PATHOlOGY

RESEARCH AND PRAcnCE

© Urban &Fischer Verlag
httpllwww.urbanfischer.defjournalslprp

Is ITF1 a Good Immunohistochemical Marker to

Distinguish Primary from Metastatic Lung
Adenocarcinomas?

Jorge S. Reis-Filho1, Carla Carrilho 2, Carla Valenti 1, Dina Leitao 1,

Carlos A. Ribeiro 3 , Silvana G. A. Ribeiro 3 and Fernando C. Schmitt 1

'Institute of Pathology and Molecular Immunology at Porto University, IPATIMUP, Porto,

Portugal;
'Department of Pathology, Medical Faculty of Eduardo Mondlane University, Maputo,
Mozambique;
JDepartment of Pathology, Medical Faculty of Minas Gerais, Brazil

Summary Introduction

To evaluate thc immunohistochcmical expression of Thyroid transcription factor 1 (TIFI) is a home-

thyroid transcription factor I (TTFI) in primary and
metastatic pulmonary adenocarcinomas, and test the
diagnostic accuracy of this antibody, two surgical
pathologists independently evaluated 34 caseS of ade-
nocarcinomas in the lung without clinical data and tried
to distinguish between primary and mctastatic cases
using histological criteria exclusively. Thirteen cases
were primary in the lung and 21 were metastases of ex-
trapulmonary adenocarcinomas: 6 from the endometri-
um, 4 from the ovary, 3from the colon, 2 from the kid-
ney, 2 from the breast, 2 from the liver and 1 from the
prostate. Afterward, the immunoreactivity of TTFI in
these neoplasms was evaluated and correlated with
morphological and clinical data. The two pathologists
were able to diagnose only 5 out of 13 cases of primary
lung adenocarcinomas (sensitivity of 38.46%) and also
misdiagnosed two primary malignancies as metastases.
After correlation with TIFI data, the sensitivity in-
creased to 61.53%. The specificity ofTIFI was 100%.
In conclusion, TIFI is a highly specific marker for pri-
mary lung adenocarcinomas, and should be included in
a panel of antibodies for the differential diagnosis be-
tween primary and metastatic adenocarcinomas of the
lung.
Key words: Thyroid transcription factor - Lung cancer
- Adenocarcinoma - Immunohistochemistry
odomain-containing nuclear transcription protein of the
Nkx2 gene family that is normally expressed in certain
areas of the central nervous system during embryogene-
sis {l, 2, 3/. TTFI expression is also well documented
in tbe thyroid and lung 11, 2, 3}. In these tissues, TIF 1
activates the transcription of thyroglobulin and thy-
roperoxidase in the thyroid and the transcription of the
surfactant B protein gene in lung epithelial cells [2/. In
fact, some authors {l, 2, 3/ have shown that Clara cells
and type II pneumocytes present high levels of this pro-
tein; these have been confirmed by immunohistochemi-
cal {l, 2J and RNAse protection assays {3j.
Recently, TIFl has received great attention among
surgical pathologists as a good immunohistochemical
marker for lung and thyroid carcinomas {l, 2}. In differ-
ent types of lung carcinomas, only adenocarcinomas,
large cell carcinomas and small cell carcinomas are posi-
tive for this protein fl, 2, 3, 4, 5, 6}. Antibodies against
this transcription factor have been widely tested as mark-
ers to disti nguish metastatic small cell carcinomas from
Merkel cell carcinomas {7/, as well as to distinguish lung
adenocarcinomas from mesotheliomas [2. 8. 9].
Address for correspondence: Fernando C. Schmitt, IPA-
TIMUP - Instituto de Pathoiogia e Imunologia Molecular da
Universidadc do Porto, R. Roberto Frias, SIN. 4200 Porto,
Portugal.

Pathol. Res. Pract. 196: 83S-B40 (2000) 0344·033812000/ 196112-835 $15.0010

Table I. Clinical. histological, and immunohistochemical profile of the adenocarcinomas evaluated in this study 00
0',
""
Case Primary site Morphological diagnosis Histological features TIFl TIFI Diagnosis considering TIFI !-
(tumor) Lung
;<l
I Breast Metastatic adenocarcinoma Poorly differentiated adenocarcinoma with solid areas + Metastatic adenocarcinoma <;;>
'""
2 Breast Adenocarcinoma* Carcinoma with solid areas + Adenocarcinoma'" :!1
3 Colon Metastatic adenocarcinoma Tubular adenocarcinoma + Metastatic adenocarcinoma 5'
0
4 Colon Metastatic adenocarcinoma Mucinous adenocarcinoma + Metastatic adenocarcinoma :!.
5 Colon Metastatic adenocarcinoma Mucinous adenocarcinoma + Metastatic adenocarcinoma ~
6 Endometrium Metastatic adenocarcinoma Poorly differentiated adenocarcinoma with solid areas + Metastatic cndometrioid
adenocarcinoma
7 Endometrium Metastatic adenocarcinoma Adenocarcinoma with mucinous foci + Metastatic adenocarcinoma
8 Endomctrium Metastatic adenocarcinoma Poorly differentiated adenocarcinoma with squamous foci + Metastatic endometrioid
adenocarcinoma
9 Endometrium Metastatic adenocarcinoma Cribriform adenocarcinoma + Metastatic adenocarcinoma
10 Endometrium Metastalic adenocarcinoma Endometrioid adenocarcinoma + Metastatic endometrioid
adenocarcinoma
II Endometrium Metastatic adenocarcinoma Endometrioid adenocarcinoma + Metastatic adenocarcinoma
12 Kidney Metastatic clear cell carcinoma Clear cell carcinoma + Metastatic clear cell carcinoma
13 Kidney Metastatic clear cell carcinoma Clear cell carcinoma + Metastatic clear cell carcinoma
14 Liver Metastatic adenocarcinoma Carcinoma with solid areas + Metastatic adenocarcinoma
15 Liver Metastatic adenocarcinoma Carcinoma with solid areas + Metastatic adenocarcinoma
16 Ovary Adenocarcinoma* Poorly differentiated adenocarcinoma with mucinous foci + Adenocarcinoma*
17 Ovary Adenocarcinoma* Poorly differentialed adenocarcinoma + Adenocarcinoma*
18 Ovary Metastatic adenocarcinoma Mucinous adenocarcinoma + Metastatic adenocarcinoma
19 Ovary Metastatic adenocarcinoma Moderately differentiated adenocarcinoma + Metastatic adenocarcinoma
20 Ovary Metastatic adenocarcinoma Mucinous adenocarcinoma + Metastatic adenocarcinoma
21 Prostate Metastatic adenocarcinoma Adenocarcinoma with cribriform and solid areas + Metastatic adenocarcinoma
22 Lung Adenocarcinoma* Solid carcinomas with mucus formation + Adenocarcinoma*
23 Lung Metastatic carcinoma Solid carcinomas with mucus formation + Metastatic clear cell carcinoma
24 Lung Adenocarcinoma* Solid carcinomas with mucus formation + Adenocarcinoma*
25 Lung Metastatic adenocarcinoma Solid carcinomas with mucus formation + Metastatic adenocarcinoma
26 Lung Adenocarcinoma* Solid carcinomas with mucus formation + Adenocarcinoma*
27 Lung Primary Adenocarcinoma Acinar adenocarcinoma with focal squamous differentiation + + Primary Adenocarcinoma
28 Lung Adenocarcinoma* Acinar adenocarcinoma + + Primary Adenocarcinoma
29 Lung Primary Adenocarcinoma Acinar adenocarcinoma + + Primary Adenocarcinoma
30 Lung Adenocarcinoma* Solid carcinomas with mucus formation + + Primary Adenocarcinoma
31 Lung Primary Adenocarcinoma Acinar adenocarcinoma + + Primary Adenocarcinoma
32 Lung Primary Adenocarcinoma Bronchioloal veolar carcinoma + + Primary Adenocarcinoma
33 Lung Metastatic adenocarcinoma Acinar adenocarcinoma + + Primary Adenocarcinoma
34 Lung Primary Adenocarcinoma Papillary adenocarcinoma + + Primary Adenocarcinoma

*: cases in which a diagnosis of primary or metastatic adenocarcinoma could not be achieved based on morphological data
 TIF-I in Lung Adenocarcinomas ' 837

The differential diagnosis of primary and metastatic
lung adenocarcinomas presents a problem in routine
surgical pathology, mainly when there is not a con-
firmed primary site {l, 5, 6, 10]. In this paper, the au-
thors evaluated the role of TIFI as an ancillary tech-
nique in the differentiation of primary and metastatic
lung adenocarcinomas, testing its sensitivity and speci-
ficity, as well as describing the improvement in the di-
agnosis of primary lung adenocarcinomas,
colon, 2 from the kidney, 2 from the breast, 2 from the
liver, and I from the prostate.
The morphological diagnoses of the cases are sum-
marized in Table 1. Using only the morphological crite-
ria,S cases were diagnosed as primary adenocarcino-
mas, 22 as metastatic adenocarcinomas, and in 7 cases
it was not possible to determine, based on morphologi-
cal criteria, if the tumor was primary or metastatic,
Immunohistochemical analysis

Methods TTFI was positive in 8 cases (Fig, I), and all of these
cases were primary lung adenocarcinomas (Table 2).
Two surgical pathologists, without any clinical data, inde- According to the WHO's classification, these cases cor-
pendently evaluated 34 cases of adenocarcinomas in the lung, responded to 4 pure acinar adenocarcinomas, one acinar
In this fim analysis, the pathologists tried to diagnose the ade- adenocarcinoma with focal squamous differentiation,
nocarcinomas as primary or metastatic using only morpholog- one solid carcinoma with mucus formation, one bron-
ical criteria [11].
chioloalveolar carcinoma, and one papillary adenocar-
A second step was to evaluate the immunocxpression of
TIFI in the neoplastic cells and in the normal lung, trying to cinoma, Five primary lung adenocarcinomas were neg-
confirm the diagnosis or primary lung carcinomas. ative, and none of the metastatic cases presented any
The immunohistochemical assays were performed accord- positivity for this antibody, The negative primary lung
ing to the avidin-biotin-peroxidase technique as previously
describcd [12/, with monoclonal antibody raised against thy-
roid transcription factor-I (8070311 culture supernatant -
Neomarkers, Union City, CAl, To exclude equivocal reac-
tions, an at least moderate staining intensity in more than 10%
of the tumor cells was required as a relevant positive reaction
'I]. Only nuctear staining was considered as positive,
After that, the two surgical pathologists and a third pathol-
ogist subclassified the primary lung adenocarcinomas accord-
ing to the WHO's classification [13].
The sensitivity and the specificity of the immunostaining
were defined as the following:
True positives
Sensitivity = - --,-,--'--,::-c----,--
True posiLives + False negatives

True negatives
Specificity
True negatives + False positives
Fig, I, Photomicrography of a primary lung adenocarcinoma
The authors also evaluated the number of cases in which showing nuclear positivity for TIFI. (Avidin-biotin-peroxi-
TIFI had changed the diagnosis, improving the definition of dase/Diaminobenzidine x400)
primary or metastatic lung adenocarcinomas.
In the present study, a definitive classification of the adeno-
carcinomas as primary or metastatic was provided according Table 2. TIFl immunoreactivity in primary and metastatic
to clinical data, radiologicaJ features, patient'S follow-up and
lung adenocarcinomas
in some cases by autopsy, The cases were considered as prima-
ry lung tumors when no other primary site had been found, TIF immunoreactivity

Cases Positive Negative Total

Results
Primary lung adenocarcinoma 8 5 13
Evaluation of the morphological data Metastatic adenocarcinomas 0 21 21
The cases evaluated in this study included 13 prima- Total 8 26 34
ry lung adenocarcinomas, and 21 metastatic adenocar-
cinomas, Among the metastases, the primary sites were: Sensitivity = 61.53%
6 from the endometrium, 5 from the ovary, 3 from the Specificity = 100%
838 . J. S. Reis-Filho et al.
Table 3. Histological and immunohistochemical features of
the primary lung adenocarcinomas
Histological diagnosis
Solid carcinomas with mucus formation
Solid carcinomas with mucus formation
Solid carcinomas with mucus formation
Solid carcinomas with mucus formation
SaUd carcinomas with mucus formation
Acinar adenocarcinoma with focal squamous
differentiation
Acinar adcnocan.:inorna
Acinar adenocarcinoma
Solid carcinomas with mucus formation
Acinar adenocarcinoma
Bronchioloalveolar carcinoma
Acinar adenocarcinoma
Papillary adenocarcinoma
Fig. 2. Normal lung alveolar epithelial cells showing nuclear
positivity for TIFt (arrows) . In the center, nests of cells from
a metastatic adenocarcinoma without nuclear staining.
(Avidin-biotin-peroxidase/Diaminobenzidine x400)
adenocarcinomas were all solid carcinomas with mucus
formation (Table 3). In all cases, the adjacent lung tis-
sue presented immunoreactivity for TTFI in the nuclei
of type II pneumocytes and Clara cells (Fig. 2).
Evaluation of the diagno5tic accuracy with and without
ITFl
When the pathologists correlated the TTF I im-
munostaining with the morphological data, the five
cases diagnosed as primary lung adenocarcinomas were
confirmed. Among the 8 cases in which differentiation
between metastatic and primary adenocarcinoma was
not possible by morphology, 2were positive and 6 were
negative for TTFI. One of the 22 cases diagnosed mor-
phologically as metastasic was positive for TIFI.
TIFI
(tumour)
Compari50n of the hi5tological and immunohi5tochemi-
cal finding5 with the clinical data
The clinical, pathological, and immunohistochemical
profile of all cases is described in Table I.
The sensitivity of the surgical pathologists without
+ any clinical or immunohistochemical data to diagnose
primary lung adenocarcinomas was 38.46% (5 out of 13
+ cases). Considering TTFI results, 3 additional cases
+ could be included among the primary cancers, leading
+ to a sensitivity of 61.53%.
+ In one case misdiagnosed as a metastasis and in 2
+
+
cases without a definition of primary or metastatic,
+ TTFI immunostaining was positive, changing the diag-
noses to primary lung adenocarcinomas. After this cor-
relation, 8 cases were considered primary malignancies,
and 20 metastases. In 6 cases in which the distinction of
primary or metastases could not be achieved, TTF I did
not contribute further to the diagnosis.
Compared to the clinical data, all 8 cases defined as
primary adenocarcinomas were correctly diagnosed.
Eighteen out of the 20 cases diagnosed as metastases
after TTFI evaluation were correctly diagnosed; 2 cases
were primary adenocarcinomas misinterpreted as
metastases. Considering the 6 adenocarcinomas in
which the diagnosis of primary or metastatic adenocar-
cinomas could not be achieved by morphology and
were negative for TTFI, 3 were metastatic and 3 were
primary lung adenocarcinomas. In all these cases, the
internal control was positive.
Discussion
Primary lung adenocarcinomas are one of the most
prevalent causes of metastatic disease with an unknown
primary site; furthermore, the lungs are frequent targets
for metastatic carcinomas, coming mainly from the kid-
ney, breast, prostate, and liver, among others [1,3,5,6,
9, II/. The distinction hetween primary and secondary
adenocarcinomas affecting the lungs is a difficult task,
especially when we have a small biopsy or fine needle
aspiration material.
Recently, several reports have been published con-
cerning new immunohistochemical antibodies against
proteins and tissue-specific transcription factors as
markers to distinguish primary from metastatic lung
adenocarcinomas [l, 5, 6, 9, 10]. TTFI is considered
one of the best markers to distinguish primary from
metastatic primary lung adenocarcinomas II , 5, 6, 10};
besides, it is commercially available and able to be em-
ployed in paraffin-embedded tissue specimens [9].

Table 4. TTF1 immunoreactivity in primary and metastatic lung adenocarcinomas in different reported series
Authors (Reference) Primary pulmonary tumors
(# of adenocarcinomas)
Kaufmann et al. [l] 138 (98)
Khoor et al. [2/ 370 (208)
Fabbro et al. [3 J* 43 (15)
Bejarano et al. [4/ 57
Harlamert et al. [51 41 (21)
Di Loreto et al. [61 33 (33)
Di Loreto et al. [81 17 (8)
Ordonez NG {9J 40 (40)
Borrinski et al. [10/ 18 (18)
Present study 13 (13)
*: studies in which only primary lung tumors were evaluated
The sensitivity of TTFI in lung adenocarcinomas in
the different reported series ranged from 57.5 % to 76%
(Table 4) [I, 2, 3, 4, 5, 6, 8,9, 1Oj. Excluding thyroid
neoplasms, the TTFI specificity ranged from 99% {4/
to 100% for primary lung carcinomas [1 , 2, 3, 5, 6, 8,
9]. In our series, the sensitivity was 61.53% and the
specificity 100%. Five cases of primary lung carcino-
mas werc negative for TTFI; in all these cases, the au-
thors observed that they were solid adenocarcinomas
with mucin production. Among the TTF I primary lung
adenocarcinomas, most of them were acinar adenocar-
cinomas and only one solid adenocarcinoma with
mucin production was positive for TTFI. These find-
ings are in agreement with those published by Kauf-
mann et a1. [II, who reported a TTFI sensitivity of 10%
for mucinous lung adenocarcinomas.
The authors also evaluated TTFl's role in the im-
provement of the morphological diagnosis of primary
lung adenocarcinomas. Without TTFI the ability of the
pathologists to diagnose an adenocarcinoma accurately
as a primary lung malignancy was just 38.46%. After
TTFI, it improved to 61.53 %. Furthermore, one case
misdiagnosed as a metastasis based solely on morpho-
logical grounds could be correctly diagnosed as a pri-
mary lung adenocarcinoma using TTFI.
Another interesting feature of the present study is the
ability of TTFI to infer a pulmonary origin for un-
known primary site adenocarcinomas in which mor-
phology does not allow the distinction between primary
and metastatic lung adenocarcinomas. In the present se-
ries, in 8 cases a distinction between primary and
metastatic adenocarcinoma could not be achieved; with
TTFL two of these cases could be diagnosed as pul-
monary malignancies, lacking any clinical data.
In conclusion, TTFI is a good marker to prove or to
exclude the lung origin of adenocarcinomas with a sen-
sitivity of61.53 % and 100% specificity (excluding thy-
roid neoplasm). However, is important to stress that
TTF-I in Lung Adenocarcinomas 839
Extrapuhnonary tumors Sensitivity (%) Specificity (%)
276 75 98
95 76 100
27
177 76 99
6 76 100
24 57.5 100
62.5
145
12 67 100
21 61.5 100
even primary lung adenocarcinomas presenting mucin
production arc usually negative for TTFI. Taking this
aspect into account, we suggest that TTFI be incorpo-
rated in a panel for the differential diagnosis of adeno-
carcinomas with an unknown primary site, mainly in
small specimens and in fine needle aspiration biopsies.
References
I. Kaufmann O. Dietel M (2000) Thyroid transcription fac-
tor-I is the superior immunohistochemical marker for
pulmonary adenocarcinomas and large cell carcinomas
compared to surfactant proteins A and B. Histopathology
36:8-16
2. Khoor A. Whitsett JA, Stahlman MT, Olson SJ, Cagle IT
(1999) Utility of surfactant protein B precursor and thy-
roid transcription factor I in differentiating adenocarcino-
ma of the lung from malignant mesothelioma. Hum
Pathol 30: 695-700
3. Fabbro 0 , Oi Loreto C, Stamerra 0, Beltrami CA, Loni-
gro R, Damante G (1996) TIF-J gene expression in
human lung tumours. Eur J Cancer 32A: 512-517
4. Bejarano PA. Baughman RP, Biddinger PW, Miller MA,
Fenoglio-Preiser C, al-Kafaji B, Oi Lauro R, Whitsett JA
(1996) Surfactant proteins and thyroid transcription fac-
tor-I in pulmonary and breast carcinomas. Mod Pathol 9:
445-452
.5. Harlamert HA, Mira J. Bejarano PA, Baughman RP,
Miller MA, Whitsett JA, Yassin R (1998) Thyroid tran-
scription factor-\ and cytokeratins 7 and 20 in pulmonary
and breast carcinoma. Acta Cytol42: 1382-1388
6. Oi Loreto C. Di Lauro Y, Puglisi F, Damante G, Fabbro
D, Beltrami CA (1997) lnununocytochemical expression
oftjssue specific transcription factor-l in lung carcinoma.
J Clin Pathol50: 30-32
7. Byrd-Gloster AL, Khoor A, Glass LF, Messina JL, Whit-
sett JA, Livingston SK, Cagle PT (2000) Differential ex-
pression of thyroid transcription factor I in small cell
lung carcinoma and Merkel cell tumor. Hum Pathol 31:
58-02
840 J. S. Reis-Filho et al.
8. Di Loreto C, Puglisi F ,Di Lauro V, Damante G, Beltrami
CA (1998) TTF-I protein expression in pleural malignant
mesotheliomas and adenocarcinomas of the lung. Cancer
Lett 124: 73-78
9. Ordonez NG (2000) Value of thyroid transcription factor-
I, E-cadherin, BG8, WTI, and CD44S immunostaining
in distinguishing epithelial pleural mesothelioma from
pulmonary and non pulmonary adenocarcinoma. Am J
Surg Pathol24: 598--{j06
10. Bohinski RJ, Bejarano PA, Balko G, Warnick RE, Whit-
sett JA (1998) Determination of lung as the primary site
or cerebral metastatic adenocarcinomas using monoclon-
al antibody to thyroid transcription factor-I. J Neurooncol
40: 227-231
11. Colby TV, Koss MN, Travis WD (1995) Tumors
metastatic to lung. In: Tumors of the respiratory tract. 3rd
ed., pp 515-538. Armed Forces Institute of Pathology,
Washington DC
12. Schmitt FC, Soares R,Cirnes L, Seruea R (1998) P53 in
breast carcinomas: association between presence of muta-
tion and immunohistochemical expression using a semi-
quantitative approach. Pathol Res Pract /94: 815-819
13. World Health Organization (1982) The World Health Or-
ganization hi stological typing of lung tumors. Am J Cliln
Pathol 77: 23- \36
Received in revi sed version: August 3 ,2000
Accepted: Augu st 3, 2000