35
Immunocytochemistry
Carcinoma of Unknown Primary
C arcinom a o f u n k n o w n p rim a ry (C U P ) accounts fo r 3 -5 % o f
all solid m alignancies. The scenarios o f CU P presentation are:
(1) m etastatic C U P p rim a rily to the liv e r o r to m u ltip le sites;
(2) m etastatic CU P to lym p h nodes in v o lv in g p rim a rily the
m e d ia stin a l-re tro p e rito n e a l, axillary, cervical, o r in g u in a l nodes;
(3) m etastatic C U P in the p eritoneal cavity inc lu d in g p a p illa ry
serous carcinom atosis in fem ales and the p eritoneal n o n p a p il-
la ry carcinom atosis in m ales o r females; (4) m etastatic CU P to
the lungs w ith parenchym al metastases o r isolated m alig n an t
p leural effusion; (5) m etastatic CU P to the bones; (6) m etastatic
C U P to the b rain; (7) m etastatic neuroend ocrine carcinom as;
and (8) m etastatic m elan om a o f an u n k n o w n prim ary. Therapeu-
tic strategies are scarce, and patients presenting w ith CU P carry a
p o o r prognosis, w ith survival rates o f 1 -2 years. Id e n tific a tio n o f
the organ o f o rig in o f the m etastatic tu m o r m ay be o f som e th e r-
apeutic significance. A d d itio n a l clinical history, such as sex, age,
rad iolog ic im pression, and pattern o f tu m o r d issem ination, is a
h e lp fu l step in evaluation. B ody effusions (p leural, peritoneal,
pericardial) are c om m o n presentations o f metastasis fro m an
u n k n o w n prim ary. Serous effusions are therefore c o m m o n ly
encountered specimens in cytology fo r d e te rm in a tio n o f CUP.
F our d iffe re n t tu m o r types are possible to recognize cytologi-
cally in effusions: adenocarcinom a (60% ), squam ous cell car-
cinom a (5% ), n on -sm all-c ell carcinom a (30% ), and sm all-cell
carcinom a (1 % ).84 C yto m o rp h o lo g y plays an im p o rta n t ro le to
gear the a ncillary studies (IC C ) fo r proper interp retation.
A d e n o c a rc in o m a
is the m ost com m on tu m o r to involve the
serous m em branes.85 The m ost c om m on sites o f adenocarcinom a
presenting as C U P are lung and colon (40% ), fo llow ed by liver,
pancreas, gastric sites, and kidney. The characteristic coexpression
patterns o f cytokeratins-7 and 20 are o f great u tility in surgical
pathology and have been u tilize d in cytology specimens as a
first panel o f im m u n o sta in to id e n tify the orig in in m alig nant
effusions.33 The m a jo rity o f lung, breast, gastric, and ovarian aden-
ocarcinomas show cytokeratin-7-positive/cytokeratin-20-negative
im m unop henotyp e. TTF-1, a m on oc lo na l antibody, has emerged
as a h ig h ly sensitive and specific m arker in d ifferentiating p u lm o -
nary adenocarcinomas fro m n o n p u lm o n a ry adenocarcinomas
in effusion cytology specimens.63,85 A strong nuclear staining o f
TTF-1 in lung adenocarcinomas has a sensitivity o f 88% and a
specificity o f 100% .61 TTF-1 in conjunction w ith p63 can be u ti-
lized to distinguish p u lm on ary sm all-cell carcinomas fro m p o orly
differentiated squam ous cell carcinomas. N ine ty-tw o percent o f
sm all-cell carcinomas express TTF-1, and none express p63.86 The
cytokeratin-7-negative/cytokeratin-20-positive im m u n o p ro file is
h ig h ly specific fo r colorectal orig in o f adenocarcinom a, although
a significant num b er o f gastric as w e ll as pancreatic adenocarcino-
mas e xhib it this im m u n o p ro file .87 C D X2 is a m arker th a t encodes
a transcription factor involved in d iffe re n tia tion o f colonic
ep ithelium . The sensitivity o f expression o f C D X2 in colorectal
adenocarcinom a is 97% .63 A weak p o sitivity o f C D X2 is seen in
gastric, biliopancreatic, and m ucinous ovarian adenocarcinomas.
A n im m un op a ne l com prising cytokeratin-7, cytokeratin-20, TTF-
1, and C D X2 is a reasonable panel fo r CU P in serous effusions.
S q u a m o u s c e ll c a rc in o m a ,
w hen p o orly differentiated, is d iffi-
cult to differentiate fro m p o o rly differentiated adenocarcinoma.
C ytokeratin-5/6, a m esothelial marker, and p63 im m unostains are
strongly expressed by squam ous cell carcinomas and can be added
to the panel.86 O ther probable sites o f origin, such as head and
neck, anorectal, and g enitourinary sites, should be considered.
In males, prostate carcinom a can present rarely as C U P in
pleural effusions. Prostate-specific antigen can be useful in this
s itu a tio n and could be added to the panel.88 In fem ales w ith peri-
toneal carcinom atosis o f u n k n o w n orig in, the m ost c om m o n site
o f o rig in is the m u lle ria n tract, fo llo w e d b y the gastrointestinal
tract, breast, and lym p h om a. In ord er to accurately determ ine
the lik e ly sites o f o rig in, accurate diagnostic m arkers are needed.
CA-125 is a m arker o f ovarian carcinom a b u t can be expressed in
a sm all percentage o f gastric and peritoneal carcinom as.89 W T1
can be added to the panel i f ovary o r breast is considered in the
d iffe re n tia l diagnosis.64 It is im p o rta n t to rem em ber th a t W T1
stains the background m esothelial cells as w ell.
Nonepithelial Malignancies Presenting as
Tumors of Unknown Origin
The c om m o n n o n e p ith e lia l neoplasm s th a t cause m alig n an t
effusions includ e m a lig n a n t m elanom a, lym p hom as, sarcomas,
germ cell tum ors, and som e pediatric m a lig n a n t tum ors. The
d e te rm in a tio n o f the o rig in o f n o n e p ith e lia l m alignancies on
p urely c ytom orp holog ic grounds is d iffic u lt. These tu m o rs often
e xh ib it a variety o f m orp h olog ic features th a t can d iffe r fro m
those o f the o rig ina l tu m o r and can overlap w ith various epithe-
lia l tu m o rs and m ay preclude the correct diagnosis. IC C plays an
im p o rta n t ro le in id e n tific a tio n o f these tum ors.
Malignant Melanoma
E ffusions due to m elan om a are relative ly rare and com prise
o n ly 2 -3 % o f the m a lig n a n t effusions.
Im m u n oc ytoc he m istry has proved m ost practical in the
id e n tific a tio n o f m elanom a. The tu m o r cells are im m unoreac-
tive fo r H M B -45 (a n tim e la n o m a ) (ab out 60% ), S-100 p rotein
(95% ), and o th e r m elan om a markers. V im e n tin , w h ile positive
in m elanom as, is also p ositive in m esotheliom as, sarcomas, and
som e carcinomas. M a lig n a n t m elan om a cells are im m unoreac-
tive fo r M ART-1 (m elan A ) in 80% o f m elanom as. C ytokeratin
im m u n o re a c tivity has been reported in m a lig n a n t m elan om a to
range fro m 1 to 27% o f cases, b u t the im m u n o s ta in in g in these
cells is u su a lly w eak and focal (Fig. 35.21).90
S a rc o m a s
account fo r o n ly 3 -6 % o f m a lig n a n t effusions and
u su a lly occur in the setting o f a k n o w n p rim a ry tu m o r; they
rarely can present as tu m o r o f u n k n o w n orig in. S ynovial sar-
coma, e p ith e lio id sarcoma, vascular tum ors, leiom yosarcom a,
end om etrial strom a l sarcoma, and GIST are som e o f the sarco-
mas th a t can present as tu m o rs o f u n k n o w n orig in.
Regarding
p e d ia tric tu m o rs ,
m a lig n a n t effusions caused by
n o n e p ith e lia l neoplasm s are m ore fre q ue ntly encountered in
child ren th a n in adults. The m ost c om m o n causes o f m a lig n a n t
effusions in child ren are lym p h o m a and leukem ia, fo llo w e d by
n o n e p ith e lia l neoplasm s inc lu d in g W ilm s ' tu m o r, neuroblas-
tom a, Ewing's sarcoma, and em b ryonal rhabdom yosarcom a.
G e rm c e ll n e o p la sm s
are c om m o n in child ren as w e ll as young
adults. Sem inom as are occasionally fo u n d in m a lig n a n t e ffu -
sions in pleura, p eritoneum , and hydrocele sac.91
D y s g e rm in o m a
is the fem ale counterp art o f the sem in o m a and rarely presents
in effusions. These tu m o rs can occur as pure tu m o rs o r as m ixed
tu m o rs w ith o th e r germ cell tum ors, and th e y have p o ten tial to
metastasize. C ell blocks are proven to be a useful to o l to p erform
the im m u n o h isto c h e m ic a l studies to diagnose and d ifferentiate
the d iffe re n t com ponents o f the m ixed germ cell tum ors.
1065
previous page 1049 ComprehensiveCytopathology 1104p 2008 read online next page 1051 ComprehensiveCytopathology 1104p 2008 read online Home Toggle text on/off