found that a skilled artisan would have expected that MRSA could be inactivated by 407–420 nm blue light without using a photosensitizer be- cause, as both parties agree, MRSA naturally produces “at least some” amount of endogenous porphyrins. Final Deci- sion, 2020 WL 3848045, at *17. The Board, relying on Ash- kenazi’s teaching that “blue light may” inactivate “other bacterial cells that produce porphyrins,” Ashkenazi at 21 (emphases added), reasoned that because light-activated porphyrin molecules were shown in Ashkenazi to cause in- activation for P. acnes (even though Ashkenazi’s experi- ments applied a photosensitizer), the fact that MRSA has some level of endogenous porphyrin suggests to a skilled artisan that MRSA, too, would exhibit some amount of in- activation after exposure to 407–420 nm blue light. But there is simply no evidence of record at the time of the ’706 patent to support this assumption. In fact, the evi- dence of record—Nitzan and an earlier publication Case: 20-2243 Document: 28 Page: 14 Filed: 11/04/2021

14 UNIVERSITY OF STRATHCLYDE v. CLEAR-VU LIGHTING LLC

authored by Dr. Nitzan in 1999 (Nitzan 1999 7)—shows the opposite, illuminating the error in the Board’s finding. In the late 1990s, Dr. Nitzan and his colleagues “exam- ined the effects of the accumulation of endogenous porphy- rins on” MRSA. Nitzan 1999, Abstract; see also id. at 270 (explaining that the tested S. aureus strain was methicil- lin-resistant). As with the later studies described in Ash- kenazi and Nitzan, Dr. Nitzan exposed both the ALA and non-ALA MRSA to doses of 400–450 nm blue light ranging from 0–50 J/cm2. The survival fraction at each light dose for the non-ALA MRSA was 1.0, id. at 274 (Fig. 3(a)), mean- ing there was no decrease in viability and thus no inactiva- tion, whereas the viability of the ALA MRSA decreased by “3–4 orders of magnitude,” id. at 273. Dr. Nitzan reported the same results for the non-ALA MRSA only a couple of years later. See Nitzan at 433 & Tbl. 5 (reporting a 1.0 sur- vival fraction for non-ALA MRSA after illumination with 50 J/cm2 blue light, meaning “no decrease in viability”). Dr. Nitzan’s experiments thus are directly contrary to the Board’s rationale for why a skilled artisan would have expected success in inactivating MRSA in the absence of a photosensitizer. Even accepting as true that MRSA natu- rally has “at least some” amount of endogenous porphyrins, the data Dr. Nitzan reported would have indicated to a per- son of ordinary skill in the art that those natural levels are insufficient to inactivate MRSA using 407–420 nm blue light without also using a photosensitizer. In finding that Clear-Vu had established a reasonable expectation of success, the Board dismissed the data re- ported in Nitzan and Nitzan 1999 because, in its view, it

7 Y. Nitzan & M. Kauffman, Endogenous Porphyrin Production in Bacteria by δ-Aminoleavulinic Acid and Sub- sequent Bacterial Photoeradication, 14 Lasers Med. Sci. 269 (1999). Case: 20-2243 Document: 28 Page: 15 Filed: 11/04/2021

UNIVERSITY OF STRATHCLYDE v. CLEAR-VU LIGHTING LLC 15

would lead a skilled artisan to the “limited conclusion that non-ALA-treated MRSA incubated for 4 hours and treated with a single dose of 50 J/cm2 light showed no decrease in activity.” Final Decision, 2020 WL 3848045, at *18 (em- phasis added); id. (“We reach a similar conclusion for the data reported in Nitzan 1999, which only shows data up to a light intensity of 50 J/cm2, the same intensity used in Nitzan.”). The Board based its finding largely on the fact that neither Nitzan nor Nitzan 1999 tested light doses above 50 J/cm2, reasoning that using an increased light dose, such as the 75 J/cm2 dose used in Ashkenazi’s exper- iments, would cause inactivation of MRSA based on Ash- kenazi’s finding that increased light doses resulted in greater inactivation for both ALA and non-ALA P. acnes. Id. at *17–18. The Board’s reasoning finds no support in the record. The only evidence before the Board on this issue was the unrebutted testimony of Strathclyde’s expert, Dr. Goodrich, who testified the opposite. See Oral Arg. at 12:22–15:05, http://oralarguments.cafc.uscourts.gov/de- fault.aspx?fl=20-2243_05032021.mp3 (Clear-Vu conceding it did not provide rebuttal testimony on this point). Dr. Goodrich explained that a skilled artisan “viewing th[e] graphical representation of Nitzan 1999’s data would have clearly expected the curve of the ALA-absent S. aureus [MRSA] data to remain flat as dosage increased to 75 J/cm2,” J.A. 2133 (Goodrich Decl. ¶ 107) (emphases added), as shown below in his annotated version of Nitzan 1999’s Figure 3(a): Case: 20-2243 Document: 28 Page: 16 Filed: 11/04/2021

16 UNIVERSITY OF STRATHCLYDE v. CLEAR-VU LIGHTING LLC

J.A. 2132–33 (Goodrich Decl. ¶ 106). Dr. Goodrich then concluded that, to a skilled artisan, “the data in Nitzan 1999 would have indicated that the response of MRSA without ALA to blue light at 75 J/cm2 would have been the same as 50 J/cm2—no bacteria would have been killed.” J.A. 2133 (Goodrich Decl. ¶ 107) (emphasis added). The Board gave no weight to Dr. Goodrich’s testimony because, in its view, Dr. Goodrich did not provide the un- derlying facts forming the basis for his opinions. See Final Decision, 2020 WL 3848045, at *18 (citing 37 C.F.R. § 42.65(a)). We disagree. In his declaration, Dr. Goodrich specifically relied on and analyzed Nitzan 1999’s data in forming his opinions. Given that the Board’s stated reason for discrediting this written testimony is unsupported by the record before us, we see no reason to “defer[] to the spe- cial province of the Board to exercise its discretion concern- ing the credibility of expert witnesses” as Clear-Vu Case: 20-2243 Document: 28 Page: 17 Filed: 11/04/2021

UNIVERSITY OF STRATHCLYDE v. CLEAR-VU LIGHTING LLC 17

suggests we do here. Appellee’s Br. 49 (citing Yorkey v. Diab, 601 F.3d 1279, 1284 (Fed. Cir. 2010)). Thus, not only is there a complete lack of evidence in the record that any bacteria were inactivated after expo- sure to 407–420 nm blue light without using a photosensi- tizer, there is also evidence showing that others had failed to inactivate MRSA—one of the claimed Gram-positive bac- teria—without using a photosensitizer, despite experi- menting with different light doses and different wavelength ranges of blue light. We have found that such failures undermine a finding of a reasonable expectation of success. For example, in OSI Pharmaceuticals, we reversed the Board’s obviousness determination because its finding of a reasonable expectation of success was not supported by substantial evidence. 939 F.3d at 1384–85. The claims at issue in OSI were directed to a method of treating non-small cell lung cancer (NSCLC) using a therapeuti- cally effective amount of a drug known as erlotinib. Id. at 1378–79, 1384. The Board found that two different prior art combinations would have provided a skilled artisan with a reasonable expectation of success in using erlotinib to treat NSCLC in a mammal. Id. at 1384. We disagreed, explaining that none of these references contained “data or other promising information regarding erlotinib’s efficacy in treating NSCLC.” Id. at 1384; see also id. at 1385. We also found it significant that, during the time of the inven- tion, there was a 99.5% failure rate for other drugs entering Phase II clinical trials that, like erlotinib, were targeted for the treatment of NSCLC. Id. at 1385. Thus, given the “fail- ure rate” and lack of “data or any other reliable indicator of success,” we found that “the only reasonable expectation at the time of the invention was failure, not success.” Id. Such is the case here. In view of Dr. Nitzan’s reported failures for MRSA and lack of any “reliable indicator of suc- cess,” we fail to see how Ashkenazi’s prophetic statement Case: 20-2243 Document: 28 Page: 18 Filed: 11/04/2021

18 UNIVERSITY OF STRATHCLYDE v. CLEAR-VU LIGHTING LLC

about what “may” happen when “other bacterial cells” are exposed to blue light would lead a skilled artisan to reason- ably expect that MRSA could be inactivated when exposed to 407–420 nm blue light without using a photosensitizer. The Board’s finding that a skilled artisan would expect at least “some” inactivation for non-ALA MRSA—in view of Ashkenazi’s teaching that increasing the light doses, the number of illuminations, and the length of time the bacte- ria are cultured can result in greater inactivation based on experiments that were conducted using a photosensitizer— is not supported by substantial evidence. In attempting to support the Board’s findings, Clear-Vu argues that support can be found in the ’706 pa- tent itself. See Appellee’s Br. 58 (“It defies logic to conclude that inactivating MRSA by applying Ashkenazi’s technique to Nitzan’s MRSA would not result in inactivating MRSA when the patentee obtained that very result using the same technique.” (emphasis added)). But “[t]he inventor’s own path itself never leads to a conclusion of obviousness; that is hindsight. What matters is the path that the person of ordinary skill in the art would have followed, as evi- denced by the pertinent prior art.” Otsuka Pharm. Co., v. Sandoz, Inc., 678 F.3d 1280, 1296 (Fed. Cir. 2012). Given the record on appeal, as with OSI, we are left to con- clude that “[i]t is only with the benefit of hindsight that a person of skill in the art would have had a reasonable ex- pectation of success in view of the asserted references.” OSI Pharms., 939 F.3d at 1385. Contrary to Clear-Vu’s arguments, we do not hold that “absolute predictability” or “guaranteed success” is re- quired to find a reasonable expectation of success. Appel- lee’s Br. 43, 45 (emphasis omitted). To be sure, we have repeatedly rejected that notion. See, e.g., Acorda Thera- peutics, Inc. v. Roxane Lab’ys, Inc., 903 F.3d 1310, 1333 (Fed. Cir. 2018) (“This court has long rejected a require- ment of ‘[c]onclusive proof of efficacy’ for obviousness.” (al- teration in original) (quoting Hoffmann-La Roche Inc. Case: 20-2243 Document: 28 Page: 19 Filed: 11/04/2021

UNIVERSITY OF STRATHCLYDE v. CLEAR-VU LIGHTING LLC 19

v. Apotex Inc., 748 F.3d 1326, 1331 (Fed. Cir. 2014))); OSI Pharms., 939 F.3d at 1385; PAR Pharm., 773 F.3d at 1198. We reaffirm today that absolute predictability of success is not required, only a reasonable expectation. In this case, where the prior art evidences only failures to achieve that at which the inventors succeeded, no reasonable fact finder could find an expectation of success based on the teachings of that same prior art. The Board’s finding is not supported by substantial evidence, and we therefore reverse its obvi- ousness determination. CONCLUSION We have considered the parties’ remaining arguments but find them unpersuasive. For the foregoing reasons, we reverse the Board’s obviousness determination. REVERSED