Immunohistochemistry in the

Diagnosis of Metastatic Carcinoma

of Unknown Primary Origin

Rodney T. Miller, M.D.

Director of Immunohistochemistry
ProPath Laboratory
1355 River Bend Drive
Dallas, TX 75247-4915
[email protected]
www.propath.com

American Academy of Oral and

Maxillofacial Pathology Annual Meeting
San Juan, Puerto Rico
Saturday, April 30, 2011, 8:30-11:30 am

Miller

IHC for Carcinoma of Unknown Primary

Page 1

INTRODUCTION

Carcinomas are undoubtedly the most frequent type of malignancy seen by diagnostic
surgical pathologists.

Making the diagnosis of carcinoma is often very simple, but

determining its origin can be very challenging. Occasionally prior medical history, clinical
findings, or x-ray findings may make the origin obvious, but as we all know there are many
cases where the primary site remains a mystery. A valid question that may arise is whether
finding the primary really matters at all. For some patients with widespread metastatic
disease and a virtually hopeless prognosis, the answer to this question may be no, and
resources would be better spent at providing palliative and comfort care. However, we all
know that clinicians, patients, and their families frequently want to know where the cancer
is coming from, and immunohistochemistry (IHC) is well suited to address this problem.
There are some who criticize the use of immunohistochemistry in this situation because it
is "expensive". However, from my standpoint, I know that IHC frequently obviates the
need for certain far more expensive diagnostic procedures that would be considered during
the search for a primary. Although a complete battery of immunostains may generate a
sizable bill, in the grand scheme of patient care it is well worth the cost and frequently
saves the patient and the health-care system a great deal of money if performed well and
early in the patients course. Furthermore, the cost of a misdiagnosis is far greater than the
cost of an appropriate battery of immunostains.

Principles of Immunophenotyping
Before discussing specifics, it is important to keep certain principles in mind at all
times, which will help to keep us out of trouble when using immunostains to assist with
diagnostic problems.

I. Immunostains must be of high quality. My late father, a very wise, talented and
practical man, told me many times when I was growing up that "You need good tools if
you want to do a good job", words that are particularly pertinent to IHC and diagnostic

Miller

IHC for Carcinoma of Unknown Primary

Page 2

pathology in general. If the immunostains employed are of substandard quality and suffer
from poor sensitivity, poor specificity, lack of reproducibility, or are performed in a
laboratory with lax quality control procedures that does not validate the methodology and
markers used in-house, it is a waste of time (and overtly dangerous) to embark on
immunophenotyping. Either fix the problems in the lab, or send the stains to a lab that can
do them right.

(Subliminal message: Send all your technical immunostaining work to Millers Lab).

Another

dangerous mindset is to assume that "automation = quality. Using a machine to do your

immunostains does not free you from the responsibility of optimizing (including
determination of optimal titers) and validating your immunostains, something that should
not be abdicated to a vendor or manufacturer (Cardinal Rule 1: There is no such thing as a
"predilute ready to use" antibody.) (Cardinal Rule 2: If you use an avidin-biotin based
detection system, you must take steps to block endogenous biotin activity with every stain
you perform, or you are asking for trouble). (Cardinal Rule 3: You should use some type
of multitumor or multitissue positive control material, and it should be mounted on the
same slide as the patient tissue). Cardinal Rule 4: Treat your tissue right (garbage in,
garbage out).

II. You must be able to generate the appropriate differential diagnosis based on
H&E, know the spectrum of reactivity of the markers used, and know the expected
immunophenotypes of the tumors in the differential diagnosis. This point may be an
obvious one (duh..), but it is worth stressing, since incomplete knowledge of the
spectrum of immunoreactivity with the various markers used has been responsible for
many diagnostic errors when employing IHC in diagnosis. For example, many people
employ cytokeratin AE1/AE3 as a pan-cytokeratin NOT!! (see discussion in the
:undifferentiated malignant tumor handout). If you plan on using these markers in your
practice, you must make a commitment to become knowledgeable about the tools that you
will be using. I can think of several genius-level pathologists who can remember all of this
important information, but I am certainly not one of them, so I rely heavily on a
comprehensive series of notes that I call my IHC peripheral brain. My personal IHC
peripheral brain is in spreadsheet format (which is readily searchable), and has a number

Miller

IHC for Carcinoma of Unknown Primary

Page 3

of different sheets in the complete workbook, organized to rapidly assist in answering

the following recurrent questions:

A. Given the clinical findings and H&E morphology of this tumor, what entities
should I be thinking about in the differential diagnosis? This sheet in the "brain" is
broken into several categories (such as pleomorphic large cell tumors, small blue cell
tumors, epithelioid tumors, spindle cell tumors, etc.), with tumors that may show
those morphologic features listed under the appropriate category. Going through this list
when I see cases has helped me on many occasions to stumble onto the correct diagnosis,
or at least point my way toward the correct path.

B. What type of tumors would be expected to stain (or not stain) with antibody X?
This sheet in the "brain" consists of a list of antibodies, with expected positive and negative
tumors listed below each antibody, along with pertinent notes and references to both the
pertinent literature and to prior personally-studied cases that can be retrieved for review if
needed.

C. What is the expected immunophenotype of tumor X? (i.e., does the

immunophenotype that I see in this case fit for tumor X?) This sheet in the workbook
consists of a list of tumors with expected immunophenotypes listed under the particular
tumors, along with pertinent notes and references as above.

III. There are no perfect markers. Or, to put it another way, there are no (or virtually
no) markers that are 100% sensitive and 100% specific. Generate a logical differential
diagnosis based on the clinical and morphologic findings, and USE PANELS of antibodies
to narrow the differential diagnosis. Dont succumb to immunohistochemical guilt
when ordering panels of antibodies for difficult cases. Unfortunately, some pathologists
suffer from intense guilt every time they order an immunostain, and their level of guilt rises
(sometimes exponentially) with each immunostain that is ordered. I have seen many cases
where this guilt has directly contributed to a misdiagnosis, secondary to insufficient and

Miller

IHC for Carcinoma of Unknown Primary

Page 4

incomplete analysis of the case. My advice to these pathologists is to Get over it (the
tough love approach), or see a therapist if needed, since as mentioned previously, the
cost of an erroneous diagnosis is far greater than the cost of an appropriate panel of
immunostains.

IV. Tumors do not read textbooks (and sometimes the textbooks are wrong or
outdated). This is an important point to keep in mind, because you will undoubtedly see
cases that exhibit immunophenotypes that they are not supposed to have (particularly if
the immunostains are not of high-quality or if you are using so-called "predilute ready to
use" antibodies). Dont allow an aberrant immunophenotype to sway you into making an
insane diagnosis, and before accepting an aberrant phenotype, make sure you look at the
positive control material (which ideally should consist of a multitumor sandwich with
expected positive and expected negative cases, and be on the same slide as the patient
tissue), to make certain that the correct antibody has been placed on that slide, and to make
certain that the slide bears the appropriate label corresponding to the actual stain that was
performed. The immunophenotype is only one piece of evidence (along with the H&E
morphology, clinical findings, and laboratory findings) that you should consider before
making a diagnosis. In addition to good immunostains, it takes common sense to be a good
pathologist.

V. You will get some cases wrong (hopefully not very many). Anyone who has not
gotten a case wrong has probably not been in practice for very long, or has a personality
disorder that prevents them from seeing (or admitting) their mistakes (i.e., those that
trained at the It is what I say it is because I say so school of pathology). Unfortunately,
we live in an imperfect world.

Miller

IHC for Carcinoma of Unknown Primary

Page 5

ANTIBODIES FOR METASTATIC CARCINOMA

OF UNKNOWN PRIMARY
ORGAN-SPECIFIC OR ORGAN-RELATED MARKERS

PROSTATE

MARKERS:

Prostate-Specific

Antigen

(PSA),

Prostatic

Acid

Phosphatase (PSAP), Prostate-Specific Membrane Antigen (PSMA), and P501S

(prostein) (all cytoplasmic reactivity(1-12); NKX3.1(nuclear reactivity) (12a): PSA and
PSAP antibodies have been around a long time, and they are very useful for detecting
prostate adenocarcinoma, although they may also be found in some other types of tumors
(see below). PSA is very specific for prostate carcinoma, although it will also stain a small
number of breast carcinomas and some salivary gland tumors. Rectal carcinoid tumors
(and less frequently non-rectal carcinoids) may show reactivity with PSAP, so that
potential trap is important to know. Parenthetically, some primary prostatic tumors may
express significant chromogranin and synaptophysin, and may look just like carcinoid
tumors (4), another potential pitfall. It is not at all unusual for PSA to be negative in
poorly-differentiated prostate tumors, and in my experience PSAP is more sensitive than
PSA. PSAP reactivity has also been described in periurethral glands in females, sweat
glands, breast carcinoma, rare islet cell tumors, some salivary gland tumors, and rare renal
cell carcinomas. Some authors report that PSA and PSAP are negative in about 5% of high
grade prostate cancers, and others report that treatment of prostate cancer may be
associated with loss of reactivity to PSA and PSAP. In these situations, the more recently
available markers PSMA and P501S (prostein) may be of particular utility. PSMA has
been shown to be positive in many prostate cancers that are negative for PSA and PSAP (710). It is a transmembrane glycoprotein that functions as a folate hydrolase, and it is
expressed at low levels in benign prostate epithelium, but shows marked upregulation in
prostate carcinoma.

The degree of expression of this marker seems to be inversely

correlated with the degree of differentiation of the tumor. As such, high grade prostate
carcinomas tend to express this marker in a very high percentage of tumor cells, whereas
lower grade tumors show more heterogeneous expression. Expression of PSMA has also
been reported in the brain (weak), salivary glands, a subpopulation of proximal renal

Miller

IHC for Carcinoma of Unknown Primary

Page 6

tubules, duodenal mucosa, and a subpopulation of neuroendocrine cells in colonic mucosa.

Interestingly, endothelial cells within tumors of many different types express PSMA,
although most normal endothelial cells are negative for PSMA (although I have seen
normal liver sinusoidal endothelial cells express PSMA on many occasions). Epstein (12)
found expression of P501S in 68 of 69 (99%) of cases of prostate cancer, and it has a
characteristic perinuclear cytoplasmic Golgi pattern of reactivity. In addition to prostate
cancers, I have seen P501S reactivity in a few breast carcinomas (not a big problem in male
patients), and in a rare lung carcinoma and an acinic cell carcinoma of the parotid.
NKX3.1 is the newest prostate-related marker on the scene (and has become commercially
available only since November 2010). Unlike the other prostate markers, it is localized to
the nucleus of the cells. Gurel et al (12a) studied tissue microarrays (TMAs) of 69 prostate
carcinomas and 349 non-prostate carcinomas, and found that nuclear reactivity with
NKX3.1 had excellent sensitivity for prostate carcinoma, staining 68 of the 69 cases,
(98.6%). Only 1 of the 349 cases of non-prostate carcinoma was positive, a case of lobular
breast carcinoma.

From a practical standpoint, if I am screening a poorly differentiated tumor for

prostate origin, I typically order NKX3.1. If that is negative and prostate origin is still
strongly expected, I will then order PSMA, P501S, PSA, and PSAP.

BREAST MARKERS: Gross Cystic Disease Fluid Protein-15 and Mammaglobin

(cytoplasmic reactivity) (13-23): GCDFP-15 is a very useful marker for the identification
of breast carcinomas, although it is positive in only about 50-60% of primary breast
carcinomas (8-9). It is important to note that the pattern of reactivity with this antibody is
often very focal, and only a small percentage of tumor cells may be immunoreactive. Some
sweat gland carcinomas, salivary gland tumors, and prostate carcinomas are positive,
but it is only rarely positive in carcinomas of other sites. In a case of widely metastatic
salivary duct carcinoma that I saw several years ago in an elderly male patient, GCDFP-15
was strongly positive, and provided a strong clue regarding the origin of the patients
tumor. I have only seen four or five lung adenocarcinomas that have expressed this

Miller

IHC for Carcinoma of Unknown Primary

Page 7

antigen, usually weak and very focal, although some authors have reported rare lung
adenocarcinomas that are strongly positive for GCDFP-15. Mammaglobin is a more
recently available antibody that is useful as a marker of breast origin. The reported
sensitivity for breast cancer varies substantially, from ~40 to >85% of cases, and in my
experience the sensitivity I observe is probably in the 50% to 60% range. However, in the
cases I have studied, its expression appears to be independent of GCDFP-15 expression, so
if breast carcinoma is being considered, I always order both GCDFP-15 and mammaglobin.
Unfortunately, mammaglobin expression is not completely specific for breast origin, and it
has been reported in skin adnexal tumors, some salivary gland tumors (particularly strong
in some pleomorphic adenomas that I have seen), normal endocervical glands, and in a
significant number of ovarian carcinomas (17% in one study), endometrial carcinomas
(40-70%), and endocervical adenocarcinomas (30% in one study).

Thyroglobulin (cytoplasmic reactivity): Thyroglobulin is useful in detecting metastatic

papillary and follicular thyroid carcinoma, although it is negative in medullary
carcinoma of the thyroid. It is also positive in some anaplastic carcinomas of the thyroid,
although it may be very focal in these tumors, and most anaplastic thyroid carcinomas are
negative.

There can be difficulty interpreting the results of the stains when tumors are

invading the thyroid gland, since some authors have found that a certain amount of antigen
diffusion (from benign thyroid tissue into adjacent tumor cells) may occur, resulting in a
risk of false positive staining of tumors of non-thyroid origin that are invading the thyroid.
TTF-1 (see below) is a more sensitive marker of thyroid tumors than thyroglobulin,
although it is not as specific for thyroid origin. Pax8 is also a good marker of thyroid
neoplasms, and is discussed below.

TTF-1 (Thyroid Transcription Factor-1) (nuclear reactivity) (24-50)): TTF-1 is a

protein involved in the regulation of surfactant proteins, and it is well established as a
useful antibody for metastatic carcinoma of unknown origin. TTF-1 is normally expressed
in the brain (diencephalon), parathyroid, C-cells of the thyroid, anterior pituitary, thyroid,
and nonciliated respiratory and alveolar epithelium. Overall, it is expressed in 75% of non-

Miller

IHC for Carcinoma of Unknown Primary

Page 8

mucinous lung adenocarcinomas, 10% of mucinous lung adenocarcinomas, and 40% of

large cell carcinomas of lung. According to some authors, TTF-1 is expressed in 100% of
non-mucinous bronchoalveolar adenocarcinoma, but is essentially absent in mucinous
bronchoalveolar carcinoma (43). However, I have seen a few mucinous bronchoalveolar
carcinomas that have been TTF-1 positive. TTF-1 is negative in squamous carcinomas of
the lung (at least when using clone 8G7G3/1), although I have observed weak to moderate
staining in pulmonary squamous tumors employing the more sensitive clone SPT24. TTF1 has been found to be more sensitive than PE-10 (see below) for detection of pulmonary
adenocarcinomas (32), and my personal experience with this antibody agrees with this
contention.

PE-10 is frequently only focally positive in lung tumors, whereas TTF-1

usually stains a much higher percentage of tumor cells, so the utility of TTF-1 is often
greater when dealing with miniscule specimens, like FNAs. Thyroid carcinomas are also
positive with TTF-1 (including papillary carcinoma, follicular carcinoma, and medullary
carcinoma), although anaplastic carcinomas are generally negative, and between 20-75% of
Hurthle cell tumors are reported to express this marker. TTF-1 is rarely expressed in
stomach carcinoma (1.7%), breast carcinoma, prostate carcinoma, mesothelioma, renal cell
carcinoma, and colon carcinoma.
endometrial

adenocarcinoma,

Some authors have described TTF-1 in 17% of

and

indeed

have

adenocarcinomas with rather striking TTF-1 reactivity.

seen

several

endometrial

In a study of 546 breast

carcinomas, TTF-1 was expressed in 13 cases (2.4%), so TTF-1 reactivity by itself can not
completely exclude breast origin (47a). In my own experience, I have seen several focally
positive colonic carcinomas and a positive pancreatic carcinoma, and urothelial carcinomas
may have scattered positive cells in some cases, sometimes quite strong. I have also
observed positivity in a case of desmoplastic small round cell tumor and in a subset of
lymphocytes. Strong nonspecific granular cytoplasmic staining can be observed with
TTF-1 when using clone 8G7G3/1, particularly in hepatoma (45) (where it may be a
clue to this diagnosis), GI tumors, and prostate tumors, but for the purposes of use as a
pulmonary and thyroid marker, this type of reactivity should be ignored, as only nuclear
reactivity is significant with this antibody. This cytoplasmic reactivity with clone 8G7G3/1
has been found to be due to cross-reactivity with antigens in hepatocyte mitochondria (50).

Miller

IHC for Carcinoma of Unknown Primary

Page 9

TTF-1 clone SPT24 does not show cytoplasmic staining of the type seen with clone
8G7G3/1. TTF-1 clone SPT24 is substantially more sensitive for lung carcinomas than
clone 8G7G3/1, but according to some authors may also be somewhat less specific, and
may stain some colonic adenocarcinomas (48-49). In my mind, I think that SPT24 is a
better clone than 8G7G3/1, and the so-called nonspecificity of SPT24 is actually a
reflection of its superior sensitivity. At the 1999 USCAP meeting, one group reported that
TTF-1 was absent in all of 82 thymic epithelial tumors, and was positive in 1 of 25 thymic
carcinomas (31), although we see weak to moderate TTF-1 reactivity in thymic tumors
with clone SPT24. TTF-1 is negative in mesothelioma, so it has utility in the differential
diagnosis of mesothelioma vs. adenocarcinoma. TTF-1 is also present in a fair number of
neuroendocrine tumors, including 90% of small cell carcinomas of the lung according to
some authors. In my personal experience, clone 8G7G3/1 stains about 50% of small cell
carcinomas, and clone SPT24 stains about 75-85% of small cell carcinomas.

80% of

atypical carcinoids of the lung are reported to be TTF-1 positive, but only 20% of typical
carcinoids of the lung. In addition to pulmonary small cell carcinoma, expression of TTF-1
has been reported in one study in 44% of non-pulmonary small cell carcinomas (4/4
prostate, 2/4 bladder, 1/7 cervix), and this study also reported absence of expression in all
of 49 cases of gastrointestinal carcinoids, all of 15 pancreatic islet cell tumors, and all of 21
paragangliomas (35). Another study found expression of TTF-1 in 81% of 37 pulmonary
small cell carcinomas, and also in 80% of 15 non-pulmonary small cell carcinomas, so
TTF-1 does not have pulmonary-related specificity in the setting of small cell carcinoma
(36).

However, TTF-1 is negative in Merkel cell tumor, which can assist in the

differential diagnosis from small cell carcinoma (often TTF-1 positive) (37). Pulmonary
sclerosing hemangiomas are positive for TTF-1 (47).

Medullary carcinomas of the

thyroid are TTF-1 positive.

Napsin A (cytoplasmic reactivity) (220-221): Napsin A is similar to TTF-1 with respect to

its sensitivity and specificity for lung carcinoma. Using tissue microarrays, Bishop et al
(220) studied 95 cases of lung carcinoma, and Napsin A was positive in 83% of case (TTF1 was positive in 73% of cases). There were 13 Napsin A positive, TTF-1 negative cases,

Miller

IHC for Carcinoma of Unknown Primary

Page 10

and 2 TTF-1 positive, Napsin A negative cases. By using both Napsin A and TTF-1, these
authors detected 85% of the cases of lung carcinoma. All 48 squamous carcinomas, 6
neuroendocrine tumors, 5 colonic, 31 pancreatic, 17 breast, and 38 mesotheliomas were
negative. Of 118 renal cell carcinomas, 79% of the papillary renal cell carcinomas,
34% of clear cell carcinomas, and 3% of chromophobe carcinomas were positive. Of 81
thyroid tumors, only 5% of papillary carcinomas (2 cases, both with tall cell morphology)
were Napsin A positive

PE-10 (Surfactant Apoprotein A) (cytoplasmic reactivity): This antibody has good

specificity for lung adenocarcinoma and thyroid carcinoma, although its sensitivity for
pulmonary origin is poor in my experience, so we rarely use it. Langel and colleagues (26)
report that PE-10 stains about 60-70% of lung adenocarcinomas, although in my laboratory
I would estimate that we have seen it positive in 20-30% or less of lung carcinomas,
similar to other reports (32).

It will also stain alveolar lining cells and alveolar

macrophages, so that must be kept in mind when interpreting the results, especially in
pleural fluids or lung FNAs, that may show strong reactivity in background normal lung
elements or in the background tissue fluid. This antibody is reported to be absent in breast,
colon, renal cell, and endometrial carcinomas. Like GCDFP-15, reactivity with PE-10 may
be quite focal, and that is a point that must be kept in mind when forced to deal with very
tiny amounts of diagnostic material. One report described observing PE-10 reactivity in 6
of 15 (40%) of prostate carcinomas and 3 of 7 (43%) thyroid carcinomas.

Estrogen Receptor (ER) and Progesterone Receptor (PR) (nuclear reactivity) (57-71):
Estrogen receptor can be very useful in determining the origin of metastatic carcinoma. It
is common knowledge that ER is positive in many breast carcinomas and also female
genital tract tumors (both epithelial and stromal), but it can also be positive in a number
of other tumors. Tumors that may express ER include thyroid tumors, salivary gland
tumors, sweat gland carcinomas, genital angiomyofibroblastoma, and 80% of aggressive
angiomyxomas.

ER also has been recently described in some cases of skull base

chordomas (63). I have seen ER expressed in normal hepatocytes and in a few hepatomas.

Miller

IHC for Carcinoma of Unknown Primary

Page 11

ER has been reported in 7% of carcinoid tumors. Most pulmonary carcinomas are ER

negative, although 4%-15% of pulmonary carcinomas may express ER, and indeed I
have this on many occasions in my consultation practice, although it is usually focal or
patchy and relatively weak, with a few exceptional cases that show strong diffuse staining.
When using the ER antibody clone 6F11 on the Ventana automated immunostainer, Dabbs
et al (69) reported ER positivity in 67% of a series of 45 primary pulmonary
adenocarcinomas, although since I do not use that clone, I cannot comment on that figure.
(They did not report any ER positive cases when using clone 1D5). ER is negative in
small cell carcinomas and gastrointestinal tract carcinomas, so ER reactivity can be
very useful in ruling out those possibilities. ER and PR reactivity is also reported in
stromal cells and rarely epithelial cells of hepatobiliary and pancreatic mucinous
cystadenocarcinomas (64), although in my experience pancreatic ductal adenocarcinomas
and the usual type of biliary tract carcinomas are ER negative. Interestingly, I have seen
progesterone receptor (PR) reactivity in rare pulmonary carcinomas, and rare GI tract
tumors, so I would not recommend using PR in the same fashion as ER if one is dealing
with the problem of a metastatic tumor of unknown primary. PR is also reported to be
positive in some medullary carcinomas of the thyroid, some melanomas, meningiomas, and
in 20% of carcinoid tumors and 20% of small cell carcinomas. I have also seen strong PR
reactivity in minute pulmonary meningothelial-like nodules (so-called chemodectoma).

HepPar 1 (cytoplasmic granular reactivity) (72-78): HepPar 1 (short for hepatocyte

paraffin 1) is a monoclonal antibody useful in the diagnosis of hepatocellular carcinoma
(HCC), where it has been found to show 82% sensitivity and 90% specificity for
hepatocellular neoplasms in one study, although I think it is vastly overrated as a hepatoma
marker. In my own experience, I see it in only about 50% of the HCCs that are sent to me
for immunophenotyping, although there is probably some selection bias in my figures,
since classic cases of HCC are generally not sent to us for immunophenotyping. It is far
more sensitive than AFP for hepatoma, as AFP stains only about 15% of cases. However,
it is not completely specific for hepatoma, and has been found to be strongly expressed in a
fairly high percentage of gastric adenocarcinomas, as well as occasional other non-liver

Miller

IHC for Carcinoma of Unknown Primary

Page 12

tumors, underscoring the importance of using appropriate panels of antibodies when

evaluating cases, and not relying on a single marker.

Arginase-1(Arg-1)(cytoplasmic and nuclear reactivity) (227): Argnase-1, an enzyme

involved in the urea cycle, has been found to be expressed normally in liver cells but few
others in the body. At the 2010 USCAP meeting, Yan et al (abstract #1668) reported on
their experience using Arginase-1 as a specific marker of hepatocytes and hepatocellular
neoplasms, and they subsequently published their findings in the August 2010 issue of Am
J Surg Pathol.

A series of 193 hepatocellular carcinomas (HCC), Arginase-1 had a

sensitivity of 96%.

Only 2 of 557 non-hepatocellular tumors expressed Arg-1 (1

cholangiocarcinoma and 1 prostate carcinoma), and the staining was focal and weak. When
compared with HepPar1, Arg-1 was clearly superior from both the standpoints of
sensitivity and specificity for HCC. Arg-1 stains both cytoplasm and nuclei, but the authors
required cytoplasmic reactivity in order to qualify for a "positive" Arg-1 stain. Arg-1 also
stains neutrophils and macrophages. I have been very impressed with this antibody, and
expect that it will relegate HepPar1 to the trash bin for the diagnosis of HCC.

Pax8 (nuclear reactivity) (217-219): Pax 8 is relatively new on the scene, but I have
found it to be one of the most useful and highly valued markers for addressing the problem
of metastatic carcinoma of unknown primary, where it can be used as a marker of thyroid
carcinomas, female genital tract carcinomas, and renal cell carcinomas. In one study
of 94 thyroid tumors (17 papillary carcinomas, 18 follicular adenomas, 16 follicular
carcinomas, 7 poorly differentiated carcinomas, 28 anaplastic carcinomas, and 8 medullary
carcinomas), Pax 8 was diffusely expressed in all of the papillary carcinomas, follicular
adenomas, and poorly differentiated carcinomas. Expression was variable in medullary
carcinomas. In contrast to TTF-1 (which stained only 18% of the anaplastic carcinomas),
Pax8 was positive to a variable degree in 79% of the anaplastic tumors.

In a study of 182 kidney tumors, Tong et al (217) found Pax8 expression in 98% of
clear cell renal cell carcinomas, 90% of papillary renal cell carcinomas, 82% of

Miller

IHC for Carcinoma of Unknown Primary

Page 13

chromophobe carcinomas, 71% of sarcomatoid renal carcinomas, and 95% of

oncocytomas. 23% of urothelial carcinomas arising from the renal pelvis were positive for
Pax8, but not those arising in the bladder.

Nonaka et al (218) found Pax8 to be very useful in distinguishing ovarian carcinomas

from breast carcinomas. Of 124 ovarian carcinomas (84 papillary serous, 18 endometrioid,
12 mucinous, 10 clear cell), Pax 8 was expressed typically in a diffuse fashion in 96% of
the papillary serous tumors, 89% of the endometrioid tumors, 100% of the clear cell
tumors, and 8% of the mucinous tumors. All 243 cases of breast cancer (178 ductal and 65
lobular) were negative for Pax8. We have also found this antibody very useful in detecting
endometrial adenocarcinomas.

Long et al (219a) reported that Pax8 is positive in a high proportion of pancreatic

endocrine tumors, in the majority of duodenal and rectal carcinoid tumors, and a minor
subset of appendiceal and gastric carcinoids and it was not expressed in the ileal and
pulmonary carcinoid tumors.

At ProPath, we see occasional non-thyroid/kidney/female genital tract tumors (e.g., lung

ca, esophageal adenoca, hepatoma, thymoma, mesothelioma) that show weak staining, and
to my knowledge this is of no significance and should be ignored. Pax 8 stains some
lymphocytes (likely B-cells), histiocytes, and also classical Hodgkin cells. I have seen
cases of Pax8 positive embryonal carcinoma and pleomorphic rhabdomyosarcoma, and a
single case of Pax8 positive breast carcinoma. A case of basaloid squamous carcinoma of
the anus was moderately positive for Pax8. Normal pancreatic islet cells and normal
adrenal cortical cells are Pax8 positive.

Wilms Tumor Gene (WT1) (nuclear or cytoplasmic reactivity) (79-90): Nuclear

reactivity with WT1 has been found to be very useful in the recognition of mesothelioma,
but it is also characteristically expressed in the nuclei of serous adenocarcinoma from the
ovary and fallopian tube (or surface serous carcinoma from the pelvic peritoneum, which

Miller

IHC for Carcinoma of Unknown Primary

Page 14

are now felt to arise from small subclinical primaries in the distal fallopian tube).
Therefore, if strong nuclear reactivity with WT1 is present in a non-mesothelial tumor, it
most likely represents a serous adenocarcinoma. Goldstein (83) has reported that WT1 is
absent in uterine papillary serous adenocarcinomas, although I have seen several cases in
my laboratory that have been clearly positive with this marker. As discussed in the prior
presentation on IHC in gynecologic lesions, there are also a number of other groups that
have identified nuclear WT1 in a certain proportion of uterine serous carcinomas. Most
authors would agree however, that the frequency of nuclear WT1 expression is lower in
uterine serous carcinoma as compared to ovarian serous carcinoma (and surface serous
carcinoma of the peritoneum). Dr. Allen Gown also reports that WT1 may be expressed in
some renal cell carcinomas and prostate carcinomas, as well as mucinous carcinoma of the
breast. Cytoplasmic reactivity is present in a large number of tumors and to my knowledge
has no particular diagnostic significance (perhaps other than the typically intense
cytoplasmic reactivity in rhabdomyosarcoma). I have also seen strong nuclear reactivity
with this marker in endometrial stromal sarcoma, granulosa cell tumor, thecoma, and
normal uterine smooth muscle cells.

p63 (nuclear reactivity) (91-100): In the past several years, p63 has found increasing
utility in a number of areas of diagnostic pathology, including its use as a marker of
myoepithelial cells in breast and elsewhere, and as a marker of prostatic basal cells that
can be used as an alternative to high molecular weight cytokeratin. In addition, it serves as
a useful marker of squamous cell carcinoma (including basaloid squamous cell carcinoma
and "lymphoepithelioma") (similar to the use of cytokeratin 5 or 5/6 for recognizing
squamous differentiation).

We have also observed strong p63 in metaplastic or

sarcomatoid breast carcinoma (a finding recently observed by others as well) (99), where
it is probably a reflection of squamous or myoepithelial differentiation (considering that
these cases have also shown strong staining with cytokeratin 5 or 5/6, also typical of
squamous and myoepithelial tumors). It also stains a significant percentage of urothelial
carcinomas, so it can be of utility in the recognition of those tumors. Not surprisingly, we
have observed strong p63 expression in Brenner tumors and transitional cell

Miller

IHC for Carcinoma of Unknown Primary

Page 15

carcinomas of the ovary. Many tumors will show occasional scattered p63-positive cells,
but that pattern of reactivity has no particular diagnostic significance. Other tumors that
may express strong and diffuse p63 include thymoma, basal cell carcinoma, and
cutaneous adnexal tumors (such as syringoma, spiradenoma, etc.). Since it is a good
myoepithelial marker, p63 also stains tumors that include a population of myoepithelial
cells or show myoepithelial differentiation, including salivary gland tumors like
pleomorphic adenoma and adenoid cystic carcinoma. Interestingly, we have observed that
benign glandular inclusions in axillary lymph nodes have an associated myoepithelial cell
layer that is highlighted nicely by p63 (and also smooth muscle myosin), a feature that can
be useful in the differential diagnosis of benign glandular inclusions vs. metastatic well
differentiated ductal carcinoma in axillary lymph nodes.

Renal cell carcinoma marker (RCC) (cytoplasmic reactivity) and Pax-2 (nuclear
reactivity) (51-56). RCC (aka gp220) has been available for a number of years, but only
since clone PN-15 became available did I have much success with it, and even then not
much success. To be honest, I think both RCC and Pax2 are overrated as markers of
kidney tumors, particularly since Pax8 became available as a kidney marker. The RCC
antibody requires enzymatic digestion for optimal staining (we use pepsin), and one thing
that we discovered when using RCC is that it is important to do the stain with several
protease digestion times (we use 5 minutes and 10 minutes), to help deal with the varying
sensitivity of different cases to protease digestion. On more than one occasion, cases of
renal cell carcinoma digested for 5 minutes have been positive and the same case digested
for 10 minutes has been negative secondary to overdigestion (and vice versa). RCC is
reportedly positive in 80% or more of renal cell carcinomas of conventional type and
papillary renal cell carcinomas, but its expression can be focal, a problem when dealing
with small biopsies. RCC is negative in chromophobe carcinoma and oncocytoma. RCC
is not completely specific, as it has also been reported in some breast carcinomas, thyroid
carcinomas, and yolk sac carcinomas.

It can also be seen in normal breast, thyroid,

epididymis and parathyroid, and I have seen it expressed in basal cell carcinoma of the skin
and parathyroid adenoma, as well as focally within a case of ovarian serous carcinoma.

Miller

IHC for Carcinoma of Unknown Primary

Page 16

More recently Pax2, a nuclear transcription factor involved in development of renal

epithelium, has been touted as a marker of renal cell carcinoma (54-55), where it has
reported in 60-88% of cases of conventional renal cell carcinoma, tending to be stronger in
the lower-grade tumors. It has also been reported in ovarian serous carcinomas, Wilms
tumor, and nephrogenic adenoma. In my lab, I have also seen it stain B-cells in lymph
nodes, normal distal tubules in the kidney, normal bile ducts, with faint staining in several
renal oncocytomas. I have not been impressed with the antibody, as it does not appear to
be particularly robust, and we have much better results with Pax8.

von Hippel-Lindau gene product (pVHL) (cytoplasmic reactivity)(222-223): Lin et al

studied a large series of tumors, and found pVHL in 99% of 79 clear cell renal cell
carcinomas, 100% of 57 papillary renal cell carcinomas, 100% of 26 chromophobe
carcinomas, 100% of 24 oncocytomas, and 95% of 37 metastatic renal cell carcinomas. Of
213 non-renal tumors studied, pVHL was found in 17.4% of cases, and 34 of these 37
cases were clear cell carcinomas of the ovary or uterus (there were 19 cases each of clear
cell carcinoma of the uterus and ovary, and pVHL stained 17 of 19 in each site, or 89%). 3
of 13 hepatomas showed focal or moderate staining. Pancreatic carcinomas (n=20),
hemangioblastomas (n=14), pheochromocytomas (n=12), colonic carcinoma (n=11), breast
carcinoma (n=42), lung adenocarcinoma (n=16), endometrioid carcinomas (n=12), serous
carcinomas (n=20), thyroid carcinomas (n=20; 10 follicular, 10 papillary), urothelial
carcinomas (n=10), and mesotheliomas (n=5) were negative.

OCT3/4 (nuclear reactivity) (101-106) and SALL4 (224-226): OCT3/4 (also known
as OCT3 or OCT4) is a nuclear transcription factor expressed in early embryonic cells,
germ cells, and stem cells. A number of studies have shown that it is a highly sensitive and
specific marker of seminoma, dysgerminoma, and embryonal carcinoma (although yolk
sac tumor is negative). Because seminoma, dygerminoma, and embryonal carcinoma can
often mimic other metastatic carcinomas, OCT3/4 is an excellent marker for screening for
these tumors. With the exception of one report of focal staining in 4 of 14 clear cell
carcinomas of the ovary (106) it stains virtually no other types of tumors, so its utility far

Miller

IHC for Carcinoma of Unknown Primary

Page 17

exceeds that of previously-used markers of germ cell tumors such as placental alkaline
phosphatase (PLAP). As might be expected, OCT3/4 is also a superb marker for detecting
in-situ germ cell neoplasia. SALL4 is a more recently described marker that stains nuclei
of seminoma, dysgerminoma, embryonal carcinoma, and yolk sac tumor, so it is an
excellent screening marker for those tumors. Ushika et al (226) report that SALL4 is also
positive in hepatoid carcinoma of the stomach, but negative in hepatoma, so it is useful
for that differential diagnosis. We have seen occasional cases of very poorly differentiated
non-germ cell carcinomas that express SALL4, and several dedifferentiated liposarcomas
have shown strong expression of SALL4.

Epithelial Membrane Antigen (membrane or cytoplasmic reactivity) (107-111):

EMA is certainly not an organ-related marker, since it is present in many carcinomas.
However, the absence of EMA does have some degree of organ-specificity.

It is

characteristically absent in adrenal carcinoma, most hepatomas (although occasional

tiny intercellular dots or canalicular patterns can be seen in some hepatomas), and certain
germ cell tumors (seminoma, embryonal carcinoma, and yolk sac tumor) (although I
have seen a few EMA-positive cells in yolk sac tumor). Papillary cystic tumors of the
pancreas, ovarian granulosa cell tumors, and Sertoli-stromal tumors are also EMA
negative. I have also seen a few EMA-negative prostate carcinomas. EMA can be positive
in choriocarcinoma.

INTERMEDIATE FILAMENTS (Cytokeratins etc.)

AND RELATED MARKERS
Cytokeratins 7 and 20 (cytoplasmic or membrane reactivity) (112-140): These are very
useful antibodies in dealing with the problem of metastatic carcinomas, since the patterns
of immunoreactivity with these two antibodies can help to substantially narrow the
likelihood of various primary sites. Common tumors positive for both CK7 and CK20
include urothelial (transitional cell) carcinoma, pancreatic carcinoma, and ovarian
mucinous carcinomas.

Invasive papillary carcinoma of the breast and about 1/3 of

mucinous breast carcinomas also co-express CK7 and CK20 (abstract, Am J Clin Pathol 110:517,

Miller

1998).

IHC for Carcinoma of Unknown Primary

Page 18

CK7 positive, CK20 negative tumors include lung, breast, non-mucinous ovarian,

endometrial, mesothelial, pancreaticobiliary tract, gallbladder, small bowel, some

squamous carcinomas, and thyroid tumors. CK7 negative, CK20 positive reactivity is
typical of colorectal carcinoma, and tumors negative for both CK7 and CK20 include
hepatoma, renal cell carcinoma, prostate carcinoma, and some squamous carcinomas.
There are a significant number of exceptions, however, particularly in bladder, stomach
and pancreaticobiliary tract tumors, so these antibodies must be used as part of a panel
approach. Despite this limitation, these antibodies are indeed extremely useful.

Villin (cytoplasmic or membrane reactivity) (141-146): Villin is a GI-related

cytoskeletal protein associated with brush border microfilaments, and it has been found to
be useful in the workup of metastatic carcinomas. In my mind, one of the most important
aspects of villin is that breast carcinomas only very rarely (<1-2% of cases) show
strong reactivity with villin. I am not aware of any reports of villin-positive invasive
breast cancer, and in all the breast carcinomas that I have seen, I remember only 6 or 7 that
have shown moderate to strong villin reactivity. Therefore, in the large majority of
instances in which strong villin reactivity is observed (particularly with a brush border
pattern), breast carcinoma is unlikely as a potential primary site. Villin is reported to be
quite sensitive in the detection of gastrointestinal tumors (including pancreas and
biliary tract), and is reported to stain nearly 100% of colon tumors, with a brush border
pattern of reactivity. It is positive in about 50% of hepatomas, and will often show a
canalicular pattern of reactivity, similar to polyclonal CEA and CD10 in some
hepatomas. Focal positivity can be seen in 26% of renal cell carcinomas and 36% of
endometrial carcinomas. A substantial number of lung adenocarcinomas express villin
(although the percentage of positive cases is certainly less than GI carcinoma), including
some cases that show a prominent brush border pattern of accentuation. Villin can be
positive in mucinous tumors of the ovary (50%), but not in serous tumors. Membranous
reactivity can be observed in carcinoid tumors. Some authors have noted that villin is
much less frequent in pancreatic endocrine neoplasms (islet cell tumors) than carcinoid
tumors (146), where it was expressed in only 7% (1of 15) of islet cell tumors, but in 82%

Miller

IHC for Carcinoma of Unknown Primary

Page 19

(18 of 22) of GI carcinoid tumors, although I have seen it expressed in a number of

pancreatic endocrine tumors. Villin was found in only 2 of 24 (8%) lung carcinoids.
These authors also found that 4 of 4 small cell carcinomas of GI origin expressed villin,
whereas all 24 lung small cell carcinomas were villin negative, as were 11 small cell
carcinomas of other sites (3 esophagus, 3 prostate, 1 bladder, 1 thyroid, 1 nose, 1 parotid, 1
ovary). Therefore, it is possible that villin expression in a metastatic small cell carcinoma
might favor a GI primary. Villin is also common in other neuroendocrine tumors,
including large cell neuroendocrine carcinomas, medullary thyroid carcinoma, and primary
neuroendocrine carcinoma (carcinoid-like) of the prostate.

Some authors report

cytoplasmic villin reactivity in up to 68% of lung carcinomas, although in my experience

this figure seems high. Normal pancreatic acini may show luminal staining for villin, and I
have seen strong villin in yolk sac tumors.

Often, the combination of immunostaining results with CK7, CK20, and villin can
substantially narrow the possibilities for most likely primary site of a metastatic carcinoma.
Tables of likely (and unlikely) possible primary sites based on the patterns of
immunoreactivity with CK7, CK20, and villin are included at the end of this handout.

Cytokeratin (AE1/AE3) (cytoplasmic reactivity): CK(AE1/AE3) should be present in

most carcinomas, but as discussed in the undifferentiated tumor talk, it is an imperfect
antibody and should NOT be viewed as a Pankeratin.

There are certain tumors

(including some cases of renal cell carcinoma, adrenal cortical carcinoma, prostatic
carcinoma, carcinoid tumors, small cell carcinomas, and pancreatic islet cell tumors)
where CK (AE1/AE3) may be weak or absent, although most of these are positive with
low molecular weight cytokeratins (i.e., cytokeratins 8 and 18). The extent and intensity of
reactivity with CK (AE1/AE3) can also be of diagnostic utility, since strong diffuse
AE1/AE3 immunoreactivity is exceedingly rare in hepatoma and seminoma (although
focal or faint reactivity may be seen in some tumors, with a perinuclear dot-like pattern in
some seminomas).

(Interestingly, AE1/AE3 may be strongly expressed in ependymoma

and in some shwannomas, and this usually parallels the degree of GFAP reactivity, but

Miller

IHC for Carcinoma of Unknown Primary

Page 20

widespread staining of these tumors with other cytokeratin antibodies is very unusual)
(156).

In addition, ependymoma is about the only tumor that I can think of (other than

some well-differentiated squamous tumors) that expresses strong AE1/AE3 but is often
negative with low molecular weight cytokeratin.

On occasion I have observed patchy

perinuclear dot-like immunoreactivity with CK (AE1/AE3) in normal myometrial smooth

muscle cells.

Low Molecular Weight Cytokeratin (CK8 & CK18) (CK-lmw) (cytoplasmic

reactivity): For detecting CK-lmw (cytokeratins 8 and 18) a number of clones can be used,
such as Zym 5.2, CAM5.2 or 5D3. CK-lmw is useful as part of an epithelial screen,
since it will nearly always detect those epithelial tumors that are negative with
Cytokeratin AE1/AE3 (particularly hepatoma, and those cases of renal cell carcinoma,
prostate carcinoma, and neuroendocrine tumors that are negative with AE1/AE3). (An
exception to this is ependymomas, which are often strongly positive with AE1/AE3, but
negative or weak with CK-lmw) (156). Finally, in most cases CK-lmw is better than
AE1/AE3 for detecting epithelial differentiation in small cell carcinomas. In practice, I
think it is always a good idea to make certain that both CK-lmw and CK-hmw are negative
before excluding the possibility of an epithelial tumor when dealing with a poorly
differentiated neoplasm. A population of CK-lmw positive spindle cells exists in normal
lymph nodes, associated primarily with the lymph node sinuses and the areas around
follicles in the cortex. These represent a normal keratin-positive population of dendritic
cells in normal lymph nodes, which are part of the accessory immune cells (149). It is
useful to keep these cytokeratin positive interstitial reticulum cells (CIRC) in mind so
that they are not misinterpreted as evidence of carcinoma, particularly when using CK-lmw
for sentinel lymph node studies on breast cancer patients.

Before proceeding further, I think it is useful to mention a pattern of cytokeratin

immunoreactivity frequently seen in neuroendocrine tumors.

Most pathologists are

probably familiar with the perinuclear globs of keratin in Merkel cell tumor, but
perhaps fewer are aware that similar but smaller perinuclear keratin dots are

Miller

IHC for Carcinoma of Unknown Primary

Page 21

characteristically seen in small cell carcinoma and sometimes in other neuroendocrine

tumors. In fact, it is so characteristic of small cell carcinoma that I am very hesitant to
diagnose a tumor as a small cell carcinoma unless I can see perinuclear dots on the
cytokeratin stain, highlighted to best advantage in most cases with CK-lmw. On a few
occasions I have noticed a similar pattern of reactivity in non-neuroendocrine tumors,
including granulosa cell tumors of the ovary, poorly differentiated (insular or primordial)
carcinoma of the thyroid gland, and microglandular adenocarcinoma of the pancreas.
Some renal oncocytomas will also show prominent cytoplasmic globs of reactivity on
CK-lmw stains. Suster et al (120) have reported similar findings in 80% of mediastinal
seminomas and 20% of testicular seminomas. Prominent globs of cytokeratin have been
described in paraganglioma of the cauda equina (but not those outside the spinal canal) and
in gangliocytomas of the pituitary (122). Additionally, perinuclear dots of low molecular
weight cytokeratin are a common pattern of aberrant reactivity observed in sarcomas that
express this marker (121). A small number of melanomas may express CK-lmw, and
indeed I have seen several melanomas express an impressive amount of CK-lmw, although
these cases have been negative with CK (AE1/AE3).

In at least one of these cases,

perinuclear dots or globs of cytokeratin was observed.

Cytokeratin 5 or 5/6 (cytoplasmic reactivity): Cytokeratin 5 antibodies are

diagnostically equivalent to cytokeratin 5/6 antibodies. These CK antibodies are very
useful because they typically stain squamous carcinomas strongly and diffusely. Many
different types of tumors contain scattered cytokeratin 5 positive cells (particularly if they
show focal areas of squamous differentiation). However, focal reactivity with cytokeratin
5 has relatively little diagnostic significance. However, when expressed in a strong diffuse
fashion it can be used as a marker of squamous differentiation (providing that
mesothelioma has been excluded) when trying to assess the nature of poorly differentiated
carcinoma.

It

also

stains

basaloid

squamous

carcinomas

and

papillary

squamotransitional carcinomas of the uterine cervix, so it can be very useful in

identifying these unusual variants of squamous carcinoma.

Other tumors that

characteristically show strong staining with CK5 include cutaneous basal cell carcinoma

Miller

IHC for Carcinoma of Unknown Primary

Page 22

and lymphoepitheliomas of the nasopharynx and other sites (which represent poorly
differentiated variants of squamous carcinoma), and thymoma. We have also observed
strong CK 5 in a number of cases of metaplastic or sarcomatoid breast carcinoma
(similar to strong p63 staining in these tumors). The only tumor with glandular features
that may show strong diffuse staining with CK5 is epithelial mesothelioma, and this fact
can be exploited in the diagnosis of mesothelioma. (Parenthetically, it is worthwhile to
mention that the mesothelial-related marker calretinin is also commonly expressed in
pulmonary squamous carcinomas, in up to 40% or 50% of cases.) Remember that p63 also
strongly stains squamous carcinomas, but unlike p63, cytokeratin 5 does not show strong
and diffuse staining in urothelial (transitional cell) carcinomas (in the absence of overt
squamous differentiation). Therefore, in the appropriate clinical context a carcinoma
that is strongly p63 positive but negative or weak for cytokeratin 5 is likely to
represent urothelial (transitional cell) carcinoma. (Parenthetically, I should mention
that I have seen a case of squamous carcinoma in situ of the eyelid that was essentially
negative for cytokeratin 5 but was strongly positive for nuclear p63, although most
squamous proliferations express both cytokeratin 5 and p63). Some authors have utilized
cytokeratin 5 in a fashion similar to high molecular weight cytokeratin for the
interpretation of difficult prostate needle biopsies (158).
Use of High and Low Molecular Weight Cytokeratin (cytoplasmic reactivity):

In

some instances, the staining results with the combination of low molecular weight
cytokeratin (keratins 8 and 18) and high molecular weight cytokeratin (employing clone
34E12) can be useful in the problem of metastatic carcinoma of unknown origin. Certain
tumors tend to strongly express both of these keratins, including carcinomas of the breast,
ovary, pancreas, bladder, stomach, and, non-squamous non-small cell carcinomas of the
lung.

However, hepatoma, "typical" clear cell carcinoma of the kidney, and

adrenocortical carcinomas characteristically lack expression of high molecular weight

cytokeratin, or express it so focally and weakly as to be forgettable. Prostate carcinoma
also typically lacks CK-HMW or expresses it only focally. Therefore, if you have a tumor
that expresses strong high molecular weight cytokeratin, you can place those possibilities
(i.e. typical renal cell carcinoma, hepatoma, adrenocrotical carcinoma, and prostate

Miller

IHC for Carcinoma of Unknown Primary

Page 23

adenocarcinoma) way down on the list of potential primary sites. Essentially all squamous
carcinomas express strong high molecular weight cytokeratin.

Well-differentiated

squamous carcinomas express low molecular weight cytokeratin weakly or not at all.
However, poorly differentiated squamous carcinomas may express significant low
molecular weight cytokeratin, but it is almost always less than (in a few cases equal to) the
degree of expression of high molecular weight cytokeratin. Because of this feature of
squamous carcinoma, if you have a tumor that expresses strong low molecular weight
cytokeratin but expresses high molecular weight cytokeratin weakly or not at all, you
are not dealing with a squamous carcinoma, and other possibilities should be
considered. Conversely, if you have a tumor that expresses substantially more high
molecular weight cytokeratin than low molecular weight cytokeratin, squamous
carcinoma should be considered.

Cytokeratin 17 (cytoplasmic reactivity): CK17 is expressed in a wide variety of

carcinomas,

including

pancreas

(58%

positive),

squamous

carcinoma

(75%),

cholangiocarcinoma (38%), ovarian serous carcinoma (73%), lung adenocarcinoma (23%),

urothelial carcinoma (70%), and endometrial carcinoma (13%) (151). This antibody can be
helpful in some situations, as it is reported to be negative (i.e., < 1% of tumor cells
positive) in stomach carcinoma, colon carcinoma, kidney carcinoma, hepatoma, prostate
carcinomas, mesotheliomas, breast lobular carcinomas, and most breast ductal carcinomas
(92%).

For a number of years I have been using it to attempt to separate gastric

adenocarcinoma (ideally cytokeratin 17 negative) from pancreatic adenocarcinoma (58%

positive). However, I have seen a number of cases of gastric carcinoma that have shown
strong expression of cytokeratin 17, so I have doubts about its true utility for this
differential diagnosis.

Vimentin (cytoplasmic reactivity): In certain situations, vimentin can be a useful

marker to assess the most likely primary site of a tumor (147). Although vimentin is
widely expressed, there are certain tumors that are characteristically vimentin positive and
others that are usually vimentin negative.

When dealing with tumors having

Miller

IHC for Carcinoma of Unknown Primary

Page 24

endometrioid morphology in a female, the differential diagnosis often includes

endometrial, ovarian, endocervical, or colonic origin. Vimentin is usually positive in
endometrial carcinomas (80%), and is positive in about 30% of ovarian endometrioid
adenocarcinomas.

In contrast, vimentin is negative or only focally positive in

adenocarcinomas arising from the colon or endocervix. Hepatomas are characteristically

negative for vimentin, and embryonal carcinomas are usually negative or only focally
positive. Adenocarcinomas of the colon, pancreas, gallbladder, and prostate are usually
vimentin negative. Transitional cell carcinomas and pancreatic carcinomas are also usually
negative. Adenocarcinomas of the thyroid and kidney are almost always vimentin
positive (except for chromophobe renal cell carcinoma, which is vimentin negative).
Vimentin is variable in adenocarcinomas of breast (12% positive), stomach (33% positive),
lung (42% positive), and ovary (44% positive). Small cell carcinomas, carcinoids, and
paragangliomas can express vimentin, but islet cell tumors are reported to be generally
negative (although I cannot confirm this with my personal experience).

An important

point to keep in mind is that virtually any type of spindle cell carcinoma also expresses
strong vimentin.

NEUROENDOCRINE MARKERS

Chromogranin and Synaptophysin (cytoplasmic reactivity) (160-162): These markers

of neuroendocrine differentiation should always be used together, since it is not uncommon
for a neuroendocrine tumor to lack one or the other of these markers. If adenocarcinomas
contain only occasional scattered positive cells, it is best not to label these as
neuroendocrine carcinomas, but rather to consider them as carcinomas that happen to
contain scattered neuroendocrine cells, which is probably of doubtful significance in most
instances. As previously mentioned, the identification of small perinuclear cytokeratin dots
may be the first hint that one is dealing with a neuroendocrine tumor, and should prompt
immunostains for neuroendocrine markers.

Miller

IHC for Carcinoma of Unknown Primary

Page 25

CD56 (NCAM) (membrane reactivity) (160-162): This marker does not have the
specificity for neuroendocrine lesions that chromogranin and synaptophysin do, but it is
more sensitive for neuroendocrine differentiation in some instances. This is particularly
true in the identification of neuroendocrine differentiation in small cell carcinoma, where it
shows strong and diffuse cytoplasmic membrane staining in nearly all cases. For this
reason, I also add this marker to the "neuroendocrine panel" in selected cases, particularly
if small cell carcinoma is in the morphologic differential diagnosis.

Neuron specific enolase (NSE) (cytoplasmic reactivity): The utility of this putative
neuroendocrine marker in the diagnosis of tumors (neuroendocrine or otherwise) can be
summarized briefly: It is worthless.

Those pathologists who use this marker as a

reflection of neuroendocrine differentiation do so at their own peril.

MISCELLANEOUS MARKERS
CEA (clone COL1) (cytoplasmic or membrane reactivity); Polyclonal CEA antibodies
are to be avoided unless you are trying to demonstrate canalicular reactivity in
hepatocellular carcinoma. I prefer the COL1 clone (Zymed) of CEA, as it shows no
reactivity with non-specific cross reacting antigen (NCA), and it performs very well (and I
have never seen a mesothelioma that has shown even a single positive cell with this
clone!). Parenthetically, if you see staining of neutrophils on a CEA immunostain, that
CEA antibody is cross reacting with NCA, and you would be well served to find another
CEA antibody that does not cross-react with NCA. Although monoclonal CEA is positive
in many types of adenocarcinomas, it should be negative in renal cell carcinoma, adrenal
carcinoma, and typical papillary or follicular thyroid carcinoma (although it may be
positive in areas of squamous differentiation) and should be negative in the cytoplasm of
hepatomas. However, I have seen a small number of cases where CEA (COL1) has shown
a beautiful canalicular pattern in hepatoma, identical to that described for polyclonal CEA,
villin, and CD10 antibodies, so that fact should be kept in mind. Medullary thyroid
carcinomas are always strongly and diffusely positive for CEA, expressing it in nearly
100% of cells. On several occasions, this finding in a TTF-1 positive tumor has given us

Miller

IHC for Carcinoma of Unknown Primary

Page 26

the first clue to the diagnosis, and allowed us to correctly diagnose a medullary thyroid
carcinoma in the setting of a metastatic tumor of unknown primary.

Most prostate

carcinomas and endometrial adenocarcinomas are CEA negative, although they may show
patchy areas of positivity. CEA can also be useful in the workup of tumors with an
endometrioid morphology, since CEA is positive in the vast majority of endocervical
adenocarcinomas (65-95%) and colonic adenocarcinomas with an endometrioid
morphology (90%), but is negative or only focally positive in the vast majority of
endometrioid adenocarcinomas arising in the ovary or endocervix . Clone Z3 of CEA has
been reported to be useful in separating the tall cell variant of papillary carcinoma of the
thyroid (CEA+ and also CD15+) from usual papillary carcinoma of the thyroid (CEA-,
CD15-) (although I have no personal experience with this) (173).

CDX2 (nuclear reactivity (163-170): This marker has been found to be positive in a
very high percentage of gastrointestinal adenocarcinomas, particularly those from the
colon and duodenum. In one study, (169), CDX2 was present in 188 of 189 (99%) of
colonic and duodenal adenocarcinomas.

Gastric and pancreatic carcinomas showed

heterogeneous expression, but no reactivity was noted in hepatocellular carcinomas or in

carcinomas from the urinary tract (except urinary bladder adenocarcinoma), female genital
tract (except for mucinous ovarian tumors), breast, lung, and head and neck. Barbareschi
et al noted similar findings (165), with CDX2 staining 98% (88/90) of colorectal
adenocarcinomas, as well as 55% (11/20) of gastric carcinomas, 60% (3/5) of pancreatic
carcinomas, 60% (3/5) of gallbladder carcinomas, and 100% (5/5) of mucinous ovarian
tumors. They did not observe CDX2 in 117 lung cancers of different types, nor in cancers
of the breast (n=30), prostate (n=8), mesothelioma (n=5), thyroid carcinoma (n=4), kidney
carcinoma (n=5), and ovarian serous carcinoma (n=5). In a tissue microarray study (166)
where a positive CDX2 stain was defined as reactivity in >10% of nuclei, positivity was
found in 84% of 1288 colonic adenocarcinomas, 29% of 45 gastric carcinomas of intestinal
type, 12% of 26 gastric carcinomas of diffuse type, 10.5% of 19 mucinous ovarian
carcinomas, 9.3% of 43 endometrial carcinomas, 2% of 49 serous carcinomas, 2% of 48
lung squamous carcinomas, and 2% of 89 bladder carcinomas. Negative tumors included

Miller

IHC for Carcinoma of Unknown Primary

Page 27

50 pancreatic carcinomas, 27 cholangiocarcinomas, 48 hepatocellular carcinomas, 49

pulmonary adenocarcinomas, 48 pulmonary large cell carcinomas, 48 pulmonary small cell
carcinomas, 112 renal cell carcinomas, 93 prostate carcinomas, 153 breast carcinomas, 94
oral cancers, 103 thyroid carcinomas, and 42 carcinomas of the uterine cervix.

Another

recent study (167) reported CDX2 reactivity in 13/13 colonic adenocarcinomas, 2/10
pancreatic carcinomas, 9/12 gastric carcinomas, 9/11 mucinous ovarian carcinomas, 0/5
non-mucinous ovarian carcinomas, 4/5 esophageal adenocarcinomas, 1/10 endometrial
carcinomas, 2/12 pulmonary adenocarcinomas, and 0/22 breast carcinomas. At ProPath,
our findings have been similar to those reported above. Like the last study mentioned, we
have identified significant CDX2 immunoreactivity in several unequivocal pulmonary
carcinomas, and also some neuroendocrine tumors (GI carcinoids, particularly those of
midgut origin, islet cell tumor of pancreas, and large cell neuroendocrine carcinoma), and
some urothelial carcinomas of bladder. Morules that may be present in endometrioid
adenocarcinomas also stain strongly with CDX2. In a sense, it seems reasonable to employ
it in a fashion similar to villin for the identification of GI primary tumors, although I
suspect as more laboratories use this antibody, we will find its expression in other non-GI
adenocarcinomas.

N-Cadherin (membrane reactivity): N-cadherin is a protein involved in intercellular

adhesion. Antibodies to the protein are useful in the workup of metastatic carcinomas in
females, since they have been found to stain a high percentage of serous and
endometrioid carcinomas of the female genital tract, although mucinous ovarian
carcinomas are negative (178). Mesotheliomas also frequently express this marker, and it is
not unusual for lung carcinoma to show expression. The literature on this marker is rather
scant, so its complete spectrum or reactivity is not well defined.

I have personally

observed this marker in a number of other tumors (and non-neoplastic tissues), including
hepatocellular carcinoma, renal cell carcinoma (both papillary and conventional clearcell types), seminoma, yolk sac tumor, thymoma, melanoma (focal), liver adenoma,
ganglioneuroma, glial tissue, thymic carcinoid, medullary carcinoma of thyroid,
extraskeletal myxoid chondrosarcoma, thyroid papillary carcinoma, thyroid follicular

Miller

IHC for Carcinoma of Unknown Primary

Page 28

adenoma, solitary fibrous tumor, endometrial stromal sarcoma, schwannoma, adrenal

adenoma, desmoplastic small cell tumor, MFH, PNET, chordoma, and nerve fibers in the
myenteric plexus of the bowel wall. In addition to being strongly expressed by normal
hepatocytes, N-cadherin is also expressed by normal benign bile ducts. As such, I think
that this marker may have some potential utility in distinguishing cholangiocarcinoma
(which might be expected to express strong N-cadherin) from some of its mimics, although
at the time of this writing I am not aware of any published literature that addresses this
issue. Parenthetically, there are a few papers that have touted parathyroid hormone-related
protein (PTH-rP) as a marker of cholangiocarcinoma, but after using that marker for this
purpose for a number of years, I abandoned it because of my perception of poor sensitivity
in the recognition of cholangiocarcinoma. To my knowledge there are no good markers
available at this time that are specific for cholangiocarcinoma.

HMBE-1 (membrane or cytoplasmic reactivity): HBME-1 is an antibody to a

mesothelioma cell line. I have not found it to be useful in the diagnosis of mesothelioma,
but before the advent of Pax8, I employed it for attempting to distinguish breast carcinomas
(HMBE-1 negative in 90% of breast cancers) versus female genital tract carcinomas
(often HBME-1 positive) (134).

HMBE-1 is reported to be consistently positive in

ovarian, endometrioid, and thyroid tumors, but is only rarely reported in tumors of the
colon, bladder, and kidney. Also, HBME-1 can be very useful in the recognition of the
follicular variant of papillary thyroid carcinoma, since it generally stains those tumors quite
strongly (similar to cytokeratin 19). HBME-1 has also been reported to stain almost all
chordomas, which can be useful in their differential diagnosis with chondrosarcomas,
which are HBME-1 negative (177).

BCL-2 (cytoplasmic or membrane reactivity): Alsabeh and colleagues (175) published

a paper in 1996 that studied the use of BCL-2 to aid in the distinction of breast carcinoma
(79% positive) from lung carcinoma (5.6% positive) and gastric carcinoma (8.3%
positive). In addition, there were also significant differences in the intensity of staining
with this marker. 70% of the breast carcinomas were moderately to intensely positive,

Miller

IHC for Carcinoma of Unknown Primary

Page 29

whereas only 1.9% of the lung carcinomas and only 0.9% of the gastric carcinomas showed
moderate to intense reactivity for BCL-2. I have seen BCL-2 positivity in female genital
tract tumors, thyroid tumors, neuroendocrine tumors, renal oncocytomas, and melanomas.
In lung carcinoma, some authors have noted an association of BCL-2 positivity with
neuroendocrine differentiation, in that most small cell carcinomas appear to be BCL-2
positive (137). BCL-2 may also have a role in helping to separate basal cell carcinoma of
the skin (BCL-2 positive) from squamous tumors (BCL-2 negative) (172), although
personally I have not been particularly impressed with its utility in this situation.

Inhibin, Calretinin, and A103 for Adrenal Tumors (cytoplasmic reactivity) (179-190):
These 3 markers have found utility in the recognition of adrenal cortical tumors,
including adrenal cortical carcinoma. They are also commonly expressed in sex cordstromal tumors of the genital tract, and calretinin is well known as a mesotheliomaassociated marker.

Some endometrioid adenocarcinomas of the ovary may resemble

Sertoli cell tumors quite closely, and since inhibin stains the tubules of Sertoli cell tumors
(and not the glands of endometrioid adenocarcinomas), it can be very useful in this
differential diagnostic problem.

(Parenthetically, it is worthwhile mentioning that

calretinin is commonly expressed in squamous carcinomas in addition to

mesothelioma.)

CD10 (CALLA) (cytoplasmic or membrane reactivity): CD10 (CALLA) has been

found to be expressed in a high percentage of renal cell carcinoma, and since it is
typically absent in adrenal carcinoma, it can be of use in the distinction of those two
tumors. However, its specificity is poor, since it stains a significant percentage of nonrenal tumors as well (191). CD10 is also useful in the diagnosis of hepatocellular
carcinoma, since it is one of several antibodies (along with polyclonal CEA and villin)
that may highlight a diagnostically useful canalicular pattern of reactivity in hepatoma.
Recent studies (192) have shown that CD10 is positive in mesonephric remnants and
tumors, but is negative in clear cell carcinomas of gynecologic origin, so this finding
can assist in the differentiation from clear cell carcinomas of renal origin.

Miller

IHC for Carcinoma of Unknown Primary

Page 30

Thrombomodulin (membrane reactivity): Collins et al (193) have studied the

expression of thrombomodulin, a cell surface glycoprotein, in a variety of tumors. They
found that a high percentage (91%) of transitional cell carcinomas of the urinary tract
expressed this marker, as did a high percentage (87%) of squamous carcinomas.
However, expression of thrombomodulin in other types of adenocarcinoma was much less
common, being found in 15% of lung adenocarcinoma, 17% of bladder adenocarcinoma,
19% of breast carcinoma, and 8% of ovarian carcinoma. An exception was pancreatic
carcinoma, which stained 1-25% of cells in 2 of 4 cases studied. There was no staining of
prostate carcinoma, endometrial adenocarcinoma, renal cell carcinoma, or colon
adenocarcinoma. This marker will also stain some histiocytes, normal endothelial cells,
some angiosarcomas, some trophoblastic tumors, and between 40-100% of epithelial
mesotheliomas. I have also observed this marker in yolk sac tumor, basal cell carcinoma,
PNET, and in basement membranes of seminiferous tubules. Personally, in my experience
p63 is better than thrombomodulin as a marker of urothelial carcinoma, and I do not find
thrombomodulin to be of great utility in my practice.

Uroplakin (cytoplasmic or membrane reactivity) (194-197): This marker has been

touted as useful for recognition of transitional cell tumors. Although its sensitivity is
modest, it is reported to be highly specific, and was not found in non-urothelial
carcinomas. I obtained this antibody several years ago, and in my experience its sensitivity
is low (and the vendors tech support was highly unpleasant, to say the least!), so I have not
found this particular antibody to be of great utility in my practice. Parker et al (197)
reported that uroplakin III was expressed in 57% of 112 urothelial carcinomas, but in none
of 498 non-urothelial carcinomas present in a tissue microarray.

CA-125 (cytoplasmic reactivity or membrane): According to Dr. Mark Wick, CA-125

is positive in Mullerian tumors and about 50% of biliary tract and pancreatic tumors. It is
also reported to be positive in clear cell carcinoma of the bladder (203). In my laboratory, I
have also seen CA-125 in amnionic epithelium, several cases of lung adenocarcinoma,

Miller

IHC for Carcinoma of Unknown Primary

Page 31

focal staining in colonic adenocarcinoma, and strong reactivity in normal reactive

mesothelial cells. I also observed rare positive cells in a case of high-grade transitional cell
carcinoma. Goldstein (204) reports that CA-125 is negative in prostate carcinoma. I have
not found this marker to be particularly useful in the types of cases that I see in my
consultation practice.

Mesothelin (clone 5B2) (cytoplasmic or membrane reactivity): This antibody was

originally raised to a mesothelioma cell line, and it stains about 55% of mesotheliomas.
Dr. Allen Gown found this marker also stains a very high percentage (approaching 100%)
of serous adenocarcinomas of female genital tract origin (unpublished observations) but
few other carcinomas, except for squamous carcinoma. A large study by Dr. Nelson
Ordonez (205) using clone 5B2 described the expression of this marker in a wide variety of
tumors. I have not found it to be particularly useful in my practice. There is some data to
suggest that expression of mesothelin favors pancreatic carcinoma when the differential
diagnosis is pancreatic carcinoma vs. reactive atypia in pancreatic epithelial cells.
However, the combination of Placental S100 (S100P) and pVHL stains are better suited to
that differential diagnosis.

CA19-9 (cytoplasmic reactivity): In my own experience I have found CA19-9 to have

very limited utility. Gatalica and Miettinen (198) examined a large series of tumors, and
found that it was positive in the majority of gastrointestinal and pancreatic carcinomas (7090%), but it also was positive in a significant proportion of other tumors as well, including
bladder tumors (64%+), lung adenocarcinomas (45%+), and up to 30% of breast ductal
carcinomas. 80% of mucoepidermoid tumors and 60% of adenoid cystic carcinomas of
salivary glands were also positive. The majority of prostate carcinomas (88%) and kidney
carcinomas (83%) were negative, and consistently negative tumors included hepatoma,
lobular breast carcinomas, GI carcinoids, islet cell tumors, mesothelioma, melanoma,
MFH, seminoma, small cell lung cancer, and squamous lung cancer. These investigators
found that CA19-9 was negative in follicular carcinoma of the thyroid, but positive in 71%
of papillary carcinomas of the thyroid, so perhaps it might be of some help in making that

Miller

IHC for Carcinoma of Unknown Primary

distinction.

Page 32

However, other authors have reported up to 24% of thyroid follicular

carcinomas positive with CA19-9, and I do not have enough personal experience at the
time of this writing with this marker in that situation to render an opinion on its utility in
this circumstance.

APPROACH TO THE INDIVIDUAL CASE

When faced with an individual case, there are several questions that we should always
try to ask ourselves, and the answers to these questions may be very easy or very difficult
to come by.

1. Did the clinicians provide us with any useful history (and if we suspect we
are missing important history, have we done the appropriate things to try to obtain
this information)? Obviously, pertinent history can be very valuable to us as we approach
individual cases, but unfortunately we all know clinicians who fail to provide us with this
information. In fact, in the past when I was still practicing in a hospital lab, a clinician
once told me that he specifically withholds information from pathologists, so as to not bias
their opinion. In my unbiased opinion, such clinicians are fools who are only hurting the
interests of their patients.

2. Is the tumor really carcinoma, or could it be something else? I have been

surprised countless times by tumors that I thought were carcinoma on H&E (see the list of
"epithelioid tumors" in the accompanying IHC peripheral brain excerpt), so before we start
trying to find a primary, it is important to make sure that we really are dealing with a
carcinoma (and not melanoma, germ cell tumor, lymphoma, sarcoma etc.).

In these

situations, it may be necessary to perform an appropriate screening battery to get some idea
of the true cell lineage of the neoplasm in question, before pulling out all the stops. As a
practical matter, I am personally unwilling to exclude carcinoma until I see negative stains
with CK-lmw and CK-hmw (and sometimes EMA) (although in some instances noncarcinomas may also show reactivity with these markers, a topic beyond the scope of this
presentation). In addition, the possibility of radiation or chemotherapy effect should also

Miller

IHC for Carcinoma of Unknown Primary

Page 33

be kept in mind, as in the absence of appropriate history, it is easy to misinterpret radiation

or chemotherapy effect as malignancy.

3. If it is a carcinoma, what kind is it (adenocarcinoma vs. squamous

carcinoma vs. transitional cell carcinoma vs. hepatoma vs. neuroendocrine, etc). By
knowing the spectrum of reactivity of the markers discussed above, these questions can be
addressed in a successful fashion.

4. Where could this tumor be arising? Each case is different, so the antibody
panel that I use is not the same for every case, and obviously the gender of the patient and
the H&E morphology guide my selection of antibodies.

I have never been particularly

successful with the algorithmic approach to metastatic carcinoma that has been advocated
by some authors, since I find that the published algorithms are not easily applied to the
large variety of situations that we see, and the demands for rapid turnaround time often
make the application of a sequential algorithmic approach impractical.

For most metastatic carcinomas, I find that the combination of cytokeratin 7,

cytokeratin 20, and villin generally allows me to narrow the possibilities substantially, so
that is a good place to start for nearly all cases. In my own practice, I use voice recognition
software on my PC (Dragon NaturallySpeaking) for all reports, that allows large
paragraphs or lists of antibodies to be inserted into reports by uttering only a few command
words. Therefore, when I get a carcinoma of unknown origin, I dictate my standard huge
carcinoma panel (which is a list of about 30-40 antibodies) into the report, and then I go
down the list and briefly think about each antibody on the list and whether or not it would
have utility in the particular case under study. Those antibodies that are of no use are
rapidly deleted from the list. Similar to using the peripheral brain to assist my imperfect
memory, the use of these lists has helped me on many occasions to think about differential
diagnostic possibilities that had previously escaped me, and to remember to order
important antibodies that I would have otherwise forgotten.

Miller

IHC for Carcinoma of Unknown Primary

Page 34

It is a very difficult task to adequately cover the topic of immunohistochemistry in

metastatic carcinoma in the time allotted, so it will be impossible for me to cover
everything in this handout. However, I hope that the information presented will be of use
to those in the audience who struggle with these cases as I do.
REFERENCES
Prostate Markers: PSA, PSAP, PSMA, and NKX3.1
1. Federspiel BH, Burke AP, Sobin LH et al: Rectal and colonic carcinoids. A clinicopathologic
study of 84 cases. Cancer 65(1): 135-140, 1990.
2. Kamoshida S, Tsutsumi Y: Extraprostatic localization of prostatic acid phosphatase and
prostate-specific antigen: Distribution in cloacogenic glandular epithelium and sex-dependent
expression in human anal gland. Hum Pathol 21: 1108-1111, 1990.
3. Azumi N, Traweek ST, Battifora H: Prostatic acid phosphatase in carcinoid tumors.
Immunohistochemical and immunoblot studies. Am J Surg Pathol 15(8): 785-790, 1991.
SantAgnese, PAD: Neuroendocrine differentiation in human prostatic carcinoma. Hum Pathol
23: 287-296, 1992.
4. Sidhu J, Sanchez RL: Prostatic acid phosphatase in strumal carcinoids of the ovary. An
immunohistochemical study. Cancer 72(5): 1673-1678, 1993.
5. Van Krieken JHJM: Prostate marker immunoreactivity in salivary gland neoplasms. A rare
pitfall in immunohistochemistry. Amer J Surg Pathol 17(4): 410-414, 1993.
6. Burke AP, Thomas RM, Elsayed AM et al: Carcinoids of the jejunum and ileum: An
immunohistochemical and clinicopathologic study of 167 cases. Cancer 79(6): 1086-1093, 1997.
7. Murphy GP et al: Current evaluation of the tissue localization and diagnostic utility of prostate
specific membrane antigen. Cancer 83 (11): 2259-2269, Dec 1, 1998.
8. Chang SS et al: Comparison of anti-prostate-specific membrane antigen antibodies and other
immunomarkers in metastatic prostate carcinoma. Urology 57(6): 1179-1113, Jun 2001.
9. Silver DA et al: Prostate-specific membrane antigen expression in normal and malignant
human tissues. Clin Cancer Res 3(1): 81-85, Jan 1997.
10. Ross JS et al: Prostate specific membrane antigen (PSMA) expression in non-prostate
cancers.Mod Pathol 17 (Supplement 1): 326A (abstract 1373), Jan 2004.
11. Fan CY et al: Expression of androgen receptor and prostatic specific markers in salivary duct
carcinoma: an immunohistochemical analysis of 13 cases and review of the literature. Am J
Surg Pathol 24(4): 579-586, Apr 2000.
12. Sheridan T, Herawi M, Epstein JL: The role of P501S and PSA in the diagnosis of metastatic
adenocarcinoma of the prostate. Am J Surg Pathol 31(9): 1351-1355, Sep 2007.
12a. Gurel B, Ali TZ, Montgomery EA et al: NKX3.1 as a marker of prostatic origin in metastatic
tumors. Am J Surg Pathol 34(8): 1097-1105, Aug 2010.
Breast markers: GCDFP-15 and Mammaglobin
13. Mazoujian G, Pinkus GS, Davis S et al: Immunohistochemistry of gross cystic disease fluid
protein (GCDFP-15) of the breast. A marker of apocrine epithelium and breast carcinomas with
apocrine features. Am J Pathol 110:105-112, 1983.
14. Wick MR, Lillemoe TJ, Copland GT et al: Gross cystic disease fluid protein-15 as a marker
for breast cancer: Immunohistochemical analysis of 690 human neoplasms and comparison with
alpha-lactalbumin. Hum Pathol 20: 281-287, 1989.

Miller

IHC for Carcinoma of Unknown Primary

Page 35

15. Swanson PE, Pettinato G, Lillemoe T et al: Gross cystic disease fluid protein-15 in salivary
gland tumors. Arch Pathol Lab Med 115(2): 158-163, 1991.
16. Raab SS, Berg LC, Swanson PE et al: Adenocarcinoma in the lung in patients with breast
cancer. A prospective analysis of the discriminatory value of immunohistology. American Journal
of Clinical Pathology 100:27-35, 1993.
17. Han JH, Shin HC, Kim HS et al: Mammaglobin expression in lymph nodes is an important
marker of metastatic breast carcinoma. Arch Pathol Lab Med 127(10): 1330-1334, Oct 2003.
18. Sjodin A, Guo D, Henriksson R et al: Mammaglobin in normal human sweat glands and
human sweat gland tumors. J Invest Dermatol 121(2): 428-429, Aug 2003.
19. Gown AM et al: Mod Pathol 15 (Supplement 1): (abstract 223), Jan 2002.
20.
Dabbs
Mod Pathol 19 (Supplement 1): (abstract 88), Jan 2006.
21.
Gown et al: Mod Pathol 19 (Supplement 1): 128A (abstract 189), Jan 2006.
22.
Mod Pathol 19 (Supplement 1): 128A (abstract 886), Jan 2006.
23. . Ciampa A, Fanger G, Khan A et al: Mammaglobin and CRxA-01 in pleural effusion
cytology: potential utility of distinguishing metastatic breast carcinomas from other
cytokeratin7-positive/cytokeratin 20-negative carcinomas. Cancer 102(6):368-372, Dec
2004.
TTF-1 and PE-10
24. Nicholson AG, McCormick CJ, Schimosato T et al: The value of PE-10, a monoclonal
antibody against pulmonary surfactant, in distinguishing primary and metastatic lung tumours
Histopathol 27: 57-60, 1995.
25. Bejarano PA et al: Surfactant proteins and thyroid transcription factor-1 in pulmonary and
breast carcinomas. Mod Pathol 9(4): 445-452, 1996.
26. Langel DJ, Gown AM, Astion ML et al: Monoclonal antibody to surfactant apoprotein A: A
highly specific marker of pulmonary adenocarcinoma. (abstract) Am J Clin Pathol 108(1):108-109,
1997.
27. DiLoreto C, DiLauro V, Puglisi F et al: Immunocytochemical expression of tissue specific
transcription factor-1 in lung carcinoma. J Clin Pathol 50: 30-32, 1997.
28. Anderson SS et al: Thyroid transcription factor-1 (TTF-1) is a more sensitive lung carcinoma
marker than surfactant APO A1 (abstract). Mod Pathol 11(1): 171A, 1998.
29. Folpe AL et al: Thyroid transcription factor-1: Immunohistochemical evaluation of pulmonary
neuroendocrine tumors. Mod Pathol 12(1): 5-8, 1999.
30. Anwar F, Schmidt RA: Thyroid transcription factor-1 (TTF-1) distinguishes mesothelioma
from pulmonary adenocarcinoma. (abstract #1061) Mod Pathol 12(1): 181, 1999
31. Gown AM, Rosai J: TTF-1, a marker of lung and thyroid carcinomas, is not expressed in
thymic epithelial tumors. (abstract #1075) Mod Pathol 12(1): 183A, 1999.
32. Kaufmann O, Dietel M: Thyroid transcription factor-1 is the superior immunohistochemical
marker for pulmonary adenocarcinomas and large cell carcinomas compared to surfactant proteins
A and B. Histopathol 36: 8-16, 2000.
33. Ordonez N: Thyroid transcription factor-1 is a marker of lung and thyroid carcinomas.
Advances in Anatomic Pathology 7 (2): 123-127, 2000.
34. Katoh R et al: Expression of thyroid transcription factor-1 (TTF-1) in human C-cells and
medullary thyroid carcinomas. Hum Pathol 31: 386-393, 2000.
35. Agoff SN et al: Thyroid transcription factor-1 is expressed in extrapulmonary small cell
carcinomas but not in other extrapulmonary neuroendocrine tumors. Mod Pathol 13 (3): 238-242,
2000.
36. Kaufman O, Dietel M: Expression of thyroid transcription factor-1 in pulmonary and
extrapulmonary small cell carcinomas and other neuroendocrine carcinomas of various primary
sites. Histopathol 36: 415-420, 2000.

Miller

IHC for Carcinoma of Unknown Primary

Page 36

37. Byrd-Gloster A, Khoor AL, Glass LF et al: Differential expression of thyroid transcription
factor-1 in small cell lung carcinoma and Merkel cell tumor. Hum Pathol 31: 58-62, 2000
38. Ordonez, NG: Value of thyroid transcription factor-1, E-cadherin, BG-8, WT-1, and CD44S
immunostaining in distinguishing epithelial pleural mesothelioma from pulmonary and nonpulmonary adenocarcinoma. Am J Surg Pathol 24 (4): 598-606, 2000.
39. Ordonez NG: Value of thyroid transcription factor-1 immunostaining in distinction of small
cell lung carcinomas from other small cell carcinomas. Am J Surg Pathol 24 (9): 1217-1223,
2000.
40. Cheuk W, Chan JKC: Thyroid transcription factor-1 is of limited value in practical distinction
between pulmonary and extrapulmonary small cell carcinomas. (Letter to the editor) Am J Surg
Pathol 25 (4): 545-546, 2001.
41. Oliveira AM et al.: Thyroid transcription factor-1 distinguishes metastatic pulmonary from
well-differentiated neuroendocrine tumors of other sites. Am J Surg Pathol 25 (6): 815-819,2001
42. Cheuk W et al: Immunostaining for thyroid transcription factor 1 and cytokeratin 20 aids the
distinction of small cell carcinoma from Merkel cell carcinoma, but not pulmonary from
extrapulmonary small cell carcinomas. Arch Pathol Lab Med 125: 228-231, 2001.
43. Goldstein NS et al: Mucinous and non-mucinous bronchioloalveolar adenocarcinomas have
distinct staining patterns with thyroid transcription factor and cytokeratin 20 antibodies. Am J
Clin Pathol 116: 319-325, 2001.
44. Lau SK et al: Expression of thyroid transcription factor-1, cytokeratin 7, and cytokeratin 20 in
bronchoalveolar carcinomas: An immunohistochemical evaluation of 67 cases. Mod Pathol 15 (5):
538-542, 2001.
45. Wieczorek TJ, Pinkus JL, Glickman JN et al: Comparison of thyroid transcription factor-1 and
hepatocyte antigen immunohistochemical analysis in the differential diagnosis of hepatocellular
carcinoma, metastatic adenocarcinoma, renal cell carcinoma, and adrenal cortical carcinoma. Am J
Clin Pathol 118(6): 911-921, 2002.
46. Nakamura N, Miyagi E, Murata S et al: Expression of thyroid transcription factor-1 in normal
and neoplastic lung tissues. Mod Pathol 15 (10): 1058-1067, 2002.
47. Nicholson AG, Magkou C, Snead D et al: Unusual sclerosing haemangiomas and sclerosing
haemangioma-like lesions, and the value of TTF-1 in making the diagnosis. Histopathology 41
(5): 404-413, 2002.
47a. Robens J, Goldstein L, Gown AM et al: Thyroid transcription factor-1 expression in breast
carcinomas. Am J Surg Pathol 34(12):1881-1885, Dec 2010.
48. Penman D, Downie J, Roberts F: Positive immunostaining for thyroid transcription factor-1
in primary and metastatic colonic adenocarcinoma: a note of caution. J Clin Pathol 59(6):
663-664, Jun 2006.
49. Comperat E, Zhang F, Perrotin C et al: Variable sensitivity and specificity of TTF-1
antibodies in lung metastatic adenocarcinoma of colorectal origin. Mod Pathol 18(10): 13711376, Oct 2005.
50. Pang Y, von Turkovich M, Wu H et al: The binding of thyroid transcription factor-1 and
hepatocyte paraffin 1 to mitochondrial proteins in hepatocytes: a molecular and
immunoelectron microscopic study. Am J Clin Pathol 125(5): 722-726, 2006.
Renal cell carcinoma antigen (RCC) and Pax-2
51. McGregor DK, . M.D.; Khurana KK, Cao CB et al: Diagnosing Primary and Metastatic
Renal Cell Carcinoma: The Use of the Monoclonal Antibody `Renal Cell Carcinoma Marker'.
Am J Surg Pathol . 25(12):1485-1492, Dec 2001.
52. Wang H-Y, Mills SE,: KIT and RCC Are Useful in Distinguishing Chromophobe Renal Cell
Carcinoma From the Granular Variant of Clear Cell Renal Cell Carcinoma. Am J Surg Pathol
29(5):640-646, May 2005.

Miller

IHC for Carcinoma of Unknown Primary

Page 37

53. Mazal PR, Stichenwirth M, Koller A et al: Expression of aquaporins and PAX-2 compared
to CD10 and cytokeratin 7 in renal neoplasms: a tissue microarray study. Mod Pathol 18:535540, 2005.
54. Tong G-X, Melamed J, Mansukhani M et al: PAX2: a reliable marker for nephrogenic
adenoma. Mod Pathol 19: 356-363, 2006.
55. . Mod Pathol 19 (Supplement 1): (abstract 660), Jan 2006.
56. Avery AK, Beckstead J, Renshaw AA: Use of antibodies to RCC and CD10 in the differential
diagnosis of renal neoplasms. Am J Surg Pathol 24 (2): 203-210, 2000.

ER and PR
57. Diaz MN, Mazoujian G, Wick MR: Estrogen-receptor protein in thyroid neoplasms. Arch
Pathol Lab Med 115: 12-3-1207, 1991.
58. Swanson PE, Mazoujian G, Mills SE et al: Immunoreactivity for estrogen receptor protein in
sweat gland tumors. Am J Surg Pathol 15(9): 835-841, 1991.
59. Voytek TM, Ricci A, Cartun RW: Estrogen and progesterone receptors in primary eccrine
carcinoma. Mod Pathol 4(5): 582-585, 1991.
60. Viale G, Doglioni C, Gambacorta M et al: Progesterone receptor immunoreactivity in
pancreatic endocrine tumors. An immunocytochemical study of 156 neuroendocrine tumors of the
pancreas, gastrointestinal and respiratory tracts, and skin. Cancer 70: 2268-2277, 1992.
61. Deamant FD, Pombo MT, Battifora H: Estrogen receptor immunohistochemistry as a predictor
of site of origin in metastatic breast cancer. Appl Immunohistochem 1(3): 188-192, 1993
62. Wallace ML, Longacre TA, Smoller BR: Estrogen and progesterone receptors and anti-gross
cystic disease fluid protein (BRST-2) fail to distinguish metastatic breast carcinoma from eccrine
neoplasms. Mod Pathol 8(9): 897-901, 1995.
63. Keel SB, Koerner FC, Efird JT et al: Estrogen and progesterone receptor status and diseasespecific survival on skull base chordomas. Mod Pathol 9(1): 8A (abstract #24), 1996.
64. Weihing RR et al: Hepatobiliary and pancreatic mucinous cystadenocarcinomas with
mesenchymal stroma: Analysis of estrogen receptors / progesterone receptors and expression of
tumor-associated antigens. Mod Pathol 10(4): 372-379, 1997.
65. Bacchi C, Garcia RL, Gown AM et al: Immunolocalization of estrogen and progesterone
receptors in neuroendocrine tumors of lung, skin, gastrointestinal and female genital tracts. Appl
Immunohistochem 5(1): 17-22, 1997.
66. Hanby AM, McKee P, Jeffrey M et al: Primary mucinous carcinomas of the skin express
TFF1, TFF3, estrogen receptor, and progesterone receptors. Am J Surg Pathol 22(9): 1125-1131,
1998.
67. Kauffmann O, Kother S. Dietel M: Use of antibodies against estrogen and progesterone
receptors to identify metastatic breast and ovarian carcinomas by conventional
immunohistochemical and tyramide signal amplification methods. Mod Pathol 11: 357-363, 1998.
68. Busam KJ et al: Epidermal growth factor, estrogen, and progesterone receptor expression in
primary sweat gland carcinomas and primary and metastatic mammary carcinomas. Mod Pathol
12 (8): 786-793, 1999.
69. Dabbs DJ et al: Immunohistochemical detection of estrogen receptor in pulmonary
adenocarcinomas is dependent upon the antibody used. Mod Pathol 13 (1): 208A (abstract #
1227), 2000.
70. Bayer-Garner I et al: Androgen receptors: A marker to increase sensitivity for identifying
breast cancer in skin metastasis of unknown primary site. Mod Pathol 13 (2): 119-122,2000.
71. DiNunno L, Larsson LG, Rinehart JJ et al: Estrogen and progesterone receptors in non-small
cell lung cancer in 248 consecutive patients who underwent surgical resection. Arch Pathol Lab
Med 124 (10): 1467-70, 2000.

Miller

IHC for Carcinoma of Unknown Primary

Page 38

HepPar1
72. Wennerberg AE, Nalesnik MA, Coleman WB et al: Hepatocyte paraffin 1: A monoclonal
antibody that reacts with hepatocytes and can be used for differential diagnosis of hepatic tumors.
Am J Pathol 143(4): 1050-1054, 1993.
73. Minervani MI, Dimitris AJ, Lee RG et al: Utilization of hepatocyte-specific antibody in the
immunocytochemical evaluation of liver tumors. Mod Pathol 10(7): 686-692, 1997.
74. Fasano M, Theise ND, Nalesnik M et al: Immunohistochemical evaluation of hepatoblastomas
with use of the hepatocyte-specific marker, hepatocyte paraffin 1, and polyclonal anticarcinoembryonic antigen. Mod Pathol 11(10): 934-938, 1998.
75. Stahl J, Voyvodic F: Biopsy diagnosis of malignant versus benign liver nodules: New
helpful markers. An Update. Advances in Anatomic Pathology 7(4): 230-239, 2000.
76. Maitra A, Murakata LA, Albores-Saavedra J: Immunoreactivity for hepatocyte paraffin 1
antibody in hepatoid adenocarcinomas of the gastrointestinal tract. Am J Clin Pathol 115: 689-694,
2001.
77. Wieczorek TJ, Pinkus JL, Glickman JN et al: Comparison of thyroid transcription factor-1 and
hepatocyte antigen immunohistochemical analysis in the differential diagnosis of hepatocellular
carcinoma, metastatic adenocarcinoma, renal cell carcinoma, and adrenal cortical carcinoma. Am J
Clin Pathol 118(6): 911-921, 2002.
78. Chu PG, Ishizawa S, Wu E et al: Hepatocyte antigen as a marker of hepatocellular carcinoma.
An immunohistochemical comparison to carcinoembryonic antigen, , and alpha-fetoprotein. Am J
Surg Pathol 26 (8): 978-988, 2002.
WT1
79. Foster MR, Allred DC, Olson SJ et al: Immunohistochemical analysis of Wilms tumor gene
expression in malignant mesotheliomas and primary pulmonary adenocarcinomas. Mod Pathol
12(1): 182A, 1999. (abstract #1072)
80. Shimizu M, Toki T, Takagi Y et al: Immunohistochemical detection of the Wilms' tumor gene
(WT1) in epithelial ovarian tumors. Int J Gynecol Pathol 19: 158-163,2000.
81. Quenneville L, Yaziji H, Gown AM: WT1 is a sensitive and specific marker of serous, but not
mucinous, ovarian carcinomas. Modern Pathology 14:145A, 2001.
82. Foster MR et al: Immunohistochemical analysis of nuclear versus cytoplasmic staining of
WT1 in malignant mesotheliomas and primary pulmonary adenocarcinomas. Arch Pathol Lab
Med 125 (10): 1316-1320, 2001.
83. Goldstein NS, Uziebo A: WT1 immunoreactivity in uterine papillary serous carcinomas is
different from ovarian serous carcinomas. Am J Clin Pathol 117(4): 541-545, 2002.
84. Egan JA, Ionescu MC, Eapen E et al: Can immunohistochemical analysis of p53 and WT1
assist in distinguishing uterine serous carcinoma from uterine endometrioid carcinoma? (abstract)
Mod Pathol 16(1): 188A (abstract #858), 2003.
85. Euscher ED, Malpic A, Deavers MT et al: Differential expression of WT-1 antibody in serous
carcinomas based on primary site. (abstract) Mod Pathol 16(1): 189A (abstract #860), 2003.
86. Jewell KD, Swanson PE, Garcia RL: Immunohistochemical profile of gynecologic neoplasms
with an emphasis on fallopian tube and serous carcinomas. (abstract) Mod Pathol 16(1): 193A
(abstract #881), 2003.
87. Wang X, Dupont J, Marshall DS et al: WT1 expression in ovarian and endometrial carcinomas
using tissue microarrays. Mod Pathol 16(1):, 213A-214A (abstract #973), 2003.
88. Zhang PJ, Williams E, Pasha T et al: WT1 is expressed in serous, but not in endometrioid,
clear cell, or mucinous carcinomas of the peritoneum, fallopian tube, ovaries, and endometrium.
Mod Pathol 16(1): 216A (abstract # 987), 2003.
89. Zhang Q, Wright T, Alexis D et al: Differential expression of WT1 in uterine serous and

Miller

IHC for Carcinoma of Unknown Primary

Page 39

endometrioid earcinoma. Mod Pathol 16(1): 217A (abstract # 988), 2003.

90. Drakos E, Rassidakis GZ, Tsiolo P et al: Differential expression of WT-1 gene product in
Non-Hodgkins Lymphomas. Mod Pathol 16(1): 231A (abstract # 1054), 2003.
p63
91. Kaufmann O, Fietze E, Mengs J et al: Value of p63 and cytokeratin 5/6 as
immunohistochemical markers for the differential diagnosis of poorly differentiated and
undifferentiated carcinomas. Am J Clin Pathol 116: 823-830, 2001.
92. Wang TY, Chen BF, Yang YC et al: Histologic and immunophenotypic classification of
cervical carcinomas by expression of the P53 homologue p63: A study of 250 cases. Hum Pathol
32 (5): 479-486, 2001.
93. Barbareschi M, Pecciarini L, Cangi MG et al: p63, a p53 homologue, is a selective nuclear
marker of myoepithelial cells of the human breast. Am J Surg Pathol 25 (8): 1054-1060, 2001.
94. Choi HR, Batsakis JG, Zhan F et al: Differential expression of p53 gene family members p63
and p73 in head and neck tumorigenesis. Hum Pathol 33 (2): 158-164, 2002.
95. Pelois G, Pasini F, Olsen Stenholm C et al: p63 immunoreactivity in lung cancer: yet another
player in the development of squamous cell carcinomas? J Pathol 198 (1): 100-109,2002.
96. Shah RB, Zhou M, LeBlanc M et al: Comparison of the basal cell-specific markers, 34E12
and p63, in the diagnosis of prostate cancer. A J Surg Pathol 26 (9): 1161-1168, 2002.
97. Wang BY, Gil J, Kaufman D et al: p63 in pulmonary epithelium, pulmonary squamous
neoplasms, and other pulmonary tumors. Hum Pathol 33 (9): 921-926, 2002
98. Dellavalle RP, Walsh P, Marchbank A et al: CUSP/p63 expression in basal cell carcinoma.
Exp Dermatol 11 (3): 203-208,2002.
98. Reis-Filho JS, Schmitt FC: Taking advantage of basic research: p63 is a reliable myoepithelial
and stem cell marker. Adv Anat Pathol 9:280-289, 2002.
99. Reis-Filho JS, Schmitt FC: p63 expression in sarcomatoid/metaplastic carcinomas of the
breast. (Letter to the editor). Histopathology 42: 94-95, 2003.
100. Werling RW, Hwang H, Yaziji H et al: Immunohistochemical distinction of invasive from
noninvasive breast lesions. A comparative study of p63 versus calponin and smooth muscle
myosin heavy chain. Am J Surg Pathol 27(1): 82090, 2003
OCT3/4
101. Looijenga LH, Stoop H, de Leeuw HP et al: POU5F1 (OCT3/4) identifies cells with
pluripotent potential in human germ cell tumors. Cancer Research 63:2244-2250, 2003.
102. Jones TD, Ulbright TM, Eble JN, Cheng L: OCT 4 immunostaining is useful in the
diagnosis of intratubular germ cell neoplasia. Modern Pathology Supplement 1, page 161A,
abstract 672, January 2004.
103. Jones TD, Ulbright TM, Eble JN, Cheng L: OCT4 is a sensitive and specific marker for
testicular seminomas and embryonal carcinoma. Modern Pathology Supplement 1, page 160A,
abstract 670, January 2004.
104. Thomas A, Roth RM, Abdul-Karim FW, Zheng W, Michael H, Cheng L: OCT4 is a specific
biomarker for dysgerminoma of the ovary. Modern Pathology Supplement 1, page 215A, abstract
905, January 2004.
105. Cheng L: Diagnosing metastatic germ cell tumors using OCT4 immunohistochemistry.
Modern Pathology Supplement 1, page 145A, abstract 60, January 2004.
106. Cheng L, Thomas A, Roth LM et al: OCT4. A novel biomarker for dysgerminoma of the
ovary. Am J Surg Pathol 28(10):1341-1346, Oct 2004.
EMA

Miller

IHC for Carcinoma of Unknown Primary

Page 40

107. Sloane JP, Ormerod MG: Distribution of epithelial membrane antigen in normal and
neoplastic tissues and its value in diagnostic tumor pathology. Cancer 47: 1786-1795, 1981.
108. Pinkus GS, Kurtin PJ: Epithelial membrane antigen: A diagnostic discriminant in surgical
pathology. Immunohistochemical profile in epithelial, mesenchymal, and hematopoetic neoplasms
using paraffin sections and monoclonal antibodies. Hum Pathol 116:929-940, 1985.
109. Wick MR, Cherwitz DL, McGlennan RC et al: Adrenocortical carcinoma. An
immunohistochemical comparison with renal cell carcinoma. (discusses role of EMA) Am J Pathol
122: 343-352, 1986.
110. Rabkin MS, Kjeldsberg CR: Epithelial membrane antigen staining patterns of histiocytic
lesions. Arch Pathol Lab Med 111: 337-338, 1987.
111. Costa, MJ, et al: Immunohistochemical phenotype of ovarian granulosa cell tumors. Absence
of epithelial membrane antigen has diagnostic value. Hum Pathol 25: 60-66, 1994.
Cytokeratin 7 and 20
112. Raemaekers F, van Niekerk C, Poels L et al: Use of monoclonal antibodies to keratin 7 in the
differential diagnosis of adenocarcinomas. Am J Pathol 136(3): 641-655, 1990.
113. Moll R, Lowe A, Laufer J et al: Cytokeratin 20 in human carcinomas. A new histodiagnostic
marker detected by monoclonal antibodies. Am J Pathol 140: 427-447, 1992.
114. Loy TS, Calaluce RD: Utility of Cytokeratin immunostaining in separating pulmonary
adenocarcinomas from colonic adenocarcinomas. Am J Clin Pathol 102: 764-767, 1994.
115. Wang NP, Zee S, Zarbo RJ et al: Coordinate expression of cytokeratins 7 and 20 defines
unique subsets of carcinomas. Appl Immunohistochem 3(2): 99-107, 1995.
116. Guerrieri C, Franlund B, Boeryd B: Expression of cytokeratin 7 in simultaneous mucinous
tumors of the ovary and appendix. Mod Pathol 8(5): 573-576, 1995.
117. Wauters CCAP, Smedts F, Gerrits LGM et al: Keratins 7 and 20 as diagnostic markers of
carcinomas metastatic to the ovary. Hum Pathol 26: 852-855, 1995.
118. Miettinen M: Keratin 20: Immunohistochemical marker for gastrointestinal, urothelial, and
Merkel cell carcinomas. Mod Pathol 8(4): 384-388, 1995.
119. Kurrer MO, Swanson PE: Immunoreactivity for HLA-DR and cytokeratins 7 and 20 in
epithelial proliferations in the thyroid. Mod Pathol 9(1): 49A (abstract #272), 1996.
120. Berezowski K, Stastny JF, Kornstein MJ: Cytokeratins 7 and 20 and carcinoembryonic
antigen in ovarian and colonic carcinoma. Mod Pathol 9(4): 426-429, 1996.
121. Savera AT, Torres FX, Linden MD et al: Primary versus metastatic pulmonary
adenocarcinoma. An immunohistochemical study using villin and cytokeratins 7 and 20. Appl
Immunohistochem 4(2): 86-94, 1996.
122. Loy TS, Calaluce RD, Keeney GL: Cytokeratin immunostaining in differentiating primary
ovarian carcinoma from metastatic colonic adenocarcinoma. Mod Pathol 9: 1041-1044, 1996.
123. Dabbs DJ, Sturtz K, Zaino RJ: The immunohistochemical discrimination of endometrioid
carcinomas. Hum Pathol 27: 172-177, 1996.
124. Maeda T et al: The expression of cytokeratins 7, 19, and 20 in primary and metastatic
carcinomas of the liver. Modern Pathology 9 (9): 901-909, 1996.
125. Alexander J, Krishnamurthy S, Kovacs D et al: Cytokeratin profile of extrahepatic
pancreaticobiliary epithelium and their carcinomas. Appl Immunohistochem 5(4): 216-222, 1997.
126. Chan JK, Suster S, Wenig BM et al: Cytokeratin 20 immunoreactivity distinguishes Merkel
cell (primary cutaneous neuroendocrine) carcinomas and salivary gland small cell carcinomas from
small cell carcinomas of variable sites. Am J Surg Pathol 21: 226-234, 1997
127. Alobeid B, Zhang PJ: Cytokeratin and villin immunoprofile of pancreatic islet cell tumors.
(abstract) Am J Clin Pathol 107(4): 474-475, 1997.

Miller

IHC for Carcinoma of Unknown Primary

Page 41

128. Tan J, Sidhu G, Greco MA et al: Villin, cytokeratin 7, and cytokeratin 20 expression in
pulmonary adenocarcinoma with ultrastructural evidence of microvilli with rootlets. Hum Pathol
29: 393-396,1998.
129. Lagendijk JH et al: Tracing the origin of adenocarcinomas with unknown primary using
immunohistochemistry: differential diagnosis between colonic and ovarian carcinomas as primary
sites. Hum Pathol 29: 491-7, 1998.
130. Duval JV, Savas L, Banner BF: Expression of cytokeratins 7 and 20 in carcinomas of the
extrahepatic biliary tract, pancreas, and gallbladder. Arch Pathol Lab Med 124:1196-1200, 2000.
131. Chu P et al: Cytokeratin 7 and cytokeratin 20 expression in epithelial neoplasms: A survey of
435 cases. Mod Pathol 13 (9): 962-972, 2000.
132. Bassily NH et al: Coordinate expression of cytokeratin 7 and 20 in prostate adenocarcinoma
and bladder urothelial carcinoma. Am J Clin Pathol 113: 383-388, 2000.
133. oldstein NS et al: Cytokeratin 7, 17, and 20 reactivity in pancreatic and ampulla of Vater
adenocarcinomas. Percentage of positivity and distribution is affected by the cut-point threshold.
Am J Clin Pathol 115: 695-702, 2001.
134. Lau SK et al: Expression of thyroid transcription factor-1, cytokeratin 7, and cytokeratin 20
in bronchoalveolar carcinomas: an immunohistochemical evaluation of 67 cases. Mod Pathol 15
(5): 538-542, 2001.
135. Lau SK et al: expression of thyroid transcription factor-1, cytokeratin 7, and cytokeratin 20
in bronchoalveolar carcinomas: an immunohistochemical evaluation of 67 cases. Mod Pathol 15
(5): 538-542, 2001.
136. Cathro HP et al: Expression of cytokeratin 7 and 20 in ovarian neoplasia. Am J Clin Pathol
117: 944-951, 2002.
137. Shah RN, Badve S, Papreddy K et al: Expression of cytokeratin 20 in mucinous
bronchioloalveolar carcinoma. Human Pathology 33 (9): 915-920, 2002.
138. Park SY, Kim HS, Hong EK et al: Expression of cytokeratins 7 and 20 in primary
carcinomas of the stomach and colorectum and their value in the differential diagnosis of
metastatic carcinomas to the ovary. Human Pathology 33:1078-1085, 2002.
139. Lau SK, Prakash S, Geller SA et al: Comparative immunohistochemical profile of
hepatocellular carcinoma, cholangiocarcinoma, and metastatic adenocarcinoma. Human Pathology
33 (12): 1175-81, 2002.
140. Parker DC, Folpe AL, Bell J et al: Potential utility of uroplakin III, thrombomodulin, high
molecular weight cytokeratin, and cytokeratin 20 in noninvasive, invasive, and metastatic
urothelial (transitional cell) carcinomas. Am J Surg Pathol 27(1): 1-10, 2003
Villin
141. Moll R, Robine S, Dudouet B et al: Villin: a cytoskeletal protein and a differentiation marker
expressed in some human adenocarcinomas. Virchows Arch B Cell Pathol Incl Mol Pathol 54:
155-169,1987.
142. Bacchi C, Gown AM: Distribution and pattern of expression of villin, a gastrointestinalassociated cytoskeletal protein, in human carcinomas: A study employing paraffin-embedded
tissue. Lab Investig 64(3): 418-424, 1991.
143. Savera AT, Torres FX, Linden MD et al: Primary versus metastatic pulmonary
adenocarcinoma. An immunohistochemical study using villin and cytokeratins 7 and 20. Appl
Immunohistochem 4(2): 86-94, 1996.
144. Alobeid B, Zhang PJ: Cytokeratin and villin immunoprofile of pancreatic islet cell tumors.
(abstract) Am J Clin Pathol 107(4): 474-475, 1997.
145. Tan J, Sidhu G, Greco MA et al: Villin, cytokeratin 7, and cytokeratin 20 expression in
pulmonary adenocarcinoma with ultrastructural evidence of microvilli with rootlets. Hum Pathol
29:393-396,1998.

Miller

IHC for Carcinoma of Unknown Primary

Page 42

146. Zhang PJ, Harris KR, Alobecid B, Brooks JJ: Immunoexpression of villin in neuroendocrine
tumors and its diagnostic implications. Arch Pathol Lab Med 123: 812-816, 1999
Cytokeratins 5, 5/6, 17, 8 & 18 (CK-lmw), 34bE12 (CK-hmw) and Vimentin
147. Azumi N, Battifora H: The distribution of vimentin and keratin in epithelial and nonepithelial
neoplasms. Am J Clin Pathol 88: 286-296, 1987.
148. Balaton AJ, Nehama-Sibony M, Gotheil C et al: Distinction between hepatocellular
carcinoma, cholangiocarcinoma, and metastatic carcinoma based on immunohistochemical staining
for carcinoembryonic antigen and for cytokeratin 19 on paraffin sections. J Pathol 156: 305-310,
1988.
149. Gould VE, Bloom KJ, Franke WW: Increased numbers of cytokeratin positive interstitial
reticulum cells (CIRC) in reactive, inflammatory, and neoplastic lymphadenopathies: Hyperplasia
or induced expression? Virchows Arch 425:617-630, 1995.
150. Clover J, Oates J, Edwards C: Anti-cytokeratin 5/6: A positive marker for epithelioid
mesothelioma. Histopathol 31:140-3, 1997
151. Miettinen M, Nobel MP, Tuma BT et al: Keratin 17. Immunohistochemical mapping of its
distribution in human epithelial tumors and its potential applications. Appl Immunohistochem
5(3): 152-159, 1997.
152. Ordonez NG: Value of cytokeratin 5/6 immunostaining in distinguishing epithelial
mesothelioma of the pleura from lung adenocarcinoma. Am J Surg Pathol 22(10): 1215-1221,
1998.
153. Suster S, Moran C, Dominguez-Malagon H et al: Germ cell tumors of the mediastinum and
testis: A comparative immunohistochemical study. Hum Pathol 29: 737-742, 1998.
154. Alobeid B, Brooks JJ, Zhang PJ et al: Aberrant cytokeratin subset immunoreactivity in
sarcomas using a large panel of cytokeratin subset antibodies. Appl Immunohistochem 6(3): 154157, 1998.
155. Geddes JF et al: Gangliocytomas of the Pituitary. A heterogeneous group of lesions with
differing histogenesis. Amer J Surg Pathol 24(4): 607-613, 2000.
156. Vege KD, Giannini C, Scheithaur BW: The immunophenotype of ependymomas. Appl
Immunohistochem Mol Morphol 8(1): 25-31, 2000
157.
Kaufmann O, Fietze E, Mengs J et al: Value of p63 and cytokeratin 5/6 as
immunohistochemical markers for the differential diagnosis of poorly differentiated and
undifferentiated carcinomas. Am J Clin Pathol 116: 823-830, 2001.
158. Abrahams NA, Ormsby AH, Brainard J: Validation of cytokeratin 5/6 as an effective
substitute for keratin 903 in the differentiation of benign for malignant glands in prostate needle
biopsies. Histopathology 41:35-41, 2002.
159. Reis-Filho JS, Martins A, Preto A et al: Distribution of p63, cytokeratin 5/6, and cytokratin
14 in 51 normal and 450 neoplastic human tissue samples using TARP-4 multi-tumor tissue
microarray. Mod Pathol 16(1): 326A (abstract # 1493), 2003.
Neuroendocrine Markers
160. Kauffman O et al: Utility of 123C3 monoclonal antibody against CD56 (NCAM) for the
diagnosis of small cell carcinomas in paraffin sections. Hum Pathol 28: 1373-1378, 1997.
161. Lyda MH, Weiss LM: Immunoreactivity for epithelial and neuroendocrine antibodies is
useful in the differential diagnosis of lung carcinomas. Hum Pathol 31: 980-987, 2000.
162. Lantuejoul S, Moro D, Michalides RJAM et al: Neural cell adhesion molecules (NCAM) and
NCAM-PSA expression in neuroendocrine lung tumors. Am J Surg Pathol 22 (10): 1267-1276,
1998.
CDX2

Miller

IHC for Carcinoma of Unknown Primary

Page 43

163. Ee HC, Erler T, Bhathal PS et al: CDX2 homeodomain protein expression in human and rat
colorectal adenoma and carcinoma. Am J Pathol 147:586-592, 1995
164. Silberg DG, Swain GP, Suh ER et al: CDX1 and CDX2 expression during intestinal
development. Gastroenterol 119:961-971, 2000
165. BaRbareschi M, Murer B, Colby TV et al: CDX-2 homeobox gene expression is a reliable
marker of colorectal adenocaricoma metastases to the lungs. Am J Surg Pathol 27(2): 141-149,
2003.
166. Kaimakchiev S, Simhofer S, Sauter G et al: Selective staining of gastrointestinal
adenocarcinomas by the homeobox intestinal differentiation factor CDX2. Mod Pathol 16(1):
123A (abstract # 556), 2003.
167. Mazziotta RM, Borczuk AC, Alexis D et al: Differential expression by immunohistochemistry, of CDX2 transcription factor in various adenocarcinomas. Mod Pathol 16(1):
127A (abstract # 577), 2003
168. Furlanetto A, Orvieto E, Laurino L et al: Utility of CDX2 in the diagnosis of metastatic
adenocarcinoma to the liver. Mod Pathol 16(1): 273A (abstract 1246), 2003.
169. Werling RW, Yaziji H, Bacchi CE, Gown AM. CDX2, a highly sensitive and specific
marker of adenocarcinomas of intestinal origin: an immunohistochemical survey of 476
primary and metastatic carcinomas. Am J Surg Pathol. 2003 Mar;27(3):303-10.
170. Vang R, Gown AM, Barry TS, Wheeler DT, Ronnett BM. Ovarian atypical proliferative
(borderline) mucinous tumors: gastrointestinal and seromucinous (endocervical-like) types are
immunophenotypically distinctive. Int J Gynecol Pathol. 2006 Jan;25(1):83-9.

BCL-2, CEA, HBME-1, N-cadherin

171. Costa, MJ, et al: Utility of immunohistochemistry in distinguishing ovarian sertoli stromal
cell tumors from carcinosarcomas. Hum Pathol 23: 787-797, 1992.
172. Fitzpatrick M, Adesokan PN, Ritter JH et al: Immunohistologic differential diagnosis of
basal cell carcinoma, squamous cell carcinoma, and trichoepithelioma in small cutaneous biopsy
specimens (abstract). Am J Clin Pathol April 1995, page 508.
173. Miettinen M, Kovatich AJ: HBME-1. A monoclonal antibody useful in the differential
diagnosis of mesothelioma, adenocarcinoma, and soft tissue and bone tumors.
Appl
Immunohistochem 3(2): 115-122, 1995.
174. Ostrowski ML, Merino M: Tall cell variant of papillary carcinoma: A reassessment and
immunohistochemical study with comparison to the usual type of papillary carcinoma of the
thyroid. Am J Surg Pathol 20: 964-974, 1996.
175. Alsabeh R, Wilson CS, Ahn SW, Vasef MA, Battifora H: Expression of BCL-2 by breast
cancer: A possible diagnostic application. Mod Pathol 9(4): 439-444, 1996.
176. Jiang S-X et al: BCL-2 protein expression in lung cancer and close correlation with
neuroendocrine differentiation. Am J Pathol 148(3): 837-846, 1996.
177.
O-Hara BJ, Paetau A, Miettinen MM et al: Chordoma - A panel allowing
immunohistochemical differential diagnosis from epithelial and non-epithelial tumors. (abstract)
Mod Pathol 9(1): 11A, abstract #46, 1996.
178. Peralta-Solter P, Knudsen KA, Tecson-Miguel A et al: Expression of E-cadherin and Ncadherin in surface epithelial-stromal tumors of the ovary distinguishes mucinous from serous and
endometrioid tumors. Hum Pathol 28(6): 734-739, 1997.
Inhibin, A103, and Calretinin (as markers of adrenal cortical and ovarian tumors)
179. Flemming P, Wellmann A, Maschek H et al: Monoclonal antibodies against inhibin represent
key markers of adult granulosa cell tumors of the ovary even in their metastases. Am J Surg
Pathol 19(8): 927-933, 1995.

Miller

IHC for Carcinoma of Unknown Primary

Page 44

180. Rishi M, Howard LN, Bratthauer GL et al: Use of monoclonal antibody against inhibin as a
marker for sex cord-stromal tumors of the ovary. Amer J Surg Pathol 21(5): 583-589, 1997.
181. Costa MJ, Ames PF, Walls J et al: Inhibin immunohistochemistry applied to ovarian
neoplasms: A novel, effective, diagnostic tool. Hum Pathol 28(11): 1247-1254, 1997.
182. Hildebrandt RH, Rouse RV, Longacre TA: Value of inhibin in the identification of granulosa
cell tumors of the ovary. Hum Pathol 28(12): 1387-1395, 1997
183. McCluggage WG, Maxwell P, Sloan JM: Immunohistochemical staining of ovarian
granulosa cell tumors with monoclonal antibody against inhibin. Hum Pathol 28(9): 1034-1038,
1997.
184. Pelkey TJ, Frierson HF, Mills SE et al: The diagnostic utility of inhibin staining in ovarian
neoplasms. Int J Gynecol Pathol 17(2): 97-105, 1998.
185. Pelkey TJ, Frierson HF, Mills SE et al: The alpha-subunit of inhibin in adrenal cortical
neoplasia. Mod Pathol 11(6): 516-524, 1998.
186. Renshaw AA, Granter SR: A comparison of A103 and inhibin reactivity in adrenal cortical
tumors: Distinction from hepatocellular carcinoma and renal tumors. Mod Pathol 11 (12): 11601164, 1998.
187. Shin SJ, Hoda RS, Ying L et al: Diagnostic utility of the monoclonal antibody A103 in fineneedle aspiration biopsies of the adrenal. Am J Clin Pathol 113: 295-302, 2000.
188. Cho E, Ahn GH: Immunoexpression of inhibin alpha-subunit in adrenal neoplasms. Appl
Immunohistochem Molec Morphol 9 (3): 222-228, 2002.
189. Loy TS, Phillips RW, Linder CL: A103 immunostaining in the diagnosis of adrenal cortical
tumors. Arch Pathol Lab Med 126: 170-172, 2002.
190. Jorda M, Madeiros BD, Nadji M: Calretinin and inhibin are useful in separating
adrenocortical neoplasms from pheochromocytomas. Appl Immunohistochem Molec Morphol 10
(1): 67-70, 2002.

CD10
191. Chu P, Arber DA: Paraffin-section detection of CD10 in 505 nonhematopoetic neoplasms.
Frequent expression in renal cell carcinoma and endometrial stromal sarcoma. Am J Clin Pathol
113 (3): 374-382,2000.
192. Ordi J, Romagosa C, Tavassoli F et al: CD10 expression in epithelial tissues and tumors of
the gynecologic tract. Am J Surg Pathol 27(2): 178-186, 2003.
Thrombomodulin
193. Collins CL, Ordonez NG, Schaefer R et al: Thrombomodulin expression in malignant pleural
mesothelioma and pulmonary adenocarcinoma. Am J Pathol 141(4): 827-833, 1992.
Uroplakin
194. Moll R, Wu XR, Lin JH et al: Uroplakins, specific membrane proteins of urothelial umbrella
cells, as histological markers of metastatic transitional cell carcinomas. Am J Pathol 147: 13 8397,1995.
195. Kaufman O, Volmerig J, Dietel M: Uroplakin III is a highly specific and moderately sensitive
immunohistochemical marker for primary and metastatic urothelial carcinomas. Am J Clin Pathol
113: 683-687, 2000.
196. Mhawech P, Uchida T, Pelte M-F: Immunohistochemical profile of high-grade urothelial
bladder carcinoma and prostate adenocarcinoma. Human Pathology 33:1136-1140, 2002.
197. Parker DC, Folpe AL, Bell J et al: Potential utility of uroplakin III, thrombomodulin, high
molecular weight cytokeratin, and cytokeratin 20 in noninvasive, invasive, and metastatic
urothelial (transitional cell) carcinomas. Am J Surg Pathol 27(1): 1-10, 2003

Miller

IHC for Carcinoma of Unknown Primary

Page 45

CA19-9, CA15-3, CA-125, MOC-31, Mesothelin

198. Gatalica Z, Miettinen M: Distribution of carcinoma antigens CA19-9 and CA15-3. An
immunohistochemical study of 400 tumors. Appl Immunohistochem 2(3): 205-211, 1994.
199.
Niemann TH, Hughes JH, DeYoung BR: MOC31 immunoreactivity aids in the
differentiation of metastatic adenocarcinoma from hepatocellular carcinoma of liver. Cancer
87(5): 295-298, 1999.
200. Proca DM, Niemann TH, Porcell AI, DeYoung BR: MOC31 immunoreactivity in primary
and metastatic carcinoma of the liver. Report of findings and review of other utilized markers.
Appl Immunohistochem Molec Morphol 8(2): 120-125, 2000.
201. Porcell AI, DeYoung BR, Proca DM et al: Immunohistochemical analysis of hepatocellular
and adenocarcinoma in the liver: MOC-31 compares favorably with other putative markers. Mod
202. Pathol 13(7):773-778, 2000.
203. Oliva E, Amin MB, Jimenez R et al: Clear cell carcinoma of the urinary bladder: A report
and comparison of 4 tumors of mullerian origin and 9 of probable urothelial origin with discussion
of histogenesis and diagnostic problems. Am J Surg Pathol 26 (2): 190-197, 2002.
204. Goldstein NS: Immunophenotypic characterization of 225 prostate adenocarcinomas with
intermediate or high Gleason scores. Am J Clin Pathol 117 (3): 471-477, 2002.
205. Ordonez NG: Application of mesothelin immunostaining in tumor diagnosis. Am J Surg
Pathol 27:1418-1428, Nov 2003.

Tissue Transfer Immunohistochemistry

206. Miller RT, Kubier P: Immunohistochemistry on Cytologic Specimens and Previously Stained
Slides (When No Paraffin Block Is Available). Journal of Histotechnology 25(4):251-257, 2002.
Additional Selected References
207. Drier JK, Swanson PE, Cherwitz DL, et al: S100 protein immunoreactivity in poorly
differentiated carcinomas. Immunohistochemical comparison with malignant melanoma. Arch
Pathol Lab Med 111: 447-452, 1987.
208. Esteban JM, Battifora H: Tumor immunophenotype: Comparison between primary neoplasm
and its metastases. Mod Pathol 3(2): 192-197, 1990.
209. Arber DA, Weiss LM: CD15: A review. Appl Immunohistochem 1(1): 17-30, 1993.
210. Raab SS, Berg L, Swanson PE et al: Adenocarcinoma in the lung in patients with breast
cancer. A prospective analysis of the discriminatory value of immunohistology. Amer J Clin
Pathol 100: 27-35, 1993.
211. Brown RW, Campagna LB, Dunn JK et al: Immunohistochemical identification of tumor
markers in metastatic adenocarcinoma. A diagnostic adjunct in the determination of primary site.
Am J Clin Pathol 107 (1): 12-19, 1997.
212. DeYoung BR, Wick MR: Immunohistologic evaluation of metastatic carcinomas of unknown
origin: an algorithmic approach. Semin Diag Pathol 17 (3): 184-193, 2000.
213. Raab SS: The cost-effectiveness of immunohistochemistry. Archives of Pathology and
Laboratory Medicine 124: 1185-1191, 2000.
214. Kaufman O, Fietze E, Dietel M: Immunohistochemical diagnosis in cancer metastasis of
unknown primary tumor. (article in German) Pathologe 23 (3): 183-197,2002.
215. Ordonez NG: Immunohistochemical diagnosis of epithelioid mesotheliomas: A critical
review of old markers, new markers. Hum Pathol 33(10): 953-967, 2002
216. Chu PG, Weiss LM: Keratin expression inhuman tissues and neoplasms (Review).
Histopathology 40: 403-439, 2002.

Miller

IHC for Carcinoma of Unknown Primary

Page 46

Pax8
217. Tong G-X, Yu WM, Beaubier NT et al: Expression of Pax8 in normal and neoplastic renal
tissues: an immunohistochemical study. Mod Pathol 22:1218-1227, Sep 2009.
218. Nonaka D, Chiriboga L, Soslow RA: Expression of Pax8 as a useful marker in distinguishing
ovarian carcinomas from mammary carcinomas. Am J Surg Pathol 32(10): 1566-1571, Oct 2008.
219. Nonaka D, Tang Y, Chiriboga L et al: Diagnostic utility of thyroid transcription factors Pax8
and TTF-2 (FoxE1) in thyroid epithelial neoplasms. Mod Pathol 21(2): 192-200, Feb 2008.
219a. Long KB, Srivastava A, Hirsch MS et al: Pax8 expression in well differentiated pancreatic
endocrine tumors: Correlation with clinicopathologic features and comparison with
gastrointestinal and pulmonary carcinoid tumors. Am J Surg Pathol 34(5): 723-729, May 2010.
Napsin A
220. Bishop JA, Sharma R, Illei PB: Napsin A and thyroid transcription factor-1 in carcinoma of
the lung, breast, pancreas, colon, kidney, thyroid, and malignant mesothelioma. Hum Pathol Sep
7, 2009 (EPub ahead of print).
221. Jagirdar J: Application of immunohistochemistry to the diagnosis of primary and metastatic
carcinoma to the lung. Arch Pathol Lab Med 132(3): 384-396, Mar 2008.
pVHL
222. Lin F, Shi J, Liu H et al: Immunohistochemical detection of the von Hippel-Lindau gene
product (pVHL) in human tissues and tumors. Am J Clin Pathol 129(4): 529-605, Apr 2008.
223. Lin F, Shi J, Liu H et al: Diagnostic utility of S100P and von Hippel-Lindau gene product
(pVHL) in pancreatic adenocarcinoma with implication of their roles in early tumorigenesis.
Am J Surg Pathol 32(1): 78-91, Jan 2008.
SALL4
224. Cao D, Guo S, Allan RW et al: SALL4 is a novel sensitive and specific marker of ovarian
primitive germ cell tumors and is particularly useful in establishing yoke sac tumor from clear cell
carcinoma. Am J Surg Pathol 33(6): 894-904, Jun 2009 .
225. Cao D, Li J, Guo S et al: SALL4 is a novel diagnostic marker for testicular germ cell tumors.
Am J Surg Pathol 33(7): 1065-1077, Jul 2009 .
226. Ushiku T, Shinozaki A, Shibahara J et al: SALL4 represents fetal gut differentiation of
gastric cancer, and is diagnostically useful in the distinguishing hepatoid gastric carcinoma from
hepatocellular carcinoma. Am J Surg Pathol 34(4): 533-540, Apr 2010
Arginase-1
227. Yan BC, Gong C, Song J et al: Arginase-1. A new immunohistochemical marker of
hepatocytes and hepatocellular neoplasms. Am J Surg Pathol 34(8):1147-1154, Aug 2010.

Rodney T. Miller, M.D.

ProPath Laboratory
January 2011

Miller

IHC for Carcinoma of Unknown Primary

Immunohistochemical
Approach to Metastatic
Carcinoma of Unknown
Primary Origin

IHC in Metastatic Carcinoma

of Unknown Primary: Goals
Discuss the spectrum of reactivity of
most useful antibodies (we will skip the

hard-core science)

Review illustrative cases

Suggest an approach to workup of
cases
Provide useful written information for
diagnostic pathologists (Handout)

Rodney T. Miller, M.D.

Director of Immunohistochemistry
ProPath Laboratory, Inc.
Dallas, Texas

Metastatic Carcinoma
of Unknown Primary
Why Immunohistochemistry?
Tumors of widely varying origins may look
identical on standard H&E sections.

Tumors of specific origins tend to express

certain markers or combinations of markers
(immunophenotypes) that can distinguish
them from other origins.

Page 47

Principles of Immunophenotyping: 1
Immunostains must be of high quality
(You need good tools to do a good job).

Generate appropriate DDx on H&E.

You must know :

A. spectrum of reactivity of Abs
B. expected immunophenotypes
(Use of IHC Peripheral Brain).

Principles of Immunophenotyping: 2
There are no perfect markers, so USE
PANELS and avoid IHC Guilt Syndrome

(Immunophenotype is only part of the picture.

Correlate with other findings).

Tumors do not read textbooks.

You will get some cases wrong.

Met Ca: Organ-Related Abs

Prostate Specific Antigen (PSA) and
Prostate Specific Acid Phosphatase (PSAP)
PSA False +: Breast, Salivary, Anal glands,
Periurethral Glands, Cystitis Cystica /
Glandularis
PSAP False +: Same as above plus Rectal
Carcinoid, Sweat Glands, rare Renal Cell ca,
Islet Cell Tumors

PSAP+ rectal carcinoid

Miller

IHC for Carcinoma of Unknown Primary

Met Ca: Organ-Related Abs

Met Ca: Organ-Related Abs: Breast

PSMA, prostate ca

Prostate-Specific Membrane Ag (PSMA)

P501S (prostein), and NKX3.1
PSMA: More sensitive than PSA and PSAP
in high-grade tumors. (also stains endothelial
cells in many non-prostate tumors) (salivary
glands, renal tubules, some GI mucosal cells,
maybe Mallory bodies??).
P501S (prostein) Characteristic perinuclear
globs in prostate ca

Page 48

PSMA on RCC

Gross Cystic Disease Fluid

Protein-15 (GCDFP-15)
- Pos in 50-60% Breast ca, also

Salivary ca, Sweat Gland ca,

Prostate ca, rare Lung ca

Met Breast ca to endomet

Mammaglobin
P501S, prostate ca

- Pos in 50-60% Breast ca, also

Endometrial ca (40-70%), Endocervical ca (30%), Ovarian (17%),

Salivary ca, Sweat Gland ca,

NKX3.1 Great nuclear marker: 98.6%

sensitive (68/69), 99.7% specific (1/349
cases, a lobular breast ca)

Mammaglob, endomet ca

NKX3.1, prostate ca

Met Ca: Organ-Related Abs

Met Ca: Organ-related Abs

Thyroglobulin

TTF-1 (Thyroid Transcription Factor-1)

- Pos: 75% Lung Non-Small Cell ca,(75%

Pos: Follicular and Papillary ca

thyroid, sometimes focal
pos in Anaplastic (less

acinar, 100% bronchoalveolar, 91% clear cell,

57% solid ca), 80% Atypical Carcinoid, 90%
Small Cell ca Lung, Thyroid ca (incl Medullary)
(17% endometrial ca, rare pancreas)
- Neg: Breast, Colon (r+), Gastric (1.5%+),
Carcinoid Lung (20%+), Prostate,
Kidney, Mesothelioma, Merkel cell

sensitive for thyroid than TTF-1)

Neg: Medullary ca thyroid,

Non-thyroid carcinomas

76F, pleural bx

Cytoplasmic in hepatoma, gastric ca,

clone 8G7G3/1 only (not SPT24)

Pax8 also a good thyroid marker, discussed later

Met Ca: Organ-related Abs

Napsin A

Lung CA

Hepatoma

Met Ca: Organ-related Abs

Estrogen Receptor (SP1 or 1D5)
Breast Ca

Lung ca in liver bx

- Pos: 70-80% Lung Adenocarcinoma,

Kidney: 79% papillary, 34% clear cell
<5% lung squamous, pancreaticobiliary,
thyroid, bladder, colon, female genital tract

Pos: Breast, Female Genital Tract,

Thyroid, Salivary, Sweat Gland,
Chordoma, 10-15% Lung ca,
few Urothelial, Hepatoma (rare)
TTF-1

Neg: GI tract, Kidney, Pancreas, Bile Ducts

TRAP: Papillary RCC (79%)

Clear cell RCC (34%)

Alveolar macrophages

ER+ lung Ca

ER+ Normal Liver

Miller

IHC for Carcinoma of Unknown Primary

Met Ca: Organ-related Abs

Page 49

Met Ca: Organ-related Abs

clear cell hepatoma

HepPar 1 (Hepatocyte Paraffin 1)

Arginase-1 (Arg-1)

- Pos in 50-90% hepatomas,

also significant # gastric cas
and hepatoid carcinomas

- Pos in 96% hepatomas

(also neutrophils, histiocytes)

- Pos in only 2/557 non-HCC

- Better than HepPar1

Reactivity tends to occur along with

strong granular cytoplasmic
reactivity with TTF-1 (8G7G3/1)

Cytoplasmic and nuclear, but

cytoplasmic reactivity required
for a positive stain
HepPar 1+ Gastric Ca

Met Ca: Organ-related Abs

Met Ca: Organ-related Abs

WT1 (Wilms Tumor Gene)

Pax8
Nuclear positivity in:

Arg-1, pancreatic bx
with metastatic HCC

Pleural fluid Met RCC

1: Kidney tumors
2: Female genital tract adenoca
3: Thyroid tumors (better than
TTF-1 and TG in anaplastic ca)

Nuclear positivity in
mesothelioma and serous
carcinoma of ovary, tube, and
peritoneum (+/- in uterine serous)

WT1 on mesothelioma

(Also in endometrial stromal sarcoma,

granulosa cell tumor, thecoma,
uterine smooth muscle).

also: mesonephric things, some

NE tumors, some lymphoid cells

(Cytoplasmic reactivity nonspecific)

LN Met Serous Ca

WT1 on serous adenoca

Case: 62F with ascites and L axillary

adenopathy undergoes L ax LN bx

GCDFP-15

Mammaglobin

ER

Miller

IHC for Carcinoma of Unknown Primary

Page 50

Met Ca: Organ-related Abs

p63
Pax8

WT1

Dx: Met. Serous

Adenocarcinoma

Pos in myoepithelial cells,

prostatic (& other) basal
cells, squamous cas,
urothelial cas, thymoma

p63 on squamous ca

- Good for sarcomatoid ca

(Scattered pos cells common in many
tumors)

Pax2

p63 on urothelial ca

Case: 49M with a Hx of L neck FNA showing

adenoca of unknown primary, developed pericardial
effusion. Cell block obtained

Met Ca: Organ-related Abs

RCC
Renal cell, breast (33%), colon (38%),
prostate 27%, Lung 14%, Ovary 12% RCC on met RCC to pleura
(overrated as a kidney marker)

pVHL
Renal cell ca, Clear cell ca of female
genital tract, some cholangioca
(neg in pancreatic ca, but pos in
benign pancreatic ducts, useful
in combination with Placental S100)
EMA

pVHL on met RCC to pleura

Calretinin

OCT3/4 and SALL4 for Germ cell Tumors

TTF-1

Napsin A

pVHL
Pax8

Highly sensitive and specific

OCT3/4: Seminoma, Embryonal ca
SALL4: Above plus Yolk Sac ca and others

Vim

Dx: Met. Pap. Renal Cell Ca

Embryonal ca in LN

in situ germ cell neoplasia

Miller

IHC for Carcinoma of Unknown Primary

Ab for Met ca:

EMA

Page 51

Met Ca: CK 7 and 20 (Wang 1995)

Pos in many epithelial tumors,

some lymphoid tumors
(ALCL, myeloma)
- Neg in Adrenal, Hepatoma
(dots), Germ Cell tumors

CK 7 pos, CK 20 pos:

CK 7 pos, CK 20 neg:

Urothelial, Pancreatic, Ovarian Mucinous, Stomach

Lung, Breast, Ovary (non-mucinous), Endometrial,
Mesothelial, Pancreaticobiliary, Stomach, Small bowel,
Thyroid

Hepatomas EMA dots

(except Chorioca, Teratoma)

CK 7 neg, CK 20 pos:

CK 7 neg, CK 20 neg:

Colon, Duodenal / Ampullary Ca, Stomach

Lung squamous, Hepatoma, Kidney, Prostate, Stomach

Mesothelioma

Metastatic Carcinoma

Met Ca: Useful Abs

Use of CK7, CK20, and Villin

Colon ca

VILLIN (actin binding protein)

Pos: Brush border: GI, Pancreas, Bile
ducts, Gallbladder, some Lung ca
Cytoplasmic: 68% Lung
Canalicular: 50% Hepatoma
Membrane: Carcinoid, other NE

Lung ca

Carcinoid

Neg: Breast, Mesothelioma

(weak in some cases)

CK7 Positive, CK20 Positive

Villin Positive
Stomach, Pancreas,
Bile ducts, Mucinous
Ovary, Small bowel
Rare: urothelium, breast,
prostate (colon unless
rectal), endomet, lung
unlikely)

Villin Negative
Mucinous Ovary,
Urothelium, Breast
(1/3 of mucinous
breast ca, most inv.
papillary ca breast)
Rare: GI, pancreas, bile
ducts (unless high
grade)

Hepatoma

Metastatic Carcinoma

Metastatic Carcinoma

Use of CK7, CK20, and Villin

Use of CK7, CK20, and Villin

CK7 Positive, CK20 Negative

Villin Positive
Lung, Pancreas/BD
Stom/SB, Endomet,
Mucinous Ovary, Sq
Rare: urothelium,
breast, serous ov,
mesothel., colon

Villin Negative
Lung, Breast, Ovary
(Serous or Mucinous)
Urothelium, Endomet,
Mesothelioma, Sq
Rare: GI, pancreas,
bile ducts

CK7 Negative, CK20 Positive

Villin Positive
Stomach, Duodenal
Ampullary, Colon
Hepatoma (canalicular)

Villin Negative
Hepatoma
Some prostate cas

Rare: breast, lung

Rare: breast (3%), lung
(rare), bladder, fem
(rare), bladder, fem
genital, mesothelioma
genital, mesothelioma

Miller

IHC for Carcinoma of Unknown Primary

Metastatic Carcinoma

Page 52

Case: 86M with a PSA of 37 with multiple

bone & lung lesions undergoes lung bx.

Use of CK7, CK20, and Villin

CK7 Negative, CK20 Negative
Villin Positive
Stomach, Renal cell,
Lung Squamous,
Hepatoma (canalicular)
Prostate? (33% villin pos)

Villin Negative
Mesothelioma, Renal
Cell, Lung Squam.,
Hepatoma, Prostate,
(Breast)

Neuroendocrine

Rare: stomach, ov.,

pancreas, urothel.

Rare: mesothelioma,
breast, ov., urothel.,
pancreas.

Additional IHC on Liver Biopsy

CK 7

CK 20

Villin

CEA

PSA

PSAP

CK7 Negative, CK20 Positive

Villin Positive
Stomach, Duodenal
Ampullary, Colon
Hepatoma (canalicular)

Villin Negative
Hepatoma
Some prostate cas

Rare: breast, lung,

Rare: breast (3%), lung,
bladder, fem genital, bladder, fem genital,
mesothelioma
mesothelioma

Dx: Metastatic Colonic Ca

(confirmed on colon bx)

Doctor reacts to the news

Met Ca: Useful Abs

Cytokeratin (AE1/AE3)
Imperfect first line epithelial screen:
not a true Pankeratin
Neg: Hepatoma, Seminoma, some
Renal cell, Adrenal, Prostate,
some Carcinoid & Islet Cell

AE1/AE3 neg Hepatoma

AE1/AE3 pos Lung ca

AE1/AE3 neg Hepatoma

Hepatoma, foc+

Miller

IHC for Carcinoma of Unknown Primary

Met Ca: Useful Abs

Page 53

Met Ca: Useful Abs

Cytokeratin LMW (8,18)

Merkel
Compliments AE1/AE3
Pos: Hepatoma, Carcinomas neg w AE1/AE3

(Renal cell ca, Carcinoid, Prostate ca, etc.)

Better for detecting Small Cell Ca than AE1/AE3 in

nearly all cases (perinuclear dots)
AE1/AE3

Pagets Disease

Hepatoma

Merkel

Pos: Squamous carcinoma

Basaloid carcinoma,
Basal cell carcinoma,
Epithelial Mesothelioma
Thymoma, Myoepithelial
Neg: Adenocarcinomas
Urothelial carcinoma
(scattered pos cells or
clusters common)

Pagets Disease

Mesothelioma

Small cell ca

Use of Cytokeratin LMW & HMW

Cytokeratin 5 or (5/6)

LMW POS, HMW NEG:

HCC, RCC (conv. type), Prostate
LMW POS, HMW POS:
Many Tumors (incl well diff Squamous)
LMW NEG, HMW POS:
Squamous Ca (well differentiated)

Squamous Cas: HMW > LMW

Strong coexpression of CK5 and p63

Met Ca: Useful Abs

Vimentin
Pos: Kidney (usual type),
Thyroid, Endometrioid,
Paraganglioma, Melanoma

Neg: Chromophobe kidney, Hepatoma,

Neuroblastoma, GI, Urothelial,
Pancreas, Prostate, Germ cell
variable

Colon

Hepatoma

Met Ca: Useful Abs

Met Ca: Useful Abs

CDX2

CEA (clone COL1)

Pos: Many Adenocarcinomas, Canalicluar CEAm (Rare)
Medullary ca Thyroid (~100% of cells)
Neg: Renal Cell ca, Hepatoma (rare canalicular),
Thyroid, Adrenal, Mesothelioma
If your CEA Ab stains neutrophils, you should switch
to another Ab (yours is cross reacting with NCA)

Hepatoma

Strong pos: Colonic & duodenal,

Muc ovarian, morules, yolk sac,
some NE, esp. midgut carcinoids
Heterogeneous: Panc/BD, Stomach

Carcinoid

TCC
TCC Bladder
Bladder

Neg: Lung, GU, Hepatoma (r+?), Breast (2.4%,

13/546 cases), Fem Gen (non-mucinous), ENT

(At ProPath, occas lung

and TCCs &1 hepatoma)

Hepatoma

Lung
Lung

Miller

IHC for Carcinoma of Unknown Primary

Met Ca: Useful Abs

N Cadherin
Pos:

Approach to the individual case

Liver

Serous ca (mucinous neg)

Endometrioid ca, Hepatoma,
Renal cell ca, Bile duct ca (also
Mesothelioma,Squamous ca, Thyroid,
Adrenal, SFT, ESS, Thymoma, Germ
cell tumors, NE tumors, Schwannoma,
etc.)

Neg:

Page 54

Clinical findings & history?

Is it really a carcinoma?

If carcinoma, what kind?

Where is it from?

Esoph-Gastric, Pancreas, Colon

Standard huge carcinoma panel

Standard huge carcinoma panel

CK7
CK20
Villin
CK5
p63
EMA
CK AE1/AE3
CK LMW
CK HMW
CDX2

TTF-1
Napsin A
Thyroglobulin
CEA (COL-1)
PSA, PSAP
PSMA, P501S
NKX3.1
Pax8
CK15
Vimentin
N-cadherin

CD56
GCDFP-15
Mammaglob
S100
ER, PR
Arginase-1
HBME-1
WT1
CA-125
Mesothelin
MUC2

MUC5AC
b-catenin
CD10
HepPar 1
CEA (poly)
MOC-31
Chromog
Synapto
Inhibin
A103
SALL4

Liver Biopsy (S97-5632)

Case History
87F with a breast mass and multiple liver
metastases. Two FNAs of breast were
negative, radiologic workup showed no other
tumor. A needle biopsy of the liver was
performed, revealing adenocarcinoma (Path
Comment: IHC can be performed if primary is
unknown). A breast bx followed, that showed

fat necrosis but no tumor. Four weeks later,

immunophenotyping of liver bx is requested by
oncologist.

CK7

CK20

Villin

Miller

IHC for Carcinoma of Unknown Primary

Page 55

Liver Biopsy (S97-5632)

CK7 Negative, CK20 Negative

Villin Positive
Stomach, Renal cell,
Lung Squamous,
Hepatoma (canalicular)
Prostate? (33% villin pos)

Villin Negative
Mesothelioma, Renal
Cell, Lung Squam.,
Hepatoma, Prostate,
(Breast)

Neuroendocrine

Rare: stomach, ov.,

pancreas, urothel.

Rare: mesothelioma,
breast, ov., urothel.,
pancreas.

CK7

AE1/AE3

CK20

Villin

Chg - Syn

CEA

Pictorial representation of this case

CK7 Negative, CK20 Negative

Villin Positive
Stomach, Renal cell,
Lung Squamous,
Hepatoma (canalicular)
Prostate? (33% villin pos)

Villin Negative
Mesothelioma, Renal
Cell, Lung Squam.,
Hepatoma, Prostate,
(Breast)

Neuroendocrine

Rare: stomach, ov.,

pancreas, urothel.

Rare: mesothelioma,
breast, ov., urothel.,
pancreas.

Stomach


 Breast

 Clinicians

Dx: Stomach most likely

(confirmed on gastric bx)

Clinicians barking up the wrong tree

85M, FNA of L1 vertebral mass (CM02-584)

Additional Immunostains

Chg

CK7

Villin

CK-hmw
Syn

PSA, PSAP, TTF1 are neg (CM02-584)

CK20

Miller

IHC for Carcinoma of Unknown Primary

Page 56

CK7
Positive, CK20 Positive
____________________________________________
CK7

Chg

Villin Positive
Stomach, Pancreas,
Bile ducts, Mucinous
Ovary, Small bowel
Rare: urothelium,
breast, prostate
(colon, endomet,
lung unlikely)

Villin Negative
Mucinous Ovary,
Urothelium, Breast

Syn

CK-hmw

CK20

Villin

(1/3 of mucinous
breast ca, most inv.
papillary ca breast)

Rare: GI, pancreas,

bile ducts

p63

CK5

Dx: Metastatic Urothelial Ca

Screening Immunostains

56M, bone & skin lesions, scalp bx.

Carcinoid vs plasmacytoid lymphoma vs

melanoma vs prostate ca??

Additional Immunostains

CK 7 and 20

Villin

Villin

Villin

CD45

Chg, Syn

S100, HMB45

PSA, PSAP

VS38

Vim

EMA

CK AE1/AE3

CK LMW

CK7 Negative, CK20 Negative

Villin Positive
Stomach, Renal cell,
Lung Squamous,
Hepatoma (canalicular)
Rare: mesothelioma,
breast, ov., urothel.,
pancreas.

Villin Negative
Mesothelioma, Renal
Cell, Lung Squam.,
Hepatoma, Prostate,
(Breast)
Rare: stomach, ov.,
pancreas, urothel.

Dx: Metastatic Hepatoma

Miller

IHC for Carcinoma of Unknown Primary

Conclusions

Immunohistochemistry

(IHC) plays
an important role in the evaluation of
metastatic tumors of unknown origin.

IHC

can save cost and discomfort of

further diagnostic procedures
(particularly with the use of tissue
transfer techniques)

IHC

can allow rapid institution of

appropriate therapy.

Immunohistochemistry

Page 57