NOTE: This disposition is nonprecedential.
`
`United States Court of Appeals
`for the Federal Circuit
`______________________
`
`NEUROGRAFIX, NEUROGRAPHY INSTITUTE
`MEDICAL ASSOCIATES, INC., IMAGE-BASED
`SURGICENTER CORPORATION, AARON
`GERSHON FILLER,
`Plaintiffs-Appellants
`
`v.
`
`BRAINLAB, INC., BRAINLAB AG, BRAINLAB
`MEDIZINISCHE COMPUTERSYSTEME GMBH,
`Defendants-Appellees
`______________________
`
`2018-2363
`______________________
`
`Appeal from the United States District Court for the
`Northern District of Illinois in No. 1:12-cv-06075, Judge
`Matthew F. Kennelly.
`______________________
`
`Decided: October 7, 2019
`______________________
`
`AARON GERSHON FILLER, Tensor Law, P.C., Santa Mon-
`ica, CA, argued for plaintiffs-appellants.
`
` JAY CAMPBELL, Tucker Ellis LLP, Cleveland, OH, ar-
`gued for defendants-appellees. Also represented by DAVID
`AARON BERNSTEIN.
`
`

`

`2
`
`NEUROGRAFIX v. BRAINLAB, INC.
`
` ______________________
`
`Before NEWMAN, O’MALLEY, and TARANTO, Circuit Judges.
`TARANTO, Circuit Judge.
`U.S. Patent No. 5,560,360, which names Dr. Aaron
`Filler as a co-inventor, describes and claims particular
`methods of generating images of nerves and other bodily
`structures by use of magnetic resonance imaging (MRI)
`technology. Dr. Filler and the three appellants named in
`the caption (collectively, NeuroGrafix) sued the appellees
`named in the caption (collectively, Brainlab), asserting in-
`fringement of the ’360 patent. The case was consolidated
`with cases filed against other defendants and assigned for
`pretrial purposes to a multidistrict litigation (MDL) court.
`The MDL court granted summary judgment of non-in-
`fringement to Brainlab, and it denied reconsideration, as
`did the original district court when the case returned from
`the MDL court. NeuroGrafix appeals. We conclude that
`the grant of summary judgment was procedurally im-
`proper, and we resolve the parties’ key disputes about
`claim construction. We reverse and remand.
`I
`A
`The ’360 patent describes methods and systems for cre-
`ating detailed images of neural tissues by using diffusion
`tensor imaging (DTI), an application of MRI technology.
`’360 patent, Abstract; see also id., col. 21, lines 35–45. DTI
`exploits certain facts about water diffusion in, e.g., brain
`structures. Notably, diffusion along white matter nerve
`tracts is anisotropic: substances such as water diffuse
`freely along the main, long axis of the nerve tract, but dif-
`fusion is very limited in a direction perpendicular to
`(across) that axis. Id., col. 5, lines 5–11. By contrast, the
`surrounding gray matter is relatively isotropic: substances
`
`

`

`NEUROGRAFIX v. BRAINLAB, INC.
`
`3
`
`diffuse at similar rates in all directions. Id., col. 5, lines
`11–12.
`In the patented method, pulsed magnetic field gradi-
`ents are applied in two orthogonal (perpendicular) direc-
`tions in a region containing the nerve tissues for which a
`precise image is sought. Id., col. 5, lines 17–21; see also id.,
`col. 15, lines 40–57. “[I]f the axis of the nerve is generally
`known to the operator,” the specification explains, “the di-
`rection of the desired orthogonal diffusional weighting gra-
`dients can be readily determined.” Id., col. 15, lines 58–62;
`see also id., col. 16, lines 34–47. “On the other hand, if the
`axis of the peripheral nerve is not known, or if many[ ]
`nerves having different axes are being imaged,” the initial
`directions for the magnetic field gradients are “arbitrarily
`selected,” and then a number of alternative directions are
`used. Id., col. 15, lines 63–67; id., col. 16, lines 48–53.
`The result of this process of applying magnetic field
`gradients depends on the types of tissue in the subject re-
`gion. In isotropic tissue, the signal reduction will be the
`same regardless of how the magnetic field gradients are
`oriented relative to the tissue, whereas in anisotropic tis-
`sue, the signal reduction will be greatest when the mag-
`netic field gradients are parallel and perpendicular,
`respectively, to the direction of the anisotropy, i.e., along
`the major, long axis of the neural tract. Id., col. 5, lines 21–
`39. Accordingly, neural tissue can be identified and visu-
`ally differentiated from the surrounding structures by de-
`termining the areas of greater relative anisotropy. Id., col.
`6, lines 46–55; see also id., col. 15, lines 52–57 (“[W]ith gra-
`dients approximately perpendicular and parallel to the
`axis of the peripheral nerve at the particular point being
`imaged, the parallel gradient image can be subtracted from
`the perpendicular gradient image to produce the desired
`‘nerve only’ image.”).
`
`

`

`4
`
`NEUROGRAFIX v. BRAINLAB, INC.
`
`Claim 36 of the ’360 patent is the only independent
`claim at issue in this appeal, and the parties have generally
`treated that claim as representative. That claim recites:
`36. A method of utilizing magnetic resonance to
`determine the shape and position of a structure,
`said method including the steps of:
`(a) exposing a region to a magnetic polarizing
`field including a predetermined arrangement of
`diffusion-weighted gradients, the region including
`a selected structure that exhibits diffusion anisot-
`ropy and other structures that do not exhibit diffu-
`sion anisotropy;
`(b) exposing the region to an electromagnetic
`excitation field;
`(c) for each of said diffusion-weighted gradi-
`ents, sensing a resonant response of the region to
`the excitation field and the polarizing field includ-
`ing the diffusion-weighted gradient and producing
`an output indicative of the resonant response; and
`(d) vector processing said outputs to generate
`data representative of anisotropic diffusion exhib-
`ited by said selected structure in the region, re-
`gardless of the alignment of said diffusion-
`weighted gradients with respect to the orientation
`of said selected structure; and
`(e) processing said data representative of ani-
`sotropic diffusion to generate a data set describing
`the shape and position of said selected structure in
`the region, said data set distinguishing said se-
`lected structure from other structures in the region
`that do not exhibit diffusion anisotropy.
`Id., col. 42, line 43, through col. 43, line 2. The central dis-
`pute in this appeal involves the “selected structure” limita-
`tion in steps (a), (d), and (e).
`
`

`

`NEUROGRAFIX v. BRAINLAB, INC.
`
`5
`
`B
`In August 2012, NeuroGrafix, Neurography Institute
`Medical Associates, Inc., and Image-Based Surgicenter
`Corporation sued Brainlab, Inc., Brainlab AG, and Brain-
`lab Medizinische Computersysteme GmbH in the Northern
`District of Illinois, and in August 2014, Dr. Filler became a
`co-plaintiff by the filing of an amended complaint. The
`plaintiffs (NeuroGrafix) alleged that users of Brainlab’s Fi-
`berTracking software directly infringed the ’360 patent and
`that Brainlab induced the direct infringement by those us-
`ers through statements in its manual and advertisements
`directing users to use the software in an infringing man-
`ner.1 In particular, NeuroGrafix asserted claims 36–37,
`39–42, 44, 46–47, and 49, all of which are method claims.
`Brainlab counterclaimed for a declaratory judgment that
`the asserted claims of the ’360 patent are invalid.
`In April 2013, the Judicial Panel on Multidistrict Liti-
`gation transferred the case to the District of Massachu-
`setts, where it was consolidated, for pretrial proceedings,
`with several cases that NeuroGrafix brought against vari-
`ous MRI equipment manufacturers and university and
`hospital end-users.
`In May 2016, Brainlab filed the first of its two motions
`for summary judgment of non-infringement. Brianlab re-
`lied on customer-protection provisions of settlement agree-
`ments NeuroGrafix had entered into with MRI-equipment
`makers Siemens, GE, and Philips. Brainlab argued that
`
`
`1 There is evidence in the record before us that a user
`of the FiberTracking software selects a region of interest
`from an anatomical image fused with DTI data and chooses
`a minimum diffusion value and a minimum length, and the
`software then displays all fibers that intersect the chosen
`region of interest and exceed the minimum diffusion and
`length parameters.
`
`

`

`6
`
`NEUROGRAFIX v. BRAINLAB, INC.
`
`its FiberTracking software is used to process the output
`from MRI systems made by those manufacturers and that
`FiberTracking users do not infringe under the terms of the
`settlement agreements. In its response, NeuroGrafix ar-
`gued, among other things, that Brainlab could still be lia-
`ble for infringement by “unauthorized independent medical
`practitioners” who use Brainlab’s software but are not cus-
`tomers of Siemens, GE, or Philips.
`The MDL court granted the motion, but only in part, in
`August 2016. It held that summary judgment of non-in-
`fringement was proper with respect to Brainlab’s custom-
`ers using Siemens MRI systems but not as to Brainlab’s
`customers using GE and Philips MRI systems, reasoning
`that only the Siemens settlement agreement, not the GE or
`Philips agreements, extended to Brainlab’s software. The
`court also held summary judgment of non-infringement
`proper as to the alleged independent medical practitioners,
`concluding that NeuroGrafix had produced “no evidence
`that any of the handful of such practitioners identified by
`[NeuroGrafix] used Brainlab products in their alleged in-
`fringement.” J.A. 51.
`Brainlab eventually filed a second motion for summary
`judgment, but before that occurred, NeuroGrafix, in Sep-
`tember 2017, sought leave to file a second amended com-
`plaint that, if allowed, would add allegations that Brainlab
`itself directly infringed the ’360 patent because the steps
`performed by Brainlab’s customers were attributable to
`Brainlab under an agency theory. In conjunction with its
`proposed second amended complaint, NeuroGrafix filed a
`declaration from Dr. Filler and attached several articles
`and other exhibits allegedly demonstrating infringement
`by several of Brainlab’s customers, such as Memorial Sloan
`Kettering Cancer Center and Akron General Hospital. The
`MDL court denied NeuroGrafix permission to file a second
`amended complaint, characterizing the new allegations as
`a “last-ditch attempt to repackage the inducement claim,”
`which it had “long alleged but neglected until the close of
`
`

`

`NEUROGRAFIX v. BRAINLAB, INC.
`
`7
`
`fact discovery,” as a direct-infringement claim under an
`agency theory. J.A. 6986.
`In February 2018, Brainlab filed its second motion for
`summary judgment of non-infringement. Brainlab’s entire
`argument was that users of the software do not commit di-
`rect infringement and therefore Brainlab could not be lia-
`ble for induced infringement; it made no argument against
`inducement liability except for the absence of direct in-
`fringement. J.A. 7309 (“without direct infringement there
`can be no induced infringement”), 7327 (“Absent direct in-
`fringement, there can be no induced infringement.”). On
`direct infringement, Brainlab argued that users of the Fi-
`berTracking software do not satisfy two limitations of
`claim 36—the “selected structure” limitation and the “do
`not exhibit the diffusion anisotropy” limitation. In support
`of that assertion, Brainlab set forth essentially three argu-
`ments in its motion.
`First, and most significantly for present purposes,
`Brainlab argued that “selected structure” requires that a
`user know the “existence and location” of the structure of
`interest before performing the claimed steps of exposing a
`region to a magnetic field, sensing a resonant response, and
`so forth. J.A. 7308. Brainlab asserted that it was impossi-
`ble for users of the FiberTracking software to “select[] [a]
`structure” because “Brainlab’s FiberTracking module does
`not permit a user to isolate or select a specific structure for
`tractography” before scanning; instead, the accused soft-
`ware “automatically generates all tracts that intersect a
`certain volume, like a tumor, if they meet certain criteria,”
`and those tracts “are not visible until after the FiberTrack-
`ing software has been run.” J.A. 7322; see J.A. 7308 (“users
`of Brainlab’s FiberTracking module cannot infringe claim
`36” because they cannot select a structure as required),
`7309 (same), 7312 (same), 7322 (same), 7324 (same), 7325
`(same), 7327 (same). Second, Brainlab contended that “se-
`lected structure” was limited to peripheral nerves, whereas
`the FiberTracking software was used to image only nerves
`
`

`

`8
`
`NEUROGRAFIX v. BRAINLAB, INC.
`
`in the brain, which are not considered peripheral nerves.
`J.A. 7317–18. Third, Brainlab argued that “do not exhibit
`diffusion anisotropy” should be construed as requiring zero
`diffusion anisotropy. J.A. 7321. Under that construction,
`Brainlab asserted, the limitation was not satisfied because
`the gray matter distinguished by the FiberTracking soft-
`ware has a small but nonzero anisotropy, J.A. 7325–27, and
`the FiberTracking software does not permit users to choose
`zero as the anisotropy threshold above which structures
`will be displayed, J.A. 7322.
`In its opposition, NeuroGrafix responded to Brainlab’s
`arguments. It argued that “selected structure” does not re-
`quire that the precise location and orientation of the chosen
`structure be known in advance. J.A. 8011–12. According
`to NeuroGrafix, users could satisfy the claim by, for in-
`stance, obtaining a preliminary MRI image, choosing a
`structure that would be “distinctive and visibly apparent”
`from the preliminary image (such as the pyramidal tract),
`and then performing the steps of the claimed method with
`the chosen structure as the subject. J.A. 8012; see J.A.
`8011–13, 8025–26. NeuroGrafix also asserted that the Fi-
`berTracking software was capable of being used in such a
`manner, pointing to Brainlab’s advertisements, which
`state that users can use the software to image the pyrami-
`dal tract, J.A. 8013, 8015, and the FiberTracking manual,
`which instructs that users can select fiber bundles to in-
`clude or exclude in the region of interest, J.A. 8020.
`The MDL court granted Brainlab’s second summary-
`judgment motion in May 2018. In re NeuroGrafix (’360)
`Patent Litig., MDL No. 13-2432, 2018 WL 2392000, at *5
`(D. Mass. May 25, 2018) (Summary Judgment Op.). It re-
`jected Brainlab’s claim-construction arguments limiting
`“selected structure” to peripheral nerves and limiting “do
`not exhibit diffusion anisotropy” to zero anisotropy. See id.
`at *3. As to Brainlab’s argument that some aspects of the
`“selected structure” must be known in advance, the court
`rejected Brainlab’s position that it was not possible to use
`
`

`

`NEUROGRAFIX v. BRAINLAB, INC.
`
`9
`
`the FiberTracking software in a manner that satisfies the
`claim limitation. See id. “[D]epending on the physician’s
`purpose and objective,” the court held, “FiberTracking is
`capable of both infringing uses and non-infringing uses,”
`though it did not identify precisely what those infringing
`and non-infringing uses would be. Id.
`Nevertheless, the court concluded, summary judgment
`was warranted because NeuroGrafix had pointed to no ev-
`idence that any FiberTracking users actually used the soft-
`ware in an infringing manner, i.e., there was “nothing in
`the record showing that either Brainlab or any of its cus-
`tomers actually uses FiberTracking in the manner hypoth-
`esized by Neuro[G]rafix.” Id. at *4; see also id. at *4 n.5
`(concluding that there was “no evidence in the record” that
`neurosurgeons used FiberTracking to “ascertain the pre-
`cise location of the pyramidal tract” to avoid injuring it dur-
`ing surgery). The court also determined that instances of
`direct infringement could not be inferred from statements
`in Brainlab’s advertisements that it was “possible” to use
`the FiberTracking software to delineate the pyramidal
`tract, noting that those materials “do[] not teach a means
`of selecting a particular ROI and FA Threshold and Mini-
`mum Length values to accomplish this, nor does it recom-
`mend this as a superior or even commensurate mode of
`use.” Id. at *4. In a footnote, the court added a conclusion
`seemingly about the absence of inducement even apart
`from the absence of direct infringement, even though
`Brainlab’s motion had not so argued. It stated that, as a
`matter of law, Brainlab did not induce infringement “for
`the same reason that a reasonable factfinder cannot infer
`instances of direct infringement,” namely, the FiberTrack-
`ing advertisements and manual “[do not] teach an infring-
`ing use of the device such that we are willing to infer from
`those instructions an affirmative intent to infringe the pa-
`tent.” Id. at *4 n.6 (quoting Takeda Pharm. U.S.A., Inc. v.
`W.-Ward Pharm. Corp., 785 F.3d 625, 631 (Fed. Cir. 2015)).
`
`

`

`10
`
`NEUROGRAFIX v. BRAINLAB, INC.
`
`In June 2018, NeuroGrafix moved for reconsideration
`of the MDL court’s grant of summary judgment, primarily
`arguing that several articles attached to NeuroGrafix’s mo-
`tion for leave to file a second amended complaint had pro-
`vided evidence of actual
`infringing uses of
`the
`FiberTracking software. The MDL court denied Neuro-
`Grafix’s motion for reconsideration, noting that Neuro-
`Grafix had not included or relied on the relevant articles in
`its opposition to Brainlab’s summary-judgment motion.
`The case was then remanded to the Northern District
`of Illinois for proceedings on Brainlab’s invalidity counter-
`claim. [A191] In July 2018, NeuroGrafix asked the Illinois
`court to reconsider the MDL court’s summary-judgment or-
`der, contending, as relevant here, that the MDL court had
`granted summary judgment on a basis not asserted in
`Brainlab’s summary-judgment motion. J.A. 8775–76,
`8781–83. The district court denied NeuroGrafix’s motion
`for reconsideration and dismissed Brainlab’s invalidity
`counterclaim without prejudice, producing a final judg-
`ment.
`NeuroGrafix appeals. We have jurisdiction under 28
`U.S.C. § 1295(a)(1).
`
`II
`A
`We begin by addressing NeuroGrafix’s procedural chal-
`lenge to the MDL court’s grant of summary judgment.
`NeuroGrafix argues that it was improper for the MDL
`court to fault it for failing to produce evidence of actual in-
`fringement because Brainlab argued only that, under its
`construction of “selected structure,” the accused software
`was not capable of infringement, not that, under the con-
`struction adopted by the MDL court, there was no evidence
`of actual infringement. We review the MDL court’s grant
`of summary judgment de novo. Momenta Pharm., Inc. v.
`Teva Pharm. USA Inc., 809 F.3d 610, 614 (Fed. Cir. 2015)
`
`

`

`NEUROGRAFIX v. BRAINLAB, INC.
`
`11
`
`(following First Circuit law); see also In re Cygnus Tele-
`comms. Tech., LLC, Patent Litig., 536 F.3d 1343, 1352 (Fed.
`Cir. 2008) (following law of MDL court’s regional circuit in
`deciding issues involving summary-judgment procedures).
`We agree with NeuroGrafix and accordingly reverse the
`grant of summary judgment.
`As Brainlab’s motion for summary judgment repeat-
`edly made clear, its non-infringement position depended on
`the premise that “select[ing] [a] structure” requires know-
`ing in advance the location of the chosen structure. Under
`that construction, Brainlab argued, the FiberTracking soft-
`ware is not capable of infringement, since the software is
`used to detect structures whose location is not already
`known. See, e.g., J.A. 7311 (“Claim 36 is focused on deter-
`mining the location and shape of an anisotropic structure
`that is already known and ‘selected’ for imaging in advance
`of scanning . . . . Conversely, Brainlab’s Fiber[T]racking
`module is focused on finding patient specific anisotropic
`structures that are not previously known.”); J.A. 7322
`(“The user certainly cannot select a structure in advance of
`scanning. The reason is simple: Brainlab’s FiberTracking
`module is used to find white matter tracts that are not vis-
`ible until after the FiberTracking software has been run
`. . . .”).
`Moreover, the expert reports cited in Brainlab’s sum-
`mary-judgment motion were also premised on this under-
`standing of “selected structure.” Dr. James Leach declared
`that “the neuroradiologist cannot select certain white mat-
`ter structures or tracts in advance for imaging” because
`“the position or orientation of white matter tracts is not
`known in advance of imaging” in cranial DTI. J.A. 7921.
`Dr. Andrew Tsung stated that “I do not select certain white
`matter structures for imaging by the MRI,” as “[t]he loca-
`tion of white matter tracts are not identifiable prior to im-
`aging.” J.A. 7912. And Dr. Michael Moseley asserted that
`“a ‘selected’ structure is one where the axis of the structure,
`such as a nerve, . . . would be known in advance of the
`
`

`

`12
`
`NEUROGRAFIX v. BRAINLAB, INC.
`
`imaging,” J.A. 7946, and using that understanding, he
`added that “there is no ‘selected structure’ when DTI imag-
`ing is performed” using Brainlab’s FiberTracking software
`because “the axes of the white matter fiber tracts are not
`known in advance,” J.A. 7947. Neither Brainlab nor its ex-
`perts argued in the alternative that, even if “selected struc-
`ture” did not include a requirement of knowing the
`position, orientation, location, or axes of a structure in ad-
`vance, the record was devoid of evidence that Brainlab’s
`customers used the FiberTracking software to image par-
`ticular chosen structures.
`In its summary-judgment opposition, NeuroGrafix dis-
`puted this claim construction, essentially arguing that “se-
`lected structure” simply requires choosing a particular
`structure as a subject for the claimed process. That is pos-
`sible in the FiberTracking software, NeuroGrafix asserted,
`because at least the pyramidal tract is visible after taking
`a preliminary image and can then be chosen for imaging
`according to the claimed method. See J.A. 8012 (“[E]ither
`visually after opening the skull or from preliminary routine
`MRI scout images, the technologist can select[] a brain
`structure called the pyramidal tract.”); J.A. 8014–15 (“With
`tractography and DTI, it is possible to select this structure
`of the brain . . . and then to provide this selected structure
`as an ROI for the FiberTracking software.”). And Neuro-
`Grafix pointed to Brainlab’s advertisements as evidence
`that such a use was possible and even encouraged by
`Brainlab. See J.A. 8015 (showing Brainlab advertisement
`that says: “It is possible to delineate major white matter
`tracts, such as the pyramidal tract, by applying fiber track-
`ing algorithms.”); see also J.A. 8013 (showing Brainlab ad-
`vertisement that says: “Waves of DTI data on exotic
`eloquent white matter specimens, like pyramidal tracts,
`now flow easily to your BrainLAB IGS.”). In other words,
`NeuroGrafix argued, and the MDL court eventually
`agreed, that the FiberTracking software is capable of in-
`fringing uses as well as non-infringing uses.
`
`

`

`NEUROGRAFIX v. BRAINLAB, INC.
`
`13
`
`That showing was sufficient for NeuroGrafix to defeat
`summary judgment, and the MDL court erred in conclud-
`ing otherwise. NeuroGrafix demonstrated that there was
`a genuine dispute of material fact on the only issue raised
`by Brainlab, namely, whether the FiberTracking software
`was capable of infringing uses. Evidence of actual infring-
`ing uses of the FiberTracking software was unnecessary to
`answer the only grounds for summary judgment asserted
`by Brainlab.2
`A court cannot grant summary judgment on a ground
`that was neither asserted by the movant nor made the sub-
`ject of judicial action under Rule 56(f) that gave the non-
`movant proper notice of the ground and of the obligation
`“to come forward with all of her evidence.” Celotex Corp. v.
`Catrett, 477 U.S. 317, 326 (1986); see Glaverbel Societe
`Anonyme v. Northlake Mktg. & Supply, Inc., 45 F.3d 1550,
`1562 (Fed. Cir. 1995) (following Seventh Circuit law); see
`also, e.g., Lusson v. Carter, 704 F.2d 646, 647 (1st Cir.
`1983). And in the specific context of patent infringement,
`we have held that summary judgment of non-infringement
`requires the accused infringer to “point[] to the specific
`ways in which accused systems did not meet the claim lim-
`itations.” Exigent Technology, Inc. v. Atrana Solutions,
`Inc., 442 F.3d 1301, 1309 (Fed. Cir. 2006). The MDL court’s
`ruling was contrary to those basic principles in that it
`granted summary judgment against NeuroGrafix for its
`failure to come forward with evidence to answer a non-
`
`
`2 Thus, we need not and do not decide whether, even
`if NeuroGrafix did not produce direct evidence of actual in-
`fringement, instances of infringement can be inferred from
`the statements and figures in Brainlab’s advertisements
`and manual. See Summary Judgment Op. at *4 (citing
`Toshiba Corp. v. Imation Corp., 681 F.3d 1358, 1364 (Fed.
`Cir. 2012); Fujitsu Ltd. v. Netgear Inc., 620 F.3d 1321, 1329
`(Fed. Cir. 2010)).
`
`

`

`14
`
`NEUROGRAFIX v. BRAINLAB, INC.
`
`infringement ground that had not been asserted and of
`which it had not been given proper notice.
`To be sure, our law is clear that, in this case, Neuro-
`Grafix could not sustain a claim of direct infringement of
`the method claims by merely showing that the accused soft-
`ware is “capable of” operating in an infringing manner.
`See, e.g., Fujitsu Ltd. v. Netgear Inc., 620 F.3d 1321, 1329
`(Fed. Cir. 2010). We assume, without questioning, that in
`this case NeuroGrafix must ultimately make a showing
`that the accused software was actually used in an infring-
`ing manner by Brainlab (for direct infringement case) or by
`one or more of Brainlab’s customers (for indirect infringe-
`ment). Moreover, it is understandable that the district
`court might be surprised that NeuroGrafix made no such
`showing after the years of litigation and discovery this
`MDL spanned. Nevertheless, the motion being considered
`by the district court in this case was one structured and
`limited by the movant. The court was not free to look down
`the road and consider what the non-movant might need to
`establish to survive a differently structured, well-sup-
`ported motion. The motion before it necessarily limited the
`court’s inquiry.
`For the same reason, the MDL court’s apparent holding
`that Brainlab’s advertisements and manual do not induce
`infringement as a matter of law also was procedurally im-
`proper. See Summary Judgment Op. at *4 n.6. Brainlab’s
`summary-judgment motion argued only that “Brainlab
`cannot induce infringement of the asserted claims of the
`’360 patent” because “[a]bsent direct infringement, there
`can be no induced infringement.” J.A. 7327. It did not ar-
`gue, as the MDL court seemed to conclude, that the rele-
`vant Brainlab materials merely suggested that an
`infringing use was possible rather than instructing how to
`use the software in an infringing manner. To the extent
`that this conclusion was an independent basis for the MDL
`
`

`

`NEUROGRAFIX v. BRAINLAB, INC.
`
`15
`
`court’s grant of summary judgment, we reverse the court’s
`decision on that ground as well.3
`B
`The MDL court’s procedural error is an adequate
`ground for reversal and does not depend on whether its
`claim construction of “selected structure” was correct. But
`we address the disputes about the proper construction of
`that term so that the district court can apply the correct
`construction on remand. We review the MDL court’s claim
`construction de novo and any underlying factual findings
`based on extrinsic evidence for clear error. Teva Pharm.
`USA, Inc. v. Sandoz, Inc., 135 S. Ct. 831, 841 (2015).
`We conclude that to “select[] [a] structure” is simply to
`choose it as a subject for placement into the claimed process
`that starts with exposing a region to a magnetic field, pro-
`ceeds to sensing a resonant response, and continues as
`claimed. That meaning follows from the language of claim
`36 itself: in step (a), the region exposed to a magnetic po-
`larizing field includes the “selected structure,” and in step
`(e), the resulting data set distinguishes the “selected struc-
`ture” from other structures in the region. ’360 patent,
`col. 42, lines 46–50; id., col. 42, line 64, through col. 43,
`line 2. The specification does not use the language of “se-
`lected structure,” but it uses “select” simply to describe
`choosing something before taking some action. See, e.g.,
`id., col. 14, lines 53–62 (discussing “select[ing]” a region of
`interest before determining the average intensity within
`that region of interest); id., col. 28, lines 23–26 (discussing
`“select[ing] a volume of interest” before rendering that vol-
`ume of interest into a projection neurogram).
`
`
`3 The MDL court’s rejection of NeuroGrafix’s induce-
`ment claim may also have been based on a construction of
`“selected structure” that, as we discuss below, was incor-
`rect.
`
`

`

`16
`
`NEUROGRAFIX v. BRAINLAB, INC.
`
`The MDL court did not set forth a precise claim con-
`struction of “selected structure” in its summary-judgment
`opinion.4 In one key respect, though, the court’s under-
`standing of the phrase fits the simple construction that we
`think is mandated. The court correctly rejected the con-
`struction that seemingly underlies Brainlab’s contention
`that infringing use of the FiberTracking software is impos-
`sible, namely, that a “selected structure” is one whose loca-
`tion, orientation, axis, or the like is known in advance of
`the claimed mapping process to the same degree it will be-
`come known upon completion of that process. And the
`court indicated that “delineat[ing] the pyramidal tract,”
`Summary Judgment Op. at *4, and “ascertain[ing] the pre-
`cise location of the pyramidal tract,” id. at *4 n.5, would
`satisfy the “selected structure” limitation. Those observa-
`tions fit the specification’s express contemplation of per-
`forming the patented method even when, for example, “the
`axis of the peripheral nerve is not known.” Id., col. 15, lines
`63–64.
`Two further points about claim construction conten-
`tions advanced by the parties—one by Brianlab, one by
`NeuroGrafix—are warranted. Brainlab has suggested that
`software that tracks all fibers in an area cannot perform
`the method, because the tracking is not limited to a partic-
`ular selected structure. That view is not supported by
`claim 36’s language. As long as a chosen structure is
`among those put into the process for distinguishing the
`data or images in the way the claim specifies, the claim is
`satisfied, even if the process used to do that results in com-
`parable data and images for other structures as well. Both
`
`
`4 The MDL court did not construe “selected struc-
`ture” in its August 2016 claim-construction order; nor did
`the parties agree to a construction of the phrase. See In re
`NeuroGrafix (’360) Patent Litig., 201 F. Supp. 3d 206, 212
`& n.4 (D. Mass. 2016).
`
`

`

`NEUROGRAFIX v. BRAINLAB, INC.
`
`17
`
`claim 36’s preamble and the claim phrase “region including
`a selected structure” use the word “including.” ’360 patent,
`col. 42, lines 45, 48. We have “consistently interpreted ‘in-
`cluding’ and ‘comprising’ to have the same meaning,
`namely, that the listed elements . . . are essential but other
`elements may be added.” Lucent Techs., Inc. v. Gateway,
`Inc., 525 F.3d 1200, 1214 (Fed. Cir. 2008). And nothing in
`the language following either of the “including” terms im-
`plies that no other structure may be mapped in the claimed
`way when a particular chosen structure is placed into the
`claimed mapping process.
`For its part, NeuroGrafix argues on appeal that “se-
`lected structure” should be construed as equivalent to “re-
`gion” and that all uses of the FiberTracking software are
`therefore infringing because Brainlab’s customers neces-
`sarily choose a region to be the subject of the claimed
`method before performing the steps of the method. That
`always-infringes contention is the polar opposite of Brain-
`lab’s never-infringes contention, and it is equally wrong.
`The argument was likely forfeited by not being adequately
`presented; indeed, in its motion for reconsideration before
`the MDL court, NeuroGrafix specifically agreed with the
`MDL court’s conclusion that “FiberTracking is capable of
`both infringing uses and non-infringing uses.” J.A. 8449
`(quoting Summary Judgment Op. at *3). In any event,
`NeuroGrafix’s construction contradicts the claim language.
`Claim 36 refers to “selected structure” and “region” as