PAM50/PROSIGNA®€¦ · Prat et al. BCRT 2012; updated analysis unpublished pCR to sequential anthracycline and paclitaxel/exabepilone-based regimens (N=957) PEPI of 0 (best prognostic

PAM50/PROSIGNA®

Aleix Prat, MD PhD

Medical Oncology Dept.,

Hospital Clínic, Univ. of Barcelona

Disclosures / COI

• Uncompensated advisory role for Nanostring Technologies

Predicting…

Biology

Prognosis Treatment benefit

Predicting…

Biology

Prognosis Treatment benefit

Luminal A

Normal Breast Basal-like Luminal B

Claudin-low HER2-enriched Intrinsic Subtypes Perou et al., Nature 2000

Sorlie et al., PNAS 2001

Sorlie et al., PNAS 2003

Hu et al., BMC Genomics 2006

Herschkowitz et al., GB 2007

Cheang et al. JNCI 2008

Parker et al., JCO, Feb 2009

Prat et al., BCR 2010

Nielsen et al., CCR 2010

Cheang et al., CCR 2012

Prat et al. Ann Oncol 2012

Chia et al., CCR 2012

Prat et al. BCRT 2012

TCGA Nature 2012 / Cell 2014

Prat et al. JCO 2013

Martín et al. BRCT 2013

Prat et al. Oncologist 2013

Prat et al. CCR 2014

Gnant et al. Annals Oncol 2014

Rel

apse

Fre

e-S

urv

ival

Hoadley et al. Cell 2014 BASAL-LIKE

2013 St Gallen International Expert Consensus

Quote: “The Panel recognized the superior accuracy and reproducibility of multi-gene molecular assays”.

Goldhirsch et al. Ann Oncol 2013 Cheang et al. JNCI 2009

Endocrine Therapy (chemo in selected cases)

Endocrine + Chemo (most)

Endocrine+ Chemo + anti-HER2

Chemo + anti-HER2

Chemo

14-20%

Prospective study of the impact of the PAM50 assay on adjuvant clinical decision-making in an unselected

population of women with ER+, HER2-negative, node-negative breast cancer: a GEICAM study

Martín et al. Curr Med Res Opin 2015

9 (60%) ROR-low and 6 (40%) ROR-interm

2 (50%) ROR-interm and 2 (50%) ROR-high

Concordancia = 42/64 = 65.6%

Translational Genomics Group

FDA 510(k)

Local IHC vs.

Basal Her2E LumA LumB

N 0 0 22 4 26

% 0% 0% 85% 15% 100%

N 2 1 15 20 38

% 5% 3% 39% 53% 100%

N 2 1 37 24 64

% 3% 2% 58% 38% 100%Total

LumB

LumA

IHC

TotalPROSIGNA

Analytical reproducibility of the PAM50 breast cancer intrinsic subtyping test and nCounter® analysis system using formalin-fixed

paraffin-embedded (FFPE) breast tumor specimens

Reproducibility from Tissue • N=43 FFPE / 3 testing sites • Intrinsic subtype: 97% concordance

• For ROR, 90% concordance

• ROR SD = 2.9

Precision from RNA • N=43 RNAs / 3 testing sites • Intrinsic subtype, 100% concordance

• For ROR, 100% concordance

• Site-to-site or operator-to-operator <1% of variance

• ROR SD = 0.67

Torsten et al. BMC Genomics 2014

GEICAM Study: 190 samples SUBTYPE CONCORDANCE (95%)

(Kappa = 0.89)

Predicting…

Biology

Prognosis Treatment benefit

Tumor size

Nodal status

Tumor size

Nodal status

PAM50 Generates an Risk Of Relapse (ROR) Score Specific to Each Patient

• Intrinsic subtype + Proliferation score + Tumor size. • “Training cohort”:

– Patients with tumors representing all subtypes. – Pre (30%) and post-menopausal patients. – None received adjuvant systemic therapy.

Parker et al. JCO 2009; Nielsen et al. CCR 2010

Patient expression

profile

PAM50 centroids

Pearson’s correlation to centroids

Proliferation score Tumor size

ATAC study

Postmenopausal

women with invasive

BC (N=9366)

Tam alone

(N=3116)

Arimidex alone

(N=3125)

Tam + Arimidex

(N = 3125)

TransATAC study

(N=1125 blocks)

ABCSG-8 study

Post-menopausal

women with HR+

BC (N=3714)

Tamoxifen

(N=1849)

Anastrozole

(N=1865)

ABCSG-8 study

(N = 1478 blocks)

Tam

2-years 3 years

PAM50 (PROSIGNA®) Validation Studies

JNCI 2009

Patient Report

Paciente Postmenopáusica PS 0 Tumor de 1.6 cm. RE 90%, RP 40%, Ki67 20%, HER2-negativo pT1cN0M0 LumA LumB

JNCI 2013

• The ROR score was the strongest molecular prognostic factor in the late follow-up period, whereas IHC4 and OncotypeDX RS were only weakly prognostic in this period.

Ivana Sestak

Jack Cuzick, Mitch Dowsett, Martin Filipits, Peter Dubsky, J. Wayne Cowens,

Sean Ferree, Carl Schaper, Christian Fesl, Michael Gnant

Centre for Cancer Prevention, Wolfson Institute of Prevention Medicine, Queen Mary University, London, UK

Academic Department of Biochemistry, Royal Marsden Hospital, London, UK

Austrian Breast and Colorectal Cancer Study Group, Vienna, Austria

Department of Surgery and Comprehensive Cancer Centre, Medical University of Vienna, Vienna, Austria

NanoString Technologies, Seattle, WA, USA

MyRAQA, Redwood Shores, CA, USA

Prediction of late distant

recurrence after 5 years of

endocrine treatment: A combined analysis of 2137 patients from the ABCSG-8

and transATAC studies using the PAM50 Risk of

Recurrence (ROR) score

Risk groups – ROR score HR (95% CI)

Low (N=1183 (55.4%)) -

Intermediate (N=538 (25.2%)) 3.26 (2.07-5.13)

High (N=416 (19.5%)) 6.90 (4.54-10.47)

2.4%

16.6%

8.3%

0

5

10

1

5

20

5 6 7 8 9 10 Follow-up time [years]

Low

Intermediate

High

Dis

tant

recurr

ence (

%)

0

5

10

1

5

20

Sestak et al. SABCS 2013 and JCO 2015

PAM50 in First-line HR+/HER2-neg Metastatic Breast Cancer

PFS following Letrozole monotherapy (n=644)

unpublished data

0 10 20 30 40

0.0

0.2

0.4

0.6

0.8

1.0

Pro

ba

bili

ty o

f Eve

nt

high

low

med Log Rank p=1.08e-05

0 10 20 30 40

0.0

0.2

0.4

0.6

0.8

1.0

Pro

ba

bility o

f E

ve

nt

LumA

LumB

Basal

Her2 Log Rank p=1.1e-07

SUBTYPE ROR

months months

LumA

LumB

Basal

Her2

Independent of age, visceral metastasis, prior endocrine therapy, performance status

Predicting…

Biology

Prognosis Treatment benefit

Multivariable Analysis (n=957) Multivariable Analysis (n=508) Variables N pCR rate OR Lower 95% Upper 95% p-value OR Lower 95% Upper 95% p-value

Age - - 1.0 0.97 1.00 0.178 1.00 0.97 1.02 0.749

Tumor size T0-T2 584 24% 1.0 - - - 1.0 - - - T3-T4 369 12% 0.6 0.50 0.84 0.001 0.5 0.30 0.81 0.005

ER IHC Positive 507 11% 1.0 - - - 1.0 - - - Negative 442 37% 1.8 0.99 3.34 0.052 1.3 0.61 2.58 0.533

PR IHC Positive 408 12% 1.0 - - - 1.0 - - - Negative 539 33% 1.1 0.65 1.89 0.716 0.8 0.40 1.51 0.463

HER2 STATUS Negative 841 22% 1.0 - - - 1.0 - - - Positive 93 35% 1.3 0.64 2.51 0.492 2.1 0.90 4.88 0.087

Grade 1 50 7% - - - - 1.0 - - - 2 285 13% - - - - 0.70 0.17 2.85 0.62

3 306 32% - - - - 1.51 0.35 6.50 0.58

Nodal status N0 157 18% - - - - 1.0 - - - N1 244 21% - - - - 0.70 0.17 2.85 0.62

N2 66 19% - - - - 1.51 0.35 6.50 0.58

N3 41 26% - - - - 1.50 0.37 6.40 0.17

PAM50 Luminal A 293 6% 1.0 - - - 1.0 - - - Luminal B 174 16% 3.3 1.72 6.44 <0.001 4.5 1.71 11.75 0.002

Basal-like 313 37% 6.1 2.94 12.66 <0.001 8.3 2.74 24.85 <0.001

HER2-E 99 38% 6.1 2.75 13.38 <0.001 13.2 4.50 38.50 <0.001

Prat et al. BCRT 2012; updated analysis unpublished

pCR to sequential anthracycline and paclitaxel/exabepilone-based regimens (N=957)

PEPI of 0 (best prognostic group) was highest in the LumA vs. LumB (27.1% v 10.7%; P = 0.004)

J Clin Oncol 2011

LumA LumB HER2E

LumA LumB

HER2E Basal-like

Association between breast cancer subtypes and response to neoadjuvant anastrozole

Steroids 2011

Anita K. Dunbier, Helen Andersona, Zara Ghazoui, Janine Salter, Joel S. Parker, Charles M. Perou, Ian E. Smith, Mitch Dowsett

The Oncologist 2013;18:123–133

Take-home messages

• Intrinsic biology (Luminal A and B, HER2-enriched and Basal-like)

captures the vast majority of the biological diversity in breast cancer.

• However, current IHC-based definitions of intrinsic subtypes are

suboptimal and the gold-standard is, in my opinion, PAM50.

• The assay is reproducible across labs and works in FFPE core biopsies

(i.e. primary tumors and metastatic tissues).

• Intrinsic subtyping and ROR, together with tumor size and nodal status,

can help identify patients:

• That do not need adjuvant polychemotherapy due to their low risk

of relapsing distantly (<10% at 10 years).

• That do not need adjuvant endocrine therapy beyond 5 years due

to their low risk of relapsing between period years 5-10 (<5% at 10

years).

• In HR+/HER2-negative disease, intrinsic subtyping at diagnosis before

neoadjuvant therapy or in first-line metastatic setting, can help decide

whether to begin with endocrine therapy or chemotherapy.

• In metastatic TNBC, distinguishing Basal-like vs. not can help predict

benefit from docetaxel vs. carboplatin.

University of North Carolina, NC, USA

Chuck Perou Lisa Carey Joel S. Parker

GRANT SEOM PARA FORMACION EN INVESTIGACION TRASLACIONAL

GEICAM, Spain

Miguel Martín Eva Carrasco Rosalía Caballero Maribel Casas

Acknowledgements

Patricia Galván Maria Vidal Ana Vivancos Javier Cortés Josep Tabernero Vicente Peg Santiago Ramon y Cajal

CAREER CATALYST GRANT FROM SUSAN G KOMEN FOUNDATION

SOLTI, Spain

Eva Ciruelos Lorena de la Peña Josep Vazquez José Baselga

Barbara Adamo Montse Muñoz Imma Alonso Blanca Farrús Pedro Fernández