ASCO/CAP Guideline Recommendations for IHC

Testing of ER and PgR in Breast Cancer

Arch Pathol Lab Med, 134:907-22, 2010

Journal of Clinical Oncology, 28:2784-2795, 2010

D. Craig Allred, M.D.

Department of Pathology and Immunology

Journal of Clinical Oncology, 28:2784-2795, 2010

ASCO/CAP Guidelines for IHC Testing of

ER and PgR in Breast Cancer

Invasive Breast Cancers (Mandatory)

Ductal Carcinoma In Situ (Optional)

IHC Assays for ER and PgRMust be Comprehensively Validated

Quantitative Scoring of ResultsPercent or Proportion of Positive Cells

Intensity of Positive Cells

≥ 1% Expressing Cells

"Positive"Expect 70-80%

Responsive

to Endocrine Therapy

Interpretation of ResultsCalibrated to Response to Endocrine Therapy

Report ResultsScores

Interpretation

<1% Expressing Cells

"Negative"Expect 20-30%Not Responsive

to Endocrine Therapy

Retest and Confirm If*:

External Control Negative

Internal Control Negative

Low Histological Grade**

Lobular Subtype**

Tubular Subtype**

Mucinous Subtype**

Other…

Comprehensive Ongoing:

Quality Assurance

Proficiency Testing

*Probably single most helpful reccommedation for improving accuracy**Not required if internal control positive

What is Comprehensive Validation?

Technical: The assay should be specific, sensitive, reproducible,

calibrated to clinical outcome, interpreted, and reported in a relativelyuniform manner. There should be comprehensive ongoing qualityassurance.

Clinical: The factor should identify groups of patients with significantly

J Natl Cancer Inst. 1991;83:154.

Cancer. 1993;72:3131.

Arch Pathol Lab Med. 1995;119:1109.

J Clin Oncol. 1996;14:2843.

J Natl Cancer Inst. 1996;88:1456.

different risks of relapse, survival, or treatment response – demonstratedin multiple large studies (ideally randomized clinical trials).

Useful: Actually used by physicians to make important treatment

decisions.

Estrogen ReceptorEstrogen ReceptorEstrogen ReceptorEstrogen Receptor

ReferenceReferenceReferenceReference AntibodyAntibodyAntibodyAntibody CutpointCutpointCutpointCutpoint forforforfor "Positive""Positive""Positive""Positive"

Harvey. J Clin Oncol 17:1474, 1999 6F11 Allred Score ≥3 (1-10% weakly positive cells)

Regan, JNCI 98:1571-81, 2006

Viale. J Clin Oncol 25:3846, 2007

Viale. J Clin Oncol 26(9):1404, 2008

1D5 1-9% (low) and ≥10% (high)

Cheang. J Clin Oncol 24:5637, 2006 SP1 ≥1%

Comprehensively Validated IHC Assaysfor Measuring ERαααα and PgR in Breast Cancer

(identified in ASCO/CAP Guidelines)

Cheang. J Clin Oncol 24:5637, 2006 SP1 ≥1%

Phillips. Appl IHC Molec Morphol15:325, 2007 ER.2.123 +

1D5

Allred Score ≥3 (1-10% weakly positive cells)

Dowsett. J Clin Oncol 26:1059, 2008 6F11 H-score >1 (≥1%)

Progesterone ReceptorProgesterone ReceptorProgesterone ReceptorProgesterone Receptor

Mohsin. Modern Pathol 17:1545, 2004 1294 Allred Score ≥3 (1-10% weakly positive cells)

Regan, JNCI 98:1571-81, 2006

Viale. J Clin Oncol 25:3846, 2007

Viale. J Clin Oncol 26(9):1404, 2008

1A6 1-9% (low) and ≥10% (high)

Phillips. Appl IHC Molec Morphol 15:325, 2007 1294 Allred Score ≥3 (1-10% weakly positive cells)

Dowsett. J Clin Oncol 26:1059, 2008 312 ≥10%

It is NOT mandatory to use the same validated

antibodies and assays identified in the Guidelines

It IS mandatory to get the same results

A Few Words about Validation

in Your Laboratory

It IS mandatory to provide ongoing proof of

equivalent results

Using the same validated antibodies and assays

is a very reasonable thing to do

FDA approval does not necessarily mean that there

has been comprehensive validation of the assay

A Few Words About Scoring Results

Examples of Satisfactory Methods:

H-Score (McCarty, Arch Pathol Lab Med 109:716, 1985)

Allred-Score (Allred, Mod Pathol 11:155, 1998)

Quick-Score (Rhodes, J Clin Pathol 53:125, 2000)

Percent and intensity positive – by computer (many methods)

Percent and intensity positive – by human (absolute, point-counting)Percent and intensity positive – by human (absolute, point-counting)

Clinical Relevance:Tam = Tam+Chemo in postmenopausal patients with HIGH ER

e.g. Albain. Lancet Oncol. 11:55, 2010

Neoadjuvant endocrine therapy = restricted to patients with HIGH ER

e.g. Ellis. JNCI 100:1380, 2008

⇒ Must avoid using IHC assays that are too sensitive (saturated)

A Few Words about Controls

Multiple purposes

confirm satisfactory performance of assay

confirm satisfactory condition of sample

Perfect control does not currently exist

Certain normal tissues are satisfactory practical external controls

(e.g. endometrium; normal breast = external and internal control)

readily availablereadily available

relatively stable reactivity

broad range of expression

∼90% normal breast tissue with positive epithelial cells...but not 100%

do NOT use breast cancers (abnormal variable expression)

Routinely score and record results to illuminate problems

utilizing same control daily can be very helpful

Every batch of cases must have positive and negative controls

Every pathologist must review the controls before scoring case

Essential Elements of Quality Assuranceand Proficiency Testing

Confirm accuracy of results

reasonable distribution (70-85% ER-positive; 60-75% PgR+)

stable with time (e.g. day of week; quarter to quarter)

>90% concordance vs. independent expert

Demonstrate reproducibility of results

>90% vs. in-house or outside expert

Demonstrate concordance of results

>90% between pathologists

> 90% by same pathologist

Comprehensive training of pathologistsComprehensive training of pathologists

educational conferences (e.g. review important new publications)

training conferences (e.g. calibrate scoring between pathologists)

Comprehensive training of other laboratory personnel

educational and training conferences

monitor reagents (e.g. lot numbers; expiration dates)

monitor and regulate fixation time

Designate an in-house expert Medical Director

with true expertise (requires substantial training and committment)

monitor quality of slides daily before disseminating to other pathologists

go-to person for general oversight; problem solving; feedback, etc.

Advertise quality of performance

reassurance to others (e.g. annual report at tumor board)

Comprehensive record keeping

all above and more

FactorProportion

Score (PS)

Intensity

Score (IS)

Total

Score

(TS)

Interpretation

Estrogen Receptor 5 2 7Positive

Progesterone Receptor 1 1 2

LEFT BREAST MASS, CORE BIOPSY:Invasive ductal carcinoma

Example of Comprehensive Report

Progesterone Receptor 1 1 2Negative

Comments:

Estrogen receptor (antibody 6F11) and progesterone receptor (antibody 1294) were

evaluated by immunohistochemistry (IHC) on routine formalin-fixed paraffin-embedded

tissue utilizing comprehensively validated assays (J Clin Oncol 17;1474, 1999; Mod

Pathol 17:1545, 2004) in compliance with ASCO/CAP guidelines (Arch Pathol Lab Med

134:907, 2010; J Clin Oncol 28:2784, 2010).

The results were scored and interpreted using the Allred Score (J Clin Oncol 17;1474,

1999). PS (proportion of positive tumor cells): 0=none; 1<1/100th; 2=1/100th-1/10th;

3>1/10th-1/3rd; 4>1/3rd-2/3rds; 5>2/3rds. IS (average intensity of positive tumor cells):

0=none; 1=weak; 2=intermediate; 3=strong. TS = PS+IS (range 0-8); positive >2.

Formalin fixation time: 20 hours.

Essential Reading

American Society of Clinical Oncology/College of American Pathologists guideline recommendations for immunohistochemical testing of estrogenand progesterone receptors in breast cancer. Arch Pathol Lab Med. 2010 Jun;134(6):907-22; JCO, 28:2784-2795, 2010

Recommendations for validating estrogen and progesterone receptor immunohistochemistry assays. Arch Pathol Lab Med 134(6):930-5, 2010

Current issues in ER and HER2 testing by IHC in breast cancer. Mod Pathol. 2008 May;21 Suppl 2:S8-S15

Estrogen receptor analysis. Correlation of biochemical and immunohistochemical methods of using monocloncal antireceptor antibodies. Arch Pathol Lab Med. 1985 Aug;109(8):716-21 (first important publication)

Estrogen receptor status by immunohistochemistry is superior to the ligand-binding assay for predicting response to adjuvant endocrine therapy in breast cancer. J Clin Oncol. 1999 May;17(5):1474-81

Re-evaluating adjuvant breast cancer trials: assessing hormone receptor status by immunohistochemical versus extraction assays. JNCI. 2006 Nov 1;98(21):1571-81.Nov 1;98(21):1571-81.

Prognostic and predictive value of centrally reviewed expression of estrogen and progesterone receptors in a randomized trial comparing letrozoleand tamoxifen adjuvant therapy for postmenopausal early breast cancer: BIG 1-98. J Clin Oncol. 2007 Sep 1;25(25):3846-52.

Chemoendocrine compared with endocrine adjuvant therapies for node-negative breast cancer: predictive value of centrally reviewed expression of estrogen and progesterone receptors--International Breast Cancer Study Group. J Clin Oncol. 2008 Mar 20;26(9):1404-10.

Immunohistochemical detection using the new rabbit monoclonal antibody SP1 of estrogen receptor in breast cancer is superior to mouse monoclonal antibody 1D5 in predicting survival. J Clin Oncol. 2006 Dec 20;24(36):5637-44. Epub 2006 Nov 20.

Development of standard estrogen and progesterone receptor immunohistochemical assays for selection of patients for antihormonal therapy. Appl Immunohistochem Mol Morphol. 2007 Sep;15(3):325-31.

Relationship between quantitative estrogen and progesterone receptor expression and human epidermal growth factor receptor 2 (HER-2) status with recurrence in the Arimidex, Tamoxifen, Alone or in Combination trial. J Clin Oncol. 2008 Mar 1;26(7):1059-65. Epub 2008 Jan 28.

Progesterone receptor by immunohistochemistry and clinical outcome in breast cancer: a validation study. Mod Pathol. 2004 Dec;17(12):1545-54.

Resistance to Hormone Therapies in

Breast Cancer: Mechanisms and

Clinical Implications

C. Kent Osborne C. Kent Osborne

Dan L. Duncan Cancer Center

Lester and Sue Smith Breast Center

Baylor College of Medicine, Houston, TX

Estrogen Action in Normal and

Cancer

• Requires binding of E to the estrogen receptor (ER).

• ER is predominantly a nuclear protein that acts as a

transcription factor.

• ER also has a non-nuclear function to activate growth • ER also has a non-nuclear function to activate growth

factor signaling .

• ER controls hundreds of genes important for

proliferation, cell survival, angiogenesis, other.

• Blocking ER is very effective treatment in appropriate

tumors.

• Proper measurement of ER is crucial.

Types of Endocrine Therapy (Block ER)

1. Estrogen deprivation

- ovarian ablation/suppression in premen

- aromatase inhibitors in postmen

2. ER blockade (SERMS)2. ER blockade (SERMS)

- tamoxifen

3. ER downregulators

- fulvestrant

Benefits of Endocrine Therapy

• Metastatic breast cancer (macrometastasis)

- Clinical benefit rate of 50-60% in ER+

- Median survival of 5-6 years (rare 10+ yrs)

• Adjuvant therapy after primary surgery • Adjuvant therapy after primary surgery

(micrometastasis)

- Reduction of recurrence of 50% if ER+

- Probably cures in many patients

But, de novo and acquired resistance common

General Types of Resistance

1. De Novo.

2. Acquired, with ER still present but not

functioning.

3. Acquired drug resistance but ER still 3. Acquired drug resistance but ER still

functioning.

4. Acquired with ER lost.

Drug resistance vs loss of E dependence and

acquisition of an alternate survival pathway.

Clinical Clues

• Loss of ER expression.

• Multiple responses to sequential endocrine therapies

over time; “drug” resistance but not loss of E

dependence.

• ER level decreases over time; gradual loss of E

dependence.

• PR is lost 50% of time with resistance to tam; tumor

progresses faster when lost.

• Tumors with high HER2 or EGFR have lower ER and

PR and less response to endo therapy.

Mechanism of Resistance

• The tumor evolves and activates other survival

pathways (escape pathways) to bypass the E

block.

Driver and Escape Pathways in

Cancer

Survival / Proliferation

2

1

ab

c3

2

Biological Systems in Normal and

Cancer Cells

1. Complex

2. Multiple control mechanisms

3. Fine tuning

4. Redundant4. Redundant

5. Evolvable

Makes cells difficult to kill!

Escape Pathways

• Little data

• Few resistant/metastatic tumors to study

• Reliance on preclinical models

ER ER

ER

ER

ER

Ras PI3K

Erk Akt

Src

Erk

Estrogen

GFRTKs:EGFR, HER2, IGFR

Akt

SrcPI3K FAK

SrcPELP1

A

B

C C

C

Molecular Action ofE2/ER

Membrane

PAK

C

ER

D

Microenvironment

Stress

p38 JNK

INGs

FAK

ER

ER

ER

ER

HATCoA CoA

AP1 SP1

ER

Adapted from:

Musgrove & Sutherland,

Nature Reviews Cancer, 2009

TFs TFs

TFs TFs

ERE SRE RE

BCL-2, MYC, CyclinsD1, E1, E2

Cytoplasm

Nucleous

Growth, Proliferation, Survival, Angiogenesis

ER ER

ER

ER

ER

Ras PI3K

Erk Akt

Src

Erk

Endocrine Therapy

GFRTKs:EGFR, HER2, IGFR

Akt

SrcPI3K FAK

SrcPELP1

A

B

D D

DMembrane

PAK

C

ER

E

Microenvironment

Stress

p38 JNK

INGs

FAK

Possible Escape Pathways toEndocrine Resistance

Cellular/signaling Kinases:MEK/MAPKPTEN/PI3K/AktP38/JNKSrcFAK

Stress Responses:TreatmentsCytokines

HypoxiaROS

TKR/GF:EGFR/HER2

IGFR, FGFR, VEGFR2EGF, TGFαααα,IGF1, IGFII, HRG, FGF

ER

ER

ER

ER

HATCoA CoA

AP1 SP1

ER

Adapted from:

Musgrove & Sutherland,

Nature Reviews Cancer, 2009

TFs TFs

TFs TFs

ERE SRE RE

BCL-2, MYC, CyclinsD1, E1, E2

Cytoplasm

Nucleous

Resistance: Growth, Proliferation, Survival, Angiogenesis

NR:ERαααα: variants, Erβ: low AR

CoA/RHigh P160 (SRC1/3)Low NCoR/SMART

TF:AP1SP1NFkB

Pathways:MycCyclin D1, E1, E2Rb, p53, p21/27

Proapoptotic/ Survival

(Bcl, BAD,XBP1)

800

1000

1200Tumorvolume(mm3)

Tam-S Tam-R

De novo Tam-R

In Vivo Model of Endocrine Resistance

Acquired Resistance to Tamoxifen is Associated with Increased Levels of

EGFR and HER2

0

200

400

600

800

Days

Tam

MCF7

MCF7/HER2

Tam-S Tam-R

EGFR

HER2

+E2

Massarweh et al., SABCS 2004

(Benz et al., Breast Cancer Res Treat, 1992)

800

1000

1200

1400

Tum

or volum

e E2 TAM

E2+gefitinib

Overcoming Tam Resistance with Gefitinib in HER2-

Positive Tumors

0

200

400

600

800

Days

Tum

or volum

e

E2+gefitinib

TAM+gefitinib

1 30 60 90 120

Shou J, JNCI 2004Massarweh S, ASCO 2002

Hypothesis: GefitinibR is due to incomplete

blockade of the HER signaling pathway

(all HER dimer pairs).

Reversal of Tam Resistance with Gefitinib in

HER2-Negative Tumors

800

1000

1200

1400

Tum

or volum

e

E2E2+ gefitinib

TAM

0

200

400

600

800

Days

Tum

or volum

e

TAM+gefitinib

Massarweh et al., SABCS 2002

Conversion from HER2- to HER2+

Study % Conversion

Gutierrez (tumor) 12

Uhr (CTCs) 37Uhr (CTCs) 37

Lipton (serum) 26

Most converters had intervening endo Rx

tamoxifen 20 mg / day +

gefitinib 250 mg / day

0225 trial: study design

Patients

• Postmenopausal women

• Age ≥ 18 years

• Stratum 1: Newly diagnosed ER- and / or PgR-positive MBC or disease recurring after adjuvant tamoxifen,

Primary• PFS in Stratum 1• CBR (CR + PR + SD for ≥24 weeks using RECIST)in Stratum 2

Response variables

N=290 (206 in Stratum 1 & 84 in Stratum 2)

tamoxifen 20 mg / day + placebo

1:1 randomisationadjuvant tamoxifen, completed ≥1 year before study entry

• Stratum 2: Disease recurring during or after AI therapy or who have failed first-line AI therapy for MBC

• PS 0-2

• No prior chemotherapy for metastatic disease

in Stratum 2

Secondary• CBR in Stratum 1• PFS in Stratum 2• ORR• PFS in patients with HER2-expressing tumours

• Safety and tolerability• PK

Osborne et al, manuscript submitted

Until disease progression or

other event requiring discontinuation

0225 trial: PFS in Stratum 1 patients

PFS similar over first 200 days

and KM curves then diverge

compatible with a delay in the

development of resistance similar

to that in preclinical models

A more substantial difference in

PFS observed for the HER2

positive subset of patients (n=37)

Osborne et al, manuscript submitted

anastrozole 1 mg / day + gefitinib 250 mg / day

0713 trial: study design

Patients

• Postmenopausal women

• Age ≥ 18 years

• Newly diagnosed ER-and / or PgR-positive metastatic breast cancer

Primary• PFS

Secondary

Response variables

N=94

anastrozole 1 mg / day +

placebo

1:1 randomisationcancer

• No prior hormonal therapy, or development of metastatic disease during / after adjuvant tamoxifen

• Measurable or non-measurable disease (via RECIST)

Secondary• ORR• CBR• OS• Safety and tolerability• Expression of biomarkersin tumour tissue sections

Cristofanilli et al, Abstract 1012, oral presentation, ASCO 2008

Until disease progression or

other event requiring discontinuation

Progression-free survival

Probability of PFS

1.0

0.8

0.6

EventsMedian PFS (months)

2214.5

328.2

Gefitinib +anastrozole

(n = 43)

Placebo +anastrozole

(n = 50)

HR (95% CI) = 0.55 (0.32, 0.94)

30

0.4

0.2

0.0

0 3 6 9 12 15 18 21 24 27

5043

3540

2328

1322

913

610

56

33

12 1

PlaceboGefitinib

At risk:Months

Reversal of Tam Resistance with Gefitinib in

HER2- Tumors

800

1000

1200

1400

Tum

or volum

e

E2E2+ gefitinib

TAM

0

200

400

600

800

Days

Tum

or volum

e

TAM+gefitinib

Massarweh et al., SABCS 2002

Other Escape Pathways Identified

1. Oxidative stress/increased AP1 activity

2. Integrin signaling/increased pSrc and pFac

3. PI3K/AKT pathway activation

Oxidative stress and cancer

•Oxidative stress has been defined as “a disturbance in the pro-oxidant-antioxidant balance in favor of the former, leading to potential damage” (Sies, 1991)

• There is increasing evidence that malignant cells are in a pro-oxidant state due to:

• increased formation of reactive oxygen species

• decreased antioxidant defenses (Halliwell B, Biochem J. 2007)

Oxidants

Antioxidants

Oxidants Antioxidants

HER/IGFR

GFRs Cytokines

Stress

Endocrine Resistance: ER/Signaling Networks Crosstalk

Integrins

PI3K/Akt

ER

ERE

TamProliferation

Invasiveness

CoR

ERK1/2 p38 JNK

PI3K/Akt

SRC

FAK

Proliferation

Invasiveness

CoA

TFs

cFos cJun

AP-1

ERTam

Endocrine resistance?

Proliferation

Invasiveness

CoA

Oxidative stress and endocrine resistance

• Tamoxifen can alter the redox status of the cell both as an oxidant and as an antioxidant

• Resistance to endocrine treatment is associated with

oxidative stress (Schiff R et al., JNCI, 2000)oxidative stress (Schiff R et al., JNCI, 2000)

Stress-related pathways and endocrine resistance

• The stress-related kinase pathways Jun N-terminal kinase (JNK)/AP-1 and p38 MAPK are key mediators of cell response to oxidative stress

• JNK overexpression is a poor prognostic factor (Yeh et al.Int J Cancer 2006)Int J Cancer 2006)

• cJun overexpression in ER+ breast cancer cells results in endocrine resistance (Smith et al. Oncogene 1999)

• JNK activity and AP-1 levels are higher in Tam resistant human tumors (Johnston et al. Clinical Caner Res, 1999)

TransactivationDNA

Binding Dimerization

cJun

MCF7 Tet-Off

Genetic approach (DN cJun)

DN cJun (Tam 67)

Inhibits AP-1 activity in different cell lines

cJun

DN cJun

(Ludes-Meyers JH et al., Oncogene, 2001)

DN cJun enhances response to Tam in vivo

DNcJun Clone 67

0.0

0.2

0.4

0.6

0.8

1.0

0 30 60 90 120 150 180 210 240Pro

po

rtio

n N

ot

Re

sp

on

de

rs

40% non-responders

ControlDN cJun

p=.04

0.0

0.2

0.4

0.6

0.8

1.0

0 50 100 150 200

ControlDN cJun

p=.014

Pro

po

rtio

n N

ot

Re

sp

on

de

rs

30% non-responders

DNcJun Clone 62

Time to Response

0 30 60 90 120 150 180 210 240Pro

po

rtio

n N

ot

Re

sp

on

de

rs

85% non-complete responders

0.0

0.2

0.4

0.6

0.8

1.0

0 50 100 150 200 250

Time to Complete Response

Su

rviv

al P

rob

ab

ilit

y

ControlDN cJun

p=.0034

0.0

0.2

0.4

0.6

0.8

1.0

0 50 100 150 200

ControlDN cJun

p=.001

Time to Complete Response

Su

rviv

al P

rob

ab

ilit

y

0 50 100 150 200Pro

po

rtio

n N

ot

Re

sp

on

de

rs

Time to Response

60% non-complete responders

DN cJun delays TamRes onset in vivo

0.6

0.8

1.0

Su

rviv

al

Pro

ba

bilit

yDNcJun Clone 67

50% tumor

progression

Control

DNcJunp=.0028

0.0

0.2

0.4

0 5 0 10 0 150 2 00 2 50

Time to tumor size doubling

Su

rviv

al

Pro

ba

bilit

y

0.4

0.6

0.8

1.0

Su

rviv

al P

rob

ab

ilit

y

ControlDN cJun

p= 0.31

0.4

0.6

0.8

1.0

Su

rviv

al P

rob

ab

ilit

y

ControlDN cJun

p= 0.6

DN cJun does not affect E2 stimulated growth in vivo

DNcJun Clone 67 DNcJun Clone 62

0.0

0.2

0.4

Su

rviv

al P

rob

ab

ilit

y

0 20 40 60 80

Time to tumor size doubling

0.0

0.2

0.4

0 10 20 30 40 50 60

Time to tumor size doubling

Su

rviv

al P

rob

ab

ilit

y

Summary and Conclusions

1. Resistance to endo therapy:

- reactivation of the ER pathway or

- activation of another escape pathway

2. Growth factor receptors, integrins, and stress pathways are

prime candidates for escape from ER blockade.prime candidates for escape from ER blockade.

3. Identification or anticipation of which pathways will take

over when ER is blocked will be crucial for personalized

treatment.

4. Treatment will require combination therapy.

5. Biopsies of resistant tumors in patients are crucial for study.

1

Molecular Classification of Oestrogen Receptor (ER) Positive/ Luminal Breast Cancers

Jorge S Reis-Filho 1, MD PhD FRCPath & Britta Weigelt 2, PhD

1 – The Breakthrough Breast Cancer Research Center/ Institute of Cancer Research,

London, UK; 2 – Cancer Research UK London Research Institute, Lincoln’s Inn Fields,

London, UK

Email: Jorge.Reis-Filho@icr.ac.uk

INTRODUCTION

Breast cancer comprises a complex and heterogeneous group of diseases. Although the

heterogeneity of breast cancer was known for a long time, it was only after seminal studies

using high throughput transcriptomic methods that this concept was brought to the forefront

of breast cancer research, and, most importantly, clinical practice [1-4]. It is currently

accepted that the heterogeneity of breast cancer is such that it encompasses different

diseases that have distinct risk factors, clinical presentation, histopathological features,

molecular characteristics, response to therapies and clinical behaviour, which happen to

affect the same anatomical site and to originate from cells in the same microanatomical

structure (i.e. terminal duct-lobular unit) [1-4].

Despite this heterogeneity, the management of breast cancers is currently based on a

constellation of clinicopathological features that are derived from careful histopathological

analysis of primary cancers, including tumour size, histological subtype and grade, lymph

node metastases, and lymphovascular invasion [1-5]. In addition, three predictive markers

have been incorporated in breast cancer patient care, namely oestrogen (ER) and

progesterone (PR) receptors, which are the predictive markers of response to endocrine

therapy, and HER2, which is the predictive marker and molecular target of Trastuzumab and

Lapatinib [2-5]. This information is then used according to guidelines or using multiple

parameter algorithms for clinical decision making (e.g. Adjuvant! Online) [1-5]. These

approaches, albeit simplistic, have been proven to work, given that the mortality of breast

cancer has declined over the last two decades despite the increase in the incidence of breast

cancer. Furthermore, the predictions made by multiparameter algorithms largely meet the

actual outcome of breast cancer patients [6]. It has become blatantly clear, however, that

these approaches only provide information about the best therapy for the average breast

cancer patient with a given constellation of clinicopathological features, and is not sufficient

for the implementation of individualised therapies.

2

In addition to the development of the concept of personalised medicine, the advent of high

throughput technologies that could be applied to the study of solid malignancies have led to a

rediscovery of the heterogeneity of cancers [7-12]. In particular, these methods have led to

the development of a molecular taxonomy for breast cancers, which was initially believed to

have histogenetic implications [7, 9, 13]. Microarray-based gene expression profiling studies

have also provided direct evidence to demonstrate that ER-positive and negative breast

cancers have remarkably distinct transcriptomes [1-3]. This has led to the realisation that ER-

positive and negative diseases are fundamentally different. Furthermore, multi-gene

predictors, which have been suggested as alternatives for the current clinicopathological

algorithms for treatment decision making, have been developed.

Oestrogen receptor positive disease comprises the vast majority of breast cancers

diagnosed, therefore it is not surprising that numerous potential prognostic and predictive

markers that can be used for the management of patients with ER-positive disease have

emerged [1, 2, 5]. In this manuscript, I will review the impact of the molecular classification of

breast cancers, its impact on our understanding of ER-positive disease, and the current

multi-gene predictors that can be used for the management of patients with ER-positive

breast cancer.

BREAST CANCER MOLECULAR CLASSIFICATION: EMPEROR’S NEW CLOTHES?

Back in the late 90s, the promise of microarrays was of apocalyptic dimensions, with one of

the proponents of this technology suggesting that treatments or cures for all human illnesses

would be found by applying this technology to the study of human cancers. In fact, a

paradigm shift in terms of our understanding of breast cancer took place with the publication

of the seminal class discovery studies published by the Stanford group [7], where the

heterogeneity and complexity of breast cancers were re-discovered at the molecular level.

Perou et al. [7] analysed 38 invasive breast cancers (36 invasive ductal and 2 lobular

carcinomas), 1 ductal carcinoma in situ, 1 fibroadenoma and 3 normal breast samples, and a

number of biological replicates of tumours from the same patients with cDNA microarrays,

and defined an ‘intrinsic gene’ list (i.e., the genes that vary more among tumours from

different patients than in samples from the same tumour) [7]. The approach employed at that

time may now sound quaint to the average microarrayer, but in early 00s they had a major

impact on how breast cancer was perceived. Hierarchical cluster analysis using this ‘intrinsic’

gene list revealed that ER-positive and ER-negative breast cancers are fundamentally

distinct at the transcriptomic level [7-11], and also demonstrated the existence of molecular

3

subtypes of breast cancer: luminal, normal breast-like, HER2 and basal-like [2, 8-10, 14-16].

In the original publication, ER-positive cancers were classified as luminal tumours, whereas

ER-negative cancers comprised the so-called basal-like, HER2 and normal breast-like

cancers. When this approach was applied to additional breast cancer microarray datasets

linked to outcome information, it was demonstrated that i) similar molecular subtypes of

breast cancer could be identified in multiple cohorts of breast cancers [8-10, 15], ii) that

luminal cancers could be subclassified into two (luminal A and B) [8] or three groups (luminal

A, B and C) [10], and iii) that different molecular subtypes were associated with distinct

clinical outcomes [8, 15, 17]. According to this classification, luminal cancers are

characterised by the expression of ER and ER-related genes, and can be subclassified into

two groups (i.e. luminal A and luminal B) according to the expression levels of proliferation

related genes. Luminal A tumours express the highest levels of ER and ER-related genes,

and the lowest levels of proliferation-related genes; on the other hand, luminal B cancers

express the opposite gene expression pattern [1, 2, 10, 18]. Basal-like breast cancers are

characterised by the expression of genes usually found in ‘basal’/ myoepithelial cells of the

normal breast, and express high levels of proliferation related genes. HER2 or HER2-

enriched breast cancers are characterised by the expression of HER2 and genes mapping to

the HER2 amplicon. Normal breast-like cancers are still poorly understood, however their

defining feature is that they consistently cluster together with samples of normal breast and

fibroadenomas. In terms of outcome, luminal A cancers were shown to have the best

prognosis, whereas basal-like tumours to have the worst outcome [8-10].

This molecular taxonomy for breast cancers has been enthusiastically embraced by

surgeons, oncologists and scientists alike. Several groups have demonstrated that some of

these subtypes (e.g. basal-like) have distinct risk factors, clinical presentation, histological

features, response to therapy and outcome [2, 19-25]. The data accumulated have led to

some experts in the field to suggest that traditional clinicopathological features and

immunohistochemical markers should be replaced by this molecular taxonomy [26].

To quote Amos Elon, “if hindsight really is twenty-twenty it is important to try as often as we

can to analyse, embrace and employ it as we make our way through life”. In 2011, if we look

back at the ‘intrinsic’ molecular subtype classification, it would be fair to temper the

enthusiasm with this taxonomy, as it has numerous limitations [4, 27]. The initial approach

employed for the identification of the molecular subtypes was based on hierarchical

clustering analysis. It should be noted, however, that this approach requires large datasets,

is to some extent subjective, and cannot be employed for the classification of individual

4

samples prospectively [28, 29]. Therefore, the proponents of the molecular taxonomy

developed ‘single sample predictors’ (SSPs). These SSPs are based on the correlation

between the expression profile of a given sample with the centroids for each molecular

subtype (i.e. average expression profile of each molecular subtype) [10]. Our group [18] and

others [30] have recently demonstrated that the identification of molecular subtypes of breast

cancer by SSPs depends on the methodology employed, and only basal-like cancers can be

reliably identified. In fact, the subclassification of luminal tumours into A and B subclasses

was strongly dependent on the SSP used and different patients were classified as A or B

depending on the methodology employed [18, 30]. In fact, even when the authors of the

molecular taxonomy themselves classified the same cohort of breast cancer patients (that is,

NKI-295 [17]) using two different methods [9, 10], one by Sorlie et al.[9, 31] and the other by

Hu et al.[10, 32], the agreement was only moderate (Kappa scores = 0.527 (95% confidence

interval 0.456 to 0.597)). Second, there are several lines of evidence to suggest that normal

breast-like cancers may constitute an artefact of gene expression profiling (that is, samples

with a disproportionately high content of normal breast epithelial cells and stromal cells) [2,

11, 12, 18]. Third, and perhaps most importantly for the main topic of this session, given that

the subdivision of luminal tumours into A and B is driven by the levels of expression of

proliferation-related genes and that several studies have demonstrated that the expression

levels of proliferation-related genes in ER-positive cancers are a continuum and do not

display a bi-modal distribution, the subclassification of luminal cancers is likely to be arbitrary

[2, 18, 30, 33, 34]. Fourth, the HER2 molecular subtype neither comprises all cases

classified as HER2-positive by FDA approved methods (i.e. immunohistochemical analysis

and chromogenic/fluorescence in situ hybridisation) and not all HER2-positive cancers by

FDA approved methods are classified as HER2 subtype by microarrays [11, 18, 35]. Finally,

contrary to the initial belief that luminal tumours would originate from luminal cells and basal-

like cancers from ‘basal’ cells of the normal breast, recent studies [36, 37] have

demonstrated that the likeliest cell of origin of basal-like breast cancer resides in the luminal

progenitor compartment.

Therefore, I argue that the ‘intrinsic’ molecular subtype classification of breast cancer is not

yet ready for clinical use in prognostic models or otherwise. Standardisation of the definitions

and the methodologies for the identification of the molecular subtypes and prospective

clinical trials to validate the contribution of these five molecular subtypes in addition to the

current clinicopathological parameters for the management of breast cancer patients are still

required [18, 38]. Although the qRT-PCR based test for the identification of the subtypes (i.e.

5

PAM50 [11, 26]) is an interesting approach, independent validation of its robustness and its

prognostic and predictive values have yet to be published.

PROGNOSTIC GENE SIGNATURES: PROLIFERATION BY ANOTHER NAME

Microarray-based prognostic gene signatures were heralded as a major breakthrough for the

management of breast cancer patients, as initial studies claimed that these signatures would

provide a more objective assessment of the risk of relapse of breast cancer patients and

would be more reproducible than the methods currently used. The first prognostic gene

signatures (70-gene signature also known as Mammaprint® [14, 17] and the 76-gene

signature [39, 40]) were developed to be applied to all breast cancer patients. Numerous

studies provided evidence to demonstrate that the prognostic information provided by these

signatures is indeed independent of the information provided by tumour size, presence of

lymph node metastasis and histological grade [14, 17, 39, 40]. Subsequent to these initial

stories of success, several groups developed their own prognostic signatures either

employing bottom-up or top-down approaches (for reviews, see [1, 2]). Furthermore,

microarray signatures to capture the information provided by histological grade were

developed and shown to be independent predictors of outcome [41, 42].

Following the initial over-hyping of microarray-based prognostic gene signatures, the

enthusiasm with this approach waned. Furthermore, re-analyses of the initial studies on

cancer prognosis with microarrays demonstrated that the overlap between gene signatures

was negligible [1, 2]; that the gene composition of first generation signatures was not stable

[43, 44]; and that these gene signatures were strongly time dependent [39]. The wave of

often unjustified scepticism that followed, was possibly an over-reaction. One expert in the

field of biomarker discovery and validation stated that “on close scrutiny, in five of the seven

largest studies on cancer prognosis, this technology performs no better than flipping a coin.

The other two studies barely beat horoscopes" [45]. Again, hindsight is a beautiful thing, in

particular when it comes to the contribution of new technologies to science. With the

availability of microarray datasets in public repositories, meta-analyses performed by

independent groups revealed that:

i) different gene signatures do identify similar (but not necessarily identical) groups of

patients as of poor outcome [32, 34, 46];

ii) the assignment of cases as of poor outcome is based on the expression of proliferation-

related genes [18, 33, 34];

6

iii) the discriminatory power of first generation signatures is restricted to ER-positive breast

cancers; in fact, <5% of ER-negative breast cancers are classified as of good prognosis

using these approaches [33, 34];

iv) first generation prognostic gene signatures do not identify prognostically significant groups

of ER-negative disease [33, 34];

v) that rare types of ER-negative breast cancers that have an indolent outcome (e.g. adenoid

cystic carcinomas) are consistently classified as of poor prognosis [47];

vi) proliferation is perhaps the strongest determinant of outcome in ER-positive disease. In

fact, when first generation prognostic gene signatures are divided into sub-signatures

composed of ‘proliferation-related genes’ and ‘non-proliferation-related genes’, and only the

former are used, the overall performance was not degraded and, improved for some

signatures. On the other hand, sub-signatures composed of ‘non-proliferation-related genes’

display suboptimal performance [34];

vi) within ER-positive breast cancers, there is a correlation between the poor prognosis

groups ascribed by Mammaprint® and OncotypeDxTM and luminal B cancers [1, 2, 32];

however, the level of agreement between these methods for the identification of poor

prognosis ER-positive cancers is yet to be determined.

In parallel with the development of first generation microarray-based prognostic gene

signatures, a 21-gene signature based on quantitative real time RT-PCR was developed

through a re-analysis of microarray datasets and a review of the literature [48, 49].

OncotypeDX™(Genomic Health, Redwood, CA, USA) was developed and validated through

a retrospective analysis of formalin-fixed, paraffin-embedded material from the prospective

clinical trials B-20 and B-14 (for reviews, see [50, 51]). The signature includes the expression

assessment of 5 reference genes for the standardisation of the relative quantification and 16

prognostic genes that are related to proliferation (KI-67, STK15, Survivin, CCNB1 and

MYBL2), invasion (Stromelysin 3 and Cathepsin L2), HER2 group (HER2 and GRB7),

oestrogen receptor signalling (ER, PR, Bcl2 and SCUBE2), and GSTM1, BAG1 and CD68

[48]. The expression of these genes is presented as a Recurrence Score (RS) ranging from 0

to 100 providing an estimate of 10-year distant recurrence-risk. For clinical use, patients are

separated in 3 categories: low RS (RS<18), intermediate RS (RS≥ 18 and <31) and high RS

(RS≥31) [48]. OncotypeDx™ was developed and validated in patients with ER-positive,

node-negative breast cancers using retrospectively the material of 2 randomised trials (i.e.

NSABP-B-20 and NSABP-B-14), and this signature was shown to outperform standard

clinico-pathological parameters for the prediction of 10-year distant recurrence-risk [48]. The

21-gene has been subsequently evaluated in other cohorts of breast cancer patients [52] and

7

shown to be an independent prognostic parameter in patients with ER-positive tumours with

up to 3 positive-nodes receiving adjuvant chemotherapy [53], and in postmenopausal

patients with ER-positive tumours treated with anastrozole [54].

OncotypeDx™ RS has also been shown to be correlated with the benefit patients derive from

adjuvant chemotherapy in samples from clinical trials [49, 55, 56]. In fact, patients with

tumours displaying high RSs, despite their poor prognosis, derive significantly more benefit

from chemotherapy than those with low RS tumours. In addition, patients with low RS

cancers appear to derive negligible benefit from the addition of chemotherapy to tamoxifen.

Therefore, OncotypeDx™ has also been used as a predictive marker of benefit from

chemotherapy.

Level II evidence in support of the prognostic role of OncotypeDx™ has already been

accrued. Therefore, it has received the approval from the American Society of Clinical

Oncology (ASCO) [57] and was included in the National Comprehensive Cancer Network

guidelines (NCCN guidelines Breast Cancer version 1.2011 - http://www.nccn.org/) as an

option to evaluate prognosis and to predict response to chemotherapy for ER-positive, node-

negative breast cancer patients, as a complement to clinico-pathological features. None of

the other prognostic signatures has been endorsed by these professional bodies.

One important point that should not be overlooked is the fact that there is evidence to

suggest that the prognostic power of OncotypeDx™, in a way akin to the other first

generation signatures, largely if not exclusively stems from the quantitative analysis of the

levels of expression of proliferation-related genes.

Hence, one could claim that first generation prognostic signatures only have prognostic

power in ER-positive disease [27], and that this prognostic information is derived from the

analysis of the expression levels of proliferation-related genes. In fact, recent comparisons of

the prognostic information provided by OncotypeDx™ or four immunohistochemical markers

(i.e., ER, PR, HER2 and Ki67 - a proliferation marker) semi-quantitatively analysed in the

material from the ATAC (Arimidex, Tamoxifen, Alone or in Combination) prospective trial

demonstrated that these four markers would provide prognostic information that is at least be

equivalent to OncotypeDx™ [58].

Given the importance of proliferation, and the fact that multiple studies have provided level III

evidence in support of Ki67 as an independent predictor of outcome, why has Ki67, a

8

proliferation marker routinely used in surgical pathology laboratories, not been incorporated

in immunohistochemistry-based prognostic panels? The answer to this question lies in the

definition of biomarkers (i.e. “a characteristic that is objectively measured and evaluated as

an indicator of normal biological processes, pathogenic processes, or pharmacologic

responses to a therapeutic intervention”)[59]. The assessment of Ki67 in different pathology

laboratories has yet to be standardised. Multiple antibodies are available and different

laboratories perform this test using different antigen retrieval methods and antibody dilutions;

furthermore, the quantification of Ki67 labelling indices has some degree of subjectivity, as it

is based on the assessment of ‘hot-spot’ areas. However, international consortia of experts in

the field of biomarker discovery and validation have now turned their attention to pre- and

analytical parameters required for a standardised Ki67 immunohistochemical test. Guidelines

should be available in 2011.

In any case, acknowledgement of the existence of a continuum of proliferation levels in the

group of ER-positive breast cancer [1, 2, 18, 33, 34] and that the extremes of this continuum

have dramatically different outcomes and probably responses to chemotherapy is of utmost

importance and clinical trials testing the therapeutic efficacy of new agents should take this

information into account.

PREDICTIVE SIGNATURES: CAN THEY BE INCORPORATED IN CLINICAL PRACTICE?

First generation prognostic signatures can also be used as predictors of response to multi-

drug chemotherapy regimens in ER-positive breast cancers. Studies have demonstrated that

OncotypeDx™ can be used to define which patients should receive chemotherapy in addition

to endocrine therapy [49]. There is also evidence to suggest that Mammaprint® [14, 60] and

Genomic Grade Index [42, 61] can also be used in this context. The clinical utility of these

signatures in the context of prediction of response to multi-drug chemotherapy regimens

stems from the fact that they are surrogates of proliferation and that proliferation is

associated with response to multi-drug chemotherapy.

In terms of predictive markers of endocrine therapy, some considerations need to be made.

ER status has a strong negative predictive value for response to endocrine therapy (i.e.

patients with ER-negative disease do not respond to endocrine therapy). ER expression,

however, is not sufficient to predict which ER-positive tumours will respond to hormone

therapies [5, 62]. Using microarrays technologies, gene expression signatures have been

developed in several studies to predict outcome in tamoxifen treated. A promising approach

is the sensitivity to endocrine therapy (SET) index, which was developed through the

9

transcriptomic analysis of a large series of ER-positive breast cancers [63]. The SET index is

based on the principle that expression of genes correlated with ER may predict better

response to endocrine treatment than ER expression alone. Microarray analysis was used to

identify 165 genes co-expressed either positively (n=109) or negatively (n=59) with ER in a

discovery cohort of 437 breast cancers. Cut-off points were determined in a validation cohort

of 245 patients to define 3 categories of sensitivity (low, intermediate and high). Association

between SET and outcome was then analysed in 3 types of ER-positive cohorts receiving

either adjuvant tamoxifen for 5 years or neo-adjuvant chemotherapy followed by endocrine

therapy (i.e. tamoxifen or aromatase inhibitors) or no adjuvant systemic treatment. The SET

index was significantly associated with the outcome of patients receiving any type of

endocrine treatment (tamoxifen or chemo-endocrine treatment) but had no prognostic value

in untreated patients. Unlike other multi-gene signatures evaluating proliferation in ER-

positive tumours, the SET index seems to be predictive of benefit from endocrine therapy

independently of the inherent prognosis of the tumour. A potential clinical application of the

SET index is in the identification of a subset of ER-positive tumours associated with an

excellent prognosis and no relapse in the tamoxifen-treated group (high SET index tumours)

and in the chemo-endocrine group (high and intermediate SET index) [63]. This type of

predictive signatures may constitute one of the ways forward for the molecular stratification

of ER-positive cancers.

MOLECULAR GENETICS OF OESTROGEN RECEPTOR POSITIVE BREAST CANCERS

Studies investigating the patterns of gene copy number aberrations and mutations in breast

cancer have demonstrated that the pattern and type of gene copy number aberrations

segregates with the ER-status breast cancers [64-68]. In fact, while the approximately 80%

of grade I ER-positive breast cancers and 50% of grade III ER-positive tumours harbour

concurrent deletions of 16q and gains of 1q, these changes are found in a small minority of

ER-negative tumours [69] (for a review see [70]). Even when present in ER-negative

cancers, the mechanisms leading to 16q losses in ER-positive and –negative diseases differ:

in ER-positive disease, 16q losses and 1q gains often stem from an unbalanced

chromosomal translocation [i.e. der(16)t(1;16)/der(1;16)], whereas in ER-negative cancers,

deletion of 16q often results from loss of chromosome 16 (i.e. 16p and 16q losses) [69-72].

These lines of evidence have been interpreted as evidence that ER-positive and ER-negative

breast cancers are distinct at the genetic level and that progression from ER-positive to ER-

negative disease is an uncommon biological phenomenon.

10

Contrary to initial observations that progression from low- to high-grade breast cancer would

be incredibly rare [73], the available genomic data suggest that progression from grade I to

grade III ER-positive tumours may happen. In fact, approximately 50% of grade III ER-

positive cancers harbour the typical pattern of gene copy number aberrations found in grade

I tumours (i.e. deletion of 16 and gain of 1q) [69, 70]. Grade III ER-positive cancers, however,

harbour additional genetic changes and more often harbour gene amplifications rarely seen

in grade I ER-positive disease (e.g. 8p11.2, 11q13-q14, 17q21, 17q23.2, and 20q13) [64-69].

Given that histological grade correlates with proliferation in ER-positive breast cancers, that

high proliferation ER-positive breast cancers have a poor prognosis, and that the so-called

luminal B cancers overlaps significantly with the group of ER-positive grade III patients, it is

not surprising that luminal B cancers have been shown to have more complex molecular

karyotypes than luminal A cancers [67, 68, 74].

Importantly, although grade I ER-positive cancers seem to have less genomic aberrations

than high grade ER-positive disease, it should be noted that these tumours do display

varying levels of genetic instability [70]. Furthermore, recent massively parallel sequencing

studies have demonstrated that there is intra-tumour genetic heterogeneity within ER-

positive cancers and that some ER-positive tumours harbour fusion genes [75]; however, no

fusion gene in ER-positive disease has been shown to be recurrent as yet.

ER-positive disease has been shown to harbour numerous gene mutations and the

repertoire is quite vast; it should be emphasised, however, that the majority of mutations

identified so far are present in a minority of lesions. One of the most prevalently mutated

genes in ER-positive breast cancers is PIK3CA [76], which is an integral component of the

PI3K-AKT-mTOR pathway. Interestingly, in vitro and clinical studies have suggested that

tumours with PIK3CA mutations may be sensitive to inhibitors of mTOR (e.g. rapalogs) and

small molecule inhibitors that inhibit PIK3CA, TORC1 and TORC2 [77].

CONCLUSION

ER positive disease comprises a spectrum of tumours, with varying degrees of proliferation

and levels of genetic aberrations. Proliferation as defined by microarray-based gene

signatures and OncotypeDxTM has been shown to be one of the main independent prognostic

markers for patients with ER-positive disease, and also to be a predictive marker of benefit

for addition of multi-drug chemotherapy to endocrine therapy. As a group, these tumours

respond to endocrine therapies, however a substantial proportion of cases are either de novo

resistant or develop resistance over time. Active research using high throughput methods is

11

currently being performed to identify the potential mechanisms of resistance to hormone

therapies and ways to circumvent them. Despite the enthusiasm with the use of the

molecular taxonomy for breast cancers and the terminology luminal A and luminal B [26],

recent studies have called into question the reproducibility of these subtypes [18, 30, 38]. In

fact, their clinical utility remains to be determined.

With the advent of massively parallel sequencing and the ability to characterise the entire

genomes of cancers, it is likely that the drivers of ER-positive disease will soon be identified.

It is anticipated that the repertoire of mutations in breast cancer will be vast with few

recurrent genetic lesions. Nevertheless, this deluge of information in conjunction with

functional genomic approaches may expedite the development of predictive classification

systems for ER-positive disease.

REFERENCES

1. Sotiriou C, Pusztai L. Gene-expression signatures in breast cancer. N Engl J Med 2009; 360: 790-800.

2. Weigelt B, Baehner FL, Reis-Filho JS. The contribution of gene expression profiling to breast cancer classification, prognostication and prediction: a retrospective of the last decade. J Pathol 2010; 220: 263-280.

3. Weigelt B, Geyer FC, Reis-Filho JS. Histological types of breast cancer: how special are they? Mol Oncol 2010; 4: 192-208.

4. Weigelt B, Reis-Filho JS. Histological and molecular types of breast cancer: is there a unifying taxonomy? Nat Rev Clin Oncol 2009; 6: 718-730.

5. Weigel MT, Dowsett M. Current and emerging biomarkers in breast cancer: prognosis and prediction. Endocr Relat Cancer 2010; 17: R245-262.

6. Mook S, Schmidt MK, Rutgers EJ, et al. Calibration and discriminatory accuracy of prognosis calculation for breast cancer with the online Adjuvant! program: a hospital-based retrospective cohort study. Lancet Oncol 2009; 10: 1070-1076.

7. Perou CM, Sorlie T, Eisen MB, et al. Molecular portraits of human breast tumours. Nature 2000; 406: 747-752.

8. Sorlie T, Perou CM, Tibshirani R, et al. Gene expression patterns of breast carcinomas distinguish tumor subclasses with clinical implications. Proc Natl Acad Sci U S A 2001; 98: 10869-10874.

9. Sorlie T, Tibshirani R, Parker J, et al. Repeated observation of breast tumor subtypes in independent gene expression data sets. Proc Natl Acad Sci U S A 2003; 100: 8418-8423.

10. Hu Z, Fan C, Oh DS, et al. The molecular portraits of breast tumors are conserved across microarray platforms. BMC Genomics 2006; 7: 96.

11. Parker JS, Mullins M, Cheang MC, et al. Supervised risk predictor of breast cancer based on intrinsic subtypes. J Clin Oncol 2009; 27: 1160-1167.

12. Peppercorn J, Perou CM, Carey LA. Molecular subtypes in breast cancer evaluation and management: divide and conquer. Cancer Invest 2008; 26: 1-10.

13. Stingl J, Caldas C. Molecular heterogeneity of breast carcinomas and the cancer stem cell hypothesis. Nat Rev Cancer 2007; 7: 791-799.

12

14. van 't Veer LJ, Dai H, van de Vijver MJ, et al. Gene expression profiling predicts clinical outcome of breast cancer. Nature 2002; 415: 530-536.

15. Sotiriou C, Neo SY, McShane LM, et al. Breast cancer classification and prognosis based on gene expression profiles from a population-based study. Proc Natl Acad Sci U S A 2003; 100: 10393-10398.

16. Correa Geyer F, Reis-Filho JS. Microarray-based gene expression profiling as a clinical tool for breast cancer management: are we there yet? Int J Surg Pathol 2009; 17: 285-302.

17. van de Vijver MJ, He YD, van't Veer LJ, et al. A gene-expression signature as a predictor of survival in breast cancer. N Engl J Med 2002; 347: 1999-2009.

18. Weigelt B, Mackay A, A'Hern R, et al. Breast cancer molecular profiling: a retrospective analysis of molecular subtype assignment using single sample predictors. Lancet Oncol 2010; 11: 339-349.

19. Foulkes WD, Smith IE, Reis-Filho JS. Triple-negative breast cancer. N Engl J Med 2010; 363: 1938-1948.

20. Fulford LG, Easton DF, Reis-Filho JS, et al. Specific morphological features predictive for the basal phenotype in grade 3 invasive ductal carcinoma of breast. Histopathology 2006; 49: 22-34.

21. Livasy CA, Karaca G, Nanda R, et al. Phenotypic evaluation of the basal-like subtype of invasive breast carcinoma. Mod Pathol 2006; 19: 264-271.

22. Reis-Filho JS, Westbury C, Pierga JY. The impact of expression profiling on prognostic and predictive testing in breast cancer. J Clin Pathol 2006; 59: 225-231.

23. Turner NC, Reis-Filho JS, Russell AM, et al. BRCA1 dysfunction in sporadic basal-like breast cancer. Oncogene 2007; 26: 2126-2132.

24. Rakha EA, Reis-Filho JS, Ellis IO. Basal-like breast cancer: a critical review. J Clin Oncol 2008; 26: 2568-2581.

25. Carey LA, Perou CM, Livasy CA, et al. Race, breast cancer subtypes, and survival in the Carolina Breast Cancer Study. JAMA 2006; 295: 2492-2502.

26. Nielsen TO, Parker JS, Leung S, et al. A comparison of PAM50 intrinsic subtyping with immunohistochemistry and clinical prognostic factors in tamoxifen-treated estrogen receptor-positive breast cancer. Clin Cancer Res 2010; 16: 5222-5232.

27. Reis-Filho JS, Weigelt B, Fumagalli D, et al. Molecular profiling: moving away from tumor philately. Sci Transl Med 2010; 2: 47ps43.

28. Pusztai L, Mazouni C, Anderson K, et al. Molecular classification of breast cancer: limitations and potential. Oncologist 2006; 11: 868-877.

29. Kapp AV, Tibshirani R. Are clusters found in one dataset present in another dataset? Biostatistics 2007; 8: 9-31.

30. Haibe-Kains B, Culhane A, Desmedt C, et al. Robustness of breast cancer molecular subtypes identification. Ann Oncol 2010; 21: iv49-iv59.

31. Chang HY, Nuyten DS, Sneddon JB, et al. Robustness, scalability, and integration of a wound-response gene expression signature in predicting breast cancer survival. Proc Natl Acad Sci U S A 2005; 102: 3738-3743.

32. Fan C, Oh DS, Wessels L, et al. Concordance among gene-expression-based predictors for breast cancer. N Engl J Med 2006; 355: 560-569.

33. Desmedt C, Haibe-Kains B, Wirapati P, et al. Biological processes associated with breast cancer clinical outcome depend on the molecular subtypes. Clin Cancer Res 2008; 14: 5158-5165.

34. Wirapati P, Sotiriou C, Kunkel S, et al. Meta-analysis of gene expression profiles in breast cancer: toward a unified understanding of breast cancer subtyping and prognosis signatures. Breast Cancer Res 2008; 10: R65.

13

35. de Ronde JJ, Hannemann J, Halfwerk H, et al. Concordance of clinical and molecular breast cancer subtyping in the context of preoperative chemotherapy response. Breast Cancer Res Treat 2010; 119: 119-126.

36. Lim E, Vaillant F, Wu D, et al. Aberrant luminal progenitors as the candidate target population for basal tumor development in BRCA1 mutation carriers. Nat Med 2009; 15: 907-913.

37. Molyneux G, Geyer FC, Magnay FA, et al. BRCA1 basal-like breast cancers originate from luminal epithelial progenitors and not from basal stem cells. Cell Stem Cell 2010; 7: 403-417.

38. Weigelt B, Reis-Filho JS. Molecular profiling currently offers no more than tumour morphology and basic immunohistochemistry. Breast Cancer Res 2010; 12 Suppl 4: S5.

39. Desmedt C, Piette F, Loi S, et al. Strong time dependence of the 76-gene prognostic signature for node-negative breast cancer patients in the TRANSBIG multicenter independent validation series. Clin Cancer Res 2007; 13: 3207-3214.

40. Wang Y, Klijn JG, Zhang Y, et al. Gene-expression profiles to predict distant metastasis of lymph-node-negative primary breast cancer. Lancet 2005; 365: 671-679.

41. Ivshina AV, George J, Senko O, et al. Genetic reclassification of histologic grade delineates new clinical subtypes of breast cancer. Cancer Res 2006; 66: 10292-10301.

42. Sotiriou C, Wirapati P, Loi S, et al. Gene expression profiling in breast cancer: understanding the molecular basis of histologic grade to improve prognosis. J Natl Cancer Inst 2006; 98: 262-272.

43. Ein-Dor L, Zuk O, Domany E. Thousands of samples are needed to generate a robust gene list for predicting outcome in cancer. Proc Natl Acad Sci U S A 2006; 103: 5923-5928.

44. Michiels S, Koscielny S, Hill C. Interpretation of microarray data in cancer. Br J Cancer 2007; 96: 1155-1158.

45. Ioannidis JP. Microarrays and molecular research: noise discovery? Lancet 2005; 365: 454-455.

46. Reyal F, van Vliet MH, Armstrong NJ, et al. A comprehensive analysis of prognostic signatures reveals the high predictive capacity of the proliferation, immune response and RNA splicing modules in breast cancer. Breast Cancer Res 2008; 10: R93.

47. Weigelt B, Horlings HM, Kreike B, et al. Refinement of breast cancer classification by molecular characterization of histological special types. J Pathol 2008; 216: 141-150.

48. Paik S, Shak S, Tang G, et al. A multigene assay to predict recurrence of tamoxifen-treated, node-negative breast cancer. N Engl J Med 2004; 351: 2817-2826.

49. Paik S, Tang G, Shak S, et al. Gene expression and benefit of chemotherapy in women with node-negative, estrogen receptor-positive breast cancer. J Clin Oncol 2006; 24: 3726-3734.

50. Kim C, Paik S. Gene-expression-based prognostic assays for breast cancer. Nat Rev Clin Oncol 2010; 7: 340-347.

51. Paik S. Development and clinical utility of a 21-gene recurrence score prognostic assay in patients with early breast cancer treated with tamoxifen. Oncologist 2007; 12: 631-635.

52. Habel LA, Shak S, Jacobs MK, et al. A population-based study of tumor gene expression and risk of breast cancer death among lymph node-negative patients. Breast Cancer Res 2006; 8: R25.

53. Goldstein LJ, Gray R, Badve S, et al. Prognostic utility of the 21-gene assay in hormone receptor-positive operable breast cancer compared with classical clinicopathologic features. J Clin Oncol 2008; 26: 4063-4071.

14

54. Dowsett M, Cuzick J, Wale C, et al. Prediction of risk of distant recurrence using the 21-gene recurrence score in node-negative and node-positive postmenopausal patients with breast cancer treated with anastrozole or tamoxifen: a TransATAC study. J Clin Oncol 2010; 28: 1829-1834.

55. Gianni L, Zambetti M, Clark K, et al. Gene expression profiles in paraffin-embedded core biopsy tissue predict response to chemotherapy in women with locally advanced breast cancer. J Clin Oncol 2005; 23: 7265-7277.

56. Albain KS, Barlow WE, Shak S, et al. Prognostic and predictive value of the 21-gene recurrence score assay in postmenopausal women with node-positive, oestrogen-receptor-positive breast cancer on chemotherapy: a retrospective analysis of a randomised trial. Lancet Oncol 2010; 11: 55-65.

57. Harris L, Fritsche H, Mennel R, et al. American Society of Clinical Oncology 2007 update of recommendations for the use of tumor markers in breast cancer. J Clin Oncol 2007; 25: 5287-5312.

58. Cuzick J, Dowsett M, Wale C, et al. Prognostic Value of a Combined ER, PgR, Ki67, HER2 Immunohistochemical (IHC4) Score and Comparison with the GHI Recurrence Score - Results from TransATAC. Cancer Res 2009; 69: 503S-503S (Abstract).

59. Biomarkers and surrogate endpoints: preferred definitions and conceptual framework. Clin Pharmacol Ther 2001; 69: 89-95.

60. Knauer M, Mook S, Rutgers EJ, et al. The predictive value of the 70-gene signature for adjuvant chemotherapy in early breast cancer. Breast Cancer Res Treat 2010; 120: 655-661.

61. Liedtke C, Hatzis C, Symmans WF, et al. Genomic grade index is associated with response to chemotherapy in patients with breast cancer. J Clin Oncol 2009; 27: 3185-3191.

62. Oakman C, Santarpia L, Di Leo A. Breast cancer assessment tools and optimizing adjuvant therapy. Nat Rev Clin Oncol 2010; 7: 725-732.

63. Symmans WF, Hatzis C, Sotiriou C, et al. Genomic index of sensitivity to endocrine therapy for breast cancer. J Clin Oncol 2010; 28: 4111-4119.

64. Natrajan R, Lambros MB, Rodriguez-Pinilla SM, et al. Tiling path genomic profiling of grade 3 invasive ductal breast cancers. Clin Cancer Res 2009; 15: 2711-2722.

65. Natrajan R, Weigelt B, Mackay A, et al. An integrative genomic and transcriptomic analysis reveals molecular pathways and networks regulated by copy number aberrations in basal-like, HER2 and luminal cancers. Breast Cancer Res Treat 2009: epub ahead of print.

66. Andre F, Job B, Dessen P, et al. Molecular characterization of breast cancer with high-resolution oligonucleotide comparative genomic hybridization array. Clin Cancer Res 2009; 15: 441-451.

67. Chin SF, Teschendorff AE, Marioni JC, et al. High-resolution aCGH and expression profiling identifies a novel genomic subtype of ER negative breast cancer. Genome Biol 2007; 8: R215.

68. Chin K, DeVries S, Fridlyand J, et al. Genomic and transcriptional aberrations linked to breast cancer pathophysiologies. Cancer Cell 2006; 10: 529-541.

69. Natrajan R, Lambros MB, Geyer FC, et al. Loss of 16q in high grade breast cancer is associated with estrogen receptor status: Evidence for progression in tumors with a luminal phenotype? Genes Chromosomes Cancer 2009; 48: 351-365.

70. Lopez-Garcia MA, Geyer FC, Lacroix-Triki M, et al. Breast cancer precursors revisited: molecular features and progression pathways. Histopathology 2010; 57: 171-192.

71. Flagiello D, Gerbault-Seureau M, Sastre-Garau X, et al. Highly recurrent der(1;16)(q10;p10) and other 16q arm alterations in lobular breast cancer. Genes Chromosomes Cancer 1998; 23: 300-306.

15

72. Tsuda H, Takarabe T, Fukutomi T, et al. der(16)t(1;16)/der(1;16) in breast cancer detected by fluorescence in situ hybridization is an indicator of better patient prognosis. Genes Chromosomes Cancer 1999; 24: 72-77.

73. Simpson PT, Reis-Filho JS, Gale T, et al. Molecular evolution of breast cancer. J Pathol 2005; 205: 248-254.

74. Bergamaschi A, Kim YH, Wang P, et al. Distinct patterns of DNA copy number alteration are associated with different clinicopathological features and gene-expression subtypes of breast cancer. Genes Chromosomes Cancer 2006; 45: 1033-1040.

75. Stephens PJ, McBride DJ, Lin ML, et al. Complex landscapes of somatic rearrangement in human breast cancer genomes. Nature 2009; 462: 1005-1010.

76. Stemke-Hale K, Gonzalez-Angulo AM, Lluch A, et al. An integrative genomic and proteomic analysis of PIK3CA, PTEN, and AKT mutations in breast cancer. Cancer Res 2008; 68: 6084-6091.

77. Janku F, Tsimberidou AM, Garrido-Laguna I, et al. PIK3CA Mutations in Patients with Advanced Cancers Treated with PI3K/AKT/mTOR Axis Inhibitor. Mol Cancer Ther 2011: Epub ahead of print.