uPA: From Pilot Studies to Recommendation for Clinical Use

Professor Joe DuffySt Vincent’s University Hospital,

Dublin and University College DublinDublin and University College Dublin

M t I t t Q ti AftMost Important Questions After a Diagnosis of Breast Cancerg

• How “bad” or aggressive is the tumor

• Will the tumor recur

I dj t h th ?• Is adjuvant chemotherapy necessary?

Node Negative Breast Cancer

• With screening, 2/3 of newly-diagnosed breast cancer patients are node-negative

• 70% of node-negative patients cured by70% of node negative patients cured by surgery + radiotherapy

• 30% develop recurrent disease by 10 yrProblem: How to differentiate those withProblem: How to differentiate those with

aggressive from those with indolent ca

Node Negative Breast Cancer:Node-Negative Breast Cancer: Treatment DilemmaTreatment Dilemma

• Circumventing the Dilemma: Treat almost all node-negative patients with adjuvant chemotherapyj py

• Problem: Only a minority of patients will benefit but most will suffer toxic side-effects

C O O S C CCHEMOTHERAPY FOR BREAST CANCER: OVERVIEW OF RANDOMISED TRIALS

• 18,000 women• 47 trials of chemotherapy vs no chemoChange in 10 yr survival (node negativeChange in 10-yr survival (node-negative

patients)• <50 yr: 71% 78%

• 50-69 yr: 67% 69%• 50-69 yr: 67% 69%Lancet 1998;352:930

ADJUVANT CHEMOTHERAPY FOR BREASTADJUVANT CHEMOTHERAPY FOR BREAST CANCER: OVERVIEW OF RANDOMISED

TRIALSTRIALS

• 145,000 women194 i l f h h h• 194 trials of chemotherapy vs no chemo

Absolute Improvement in Mortality (%)5 yr 10 yr 15 yr

Women <50 yr 4.7 7.9 10Women 50-69 yr 2.6 2.9 3.0

Lancet 2005;365:1687Lancet 2005;365:1687

Side Effects of ChemotherapySide Effects of Chemotherapy (CMF)(CMF)

Side effect % affected• Nausea 43• Vomiting 42• Alopecia 40• Ovarian failure 70• Weight gain 12• Diarrhea 4.5

Shapiro & Recht, N Eng J Med 2001;344:1997

QUESTION

Should most node-negative breast gcancer patients be treated with h th th t llchemotherapy so that a small

minority benefit while a largeminority benefit while a large proportion suffer from adverse toxic effects ?

Rational Way Forward

Develop and validate markers that will reliably differentiate between patients with aggressive and indolent diseasegg

Metastasis is the main causes of mortality in cancer Since uPA is causally involvedin cancer. Since uPA is causally involved in this processes, it should be a strong marker of metastatic potential and thus of prognosis p g

uPA• A 53 kDa serine protease• It degrades basement membranes

d ECMand ECM• Multiple studies with animal modelsMultiple studies with animal models

suggest that uPA is causally involved i iin metastasis

Invasion & MetastasisInvasion & MetastasisInvasiveInvasiveCarcinomaCarcinoma

Normal EpitheliumNormal Epithelium

1

pp

Stromal cellsStromal cells& Matrix& Matrix

22

3

Distant SiteDistant Site

Blood VesselBlood Vessel

4

5PROTEASES

5

MetastaticMetastatic

ADHESION MOLECULESANGIOGENIC FACTORSMOTILITY FACTORS

LesionLesion

uPA vs DFIuPA vs DFI

1

0.8

0.9

1su

rviv

al uPA <= 0.81 (n=142)uPA > 0.81 (n=42)

0.6

0.7

se-fr

ee s

0 3

0.4

0.5

of d

isea

0.1

0.2

0.3

babi

lity

P=0.0007, RR=2.4

00 2 4 6 8 10 12

Ti ( )

Pro

Time (years)

uPA vs overall survival

1

0.8

0.9

1rv

ival

uPA <= 0.81 (n=142)uPA > 0.81 (n=42)

0.6

0.7

0.8

vera

ll su

r

0.4

0.5

lity

of o

v

0 1

0.2

0.3

Prob

abil

P=0.0009; RR=2.5

0

0.1

0 2 4 6 8 10 12 14

P

Time (years)

uPA vs OS

Univariate Multivariatep p

LN <0 0001 <0 0001LN <0.0001 <0.0001uPA 0.0004 0.0003Size 0.003 0.006ER 0 020 NSER 0.020 NS

uPA vs DFI; NODE-NEGATIVE PATIENTS

1al

0.80.9

1e

surv

iva

0 50.60.7

ease

-free

0.30.40.5

y of

dis

e

PA 0 81 ( 72)

00.10.2

roba

bilit

y uPA <= 0.81 (n=72)uPA > 0.81 (n=15)

P=0003; RR=4.41

Time (years)

00 2 4 6 8 10 12

Pr

uPA vs overall survival; NODE NEGATIVE PATIENTS

0 80.9

1su

rviv

al

0.60.70.8

over

all s

0 30.40.5

bilit

y of

o

PA 0 81 ( 72)

0.10.20.3

Prob

ab uPA <= 0.81 (n=72)uPA > 0.81 (n=15)

P=0.0004; RR=5.7

00 2 4 6 8 10 12

Time (years)Time (years)

PAI-1: Inhibitors of uPA

• If uPA is involved in metastasis and predicts poor patient

t hi h l l f PAI 1outcome, high levels of PAI-1 might be expected to blockmight be expected to block metastasis and thus predict good outcome

PAI-1 vs DFI

0 8

0.9

1su

rviv

al PAI-1<=0.647 (n= 104)PAI-1>0.647 (n= 80)

0.6

0.7

0.8

se-fr

ee s

0.4

0.5

of d

isea

s

0 1

0.2

0.3

babi

lity

o

P=0.0003; RR=2.37

0

0.1

0 2 4 6 8 10 12

Prob

Time (years)

PAI-1 vs overall survival

1 PAI-1<=0.647 (n= 104)

0.8

0.9

1ur

viva

l PAI 1< 0.647 (n 104)PAI-1>0.647 (n= 80)

0.6

0.7

vera

ll su

0 3

0.4

0.5

ity o

f ov

0.1

0.2

0.3

roba

bili

P <0.0001; RR=2.96

00 2 4 6 8 10 12 14

Pr

Time (years)

PA d PAI 1 P ti M kuPA and PAI-1 as Prognostic Markers in Breast Cancer

• Prognostic impact of both confirmed by > 20 i d d t20 independent groups

• Prognostic value independent of g ptraditional factors

• Prognostic value stronger than HER2,Prognostic value stronger than HER2, EGFR, CB, CD, p53

• Prognostic in patients with both N+ and• Prognostic in patients with both N+ and N- disease

uPA and PAI 1: Transfer touPA and PAI-1: Transfer to ClinicClinic

• Clinical value must be validated in a ld t d d l l I idgold-standard or level I evidence

studyy

LOE for Grading Clinical utilityLOE for Grading Clinical utility of Tumor Markers

• I. High powered randomized prospective t i l t / l d l itrial or meta/pooled analysis

• II. Therapeutic trial to test therapeutic p phypothesis but not marker utility

• III. Large retrospective studyIII. Large retrospective study• IV. Small retrospective study

V Pil t t d• V. Pilot studyHayes et al JNCI 1996;88:1456y ;

Pooled Analysis of uPA/PAI 1:Pooled Analysis of uPA/PAI-1: EORTC Studyy

• Raw data: 18 data sets, 8377 patients• Published, 11; unpublished 7• Median follow-up: 79 months• 35% of patients relapsed• 35% of patients relapsed• 27% had diedLook et al. JNCI 2002;94:116

Participating laboratories (9 t i )(9 countries)

uPA/PAI-1 Pooled Analysis

PIPIM Look, Rotterdam

STEERING COMMITTEEN Harbeck, Munich,

K Ul M i hK Ulm, Munich,MJ Duffy, Dublin (Chairman)y, ( )

Pooled Analysis: Summary

• Both uPA and PAI-1 were independent i fprognostic factors

• uPA/PAI-1 ranked second to nodal status but stronger than size, grade, HR, age

• uPA/PAI-1 prognostic in N- and N+ patientsp g p• uPA/PAI-1 prognostic in untreated N- patients• uPA/PAI 1 together better than either alone• uPA/PAI-1 together better than either alone

Look et al. JNCI 2002;94:116

Prospective Randomized German MulticenterTherapy Trial in Node-negative Breast Cancer

ON uPA and PAI-1 low

Node-negativeC C

ATIO

Adjuvant

Observation

Breast Cancer

T > 1 cm and < 5 cm10 LN l d AT

IFIC

ATIO

N Adjuvant Chemo(6 x CMF)

10 LN analyzedpre- / postmenopausal

STR

A

OM

ISA

uPA and / or PAI-1 high

RA

ND

O

Observation

R

uPA/PAI 1: Prognostic ValueuPA/PAI-1: Prognostic Value in Breast Cancerin Breast Cancer

Prognostic value for N- patients confirmed in 2 different LOE1 studies

• Randomised prospective trial (556)Randomised prospective trial (556)• Pooled analysis (n=8377)

Janicke et al JNCI 2001;93:913Janicke et al. JNCI 2001;93:913Look et al JNCI 2002;94:116

PA d PAI 1 ASCOuPA and PAI-1: ASCO Recommendation for Clinical Use

• uPA and PAI-1 may be used for the d i i f i i i i hdetermination of prognosis in patients with newly diagnosed, node negative breast cancer.

• Low levels of both markers are associated with a sufficiently low risk of recurrence, especially in HR–positive women who will receive adjuvant endocrine therapy, that chemotherapy will only contribute minimal additional benefit.

ASCO Recommendations: 2007

• ER/PR• HER-2

• uPA/PAI 1• uPA/PAI-1• Oncotype DX

H i t l J Cli O l 2007 25 5287Harris et al. J Clin Oncol 2007;25:5287

Oth E t P l R diOther Expert Panels Recommending uPA/PAI-1 for Clinical Use

• American Society of Clinical Oncology• National Academy of Clinical

Biochemistry (US)Biochemistry (US)• European Group on Tumor Markers

(EGTM)• German Gynecological SocietyGerman Gynecological Society

PA d PAI I f D t i i P iuPA and PAI-I for Determining Prognosis in Breast Cancer: NACB Recommendation

The NACB Panel states that testing for uPA and PAI 1 b i d id if l hPAI-1 may be carried out to identify lymph node-negative breast cancer patients that do

t d lik l t b fit fnot need or are unlikely to benefit from adjuvant chemotherapy. Measurement of both

t i h ld b f d thproteins should be performed as the information provided by the combination is

i t th t f ith lsuperior to that from either alone.Sturgeon et al, Clin Chem 2008;54:e11-e79

PA d PAI I f D t i i P iuPA and PAI-I for Determining Prognosis in Breast Cancer: NACB Recommendation

Lymph node-negative breast cancer patients with l l l f b h PA d PAI 1 h llow levels of both uPA and PAI-1 have a low risk of disease relapse and thus may be spared f th t i id ff t d t f dj tfrom the toxic side effects and costs of adjuvant chemotherapy. Lymph node-negative women

ith hi h l l f ith PA PAI 1 h ldwith high levels of either uPA or PAI-1 should be treated with adjuvant chemotherapy.

Sturgeon et al, Clin Chem 2008;54:e11-e79

Acknowledgements

• HRB• Irish Cancer Society

• European Union• European Union

All the researchers, surgeons, medical oncologists histopathologists and patientsoncologists, histopathologists and patients who helped with this work