Before JOLLY, DENNIS, and HIGGINSON, Circuit Judges.STEPHEN A. HIGGINSON, Circuit Judge:The substantive issue in this appeal is whether a jury should hear Xitronix Corporation’s claim that KLA-Tencor Corporation violated the Sherman Act’s prohibition of monopolies by obtaining a patent through a fraud on the U.S. Patent and Trademark Office (“PTO”). What must first be decided, however, is whether we can reach that issue despite the Federal Circuit’s exclusive jurisdiction over cases arising under federal patent law. That court transferred this case to us, but we find it implausible that we are the proper court to decide this appeal. With respect, therefore, we transfer it to the U.S. Court of Appeals for the Federal Circuit.IThis is the third round of litigation between Plaintiff-Appellant Xitronix Corporation and Defendant-Appellee KLA-Tencor Corporation (“KLA”), competitors in the “semiconductor wafer optical inspection market.” Optical inspection technology is used for quality control in the production of semiconductor wafers, which are essential components of circuits in computers and other electronic devices. We understand from the parties that an optical inspection device employs two lasers, a “pump” beam and a “probe” beam, in tandem. The pump beam heats the surface of a semiconductor sample. The probe beam, in turn, detects changes in the semiconductor surface. The device converts the changes detected by the probe beam into an electrical signal, which it then measures. The device can thereby precisely observe the composition of the semiconductor sample, helping manufacturers ensure that their processes are working as intended.ALitigation began in 2008 with Xitronix seeking a declaratory judgment against KLA. According to Xitronix, KLA was and is the dominant player in the semiconductor optical inspection market, with approximately eighty- percent market share. KLA had examined the technology that Xitronix was then bringing to market and amended a pending patent application to cover Xitronix’s technology. This application resulted in the issuance of U.S. Patent 7,362,441 (“the ’441 patent”). In this first lawsuit, Xitronix sought a declaration of non-infringement and of the ’441 patent’s invalidity.In November 2010, a jury entered a verdict in Xitronix’s favor. When the district court entered final judgment in January 2011, it explained that the central issue at trial was the wavelength of the probe beam used by Xitronix. The claims of the ’441 patent at issue in the case specified a wavelength between 335 and 410 nanometers (nm) and said that such wavelength “is selected to substantially maximize the strength of the output signals corresponding to the modulated optical reflectivity response.”[1] The probe beam in Xitronix’s device was fixed at a wavelength of 373 nm, putting it and KLA’s patent squarely in conflict. The jury found that Xitronix’s technology infringed one claim of KLA’s ’441 patent but that this claim was anticipated by prior art: the “Therma-Probe” device and an earlier patent, the ’611 or “Alpern” patent. The jury also found three other claims of KLA’s ’441 patent invalid due to obviousness. The district court ruled that ample evidence supported the jury’s findings. It identified two additional pieces of prior art, Batista and Mansanares: “[E]ach and every element of the asserted claims were present in the combination of prior art Therma-Probe, Batista, Mansanares, and the ’611 [Alpern] patent.” The district court also ruled one of the claims invalid as indefinite. KLA did not appeal the judgment in the ’441 litigation.In March 2011, Xitronix commenced the second suit, bringing business tort claims against KLA for publicly accusing Xitronix of patent infringement. The district court, which remanded the case to Texas state court, later explained that the state district court ruled in favor of KLA “for unspecified reasons.” Neither party advises that this second litigation has any bearing on the present appeal.BThe present case began in December 2014. Xitronix alleged a single Walker Process claim: monopolization through use of a patent obtained by fraud on the PTO.[2] The patent purportedly resulting from KLA’s fraud on the PTO is U.S. Patent No. 8,817,260 (“the ’260 patent”). It is a continuation of an earlier patent, the ’486 patent, which was itself a continuation of the ’441 patent at issue in the parties’ first litigation. KLA filed the application that yielded the ’260 patent in November 2009, U.S. Application No. 12/616,710,[3] a year before the jury entered its verdict invalidating the ’441 patent. The litigation of ’441 and the prosecution of ’260 unfolded in tandem. It is KLA’s representations to the PTO concerning the ’441 litigation while conducting the ’260 prosecution that are at issue here.In February 2010, KLA submitted an Informational Disclosure Statement (“IDS”) with sixty works potentially containing relevant prior art. This IDS included the key sources on which the jury would invalidate the ’441 patent later that year as well as summary judgment briefing from the litigation. In August 2010, the PTO examiner, Layla Lauchman, initialed and signed the IDS, thereby acknowledging these sources. On November 5, 2010, the jury returned its verdict invalidating the ’441 patent. On November 18, Michael Stallman, KLA’s patent prosecution attorney, submitted the jury’s verdict in the ’441 litigation to the PTO and sought to explain its meaning. He acknowledged an Office Action of August 18 that rejected the claims in the ’260 application on the grounds of “non-statutory obviousness-type double patenting” in view of the ’441 patent.[4] This means that, as of 2010, the PTO saw claims in the ’260 application as obvious in light of claims later invalidated in the ’441 litigation. Stallman responded to this rejection by agreeing to a “terminal disclaimer” of the claims in the pending ’260 application.[5]The district court entered final judgment in the ’441 litigation on January 31, 2011. On February 2, Lauchman issued a Notice of Allowability as to the ’260 patent application predicated on the terminal disclaimer to which KLA had agreed. On February 10, Stallman filed another IDS, bringing the final judgment in the ’441 litigation and the district court’s accompanying order to the PTO’s attention. He submitted a Request for Continued Examination as well.The PTO did not act on the application again for two years, by which time a new examiner, Willie Merrell, was handling it. His initials, dated July 12, 2013, appear on the IDS from February 2011 containing the final judgment and related documents, suggesting he had seen and considered the references. In an Office Action dated July 25, 2013, Merrell rejected much of the ’260 application. He did so without reference to the final judgment in the ’441 litigation, to the PTO’s prior non-statutory double patenting rejection, or to the materials on which the judgment in the ’441 litigation was based. Instead, he conducted a novel analysis based on other prior art further afield.Despite this rejection, the ’260 patent did eventually issue. Xitronix’s claims concern the actions by KLA making that possible. Stallman filed a response to Merrell in October 2013. The remarks in that filing are one basis for the fraud element of Xitronix’s Walker Process claim. Merrell responded in January 2014, standing by his previous rejection. Stallman responded in March 2014 with more remarks—another filing central to Xitronix’s Walker Process claim. A Notice of Allowability soon issued, in which Merrell briefly explained that KLA’s arguments “have been fully considered and are persuasive.” The ’260 patent issued in August 2014, and Xitronix’s Walker Process suit followed that December.Xitronix alleged that KLA’s procurement of the ’260 patent impeded its ability to finance its entrance into the market for optical inspection technology. At summary judgment, the litigation focused on whether Xitronix had created issues of material fact as to two elements of fraud on the PTO: whether KLA had made material misrepresentations or omissions to the PTO, and whether those were a but-for cause of the ’260 patent’s issuance. The district court found that Stallman’s remarks in his October 2013 and March 2014 filings were confined to those pieces of prior art specifically addressed by Merrell in previous Office Actions and contained no broader misrepresentations. To the extent Stallman mischaracterized the prior art, the district court reasoned, this was permissible attorney argument, not fraud. Stallman was free to make such argument, and the examiner was free to reject it, because Stallman had submitted all relevant materials from the ’441 litigation already.The district court also found no but-for causation. Notably, it was not because the court viewed the ’441 and ’260 patents as dissimilar, such that the former would not control the latter. Indeed, the court saw them as similar. Rather, it saw the PTO as making a fully conscious and informed choice. Granting summary judgment to KLA, the district court wrote:Although Xitronix has repeatedly argued that the examiner was unaware of the jury verdict and final judgment invalidating the claims at issue, the Court suspects the examiner was in fact aware of the Court’s holding but chose to ignore it. It would not be the first time the PTO, an administrative agency, overrode a final judgment of an Article III court, and it will likely not be the last.That is, according to the district court, it could not be said that the PTO relied upon, and was thus defrauded by, KLA’s alleged misrepresentations; rather, the PTO had a mind of its own. The district court’s grant of summary judgment to KLA brought the case to a close, precipitating this appeal.CXitronix’s appeal went originally to the Federal Circuit. Before oral argument in the case, the panel of that court ordered briefing on transferring the case to our court for lack of jurisdiction. Xitronix Corp. v. KLA-Tencor Corp., 882 F.3d 1075, 1076 (Fed. Cir. 2018). The parties, who agreed that the case belonged in the Federal Circuit, spent the bulk of oral argument on the subject and briefed the issue further afterwards.[6] Despite the parties’ consensus, the panel was not persuaded, ruling based on Gunn v. Minton, 568 U.S. 251 (2013), that it lacked jurisdiction. Following the transfer order, KLA petitioned for en banc rehearing,[7] which the Federal Circuit rejected by a vote of ten to two. Xitronix Corp. v. KLA-Tencor Corp., 892 F.3d 1194 (Fed. Cir. 2018). Judge Pauline Newman dissented from that ruling, taking the panel to task for initiating “a vast jurisdictional change for the regional circuits as well as the Federal Circuit.” Id. at 1196.II“Federal courts are courts of limited jurisdiction,” possessing “only that power authorized by Constitution and statute.” Kokonnen v. Guardian Life Ins. Co. of Am., 511 U.S. 375, 377 (1994). “The requirement that jurisdiction be established as a threshold matter ‘spring[s] from the nature and limits of the judicial power of the United States’ and is ‘inflexible and without exception.’” Steel Co. v. Citizens for a Better Env’t, 523 U.S. 83, 94-95 (1998) (quoting Mansfield, C. & L. M. Ry. Co. v. Swan, 111 U.S. 379, 382 (1884)). Consequently, “[w]e must always be sure of our appellate jurisdiction and, if there is doubt, we must address it, sua sponte if necessary.” Casteneda v. Falcon, Jr., 166 F.3d 799, 801 (5th Cir. 1999).IIIUnder the law that prevailed for many years, it was clear that a standalone Walker Process claim such as this would belong in the Federal Circuit. See Christianson v. Colt Indus. Operating Corp., 486 U.S. 800, 809 (1988); Ritz Camera & Image, LLC v. SanDisk Corp., 700 F.3d 503 (Fed. Cir. 2012). The parties have not identified any case that involved solely Walker Process claims and that was decided by a circuit court other than the Federal Circuit. Indeed, the last Walker Process case decided by the Fifth Circuit was in 1975, before the Federal Circuit was created. See Becton, Dickinson, & Co. v. Sherwood Med. Indus., Inc., 516 F.2d 514 (5th Cir. 1975); Federal Courts Improvement Act of 1982, Pub. L. No. 97-164, § 127(a), 96 Stat. 25, 37 (1982).The Federal Circuit nevertheless transferred the case to us, based on a jurisdictional analysis that we must accept if it is at least “plausible.” See Christianson, 486 U.S. at 819. The Federal Circuit transferred the case because it understood Gunn v. Minton, 568 U.S. 251 (2013), to change the law governing the allocation of cases between it and the regional circuits. There are compelling reasons to think that Gunn did not, but the answer to this question is not determinative here. Under any reading of Gunn, we deem it implausible that we can decide this appeal.AThe Federal Circuit has “exclusive jurisdiction of an appeal from a final decision of a district court of the United States . . . in any civil action arising under . . . any Act of Congress relating to patents or plant variety protection.” 28 U.S.C. § 1295(a)(1). Before 2011, the statute read differently, conferring exclusive jurisdiction “of an appeal from a final decision of a district court of the United States . . . if the jurisdiction of that court was based, in whole or in part, on section 1338 of this title.” Section 1338(a), in turn, gave district courts “original jurisdiction of any civil action arising under any Act of Congress relating to patents,” among other subjects. Id. § 1338(a).The Supreme Court construed the earlier version of the statute in Christianson v. Colt Industries Operating Corp., 486 U.S. 800 (1988). Christianson is the primary guide to our decision here because it furnished several rules that control the present case. Christianson was a former employee of Colt, the famous gunmaker, and had gone into business selling M- 16 replacement parts. Id. at 804. Colt was telling customers that Christianson was illegally misappropriating its trade secrets, leading Christianson to sue both for tortious inference with business relations and for violations of the Sherman Act. Id. at 805. Christianson argued that Colt could not claim trade- secret protection because its patents were invalid, and indeed, the district court invalidated nine Colt patents. Id. at 806.The Supreme Court had to decide whether the appeal belonged in the Seventh Circuit or the Federal Circuit. Id. at 806-07. Appeal had been taken to the Federal Circuit, which transferred it to the Seventh Circuit, and that court then transferred it back. Id. Under protest, the Federal Circuit then decided the case in the “interest of justice.” Id. at 807.With the Federal Circuit’s jurisdictional statute, § 1295, tied to § 1338, the Court had to construe the latter provision. 486 U.S. at 807. It noted that § 1338 contained an “arising under” formulation quite like the federal-question statute, § 1331, and was therefore susceptible to a complication that has bedeviled the latter statute: what to do with causes of action not created by federal law that nevertheless turn on substantial questions of federal law? The Federal Circuit’s equivalent dilemma was deciding what to do with causes of action not created by federal patent law that nevertheless implicate it.The Court noted that federal-question jurisdiction had long included state claims turning on substantial federal questions. 486 U.S. at 808; see Merrell Dow Pharm. Inc. v. Thompson, 478 U.S. 804 (1986); Franchise Tax Bd. of Cal. v. Constrn. Laborers Vacation Trust, 463 U.S. 1 (1983); Gully v. First Nat’l Bank in Meridian, 299 U.S. 109 (1936). The Court then announced the following rule:Linguistic consistency, to which we have historically adhered, demands that § 1338(a) jurisdiction likewise extend only to those cases in which a well-pleaded complaint establishes either that federal patent law creates the cause of action or that the plaintiffs right to relief necessarily depends on resolution of a substantial question of federal patent law, in that patent law is a necessary element of one of the well-pleaded claims.486 U.S. at 809.That did not resolve the case before it, however, because not all of Christianson’s claims depended on resolving substantial questions of patent law. 486 U.S. at 810-11. Consequently, the Court held that lower courts should determine whether all claims in the plaintiff’s well-pleaded complaint necessarily depended on the resolution of a substantial question of federal patent law. Id. “[A] claim supported by alternative theories in the complaint may not form the basis for § 1338(a) jurisdiction unless patent law is essential to each of those theories.” Id. at 810. Accordingly, Christianson’s case did not belong in the Federal Circuit.The Court also addressed the “peculiar jurisdictional battle” between the Seventh and Federal Circuits. 486 U.S. at 803. As noted, the Federal Circuit had the case first but transferred it. Id. at 817. This ruling was the law of the case, from which the Seventh Circuit departed. Id. This was not impermissible: “A court has the power to revisit prior decisions of its own or of a coordinate court in any circumstance, although as a rule courts should be loath to do so in the absence of extraordinary circumstances, such as where the initial decision was clearly erroneous and would work a manifest injustice.” Id. (quotation omitted). Receiving the case again, the Federal Circuit disputed it had jurisdiction but decided the case anyway. The Court ruled that this was error. Id. at 818. But if the Federal Circuit erred by deciding the case, how then to bring this interminable “game of jurisdictional ping-pong,” id., to a close? The Court gave the following guidance: “Under law-of-the-case principles, if the transferee court can find the transfer decision plausible, its jurisdictional inquiry is at an end.” Id. at 819 (emphasis added).Following Christianson, the Federal Circuit has regularly exercised jurisdiction over Walker Process claims. See, e.g., Ritz Camera & Image, 700 F.3d at 506; Nobelpharma AB v. Implant Innovations, Inc., 141 F.3d 1059, 1072-73 (Fed. Cir. 1998). In so doing, and vital to our analysis here, the Federal Circuit has been clear in its view that “the determination of fraud before the PTO necessarily involves a substantial question of patent law.” In re Ciprofloxacin Hydrochloride Antitrust Litig., 544 F.3d 1323, 1330 n.8 (Fed. Cir. 2008) (“Cipro”) (citing Christianson, 486 U.S. at 808), abrogated on other grounds by F.T.C. v. Actavis, Inc., 570 U.S. 136 (2013).Other circuits have decided Walker Process cases, it should be said. Such cases have ended up in the regional circuits because of the line drawn in Christianson: cases depending solely on patent theories go to the Federal Circuit; cases not so dependent stay in the regional circuits. See, e.g., In re Lipitor Antitrust Litig., 855 F.3d 126, 146 (3rd Cir. 2017); In re DDAVP, 585 F.3d 677, 685 (2nd Cir. 2009); Kaiser Found. Health Plan, Inc. v. Abbott Labs., Inc., 552 F.3d 1033, 1041-42 (9th Cir. 2009).BThe foregoing is the backdrop to Gunn v. Minton, on which the Federal Circuit relied here. See Xitronix, 882 F.3d at 1077. The present question is whether, and how, Gunn altered the landscape just described.Gunn called for the Supreme Court to decide whether a state-law legal malpractice case arising from a patent infringement suit could be brought only in federal court. 568 U.S. at 253-56. Minton, a developer of software for trading securities, had hired Gunn, a patent lawyer, to sue NASDAQ and others for infringing Minton’s patent. Id. at 253-54. In the infringement case, the federal court had granted summary judgment against Minton, declaring his patent invalid. Id. at 254. Minton then sued Gunn for legal malpractice, arguing that Gunn had failed to raise a key argument in a timely manner. Id at 255. The state district court ruled for the lawyer. Id. On appeal, Minton made a novel argument: though he had filed the suit in state court, federal courts had exclusive jurisdiction because the suit raised a substantial question of federal patent law. Id. A divided Texas court of appeals disagreed, but a divided Texas Supreme Court ruled that Minton was right. Id. at 255-56.The United States Supreme Court reversed, holding that the case could be brought in state court. The Court noted that both the federal-question statute, § 1331, and the district courts’ patent jurisdiction statute, § 1338, were implicated. 568 U.S. at 257. The Court then applied a four-factor test that it had developed over the years to decide federal-question issues. “[F]ederal jurisdiction over a state law claim will lie if a federal issue is: (1) necessarily raised, (2) actually disputed, (3) substantial, and (4) capable of resolution in federal court without disrupting the federal-state balance approved by Congress.” Id. at 258 (citing Grable & Sons Metal Prods. Inc. v. Darue Eng. & Manuf., 545 U.S. 308, 314 (2005)).The Court acknowledged that Minton’s lawsuit against Gunn necessarily raised the disputed issue of Gunn’s handling of a patent case. 568 U.S. at 259. It focused on the third and fourth parts of the test, and there it found the case lacking. “The substantiality inquiry under Grable looks . . . to the importance of the issue to the federal system as a whole.” Id. at 260. The patent issue in Gunn and Minton’s case was “backward-looking,” “merely hypothetical,” and not likely to “change the real-world result of the prior federal patent litigation.” Id. at 261. As such, it had no importance for the federal system writ large.[8] The Court emphasized that it would upset the balance between state and federal judiciaries to move such legal malpractice cases exclusively into federal court, given the states’ “special responsibility for maintaining standards among members of the licensed professions.” Id. at 264 (quotation omitted).Gunn gave no indication that it meant to alter Christianson or the allocation of cases among the circuit courts. There was no occasion for it, because the case was appealed from a state’s highest court. On the contrary, the centrality of the Federal Circuit to patent adjudication was a premise of Gunns reasoning. 568 U.S. at 261-62. Against the argument that state-court adjudication of the patent issue in Gunn would undermine the uniformity of federal patent law, the Court said that “Congress ensured such uniformity” by vesting exclusive appellate jurisdiction in the Federal Circuit. Id.CSince Gunn, the Federal Circuit has incorporated a substantiality inquiry into determinations of its own jurisdiction. See, e.g., Neurorepair, Inc. v. The Nath Law Group, 781 F.3d 1340, 1345-49 (Fed. Cir. 2015); Jang v. Boston Sci. Corp., 767 F.3d 1334, 1336-38 (Fed. Cir. 2014). In this case, the court acknowledged that the case would require applying patent law but disputed the case’s substantiality. 882 F.3d at 1078. “Patent claims will not be invalidated or revived based on the result of this case,” and “the result [of the case] is limited to the parties and the patent involved in this matter.” Id. at 1078. It viewed any threat to the uniformity of patent law as insubstantial. Id.The court’s reasoning depended on several premises that we find implausible. First, the court said that there was no dispute about the validity of the patent at issue. 882 F.3d at 1078. In her dissent from denial of rehearing, Judge Newman responded that this was “a puzzling statement, for that is the dispute.” 892 F.3d at 1199. A finding of fraud on the PTO would render KLA’s ’260 patent effectively unenforceable in future cases. See C.R. Bard, 157 F.3d at 1367 (“Fraud in obtaining a United States patent is a classical ground of invalidity or unenforceability of the patent.”). Inequitable conduct is a defense to a claim of patent infringement that bars enforcement of the patent. Therasense Inc. v. Becton, Dickinson & Co., 649 F.3d 1276, 1285 (Fed. Cir. 2011) (en banc). Inequitable conduct resembles the fraud element of Walker Process, in that it requires proof of misrepresentation, scienter, and a showing of materiality or causation. Id. at 1290. Over time, it has evolved to be “virtually congruent with intentional fraud under Walker Process.” J. Thomas Roesch, Patent Law and Antitrust Law: Neither Friend nor Foe, but Business Partners, 13 Sedona Conf. J. 95, 100 (2012). Consequently, if this litigation determines that KLA defrauded the PTO in obtaining the ’260 patent, collateral estoppel principles would furnish a readymade inequitable conduct defense to any potential infringer whom KLA might sue. See Blonder-Tongue Labs., Inc. v. Univ. of Ill. Found., 402 U.S. 313, 330-34 (1971).Next, the Federal Circuit read its precedent predating Gunn in a manner at odds with our reading of that caselaw. In Nobelpharma, a Walker Process case, the en banc Federal Circuit said that “[w]hether conduct in the prosecution of a patent is sufficient to strip a patentee of its immunity from the antitrust laws is one of those issues that clearly involves our exclusive jurisdiction over patent cases.” 141 F.3d at 1067. The Federal Circuit distinguished Nobelpharma here, reasoning that Nobelpharma was not deciding the venue of the appeal, but whether to apply regional circuit or Federal Circuit precedent to various issues. 882 F.3d at 1078-79. This distinction strikes us as immaterial. The tests for both questions turn on the Federal Circuit’s exclusive jurisdiction over a given issue. It does not matter that Nobelpharma analyzed its jurisdiction for one purpose rather than the other.The court also sought to distinguish Cipro, a Walker Process case transferred from the Second to the Federal Circuit. 544 F.3d at 1323. Accepting the transfer, the Federal Circuit observed that “the determination of fraud before the PTO necessarily involves a substantial question of patent law.” Id. at 1330 n.8 (citing Christianson, 468 U.S. at 808). The court distinguished Cipro here because, as a transferred case, the Second Circuit’s jurisdictional analysis had only to meet the Christianson plausibility standard. But the Federal Circuit in Cipro stated its unqualified agreement with the Second Circuit’s analysis, making no reference to plausibility.We note one more case indicating that, before Gunn, the Federal Circuit understood fraud on the PTO to present a substantial question of federal patent law implicating its exclusive jurisdiction. In Ritz Camera & Image, an interlocutory appeal arose from a suit in the Northern District of California about the standing of certain plaintiffs to bring a Walker Process action. 700 F.3d at 505. Like the present case, this appeal presented solely a Walker Process issue, and the Federal Circuit did not even pause to consider its jurisdiction.Another basis for the Federal Circuit’s transfer decision is its interpretation of post-Gunn decisions from other circuit courts. The court cited In re Lipitor from the Third Circuit, which resolved a Walker Process claim in 2017. 882 F.3d at 1079 (citing In re Lipitor, 885 F.3d at 146). That case involved non-patent antitrust theories, however, so the Christianson rule clearly allocated it to the regional circuit. 855 F.3d at 146. Another case was Seed Co. Ltd. v. Westerman, 832 F.3d 325 (D.C. Cir. 2016), a legal malpractice case concerning the unsuccessful prosecution of a patent. Not being a Walker Process case, Seed Co. sheds no light on whether cases solely alleging fraud on the PTO no longer belong in the Federal Circuit. The court also cited one of our decisions, in which the court understood us to hold that we “had appellate jurisdiction in a case involving a state law claim based on fraud on the PTO.” 882 F.3d at 1080 (citing USPPS, Ltd. v. Avery Dennison Corp., 541 F. App’x 386, 390 (5th Cir. 2013)). On the contrary, USPPS involved fraud claims against a business and its lawyers following a failed patent prosecution, but it did not involve fraud on the PTO itself. 541 F. App’x at 388-90.Finally, the panel relied on an Eleventh Circuit case, MDS (Canada) Inc. v. Rad Source Techs., Inc., 720 F.3d 833 (11th Cir. 2013), that provides perhaps the strongest support for its decision to transfer this case to us. MDS was a breach of contract action concerning a licensing agreement between Nordion and Rad Source. Rad Source had three patents for blood irradiation devices, which it licensed to Nordion, such that Nordion would market and sell Rad Source’s RS 3000 device. Id. at 838. After a falling out, Rad Source began to develop a new product based on the same patents, the RS 3400, which it would sell independently of Nordion. Id. at 840. Litigation ensued, with Nordion alleging that Rad Source had breached their agreement by independently developing a product covered by one of the patents subject to the license agreement. Id. at 840. This injected an infringement issue into the case. During the litigation, Nordion learned that Rad Source had allowed that patent to lapse. Id. at 840-41. This gave Nordion an additional breach of contract theory. Id.The Eleventh Circuit ruled that it, and not the Federal Circuit, had jurisdiction over the appeal. 720 F.3d at 841. It reasoned that the case presented claims under state contract law and thus that the district court had exercised diversity jurisdiction, not original jurisdiction by virtue of a federal question or a federal patent issue. Id. There was a question of patent infringement in the case, but, like the patent issues in Gunn, it was backward- looking and insubstantial because it indisputably concerned a since-expired patent. Id at 842-43. Therefore, the issue was not substantial enough to implicate the district court’s “arising under” patent jurisdiction. Id. The court then proceeded to resolve the patent infringement issue. Id. at 846-48.In the present case, the Federal Circuit noted the Eleventh Circuit’s substantiality analysis in MDS and said that it “confirm[ed] the correctness of [their] decision” to transfer the case. 882 F.3d at 1079-80. Judge Newman pointed out two distinctions: first, that the patent at issue in MDS was expired, whereas KLA’s patent in the present case remains operative; and second, that MDS did not address whether the Walker Process element of fraud on the PTO implicates federal patent law. 892 F.3d at 1201. As such, it is only so helpful in figuring out whether Gunn requires the present case to be transferred away from the Federal Circuit.DTo reject the Federal Circuit’s transfer decision, we must not only disagree with its reasoning; we must find it implausible. We do not take this step lightly. With due regard for our colleagues on a coordinate court, we nevertheless conclude that it is implausible for us to resolve this appeal.The initial question is whether Gunn meant to change the Federal Circuit’s jurisdiction, in addition to changing district courts’ jurisdiction. Assuming that it did, we think that this appeal presents a substantial question in the sense that the Supreme Court has articulated. Under Gunn, “[t]he substantiality inquiry. . . [looks] to the importance of the issue to the federal system as a whole.” 568 U.S. at 260. The Court in Gunn relied on two examples: Grable, 545 U.S. 308, and Smith v. Kansas City Title & Trust Co., 255 U.S. 180 (1921). Grable concerned the validity of a foreclosure and sale by the IRS of a delinquent taxpayer’s property. 545 U.S. at 315. Smith concerned the constitutionality of certain federal bonds, challenged by a shareholder seeking to block a company from buying them. 255 U.S. at 201. Both cases put the legality of a federal action in question, in a manner that would have broader ramifications for the legal system. By contrast, Gunn, as a legal malpractice case, entailed a “merely hypothetical,” “backward-looking” review of a lawyer’s conduct regarding a now-invalid patent. 568 U.S. at 261. Nothing broader was at stake. Gunn also perceived no precedential or preclusive implications if a state court resolved the case.This case concerns a patent that is currently valid and enforceable, issued following a PTO proceeding heretofore viewed as lawful. This litigation has the potential to render that patent effectively unenforceable and to declare the PTO proceeding tainted by illegality. This alone distinguishes the present case from Gunn.[9] The adjudication of this Walker Process claim also implicates the interaction between the PTO and Article III courts. The district court’s acerbic statements about the PTO at summary judgment point to the complexity of relations between proceedings in federal court and before the PTO.Moreover, the fraud element of Xitronix’s claim can be adjudicated only with reference to patent law. Walker Process requires showing that a given statement or omission was “material to patentability.” C.R. Bard, Inc., 157 F.3d at 1364. Here, that requires reference to the bases of the ’441 patent’s invalidation (anticipation, obviousness, and indefiniteness), the significance of non-statutory double patenting rejections, the nature of prior art analysis by patent examiners, and more. Xitronix’s theories of fraud also put certain rules in issue. For example, Xitronix bases some of its theories on the regulations governing patent practitioners’ duties of candor to the PTO. See 37 C.F.R. §§ 1.56, 10.85 (2013), 11.301. This case therefore has the potential to set precedent on the precise scope of those duties. Compare KangaROOS U.S.A., Inc. v. Caldor, Inc., 778 F.2d 1571, 1576 (Fed. Cir. 1985), with Young v. Lumenis, Inc., 492 F.3d 1336, 1349 (Fed. Cir. 2007). Such precedent can profoundly affect the future conduct of practitioners before the PTO. Cf. Therasense, 649 F.3d at 1289-90 (explaining the ways that the inequitable conduct doctrine had altered patent practitioners’ behavior in PTO proceedings). To the extent we or other circuit courts differ from the Federal Circuit on these matters, we risk confusion for current practitioners and forum-shopping by future litigants.The foregoing assumes that Gunn changed the scope of the Federal Circuit’s jurisdiction, but there are compelling reasons to think that it did not. Gunn concerned the district courts’ jurisdictional statute, § 1338, not the Federal Circuit’s jurisdictional statute, § 1295. The Supreme Court never said it was changing the Federal Circuit’s caseload. It spoke only to the allocation of cases between the state and federal systems, not to the allocation of cases between the circuit courts. The Court has said elsewhere of Congress that it does not “hide elephants in mouseholes.” Whitman v. Am. Trucking Ass’n, 531 U.S. 457, 468 (2001). We may say the same of the Court. The elephant in the room, as it were, is the propensity of this jurisdictional issue, if left variable, to consume time, expense, judicial resources, and legal certainty. This propensity is well known from the history of federal-question jurisprudence, and excising it has been the Court’s aim in Gunn, Grable, and other decisions.[10]Given that history, we therefore disagree that the Supreme Court inserted sub silentio such a nettlesome issue into more cases than before.The four-factor test applied in Gunn was developed to sort cases between state and federal courts, and it is not a tool for the task of sorting cases between the circuits. See Gunn, 568 U.S. at 258 (“[F]ederal jurisdiction over a state law claim will lie if a federal issue is: (1) necessarily raised, (2) actually disputed, (3) substantial, and (4) capable of resolution in federal court without disrupting the federal-state balance approved by Congress.”) (citing Grable, 545 U.S. at 314). As noted, substantiality concerns “the importance of the issue to the federal system as a whole.” Gunn, 568 U.S. at 260. That concern, however, does not exist once the choice is between two federal circuit courts. The formulation could be tweaked to reflect the patent-specific context, but Gunn did not tell us to do so and thereby displace Christianson s time-tested rule. The fourth element is even less suited to the task of sorting cases between the circuits. The choice between circuits for a given appeal is irrelevant to the congressionally-approved balance of state and federal judiciaries. This is not to say that the Gunn-Grable framework could not be adapted to the present task, but the fact that adaptation would be necessary militates against overreading Gunn.Perhaps the strongest point in favor of incorporating Gunn into cases like this one is that Christianson linked § 1295 to § 1338 and § 1331. Gunn construed the latter two statutes together, so, under Christianson, Gunn‘s holdings arguably are automatically incorporated into § 1295. When Christianson was decided, § 1295 referred to § 1338 expressly. By the time of Gunn, § 1295 had been amended to stand on its own; the phrase “any civil action arising under . . . any Act of Congress relating to patents” replaced the reference to § 1338. It is therefore not automatic that a change to § 1338 entails a change to § 1295.To be sure, § 1295 retains the “arising under” formulation in common with the other two statutes, and the Supreme Court prefers to construe like text alike. It has refused to give identical terms the same meaning, however, when contexts and considerations differ. See, e.g., Wachovia Bank v. Schmidt, 546 U.S. 303, 315-17 (2006) (declining to read the term “located” in venue and subject-matter jurisdiction rules in pari materia because the rules serve purposes that are too different). Different considerations, including constitutional and statutory imperatives, attend the sorting of cases between state and federal systems and among the federal circuits. All the federalism concerns associated with the former have no bearing on the latter, as explained. With those set aside, the interests of uniformity and competent application of the law, which failed to carry the day in Gunn, 568 U.S. at 26163, are left as determinative concerns.[11]It would be quite reasonable to have a system that imposes different restrictions at the entrance to the federal system and at the fork in the road leading to different circuits. The exclusionary Gunn-Grable test, screening out most potential cases at the entrance, protects federal district courts from overload and reflects constitutional respect for state courts and state prerogatives. As to those cases that do make it into the federal system, preservation of uniformity comes to the fore, furthered by Christianson‘s inclusionary test for routing appeals to the Federal Circuit. Such a test also promotes judicial economy by simplifying the jurisdictional inquiry and avoiding the jurisdictional ping-pong that Christianson aimed to end.Supposing Gunn did not change the inquiry, the answer to the present question is simple and settled. According to Christianson, the Federal Circuit’s jurisdiction includes “cases in which a well-pleaded complaint establishes . . . that the plaintiff’s right to relief necessarily depends on resolution of a substantial question of federal patent law, in that patent law is a necessary element of one of the well-pleaded claims.” 486 U.S. at 809. Patent law is a necessary element of Walker Process claims. See Ritz Camera & Image, 700 F.3d at 506; Cipro, 544 F.3d at 1330 n.8. Because this case presents a standalone Walker Process claim, there are no non-patent theories in the case that would divert it to our court. Consequently, it belongs in the Federal Circuit.IVWe undertake the preceding analysis with respect for our judicial colleagues and gratitude for the litigants’ patience over the long pendency of this appeal. We nevertheless cannot conclude that the Federal Circuit’s decision to transfer this case to us was plausible, given the Supreme Court’s and Congress’s decisions to the contrary. Accordingly, IT IS ORDERED that this case is TRANSFERRED to the United States Court of Appeals for the Federal Circuit.
