Human Pathology (2014) 45, 970–975

www.elsevier.com/locate/humpath

Original contribution

Matched biopsy and resection specimens of gastric

and gastroesophageal adenocarcinoma show high
concordance in HER2 status☆,☆☆
Tao Wang MD a , Eugene T. Hsieh MDCM a,b , Pauline Henry MD, PhD a,c ,
Wedad Hanna MBCh, MD a,b , Catherine J. Streutker MSc, MD a,d , Andrea Grin MD a,d,⁎
a
Department of Laboratory Medicine and Pathobiology, University of Toronto, Toronto, ON, Canada M5S 1A8
b
Department of Anatomic Pathology, Sunnybrook Health Sciences Centre, Toronto, ON, Canada M4N 3M5
c
Department of Pathology, Toronto East General Hospital, Toronto, ON, Canada M4C 3E7
d
Department of Laboratory Medicine, St Michael's Hospital, Toronto, ON, Canada M5B 1W8

Received 10 October 2013; revised 12 December 2013; accepted 18 December 2013

Keywords:
Summary In advanced gastric and gastroesophageal junction (GEJ) adenocarcinomas that overexpress
HER2;
human epidermal growth factor receptor 2 (HER2), treatment with trastuzumab confers a survival
Gastric carcinoma;
benefit. To select patients for treatment, HER2 status is evaluated by immunohistochemistry (IHC) and
Biopsy;
in situ hybridization. Gastric and GEJ adenocarcinomas demonstrate heterogeneity in HER2 expression.
Resection;
Nonetheless, testing is often performed on biopsies alone, which raises the issue of nonrepresentative
Concordance
sampling. We investigated the correlation of HER2 status between matched biopsy and resection
specimens and the role of tumor heterogeneity in contributing to discrepancy. A total of 128 patients
with gastric or GEJ adenocarcinoma had tissue available from a biopsy and subsequent resection. HER2
IHC was performed and evaluated by the criteria used in the Trastuzumab for Gastric Cancer clinical
trial. In situ hybridization was performed if IHC was equivocal (2+) in either the biopsy or resection and
in discrepant cases. Tumor heterogeneity was defined as 3+ or 2+ staining in 10% to 60% of tumor
cells. Overall, HER2 was overexpressed in 18 tumors (14%), with a biopsy-resection concordance of
96.1%. Five cases were discrepant; 2 were positive on biopsy only, and 3 were positive on resection
only. Tumor heterogeneity was seen in 80% of discrepant biopsies and resections, compared with 24%
of concordant cases (P = .016). Our study demonstrates strong concordance between biopsy and
resection specimens for HER2 overexpression in gastric cancer. Discordance was correlated with tumor
heterogeneity. Overall, both biopsy and resection specimens are appropriate for HER2 testing, but
generous sampling for biopsy specimens is necessary to ensure accurate assessment.
© 2014 Elsevier Inc. All rights reserved.

☆
Conflict of interest statement: Drs Hsieh, Hanna, and Streutker have
previously received honorariums from Roche Canada. 1. Introduction
☆☆
Funding disclosures: The authors wish to thank the Physicians'
Sevices Incorporated Foundation and Ventana Medical Systems for funding
Adenocarcinomas of the stomach and gastroesophageal
this research.
⁎ Corresponding author. Department of Laboratory Medicine, St junction (GEJ) are aggressive malignancies that respond
Michael's Hospital, 30 Bond St, Toronto, ON, Canada M5B 1W8. poorly to traditional systemic chemotherapy. These patients
E-mail address: grina@smh.ca (A. Grin). have an average 5-year survival of only 20%, and many

0046-8177/$ – see front matter © 2014 Elsevier Inc. All rights reserved.
http://dx.doi.org/10.1016/j.humpath.2013.12.010
Biopsy-resection concordance of HER2 in gastric cancer 971

present as late-stage tumors beyond resectability [1]. complete, or basolateral positive staining in at least 10% of
Recently, the Trastuzumab for Gastric Cancer trial found cells (3+, positive). In biopsies, a cluster of at least 5 positive
that patients with advanced gastric and GEJ cancers, which tumor cells was required to qualify as 3+.
have human epidermal growth factor receptor 2 (HER2) gene Cases that were scored 2+ by IHC in either the biopsy or
overexpression by immunohistochemistry (IHC) or equivo- resection were further characterized by ISH. Brightfield
cal IHC plus amplification by in situ hybridization (ISH), had dual-color ISH (Dual ISH; Ventana; St Michael's Hospital)
a significant gain in overall survival when treated with or silver ISH (SISH; Ventana; Sunnybrook Health Sciences
trastuzumab [2]. Centre) was performed as per manufacturer's instructions
Currently, HER2 testing is often done on biopsy with appropriate controls [6]. ISH results were scored by a
specimens, which may not be representative of the tumor gastrointestinal pathologist as follows: the entire slide was
as a whole. Gastric and GEJ tumors have been noted to show scanned for areas of amplification, using the HER2 IHC slide
significant tumor heterogeneity for HER2 expression [3-5]. as a guide, and the HER2 and chromosome 17 signals
Expression heterogeneity may result in inaccurate results, (CEP17) were counted in 20 tumor nuclei. If the ratio of
particularly if testing is performed on biopsy alone. The HER2 to CEP17 fell between 1.8 and 2.2, a further 20
resultant false negatives could lead to undertreatment nuclei were counted. As per the criteria used in the
of patients. Trastuzumab for Gastric Cancer trial, HER2 was considered
Despite the potential significance of false-negative re- amplified if the HER2/CEP17 ratio was at least 2.0 and
sults, very few studies have examined the correlation of nonamplified if less than 2.0 [2,3,7].
HER2 status between biopsy and corresponding resection Well-established criteria used in clinical practice were
specimens [4-6]. Better understanding of the frequency of used to define HER2 positivity. An IHC score of 3+, or 2+
discrepancies and characteristics which may lead to combined with positive ISH, were considered positive for
discrepancy can impact clinical decision making. HER2 overexpression and/or amplification, whereas all other
scenarios were negative. Discordance was defined as a
scenario where the overall interpretation (using combined
IHC and ISH results) is negative in the biopsy but positive in
2. Materials and methods the resection specimen from the same patient, or vice versa.
Percent agreement for HER2 status in biopsy versus
Approval for the study was obtained from the Institutional resection specimens was calculated.
Research Ethics Boards from St Michael's Hospital and Tumor IHC expression heterogeneity for HER2 was
Sunnybrook Health Sciences Centre. The pathology data- assessed by estimating the percentage of tumor cells that
bases from these 2 institutions were searched for biopsies of were 2+ or 3+ by IHC. Significant heterogeneity was defined
gastric or gastroesophageal adenocarcinoma with matching as between 10% and 60% (inclusive) of tumor cells showing
resections between 2000 and 2012. The original hematoxylin 2+ or 3+ score by IHC. This definition aims to capture cases
and eosin (H&E) slides were reviewed by gastrointestinal that are equivocal or positive for HER2 using well-
pathologists (A.G., E.H.) and classified by differentiation established criteria but include significant areas of negative
(well, moderate, poor) and Lauren subtype (intestinal, IHC expression and thus would be prone to nonrepresenta-
diffuse, mixed). Differentiation was defined by the degree tive sampling.
of glandular formation (N95% for well, 50%-95% for
moderate, b50% for poor). Tumors with any areas of diffuse
or signet ring morphology combined with intestinal/
glandular morphology were classified as mixed. For biopsy 3. Results
specimens, the total number of biopsy fragments and the
number of fragments containing tumor were recorded. A total of 128 patients were found with paired biopsy and
Patient age and sex were also noted. resection specimens. The majority were of intestinal subtype
One representative tissue block was stained for HER2 (4B5 (n = 78; 61%). Tumor characteristics are described in
clone; Ventana, Tuscon, AZ, USA) by IHC as per manufac- Table 1. Eighteen tumors (14%) were classified as HER2
turer's instructions. In cases with mixed morphology, a block positive, including 1 diffuse adenocarcinoma and 1 mixed
that included different morphologies was selected. HER2 IHC carcinoma; the remaining 16 positive cases were of intestinal
was scored by a gastrointestinal pathologist using well- subtype. There were 17 (13%) tumors from the GEJ.
established gastric HER2 criteria [2,3,7]. Specifically, scores The overall biopsy-resection concordance was 96.1%.
were assigned as follows: no positive staining or staining of Testing the biopsy alone would have resulted in the correct
only a part of cell membrane in less than 10% of cells (0, diagnosis in 97.7% of patients. Testing the resection alone
negative); barely visible staining of only a part of cell would have been correct in 98.4% of cases. Overall IHC
membrane in at least 10% of cells (1+, negative); a weak-to- staining categories are listed in Table 2. The staining
moderate, complete, or basolateral positive staining in at least characteristics of concordant positive cases are listed in
10% of cells (2+, equivocal); and a moderate-to-strong, Table 3.
972 T. Wang et al.

Table 1 Tumor characteristics to provide additional therapeutic benefit in addition to

Parameter n (%) traditional chemotherapy in patients with advanced gastric
and GEJ adenocarcinoma.
Mean age (y; range) 64 (31-87) HER2 testing in gastric cancer is often only performed
Male, n (%) 75 (59)
once and frequently only on biopsy specimens. Thus, tumor
Differentiation, n (%)
heterogeneity can lead to false negatives that could deprive
Well 13 (10)
Moderate 54 (42) patients of trastuzumab therapy. This issue has been
Poor 61 (48) explored in breast cancer, with some authors advocating
Subtype, n (%) testing of both core biopsies and resection specimens due to
Intestinal 78 (61) the possibility of tumor heterogeneity [8-11]. This may not
Diffuse 25 (20) be possible in many gastric cancer cases that are not
Mixed 25 (20) surgically resectable.
Location, n (%) The most commonly endorsed definition of heterogeneity
GEJ 17 (13) in breast cancer is based on ISH. When between 5% and 50%
Stomach 111 (87) of tumor cells exhibit an HER2 to chromosome 17 ratio of
greater than 2.2, the tumor is said to display genetic
heterogeneity [12]. One problem with this definition is that
there is great overlap between cases with genetic heteroge-
Five cases were discordant (Table 4); 2 were positive on
neity and those with ratios in the equivocal range. It also does
biopsy only, and 3 were positive on resection only. Overall,
not account for regional heterogeneity, where one region
HER2 IHC heterogeneity was seen in 33 (26%) of 128 cases.
may be markedly overexpressing HER2 while another has
Heterogeneity was seen in either the biopsy or resection in 4
minimal expression. Thus, the value of separately reporting
(80%) of 5 discordant cases, one of which was mixed
genetic heterogeneity is questionable and inconsistently
subtype. Concordant cases displayed significantly less
adopted [11]. Others studies using tissue microarrays have
heterogeneity compared with the discordant ones, with a
defined heterogeneity as any clinically significant changes in
rate of 29 in 123 (24% versus 80%, P = .016).
hormone receptor or HER2 status among tissue microarray
There was no significant difference between the average
cores from the same tumor [8,9]. No definition of
number of biopsy fragments in discordant versus concordant
heterogeneity currently exists for gastric cancer.
cases (5.6 versus 5.8; P = .83), as well as the average number
Our study of gastric and GEJ adenocarcinoma demon-
of biopsy fragments containing tumor (3.6 versus 3.9; P =
strated strong concordance (96.1%) between biopsy and
.76). However, in one discrepant case with an HER2-
resection specimens for HER2 status. Interestingly, biopsy
negative biopsy, only 1 of 7 fragments contained tumor.
was not inferior to resection in our sample.
Most concordant positive cases had areas of 3+
expression in more than 50% of the tumor. There were 3
4. Discussion cases (K, L, M) that demonstrated borderline HER2
positivity (Table 3). Case K met the biopsy criteria for
HER2 is a member of the epidermal growth factor IHC positivity and was ISH amplified; however, the follow-
receptor family, whose overexpression has been associated up endoscopic mucosal resection (EMR) was borderline IHC
with numerous cancers. The gene that encodes HER2 is on positive (10% of tumor 3+) and ISH negative with polysomy
chromosome 17, and gene amplification has been found to be resulting in an overall ratio of only 0.9. Although case K was
the most common mechanism for overexpression [1]. classified overall as concordant, the discrepancy in ISH
Trastuzumab, a monoclonal antibody against HER2, has results may be due to sampling because it is uncertain
been used in the treatment of breast and more recently gastric whether the area biopsied was included in the EMR.
cancer with HER2 overexpression [1]. This has been shown Polysomy of chromosome 17 was also observed, and this
may account for some cancers with 3+ IHC expression and
negative ISH amplification. The significance of this finding
Table 2 HER2 staining patterns is presently unclear. The second borderline case, case L, had
IHC ISH No. a focal cluster of 3+ cells on biopsy sufficient for positivity
Biopsy Resection but was not ISH amplified (1.18); the resection showed
mostly 2+ staining (50%) and had insufficient 3+ (5%) to
0 66 71
meet current resection specimen criteria (10%), but was
1+ 28 22
found to be low amplified (2.3). The biopsy in this borderline
2+ Nonamplified 19 19
2+ Amplified 2 1 case exemplifies a cause of IHC-ISH discrepancy. With
3+ 13 15 respect to biopsies, only a small cluster of 5 or more cohesive
Total 128 128 3+ cells is required for positivity. However, because ISH
requires counting at least 20 nonoverlapping cells, cases
Biopsy-resection concordance of HER2 in gastric cancer 973

Table 3 Characteristics of concordant cases

Case Type No. of No. with Lauren Grade a IHC (%) b ISH
fragments tumor subtype 3+ 2+ 1+ or 0
A Biopsy 3 1 Intestinal 2 100 – –
Resection 100 – –
B Biopsy 6 3 Intestinal 2 100 – –
Resection 100 – –
C EMR NA Intestinal 1 70 15 15
Resection 100 – –
D Biopsy 5 5 Intestinal 3 50 50 –
Resection 2 80 20 –
E Biopsy 10 6 Intestinal 2 80 10 10
Resection 60 30 10
F Biopsy 9 8 Intestinal 2 90 10 –
Resection 90 10 –
G Biopsy 5 5 Intestinal 2 90 10 –
Resection 90 10 –
H Biopsy 9 5 Intestinal 2 50 20 30
Resection 70 20 10
I Biopsy 4 4 Intestinal 2 100 – –
Resection 100 – –
J Biopsy 4 4 Intestinal 3 – 15 85 4.8
Resection – 30 70 10
K Biopsy 3 3 Intestinal 2 5 30 65 8.9
Resection 10 15 75 0.9
L Biopsy 7 2 Diffuse 3 5 95 – 1.18
Resection 5 50 45 2.3
M Biopsy 6 4 Intestinal 3 – 50 50 2.18
Resection 25 25 50 1.3
a
Grade: 1, well differentiated; 2, moderately differentiated; 3, poorly differentiated.
b
Percentage of tumor showing each IHC score.

with only a small cluster of 3+ cells may be nonamplified borderline positivity. All 3 were captured within our
overall. The third borderline case, case M, had equivocal definition of IHC expression heterogeneity. These cases
IHC on biopsy with low-level amplification (2.18); it was represented situations where discordance could easily
clearly 3+ on resection but was not amplified (1.3), despite occur; thus, representative and adequate sampling is
repeat ISH and verification by a second observer. essential for proper interpretation.
Regardless, the aforementioned cases all met the criteria Discordant cases are shown in Table 4. Surprisingly, these
for concordant positivity but demonstrate scenarios of include 2 cases that were classified as HER2 positive on

Table 4 Characteristics of discordant cases

Case Type No. of No. with Lauren Grade a IHC (%) b ISH HER2
fragments tumor subtype 3+ 2+ 1+ 0 status

1 Biopsy 8 6 Mixed 3 20 80 13 +
Resection 10 90 1.85 −
2 Biopsy 3 2 Intestinal 2 10 90 2.93 +
Resection 10 90 1.58 −
3 Biopsy 7 1 Intestinal 1 100 NA −
Resection 60 40 7.6 +
4 Biopsy 5 4 Intestinal 3 100 0.88 −
Resection 2 50 50 4.2 +
5 Biopsy 5 5 Intestinal 1 60 40 1.14 −
Resection 10 50 40 1.67 +
a
Grade: 1, well differentiated; 2, moderately differentiated; 3, poorly differentiated.
b
Percentage of tumor showing IHC score.
974 T. Wang et al.

biopsy but were shown to be negative in the resection. A tumor was present in 4 of 5 biopsy fragments, but tumor
possible cause for these discrepancies could be poor fixation sampled in the biopsy showed only 1 type of morphology
or increased ischemic time in the resection specimens. that was observed in the resection specimen. Tumor in the
Unfortunately, owing to the retrospective nature of the study, biopsy was solid or poorly differentiated intestinal mor-
these data were not available. An alternative cause also phology, which was also seen in the resection specimen, but
includes tissue sampling. Although the standard grossing the resection also included areas of more well-differentiated
procedure at our institutions requires histologic examination tumor that was strongly positive for HER2 (Fig.).
of one tissue block per centimeter of tumor, areas of HER2 Discrepant case 5 showed similar morphology in the biopsy
positivity sampled in the biopsy may not have been and the resection, but only a small area of tumor was
represented in tissue sections. This may be particularly true positive for HER2 (IHC 3+) in the resection. This area
for case 1, which was of mixed subtype on biopsy, but comprised 10% of the tumor sampled, reaching minimum
sections taken from the resection specimen sampled tumor of diagnostic criteria; therefore, the discrepancy in this case
diffuse type only. Owing to the retrospective nature of the may be due to the inherent borderline overexpression of this
study, further sampling was not possible. specific tumor.
Cases 3 to 5 had HER2-negative biopsies with HER2- There was no significant relationship between the number
positive resections (Table 4). In case 3, only 1 of 7 biopsy of biopsy fragments and concordance. Nonetheless, owing to
fragments contained tumor. The discrepancy in this case is regional HER2 expression heterogeneity, the logical ap-
likely due to tumor sampling, and in these cases, further proach would be to take multiple biopsies from different
sampling of the tumor may be recommended. In case 4, areas of the tumor. Rüschoff et al [7] recommended that,

Fig Discordant pair (case 4) where biopsy (A) is negative for HER2 by IHC (B) and ISH (C). Black ISH signals are HER2; red signals are
centromere of chromosome 17. The matching resection (D) is 3+ on IHC (E) and ISH amplified (F). The resection showed variable grade with
poorly differentiated areas, similar to the biopsy.
Biopsy-resection concordance of HER2 in gastric cancer 975

ideally, at least 6 to 8 fragments of tumor should be present in Importantly, our study demonstrates that both biopsy and
biopsies for adequate assessment. resection specimens are appropriate for HER2 testing. The
There is currently no standard definition for HER2 degree of concordance is quite high, and biopsies are not
expression heterogeneity in gastric cancer. In our study, we inferior to resection specimens despite high rates of
defined a case as being heterogenous if 10% to 60% heterogeneity reported in gastric and GEJ tumors. In our
(inclusive) of the tumor showed 2+ and/or 3+ staining by cohort, tumors with heterogeneous staining had a higher
IHC. The lower limit of 10% represents the lowest incidence of discordance, which suggests that some of these
proportion of 2+/3+ cells to warrant classification as patients may benefit from additional testing; however,
equivocal or positive on resection. The upper limit of 60% validation of this finding as well as our definition of
was based on the interim observation that most positive heterogeneity is required prior to clinical application.
cases either were diffusely positive (N90%) or had focal
areas of equivocal or positive staining that usually totalled
less than 60%.
References
Our definition led to 26% of all cases being classified as
[1] Jorgensen JT. Targeted HER2 treatment in advanced gastric cancer.
heterogeneous, but despite this, only 5 cases (4%) were Oncology 2010;78:26-33.
discordant. Nonetheless, heterogeneity was observed signif- [2] Bang YJ, Van Cutsem E, Feyereislova A, et al. Trastuzumab in
icantly more often in discordant cases compared with combination with chemotherapy versus chemotherapy alone for
concordant ones. Based on our findings, cases that meet treatment of HER2-positive advanced gastric or gastro-oesophageal
our definition of heterogeneity are at higher risk for junction cancer (ToGA): a phase 3, open-label, randomised controlled
trial. Lancet 2010;376:687-97.
discordance. This suggests that patients with tumors that [3] Hofmann M, Stoss O, Shi D, et al. Assessment of a HER2 scoring
are heterogenous 2+ by IHC but nonamplified may benefit system for gastric cancer: results from a validation study. Histopathol
from additional sampling and testing, but this requires further 2008;52:797-805.
study prior to clinical application. Validation of our [4] Yang J, Luo H, Li Y, et al. Intratumoral heterogeneity determines
discordant results of diagnostic tests for human epidermal growth
definition of heterogeneity in a larger cohort is required to
factor receptor (HER) 2 in gastric cancer specimens. Cell Biochem
determine if our definition of heterogeneity should prompt Biophys 2012;62:221-8.
retesting on subsequent samples. [5] Lee S, Bastiaan de Boer W, Fermoyle S, et al. Human epidermal
There are relatively few studies performed thus far growth factor receptor 2 testing in gastric carcinoma: issues related to
examining biopsy-resection concordance in gastric cancer. heterogeneity in biopsies and resections. Histopathol 2011;59:832-40.
One previous study with 54 matched pairs found IHC [6] Yan B, Yau EX, Choo SN, et al. Dual-color HER2/chromosome 17
chromogenic in situ hybridisation assay enables accurate assessment of
concordance rates of 74%, but they did not provide HER2 genomic status in gastric cancer and has potential utility in
concordance rates for overall interpretations that account HER2 testing of biopsy samples. J Clin Pathol 2011;64:880-3.
for both IHC and ISH, as is done in clinical practice [5]. It [7] Rüschoff J, Hanna W, Bilous M, et al. HER2 testing in gastric cancer: a
was also unclear as to exactly how many of the paired practical approach. Mod Pathol 2012;25:637-50.
specimens had ISH performed. Another study found an [8] Greer LT, Rosman M, Mylander WC, et al. Does breast tumor
heterogeneity necessitate further immunohistochemical staining on
overall HER2 status concordance rate of 92% among 61 surgical specimens? J Am Coll Surg 2013;216:239-51.
pairs, which is closer to the findings in our study [13]. [9] Seol H, Lee HJ, Choi Y, et al. Intratumoral heterogeneity of HER2
There are limitations to consider in our analysis. The low gene amplification in breast cancer: its clinicopathological signifi-
number of discrepant cases, although indicative of good cance. Mod Pathol 2012;25:938-48.
[10] Yang YL, Fan Y, Lang RG, et al. Genetic heterogeneity of HER2 in
concordance, conversely makes it difficult to perform case
breast cancer: impact on HER2 testing and its clinicopathologic
comparisons. This is a limitation of sample size. Nonethe- significance. Breast Cancer Res Treat 2012;134:1095-102.
less, to our knowledge, this is the largest study to date to [11] Chang MC, Malowany JL, Mazurkiewicz J, et al. ‘Genetic
specifically address the issue of concordance. We also heterogeneity’ in HER2/neu testing by fluorescence in situ hybridi-
acknowledge that different methodologies were applied for zation study of 2,522 cases. Mod Pathol 2012;25:683-8.
ISH at the 2 contributing institutions. However, both [12] Vance GH, Barry TS, Bloom KJ, et al. Genetic heterogeneity in HER2
testing in breast cancer: panel summary and guidelines. Arch Pathol
methodologies have been validated against fluorescence Lab Med 2009;133:611-2.
ISH, and both centers are government-designated HER2 [13] Pirrelli M, Caruso ML, Di Maggio M, et al. Are biopsy specimens
testing sites for the region. All paired specimens were tested predictive of HER2 status in gastric cancer patients? Dig Dis Sci
at the same institution with the same protocols. 2013;58:397-404.