Comparison of 5 Immunohistochemical Markers
of Hepatocellular Differentiation for the Diagnosis
of Hepatocellular Carcinoma
Thuy Nguyen, MD; Daniel Phillips, MD; Dhanpat Jain, MD; Michael Torbenson, MD; Tsung-Teh Wu, MD, PhD;
Matthew M. Yeh, MD, PhD; Sanjay Kakar, MD
 Context.—Several immunohistochemical markers are
available to establish the diagnosis of hepatocellular
carcinoma. Judicious selection is essential to achieve a
reliable diagnosis in limited tissue provided by liver biopsy.
Objective.—To compare the efficacy of 5 hepatocellular
markers for the diagnosis of hepatocellular carcinoma
across various levels of differentiations.
Design.—Immunohistochemistry for hepatocyte paraffin
antigen 1 (Hep Par 1), polyclonal carcinoembryonic
antigen (CEA), glypican-3, arginase-1, and bile salt export
pump transporter was performed in 79 hepatocellular
carcinomas, yielding 93 observations (13 well-differenti-
ated [14%], 41 moderately differentiated [44%], and 39
poorly differentiated [42%] tumors).
Results.—Arginase-1 and Hep Par 1 had the highest
sensitivity for well-differentiated hepatocellular carcino-
ma, whereas arginase-1 and glypican-3 had the highest
sensitivity for poorly differentiated hepatocellular carci-
noma. When staining of more than 50% of the tumor was
considered a positive result, arginase-1 remained the most
sensitive marker for all differentiations, whereas sensitivity
T he diagnosis of hepatocellular carcinoma (HCC) and its
differentiation from metastatic or nonhepatocellular
tumors has important therapeutic implications. This dis-
tinction can be challenging if the tumor is poorly
differentiated and when limited amounts of tumor are
available from needle biopsies. Many immunohistochemical
markers are available for establishing the diagnosis, each
with its strength and limitations. Hepatocyte paraffin
Accepted for publication November 26, 2014.
From the Department of Anatomic Pathology, University of
California, San Francisco, (Drs Nguyen, Phillips, and Kakar); the
Department of Pathology, Yale University School of Medicine, New
Haven, Connecticut (Dr Jain); the Department of Pathology, Johns
Hopkins, Baltimore, Maryland (Dr Torbenson); the Division of
Anatomic Pathology, Mayo Clinic, Rochester, Minnesota (Dr Wu);
the Department of Pathology, University of Washington, Seattle, (Dr
Yeh); and the Department of Anatomic Pathology, Veteran Affairs
Medical Center, San Francisco (Dr Kakar).
The authors have no relevant financial interest in the products or
companies described in this article.
Reprints: Sanjay Kakar, MD, Department of Anatomic Pathology,
University of California, San Francisco, 4150 Clement St, San
Francisco, CA 94121 (e-mail: sanjay.kakar@ucsf.edu).
1028 Arch Pathol Lab Med—Vol 139, August 2015
for Hep Par 1 in poorly differentiated hepatocellular
carcinoma dropped to 30% and that of glypican-3 in
well-differentiated hepatocellular carcinoma was 15%.
The addition of Hep Par 1 and/or polyclonal CEA to
arginase-1 did not lead to an increase in sensitivity for any
differentiation. The combined use of arginase-1 and
glypican-3 yielded 100% sensitivity for poorly differenti-
ated hepatocellular carcinoma.
Conclusion.—Arginase-1 was the most sensitive marker
in all differentiations of hepatocellular carcinoma. Glypi-
can-3 had high sensitivity for poorly differentiated cases
and its combined use with arginase-1 enabled identifica-
tion of nearly all cases of poorly differentiated hepatocel-
lular carcinoma. Although bile salt export pump
transporter has good overall sensitivity, it has a limited
role in establishing hepatocellular differentiation when
added to a panel of arginase-1 with either glypican-3 or
Hep Par 1.
(Arch Pathol Lab Med. 2015;139:1028–1034; doi:
10.5858/arpa.2014-0479-OA)
antigen 1 (Hep Par 1), which recognizes the urea cycle
enzyme carbamoyl phosphate synthetase,1 and polyclonal
antibody directed against carcinoembryonic antigen (p-
CEA) are among the most commonly used hepatocellular
markers, but they have less than 50% sensitivity for poorly
differentiated HCC.2–5 In addition, Hep Par 1 can be positive
in gastric, lung, and esophageal adenocarcinomas3 and, less
commonly, in a wide variety of tumors.3,4 The canalicular
pattern observed with p-CEA is considered to be specific for
hepatocellular differentiation, but can be difficult to
distinguish from membranous or luminal staining patterns
seen in adenocarcinomas.4,6 Other markers, such as CD10
and villin, also yield a canalicular pattern of staining, but
suffer from the same limitations.6,7 Serum levels of a-
fetoprotein are routinely used in screening for HCC;
however, immunohistochemical staining for a-fetoprotein
has 30% to 50% sensitivity and often shows background
staining.3,6,7
In the past few years, glypican-3 (GPC-3) and arginase-1
(Arg-1) have emerged as promising markers of hepatocel-
lular differentiation. The GPC-3 marker is a membrane-
anchored, heparin-sulfate proteoglycan normally expressed
in fetal liver and placenta. Its advantages include negative
Comparison of 5 Immunohistochemical Markers—Nguyen et al
 Table 1. Antibodies Used for Immunohistochemistry in the Study
Antibody Clone
Hep Par 1 OCH1E5
Polyclonal CEA Polyclonal
Glypican-3 1G12
Arginase-1 Rabbit polyclonal
BSEP Mouse monoclonal F-6
Abbreviations: BSEP, bile salt export pump; CEA, carcinoembryonic antigen; Hep Par-1, hepatocyte paraffin antigen 1.
results in nonneoplastic liver and high sensitivity in poorly
differentiated HCC.8–11 Arginase-1 is a binuclear manganese
metalloenzyme that catalyzes the hydrolysis of arginine to
ornithine and urea. Initial studies have shown that Arg-1 is
a highly sensitive and specific marker for hepatocellular
differentiation.12–14
Bile salt export pump (BSEP) is a liver-specific ATP-
binding cassette transporter encoded by the ABCB11 gene,
which is expressed exclusively in the liver canalicular
membrane and involved with bile acid transport.15 The
BSEP marker has been shown to be specific and sensitive for
hepatocellular differentiation, but its diagnostic utility in
HCC has not been explored in detail.16
The choice of hepatocellular markers used for diagnosis of
HCC in liver biopsies varies widely. The choice of the
markers requires careful evaluation of the clinical setting
and pathologic features, generation of an appropriate
differential diagnosis, and framing specific questions to be
answered by application of immunohistochemical markers.
Indiscriminate or inappropriate use of markers can
exhaust the tissue in the block, precluding a definite
diagnosis. This study attempted to identify an optimum
combination of hepatocellular markers by comparing the
sensitivity of 5 immunohistochemical markers (Hep Par 1,
p-CEA, GPC-3, Arg-1, and BSEP) in well-differentiated,
moderately differentiated, and poorly differentiated HCC.
MATERIALS AND METHODS
The study population comprised 79 cases of HCC from the
surgical pathology files of the University of California, San
Francisco, Medical Center, the University of Washington Medical
Center (Seattle), the Yale-New Haven Hospital (New Haven,
Connecticut), the Mayo Clinic (Rochester, Minnesota), and the
Johns Hopkins School of Medicine (Baltimore, Maryland). The
slides were reviewed to confirm the diagnosis in all cases, and the
level of differentiation was assessed based on the 2010 World
Health Organization criteria.17 Well-differentiated HCC showed
obvious hepatocellular differentiation with mild cytologic atypia,
mildly increased nuclear to cytoplasmic ratios, thin trabeculae (3
cells thick) in most of the tumor, and a frequent pseudoglandular
pattern. Moderately differentiated HCC also showed morphologic
resemblance to hepatocytes with moderate cytologic atypia and/or
thick cell plates (.3 cells thick in most of the tumor). Poorly
differentiated HCC showed one or more of the following: little or no
resemblance to hepatocytes, marked cytologic atypia, a solid
growth pattern, or markedly increased nuclear to cytoplasmic
ratios.
Formalin-fixed, paraffin-embedded tissue was used to perform
immunohistochemistry for Hep Par 1, p-CEA, GPC-3, Arg-1, and
BSEP (Table 1). The incubation time was 15 minutes for the
primary antibody. For Hep Par 1 and pCEA, heat-induced, epitope-
retrieval pretreatment was performed in a Dako (Carpinteria,
California) Pascal pressure cooker for 30 to 45 minutes at 1258C
until reaching pressure of 22 to 25 psi in 0.01M citrate buffer (pH
6.0). The slides were then cooled to room temperature, rinsed, and
quenched with 3% hydrogen peroxide for 10 minutes. This was
followed by incubation with the primary antibody at room
Arch Pathol Lab Med—Vol 139, August 2015
Vendor Dilution
Dako, Carpinteria, California 1:200
Dako, Carpinteria, California 1:20
BioMosaics, Burlington, Vermont 1:2
Sigma-Aldrich, St Louis, Missouri 1:400
Santa Cruz Biotechnology, Dallas, Texas 1:200
temperature (Hep Par 1, 30 minutes; pCEA, 15 minutes). For
detection, a biotin-free, horseradish peroxidase–labeled, dextrose-
based polymer complex bound to a secondary antibody (K4061,
Dako EnVision Plus dual link system-HRP) was used with an
incubation time of 15 minutes. Slides were then developed for 10
minutes with 3-3 0 -diaminobenzidine chromogen (DABþ, K3468,
Dako), and counterstained with Gill hematoxylin.
For GPC-3 and Arg-1 immunohistochemistry, conventional
tissue sections were subjected to heat-induced, epitope-retrieval
pretreatment for 20 minutes at 1008C in epitope retrieval solution 1
(pH 6.0; AR9961, Leica Microsystems, Newcastle, United King-
dom), and cooled to room temperature. The Leica Bond Polymer
Refine detection kit (DS9800, Leica Microsystems) was used for the
blocking and detection steps. The slides were incubated with the
primary antibody for 15 minutes, incubated in postprimary reagent
(rabbit anti-mouse immunoglobulin G [IgG]) for 8 minutes,
followed by incubation in polymer reagent (anti-rabbit poly-
HRP-IgG) for 8 minutes. For the blocking step, tissue sections were
quenched with peroxide block (3%–4% hydrogen peroxide) for 10
minutes. Slides were then developed for 10 minutes with DAB
chromogen, and counterstained with hematoxylin for 1 minute.
Immunohistochemistry for BSEP was performed using a
monoclonal mouse anti-human antibody BSEP clone F-6 (Santa
Cruz Biotechnology, Santa Cruz, California) at 1:200 dilution. The
entire process was performed on a Leica BOND III platform, with
antigen retrieval using heat-induced epitope retrieval 2. The
antibody incubation time was 15 minutes. Slides were then
developed for 5 minutes with DAB chromogen.
The staining intensity was recorded as 0 to 3 (negative, mild,
moderate, and strong), and the extent of staining was recorded into
3 groups based on percentage of tumor showing positive staining
(,5%, 5%–50%, and .50%). Cytoplasmic staining was considered
positive for Hep Par 1 and GPC-3, whereas nuclear and/or
cytoplasmic staining was considered positive for Arg-1. For pCEA
and BSEP, a canalicular staining pattern was considered positive
(Figure 1, A and B); membranous or luminal positivity without
canalicular staining was considered negative.
For data analysis, cases were considered positive if moderate or
strong staining was seen in 5% or more of the tumor cells. All other
situations were considered negative. The positive cases were
divided into focal positive if moderate or strong staining was seen
in 5% to 50% of the tumor and diffuse positive if moderate or strong
staining was seen in more than 50% of the tumor cells. The 50%
cutoff is not necessarily relevant from a diagnostic standpoint, but
was chosen to compare the extent of staining with different
markers. The sensitivity for each marker as well as combination of
markers for well differentiated, moderately differentiated, and
poorly differentiated HCC was determined.
RESULTS
Clinical Data
Seventy-nine cases were examined. The age range was 31
to 83 years (62 men [78%], 17 women [22%]). Cirrhosis was
present in 58 (73%) of the cases, whereas 17 HCC cases
(22%) arose in noncirrhotic liver (status not known in 4
cases [5%]). Of the 58 cases with cirrhosis, the etiology of
the cirrhosis was known in 50 cases (86%): hepatitis C in 35
cases (60%), hepatitis B in 7 cases (12%), hepatitis B and C
Comparison of 5 Immunohistochemical Markers—Nguyen et al 1029

Figure 1. A, Well-differentiated hepatocellular carcinoma. B, Canalic-
ular pattern of staining with bile salt export pump (hematoxylin-eosin,
original magnification 3200 [A]; original magnification 3200 [B]).
coinfection in 2 cases (3%), alcohol use in 4 cases (7%),
hepatitis C and alcohol use in 1 case (2%), and steatohep-
atitis in 1 case (2%).
Immunohistochemistry
Of the 79 HCC cases, 14 (18%) showed more than one
differentiation pattern; these areas were scored separately
yielding 93 observations (13 well differentiated cases [14%],
41 moderately differentiated cases [44%], and 39 poorly
differentiated cases [42%]). The results for all antibodies
were categorized into focal and diffuse positive cases (Tables
2 and 3, Figures 2, A through E; 3, A through D; and 4, A
through D).
The Hep Par 1 staining was positive in all well-
differentiated cases (13 of 13; 100%) and most moderately
differentiated cases (40 of 41; 98%) (Figures 2, A and C; and
Table 2. Immunohistochemical Markers in Hepatocellular Carcinoma With 5%a Staining Considered Positive
Differentiation Arginase-1, No. (%) Glypican-3, No. (%)
Well, n ¼ 13 13 (100) 8 (62)
Moderately, n ¼ 41 41 (100) 33 (80)
Poorly, n ¼ 39 38 (97) 33 (85)
Abbreviations: BSEP, bile salt export pump; Hep Par-1, hepatocyte paraffin antigen 1; pCEA, polyclonal carcinoembryonic antigen.
a
Staining in 5% or more of the tumor cells was considered positive.
b
BSEP results were not available in every case; n ¼ 39.
c
BSEP results were not available in every case; n ¼ 31.
1030 Arch Pathol Lab Med—Vol 139, August 2015
3, A and D), whereas sensitivity in poorly differentiated
cases was modest (25 of 39; 64%). Similarly, pCEA showed
canalicular positivity in most well-differentiated cases (12 of
13; 92%) (Figure 2, E) and most moderately differentiated
cases (36 of 41; 88%), but the sensitivity was lower in poorly
differentiated HCC (21 of 39; 54%). For both Hep Par 1 and
pCEA, the numbers were slightly lower for well and
moderately differentiated areas when only diffuse positive
staining was considered, but fell to below one-third of cases
for poorly differentiated areas (Table 3).
The Arg-1 staining was positive in all well-differentiated
cases (13 of 13; 100%), all moderately differentiated cases
(41 of 41; 100%), and most poorly differentiated cases (38 of
39; 97%) (Figures 2, A and B; 3, A and B; and 4, A and B). All
positive cases demonstrated cytoplasmic staining, with
nuclear positivity also seen in most instances (65 of 79;
82%). The results were similar when only diffusely positive
cases were considered (Table 3). The GPC-3 staining had a
modest sensitivity in well-differentiated cases (8 of 13; 62%)
and a high sensitivity for moderately differentiated cases (33
of 41; 80%) and poorly differentiated HCC (33 of 39; 85%)
(Figures 2, E; 3, C; and 4, C). When cases with diffuse
staining were considered positive, the results were similar
for poorly differentiated HCC, but lower for well and
moderately differentiated HCC.
Although BSEP results were not available in every case, it
yielded 83 observations. It showed canalicular positivity in
most well-differentiated cases (12 of 13; 92%) and
moderately differentiated cases (37 of 39; 95%) (Figure 1,
B), whereas sensitivity was low for poorly differentiated
HCC cases (14 of 31; 45%) (Figure 4, D). The numbers were
lower for all 3 categories when only cases with diffuse
staining were considered positive. Membrane staining with
BSEP was seen in 8 of 83 HCCs (10%), which partially
obscured the canalicular staining pattern.
When a combination of GPC-3 and Arg-1 was evaluated,
at least one marker was positive in all poorly differentiated
HCCs (Figures 3 and 4). Even when only diffusely positive
cases were considered, GPC-3 and/or Arg-1 were positive in
37 of 39 cases (95%) of poorly differentiated HCC (Table 4).
The Hep Par 1 marker performed better than p-CEA and
BSEP in well-differentiated and moderately differentiated
cases. However, adding Hep Par 1 to Arg-1 did not improve
the sensitivity because Arg-1 was positive in all cases.
Hence, with the exception of GPC, addition of other
markers to Arg-1 did not improve sensitivity at any
differentiation.
COMMENT
Many tumors can closely mimic HCC, including poorly
differentiated cholangiocarcinoma; metastatic, poorly differ-
entiated adenocarcinomas; nonhepatocellular tumors, such
as epithelioid angiomyolipoma; and other metastatic tu-
Hep Par-1, No. (%) pCEA, No. (%) BSEP, No. (%)
13 (100) 12 (92) 12 (92)
40 (98) 36 (88) 37 (95)b
25 (64) 21 (54) 14 (45)c
Comparison of 5 Immunohistochemical Markers—Nguyen et al
 Table 3. Immunohistochemical Markers in Hepatocellular Carcinoma With 50%a Staining Considered Positive
Differentiation Arginase-1, No. (%) Glypican-3, No. (%) Hep Par-1, No. (%) pCEA, No. (%) BSEP, No. (%)
Well, n ¼ 13 13 (100) 2 (15) 13 (100) 10 (77) 9 (69)
Moderately, n ¼ 41 40 (98) 24 (58) 34 (83) 26 (63) 28 (72)b
Poorly, n ¼ 35 34 (88) 29 (74) 12 (30) 7 (18) 6 (6)c
Abbreviations: BSEP, bile salt export pump; Hep Par-1, hepatocyte paraffin antigen 1; pCEA, polyclonal carcinoembryonic antigen.
a
Staining in 50% or more of the tumor cells was considered positive.
b
BSEP results were not available in every case; n ¼ 39.
c
BSEP results were not available in every case; n ¼ 31.

Figure 2. Well-differentiated hepatocellular carcinoma (A) showing cytoplasmic staining with arginase-1 (B) and hepatocyte paraffin antigen 1 (C).
D, Canalicular pattern of staining with bile salt export pump. E, Staining for glypican-3 is negative (hematoxylin-eosin, original magnification 3200
[A]; original magnification 3200 [B]; original magnification 3200 [C]; original magnification 3200 [D]; original magnification 3200 [E]).
Arch Pathol Lab Med—Vol 139, August 2015 Comparison of 5 Immunohistochemical Markers—Nguyen et al 1031
Figure 3. Moderately differentiated hepatocellular carcinoma (A) showing cytoplasmic staining with arginase-1 (B) and glypican-3 (C), whereas
hepatocyte paraffin antigen 1 (Hep Par 1) is negative (D). This was the only moderately differentiated case that was negative for Hep Par 1. Staining
for bile salt export pump and polyclonal carcinoembryonic antigen was also negative (not shown) (hematoxylin-eosin, original magnification 3200
[A]; original magnification 3200 [B]; original magnification 3200 [C]; original magnification 3200 [D]).

mors, such as neuroendocrine tumors, renal cell carcinoma,
adrenocortical carcinoma, and melanoma. A large arma-
mentarium of immunohistochemical markers is available to
help in establishing hepatocellular differentiation in tumors.
As a limited amount of tissue is available in needle biopsies,
judicious evidence-based use of these markers is imperative.
Our results show that Arg-1 is the most sensitive marker
across all differentiations of HCC. Several studies have
shown the utility of Arg-1 in HCC. Similar to our finding of
98% sensitivity, figures ranging from 84% to 96% have been
reported.13,14,18 The high sensitivity of Arg-1 (81%–84%) has
also been observed in cytology material.19,20 The high
sensitivity of Arg-1 was also observed in poorly differenti-
ated HCC (95% in our study, which is similar to the 90%
reported by Yan et al13) as well as in scirrhous HCC.21 In
addition, Arg-1 was expressed in all well and moderately
differentiated cases, making it an excellent hepatocellular
marker for both ends of the differentiation spectrum. In
most cases that were Arg-1þ, the staining was diffuse,
indicating that this marker will yield similarly high
sensitivity in biopsies. The high sensitivity of Arg-1 in a
microarray study also indicates that the staining pattern is
diffuse in most cases and will have high utility in liver
biopsies.22 Arginase-1 is a component of the urea cycle and
is found only in the liver. Although no nonhepatocellular
1032 Arch Pathol Lab Med—Vol 139, August 2015
tumors were examined in this study, high specificity (.95%)
has been reported.13,18 Rare cases of positive staining in
adenocarcinomas of the prostate, pancreas, colon, breast,
and biliary tree have been reported.13,19 The only exception
is a study by Iida et al,23 which reported Arg-1 positivity in
78% of intrahepatic cholangiocarcinomas. However, that
result has not been reproduced in other studies and is not
supported by our anecdotal experience.
Although GPC-3 emerged as a promising hepatocellular
marker a few years ago, further experience has shown that it
can be expressed in a wide variety of other tumors.10 Even
though GPC-3 is not a specific hepatocellular marker, it can
be useful because of its high sensitivity in poorly differen-
tiated HCC.9,11 Among the tumors that can resemble HCC,
GPC-3 can be rarely positive in cholangiocarcinoma.10,11,21,24
In the present study, GPC-3 was positive in 85% (33 of 39)
of poorly differentiated HCC, which is similar to 85% and
89% reported in previous studies.9,11 Use of GPC-3 has also
been advocated for distinguishing well-differentiated HCC
from hepatocellular adenoma.25–27 Based on our results, the
utility of GPC-3 in this setting is limited because it is
positive in only 62% (8 of 13) of well-differentiated HCC.28
Even among positive cases, the diffuse staining was seen in
only 15% (2 of 13) of the cases, further confirming its limited
utility in biopsies.
Comparison of 5 Immunohistochemical Markers—Nguyen et al
Figure 4. Poorly differentiated hepatocellular carcinoma (A) showing cytoplasmic staining with arginase-1 (B) and glypican-3 (C), whereas bile salt
export pump was negative (D). Staining for hepatocyte paraffin antigen 1 and polyclonal carcinoembryonic antigen was also negative (not shown)
(hematoxylin-eosin, original magnification 3200 [A]; original magnification 3200 [B]; original magnification 3200 [C]; original magnification 3200
[D]).

As previously reported,2–7 Hep Par 1 and pCEA are widely
used in the diagnosis of HCC, and high sensitivity and
specificity has been reported. However, both these markers
have poor sensitivity for poorly differentiated and scirrhous
HCC.5,9,13,21 Our study also shows similar findings, with
sensitivity of Hep Par 1 and pCEA for poorly differentiated
HCC being 64% (25 of 39) and 54% (21 of 39), respectively.
Even among the positive cases, the staining was focal in
nearly one-half of the cases, further reducing their utility in
biopsies. In addition, Hep Par 1 staining is common in
esophageal, gastric, and lung adenocarcinomas2,3 and rare in
neuroendocrine, biliary, and other tumors.7,29
Because BSEP is localized to the canalicular membrane, it
is an attractive option for establishing hepatocellular
differentiation. A preliminary study16 showed that BSEP
had 100% specificity and 90% sensitivity for HCC, and its
use was advocated for the diagnosis of HCC. However, that
was a small series, and it is not clear whether any poorly
differentiated HCCs were included. In our study, BSEP had
high sensitivity for well-differentiated and moderately
differentiated HCC, but fell to less than one-half for poorly
differentiated cases. If only diffusely positive cases were
considered, the sensitivity in poorly differentiated HCC fell
even further to 6% (2 of 31). In view of the availability of
better markers, such as Arg-1 and GPC-3, for poorly
differentiated HCC, and better or equally good markers,
such as Hep Par 1 and Arg-1, for moderately differentiated
and well-differentiated HCC, our results suggest that the
addition of BSEP is not useful for diagnosis.

Table 4. Sensitivity of Different Combinations of Immunohistochemical Markers in Hepatocellular Carcinoma

Hep Par-1þ and/or Glypican-3þ, Hep Par-1þ and/or Arginase-1þ, Glypican-3þ and/or Arginase-1þ,
Differentiation No. (%) No. (%) No. (%)
Tumor cells staining, % 5 50 5 50 5 50
Well, n ¼ 13 13 (100) 13 (100) 13 (100) 13 (100) 13 (100) 13 (100)
Moderately, n ¼ 41 41 (100) 40 (98) 41 (100) 40 (98) 41 (100) 41 (100)
Poorly, n ¼ 39 36 (97) 34 (87) 37 (97) 34 (88) 39 (100) 37 (95)
Abbreviation: Hep Par-1, hepatocyte paraffin antigen 1.
Arch Pathol Lab Med—Vol 139, August 2015 Comparison of 5 Immunohistochemical Markers—Nguyen et al 1033
 Our study indicates that different combinations of
hepatocellular markers can be considered based on the
differentiation of the tumor. If the morphologic differential
diagnosis includes well differentiated or moderately differ-
entiated HCC, Arg-1 is likely to identify most cases. Because
Arg-1 can show focal staining in rare cases, and its
sensitivity and specificity may decrease with further
experience, it is prudent to employ a panel of markers in
these situations. If poorly differentiated HCC is a morpho-
logic consideration, a combination of Arg-1 and GPC-3
yields the highest sensitivity with positive staining of at least
one marker in all the cases. Even when cases with staining
of 50% of cells are considered positive, this combination
identifies 95% of poorly differentiated HCCs, indicating that
it will be a robust combination in needle biopsies. Along
with the data presented in this study, each laboratory should
take its practice environment and technical performance of
stains into consideration when choosing an optimal
combination of markers for specific clinical situations.
We thank the University of California, San Francisco, Liver
Center for their help in this study (NIH P30 DK026743).
References
1. Butler SL, Dong H, Cardona D, et al. The antigen for Hep Par 1 antibody is
the urea cycle enzyme carbamoyl phosphate synthetase 1. Lab Invest. 2008;
88(1):78–88.
2. Fan Z, van de Rijn M, Montgomery K, Rouse RV. Hep Par 1 antibody stain
for the differential diagnosis of hepatocellular carcinoma: 676 tumors tested using
tissue microarrays and conventional tissue sections. Mod Pathol. 2003;16(2):
137–144.
3. Kakar S, Muir T, Murphy LM, Lloyd RV, Burgart LJ. Immunoreactivity of
Hep Par 1 in hepatic and extrahepatic tumors and its correlation with albumin in
situ hybridization in hepatocellular carcinoma. Am J Clin Pathol. 2003;119(3):
361–366.
4. Morrison C, Marsh W Jr, Frankel WL. A comparison of CD10 to pCEA,
MOC-31 and hepatocyte for the distinction of malignant tumors in the liver. Mod
Pathol. 2002;15(12):1279–1287.
5. Kakar S, Gown AM, Goodman ZD, Ferrell LD. Best practices in diagnostic
immunohistochemistry: hepatocellular carcinoma versus metastatic neoplasms.
Arch Pathol Lab Med. 2007;131(11):1648–1654.
6. Lau SK, Prakash S, Geller SA, Alsabeh R. Comparative immunohistochem-
ical profile of hepatocellular carcinoma, cholangiocarcinoma, and metastatic
adenocarcinoma. Hum Pathol. 2002;33(12):1175–1181.
7. Chu PG, Ishizawa S, Wu E, Weiss LM. Hepatocyte antigen as a marker of
hepatocellular carcinoma: an immunohistochemical comparison to carcinoem-
bryonic antigen, CD10, and alpha-fetoprotein. Am J Surg Pathol. 2002;26(8):
978–988.
8. Yamauchi N, Watanabe A, Hishinuma M, et al. The glypican 3 oncofetal
protein is a promising diagnostic marker for hepatocellular carcinoma. Mod
Pathol. 2005;18(12):1591–1598.
9. Shafizadeh N, Ferrell LD, Kakar S. Utility and limitations of glypican-3
expression for the diagnosis of hepatocellular carcinoma at both ends of the
differentiation spectrum. Mod Pathol. 2008;21(8):1011–1018.
10. Baumhoer D, Tornillo L, Stadlmann S, Roncalli M, Diamantis EK,
Terracciano LM. Glypican 3 expression in human nonneoplastic, preneoplastic,
and neoplastic tissues: a tissue microarray analysis of 4,387 tissue samples. Am J
Clin Pathol. 2008;129(6):899–906.
1034 Arch Pathol Lab Med—Vol 139, August 2015
11. Shirakawa H, Kuronuma T, Nishimura Y, et al. Glypican-3 is a useful
diagnostic marker for a component of hepatocellular carcinoma in human liver
cancer. Int J Oncol. 2009;34(3):649–656.
12. Multhaupt H, Fritz P, Schumacher K. Immunohistochemical localization of
arginase in human liver using monoclonal antibodies against human liver
arginase. Histochemistry. 1987;87(5):465–470.
13. Yan BC, Gong C, Song J, et al. Arginase-1: a new immunohistochemical
marker of hepatocytes and hepatocellular neoplasms. Am J Surg Pathol. 2010;
34(8):1147–1154.
14. Radwan NA, Ahmed NS. The diagnostic value of arginase-1 immunostain-
ing in differentiating hepatocellular carcinoma from metastatic carcinoma and
cholangiocarcinoma as compared to HepPar-1. Diagn Pathol. 2012;7:149. doi:
10.1186/1746-1596-7-149.
15. Strautnieks SS, Bull LN, Knisely AS, et al. A gene encoding a liver-specific
ABC transporter is mutated in progressive familial intrahepatic cholestasis. Nat
Genet. 1998;20(3):233–238.
16. SM Lagana, H Remotti, RK Moreira. Bile salt export pump (BSEP): a
sensitive and specific marker of hepatocytic differentiation in liver tumors
[abstract]. Mod Pathol. 2012;25:417A.
17. Bosman FT, Carneiro F, Hruban RH, Theise ND. WHO Classification of
Tumours of the Digestive System. 4th ed. Lyon, France: IARC Press; 2010. World
Health Organization Classification of Tumours; vol 3.
18. Timek DT, Shi J, Liu H, Lin F. Arginase-1, HepPar-1, and glypican-3 are the
most effective panel of markers in distinguishing hepatocellular carcinoma from
metastatic tumor on fine-needle aspiration specimens. Am J Clin Pathol. 2012;
138(2):203–210.
19. Fujiwara M, Kwok S, Yano H, Pai RK. Arginase-1 is a more sensitive marker
of hepatic differentiation than HepPar-1 and glypican-3 in fine-needle aspiration
biopsies. Cancer Cytopathol. 2012;120:230–237.
20. McKnight R, Nassar A, Cohen C, Siddiqui MT. Arginase-1: a novel
immunohistochemical marker of hepatocellular differentiation in fine needle
aspiration cytology. Cancer Cytopathol. 2012;120(4):223–229.
21. Krings G, Ramachandran R, Jain D, et al. Immunohistochemical pitfalls
and the importance of glypican 3 and arginase in the diagnosis of scirrhous
hepatocellular carcinoma. Mod Pathol. 2013;26(6):782–791.
22. Ramachandran R BL MI, Chen Y, Kakar S. Evidence-based immunohisto-
chemical panel for the distinction of hepatocellular carcinoma and metastatic
carcinoma [abstract]. Mod Pathol. 2010;23:370A.
23. Iida H, Hata M, Kakuno A, Hirano H, et al. Expression of hepatocyte
markers in mass-forming peripheral and periductal-infiltrating hilar intrahepatic
cholangiocarcinomas. Oncol Lett. 2011;2(6):1041–1046.
24. Wang FH, Yip YC, Zhang M, et al. Diagnostic utility of glypican-3 for
hepatocellular carcinoma on liver needle biopsy. J Clin Pathol. 2010;63(7):599–
603.
25. Nassar A, Cohen C, Siddiqui MT. Utility of glypican-3 and survivin in
differentiating hepatocellular carcinoma from benign and preneoplastic hepatic
lesions and metastatic carcinomas in liver fine-needle aspiration biopsies. Diagn
Cytopathol. 2009;37(9):629–635.
26. Wang HL, Anatelli F, Zhai QJ, Adley B, Chuang ST, Yang XJ. Glypican-3 as
a useful diagnostic marker that distinguishes hepatocellular carcinoma from
benign hepatocellular mass lesions. Arch Pathol Lab Med. 2008;132(11):1723–
1728.
27. Lagana SM, Salomao M, Bao F, Moreira RK, Lefkowitch JH, Remotti HE.
Utility of an immunohistochemical panel consisting of glypican-3, heat-shock
protein-70, and glutamine synthetase in the distinction of low-grade hepatocel-
lular carcinoma from hepatocellular adenoma. Appl Immunohistochem Mol
Morphol. 2013;21(2):170–176.
28. Shafizadeh N, Kakar S. Diagnosis of well-differentiated hepatocellular
lesions: role of immunohistochemistry and other ancillary techniques. Adv Anat
Pathol. 2011;18(6):438–445.
29. Minervini MI, Demetris AJ, Lee RG, Carr BI, Madariaga J, Nalesnik MA.
Utilization of hepatocyte-specific antibody in the immunocytochemical evalu-
ation of liver tumors. Mod Pathol. 1997;10(7):686–692.
Comparison of 5 Immunohistochemical Markers—Nguyen et al