DiscernDx, Inc.

Patent Trials and Appeals Board

Nov 18, 2020

2020000640 (P.T.A.B. Nov. 18, 2020)

UNITED STATES PATENT AND TRADEMARK OFFICE
UNITED STATES DEPARTMENT OF COMMERCE
United States Patent and Trademark Office
Address: COMMISSIONER FOR PATENTS
P.O. Box 1450
Alexandria, Virginia 22313-1450
www.uspto.gov
APPLICATION NO. FILING DATE FIRST NAMED INVENTOR ATTORNEY DOCKET NO. CONFIRMATION NO.
15/414,456 01/24/2017 John Blume 53897-747.201 7289
21971 7590 11/18/2020
WILSON SONSINI GOODRICH & ROSATI
650 PAGE MILL ROAD
PALO ALTO, CA 94304-1050
EXAMINER
AEDER, SEAN E
ART UNIT PAPER NUMBER
1642
NOTIFICATION DATE DELIVERY MODE
11/18/2020 ELECTRONIC
Please find below and/or attached an Office communication concerning this application or proceeding.
The time period for reply, if any, is set in the attached communication.
Notice of the Office communication was sent electronically on above-indicated "Notification Date" to the
following e-mail address(es):
patentdocket@wsgr.com
PTOL-90A (Rev. 04/07)

UNITED STATES PATENT AND TRADEMARK OFFICE
BEFORE THE PATENT TRIAL AND APPEAL BOARD
Ex parte JOHN BLUME, ATHIT KAO,
ROSLYN DILLON, LISA CRONER, and BRUCE WILCOX
Appeal 2020-000640
Application 15/414,456
Technology Center 1600
Before FRANCISCO C. PRATS, TAWEN CHANG, and
CYNTHIA M. HARDMAN, Administrative Patent Judges.
CHANG, Administrative Patent Judge.
DECISION ON APPEAL
Pursuant to 35 U.S.C. § 134(a), Appellant1 appeals from the
Examiner’s decision to reject claims 1, 3, 4, 6, 7, and 18–37. We have
jurisdiction under 35 U.S.C. § 6(b).2
We AFFIRM.

1 We use the word “Appellant” to refer to “applicant” as defined in 37
C.F.R. § 1.42. Appellant identifies the real party in interest as DiscernDx,
Inc. Appeal Br. 4.
2 We heard oral argument relating to this appeal on October 1, 2020.
Appeal 2020-000640
Application 15/414,456

2
STATEMENT OF THE CASE
“Colorectal cancer [(CRC)] is a leading cause of cancer-related deaths
in the United States.” Spec. ¶ 2. The Specification states that colonoscopy
and sigmoidoscopy, “the gold standard for detecting colon cancer,” have low
patient acceptance due in part to their expense and highly invasive nature,
and also may “expose subjects to risk of complications such as infection.”
Id. ¶ 8. The Specification further states that fecal occult blood test (FOBT),
the most common non-invasive test for colorectal cancer, has a high false-
positive rate and a sensitivity of only around 50%, and also may be less
sensitive for detecting early stage CRC. Id. ¶ 9. The Specification also
states that numerous serum markers that have been investigated in
connection with colon cancer have exhibited low sensitivity. Id. ¶ 9.
According to the Specification, the invention relates to methods of assessing
CRC status in an individual by determining blood levels for a panel of
biomarkers comprising the proteins C9, CEA, DPP4, MIF, ORM1, PKM,
SAA, and TFRC, as well as the individual’s age and gender.3 Id. ¶ 12.
CLAIMED SUBJECT MATTER
The claims are directed to a method of assessing colorectal health of
an individual. Claim 1 is illustrative:
1. A method of assessing colorectal health of an
individual, comprising:
obtaining a circulating blood sample from said
individual;

3 C9 and CO9 both refer to the protein complement C9. Spec. 13, Table 1.
Likewise, DPP4 is synonymous with DPPIV, which refers to the protein
dipeptidyl peptidase 4. Id.
Appeal 2020-000640
Application 15/414,456

3
detecting protein levels for each member of a list of
proteins in sample, said list of proteins comprising DPPIV,
CO9, TFRC, ORM1, MIF, PKM, SAA, and CEA; and
obtaining age information and gender information for
said individual.
Appeal Br. 30 (Claims App.).
REJECTIONS
A. Claims 1, 3, 4, 25–29, 32, and 35–37 are rejected under 35 U.S.C.
§ 103 as being unpatentable over Cordero,4 Brock,5 Chen,6 Thorsen,7
Meng,8 Etzioni,9 and Mercer.10 Ans. 3.

4 OJ Cordero et al., Preoperative Serum CD26 Levels: Diagnostic Efficiency
and Predictive Value for Colorectal Cancer, 83 BRITISH J. CANCER 1139
(2000).
5 Randall Brock et al., A Multiplex Serum Protein Assay for Determining the
Probability of Colorectal Cancer, 2 AM. J. CANCER RES. 598 (2012).
6 William Tzu-Liang Chen et al., Identification of Biomarkers to Improve
Diagnostic Sensitivity of Sporadic Colorectal Cancer in Patients with Low
Preoperative Serum Carcinoembryonic Antigen by Clinical Proteomic
Analysis, 412 CLINICA CHIMICA ACTA 636 (2011).
7 Stine Buch Thorsen et al., Detection of Serological Biomarkers by
Proximity Extension Assay for Detection of Colorectal Neoplasias in
Symptomatic Individuals, 11:253 J. TRANSLATIONAL MED. (2013), available
at http://www.translational-medicine.com/content/11/1/253.
8 Wen Meng et al., Serum M2-Pyruvate Kinase: A Promising Non-Invasive
Biomarker for Colorectal Cancer Mass Screening, 4 WORLD J.
GASTROINTESTINAL ONCOLOGY 145 (2012).
9 Ruth Etzioni et al, The Case for Early Detection, 3 NATURE REVIEWS
CANCER 243 (2003), available at https://www.nature.com/articles/nrc1041.
References in the opinion to page numbers in Etzioni refer to the online
version of the article, which is numbered consecutively beginning with page
1 on the first page.
10 Donald W. Mercer, Use of Multiple Markers to Enhance Clinical Utility,
in IMMUNODIAGNOSIS OF CANCER 39 (Ronald B. Herberman & Donald W.
Mercer eds., 2nd ed. 1990).
Appeal 2020-000640
Application 15/414,456

4
B. Claims 1, 3, 4, 6, 7, and 18–37 are rejected under 35 U.S.C. § 103 as
being unpatentable over Cordero, Brock, Chen, Thorsen, Meng,
Etzioni, Mercer, and Vokes.11 Ans. 7.

OPINION
A. Issues
The same issues are dispositive for both of the rejections on appeal.12
We therefore consider the rejections together.
The Examiner concludes that the claims are drawn to methods
comprising detecting levels in blood of known colorectal cancer biomarkers,
i.e., DPPIV, CO9, TFRC, ORM1, MIF, PKM, SAA, and CEA. Ans. 3.
In particular, the Examiner finds that Cordero, Brock, Chen, Thorsen,
and Meng teach using, respectively, (1) DPPIV or CEA levels, (2) CO9,
ORM1, and SAA levels, (3) a combination of CEA and MIF levels, (4)
CEA, TFRC, and MIF levels, and (5) PKM or CEA levels in a blood sample
to detect an individual’s colorectal cancer risk. Ans. 3–5. The Examiner
finds that Cordero, Brock, Thorsen, and Meng also teach the relevance of
age and/or gender to detecting colorectal cancer risk. Id. at 3, 4. Finally, the
Examiner finds that Meng, Etzioni, and Mercer suggest the advantages of

11 Everett E. Vokes & Ralph R. Weichselbaum, Concomitant
Chemoradiotherapy: Rationale and Clinical Experience in Patients with
Solid Tumors, 8 J. CLINICAL ONCOLOGY 911 (1990).
12 In the Answer, the Examiner withdrew the nonstatutory double patenting
rejection over claims 1–21 of US 9,689,874 B2 in view of Cordero, Etzioni,
and Mercer and the provisional nonstatutory double patenting rejection over
claims 31–36 of copending application No. 15/629,593 in view of Cordero,
Etzioni, and Mercer. Ans. 8.
Appeal 2020-000640
Application 15/414,456

5
using multiple tumor markers to diagnose cancer, namely early diagnosis
and/or increased sensitivity and specificity, and that Mercer teaches that
“incorporating CEA into a colorectal cancer detection panel . . . increase[s]
sensitivity.” Id. at 5–6.
The Examiner concludes that a skilled artisan “would have been
motivated with an expectation of success” to combine the teachings of the
cited references to arrive at the claimed invention, because all the proteins
recited in the claims are known colorectal cancer biomarkers in circulating
blood, and using multiple markers for diagnosis has the advantages of early
diagnosis, increased sensitivity and specificity, and greater acceptability to
patients as compared to an invasive procedure such as colonoscopy. Ans. 6.
The Examiner further concludes that the claims would have been obvious
because they recite combinations that are “merely a ‘predictable use of prior
art elements according to their established functions.’” Id.
In a second rejection, the Examiner additionally cites Vokes as
rendering obvious the chemotherapy or radiation treatment recited in certain
dependent claims. Ans. 7.
Appellant contends that the Examiner fails to show that a skilled
artisan would have had reason to combine the cited references to arrive at
the claimed invention, with a reasonable expectation of success. Appeal Br.
11–19. Appellant further contends that the claimed subject matter exhibits
unexpected results. Id. at 19–24; see also Reply Br. 7–8. Finally, with
respect to claims 25, 26, 36, and 37, Appellant contends that the cited
references do not disclose the additional dependent limitations regarding
“wherein said list of proteins comprises no more than 15 [or 8] proteins.”
Appeal Br. 17–18.
Appeal 2020-000640
Application 15/414,456

6
The issues with respect to these rejections are (1) whether a
preponderance of evidence supports the Examiner’s conclusion that a skilled
artisan would have had a reason to combine the cited references to arrive at
the claimed invention, with a reasonable expectation of success; (2) whether
a preponderance of evidence supports the Examiner’s conclusion that the
cited prior art combination teaches or suggests the additional limitations in
dependent claims 25, 26, 36, and 37; and (3) whether Appellant has
presented evidence of unexpected results that, when considered together
with evidence of obviousness, shows the claims to be nonobvious.
B. Analysis
Claims 1, 3, 4, 6, 7, 18–24, and 27–35
With the exception of claims 25, 26, 36, and 37, Appellant does not
separately argue the claims.13 We therefore focus our analysis on claim 1 as
representative of the remaining claims. Unless otherwise noted, we adopt
the Examiner’s findings of fact and reasoning on scope and content of the
prior art with respect to claim 1 (Non-Final Act.14 7–19; Ans. 3–7, 9–24) and
agree that claim 1 is obvious over Cordero, Brock, Chen, Thorsen, Meng,

13 Appellant also refers to claim 27, the other independent claim on appeal,
in the Appeal Brief. However, to present separate argument regarding
appealed claims, “[u]nder each heading identifying the ground of rejection
being contested, any claim(s) argued separately or as a subgroup shall be
argued under a separate subheading that identifies the claim(s) by
number.” 37 C.F.R. 41.37(c)(1)(iv). Appellant has not provided a separate
subheading for claim 27; accordingly, we understand that Appellant is not
separately arguing claim 27. In any event, even if Appellant properly argued
claim 27 separately, our analysis for claim 1 set forth below would apply to
claim 27 as well.
14 Office Action mailed July 6, 2018.
Appeal 2020-000640
Application 15/414,456

7
Etzioni, and Mercer. We respond to Appellant’s arguments below. Only
those arguments timely made by Appellant in the briefs have been
considered; arguments not so presented are waived. See 37 C.F.R.
§ 41.37(c)(1)(iv) (2015); see also Ex parte Borden, 93 USPQ2d 1473, 1474
(BPAI 2010) (informative) (“Any bases for asserting error, whether factual
or legal, that are not raised in the principal brief are waived.”).
Appellant contends that, “[w]hile elements of claims 1 and 27 may be
found scattered throughout some of [the cited prior art] references, there is
no context or disclosure which would lead a person of skill in the art to
select these particular elements.” Appeal Br. 17. In particular, Appellant
contends that
Cordero is limited to a DPPIV diagnostic, and Brock
teaches CO9 only in the context of a particular 7-marker
panel. Chen teaches CEA in combination with MIF or
alpha-enolase. Thorsen describes CRC risk detection
using CEA, TFRC, and MIF only within the context of a
5-marker panel that also includes OPN and CA242.
Meng teaches only the use of PKM as a single marker for
CRC detection in place of CEA. Etzioni and Mercer fail
to provide guidance to arrive at the combination of
markers of claims 1 and 27.
Id.; see also id. at 11–15.
We are not persuaded. As the Examiner points out, whether taught
individually or as part of a biomarker panel, each of the proteins recited in
claim 1 is known in the art to be useful in assessing the colorectal health of
an individual. Ans. 6–7. Based on the disclosures in Etzioni and Mercer
that combining sensitive biomarkers may achieve greater diagnostic
accuracy, a skilled artisan would have had reason to combine these
biomarkers in a method to assess colorectal health. Id. at 5–6; see also, e.g.,
Appeal 2020-000640
Application 15/414,456

8
Chen 638–639, bridging paragraph (“[c]ombining the determinations of
serum CEA . . . with Alpha-enolase or MIF levels to diagnose primary CRC
increased the number of diagnoses”); Thorsen 2, left column (stating that “a
number of studies . . . demonstrated that higher discrimination power could
be obtained by combining biomarkers”).
We further agree with the Examiner that a skilled artisan would have
additional reason to combine the cited references to arrive at the claimed
invention, because the “combination is merely a ‘predictable use of prior art
elements according to their established functions.’” Ans. 6 (quoting KSR
Int’l Co. v. Teleflex, Inc., 550 U.S. 398, 417 (2007)); cf. In re Kerkhoven,
626 F.2d 846, 850 (CCPA 1980) (“It is prima facie obvious to combine two
compositions each of which is taught by the prior art to be useful for the
same purpose, in order to form a third composition which is to be used for
the very same purpose.”).
We are not persuaded by Appellant’s arguments that a skilled artisan,
in view of the prior art, would have understood the recited biomarkers to be
useful only in the specific context or combinations in which the biomarkers
are discussed in each of the references. Appeal Br. 11–15. A reference is
prior art for all that it teaches to a person of ordinary skill in the art, who is
also “a person of ordinary creativity, not an automaton.” Beckman
Instruments, Inc. v. LKB Produkter AB, 892 F.2d 1547, 1551 (Fed. Cir.
1989); KSR Int’l Co. v. Teleflex Inc., 550 U.S. 398, 418 (2007).
Furthermore, “[n]on-obviousness cannot be established by attacking
references individually where the rejection is based upon the teachings of a
combination of references. . . . [The reference] must be read, not in
isolation, but for what it fairly teaches in combination with the prior art as a
Appeal 2020-000640
Application 15/414,456

9
whole.” In re Merck & Co., 800 F.2d 1091, 1097 (Fed. Cir. 1986). Thus,
we agree with the Examiner that a skilled artisan reading the combination of
cited references would have understood that each of the biomarkers recited
in claim 1 is useful in assessing colorectal health of an individual in
combination with other sensitive biomarkers for, e.g., detecting CRC, even if
the other biomarkers are not disclosed in the same reference.
Appellant contends that, “[w]hile elements of claims 1 and 27 may be
found scattered throughout some of these various references, there is no
context or disclosure which would lead a person of skill in the art to select
these particular elements and expect their combination to be successful.”
Appeal Br. 17. Appellant contends that, thus, “to arrive at the combination
of markers recited . . . , one skilled in the art must ‘vary all parameters or try
each of the numerous possible choices’ of the references without ‘direction
as to which of the many choices is likely to be successful,’” which “courts
have held not to be a reasonable expectation of success.” Id. (quoting
Medichem, S.A. v. Rolabo, 437 F.3d 1157, 1165 (Fed. Cir. 2006)).
More particularly, Appellant contends that
Brock, Chen, and Thorsen highlight how a combination
of promising markers cherry-picked from the art is likely
to be found ineffective upon further evaluation. The
Office fails to offer any explanation for how the low
success rate demonstrated by the references of record
(e.g. ⁓6.8% success rate in Thorsen) could suggest a
particular panel of markers out of countless possible
panels would have a reasonable expectation of success.
By arbitrarily picking markers in isolation from disparate
references to arrive at the claimed panel of markers that
has been experimentally validated only in the pending
application, the Office is engaging in impermissible
hindsight analysis.
Appeal 2020-000640
Application 15/414,456

10
Id. at 18–19; see also id. at 13–14.
We are not persuaded. As the Examiner notes, claim 1 does not
require the claimed method of assessing colorectal health to achieve any
particular level of sensitivity, specificity, or efficiency. Ans. 16. As the
Examiner also notes, all of the recited biomarkers in claim 1 have been
validated in the prior art as blood biomarkers for assessing colorectal health
(e.g., by detecting colorectal cancer). Id. at 20. Nothing cited by Appellant
suggests that biomarkers known to be useful in assessing colorectal health of
an individual would become ineffective when combined with other known
sensitive biomarkers, even if there may be variations in sensitivity and
specificity based on the composition of a particular panel. Thus, contrary to
Appellant’s assertion, to arrive at the combination of claim 1 does not
require “‘vary[ing] all parameters or try each of the numerous possible
choices’ of the references without ‘direction as to which of the many choices
is likely to be successful.’” Finally, we note that “the expectation of success
need only be reasonable, not absolute.” Pfizer, Inc. v. Apotex, Inc., 480 F.3d
1348, 1364 (Fed. Cir. 2007).
Brock, Chen, and Thorsen do not suggest the contrary. Appellant
contends that
Brock taught that an initial list of 14 promising candidate
proteins identified by mass spectrometry had to be culled
to arrive at a 7-marker panel to improve its diagnostic
effectiveness. Brock demonstrates that a panel is not
necessarily improved by having more biomarkers and, in
fact, is diminished by having the wrong combination of
markers — thus highlighting the unpredictability of
assembling biomarker panels.
Appeal Br. 18.
Appeal 2020-000640
Application 15/414,456

11
We are not persuaded. As the Examiner points out and as discussed
above, claim 1 does not require the method to have any particular degree of
sensitivity, specificity, or efficiency. Ans. 20. Brock teaches that the
peptides reported in its final panel of 6 biomarkers “provide the model with
the best sensitivity, specificity and negative predictive value.” Brock 599,
left column (emphasis added). However, Appellant has not pointed us to
any portion of Brock that persuasively suggests that the other 46 peptides
from the 14 proteins identified in its biomarker discovery project would not
have been useful in methods to assess colorectal health.
Moreover, the biomarkers recited in claim 1 were not merely
identified in the prior art as differentially expressed in normal versus CRC
serum samples, as were the initial 46 peptides from 14 proteins identified in
Brock’s biomarker discovery project. Brock 598–599, bridging paragraph.
Instead, they were specifically mentioned in their respective reference as
useful in assessing colorectal health (e.g., in diagnosing CRC) either
individually or as part of a panel.
With respect to Chen, Appellant contends that “Chen initially
investigated HSP27 as a potential marker for colon cancer” but that, “despite
being identified as being overexpressed an average of 4.2-fold in CRC tissue
using mass spectrometry analysis . . . , HSP27 ultimately failed to
demonstrate predictive efficacy during validation experiments.” Appeal Br.
18. We are not persuaded. As the Examiner points out, unlike HSP27,
which was known to be overexpressed in tumor tissue but is not significantly
overexpressed in serum levels of the CRC patients examined, each of the
biomarkers recited in claim 1 was known to be a serum biomarker. Ans. 20;
Chen Abstract. Thus, the fact that “HSP27 did not appear to be a suitable
Appeal 2020-000640
Application 15/414,456

12
candidate serum biomarker” in Chen does not suggest that a skilled artisan
would not have found it obvious to use known serum biomarkers, such as
those recited in claim 1, in a method for assessing colorectal health of an
individual that comprises detecting protein levels in a circulating blood
sample of the individual. Chen 640, left column.
Finally, as discussed above and as the Examiner points out, each of
the recited biomarkers in claim 1 have been “experimentally validated as
colorectal cancer blood biomarkers by cited prior art related to colorectal
cancer diagnosis.” Ans. 20. In contrast, the initial pool of 74 potential CRC
biomarkers screened in Thorsen were selected from literature using the more
general criteria of “proteins present in the pathways involved in CRC,
proteins involved in inflammation and cancer, proteins found in biomarker
screening studies of CRC tumor tissue and general cancer markers.”
Thorsen 3, left column. Thus, we are not persuaded by Appellant’s
argument that the alleged “low success rate demonstrated by the references
of record (e.g., ~6.8% success rate in Thorsen)” shows that a skilled artisan
would not have had a reasonable expectation that a particular biomarker,
such as those claimed in claim 1, would be successful in a method for
assessing colorectal health. Appeal Br. 13–14, 19.
Appellant also contends that, with respect to the limitation regarding
“obtaining age . . . and gender information for said individual,” none of
Cordero, Thorsen, and Meng suggests that age and gender are useful for
diagnosing colorectal cancer risk, that Meng found that gender and age did
Appeal 2020-000640
Application 15/414,456

13
not affect PKM levels,15 and that Cordero in fact “teaches that age and
gender are not useful for assessing colorectal health . . . and discourages
combining these elements into a panel with DPPIV.” Appeal Br. 12, 14.
We are not persuaded. As the Examiner noted, claim 1 does not recite
any limitations regarding how age and gender information should be used in
the claimed method. Ans. 14. Appellant does not dispute that Cordero,
Thorsen, and Meng all disclose obtaining age and gender information from
the subjects in their study. In addition, the Examiner cites in the Answer
evidence that relationships between age/gender and colorectal cancer are
known in the art. Id. at 14–15; see also, e.g., Brock 598, left column
(describing a recommendation that screening for CRC begin at age 50 and
continue until age 75). Appellant has not persuasively disputed such
evidence in the Reply Brief.
We agree with the Examiner that the cited prior art does not suggest
that age and gender are irrelevant to assessing colorectal health of an
individual. As the Examiner points out, Cordero in fact teaches that “[t]here
were no significant differences between the two sex- and age-groups for
both donor and patient samples” in terms of serum CD26 concentration.
Ans. 15; Cordero 1140, right column; 1141, Table 1. In short, Cordero and
Meng merely teach that there may not be a significant relationship between
gender/age and levels of particular biomarkers, not that gender and age is
irrelevant to assessing colorectal health.

15 We note that Meng appears to teach that “[a]mong the normal group,
there was no significant difference in serum M2-PK between men and
women . . . or between different age groups.” Meng 147, right column
(emphasis added).
Appeal 2020-000640
Application 15/414,456

14
Finally, citing to Figure 7 of the Specification and the Wilcox
Declaration,16 Appellant contends that the particular biomarker panels
recited in claims 1 and 27 exhibit unexpected results. Appeal Br. 19.
To facilitate our discussion, Exhibit A to the Wilcox Declaration is
reproduced below:

Wilcox Decl. Ex. A. Exhibit A to the Wilcox Declaration depicts a modified
version of Figure 7 of the Specification, which plots, for 1,000 biomarker
panels, the frequency of panels having a given area under the curve (AUC)

16 Declaration of Bruce Wilcox, Ph.D. under 37 C.F.R. § 1.132 (Aug. 16,
2017).
Appeal 2020-000640
Application 15/414,456

15
value as a function of the AUC value.17 Wilcox Decl. ¶¶ 14–17; Spec.
¶ 225. Exhibit A of the Wilcox Declaration further indicates the subset of
biomarker panels comprising DPPIV, CEA, and CO9 (i.e., the biomarkers
recited in claim 27) by solid black shading. Wilcox Decl. ¶ 20.
Our reviewing court has explained that “[m]ere improvement in
properties does not always suffice to show unexpected results,” but “when
an applicant demonstrates substantially improved results . . . and states that
the results were unexpected, this should suffice to establish unexpected
results in the absence of evidence to the contrary.” In re Soni, 54 F.3d 746,
751, 34 USPQ2d 1684, 1688 (Fed. Cir. 1995) (emphasis in original).
Moreover, “when unexpected results are used as evidence of
nonobviousness, the results must be shown to be unexpected compared with
the closest prior art” and “be commensurate in scope with the claims.” In re
Baxter Travenol Labs., 952 F.2d 388, 392 (Fed. Cir. 1991); In re Dill, 604
F.2d 1356, 1361 (CCPA 1979).
In this case, we agree with the Examiner that Appellant has not
demonstrated that the claimed combination of biomarkers exhibits

17 We understand that the area under a receiver operating characteristic
(ROC) curve is one way to measure the performance of a biomarker panel in
assessing a health condition. Spec. ¶ 14 (describing “a Validation Area
Under curve (AUC)” as “a parameter of panel test success”), ¶ 97 (stating
that “[t]he overall performance of a classifier is assessed in some cases via
the AUC of the ROC”). More specifically, we understand that an ROC
curves may be used to represent “[t]he quantitative relationship between
sensitivity and specificity,” wherein “[t]he x-axis of a ROC curve shows the
false-positive rate of an assay” and “[t]he y-axis of ROC curve reports the
sensitivity for an assay.” Spec. ¶ 256. “This allows one to . . . determine a
sensitivity of an assay for a given specificity, and vice versa.” Id.
Appeal 2020-000640
Application 15/414,456

16
substantially improved results that were unexpected as compared to the
closest prior art. Ans. 21.
As an initial matter, we acknowledge Appellant’s argument that the
comparison of biomarker panels comprising DPPIV, CEA, and CO9 to other
panels in Figure 7 is a comparison to the prior art because the panels are
generated by reference to the art, and further acknowledge Appellant’s
argument that Figure 7 in fact overestimates AUC values that may be
generated from a random selection of markers in the art because “the pool
from which the panels of Figure 7 draws is already enriched for higher-
performing markers.” Appeal Br. 20–21; Wilcox Decl. ¶ 18. We similarly
acknowledge Dr. Wilcox’s statements that “[p]anels comprising DPPIV,
CEA, and CO9 substantially outperform other panels in the set” and that
“[n]othing in the art as [he] understands it suggests that panels comprising
DPPIV, CEA and CO9 would out-perform other panels for the assessment of
colorectal health.” Wilcox Decl. ¶¶ 22, 25.
Nevertheless, as the Examiner points out, neither the Specification nor
the Wilcox Declaration provides sufficient evidence as to the composition of
the randomly generated panels (or, for that matter, how the panels used in
Figure 7 were “enriched for higher-performing markers”). Ans. 21. Thus,
we are not persuaded that Appellant has shown that any improved results
were unexpected. See Pfizer, Inc. v. Apotex, Inc., 480 F.3d 1348, 1371 (Fed.
Cir. 2007) (“[I]n order to properly evaluate whether a superior property was
unexpected, the [fact-finder] should have considered what properties were
expected. Here, Pfizer’s evidence must fail because the record is devoid of
any evidence of what the skilled artisan would have expected.”). In this
regard, we also note that
Appeal 2020-000640
Application 15/414,456

17
[t]he Board has broad discretion as to the weight to give to
declarations offered in the course of prosecution. See Velander v.
Garner, 348 F.3d 1359, 1371 (Fed. Cir. 2003) (‘[A]ccord[ing] little
weight to broad conclusory statements [in expert testimony before the
Board] that it determined were unsupported by corroborating
references [was] within the discretion of the trier of fact to give each
item of evidence such weight as it feels appropriate.’).
In re American Acad. of Science Tech Ctr., 367 F.3d 1359, 1368 (Fed. Cir.
2004) (alterations in original); see also Perreira v. Secretary of the Dept. of
HHS, 33 F.3d 1375, 1377 (Fed. Cir. 1994) (“An expert opinion is no better
than the soundness of the reasons supporting it.”).18
In its Reply Brief, Appellant contends that, with respect to the data in
Figure 7, the Specification describes that “a 187 member original candidate
pool was first selected from the literature,” that “[t]his 187 marker pool was

18 According to Dr. Wilcox, the majority – although not all – of the panels
comprising DPPIV, CEA, and CO9 have AUC values of at least 0.80, while
for the complete randomized panel the most frequent AUCs were 0.77 to
0.79. Wilcox Decl. ¶ 22. While we need not and do not rely on the below in
order to affirm the Examiner’s rejection, we note that Dr. Wilcox provides
no explanation why an alleged improvement of 0.03 to 0.01 in the AUC
would be considered substantial, particularly given that some panels that do
not comprise DPPIV, CEA, and CO9 also appear to have AUCs of 0.80 or
above. Wilcox Decl. Ex. A. Similarly, even though evidence of unexpected
results must be commensurate with the scope of the claims (and claims 1, 3,
4, 6, 7, 18–24, and 27–35 are not limited as to the size of the panel), all of
the panels tested in Figure 7 had the same size – i.e., contained the same
number (10) of classifiers/markers. Although Dr. Wilcox states that “we
understand that the results can be generalized to panels of other sizes” and
“expect that, independent of classifier number, panels comprising DPPIV,
CEA and CO9 will be disproportionately represented among the higher-
performing classes,” Wilcox Decl. ¶ 23, no evidence or explanation was
provided to support this understanding or expectation. As discussed above,
“[a]n expert opinion is no better than the soundness of the reasons
supporting it.” See Perreira, 33 F.3d at 1377.
Appeal 2020-000640
Application 15/414,456

18
then analyzed for efficacy in determining health status to identify a much
smaller pool of 31 biomarkers,” that “[t]his narrowed list of 31 biomarkers
was then further narrowed to 27, and then to a set of 25 biomarkers, from
which random panels were generated that performed ‘much better than those
generated from the unenriched parent 187 marker set.’” Reply Br. 7.
Appellant thus argues that,
contrary to the Office’s suggestion that the claimed panel may not
exhibit superlative performance on grounds that the members of the
25 marker set may be ineffective blood biomarkers, this shortened set
of 25 biomarkers was experimentally validated for the claimed
methodology, having been identified by culling the 187 marker pool
and yielding superior performance.
Id.
We are not persuaded. Although the Specification describes the 25
biomarkers used in Figure 7 as being culled from a pool of 187 markers,
Appellant points to only the most general description of how such culling
was accomplished. “[T]he burden of showing unexpected results rests on he
who asserts them.” In re Klosak, 455 F.2d 1077, 1080 (CCPA 1972).
Moreover, “[i]t is well settled that unexpected results must be established by
factual evidence.” In re De Blauwe, 736 F.2d 699, 705 (Fed. Cir. 1984).
We agree with the Examiner that the cursory description in the Specification
and the Wilcox Declaration regarding Figure 7 does not suffice to show that
the claimed panels exhibits unexpected results as compared to the closest
prior art.19

19 In the Reply Brief, Appellant cites for the first time an article that
Appellant contends to be “[a] corresponding scientific publication” of the
experiment described in Figure 7 of the Specification, in order to provide
additional details regarding the 25 markers used in that figure. Reply Br. 8.
Appeal 2020-000640
Application 15/414,456

19
We are further unpersuaded by Appellant’s unexpected results
arguments because, as the Examiner points out, Meng, Brock, and Thornsen
each teach biomarker or biomarker panels with AUC values that appear to
be higher than the AUC values of the panels in Figure 7. Ans. 24. For
example, Brock reported that the ROC curve of its model relative to the test
data had an AUC of 0.8989. Brock 602, left column. Likewise, Thorsen
noted that “when applying the indicators to identify early-stage CRC a
combination of CEA, TFRC and CA242 resulted in a ROC curve with an
area under the curve of 0.861.” Thorsen Abstract. Finally, Meng reported
that the AUC for serum M2-PK (i.e., PKM) was 0.89 for CRC stages I–III.
Meng 145, right column.
In response, Appellant argues in its Reply Brief that
the Office’s comparison is made across different samples
(e.g., different cancer stages) obtained using different
methods (e.g., different methods of collection, storage,
sample processing, etc) by different research groups (e.g.,
variation in protocols and other factors between groups
that may not even be accounted for) and evaluated using
different experimental techniques (e.g., various types of
mass spectrometry analysis, ELISA, etc), and as such,
any direct comparison using performance metrics such as
AUC is meaningless. Instead, the available data includes

In particular, Appellant cites to the publication to show that the 25 markers
used were blood plasma markers rather than, for instance, tissue markers.
Id. No showing of good cause was made by Appellant to explain why the
article and any accompanying arguments were not presented earlier during
prosecution, thereby denying the Board the benefit of the Examiner’s
response. Accordingly, we decline to consider the article and arguments
related thereto, which are waived. See 37 C.F.R. § 41.41(b)(2); cf. Optivus
Tech., Inc. v. Ion Beam Applications S.A., 469 F.3d 978, 989 (Fed. Cir.
2006).
Appeal 2020-000640
Application 15/414,456

20
Figure 7 and the Wilcox Declaration, dated August 17,
2017, which show that the specific biomarkers recited in
the instant claims outperformed other panels selected
from the 25 blood biomarkers.
Reply Br. 8.
We are not persuaded. Even if comparisons of AUCs disclosed in the
prior art and in Figure 7 are inapposite, Appellant’s attorney arguments do
not suffice to show that the comparative panels of Figure 7 are, in fact, the
closest prior art as compared to, for example, Brock, Thorsen, or Meng.
Baxter, 952 F.2d at 392 (unexpected results must be shown to be unexpected
compared with the closest prior art); Johnston v. IVAC Corp., 885 F.2d
1574, 1581 (Fed. Cir. 1989) (“Attorneys’ argument is no substitute for
evidence.”).
Accordingly, for the reasons above, we affirm the Examiner’s
rejection of claim 1 as obvious over Cordero, Brock, Chen, Thorsen, Meng,
Etzioni, and Mercer. Claims 3, 4, 6, 7, 18–24, and 27–35, which are not
separately argued, fall with claim 1.
Claims 25, 26, 36, and 37
Appellant contends that claims 25, 26, 36, and 37 “recite an upper
limit of no more than 15 or no more than 8 proteins, respectively.” Appeal
Br. 17. Appellant contends that “[t]he cited art does not specifically teach or
suggest the recited upper limit to the number of proteins of any particular
panel.” Id. at 17. Appellant contends that Etzioni “teaches away from the
limits on the number of proteins in a given panel,” because it teaches that
“larger panels are uniformly better at colorectal health assessments.” Id.
We are not persuaded. Claims 25 and 26 depend from claim 1 and
further recite “wherein said list of proteins comprises no more than 15
Appeal 2020-000640
Application 15/414,456

21
proteins” or 8 proteins, respectively. Appeal Br. 31. Claims 36 and 37
depend from claim 27 and similarly recite “wherein said list of proteins
comprises no more than 15 proteins” or 8 proteins, respectively. As the
Examiner points out (Ans. 19), however, these claims recite methods
“comprising” detecting protein levels from a list of proteins. While the
claims recite that the “list” comprises no more than 15 or 8 proteins, the
transition term “comprising” encompasses methods of detecting any number
of additional proteins that are not on the recited list. CIAS, Inc. v. Alliance
Gaming Corp., 504 F.3d 1356, 1360 (Fed. Cir. 2007) (“In the patent claim
context, the term ‘comprising’ is well understood to mean ‘including but not
limited to.’”). Accordingly, under the broadest reasonable interpretation of
the claims, claims 25, 26, 36, and 37 do not “recite an upper limit” of
proteins as Appellant alleges and are invalid as obvious for the same reasons
discussed above with respect to claim 1.20

20 Although we affirm the Examiner’s rejection based on the broadest
reasonable interpretation of the claims, we note that, even assuming that
claims 25/36 and 26/37 in fact recite biomarker panels having no more than
15 or 8 proteins, respectively, the claims would be obvious in light of cited
prior art. Cordero, Brock, Chen, Thorsen, Meng, and Mercer all teach
methods of assessing colorectal health using less than 8 or 15 proteins.
More, the prior art teaches that the number of biomarkers used in a panel is a
result effective variable – for instance, Mercer teaches that measuring
several markers simultaneously increases sensitivity but may, at the same
time, decrease specificity. Mercer 40–41. “[D]iscovery of an optimum
value of a result effective variable in a known process is ordinarily within
the skill of the art.” In re Boesch, 617 F.2d 272, 276 (CCPA 1980).
Appeal 2020-000640
Application 15/414,456

22
CONCLUSION
In summary:
Claims
Rejected
35 U.S.C. § Reference(s)/Ba
sis
Affirmed Reversed
1, 3, 4, 25–29,
32, 35–37
103 Cordero, Brock,
Chen, Thorsen,
Meng, Etzioni,
Mercer
1, 3, 4, 25–
29, 32, 35–
37

1, 3, 4, 6, 7,
18–37
103 Cordero, Brock,
Chen, Thorsen,
Meng, Etzioni,
Mercer, Vokes
1, 3, 4, 6, 7,
18–37

Overall
Outcome
1, 3, 4, 6, 7,
18–37

TIME PERIOD FOR RESPONSE
No time period for taking any subsequent action in connection with
this appeal may be extended under 37 C.F.R. § 1.136(a). See 37 C.F.R.
§ 1.136(a)(1)(iv).

AFFIRMED