2015, Volume 31, Number 2, Page(s) 081-088
The Value of Cytokeratin 5/6, p63 and Thyroid Transcription Factor-1 in Adenocarcinoma, Squamous Cell Carcinoma and Non-Small-Cell Lung Cancer of the Lung
Asuman ARGON1, Deniz NART2, Ali VERAL2
1Department of Medical Pathology, Ýzmir Bozyaka Education and Research Hospital , ÝZMÝR, TURKEY
2Department of Pathology, Ege University Faculty of Medicine, ÝZMÝR TURKEY
Keywords: Adenocarcinoma, Squamous cell carcinoma, Immunohistochemistry, Lung cancer
It is now important to distinguish between adenocarcinoma
and squamous cell carcinoma of the lung because of target-specific
treatments. Our study aimed to study the efficiency of Thyroid
Transcription Factor-1 (TTF-1), cytokeratin 5/6 (CK5/6) and p63
in distinguishing between adenocarcinoma and squamous cell
carcinoma and to study the contribution of these markers to the
diagnosis in non-small cell lung cancer.
Material and Method: Immunohistochemically, TTF-1, CK 5/6 and
p63 were used in 72 cases of squamous cell carcinoma, 19 cases of
adenocarcinoma, and 29 cases of non-small cell lung cancer whose
final diagnosis was decided with the subsequent resection material.
The specificity, sensitivity, and positive and negative predictive value
were calculated for each marker.
Results: TTF-1 positivity was seen in none of the 72 squamous cell
carcinomas but in all of 19 adenocarcinoma cases. CK5/6 negativity
was seen in all cases of adenocarcinoma and in two cases of squamous
cell carcinoma. p63 was positive in all squamous cell carcinomas and
in 4 adenocarcinomas. Cytokeratin 5/6, p63 positivity and TTF-1
negativity were observed in 17 non-small cell lung cancers whose final
diagnosis was squamous cell carcinoma. None of the 12 non-small
cell lung cancers whose final diagnosis was adenocarcinoma
exhibited positive staining for CK5/6. However, p63 staining was not
seen in the biopsy but was focal in the surgical specimen in one case.
All the 12 non-small cell lung cancers whose certain diagnosis was
adenocarcinoma were positive for TTF-1. TTF-1, CK 5/6 and p63
seem to be useful for differentiating adenocarcinoma from squamous
cell carcinoma with 100% specificity, 100% sensitivity and 100%
specificity, 97% sensitivity and 87% specificity, and 100% sensitivity,
Conclusion: We concluded that TTF-1 is a reliable marker for
subtyping lung cancer. Different staining patterns can be seen with
CK5/6 and p63; however, if they are used together with TTF-1 and
interpreted correctly, they can be of help for the final diagnosis even
in cases in which the morphology is unclear.
Lung cancer is the number one cause of cancer-related
mortality in men and women worldwide1
histological types of lung cancer have been described.
Historically, lung cancers were divided into two major
groups as small-cell lung cancer and non-small-cell lung
. Adenocarcinoma (AC), squamous
cell carcinoma (SCC) and large cell carcinoma are the
most common types of NSCLC3
. It is important to
distinguish between AC and SCC because effectiveness of
target-specific treatments (such as monoclonal antibody
inhibitor of vascular endothelial growth factor (VEGF)
and epidermal growth factor receptors–tyrosine kinase
inhibitors) is different for AC and SCC. For example,
bevacizumab, a VEGF drug/inhibitor, is an important drug
in the treatment of some ACs. However, when it is used for
SCC, it may cause life-threatening hemorrhage3,4
the pulmonary carcinomas are classified, their histological
features are considered first. However, morphology alone
may not be enough in some cases due to small biopsies,
poor differentiation, artifacts, and the heterogeneity
of the tumor5,6
. Immunohistochemistry (IHC) and
gene expression profiling have been studied to solve this
problem, but usually immunohistochemistry is the most
Our study aimed to study the efficiency of Thyroid
Transcription Factor-1 (TTF-1), cytokeratin 5/6 (CK5/6)
and p63 in distinguishing between AC and SCC and to
study the contribution of these markers to the diagnosis in
|Study Design and Specimen Characteristics
Malignant lung specimens were retrieved between
December 2008- December 2010, using an electronic
database. Nonepithelial tumors, metastatic tumors,
large cell carcinoma and small cell lung carcinomas were
excluded. All the tumor slides stained with Hematoxylin
and Eosin (H&E) of the 175 specimens were reviewed by
three pathologists, at least one of whom was a pulmonary
pathologist. In biopsy specimens, the tumors were classified
as SCC if they contained intercellular bridges and/or
showed keratinization, and as AC if they clearly showed
gland formation or mucin production. Extracellular mucin
or a goblet cell-like appearance were considered as mucin
production. To be certain about the tumor classification for
NSCLC in biopsy, we required a surgical specimen as the
gold standard for diagnosis. Three pathologists blinded to
the immunohistochemical findings subtyped the resected tumors on the basis of H&E slides according to the World
Health Organization classification of lung tumors9
they changed the first diagnosis if necessary.
The cases were included in the study1 if the biopsy
had typical AC morphology2, if the biopsy had
typical SCC morphology3 when the pathologists
agreed on the diagnosis of NSCLC and if the tumor was
subsequently resected at our institution4 if blocks
from biopsy and lobectomy specimens were available for
immunohistochemical study. Fifty-five of the 175 cases did
not meet these criteria and were excluded. The remaining
120 cases were included in this study. Using morphological
evaluation, the cases were divided into four groups: AC,
SCC, NSCLC-SCC whose final diagnosis was SCC in the
subsequent resection material, and NSCLC-AC whose final
diagnosis was AC in the subsequent resection material.
Four-micron-thick sections were obtained for IHC
investigation. Immunohistochemistry was performed by a
standard protocol on a Ventana Discovery XT automated
stainer (Ventana Medical Systems, Tucson, AZ, USA). Only
primary antibodies, cytokeratin 5/6 (Clone D5/16B4, Dako,
1:200 dilution), p63 (Clone 4A4, Dako, 1:700 dilution) and
TTF-1 (Clone SPT24, NovoCastra, 1:50 dilution) were
added manually and incubated in 37°C for 32 minutes.
Pneumocytes were considered as internal controls for TTF-
1 positivity, and bronchial basal cells for p63 and CK5/6
Immunohistochemically, cytoplasmic staining was
considered positive for cytokeratin 5/6 and nuclear staining
was considered positive for TTF-1 and p63. All the areas of
biopsy materials and at least 10 high power fields in tumor
surgical materials were evaluated and the mean percentage
of the positive tumor cells was determined. The staining
rate was divided into four categories as follows: 0, <1%; 1+,
1–33%; 2+, 34–66%; 3+, >66%. Intensity of staining was
graded as negative (0, no reactivity); weak (1+, less than
normal cells); moderate (2+, same as normal cells); and
strong (3+, stronger than normal cells)9,10. When the
staining was heterogeneous, the prominent pattern became
the basis. The demographic features (age, gender) were
obtained from the hospital archive system. For NSCLC
cases, we also assessed whether immunohistochemical
staining on biopsies correlated with staining on the
subsequent surgical specimens as lung cancers are known
to show morphologic heterogeneity in different areas of the
The statistical analyses were performed with SPSS software,
version 20.0. The specificity, sensitivity, positive predictive
value and negative predictive value were calculated for each
marker. In addition, the relationship between the staining on
biopsy and resection materials was analyzed using Pearson
Chi-Square test, Chi-Square test and Fisher’s Exact Test. p
values of ≤0.05 were considered statistically significant.
There were 104 (86.7%) males and 16 (13.3%) females
included in the study with a mean age of 62.30±8.74 and a
median age of 62.00 (Min: 37, max: 82) years. Histologically, 72 (60.0%) cases were SCC and 19 (15.8%) cases AC. The
remaining 29 cases were NSCLC, 17 (14.2%) of which were
NSCLC-SCC and 12 (10.0%) were NSCLC-AC. The results
of immunohistochemical staining for each histological
subtype of carcinoma are summarized in Table I. The
relationship between the stainings on biopsy and resection
materials and p values are shown in Table II.
Click Here to Zoom
|Table I: The results of immunohistochemical staining for each histological subtype of carcinoma and the specificity, sensitivity, positive
predictive value and negative predictive value
Click Here to Zoom
|Table II: The results of immunohistochemical staining for each histological subtype of carcinoma in biopsy and resection material and
the p value
Cytokeratin 5/6: Of the 72 SCC cases, 56 had 3+ staining, 4
had 2+ and 10 had 1+. CK5/6 negativity was encountered
in two SCC cases that had TTF-1(-)/p63(+) (Figure 1A-C).
In the cases that stained positive, intensity of staining was
strong in 54 cases, moderate in 13 cases and weak in 3 cases.
However, no staining was seen in any of the 19 AC cases.
Click Here to Zoom
|Figure 1: A) SCC case whose morphology is typical in biopsy (H&E; x20), B) CK5/6 negativity (CK5/6; x10), C) p63 positivity (p63; x10),
D) TTF-1 negativity (TTF-1; x10).
When focal staining was considered positive, the specificity,
sensitivity, positive predictive value and negative predictive
value were 1.000 (95% Confidence Interval: 0.1000-1.0000),
0.978 (95% Confidence Interval: 0. 9467-1.0000), 0.939
(95% Confidence Interval: 0.8580-1.0000) and 1.000 (95%
Confidence Interval: 0.1000-1.0000), respectively.
Positive staining was observed in all of the 17 NSCLCSCC
cases (Figure 2A-B). While the staining rate did not
change in the biopsy and the surgical specimen in 15 of
the 17 cases, it increased in the surgical specimen in the
remaining 2 cases. No statistically significant difference
was found between the staining rates of the biopsy and
surgical specimens (p=0.500). None of the 12 NSCLC-AC
cases exhibited positive staining either in biopsy or surgical
specimens so no statistical calculations were performed.
When the staining intensity in the NSCLC-SCC cases and
that in surgical specimens were compared, it did not change
in 15 cases but it increased in the surgical specimens of 2
cases. In the NSCLC-SCC cases, the intensity of staining
displayed no statistically significant difference between
biopsy and surgical specimens (p=0.500).
Click Here to Zoom
|Figure 2: A) NSCLC case whose final diagnosis is SCC in its subsequent resection material (H&E; x20), B) CK5/6 focal-weak positivity
(CK5/6; x20), C) p63 positivity (p63; x20), D) TTF-1 negativity (TTF-1; x20).
p63: We observed 3+ staining in 71 of the 72 SCC cases and
2+ staining in one. There was no staining in 15 of the 19 AC
cases. However, p63 positivity was observed in 4 AC cases
that had TTF-1(+)/ CK5/6 (-) (Figure 3A-D). Two cases had
1+ staining. One of the remaining two had 2+ staining and
the other had 3+ staining. Intensity of p63 staining was weak
in 3 of the 4 AC cases and it was moderate in one. When
the focal staining was considered positive, the specificity,
sensitivity, positive predictive value and negative predictive value were 0.871 (95% Confidence Interval: 0.7530-1.0000),
1.0000 (95% Confidence Interval: 1.0000-1.0000), 1.000
(95% Confidence Interval: 1.0000-1.0000) and 0.957 (95%
Confidence Interval: 0,9158- 1.0000), respectively.
Click Here to Zoom
|Figure 3: A) AC case whose morphology is typical in biopsy (H&E; x10), B) CK5/6 negativity (CK5/6; x10), C) p63 positivity (p63; x10),
D) TTF-1 positivity (TTF-1; x10).
In all of the 17 NSCLC-SCC cases (Figure 2C), the staining
rate and intensity were the same both in the NSCLC areas
and surgical specimens so no statistical assessment was
conducted. Eleven of the 12 NSCLC-AC had no staining
both in biopsy and surgical specimens. In the other case,
staining was not seen in the biopsy but 1+ staining was
seen in the surgical specimen. That finding was not found
statistically significant (p=1.000).
TTF-1: TTF-1 staining was not seen in any of the 72 SCC
cases while it was seen in all AC cases. In 16 of the AC cases,
the staining was 3+ but 2+ in one of them and 1+ in two of
them. Staining intensity was strong in 17 cases whereas it was weak in one case. When the focal staining was considered
positive, the specificity, sensitivity, positive predictive value
and negative predictive value were 1.000 (95% Confidence
Interval: 1.0000-1.0000), 1.000 (95% Confidence Interval:
1.0000-1.0000), 1.000 (95% Confidence Interval: 1.0000-
1.0000) and 1.000 (95% Confidence Interval: 1.0000-
In all of the 12 NSCLC-AC cases, the rate and intensity
of staining in the NSCLC areas were the same as those in
the surgical specimens. In all of the 17 NSCLC-SCC cases
(Figure 2D), staining was seen in neither NSCLC areas nor
surgical specimens. Since TTF-1 staining in NSCLC- SCC
and NSCLC-AC was the same, no statistical analyses were
In 29 NSCLC cases whose diagnosis was confirmed by the
surgical specimen, no difference was found in the rate and intensity of staining in both biopsy and surgical specimens
for all three immunohistochemical markers (p>0.050).
Statistically, no difference was found between the
sensitivity of immunohistochemical markers (p=0.506)
but the specificity of TTF-1 and CK5/6 was higher than
that of p63 (p=0.001). The specificity, sensitivity, positive
predictive value and negative predictive value for each
immunohistochemical marker are shown in Table III.
Click Here to Zoom
|Table III: The specificity, sensitivity, positive predictive value and negative predictive value for each immunohistochemical marker
With target-specific therapies, precise histologic
subtyping of lung cancer is needed more than ever before.
Morphological misdiagnosis is more common in biopsy
because of more limited tissue than in surgical specimens5
. For accurate diagnosis, immunohistochemistry is
usually employed in both surgical and biopsy10,12,13
The immunohistochemical use of CK5/6, TTF-1 and p63 together is recommended by the International Association
for the Study of Lung Cancer/American Thoracic Society/
European Respiratory Society14
. Thyroid transcription
factor 1 is an indication of the lung and thyroid origin. The
high sensitivity and specificity of TTF-1 for pulmonary
adenocarcinoma (AC) have been noted by earlier studies13-15
. Most studies dealt with the sensitivity and
specificity of TTF-1 for AC of the lung. In a study conducted
on a large series by Warth et al.2,
it was reported that
TTF-1 positivity was seen in 87.7% of AC cases. However,
in the studies conducted on a limited series, positivity went
down to 70%14
. TTF-1 positivity was seen in all of the
AC cases in our study while it was focal in three cases and
weak in one. A few studies suggest that some squamous
cell carcinomas (SCC) (3% to 21%) may be TTF-1 positive17,18,19
. As already known, TTF-1 positivity is not an
anticipated finding in SCC. If TTF-1 positivity exists in SCC, it could be entrapped lung epithelium or solid AC
may have been misinterpreted as SCC13,20
. There was
no staining in any SCC case in our study. Moreover, TTF-
1 positivity was observed in neither biopsy nor surgical
specimen of non-small-cell lung cancer (NSCLC-SCC)
cases. According to the results of our study, TTF-1 is the
most reliable marker to distinguish between AC and SCC.
Besides, TTF-1 positivity can be considered as a finding of
AC differentiation even if it is focal and weak in the cases
where histological differentiation is not clear.
The reported sensitivity of CK5/6 for pulmonary SCC
ranges from 73% to 100% and the sensitivity of CK5/6 is
reported to go down as there is a decrease in the tumor
differentiation7,13,21. CK5/6 positivity was observed in
43 of 48 SCC cases in the study by Kim et al.16, but in all
32 SCC cases in the study by Nicholson et al.22. One of
the reasons for the wide range of CK5/6 sensitivity can be
the fact that some studies regard focal or weak staining as
positive but others do not. In our study, no CK5/6 staining
was detected in only two cases while 75% of SCC cases were
diffuse and strong positive. All the NSCLC-SCC cases were
CK5/6 positive but six of them were focal. In addition, the
staining was diffuse in two surgical specimens but it was
focal in their former biopsies. According to our findings,
CK5/6 was a highly specific but not very sensitive marker
in distinguishing between SCC and AC when focal and
weak staining were considered to be positive. Additionally,
CK5/6 positivity can be considered as a finding supporting
SCC even if it is focal in the cases where histological
differentiation is not clear.
The utility of p63 for the identification of lung SCCs is
well-known, and its sensitivity ranges from 73% to 100%5,7,13. However, p63 specificity is not high because p63
can be stained focal and/or weak in some ACs12,13,16. In
our study, p63 positivity was observed in 4 AC cases which
had TTF-1(+)/ CK5/6 (-). Staining was focal in three cases
but diffuse in one. Nevertheless, intensity of staining was weak in the case that had diffuse staining like the other two
cases that had focal staining. This finding is in agreement
with the literature. All of the NSCLC-SCC in our study were
positive for p63 and their staining rate and intensity were
the same both in NSCLC areas and surgical specimens.
None of the NSCLC-AC cases had staining except one,
in which staining was not seen in the biopsy while focal
(1–33%) staining was seen in the surgical specimen. In a
study contucted by Warth et al.2, p63 positivity was
seen in 16.3% of the histomorphologically unclear ACs
and staining intensity was lower than in SCC. Our study
suggests that an AC diagnosis can be made despite p63
positivity in some cases, as in the literature. If the p63
positivity is focal and weak or moderate, TTF-1 positivity
and CK5/6 negativity are enough for AC diagnosis even if
AC is not clear histomorphologically. In our study, SCC had
CK5/6 negativity and AC had p63 positivity in a few cases.
However, TTF-1 was positive in all the ACs and negative
in all the SCCs. Although the specificity, sensitivity, PPV
and NPV of TTF-1 were 100%, a statistically significant
difference was found in only the specificity value due to the
limited number of cases. Our results need to be supported
by studies including large series.
Subtyping of NSCLCs is very important due to current
treatment modalities. Immunohistochemistry can be used
in subtyping of samples with limited tissue and cases in
which the morphology is unclear. TTF-1, CK 5/6 and p63
are the markers that can be used for this purpose. Different
staining patterns can be seen with CK5/6 and p63; however
they can be of help for the final diagnosis even in the cases
in which morphology is unclear, if they are used together
with TTF-1 and interpreted correctly.
We gratefully acknowledge Hatice Uluer in the Department
of Biostatistics and Medical Communication of Ege
University for performing the statistical analyses.
1) Jemal A, Siegel R, Xu J, Ward E. Cancer statistics, 2010. CA
Cancer J Clin. 2010;60:277-300.
2) Warth A, Muley T, Herpel E, Meister M, Herth FJ, Schirmacher P,
Weichert W, Hoffmann H, Schnabel PA. Large-scale comparative
analyses of immunomarkers for diagnostic subtyping of nonsmall-
cell lung cancer biopsies. Histopathology. 2012;6:1017-25.
3) Whithaus K, Fukuoka J, Prihoda TJ, Jagirdar J. Evaluation of
napsin A, cytokeratin 5/6, p63, and thyroid transcription factor 1
in adenocarcinoma versus squamous cell carcinoma of the lung.
Arch Pathol Lab Med. 2012;136:155-62.
4) Gressett SM, Shah SR. Intricacies of bevacizumab-induced
toxicities and their management. Ann Pharmacother.
5) Pelosi G, Rossi G, Bianchi F, Maisonneuve P, Galetta D, Sonzogni
A, Veronesi G, Spaggiari L, Papotti M, Barbareschi M, Graziano P,
Decensi A, Cavazza A, Viale G. Immunhistochemistry by means
of widely agreed-upon markers (cytokeratins 5/6 and 7, p63,
thyroid transcription factor-1, and vimentin) on small biopsies of
non-small cell lung cancer effectively parallels the corresponding
profiling and eventual diagnoses on surgical specimens. J Thorac
6) Rossi G, Pelosi G, Graziano P, Barbareschi M, Papotti M. A
reevaluation of the clinical significance of histological subtyping
of non--small-cell lung carcinoma: Diagnostic algorithms in the
era of personalized treatments. Int J Surg Pathol. 2009;17:206-18.
7) Kargi A, Gurel D, Tuna B. The diagnostic value of TTF-1, CK5/6,
and p63 immunostaining in classification of lung carcinomas.
Appl Immunohistochem Mol Morphol 2007;15:415–20.
8) Noh S, Shim H. Optimal combination of immunohistochemical
markers for subclassification of non-small cell lung carcinomas:
A tissue microarray study of poorly differentiated areas. Lung
Cancer 2012;76: 51–5.
9) World Health Organization classification of tumours. Pathology
and genetics of tumours of the lung, pleura, thymus and heart.
Travis WD, Brambilla E, Mqller-Hermelink HK, Harris CC,
editors. Lyon7: IARC Press; 2004.
10) Fatima N, Cohen C, Lawson D, Siddiqui MT. TTF1 and
napsin A double stain: A useful marker for diagnosing lung
adenocarcinoma on fine-needle aspiration cell blocks. Cancer
11) Rekhtman N, Ang DC, Sima CS, Travis WD, Moreira AL.
Immunohistochemical algorithm for differentiation of lung
adenocarcinoma and squamous cell carcinoma based on large
series of whole-tissue sections with validation in small specimens.
Mod Pathol. 2011;24:1348-59.
12) Conde E, Angulo B, Redondo P, Toldos O, García-García E,
Suárez-Gauthier A, Rubio-Viqueira B, Marrón C, García-
Luján R, Sánchez-Céspedes M, López-Encuentra A, Paz-Ares
L, López-Ríos F. The use of p63 immunohistochemistry for the
identification of squamous cell carcinoma of the lung. PLoS One
13) Mukhopadhyay S, Katzenstein AL. Subclassification of nonsmall
cell lung carcinomas lacking morphologic differentiation
on biopsy specimen: Utility of an immunohistochemical panel
containing TTF-1, napsin A, p63, and CK5/6. Am J Surg Pathol
14) Travis WD, Brambilla E, Noguchi M, Nicholson AG, Geisinger
KR, Yatabe Y, Beer DG, Powell CA, Riely GJ, Van Schil PE, Garg
K, Austin JH, Asamura H, Rusch VW, Hirsch FR, Scagliotti G,
Mitsudomi T, Huber RM, Ishikawa Y, Jett J, Sanchez-Cespedes
M, Sculier JP, Takahashi T, Tsuboi M, Vansteenkiste J, Wistuba I,
Yang PC, Aberle D, Brambilla C, Flieder D, Franklin W, Gazdar
A, Gould M, Hasleton P, Henderson D, Johnson B, Johnson
D, Kerr K, Kuriyama K, Lee JS, Miller VA, Petersen I, Roggli
V, Rosell R, Saijo N, Thunnissen E, Tsao M, Yankelewitz D.
International association for the study of lung cancer/american
thoracic society/european respiratory society international
multidisciplinary classification of lung adenocarcinoma. J Thorac
15) Loo PS, Thomas SC, Nicolson MC, Fyfe MN, Kerr KM. Subtyping
of undifferentiated non-small cell carcinomas in bronchial biopsy
specimens. J Thorac Oncol 2010;5:442-7.
16) Kim MJ, Shin HC, Shin KC, Ro JY. Best immunohistochemical
panel in distinguishing adenocarcinoma from squamous cell
carcinoma of lung: Tissue microarray assay in resected lung
cancer specimens. Ann Diagn Pathol. 2013;17:85-90.
17) Chang YL, Lee YC, Liao WY, Wu CT. The utility and limitation of
thyroid transcription factor-1 protein in primary and metastatic
pulmonary neoplasms. Lung Cancer. 2004;44:149–57.
18) Jerome Marson V, Mazieres J, Groussard O, Garcia O, Berjaud
J, Dahan M, Carles P, Daste G. Expression of TTF-1 and
cytokeratins in primary and secondary epithelial lung tumours:
Correlation with histological type and grade. Histopathology.
19) Tan D, Li Q, Deeb G, Ramnath N, Slocum HK, Brooks J, Cheney
R, Wiseman S, Anderson T, Loewen G. Thyroid transcription
factor-1 expression prevalence and its clinical implications in
nonsmall cell lung cancer: A high-throughout tissue microarray
and immunohistochemistry study. Hum Pathol. 2003;34:597–
20) Monica V, Ceppi P, Righi L, Tavaglione V, Volante M, Pelosi
G, Scagliotti GV, Papotti M. Desmocollin-3: A new marker of
squamous differentiation in indifferentiated large-cell carcinoma
of the lung. Mod Pathol. 2009;22:709–17.
21) Downey P, Cummins R, Moran M, Gulmann C. If it’s not CK5/6
positive, TTF-1 negative it’s not a squamous cell carcinoma of
lung. APMIS. 2008;116:526–9.
22) Nicholson AG, Gonzalez D, Shah P, Pynegar MJ, Deshmukh M,
Rice A, Popat S. Refining the diagnosis and EGFR status of nonsmall
cell lung carcinoma in biopsy and cytologic material, using
a panel of mucin staining, TTF-1, cytokeratin 5/6, and P63, and
EGFR mutation analysis. J Thorac Oncol. 2010;5:436-41.