LKQ CORPORATION v. GM GLOBAL TECHNOLOGY OPERATIONS

United States Court of Appeals

for the Federal Circuit

______________________

LKQ CORPORATION, KEYSTONE AUTOMOTIVE

INDUSTRIES, INC.,

Appellants

GM GLOBAL TECHNOLOGY OPERATIONS LLC,

Appellee

______________________

2021-2348

______________________

Appeal from the United States Patent and Trademark

Office, Patent Trial and Appeal Board in No. IPR2020-

00534.

______________________

Decided: May 21, 2024

______________________

MARK A. LEMLEY, Lex Lumina PLLC, New York, NY,

argued for appellants. Also represented by MARK P.

MCKENNA; ANDREW HIMEBAUGH, BARRY IRWIN, ARIEL H.

KATZ, IFTEKHAR ZAIM, Irwin IP LLP, Chicago, IL.

JOSEPH HERRIGES, JR., Fish & Richardson P.C., Minne-

apolis, MN, argued for appellee. Also represented by JOHN

A. DRAGSETH, SARAH JACK; NITIKA GUPTA FIORELLA, Wil-

mington, DE.

FARHEENA YASMEEN RASHEED, Office of the Solicitor,

Case: 21-2348 Document: 227 Page: 1 Filed: 05/21/2024

LKQ CORPORATION v.

GM GLOBAL TECHNOLOGY OPERATIONS LLC

United States Patent and Trademark Office, Alexandria,

VA, argued for amicus curiae United States. Also repre-

sented by THOMAS W. KRAUSE, WILLIAM LAMARCA, BRIAN

RACILLA; BRIAN JAMES SPRINGER, Civil Division, United

States Department of Justice, Washington, DC.

JAMES RICHARD FERGUSON, Mayer Brown, LLP, Chi-

cago, IL, for amici curiae Alliance for Automotive Innova-

tion, Rivian Automotive Inc. Also represented by RYAN T.

BABCOCK, BRIAN W. NOLAN, New York, NY.

CHRISTOPHER V. CARANI, McAndrews, Held & Malloy,

Ltd., Chicago, IL, for amicus curiae American Intellectual

Property Law Association. Also represented by DUNSTAN

BARNES; SALVATORE ANASTASI, Barley Snyder LLC, Mal-

vern, PA.

JOHN LOUIS CORDANI, Robinson & Cole LLP, Hartford,

CT, for amici curiae American Property Casualty Insur-

ance Association, Certified Automotive Parts Association,

National Association of Mutual Insurance Companies.

Also represented by BENJAMIN M. DANIELS; KYLE GLENDON

HEPNER, Washington, DC.

MATTHEW JAMES DOWD, Dowd Scheffel PLLC, Wash-

ington, DC, for amici curiae Anthony Andre, Roger Belveal,

Peter Bressler, Gordon Bruce, Pieter den Heten, Michael

Derocher, Sridhar Dhulipala, Edward Dorsa, Valerie Fen-

ster, Timothy Fletcher, Esteban Fridman, Michael Garten,

Deborah George, Joellyn Gray, Jon Christopher Hacker,

David Hodge, Institute for Design Science and Public Pol-

icy, James Kaufman, Haig Khachatoorian, Carlton Lay,

Gustavo Lopes Jota, Theo Mandel, Charles Lee Mauro,

George McCain, Christopher Daniel Morley, Adrian

O'Heney, Nancy Perkins, Scott Peterson, Lloyd Philpott,

Jennifer Pierce, RitaSue Siegel, John Stram, Joyce

Thomas, Steven Visser, Angela Yeh, Sal Ziauddin. Also

represented by ROBERT JAMES SCHEFFEL; PERRY SAIDMAN,

Case: 21-2348 Document: 227 Page: 2 Filed: 05/21/2024

LKQ CORPORATION v.

GM GLOBAL TECHNOLOGY OPERATIONS LLC

Perry Saidman, LLC, Miami Beach, FL.

MARK D. SELWYN, Wilmer Cutler Pickering Hale and

Dorr LLP, Palo Alto, CA, for amicus curiae Apple Inc. Also

represented by SHARCHUN DENNIS WANG; BENJAMIN S.

FERNANDEZ, Denver, CO; MARK CHRISTOPHER FLEMING,

Boston, MA; LAURA E. POWELL, Washington, DC; TRACY-

GENE GRAVELINE DURKIN, Sterne Kessler Goldstein & Fox

PLLC, Washington, DC.

SETH DAVID GREENSTEIN, Constantine Cannon LLP,

Washington, DC, for amicus curiae Auto Care Association.

ROBERT GLENN OAKE, JR., Oake Law Office, Allen, TX,

for amicus curiae Automotive Body Parts Association.

PHILLIP R. MALONE, Juelsgaard Intellectual Property

and Innovation Clinic, Mills Legal Clinic, Stanford Law

School, Stanford, CA, for amici curiae Mark Bartholomew,

Digital Right to Repair Coalition, Shubha Ghosh, iFixit,

Aaron Perzanowski, Ana Santos Rutschman, Joshua D.

Sarnoff, Securepairs, Katherine J. Strandburg, United

States Public Interest Research Group, Inc.

ALEXANDER CHEN, InHouse Co Law Firm, Irvine, CA,

for amicus curiae Eagle Eyes Traffic Industrial Co., Ltd.

FRANK A. ANGILERI, Brooks Kushman PC, Royal Oak,

MI, for amicus curiae Ford Motor Company. Also repre-

sented by MARC LORELLI.

WILLIAM ADAMS, Quinn Emanuel Urquhart & Sullivan,

LLP, New York, NY, for amici curiae Hyundai Motor Com-

pany, Kia Corp. Also represented by DONGKWAN JAMES

PAK, Los Angeles, CA.

ROBERT S. KATZ, Banner & Witcoff, Ltd., for amicus cu-

riae Industrial Designers Society of America, Inc. Also

Case: 21-2348 Document: 227 Page: 3 Filed: 05/21/2024

LKQ CORPORATION v.

GM GLOBAL TECHNOLOGY OPERATIONS LLC

represented by SONIA MARY OKOLIE; ERIK S. MAURER, Chi-

cago, IL.

PAUL BERGHOFF, McDonnell Boehnen Hulbert &

Berghoff LLP, Chicago, IL, for amicus curiae Intellectual

Property Owners Association.

BRUCE EWING, Dorsey & Whitney LLP, New York, NY,

for amicus curiae International Trademark Association.

Also represented by MARTIN SCHWIMMER, Leason Ellis,

LLP, White Plains, NY.

CHARLES R. MACEDO, Amster Rothstein & Ebenstein

LLP, New York, NY, for amicus curiae New York Intellec-

tual Property Law Association. Also represented by DAVID

GOLDBERG; PAUL A. COLETTI, Johnson & Johnson, New

Brunswick, NJ; ROBERT J. RANDO, Greenspoon Marder

LLP, New York, NY; KSENIA TAKHISTOVA, E. Brunswick,

NJ.

ROBERT J. FIGA, Dean & Fulkerson, PC, Troy, MI, for

amicus curiae Taiwan Auto Body Parts Assoc.

______________________

Before MOORE, Chief Judge, LOURIE, DYK, PROST, REYNA,

TARANTO, CHEN, HUGHES, STOLL, and STARK,

Circuit Judges.

Opinion for the court filed by Circuit Judge STOLL, in

which Chief Judge MOORE and Circuit Judges DYK,

PROST, REYNA, TARANTO, CHEN, HUGHES, and STARK join.

Circuit Judge Newman and Circuit Judge Cun-

ningham did not participate. Circuit Judge Clevenger, who

was a member of the original panel, also did not partici-

pate.

Case: 21-2348 Document: 227 Page: 4 Filed: 05/21/2024

LKQ CORPORATION v.

GM GLOBAL TECHNOLOGY OPERATIONS LLC

Opinion concurring in the judgment filed by

Circuit Judge LOURIE.

STOLL, Circuit Judge.

This case involves the standards for assessing nonobvi-

ousness of design patents under 35 U.S.C. § 103.

The principal question that this case presents is

whether Supreme Court precedent, including KSR Interna-

tional Co. v. Teleflex Inc., 550 U.S. 398 (2007), should cause

us to rethink the long-standing Rosen-Durling test used to

assess nonobviousness of design patents. We answer in the

affirmative and overrule the Rosen-Durling test require-

ments that the primary reference must be “basically the

same” as the challenged design claim and that any second-

ary references must be “so related” to the primary refer-

ence that features in one would suggest application of those

features to the other. We adopt an approach consistent

with Congress’s statutory scheme for design patents, which

provides that the same conditions for patentability that ap-

ply to utility patents apply to design patents, as well as Su-

preme Court precedent which suggests a more flexible

approach than the Rosen-Durling test for determining non-

obviousness.

BACKGROUND

GM Global Technology LLC (“GM”) owns U.S. Design

Patent No. D797,625, which claims a design for a vehicle’s

front fender. This design is used in GM’s 2018–2020 Chev-

rolet Equinox.

Appellants LKQ Corporation and Keystone Automo-

tive Industries, Inc. (collectively “LKQ”) filed a petition to

institute an inter partes review of GM’s D’625 patent, as-

serting that the challenged claim is unpatentable under

35 U.S.C. § 102 based on U.S. Design Patent No. D773,340

(“Lian”) or under 35 U.S.C. § 103 based on Lian alone or

Lian as modified by a promotional brochure depicting the

design of the front fender on the 2010 Hyundai Tucson

Case: 21-2348 Document: 227 Page: 5 Filed: 05/21/2024

LKQ CORPORATION v.

GM GLOBAL TECHNOLOGY OPERATIONS LLC

(“Tucson”). LKQ presented the following comparisons be-

tween the D’625 patent, Lian, and Tucson:

Appellants’ Br. at 5–6. Figures 1–4 of the D’625 patent il-

lustrate the claimed front fender from a perspective view,

a side view, a front view, and a top view, respectively.

The Board held that LKQ had not established by a pre-

ponderance of evidence that Lian anticipates the claim of

the D’625 patent. LKQ Corp. v. GM Glob. Tech. Operations

Case: 21-2348 Document: 227 Page: 6 Filed: 05/21/2024

LKQ CORPORATION v.

GM GLOBAL TECHNOLOGY OPERATIONS LLC

LLC, IPR2020-00534, 2021 WL 3411458, at *16 (P.T.A.B.

Aug. 4, 2021) (“Board Decision”). Specifically, the Board

applied the ordinary observer test for anticipation of a de-

sign patent set forth in Gorham Manufacturing Co.

v. White, 81 U.S. 511 (1871), which provides that “if, in the

eye of an ordinary observer, giving such attention as a pur-

chaser usually gives, two designs are substantially the

same, if the resemblance is such as to deceive such an ob-

server, inducing him to purchase one supposing it to be the

other, the first one patented is infringed by the other.”

Board Decision, 2021 WL 3411458, at *7 (quoting Gorham,

81 U.S. at 528) (alteration omitted). The Board found that

while “there are certain articulable and visible similarities

in the overall appearance of the claimed design and Lian

that would be apparent to an ordinary observer,” differ-

ences between the designs, including (1) the wheel arch

shape and the terminus; (2) door cut line; (3) protrusion;

(4) sculpting; (5) inflection line; (6) the first and second

creases; and (7) the concavity line, affect the overall visual

impression of each design such that they are not substan-

tially the same. Id. at *11–16.

The Board then applied the long-standing Rosen-Durl-

ing test to assess the nonobviousness of the claimed design.

Under this two-part test, first, In re Rosen requires that

“[b]efore one can begin to combine prior art designs . . . one

must find a single reference, ‘a something in existence, the

design characteristics of which are basically the same as

the claimed design.’” Durling v. Spectrum Furniture Co.,

Inc., 101 F.3d 100, 103 (Fed. Cir. 1996) (quoting In re

Rosen, 673 F.2d 388, 391 (CCPA 1982)). This primary ref-

erence is commonly referred to as the Rosen reference. As

the Board explained, Rosen’s “basically the same” test re-

quires consideration of the visual impression created by the

patented design as a whole. Rosen, 673 F.2d at 391. If no

Rosen reference is found, the obviousness inquiry ends

without consideration of step two. Durling, 101 F.3d

Case: 21-2348 Document: 227 Page: 7 Filed: 05/21/2024

LKQ CORPORATION v.

GM GLOBAL TECHNOLOGY OPERATIONS LLC

at 105 (“Without such a primary reference, it is improper

to invalidate a design patent on grounds of obviousness.”).

At step two, Durling requires that “[o]nce this primary

reference is found, other references may be used to modify

it to create a design that has the same overall visual ap-

pearance as the claimed design.” Id. at 103. But another

threshold must be met. Specifically, any secondary refer-

ences must be “so related [to the primary reference] that

the appearance of certain ornamental features in one

would suggest the application of those features to the

other.” Id. (alteration in original) (quoting In re Borden,

90 F.3d 1570, 1575 (Fed. Cir. 1996)).

In this case, the Board determined that LKQ failed to

establish that the challenged claim would have been obvi-

ous because LKQ failed to identify a Rosen reference. LKQ

had proffered Lian as its Rosen reference. Identifying the

same differences it emphasized with regard to anticipation,

the Board found that Lian differs from the claimed design

in “the upper protrusion, the u-shaped notch, the door cut

line, a circular wheel arch, the lower rear terminus, . . .

first and second creases along with the concavity line, and

the inflection line.” Board Decision, 2021 WL 3411458,

at *20. Considering these differences together, the Board

found that Lian does not “create[] ‘basically the same’ vis-

ual impression” as the patented design. Id. (quoting Durl-

ing, 101 F.3d at 103). Accordingly, the Board ended its

obviousness analysis without further consideration.

LKQ appealed, and a panel of this court affirmed the

Board decision. LKQ Corp. v. GM Global Tech. Operations

LLC, No. 2023-2348, 2023 WL 328228 (Fed. Cir. Jan. 20,

2023) (“Panel Decision”). The panel concluded that sub-

stantial evidence supports the Board’s finding that Lian

and the claimed design “created different overall impres-

sions” and, thus, Lian did not anticipate the D’625 patent.

Id. at *4. The panel then addressed LKQ’s argument that

the Supreme Court’s decision in KSR, implicitly overruled

Case: 21-2348 Document: 227 Page: 8 Filed: 05/21/2024

LKQ CORPORATION v.

GM GLOBAL TECHNOLOGY OPERATIONS LLC

the Rosen and Durling tests. The panel rejected GM’s as-

sertion that LKQ forfeited this argument by not raising it

before the Board, determining that LKQ’s petition to the

Board was “sufficient to preserve the argument” and that

“this challenge to Rosen and Durling is a pure question of

law that undoubtedly presents a significant question of

general impact, favoring the exercise of our discretion to

hear and decide this issue.” Id. at *4–6. On the merits, the

panel explained that it could not “overrule Rosen or Durl-

ing without a clear directive from the Supreme Court” and

“it is not clear” that the Supreme Court had done so. Id.

at *6. In applying Rosen and Durling, the majority af-

firmed the Board’s nonobviousness determination, con-

cluding that substantial evidence supports the Board’s

findings that “LKQ failed to identify ‘the correct visual im-

pression created by the patented design as a whole,’” and

that “LKQ failed to show that Lian creates ‘basically the

same’ visual impression as the claimed design.” Id. at *6–

7 (first quoting Durling, 101 F.3d at 103; and then quoting

Rosen, 673 F.2d at 390–91).

Judge Stark wrote a concurrence-in-part, disagreeing

with the majority on the issue of forfeiture. He agreed,

however, that it is not clear whether KSR overruled Rosen

or Durling, but noted that there exists “at minimum, sub-

stantial tension between the Supreme Court’s holding in

KSR and our . . . test.” Id. at *13 (Stark, J., concurring-in-

part). The concurrence-in-part elaborated that the Rosen

reference requirement may be “precisely the type of limit-

ing, rigid rule KSR faulted,” as the obviousness analysis

stops without a Rosen reference, and, similarly, the “so re-

lated” requirement of Durling presents “[a]dditional rigid-

ity that might not be appropriate in a post-KSR world.” Id.

Judge Lourie presented additional views, disagreeing

with LKQ’s argument that KSR overruled Rosen because

“KSR did not involve design patents,” which are distinct

from utility patents. Id. at *8 (Lourie, J., additional views).

Judge Lourie further noted that “[w]hile 35 U.S.C.

Case: 21-2348 Document: 227 Page: 9 Filed: 05/21/2024

LKQ CORPORATION v.

GM GLOBAL TECHNOLOGY OPERATIONS LLC

§ 103 . . . does not differentiate . . . [and] applies to all types

of patents, the considerations involved in determining ob-

viousness are different” for utility patents versus design

patents. Id. Specifically, for design patents, the consider-

ations include “the overall appearance, visual impressions,

artistry, and style of ornamental subject matter” while for

utility patents, we consider “unexpected properties, utility,

and function.” Id.

This court granted rehearing en banc and vacated the

panel opinion. See LKQ Corp. v. GM Glob. Tech. Opera-

tions LLC, 71 F.4th 1383, 1384 (Fed. Cir. 2023). Among

other questions, our order requested briefing

on: (1) whether KSR overrules or abrogates the Rosen-

Durling test; (2) if not, whether the court should nonethe-

less eliminate or modify the Rosen-Durling test; and (3) if

we answer either of these two questions affirmatively,

what test should apply for evaluating design patent obvi-

ousness challenges. Id. at 1384–85. We reinstate the por-

tion of the panel opinion holding that substantial evidence

supports the Board’s finding of no anticipation. The panel

decision remains vacated in all other respects. We vacate

the final written decision of the Board as to the nonobvi-

ousness determination and remand for further proceedings

consistent with this opinion.

DISCUSSION

A design patent protects a “new, original and ornamen-

tal design for an article of manufacture.” 35 U.S.C.

§ 171(a). Under the Patent Act, the statutory provisions

“relating to patents for inventions,” or utility patents,

“shall apply to patents for designs, except as otherwise pro-

vided.” Id. § 171(b). Our discussion of nonobviousness of a

design or utility patent claim should start with the statu-

tory language of § 103, which states in pertinent part:

Case: 21-2348 Document: 227 Page: 10 Filed: 05/21/2024

LKQ CORPORATION v.

GM GLOBAL TECHNOLOGY OPERATIONS LLC

A patent for a claimed invention may not be ob-

tained, notwithstanding that the claimed invention

is not identically disclosed as set forth in section

102, if the differences between the claimed inven-

tion and the prior art are such that the claimed in-

vention as a whole would have been obvious before

the effective filing date of the claimed invention to

a person having ordinary skill in the art to which

the claimed invention pertains.

Id. § 103. The Supreme Court addressed the meaning of

the statute in Graham v. John Deere Co. of Kansas City,

which involved utility patents. 383 U.S. 1 (1966). The

Court in Graham explained that the ultimate question of

obviousness is one of law based on “several basic factual

inquiries.” Id. at 17. The Court elaborated that under

§ 103, these factual inquiries include “the scope and con-

tent of the prior art”; “differences between the prior art and

the claims at issue”; and “the level of ordinary skill in the

pertinent art.” Id. In addition to these factors, “[s]uch sec-

ondary considerations as commercial success, long felt but

unsolved needs, failure of others, etc., might be utilized to

give light to the circumstances surrounding the origin of

the subject matter sought to be patented” and may be rele-

vant as “indicia of obviousness or nonobviousness.” Id.

at 17–18. The ultimate question is whether differences be-

tween the claimed invention and the prior art are such that

the claimed invention as a whole would have been obvious

to a person of ordinary skill in the pertinent art at the time

of the invention.

More recently, the Supreme Court again addressed the

standards for assessing nonobviousness in KSR. 550 U.S.

398. There, the district court had held Teleflex’s utility pa-

tent claim obvious based on its findings that (1) a first prior

art reference taught every limitation of the claim except

one; (2) this missing aspect was taught in two other prior

art references; and (3) a person of ordinary skill in the art

would have been motivated to modify the first prior art

Case: 21-2348 Document: 227 Page: 11 Filed: 05/21/2024

LKQ CORPORATION v.

GM GLOBAL TECHNOLOGY OPERATIONS LLC

reference to include the missing aspect taught by the sec-

ondary prior art references. Id. at 413. The Supreme

Court in KSR specifically addressed the “teaching, sugges-

tion, or motivation” or “TSM” test recited in our precedent,

under which a patent claim was only proved obvious if

“some motivation or suggestion to combine the prior art

teachings” existed in “the prior art, the nature of the prob-

lem, or the knowledge of a person having ordinary skill in

the art.” Id. at 407. We had previously applied the TSM

test to prevent factfinders from falling victim to hindsight

bias in determining whether there was a reason to combine

prior art references to arrive at the claimed invention. Id.

at 421.

The Supreme Court rejected our rigid application of the

TSM test in the case before it. The Court explained that

Graham “set forth an expansive and flexible approach in-

consistent with the way the Court of Appeals applied its

TSM test here.” Id. at 415. The Court reasoned that while

factfinders should guard against hindsight, “[r]igid preven-

tative rules that deny factfinders recourse to common

sense” are improper under § 103. Id. at 421. The Court

further explained that a person of ordinary skill in the art

could find a motivation to combine prior art references in

ways other than meeting the TSM test as rigidly applied.

The Court noted, for example, that design incentives and

market forces could be a reason to apply an invention from

one field into another, and that a patent claim might be

proved obvious with a showing that it would have been “ob-

vious to try” a “finite number of identified, predictable so-

lutions” to a problem where there is “a design need or

market pressure to solve.” Id. at 417, 421. The Court ex-

plained that while the TSM test might provide helpful in-

sight on the obviousness of an invention, such insights

“need not become rigid and mandatory formulas” as the

“obviousness analysis cannot be confined by a formalistic

conception.” Id. at 419.

Case: 21-2348 Document: 227 Page: 12 Filed: 05/21/2024

LKQ CORPORATION v.

GM GLOBAL TECHNOLOGY OPERATIONS LLC

The Supreme Court has also addressed the validity of

design patent claims in view of combined prior art designs.

Most notably, in Smith v. Whitman Saddle Co., the Su-

preme Court considered whether a design patent claim di-

rected to an ornamental design for a saddle was patentable

in view of the combination of two prior art saddles, known

as the Granger tree and the Jenifer or Jenifer-McClellan

tree. 148 U.S. 674, 680 (1893).

Comparing the claimed

saddle design to the prior art, the Court explained that:

The saddle design described in the [patent] specifi-

cation differs from the Granger saddle in the sub-

stitution of the Jenifer cantle [raised rear part] for

the low, broad cantle of the Granger tree. In other

words, the front half of the Granger and the rear

half of the Jenifer, or Jenifer-McClellan, make up

the saddle in question, though it differs also from

the Granger saddle in that it has a nearly perpen-

dicular drop of some inches at the rear of the pom-

mel [upper front part], that is, distinctly more of a

drop than the Granger saddle had.

Id. In this way, the Supreme Court explained that the

claimed design was made up of the front half of one known

saddle and the back half (or cantle) of another, with a fur-

ther modification to one half. The Supreme Court ex-

plained that the record evidence showed that there were

several hundred types of saddles, “and that it was custom-

ary for saddlers to vary the shape and appearance of sad-

dletrees [saddle frames] in numerous ways, according to

At the time, patent law did not speak of obvious-

ness. The Whitman Saddle Court addressed the matter by

reference to “the inventive faculty.” 148 U.S. at 679. The

reasoning of Whitman Saddle carries over to the modern

§ 103 standard of obviousness.

Case: 21-2348 Document: 227 Page: 13 Filed: 05/21/2024

LKQ CORPORATION v.

GM GLOBAL TECHNOLOGY OPERATIONS LLC

the taste and fancy of the purchaser.” Id. at 681. Address-

ing the combination of the Granger and Jenifer saddles, the

Court determined that “[n]othing more was done in this in-

stance (except as hereafter noted) than to put the two

halves of these saddles together in the exercise of the ordi-

nary skill of workmen of the trade, and in the way and

manner ordinarily done,” and that “the addition of a known

cantle to a known saddle, in view of the fact that such use

of the cantle was common,” did not make the claimed sad-

dle design inventive. Id.

The Court then turned to the primary “difference be-

tween the pommel of [the claimed] saddle [design] and the

pommel of the Granger saddle, namely, the drop at the rear

of the pommel.” Id. The Court withheld judgment on the

validity of the overall design given this secondary differ-

ence in design. Rather, the Court held that if this second-

ary difference was “what was new and to be material,” it

was not present in the potentially infringing saddle, and as

such, no infringement could occur. Id. at 682.

In short, the Court considered the prior art in the field

of the article of manufacture, the knowledge of an ordinary

saddler, and the differences between the prior art and the

claimed design, and concluded that combining the two

known saddle designs was nothing more than an “exercise

of the ordinary skill of workmen of the [saddle] trade.” Id.

at 681. The Court refrained from drawing the same con-

clusion for a secondary modification without evidence that

it was known in the prior art or similar to the existing prior

art references.

III

Our test for design patent obviousness, in its present

form, does not adequately align with KSR, Whitman Sad-

dle, and other precedent, both in terms of its framework

and threshold rigidity. Rosen-Durling is out of keeping

with the Supreme Court’s general articulation of the

Case: 21-2348 Document: 227 Page: 14 Filed: 05/21/2024

LKQ CORPORATION v.

GM GLOBAL TECHNOLOGY OPERATIONS LLC

principles underlying obviousness, as well as its specific

treatment of validity of design patents.

LKQ urges us to conclude that KSR overruled or abro-

gated Rosen and Durling. LKQ made the same argument

to the panel, which rejected the argument. See Panel Deci-

sion, 2023 WL 328228, at *6 (“[I]t is not clear the Supreme

Court has overruled Rosen or Durling.”). While a panel has

the ability only in certain narrow circumstances to con-

clude that this court’s own precedent “has been implicitly

overruled as inconsistent with intervening Supreme Court

authority,” Troy v. Samson Mfg. Corp., 758 F.3d 1322,

1326 (Fed. Cir. 2014), that demanding standard does not

bind this court sitting en banc. See South Corp. v. United

States, 690 F.2d 1368, 1370 n.2 (Fed. Cir. 1982); Deckers

Corp. v. United States, 752 F.3d 949, 959 (Fed. Cir. 2014).

We therefore have no need to decide whether that standard

is met here.

The en banc court has never considered the merits of

the Rosen-Durling test, and we now take this opportunity,

sitting en banc, to do so in light of § 103, Supreme Court

precedent, and our own precedent guiding the nonobvious-

ness inquiry for utility patents. We do so bearing in mind

the respect due to long-standing panel precedents even

when the en banc court is newly considering an issue. See

Robert Bosch, LLC v. Pylon Mfg. Corp., 719 F.3d 1305,

1316–17 (Fed. Cir. 2013) (en banc).

We conclude that the Rosen-Durling test require-

ments—that (1) the primary reference be “basically the

same” as the challenged design claim; and (2) any second-

ary references be “so related” to the primary reference that

features in one would suggest application of those features

to the other—are improperly rigid. The statutory rubric

along with Supreme Court precedent including Whitman

Saddle, Graham, and KSR, all suggest a more flexible ap-

proach than the Rosen-Durling test when determining ob-

viousness.

Case: 21-2348 Document: 227 Page: 15 Filed: 05/21/2024

LKQ CORPORATION v.

GM GLOBAL TECHNOLOGY OPERATIONS LLC

Design patents and utility patents are, of course, dif-

ferent. Nevertheless, “[d]esign patents, like utility patents,

must meet the nonobviousness requirement of 35 U.S.C.

§ 103.” Titan Tire Corp. v. Case New Holland, Inc.,

566 F.3d 1372, 1380, 1384–85 (Fed. Cir. 2009); see also

35 U.S.C. § 171(b) (“The provisions of this title relating to

patents for inventions shall apply to patents for designs,

except as otherwise provided.”).

The § 103 obviousness language sets forth an expan-

sive and flexible approach for assessing obviousness, in-

volving assessment of the “differences between the claimed

invention and the prior art” and whether those differences

are such that the invention as a whole would have been

obvious to a person of ordinary skill in the field to which

the claimed design pertains. Rosen’s rigid requirement

limiting a primary reference to designs that are “basically

the same” as the claimed design—and abruptly ending the

analysis in the absence of such a reference—imposes limi-

tations absent from § 103’s broad and flexible standard.

The Rosen “basically the same” requirement is also in-

consistent with the Supreme Court’s analysis in Whitman

Saddle. There, the Court did not ask whether the design

of the prior art Granger saddle was “basically the same” as

the claimed saddle design. Instead, the Supreme Court

acknowledged that the claimed design combined the front

half of Granger with the rear half of Jenifer, 148 U.S.

at 680, suggesting that neither Granger nor Jenifer would

qualify as a Rosen reference. Moreover, even though nei-

ther prior art saddle design was “basically the same” as the

claimed design, the Supreme Court continued its analysis

and held that combining the front half of the Granger sad-

dle with the back half of the Jenifer saddle was not in-

ventive because it was “customary for saddlers” to make

such combinations. Id. at 681. This analysis reflects the

reality that the one-size-fits all approach of Rosen—in

Case: 21-2348 Document: 227 Page: 16 Filed: 05/21/2024

LKQ CORPORATION v.

GM GLOBAL TECHNOLOGY OPERATIONS LLC

which a single primary reference discloses nearly every as-

pect of the claimed design—does not fit all obviousness sce-

narios for designs and cannot be reconciled with Supreme

Court precedent.

Indeed, as the Supreme Court explained in KSR, “when

a court transforms [a] general principle into a rigid rule

that limits the obviousness inquiry, . . . it errs.” KSR,

550 U.S. at 419. The Court in KSR emphasized that both

§ 103 and Graham “set forth an expansive and flexible ap-

proach” to obviousness and reiterated “the need for caution

in granting a patent based on the combination of elements

found in the prior art.” Id. at 415. Continuing, the Court

explained, “[r]igid, preventative rules that deny factfinders

recourse to common sense, . . . are neither necessary under

our case law nor consistent with it.” Id. at 421. Because

the strict Rosen reference requirement—in which an obvi-

ousness analysis fails absent a prior art design that is “ba-

sically the same” as the claimed design—is such a rigid

rule, we conclude that it is inconsistent with the Supreme

Court’s general guidance in KSR.

We reach the same conclusion for step two of the Rosen-

Durling test. Under step two, a fact finder may consider

ornamental features missing in a primary reference but

shown in secondary reference(s) only when the secondary

reference(s) are “so related to the primary reference that

the appearance of certain ornamental features in one

would suggest the application of those features to the

other.” Durling, 101 F.3d at 103 (alterations omitted)

(quoting Borden, 90 F.3d at 1575). Like the Rosen refer-

ence requirement, this “so related” requirement is at odds

with the broad standard for obviousness set forth in § 103

and Supreme Court precedent.

First, § 103 requires consideration of the differences

between the claimed design and prior art designs and an

inquiry into whether those differences would have been

Case: 21-2348 Document: 227 Page: 17 Filed: 05/21/2024

LKQ CORPORATION v.

GM GLOBAL TECHNOLOGY OPERATIONS LLC

obvious to a person of ordinary skill in the relevant field.

When the answer to this question is affirmative, the stat-

ute provides that a patent may not be obtained. The stat-

ute gives no indication that a secondary prior art reference

need be “so related” to the primary that—as the Durling

test is too readily understood to demand—it creates its own

motivation to combine the two prior art references.

We agree with LKQ that Durling’s inflexible “so re-

lated” requirement is analogous to the rigid application of

the teaching-suggestion-motivation test rejected by the Su-

preme Court in KSR. Just as the Supreme Court in KSR

explained that the TSM test had sometimes been applied

as if it had evolved from a useful insight to a rigid rule, the

Durling “so related” requirement has evolved into a rigid

rule that limits the broad standard set forth in § 103 and

“den[ies] factfinders recourse to common sense” when as-

sessing a motivation to combine prior art references. KSR,

550 U.S. at 421.

The Durling so-related test is also inconsistent with

the Supreme Court’s decision in Whitman Saddle. There,

in considering a claimed design that essentially combined

the front half of the Granger saddle with the rear half (spe-

cifically, the cantle) of the Jenifer saddle, the Supreme

Court did not ask whether the Granger and Jenifer refer-

ences were “so related” that the use of certain ornamental

features in one would suggest the application of those fea-

tures to the other. Whitman Saddle, 148 U.S. at 680–81.

Rather, in a manner consistent with the Graham frame-

work for assessing obviousness of utility patents, the Su-

preme Court explained that it was not inventive for a

saddler to combine the front half of Granger with the cantle

of Jenifer because record evidence established that “it was

customary for saddlers to vary the shape and appearance

of saddletrees in numerous ways, according to the taste

and fancy of the purchaser” and “the Jenifer cantle was

used upon a variety of saddles.” Id. at 681. The Court rea-

soned that, based on this evidence, “[n]othing more was

Case: 21-2348 Document: 227 Page: 18 Filed: 05/21/2024

LKQ CORPORATION v.

GM GLOBAL TECHNOLOGY OPERATIONS LLC

done in this instance (except as hereafter noted) than to

put the two halves of these saddles together in the exercise

of the ordinary skill of workmen of the trade, and in the

way and manner ordinarily done.” Id. Continuing, the

Court explained the “addition of a known cantle to a known

saddle,” was not inventive “in view of the fact that such use

of the cantle was common.” Id. The Court thus relied on

the knowledge and practice of a saddler or ordinary work-

man of the trade instead of a “so related” requirement to

combine prior art references. Id.

Having overruled Rosen and Durling, we consider what

the framework for evaluating obviousness of design patent

claims should be. We are guided by the language of § 103,

as well as the Supreme Court’s and our court’s precedent

on obviousness in both the design and utility patent con-

texts. We are also mindful of the differences between de-

sign and utility patents, as well as the policy concerns

emphasized by the parties and amici. Ultimately, we agree

with our precedent holding that “[i]nvalidity based on ob-

viousness of a patented design is determined [based] on fac-

tual criteria similar to those that have been developed as

analytical tools for reviewing the validity of a utility patent

under § 103, that is, on application of the Graham factors.”

Hupp v. Siroflex of Am., Inc., 122 F.3d 1456, 1462

(Fed. Cir. 1997).

Applying Graham factor one, the fact finder should

consider the “scope and content of the prior art” within the

knowledge of an ordinary designer in the field of the design.

See Graham, 383 U.S. at 17. In determining the scope of

the prior art, there is no threshold similarity or “basically

the same” requirement to qualify as prior art. Rather, an

analogous art requirement applies to each reference. The

analogous art requirement reins in the scope of prior art

and serves to guard against hindsight. We reaffirm that

Case: 21-2348 Document: 227 Page: 19 Filed: 05/21/2024

LKQ CORPORATION v.

GM GLOBAL TECHNOLOGY OPERATIONS LLC

an analogous art requirement applies for obviousness of de-

sign patents.

We have stated that a “reference qualifies as prior art

for an obviousness determination only when it is analogous

to the claimed invention.” Airbus S.A.S. v. Firepass Corp.,

941 F.3d 1374, 1379 (Fed. Cir. 2019) (citing In re Klein,

647 F.3d 1343, 1348 (Fed. Cir. 2011)). The analogous art

requirement is tied to the statutory language of § 103 in

that it provides the perspective of the person of ordinary

skill in the art by defining that person’s scope of knowledge.

The rationale is that a person of ordinary skill “could not

possibly be aware of every teaching in every art” and thus

we limit the scope of the prior art to analogous arts. See

Airbus, 941 F.3d at 1381–82 (quoting In re Wood, 599 F.2d

1032, 1036 (CCPA 1979)). The analogous art requirement

is also “meant to defend against hindsight.” See Netflix,

Inc. v. DivX, LLC, 80 F.4th

1352, 1358 (Fed. Cir. 2023)

(quoting Kahn, 441 F.3d at 987); see also In re Oetiker,

977 F.2d 1443, 1447 (Fed. Cir. 1992) (“The combination of

elements from non-analogous sources, in a manner that re-

constructs the applicant’s invention only with the benefit

Neither party has argued at any stage of this ap-

peal that an analogous art requirement should not apply

when determining obviousness of design patent claims.

See Oral Arg. at 0:30–0:57 (LKQ agreed that a design pa-

tent analogous arts test should exist); id. at 31:42–31:47

(United States agreed with the same); and id. at 1:09:56–

1:10:11 (GM agreed regarding the potential scope of analo-

gous art), https://oralarguments.cafc.uscourts.gov/de-

fault.aspx?fl=21-2348_02052024.mp3. It is also well

established that the analogous art requirement is part of

the Graham analysis. See In re Kahn, 441 F.3d 977, 986

(Fed. Cir. 2006) (“[T]he ‘analogous art’ test . . . has long

been part of the primary Graham analysis articulated by

the Supreme Court.”).

Case: 21-2348 Document: 227 Page: 20 Filed: 05/21/2024

LKQ CORPORATION v.

GM GLOBAL TECHNOLOGY OPERATIONS LLC

of hindsight, is insufficient to present a prima facie case of

obviousness.”).

An important precedent that preceded adoption of the

Rosen-Durling test, and introduced the “so related” phrase,

supports application of an analogous art requirement to de-

sign patents. In In re Glavas, 230 F.2d 447 (CCPA 1956),

the patent at issue related to the ornamental design for a

float (e.g., a life preserver) and the Board affirmed a final

rejection of the patent based on a patent relating to a float

in view of patents relating to a cushion and a pillow. The

Board added an additional ground for rejection based on

other secondary references—patents relating to bottles, a

razor blade sharpener, and a bar of soap. On appeal, the

court grappled with whether the analogous art require-

ment for utility patents applies to obviousness of design pa-

tents. Ultimately, the court concluded that “[t]he principle

of nonanalogous arts . . . cannot be applied to design cases

in exactly the same manner as to mechanical cases,” be-

cause design patent combinations rely on appearance ra-

ther than use. Glavas, 230 F.2d at 450. In the very next

sentence, the Glavas court then created the “so-related” re-

quirement for design cases: “The question in design cases

is not whether the references sought to be combined are in

analogous arts in the mechanical sense, but whether they

are so related that the appearance of certain ornamental

features in one would suggest the application of those fea-

tures to the other.” Id. In applying the “so related” test,

the court addressed the analogous art issue by noting that

while pillows may plausibly be related art to a float, a de-

signer of a float would not likely turn to bottles, soap, or

razor blade sharpeners for design inspiration. Id. at 450–

51.

Glavas thus created the “so related” test as a design-

context proxy or replacement for the analogous art require-

ment of utility cases. Unfortunately, application of the

Rosen-Durling test over time has become rigid to the point

where we abandon it today. But there is no basis for

Case: 21-2348 Document: 227 Page: 21 Filed: 05/21/2024

LKQ CORPORATION v.

GM GLOBAL TECHNOLOGY OPERATIONS LLC

abandoning the underlying analogous art requirement for

assessing nonobviousness of design patents and we return

to a more flexible fact-based analysis of whether the refer-

ences are analogous art in a manner similar to utility pa-

tents. This approach casts aside a threshold “so-related”

requirement but maintains the threshold analogous art re-

quirement.

For utility patents, we use a two-part test to determine

the scope of analogous art: (1) whether the art is from the

same field of endeavor as the claimed invention; and (2) if

the reference is not within the field of the inventor’s en-

deavor, whether the reference still is reasonably pertinent

to the particular problem with which the inventor is in-

volved. See Airbus S.A.S, 941 F.3d at 1379. This is a fact-

dependent inquiry that requires a case-by-case determina-

tion. See id. (“Whether a reference qualifies as analogous

prior art is a question of fact . . . .”).

We conclude that the first part of this two-part test ap-

plies to design patents in a straightforward manner. In

other words, analogous art for a design patent includes art

from the same field of endeavor as the article of manufac-

ture of the claimed design. See 35 U.S.C. § 171(a) (“Who-

ever invents any new, original and ornamental design for

an article of manufacture may obtain a patent there-

for . . . .” (emphasis added)). As we have previously held,

“[t]he scope of the prior art is not the universe of abstract

design and artistic creativity, but designs of the same arti-

cle of manufacture or of articles sufficiently similar that a

person of ordinary skill would look to such articles for their

designs.” Hupp, 122 F.3d at 1462.

On the other hand, the second part of the two-part

analogous art test for utility patents would not seem to ap-

ply to design patents in the same way, and how to translate

this part of the test into the design context is less apparent.

Unlike a utility patent, a design patent itself does not

clearly or reliably indicate “the particular problem with

Case: 21-2348 Document: 227 Page: 22 Filed: 05/21/2024

LKQ CORPORATION v.

GM GLOBAL TECHNOLOGY OPERATIONS LLC

which the inventor is involved.” See Airbus S.A.S, 941 F.3d

at 1379 (emphasis added). Indeed, design patents have no

written description or written claims to define their scope;

the invention is defined by the overall visual impression

that the drawings convey. See Egyptian Goddess, Inc.

v. Swisa Inc., 543 F.3d 665, 679–80 (Fed. Cir. 2008) (en

banc) (a design claim is better represented by an illustra-

tion). Moreover, at oral argument, the parties disputed

whether an ornamental design for an article of manufac-

ture could be viewed as solving a problem. See Oral Arg.

at 2:08–3:11, 1:09:12–1:10:29, https://oralarguments.cafc.

uscourts.gov/default.aspx?fl=21-2348_02052024.mp3.

In this opinion, we do not delineate the full and precise

contours of the analogous art test for design patents. Prior

art designs for the same field of endeavor as the article of

manufacture will be analogous, and we do not foreclose

that other art could also be analogous. Whether a prior art

design is analogous to the claimed design for an article of

manufacture is a fact question to be addressed on a case-

by-case basis and we “leave it to future cases to further de-

velop the application of this standard.” Cf. Egyptian God-

dess, 543 F.3d at 679 (quoting In re Seagate Tech., LLC,

497 F.3d 1360, 1371 (Fed. Cir. 2007) (en banc)).

In applying Graham factor one, the scope and content

of the prior art, a primary reference must be identified.

Identifying a primary reference protects against hindsight

because “almost every new design is made up of elements

which, individually, are old somewhere in the prior art.”

Glavas, 230 F.2d at 450. As noted above, this primary ref-

erence need not be “basically the same” as the claimed de-

sign. Rather, the primary reference need only be

“something in existence—not . . . something that might be

brought into existence by selecting individual features

from prior art and combining them, particularly where

combining them would require modification of every indi-

vidual feature.” In re Jennings, 182 F.2d 207, 208 (CCPA

1950). The primary reference will likely be the closest prior

Case: 21-2348 Document: 227 Page: 23 Filed: 05/21/2024

LKQ CORPORATION v.

GM GLOBAL TECHNOLOGY OPERATIONS LLC

art, i.e., the prior art design that is most visually similar to

the claimed design. The more visually similar the primary

reference design is to the claimed design, the better posi-

tioned the patent challenger will be to prove its § 103 case.

The primary reference will typically be in the same field of

endeavor as the claimed ornamental design’s article of

manufacture, but it need not be, so long as it is analogous

art. It is often the case that the primary reference alone

does not render the claimed design obvious, and design el-

ements from secondary prior art references may be consid-

ered as discussed below.

After considering the scope and content of the prior art,

we turn to Graham factor two: determining the differences

between the prior art designs and the design claim at issue.

Graham, 383 U.S. at 17. This approach casts aside a

threshold “similarity” requirement.

The Supreme Court has explained that design patents

protect the appearance of the design on an article of man-

ufacture, and thus, in the context of design patent infringe-

ment, we compare the visual appearance of the claimed

design with that of allegedly infringing design. See

Gorham, 81 U.S. at 526–27 (“[T]he true test of identity of

design . . . must be sameness of appearance”); see also

Egyptian Goddess, 543 F.3d at 676 (focusing on the “over-

all appearance of the design” when assessing infringe-

ment). Likewise, in addressing the differences between the

claimed design and prior art designs for validity purposes,

we compare the visual appearance of the claimed design

with prior art designs, albeit from the perspective of an or-

dinary designer in the field of the article of manufacture.

See Apple, Inc. v. Samsung Elects., Co., 678 F.3d 1314,

1329 (Fed. Cir. 2012) (characterizing obviousness as

whether “one of ordinary skill would have combined teach-

ings of the prior art to create the same overall visual ap-

pearance as the claimed design”) (quoting Titan Tire,

Case: 21-2348 Document: 227 Page: 24 Filed: 05/21/2024

LKQ CORPORATION v.

GM GLOBAL TECHNOLOGY OPERATIONS LLC

566 F.3d at 1381); Borden, 90 F.3d at 1574, 1576 (affirm-

ing the Board’s affirmance of an obviousness rejection

where the examiner found “differences between the [prior

art] and appellant’s design had little or no effect on the

overall appearance of the design”); Dobson v. Dornan,

118 U.S. 10, 15 (1886) (“Undoubtedly, the claim in this case

covers the design as a whole, and not any part of it as a

part, and it is to be tested as a whole as to novelty and in-

fringement.”).

Under Graham factor three, “the level of ordinary skill

in the pertinent art [must be] resolved.” Graham, 383 U.S.

at 17. This requirement comes directly from § 103, which

provides that a claimed invention is unpatentable if it

“would have been obvious before the effective filing date of

the claimed invention to a person of ordinary skill in the

art to which the claimed invention pertains.” 35 U.S.C.

§ 103.

In the design patent context, we have interpreted “a

person of ordinary skill in the art to which the invention

pertains” in § 103 as meaning that obviousness of a design

patent claim is assessed from the viewpoint of an ordinary

designer in the field to which the claimed design pertains.

In In re Nalbandian, our predecessor court explained that,

consistent with Graham, “[i]n design cases we will consider

the fictitious person identified in § 103 as ‘one of ordinary

skill in the art’ to be the designer of ordinary capability who

designs articles of the type presented in the application.”

661 F.2d 1214, 1216 (CCPA 1981). In other words, for Gra-

ham factor three, we consider the knowledge of “a designer

of ordinary skill who designs articles of the type involved.”

Borden, 90 F.3d at 1574 (citation omitted); Hupp, 122 F.3d

at 1462 (“The determination of the ultimate question of ob-

viousness is made from the viewpoint of a person of ordi-

nary skill in the field of the patented design.”); Whitman

Saddle, 148 U.S. at 681 (determining patentability from

Case: 21-2348 Document: 227 Page: 25 Filed: 05/21/2024

LKQ CORPORATION v.

GM GLOBAL TECHNOLOGY OPERATIONS LLC

the perspective of “the ordinary skill of workmen of the

trade” and an ordinary “saddler”).

After ascertaining the knowledge of an ordinary de-

signer in the relevant field, the scope and content of the

prior art, and the differences between the prior art and the

claimed design, the obviousness or nonobviousness of the

claimed design is evaluated. We have previously described

this inquiry as whether an ordinary designer in the field to

which the claimed design pertains would have been moti-

vated to modify the prior art design “to create the same

overall visual appearance as the claimed design.” Camp-

bell Soup, Co. v. Gamon Plus, Inc., 10 F.4th 1268, 1275

(Fed. Cir. 2021) (citation and quotation marks omitted); see

also Borden, 90 F.3d at 1574. We reaffirm that the “in-

quiry focuses on the visual impression of the claimed de-

sign as a whole and not on selected individual features.”

Borden, 90 F.3d at 1574.

Where a primary reference alone does not render the

claimed design obvious, secondary references may be con-

sidered. The primary and secondary references need not

be “so related” such that features in one would suggest ap-

plication of those features in the other but they must both

be analogous art to the patented design. Consistent with

KSR, the motivation to combine these references need not

come from the references themselves. KSR, 550 U.S.

at 418–19 (rejecting a rigidly applied TSM test). But there

must be some record-supported reason (without hindsight)

that an ordinary designer in the field of the article of man-

ufacture would have modified the primary reference with

the feature(s) from the secondary reference(s) to create the

same overall appearance as the claimed design. See Camp-

bell Soup, 10 F.4th at 1275 (discussing the question of

whether “an ordinary designer would have modified the

primary reference to create a design that has the same

overall visual appearance as the claimed design”); see also

Case: 21-2348 Document: 227 Page: 26 Filed: 05/21/2024

LKQ CORPORATION v.

GM GLOBAL TECHNOLOGY OPERATIONS LLC

Amicus United States Br. at 12–13 (describing factors to

consider including “ordinarily skilled designer’s experi-

ence, creativity,” as well as “what market demands and in-

dustry customs exist” in the relevant field and, “which

ornamental features are commonplace in the relevant

field”). Just as with the analogous art inquiry, in the area

of motivation to combine, the problem to be solved may

have less relevance in the design patent context than in the

utility patent context. Of course, it follows that the more

different the overall appearances of the primary reference

versus the secondary reference(s), the more work a patent

challenger will likely need to do to establish a motivation

to alter the primary prior art design in light of the second-

ary one and demonstrate obviousness without the aid of

hindsight.

Consistent with Graham, the obviousness inquiry for

design patents still requires assessment of secondary con-

siderations as indicia of obviousness or nonobviousness,

when evidence of such considerations is presented. Gra-

ham, 383 U.S. at 17–18 (“Such secondary considerations as

commercial success, long felt but unsolved needs, failure of

others, etc., might be utilized to give light to the circum-

stances surrounding the origin of the subject matter sought

to be patented.”). In prior cases involving design patents,

we have confirmed that commercial success, industry

praise, and copying may demonstrate nonobviousness of

design patents. Campbell Soup, 10 F.4th at 1276–79 (con-

sidering evidence of commercial success, industry praise,

and copying); MRC Innovations, Inc. v. Hunter Mfg., LLP,

747 F.3d 1326, 1335–36 (Fed. Cir. 2014) (same). We do not

disturb our existing precedent regarding the application of

secondary considerations such as commercial success, in-

dustry praise, and copying to the obviousness analysis in

design patents. It is unclear whether certain other factors

such as long felt but unsolved needs and failure of others

apply in the design patent context. We leave to future

Case: 21-2348 Document: 227 Page: 27 Filed: 05/21/2024

LKQ CORPORATION v.

GM GLOBAL TECHNOLOGY OPERATIONS LLC

cases the determination of whether considerations such as

long felt but unsolved need and failure of others will have

significance to the obviousness inquiry in the design patent

context.

We recognize that GM and several amici have raised

concerns as to the uncertainty that may result from over-

ruling the Rosen-Durling test and applying a new test. GM

argues that eliminating the test will “revert to a rudderless

free-for-all [and] will increase confusion, disrupt settled ex-

pectations, and leave lower courts and factfinders without

the necessary guidelines to properly conduct the obvious-

ness analysis.” Appellee’s Br. at 13. Along the same lines,

amici argue that “[e]liminating this framework now carries

great risk, including exposing up to 400,000 U.S[.] design

patents to substantial uncertainty [and] upsetting settled

expectations for applicants before the USPTO and the pub-

lic at large.” Amicus Am. Intell. Prop. L. Ass’n Br. at 26;

see also Amicus Intell. Prop. Owners Ass’n Br. at 15 (“Elim-

ination would leave USPTO examiners and design patent

owners without any analytical framework for evaluating

obviousness tailored to design patents.”).

We do not agree. The Graham four-part obviousness

test for utility patents has existed for a very long time and

there is considerable precedent from which the PTO and

the courts can draw when assessing obviousness in the de-

sign patent context. As the Supreme Court acknowledged

in Graham, there may well be some difficulties with this

It is possible that the rigid requirements of Rosen-

Durling have, to date, prevented the development of a sub-

stantial body of caselaw considering the application of sec-

ondary considerations to obviousness analyses of design

patents.

Case: 21-2348 Document: 227 Page: 28 Filed: 05/21/2024

LKQ CORPORATION v.

GM GLOBAL TECHNOLOGY OPERATIONS LLC

fact-based non-rigid test, but these difficulties are not un-

usual:

This is not to say, however, that there will not be

difficulties in applying the nonobviousness test.

What is obvious is not a question upon which there

is likely to be uniformity of thought in every given

factual context. The difficulties, however, are com-

parable to those encountered daily by the courts in

such frames of reference as negligence and scien-

ter, and should be amendable to a case-by-case de-

velopment.

Graham, 383 U.S. at 18. This test has proven workable for

utility patents and we see no reason why it would not be

similarly workable for design patents. As with any change,

there may be some degree of uncertainty for at least a brief

period, but our elimination of the rigid Rosen-Durling test

is compelled by both the statute and Supreme Court prec-

edent.

Turning to the merits of the specific case before us, we

vacate the Board’s determination of nonobviousness of the

D’625 patent. We remand for the Board to address in the

first instance whether the D’625 patent would have been

nonobvious under 35 U.S.C. § 103 based on Lian alone or

Lian as modified by Tucson, applying the framework for

evaluating obviousness of design patent claims set forth in

this opinion.

As an alternative ground for holding the D’625 patent

nonobvious, the Board held that LKQ’s failure to identify

“the correct visual impression created by the patented de-

sign as a whole” in its claim construction, “standing alone,

justifies [the Board’s] conclusion that LKQ fails also to

prove by a preponderance of the evidence that the chal-

lenged claim of the ’625 patent is unpatentable.” Board

Decision, 2021 WL 3411458, at *17 (quoting Durling,

Case: 21-2348 Document: 227 Page: 29 Filed: 05/21/2024

LKQ CORPORATION v.

GM GLOBAL TECHNOLOGY OPERATIONS LLC

101 F.3d at 103). We have repeatedly held that tribunals

“should not treat the process of claim construction [in de-

sign patent cases] as requiring a detailed verbal descrip-

tion of the claimed design, as would typically be true in the

case of utility patents.” Egyptian Goddess, 543 F.3d at 680;

see also Crocs, Inc. v. Int’l Trade Comm’n, 598 F.3d 1294,

1302–03 (Fed. Cir. 2010) (stating that for design patents,

“claim construction must be adapted to a pictorial setting”

and warning of “the dangers of reliance on a detailed verbal

claim construction”). Rather, the tribunal is not “obligated

to issue a detailed verbal description of the design” as “a

design is better represented by an illustration.” Egyptian

Goddess, 543 F.3d at 679. We reaffirm this principle and

hold that the Board erred in rejecting LKQ’s asserted obvi-

ousness grounds on this basis.

CONCLUSION

As explained above, we reinstate the portion of the

panel opinion affirming the Board’s finding of no anticipa-

tion. We also vacate the Board’s nonobviousness determi-

nation and remand for further proceedings consistent with

this opinion.

AFFIRMED-IN-PART, VACATED-IN-PART, AND

REMANDED

COSTS

No costs.

Case: 21-2348 Document: 227 Page: 30 Filed: 05/21/2024

United States Court of Appeals

for the Federal Circuit

______________________

LKQ CORPORATION, KEYSTONE AUTOMOTIVE

INDUSTRIES, INC.,

Appellants

GM GLOBAL TECHNOLOGY OPERATIONS LLC,

Appellee

______________________

2021-2348

______________________

Appeal from the United States Patent and Trademark

Office, Patent Trial and Appeal Board in No. IPR2020-

00534.

______________________

LOURIE, Circuit Judge, concurring in judgment.

I concur in the court’s decision to vacate and remand

the Board’s decision for it to reevaluate the validity of the

claimed design in light of the majority opinion.

I write separately because I do not agree that we need

to “overrule” In re Rosen, 673 F.2d 388 (CCPA 1982) and

Durling v. Spectrum Furniture Co., 101 F.3d 100 (Fed. Cir.

1996). To the extent that KSR International Co. v. Teleflex

Inc., 550 U.S. 398 (2007) abjured what it called rigid tests

for evaluating patent issues, all we needed to do in this case

was to make the Rosen and Durling precedents less rigid.

Their use of “must” and “only” in their analyses are the only

signs of rigidity in those opinions. Replacing them with

Case: 21-2348 Document: 227 Page: 31 Filed: 05/21/2024

LKQ CORPORATION v.

GM GLOBAL TECHNOLOGY OPERATIONS LLC

words such as “generally,” “usually,” or “typically” would

have sufficed. Otherwise, they were basically correct on

the law.

Rosen was decided by a Court of Customs and Patent

Appeals (“CCPA”) panel that consisted in part of Judges

Markey, Rich, and Nies, a group that one might say knew

a bit about patent law, including the analysis of obvious-

ness. And Durling, decided by an equally knowledgeable

panel of this court, merely cemented an already long-stand-

ing principle likewise established by the CCPA in In re Gla-

vas, 230 F.2d 447, 450 (CCPA 1956). The basics of the

Rosen-Durling framework have therefore existed, and been

consistently applied to design patent validity, for decades

and long after Smith v. Whitman Saddle Co., 148 U.S. 674

(1893).

Perhaps, in the fullness of time and in light of KSR, it

can be seen that Rosen and Durling may have used overly

strong language. But, “overruled”? It only needed to be

modified. Overruling is destabilizing. Courts should, if

possible, rule in modest steps, not abrupt acts of overrul-

ing. Even post-KSR, we did not wholly abandon the re-

quirement that there be a motivation to combine simply

because the Supreme Court criticized one instance of its

application. And since KSR, we have continued to apply

tests and frameworks for validity. See, e.g., Graham v.

John Deere Co., 383 U.S. 1 (1965).

The majority walks away from Rosen and Durling as if

they violated basic tenets of patent law. In reality, all they

needed was a bit of tinkering. It is important to note that

the U.S. Patent and Trademark Office, which also knows a

good deal about evaluating design patents, basically en-

dorsed Rosen and Durling, but found their application

needed a little more flexibility. Oral Arg. at 32:25–48

(“We’re saying that the existing test is fine . . . . The Rosen-

Durling test is fine, but we do recommend some changes to

it.”).

Case: 21-2348 Document: 227 Page: 32 Filed: 05/21/2024

LKQ CORPORATION v.

GM GLOBAL TECHNOLOGY OPERATIONS LLC

The substantial briefing by amici in this en banc appeal

was primarily dominated by entities with economic inter-

est in the outcome of the case—supporters of automotive

manufacturers, automotive part replacement companies,

and insurance companies, each advocating for results and

legal concepts supporting their economic interests. That is

fine, but we most benefit from entities without an economic

bias.

Broad-based organizations with a primary interest in

the stability and clarity of the law have told us that the

Rosen-Durling test is basically correct. They include the

American Intellectual Property Law Association, the Insti-

tute for Design Science and Public Policy, the International

Trademark Association, the Industrial Designers Society of

America, and others, all of whom supported this precedent,

which the court now “overrules.” Am. Intellectual Prop.

Ass’n Amicus Br., ECF 136 at 31 (“[T]he Rosen-Durling

framework . . . allows for an expansive and flexible ap-

proach consistent with the principles of KSR” and “elimi-

nating [it] would cause unnecessary uncertainty in an area

of design patent law that has worked reasonably well for

over 40 years.”); Inst. for Design Sci. and Pub. Policy et al.

Amicus Br., ECF 144 at 37 (“Eliminating the Rosen-Durl-

ing test will unnecessarily harm design innovation.”); In-

dus. Designers Soc’y of Am. Amicus Br., ECF 193 at 5 (“[I]t

is evident that this Court should uphold the existing Rosen-

Durling framework.”); Int’l Trademark Ass’n Amicus Br.,

ECF 201 at 12 (“The Rosen-Durling test sets the appropri-

ate balance for determining the obviousness of a design,

and the Court should not abandon it.”).

The essence of Rosen and Durling is that for obvious-

ness, which generally assumes that one reference does not

anticipate the claimed design, one ought to have a basic

starting reference with “basically the same” “design char-

acteristics” as the challenged design and a second one that

is “so related” as to result in a conclusion of obviousness.

Rosen, 673 F.2d at 391; Durling, 101 F.3d at 103 (citation

Case: 21-2348 Document: 227 Page: 33 Filed: 05/21/2024

LKQ CORPORATION v.

GM GLOBAL TECHNOLOGY OPERATIONS LLC

omitted). What the majority is doing here is confirming

these concepts while overruling their cases of origin, just to

make minor modifications in their language.

Presumably, such a basic starting reference, just as

with utility patents, should be the closest prior art. The

majority uses that phrase only once, Maj. Op. at 23–24, but

it ought to be central, as it is with utility patents. It does

state that “a primary reference must be identified” and it

“will likely be . . . the prior art design that is most visually

similar to the claimed design.” Id. But we did not need to

“overrule” a case with language that a reference’s “design

characteristics” must be “basically the same” to substitute

“most visually similar.” Those two phrases themselves are

basically the same or similar.

Maybe the word “must”

used in Rosen was excessive, but the thought expressed

was sound.

As the majority acknowledges, the primary reference

must fall within the range of what may be called analogous

art. Maj. Op. at 24 (“The primary reference will typically

be in the same field of endeavor as the claimed ornamental

design’s article of manufacture, but it need not be, so long

“Basically the same” did not mean that a reference

and the challenged design must be near-identical or even

substantially the same. Indeed, such an interpretation

would conflate anticipation and obviousness. See Gorham

Co. v. White, 81 U.S. 511, 528 (1871) (explaining that in-

fringement, which has the same test as anticipation, Int’l

Seaway Trading Corp. v. Walgreens Corp., 589 F.3d 1233,

1240 (Fed. Cir. 2009), considers whether or not the “two

designs are substantially the same”). Rather, “basically”

meant “generally” or “roughly.” Moreover, Rosen did not

require that the primary reference be “basically the same”

as the claimed design, but only that the primary reference

have “design characteristics of which are basically the same

as the claimed design.” 673 F.2d at 391 (emphasis added).

Case: 21-2348 Document: 227 Page: 34 Filed: 05/21/2024

LKQ CORPORATION v.

GM GLOBAL TECHNOLOGY OPERATIONS LLC

as it is analogous art.”). But the majority sets forth the test

for analogous art as part of the obviousness inquiry. See

Maj. Op. Sec. IV.1 (considering whether or not a reference

is analogous art as part of its Graham factor one analysis).

Properly understood, the analogous art test is not a test for

obviousness. It is a gateway. Determining whether art is

analogous is determining what may properly be considered

in an obviousness inquiry, not part of the test for what

would have been considered obvious. It is a threshold for

qualifying art, not an indicator of obviousness. E.g., In re

Klein, 647 F.3d 1343, 1348 (Fed. Cir. 2011) (“A reference

qualifies as prior art for an obviousness determination un-

der § 103 only when it is analogous to the claimed inven-

tion.”). In my view, it should not be defined as a test for

considering obviousness. That misconstrues its role in the

analysis.

Moving to the second step of the Rosen-Durling test, an

examiner or challenger might find another reference that,

combined with the basic reference, might lead a skilled de-

signer to what is claimed. That is where the “so-related”

language in Durling comes in. The additional reference, or

references, in order to lead to obviousness, must be “so re-

lated” to the primary reference that “the appearance of cer-

tain ornamental features in one would suggest the

application of those features to the other.” Durling,

101 F.3d at 103 (quoting In re Borden, 90 F.3d 1570, 1575

(Fed. Cir. 1996) (quoting Glavas, 230 F.2d at 450)). Such a

reference or references would also be within the scope of

analogous art in order to be combinable with the basic ref-

erence. Again, perhaps the word “only” in Durling is exces-

sive. But it just needed to be fine-tuned, not “overruled.”

And that is essentially what the majority did, merely swap-

ping the “so related” language of Durling for “some record-

supported reason . . . that an ordinary designer . . . would

have modified the primary reference . . . .” Maj. Op. at 26.

Rosen and Durling therefore expressed the same con-

cepts now adopted by the majority, albeit with a couple of

Case: 21-2348 Document: 227 Page: 35 Filed: 05/21/2024

LKQ CORPORATION v.

GM GLOBAL TECHNOLOGY OPERATIONS LLC

unnecessarily strong words. Such a basic analysis was

simply the way obviousness is examined and should be con-

sidered with design as well as with utility patents. Indeed,

the Rosen-Durling framework has already been long-re-

garded as the application of the relevant Graham factors

to design patents. See, e.g., Campbell Soup Co. v. Gamon

Plus, Inc., 10 F.4th 1268, 1275 (Fed. Cir. 2021); Hupp v.

Siroflex of Am., Inc., 122 F.3d 1456, 1462 (Fed. Cir. 1997).

With those thoughts, Rosen and Durling should have

been allowed to stand, as modified, rather than been deci-

sively overruled.

With that being said, I agree that the Board’s decision

should be vacated and remanded.

Case: 21-2348 Document: 227 Page: 36 Filed: 05/21/2024