Characterization of Breast Cancer Subtypes by Immunohistochemistry in a Large Retrospective Study

businessunknownInternet και Εφαρμογές Web

12 Νοε 2013 (πριν από 7 χρόνια και 8 μήνες)

406 εμφανίσεις

In:
Tumor Markers Research Perspectives


ISBN:
978
-
1
-
60021
-
577
-
3

Editor:
G. A. Sinise
, pp.
-



© 2007 Nova Science Publishers, Inc.








Chapter
X
I
V



Characterization of Breast Cancer
Subtypes

by Immunohistochemistry

in
a Large Retrospective Study



J.

Decock
*
1
, W.

Hendrickx

1
,2
,
C. Stefan
1
, P. Neven
2
, H. Wildiers
1,2
,
MR. Christiaens
2
, A. Smeets
1,
2


and
R.

Paridaens
1,2

1

Labo
ratory for Experimental Oncology (LEO), Department of
General Medical

Oncology, University Hospital Gasthuisberg, Herestraat 49, 3000 Leuven, Belgium

2

Multidisciplinary Breast Center (MBC), University Hospital Gasthuisberg,

Herestraat 49, 3000 Leuven, Be
lgium



Abstract


Background:

DNA microarray studies identified
distinct molecular subtypes
that

are
associate
d

with different clinical outcome,
defined as
luminal

A
, luminal B, basal
-
like
and Her2
+
ER
-
. This study aimed
to evaluate the immunohistochemistr
y markers ER, PR
and Her2 as surrogate markers for the previously identified molecular subtypes
in a large
retrospective cohort (n=1678) with invasive breast carcinomas

with

regard to

various
demographic,
clinical and pathological features.

Methods:
All p
atients were diagnosed with primary breast cancer between 2000 and
200
5

at the University Hospital Leuven.
None of them received neo
-
adjuvant therapy, had
bilateral cancer
,

tumor
s with direct extension to chest wall or skin
, or nipple Paget’s
disease
.
ER,
PR and Her2 expression was determined by immunohistochemistry and cases
with intermediary staining for Her2 were further subjected to two
-
color fluorescence in
situ hybridization analysis. Clinical and pathological features included age at diagnosis,
menop
ausal status,
maximum
tumor size, tumor grade, lymph node status and Nottingham
Prognostic Index (NPI).





*

Corresponding author: Decock J., Laboratory for Experimental Oncology, University Hospital Gasthuisberg

-

KULeuven
, O
&
N
1
, Room

815, Herestraat 49, 3000 Leuven, Belgium
. Phone: +
32 (0)16.34.62.93; Fax:
+
32
(0)16.34.69.01;
E
-
mail:
julie.decock@med.kuleuven.be

J.

Decock, W.

Hendrickx
,
C. Stefan

et al.

2

Results:

Of the 1678 patients with invasive breast carcinoma, 80% (n=1342) had
luminal A subtype, 7% (n=119) had luminal B subtype, 9% (n=144) had bas
al
-
like type
and 4% (n=73) had Her2
+
/ER
-

carcinomas. Within the luminal type, the majority of cases
were PR
+
, with 87% and 83% for the A and B subtypes, respectively. Luminal A
carcinomas were more often small in size (p=0.037), well/moderately differentia
ted
(p<0.00001) and associated with a lower NPI value (p<0.00001). Patients with a luminal
B
-
type tumor presented more frequently with
lymph node involvement (p=0.0
4
) and were
significantly younger at diagnosis (p
=
0.0001
2
) compared to the other patients.
M
ost
basal
-
like and Her2
+
/ER
-

carcinomas were poorly differentiated (92% and 85%
respectively).
Furthermore, patients with luminal A or Her2+/ER
-

carcinomas were more
often

postmenopausal than patients with a luminal B or basal
-
like
tumor

(p=0.0
16
).


Conclu
sions:
In our large retrospective cohort of 1678 invasive breast cancers, the
luminal A carcinomas were by far best represented in comparison with the other

immunohistochemistry
-
defined
(sub)types, and significantly associated with good
prognostic clinica
l and pathological parameters. By contrast, although less represented,
the luminal B, the basal
-
like and the Her2
+
ER
-

tumor (sub)types were more freque
ntly
associated with unfavourable prognostic factors

such as younger age at diagnosis, lymph
node involve
ment or poor histological grade.


Keywords:

breast cancer, subtypes, clinicopathological parameters
, immunohistochemistry
.



Introduction


Breast cancer is a heterogeneous disease consisting of several histological subtypes with
different clinical outcome
and with patients showing a diverse range of responses to a given
treatment. Although lymph node involvement, histologic grade, tumour size, steroid hormone
receptor expression and Her2 status all have been strongly correlated to prognosis, their
assessmen
t does not alter our inability to accurately predict relapse or response to therapy. A
broad variety of genes and proteins has been analyzed in single
-
marker prognostic and
predictive studies but most of them were hampered by small size, patient selection
criteria,
tumor heterogeneity or treatment diversity and consequentially lacked power to identify all
patients at risk for recurrence and those who would benefit from a new therapy
[1; 2]
. A large
number of genes are involved in cell growt
h, death and differentiation and account as such for
much of the phenotypic diversity
between

tumors
. The growing knowledge of these complex
gene signal transduction pathways emphasizes the importance of studying multiple genetic
alterations simultaneously

and promotes the emergence of oncogenomics. A major goal in the
field of oncogenomics is to try to answer the clinically important questions about which
tumours will behave aggressively, which tumours will remain dormant, which patients do and
do not requ
ire systemic therapy and what type of drugs should be used
[3]
.

Analyses of thousands of genes using microarrays have
classified breast carcinomas into
distinct subtypes
, associated with different clinical outcomes. Perou and colleag
ues initially
classified 38 invasive breast carcinomas based on their distinct gene expression profile into
four subtypes: luminal
-
like, basal
-
like, Her2
+
/ER
-

and normal breast
-
like
[4]
. Later the same
research gr
oup analyzed the expression profiles of 78 breas
t carcinomas and categorized them
into five subgroups
based on variations in gene expression
: luminal A, luminal B, Her2
+
/ER
-
,
Breast Cancer Subtypes

3

basal
-
like and normal breast
-
like,
[5]
. Since then, various microarra
y studies confirmed the
segregation of breast tumors into distinct subtypes with different clinical outcomes in various
ethnic populations
[6
-
11]
. Patients with luminal A type tumors are facing a relatively good
prognosis, whereas patients with basal
-
like tumors experience a much shorter overall and
disease free survival period. Some studies also reported on the clinical and pathological
ch
aracteristics of the different subtypes
[6; 7; 12
-
16]
. Furthermore, it has been shown that the
various subtypes remarkably differ in their response to treatment, as reviewed by Brenton et
al.
[17]
.

Not surpris
ingly, hope has risen that this new classification system might improve the
accuracy of prognostic stratification and lead to the development of new tumor
-
tailored
therapeutic strategies. Although microarrays have the potential to be used as diagnostic and

prognostic tools, they have not yet come into routine clinical practice due to their time
-
consuming character and high cost.

In this context,

t
he use
fullness

of immunohistochemical
surrogates for
separation of the distinct s
ubtype
s
should

be evaluated, as

indicated by the
group of Carey LA

[
7
]
.

They evaluated the use of immunohistochemistry (IHC) markers that
were previously verified against gene expression profiles to estimate the pre
valence of the
intrinsic subtypes in a large population
-
based epidemiological study of African American and
white women. They identified IHC profiles that best matched the gene expression profiles by
performing both microarray analysis and IHC and validate
d these IHC surrogate
s

using a
930
-
case tissue microarray. In this way, IHC
-
based definitions for the various molecular
subtypes were ER
+
PR
+/
-
Her2
-

for luminal

A
,
ER
+
PR
+/
-
Her2
+

for luminal B,
ER
-
PR
-
Her2+ for
Her2
+


subtype and ER
-
PR
-
Her2
-

for basal
-
like

[
5
;
8
; 16]
.

We aimed to
cha
racterize the
se

IHC
-
defined subgroups

on a large retrospective cohort of
1678 invasive breast cancer patients
with regard to
classic clinical and pathological features.



Materials
a
nd Methods


Patient
s


A total of
1678

patients with primary operable breas
t cancer

were included in this study.
Patient characteristics were extracted from clinical files
,
tumour characteristics
and
lymph
node status were
retriev
ed from pathology reports
,
and

all data were
collected in
our

central
breast cancer
database
. All pat
ients were newly diagnosed at the
Multidisciplinary Breast
Center of the University Hospital of Leuven between
2000

and 200
5
, underwent mastectomy
or local wide excision of their primary breast tumour and axillary lymph node dissection for
staging and trea
tment.
None of them received neo
-
adjuvant therapy, had bilateral cancer
,

tumor
s with direct extension to chest wall or skin
, or nipple Paget’s disease
.
For all patients
included, data are available on age at diagnosis, menopausal status, tumour size and gr
ade,
lymph node involvement, steroid hormone receptor expression and
Her2

status.




J.

Decock, W.

Hendrickx
,
C. Stefan

et al.

4

Pathological
Assessment
o
f Tumour Tissue


Tumour t
yping
and grading were

performed
on
paraffin embedded
haematoxylin
-
eosin

(H
&
E) slides according to the WHO
-
classificatio
n and the Ellis and Elston grading system
respectively.

Lymph nodes were examined with H
&
E using 3 sections per node
.

The
Nottingham Prognostic Index was calculated
using the equation NPI = 0.2 x tumor size (cm)
+ grade (I

III) + lymph node score.


Immunoh
istochemical

staining for ER, PR and Her2 was performed on
4µm
thick serial
paraffin sections. Heat
-
induced epitope retrieval was carried out in a calibrated water bath
(95
-
99°C) and antibody complexes were visualised by Envision (Dako, Glostrup, Denmark)
and DAB. A broad variety of antibodies for ER and PR has been used in the period of 2000
to 2005 and semiquantitative evaluation has been performed using the H
-
score for tumors
resected between January 2000 and august 2003 or the Allred score system for tu
mors
resected after august 2003. The primary mouse monoclonal antibody CB11 (Novocastra
Laboratories, Newcastle
-
upon
-
Tyne, UK) directed against
Her2

was applied in a dilution of
1/40 and expression was evaluated with the DAKO scoring system.



Classificati
on
o
f Breast Cancer Subtypes


Based upon ER, PR and
Her2

immunohistochemical stainings, breast tumors were
classified into
4 subgroups ER
+
PR
+/
-
Her2
-
, ER
+
PR
+/
-
Her2
+
, ER
-
PR
-
Her2
-

and ER
-
PR
-
Her2
+
,
which resemble the previously identified l
uminal A, luminal B,

basal
-
like and Her2+/ER
-

subgroups
.
Both the H
-
scoring and Allred scoring systems take in account the staining
intensity and percentage of cells positively stained.
A complete H
-
score was calculated by
summing the products of the percentage
positively sta
ined
cells

at a given staining intensity

(0

100) and the
different
staining intensit
ies

(0

3).

The Allred score was determined by
summing the score representing the
proportion of positive tumour cell
s

(0
-
5) with the staining
intensity (0
-
3). For both scori
ng systems, ER and PR expression was considered negative
when staining was completely absent.
Using the H score, expression was defined positive
when the final score was 1
-
300, while for the Allred score expression

was considered positive
when the final sc
ore was 2
-
8.

The DAKO
scoring system

for
Her2

immunostaining was
applied, taking both the proportion of tumour cells with positive membrane staining and the
staining intensity into account. A score of 0 or 1
was
defined negative, while a score of 2 or 3
wa
s considered positive for
Her2

overexpression
.

Cases with a DAKO score of 2 were further
analyzed by 2
-
color Fluorescence In situ Hybridisation
(PathVysion, Vysis, Downers Grove,
IL, USA)

in order to distinguish true amplification from polysomie.
A mean
H
e
r
2
/
chromosome 17 ratio > 2 was considered amplified for H
er
2.


Statistical
Analys
e
s


Differences between breast cancer subtypes with regard to clinicopathological
characteristics were examined using 1
-
way ANOVA for age at diagnosis and NPI value and
Chi S
quare tests for the remaining variables. Statistical analyses were performed

using

the
software package
SPSS

version
13
, the level of significance being set at p ≤ 0,05.

Breast Cancer Subtypes

5



Results


Study
Population


The median age at diagnosis of all patients was 58 (range

26
-
95
). The majority (1177) of
the patients were postmenopausal, while 498 patients were premenopausal.

Small and large
tumours were equally distributed, 857 (51%) and 821 (49%) tumours respectively. A total of
231 tumours (14%) were well differentiated,
797 (47%) moderately and 650 (39%) poorly.
No nodal involvement was found in 1069 (64%) patients while 609 (36%) were lymph node
positive. Stratification of tumours into
molecular subtype
-
l
ike groups;

based on
ER, PR and
Her2 expression
;

resulted in 1342 (
80%) luminal A, 119 (7%) luminal B, 144 (9
%) basal
-
like
and 73 (4%) Her2
+
/ER
-

carcinomas. Within the luminal tumor type, the majority of cases
were PR
+
,
with 87% and 83% for the A and B subtypes, respectively (table 1). According to
the American Joint Comm
ittee on Cancer staging system, 664 patients had stage I disease,
755 had stage II disease and 259 had stage III disease.


Table 1. Distribution of
immunohistochemistry
-

defined
breast cancer subtypes
according to clinical and pathological features (n=1678
).


variable

A
ll

n

luminal A

%(n)

luminal B

%(n)

basal
-
like

%(n)

Her2+/ER
-

%(n)

p value*

all

1678

80(1342)

7(119)

9(144)

4(73)


M
enopausal status

premenopausal

498

28

39

37

26

0,015

postmenopausal

1177

72

61

63

74


Histologic
grade







G1+2

1028

72

31

8

15

<0,00001

G3

650

28

69

92

85


Tumour size







pT1

857

53

45

44

42

0,037

pT2+3

821

47

55

56

58


A
xillary lymph node status

negative

1069

65

52

65

60

0,044

positive

609

35

48

35

40


ERPR







ER+PR+

1272

87

83

0

0

<0,00001

ER+PR
-

189

13

17

0

0


ER
-
PR
-

217

0

0

100

100


A
ge, mean

58

59

54

56

57

0,00012

NPI, mean

4,27

4,1

4,85

4,97

5,02

<0,00001

*

Associations between
immunohistochemistry
-
defined subtypes

and the continuous variables age
and NPI were analyzed using one
-
way ANOVA,

whil
e associations with the remaining
characteristics were analyzed using the Pearson Chi Square test
.



J.

Decock, W.

Hendrickx
,
C. Stefan

et al.

6



Steroid
Hormone Receptor Status of Tumor Subtypes


Based on
ER, PR and Her2 expression of the tumor
, breast tumors can be classified into
immunohistoche
mistry
-
defined subtypes.

Estrogen receptor negative tumors
can be

subdivided into basal
-
like (ER
-
PR
-
Her2
-
) and Her2
+
/ER
-

(ER
-
PR
-
Her2
+
) tumors, while ER
+

tumors
can

further
be
classified as luminal A (ER
+
PR
+
/
-
Her2
-
) or luminal B (ER
+
PR
+
/
-
Her2
+
).
Within the
luminal type, we determined the proportion of PR
+

cases and found 87% and 83%
of luminal A and luminal B tumors to be PR
+

(table 1).



Characterization
of Tumor
Subtypes


Characteristics of the IHC
-
defined breast cancer subtypes are presented in table 1.
Breast
cancer subtypes differed significantly by age (p=0.0001), menopausal status (p=0.015),
histologic grade (p< 0.00001), tumor size (p=0.037), axillary lymph node status (p= 0.044)
and NPI (p< 0.00001). As the NPI value is based on tumor size, histolog
ic grade and lymph
node involvement, this result reflects the significant difference found among the various
subtypes for each of these parameters.

The proportion of smaller tumors
(< 21 mm)
was not
dramatically

different among the
various groups, with th
e exception
again
of the luminal A group in which the tumors tended
to be smaller.

Luminal carc
inomas
and in particular luminal A
-
type tumors
were more likely
well
/moderately differentiated, whereas the majority of basal
-
like (92%) and Her2
+
/ER
-

(85%) carc
inomas were poorly differentiated. Patients with a luminal B type tumor
had the
highest prevalence of lymph node involvement and were significantly younger than the other
patients.
All 4 breast tumor
subtypes were associated with a NPI value

indicative for

a
moderately good prognosis

(
3.4
<NPI<
5.4
)
with the lowest value observed in the luminal A
-
type

tumors.
This result for luminal A tumors is not surprisingly since the tumor size, grade
and lymph node involvement in this group of tumors all indicated a good

prognosis.

T
he
number of premenopausal patients was higher in the groups with luminal B and basal
-
like
tumors compared to those with luminal A or Her2
+
/ER
-

tumors.

Moreover, each of these findings was also observed in the subgroup of patients with early
s
tage disease
(stage I and II)
at an even higher significance
, with exception of the lymph node
status (table 2)
.










Breast Cancer Subtypes

7



Table 2. Characteristics of
immunohistochemistry
-
defined

subtypes in breast cancer

patients with early stage disease (n=1419)
.


va
riable

all

n

luminal A

%(n)

luminal B

%(n)

basal
-
like

%(n)

Her2+/ER
-

%(n)

p value*

all

1419

81(1145)

7(94)

7(123)

4(57)


M
enopausal status

premenopausal

399

26

40

37

25

0,002

postmenopausal

1020

74

60

63

75


Histologic grade







G1+2

891

74

35

6

16

<0,00001

G3

528

26

65

94

84


A
xillary lymph node status

negative

1069

76

66

76

77

0,19

positive

350

24

34

24

23


ERPR







ER+PR+

1078

87

82

0

0

<0,00001

ER+PR
-

161

13

18

0

0


ER
-
PR
-

180

0

0

100

100


A
ge, mean

59

59

54

56

58

<0,00001

NPI, mean

3,91

3,75

4,4

4,69

4,57

<0,00001

*

Associations between
immunohistochemistry
-
defined
subtypes and the continuous variables age
and NPI were analyzed using one
-
way ANOVA,

while associations with the remaining
characteristics were analyzed using the Pear
son Chi Square test
.



Conclusion


We examined the presence and characteristics
of various
immunohistochemistry
-
defined

subtypes

in a large retrospective cohort of invasive breast cancer patients by means of routine
immunohistochemical staining for ER, PR
and Her2.

We were able to confirm the existence
of four tumor
subgroups, resembling the microarray identified subtypes
luminal A, luminal B,
basal
-
like and Her2
+
/ER
-
. On basis of their clinical and pathological profile, patients with
luminal A tumors seem
to have a good prognosis which was reflected in menopausal status,
small tumor size, high proportion of well/moderately differentiated tumors and consequently a

lower NPI value. In contrast, patients with luminal B tumors were more likely associated with
p
oor prognosis as they were significantly younger than the others and were often associated
with lymph node involvement.
However, any conclusion regarding menopausal status rather
than age in association with breast cancer phenotypes may have been biased by

use of
contraceptives or hormone replacement therapy.
The differences in clinical prognostic
features for both luminal subtypes can also in part be explained by microarray analyses which
showed that luminal A type tumors have, in general, higher expressio
n of estrogen responsive
genes and lower expression of proliferative genes than luminal B
[5; 8]
.

Just as others, we were not able to clearly characterize the ER
-

tumor subtypes, basal
-
like
and Her2
+
/ER
-
, except for the observation that they were more l
ikely poorly differentiated
[5;
7; 17]
. The poor prognosis of the Her2
+
/ER
-

and basal
-
like tumor subtypes is most probably
J.

Decock, W.

Hendrickx
,
C. Stefan

et al.

8

due to the fewer treatment options available for ER
-

and PR
-

tumors. Moreover, although
treatment with the anti
-
Her2 monoclonal antibody trastuzumab in
combination with
chemotherapy significantly improves disease
-
free
-
survival among women with advanced
breast cancer and remarkable reduces the number of patients with relapse, not all Her2
+

tumors respond to trastuzumab
[18]
.

To our knowledge, this is by far the largest retrospective cohort of

invasive breast cancer
patients with clinicopathological characterization of the various tumor subtypes. Although the
molecu
lar breast tumor subtypes were initially identified by microarray gene expression
studies, we
support Carey LA and colleagues in th
eir hypothesis
that IHC assessment of ER,
PR and Her2 is a good alternative approach for
analysis

of the subtypes in a large
retrospective cohort. Moreover, ou
r findings are in line with the results from Calza et al. who
recently reported the largest micro
array study on a population
-
based cohort of 412 breast
cancer patients
[6]
. They observed that luminal A carcinomas were mostly found in
postmenopausal women, tended to be small and were more likely good/moderately
differentiated. Furthermore, they found that in comparison with lumi
nal A tumors, the luminal
B group had a higher proportion of poorly differentiated and lymph node positive tumors. In
their cohort, basal
-
like tumors were most often found in younger, premenopausal patients
with poorly differentiated tumors, while Her2
+
/ER
-

tumors were more frequently observed in
elderly women with large and poorly differentiated tumors.

The strength of our study is that
the cohort consisted of as well 1419 (85 %) patients with early stage (stage I, II) disease as
259 (15%) patients with ad
vanced disease (stage III). Interestingly, analysis of early stage
disease revealed the presence of the same four distinct tumor subtypes which were similarly
characterized as those in the overall study population. Other groups analyzed the gene
expression

in DCIS for comparison with normal and invasive carcinomas in various ethnic
populations and found that DCIS tumors could be similarly divided into distinct subtypes with

different clinical outcomes
[19
-
21]
. All these studies imply that the molecular signatures
defining the subtype of a tumor and its clinical characteristics may already be set in early
stage disease and even at the pre
-
invasive

stage of carcinogenesis.

We would like to emphasize the importance of the methodology used for steroid hormone
and Her2 assessment
; immunohistochemical analyses have their limitations
.
Also, the current
reproducibility and clinical relevance of m
icroarray

generated data have recently been
criticized by Dupuy and Simon
[2
2
]
.
Both
methodologies

are used to measure the gene or
protein expression o
f ER, PR and Her2 and not necessarily reflect the activity of the receptor
and its downstream pathways. Moreover, immunohistochemical staining of Her2 does not
allow us to distinguish between true gene amplification and polysomy and hence FISH
analysis is
required for cases with intermediary Her2 staining. In our institute, dual
-
color
FISH is therefore performed as routine clinical practice for Her2 overexpressing tumors with
an IHC score of 2. Expression analysis of downstream signalling pathways might enl
ighten
the knowledge on breast cancer behaviour and improve prognostic accuracy.

In conclusion, using routine immunohistochemistry we were able to confirm the presence
of four
subgroups, resembling the microarray identified molecular subtypes,

in a large
p
opulation
-
based cohort of breast cancer patients with early or late stage disease.
Furthermore, our study supports the hypothesis that these
immunohistochemistry
-
defined
tumor subtypes are distinct biological entities with distinct clinical characteristics
.

Breast Cancer Subtypes

9





Acknowledgements


We gratefully thank Prof. M. Drijkoningen for

the

pathological assessment of all tumor
specimens.
Further, w
e acknowledge all collaborators of the Multidisciplinary Breast Cente
r
,
General Medical Oncology and Gynecology for their he
lp with data input.

Grant support: This work is funded by the “Vlaamse Liga tegen Kanker” and the EU
Framework Programme 6 (LSHC
-
CT
-
2003
-
503297, Cancerdegradome).



References


[1]

Chatterjee, S.K. and Zetter, B.R. (2005). Cancer bioma
rkers: knowing the present and
predicting the future.
Future. Oncol., 1,

37
-
50.

[2]

Coradini, D. and Daidone, M.G. (2004). Biomolecular prognostic factors in breast
cancer.
Curr. Opin. Obstet. Gynecol., 16,

49
-
55.

[3]

Strausberg, R.L., Simpson, A.J., Old, L.J. and

Riggins, G.J. (2004). Oncogenomics and
the development of new cancer therapies.
Nature, 429,

469
-
474.

[4]

Perou, C.M., Sorlie, T., Eisen, M.B., van de Rijn, M., Jeffrey, S.S., Rees, C.A., Pollack,
J.R., Ross, D.T., Johnsen, H., Akslen, L.A., Fluge, O., Pergam
enschikov, A., Williams,
C., Zhu, S.X., Lonning, P.E., Borresen
-
Dale, A.L., Brown, P.O. and Botstein, D.
(2000). Molecular portraits of human breast tumours.
Nature, 406,

747
-
752.

[5]

Sorlie, T., Perou, C.M., Tibshirani, R., Aas, T., Geisler, S., Johnsen, H.,
Hastie, T.,
Eisen, M.B., van de Rijn, M., Jeffrey, S.S., Thorsen, T., Quist, H., Matese, J.C., Brown,
P.O., Botstein, D., Eystein, L.P. and Borresen
-
Dale, A.L. (2001). Gene expression
patterns of breast carcinomas distinguish tumor subclasses with clinical

implications.
Proc. Natl. Acad. Sci. U S A, 98,

10869
-
10874.

[6]

Calza, S., Hall, P., Auer, G., Bjohle, J., Klaar, S., Kronenwett, U., Liu, E.T., Miller, L.,
Ploner, A., Smeds, J., Bergh, J. and Pawitan, Y. (2006).
Intrinsic molecular signature of
breast canc
er in a population
-
based cohort of 412 patients.
Breast Cancer Res., 8,

R34
-

[7]

Carey, L.A., Perou, C.M., Livasy, C.A., Dressler, L.G., Cowan, D., Conway, K.,
Karaca, G., Troester, M.A., Tse, C.K., Edmiston, S., Deming, S.L., Geradts, J., Cheang,
M.C., Nielse
n, T.O., Moorman, P.G., Earp, H.S. and Millikan, R.C. (2006). Race,
breast cancer subtypes, and survival in the Carolina Breast Cancer Study.
JAMA, 295,

2492
-
2502.

[8]

Sorlie, T., Tibshirani, R., Parker, J., Hastie, T., Marron, J.S., Nobel, A., Deng, S.,
Johns
en, H., Pesich, R., Geisler, S., Demeter, J., Perou, C.M., Lonning, P.E., Brown,
P.O., Borresen
-
Dale, A.L. and Botstein, D. (2003). Repeated observation of breast
tumor subtypes in independent gene expression data sets.
Proc. Natl. Acad. Sci. U S A,
100,

8
418
-
8423.

J.

Decock, W.

Hendrickx
,
C. Stefan

et al.

10

[9]

Sorlie, T., Wang, Y., Xiao, C., Johnsen, H., Naume, B., Samaha, R.R. and Borresen
-
Dale, A.L. (2006). Distinct molecular mechanisms underlying clinically relevant
subtypes of breast cancer: gene expression analyses across three different platforms
.
BMC. Genomics, 7,

127

[10]

Sotiriou, C., Neo, S.Y., McShane, L.M., Korn, E.L., Long, P.M., Jazaeri, A., Martiat,
P., Fox, S.B., Harris, A.L. and Liu, E.T. (2003). Breast cancer classification and
prognosis based on gene expression profiles from a population
-
b
ased study.
Proc. Natl.
Acad. Sci. U S A, 100,

10393
-
10398.

[11]

West, M., Blanchette, C., Dressman, H., Huang, E., Ishida, S., Spang, R., Zuzan, H.,
Olson, J.A., Jr., Marks, J.R. and Nevins, J.R. (2001). Predicting the clinical status of
human breast cancer by

using gene expression profiles.
Proc. Natl. Acad. Sci. U S A,
98,

11462
-
11467.

[12]

Abd El
-
Rehim, D.M., Pinder, S.E., Paish, C.E., Bell, J., Blamey, R.W., Robertson, J.F.,
Nicholson, R.I. and Ellis, I.O. (2004). Expression of luminal and basal cytokeratins in
human breast carcinoma.
J. Pathol., 203,

661
-
671.

[13]

Jacquemier, J., Ginestier, C., Rougemont, J., Bardou, V.J., Charafe
-
Jauffret, E.,
Geneix, J., Adelaide, J., Koki, A., Houvenaeghel, G., Hassoun, J., Maraninchi, D.,
Viens, P., Birnbaum, D. and Bertucci, F.
(2005). Protein expression profiling identifies
subclasses of breast cancer and predicts prognosis.
Cancer Res., 65,

767
-
779.

[14]

Kusinska, R., Potemski, P., Jesionek
-
Kupnicka, D.
and

Kordek, R. (2005).
Immunohistochemical identification of basal
-
type cytokera
tins in invasive ductal breast
carcinoma
--
relation with grade, stage, estrogen receptor and HER2.
Pol. J. Pathol., 56,

107
-
110.

[15]

Livasy, C.A., Karaca, G., Nanda, R., Tretiakova, M.S., Olopade, O.I., Moore, D.T. and
Perou, C.M. (2006). Phenotypic evaluation
of the basal
-
like subtype of invasive breast
carcinoma.
Mod. Pathol., 19,

264
-
271.

[16]

Nielsen, T.O., Hsu, F.D., Jensen, K., Cheang, M., Karaca, G., Hu, Z., Hernandez
-
Boussard, T., Livasy, C., Cowan, D., Dressler, L., Akslen, L.A., Ragaz, J., Gown,
A.M., Gilks
, C.B., van de, R.M. and Perou, C.M. (2004). Immunohistochemical and
clinical characterization of the basal
-
like subtype of invasive breast carcinoma.
Clin.
Cancer Res., 10,

5367
-
5374.

[17]

Brenton, J.D., Carey, L.A., Ahmed, A.A. and Caldas, C. (2005). Molecula
r
classification and molecular forecasting of breast cancer: ready for clinical application?
J. Clin. Oncol., 23,

7350
-
7360.

[18]

Nahta, R., Yu, D., Hung, M.C., Hortobagyi, G.N. and Esteva, F.J. (2006). Mechanisms
of Disease: understanding resistance to HER2
-
ta
rgeted therapy in human breast cancer.
Nat. Clin. Pract. Oncol., 3,

269
-
280.

[19]

Bryan, B.B., Schnitt, S.J. and Collins, L.C. (2006). Ductal carcinoma in situ with basal
-
like phenotype: a possible precursor to invasive basal
-
like breast cancer.
Mod. Pathol.,
1
9,

617
-
621.

[20]

Kepple, J., Henry
-
Tillman, R.S., Klimberg, V.S., Layeeque, R., Siegel, E., Westbrook,
K. and Korourian, S. (2006). The receptor expression pattern in ductal carcinoma in situ
predicts recurrence.
Am. J. Surg., 192,

68
-
71.

Breast Cancer Subtypes

11

[21]

Yu, K., Lee, C.H., Tan
, P.H. and Tan, P. (2004). Conservation of breast cancer
molecular subtypes and transcriptional patterns of tumor progression across distinct
ethnic populations.
Clin. Cancer Res., 10,

5508
-
5517.

[22]

Dupuy A. and Simon RM. (2007). Critical review of published

microarray studies for
cancer outcome and guidelines on statistical analysis and reporting.
J. Natl. Cancer
Inst., 99,
147
-
57
.