[PPT] - Integrating mol Integrating mol ecular Profiling ecular Profiling PowerPoint Presentation

SLIDE 1

Integrating mol Integrating mol Into Patient Se Treatm N S ll C l Non-Small Cel

Frances A. Sheph Scott Taylor Chair in L Scott Taylor Chair in L

Princess Mar

P f f M di i Professor of Medicine

ecular Profiling ecular Profiling election for the ment of ll L C ll Lung Cancer

herd, MD FRCPC Lung Cancer Research Lung Cancer Research

rgaret Hospital U i i f T

e, University of Toronto

SLIDE 2

Predictive and Pr

A prognostic facto

patient or tumour c p identifies a better o

f treatment
A predictive factor

patient or tumour c patient or tumour c identifies a better o treatment (respons treatment (respons

rognostic Factors

r is a baseline

characteristic that

utcome regardless

is a baseline characteristic that characteristic that

utcome from

se or survival) se or survival)

SLIDE 3

In Other In Other That is to

A prognostic ma

the effect of the the effect of the patient A di ti

A predictive mar

the effect of the tumour

r Words r Words,

Say…….

arker determines tumour on the tumour on the k d t i rker determines treatment on the

SLIDE 4

Exam

Untreated patients with PS
Untreated patients with PS

PS i

ti

PS is prognostic
Treated patients with PS 0

compared to 6 months for compared to 6 months for

Treated patients with PS 2

compared to 3 months for

PS is NOT predictive of a

treatment as the relative same in both PS 0/1 and P same in both PS 0/1 and P

absolute benefit is less

mple

S 0/1 survive 6 months S 2/3 survive 3 months 0/1 survive 8 months control control 2/3 survive 4 months control a better outcome from

benefit (i.e. the HR) is the

PS 2/3 patients although the PS 2/3 patients although the

SLIDE 5

Predictive

It is possible to ide

factors for respons p trial

However, predictive

, p benefit can only be untreated control g

This is because sup

single arm trial may ti f t prognostic factors effect of therapy

Factors

ntify predictive se from a single arm g e factors for survival determined with an roup perior survival in a y be due to inherent d t d t th and not due to the

SLIDE 6

When Using Su When Using Su Bewa

Implication is that the

give a better estimate g will benefit than the ov

Two ways this could h

Two ways this could h

– Subgroup does so wel treatment may not be n treatment may not be n – Subgroup is biological differently to treatment d e e t y to t eat e t

bgroup Results bgroup Results are!!!!

results in the subgroup

f whether the patient

p verall result happen happen

l without treatment that needed - prognostic factor needed prognostic factor lly distinct and responds t – predictive factor t p ed ct e acto

SLIDE 7

Differential Treat

Saying that a treatmen

ti l b a particular subgroup saying there is a treatm “interaction” interaction

A patient or tumour ch

identifies a subgroup w identifies a subgroup w effect is different from a “predictive” factor a predictive factor

In statistical terms, this

a treatment by baseline

tment Response p

nt “works” differently in i i t ti ti l t is, in statistical terms, ment by subgroup haracteristic that within which a treatment within which a treatment m that in other patients is

s is the same as saying there is covariate interaction

SLIDE 8

Intera Intera

Quantitative interaction

Quantitative interaction effect is in the same dir magnitude (scale used groups groups

– Considered less importa is beneficial in both sub – EGFR mutation status

Qualitative interaction:

effect is in opposite dir effect is in opposite dir groups

– Obviously very importan – ERCC1 and p53 protein

action action

n: the relative treatment n: the relative treatment rection, but of different is important) in the two

ant since the treatment effect groups

the relative treatment rections in the two rections in the two

nt when present

SLIDE 9

Testing for th Testing for th Significance of

When we are testing fo

interaction we are askin rates different from eac

This test is less powerf

iti ) b positive) because:

There are two sources of
The numbers in the group

The numbers in the group

Always look for

value value

he Statistical he Statistical f an Interaction

r the significance of an ng: “are the two hazard ch other?” ful (less likely to be

variation ps being compared are smaller ps being compared are smaller

the interaction p

SLIDE 10

Testing for th Testing for th Significance of

Saying the test for intera

another way of saying th treatment effects within arise by chance Thi bl i t h l

This problem is not help

worse) by doing multiple within subgroups within subgroups

Always look for interacti

evaluating predictive ma evaluating predictive ma

he Statistical he Statistical f an Interaction

action is not powerful is hat apparent variations in subgroups can easily d (i f t it i d ed (in fact it is made e tests for significance

n p values in studies

arkers of survival!!! arkers of survival!!!

SLIDE 11

Treatment Factors in

Clinical selection parameters

S

Sex Age PS PS

Future investigations Future investigations

Histology Folate pathway markers

n Advanced NSCLC

Biologic ma ke s Biologic markers

ERCC-1 RRM 1 RRM-1 Ras mutation Beta-tubulin

eta tubu

EGFR p53 BRCA1 TS Micro-array Micro array

profiling

SLIDE 12

ERCC1 Ex

Excision Re C l Complemen

xpression

epair Cross ti 1 G nting 1 Gene

SLIDE 13

A B

Mechanism

f Repair
f Repair
f Platinum

Adducts

C

Adducts

D E

Adduct formation Damage recognition recognition Assembly of the nucleotide excision repair complex Dual incision Damage excision Repair synthesis DNA ligation

Gazdar AF. NEJM 2 0 0 7 ;3 5 6 ( 8 ) :7 7 1 -3 .

Repaired DNA

LKP5

SLIDE 14

Slide 13 LKP5 Some words on figure look blurry. Can this be fixed, or does it project fine?

Lori K. Pender, 7/12/2007

SLIDE 15

Low ERCC 1 Express Prolonged Survival afte Prolonged Survival afte Gemcitabine & Cisplat

1.0 .8 al .6 ive Surviva

E

.4 .2 Cum ulati 0.0 20 40 60 80

E

20 40 60 80 Overall Survival ( w ee

sion Correlates with er Chemotherapy with er Chemotherapy with in in Advanced NSCLC

P=0.009 Log rank test

ERCC1 mRNA < median value (6.7)

100 120

ERCC1 mRNA > median value (6.7)

Lord et al. Clin Cancer Res 2 0 0 2 ;8 :2 2 8 6 .

100 120 eks)

SLIDE 16

ERCC1 in A Double-Edg

1 Survival Distribution Function

ERCCI > 50

0.75 0.50

ERCCI > 50

(94.6 months)

0.25

ERCCI < 50

(35.5 months)

P=0.01

20 40 60 80 100 120 140 Survival ( m onths)

P=0.01

NSCLC: ged Sword?

ERCC1 i i l l ERCC1 expression is clearly linked to platinum-resistance At the same time it is a At the same time, it is a favorable prognostic factor in untreated patients with early stage NSCLC early stage NSCLC

High expression associated with better survival with better survival Possibly due to its role in cancer susceptibility

Simon et al. Chest 2005; 127: 978-83.

SLIDE 17

Prognostic Effect of Prognostic Effect of

International Adjuvant In the control group, ER h f b

Hazard Ratio

have a favourab

Control Group ERCC1 +ve ERCC1 -ve 1 0 66 ERCC1 -ve 0.66 Chemotherapy ERCC1 +ve 1 ERCC1 -ve 1.16 All Patients ERCC1 +ve 1 ERCC1 ve ERCC1 -ve 1 0.88

f f ERCC1 Status in the

Lung Cancer Trial (IALT) RCC1 positive patients bl i !!

95% CI P Value

ble prognosis!!

0 40-0 90 0.009 0.40-0.90 0.34 0.86-1.56 0.26 0.71-0.10 0.26

Olaussen et al. NEJM 2 0 0 6 ;3 5 5 :9 8 3 -9 1 .

SLIDE 18

Overall Survival by E International Adjuva International Adjuva

(IAL

Patients with ERCC1-Negative Tumors

100 80 60

vival ( % )

Chemotherapy

(105 deaths)

60 40 20

Overall Surv

Control (113 deaths) 1 2 3 4 5

Years

Adjusted HR= 0 .6 5 P= 0 0 0 2

Patients with ERCC1- resected NSCLC benefit from

P= 0 .0 0 2

Olaussen e

but those with ERCC1-

Interactio

ERCC1 Status in the ant Lung Cancer Trial

Patients with ERCC1-Positive Tumors

ant Lung Cancer Trial

LT)

100 80 60

vival ( % )

Control (80 deaths) 60 40 20

Overall Surv

Chemotherapy

(92 deaths)

1 2 3 4 5

Years

Adjusted HR= 1 .1 4 P= 0 .4 0

negative, completely

m adjuvant cisplatin-based CT

et al. NEJM 2 0 0 6 ;3 5 5 :9 8 3 -9 1 .

positive tumors do not

n p=0.009

SLIDE 19

ITACA Trial in Ea ITACA Trial in Ea ITACA Trial in Ea ITACA Trial in Ea

TS TS High High High High TS TS Low Low ERCC 1 ERCC 1 Low Low High High Low Low TS TS Low Low

High/Low ERCC1 & TS selected accordin historical series ; * Control arm – Investig High/Low ERCC1 & TS selected accordin historical series ; * Control arm – Investig

arly Stage NSCLC arly Stage NSCLC arly Stage NSCLC arly Stage NSCLC

Taxane Taxane Profile 4 Profile 4 Taxane Taxane

Control* Control*

Profile 3 Profile 3 Pemetrexed Pemetrexed

Control* Control*

Profile 2 Profile 2 Cis/Gem Cis/Gem

Control* Control*

Profile 1 Profile 1 Cis/Pem Cis/Pem

Control* Control* Control Control

ng to median level of mRNA expression in gator choice of a cisplatin-based doublet ng to median level of mRNA expression in gator choice of a cisplatin-based doublet

SLIDE 20

RRM1 Ex

Ribonucleotid S b Subun

xpression

de Reductase it M1 nit M1

SLIDE 21

RRM1 in RRM1 in

Essential to nucleotide exc High levels linked to platin resistance in advanced NSC High expression is a favora stage disease

Correlates with ERCC1 exp

advanced2 NSCLC

OS significantly longer with OS significantly longer with (>120 months) relative to lo

in surgically treated patien

NSCLC NSCLC

ision repair (NER) p ( ) um- and gemcitabine- CLC able prognostic factor in early

ression in early stage1 and h high RRM1 expression h high RRM1 expression

w expression (60 months)

nts (p=0.02)1

1Zheng et al. NEJM 2007; 356: 800-7. 2Simon et al. JCO 2007; 25: 2741-6.

KP27

SLIDE 22

Slide 20 LKP27 slide updated, replace previous version

Lori K. Pender, 7/25/2007

SLIDE 23

RRM1 Expressio NSCLC: Impac NSCLC: Impac

80 100 %)

High leve f RRM1

60 urvival (%

f RRM1

(>40.5)

20 40 Overall S

P=0.02 Low

f

(<

24 48 72 96 120

( *187 patients with completely resected NSC underwent surgical resection between 1991

n in Early-stage

ct on Survival* ct on Survival*

el 1 ) w level RRM1 <40.5)

Months

)

Zheng et al. NEJM 2 0 0 7 ;3 5 6 :8 0 0 -7 .

CLC who 1 and 2001

SLIDE 24

RRM1 Exp Early-stag

RRM1 level has been

response to gemcitab therapy in advanced

RRM1 is prognostic i
RRM1 may not add to
There are no studies

predictive value of RR

perative adjuvant ch

ression in ge NSCLC

shown to predict bine and cisplatin NSCLC n early stage NSCLC

ERCC1

to determine the RM1 in the post- hemotherapy setting

SLIDE 25

P5

Mutation Protein ex

53

n status xpression

SLIDE 26

P53 St

Mutation or loss of p Mutation or loss of p result in:

Genomic instability progression and a patient patient Defective stress se and resistance (ins and resistance (ins to anti-growth / a

f chemotherapy

tatus

p53 gene may p53 gene may

y leading to tumor poorer prognosis for ensing mechanisms sensitivity) sensitivity) poptotic effects

SLIDE 27

Meta-Analysis of P53 Status as a Pro Status as a Pro

HR 1.25, 95% CI 1.09-1.43

Chen [21] Dosaka-akita [25] Ebina [26] Esposito [27] Fujino [32] Geradts [34] Hayakawa [41] Ishida [45] Konishi [52] Langendijk [55] Langendijk [55] Laudanski [56] Lee [58] Nishio [71] Ohsaki [74] Pastorino [76]

4.5 6.0 1.5 3.0 Hazard ratio

Pastorino [76] Tanaka [85] Overall

IHC: DO7 Antibody (n=16)

3 Protein & Mutation

gnostic Marker
gnostic Marker

HR 1.65, 95% CI 1.35-2.00

Carbone [20] Fukuyama [33] Greatens [36] H i [42] Horio [42] Huang [43] Kandioler [46] Kondo [51] Mitsudomi [65] [ ] Murakami [70] Tagawa [83] Tomizawa [86] Top [88] Vega [90]

4.5 6.0 1.5 3.0 Hazard ratio

Vega [90] Overall

Steels E, et al. Eur Respir J 2 0 0 1 ;1 8 ( 4 ) :7 0 5 -1 9

Gene Mutation (n=13)

SLIDE 28

BR.10 P53 Mutation & (Observa (Observa

Positive P53 IHC (>15%) is a marker for poor prognosis

100

P53 IHC results, observation

80 60 ility (%)

IHC negative

40 20 Probab

IHC positive

2 4 6 8 10 Years P=0.03 HR=1.89 (1.07-3.34)

& Protein Expression ation Arm) ation Arm)

P53 mutation is not prognostic for survival

100

P53 mutation results, observation

80 60 ility (%)

P53 Wild Type

40 20 Probab

P53 Mutant

2 4 6 8 10 Years P=0.45 HR=1.15 (0.75-1.77)

Tsao et al. J Clin Oncol 2 5 : 5 2 4 0 , 2 0 0 7

SLIDE 29

Positive P53 IHC P B fit f Adj Benefit from Adjuva

132/253 (52%) patien

100 80

P53 IHC positive

Chemotherapy

80 60 40 ability (%) 40 20 Proba P=0.02 HR=0.54 (0.32-0.92)

Observation

2 4 6 8 10 Years

I t ti P Interaction P

Predicts Survival t Ch th ant Chemotherapy

nts were P53 IHC positive

100 80

P53 IHC negative

Observation

80 60 40 ability (%)

Chemotherapy Observation

40 20 Proba P=0.26 HR=1.4 (0.78-2.52)

Chemotherapy

2 4 6 8 10 Years

P l 0 02

Tsao et al. J Clin Oncol 2 5 : 5 2 4 0 , 2 0 0 7

P value 0.02

SLIDE 30

P53 Mutation is No for Adjuvant for Adjuvant

125/397 (31%) patie

100 80

Mutant P53

80 60 40 ability (%)

Chemotherapy

40 20 Proba P=0.35 HR=0.78 (0.46-1.32)

Observation

Interaction P

2 4 6 8 10 Years

Interaction P

t a Predictive Marker Chemotherapy Chemotherapy

ents had P53 mutations

100 80

Wild Type P53

Chemotherapy

80 60 40 ability (%)

py Ob ti

40 20 Proba P=0.04 HR=0.67 (0.46-0.98)

Observation

2 4 6 8 10 Years

P value 0 65

Tsao et al. J Clin Oncol 2 5 : 5 2 4 0 , 2 0 0 7

P value 0.65

SLIDE 31

Beta Tubulin Levels

SLIDE 32

Influence of ß-Tubul In Advanc In Advanc

100

P = 0 0 0 3 9

Paclitaxel

80 60 40 nt Disease Free

TubIII <50% n=22 TubIII >50% n=25

P = 0 .0 0 3 9

20 Percen 200 400 600 800 1000 1200 100 80 60 ent alive

TubIII <50% n=22 TubIII >50% n=25

P = 0 .0 0 2 3

40 20 Perce 200 400 600 800 1000 1200 1400

bTubIII level predicts for paclitax

Tim e ( days)

lin Levels on Survival ced NSCLC ced NSCLC

100

P = 0 7 6

Gemcitabine

80 60 40 nt Disease Free

TubIII <50% n=17 TubIII >50% n=27

P = 0 .7 6

20 Percen 100 200 300 400 500 600 100 80 60 nt alive

TubIII <50% n=17 TubIII >50% n=27

P = 0 .8 5

40 20 Percen 100 200 300 400 500 600

Seve et al. Mol Cancer Ther 2005; 4: 2001-7.

xel sensitivity in advanced disease

Tim e ( days)

SLIDE 33

Influence of ß-Tubul In Operable N In Operable N

OS in observation arm by bTubIII expression level

100 80

e

< Med. H-score

Possibly prognostic

60 40 20

Percentage

> Med. H-score

P 0 08

2 4 6 8 10 Time (Years)

P=0.08

Adjuvant cisplatin/vinorelbi prognostic effect of high ßTubIII p g g Interactio

lin Levels on Survival NSCLC: JBR 10 NSCLC: JBR.10

OS in cisplatin/vinorelbine arm by bTubIII expression level

100 80

e

< Med. H-score

60 40 20

Percentage

> Med. H-score

2 4 6 8 10 Time (Years)

p=0.70

ine overcomes the negative expression in early stage NSCLC

Seve et al. Clin Cancer Res 2 0 0 7 ;1 3 :9 9 4 -9 .

p y g

n p=0.25

SLIDE 34

EG

Predictive versus pr Predictive versus pr EGFR mu EGFR protein EGFR gene co

GFR

rognostic markers rognostic markers utations expression

py number

SLIDE 35

Somatic Mutatio Tyrosine-Kinas Tyrosine-Kinas

TM K Tyrosi EGF ligand binding K R H DFG GXGXXG 718 745 776 835 8 18 19 20 21 Exon: 757-750 719 757 750 Paez: Lynch: 719 Pao: Tumor with point mutation (amino acid substitu

Lync

Tumor with in-frame deletion

ns In The EGFR

se (TK) Domain se (TK) Domain

DFG Y Y Y Y Autophosphorylation ne kinase L L Y 58 861 869 964

747-750

22 23 24 858 858 ution)

Adapted from: Pao et al. Proc Natl Acad Sci U S A. 2004; 101: 13306; ch et al. N Engl J Med. 2004; 350: 2129; Paez et al. Science. 2004; 304: 1497.

SLIDE 36

BR.21: Survival Ac

EGFR M t EGFR Muta

Exon 19 Deletions and L858R Mutations

ge

100 80 P=0.12 Hazard ratio, 0.55 (95% CI, 0.25-1.19)

Percentag

60 40 20

Erlotinib

20 6 12 18 24

Time (months)

Placebo

Although the HR for EGFR lower than those of EGFR WT p i i ibl d

Interaction P

interaction, possibly due to sam

ccording to Updated ti St t ation Status

Wild-Type EGFR and Indeterminate Variants

ge

100 80 P=0.09 Hazard ratio, 0.74 (95% CI, 0.52-1.05)

Percentag

60 40 20

Erlotinib

20 6 12 18 24

Time (months)

Placebo

mutation +ve patients is atients, there is no significant l i (I i 0 47)

Tsao et al J Clin Oncol: Sept, 2 0 0 8

value = 0.47

mple size (Interaction p=0.47)

SLIDE 37

RESPONSE ACCOR ISEL, IDEA

EGFR EGFR Status ISEL I

ORR (%) O N=158

EGFR +ve

N=158 13 (8.2%) (

EGFR -ve

N=69 1 (1 5%) (1.5%)

RDING TO EGFR IHC AL & BR.21

DEAL BR.21 TOTAL*

ORR (%) ORR (%) ORR (%) N=84 N=106 N=348 N=84 13 13.4%) N=106 12 (11.3%) N=348 38 (10.9%) N=17 1 (5 6%) N=80 3 (3 8%) N= 166 5 (3 0%) (5.6%) (3.8%) (3.0%) *P=0.003

SLIDE 38

Survival Accor Protein Ex

EGFR+

100 80 Erlotinib Placebo Log-rank: p=0 02 60 40 ercentage Log rank: p 0.02 HR=0.68 (0.49, 0.95) 20 Pe 6 12 18 24 30 At risk Erlotinib117 71 43 5 5 Months Placebo 67 23 12 5

Interaction p=0.25

rding to EGFR xpression

EGFR–

100 80 Erlotinib Placebo Log-rank: p=0 70 60 40 ercentage Log rank: p 0.70 HR=0.93 (0.63, 1.36) 20 Pe 6 12 18 24 30 At risk Erlotinib 93 42 22 8 3 Months Placebo 48 24 14 3 Tsao M-S et al. N Engl J Med 2005;353:133–44

SLIDE 39

BR.21 Survivals Acc EGFR Gene Copy EGFR Gene Copy

Disomy, trisomy or low polysomy

80 100 ge P=0.35 Hazard ratio, 0.80 (95% CI, 0.49-1.29) 20 40 80 Percentag

Erlotinib

6 12 18 24 30 20 Time (months)

Placebo

In m ultivariate analysis EGFR prognostic marker for poorer su significant predictive marker o significant predictive marker o from erlotinib (Inte

cording to Updated y Number Status y Number Status

High polysomy + Amplification

80 100 ge P=0.004 Hazard ratio, 0.43 (95% CI, 0.23-0.78) 20 40 80 Percentag

Erlotinib

6 12 18 24 20 Time (months)

Placebo

copy number is a significant urvival (p=0.025) and the only

f differential survival benefit

Tsao et al J Clin Oncol: Sept, 2 0 0 8

f differential survival benefit

eraction p=0.005)

SLIDE 40

What About EGFR in

1.0 0.8 0 6

EGFR Mut (n=35)

0.6 0.4

EGFR WT (n=60)

0.2

0 0143 L k t t

20 40 60 80

p=0.0143 Log-rank test p=0.0220 Breslow-Gehan-Wilcoxon tes

Early Stage NSCLC?

EGFR mutations are

prognostic prognostic

Intergroup BR.19 did

not select for any y EGFR marker

The OSI RADIANT trial

mandates EGFR IHC expression

Correlative studies

planned in both studies

Sasaki H, et al. I nt J Cancer 2 0 0 6 ;1 1 8 :1 8 0 – 4 .

st

SLIDE 41

K-RAS

SLIDE 42

RAS is a Key Reg Func Func

RTK ras

raf PI3K erk mek PI3K akt rac erk akt rac

SURVIVAL PROLIFERAT MOTILITY TRANSLATI

gulator of Cellular tions tions

GPR

mutation

Ral-GDS

mutation

Ral GDS

TION TRANSCRIPTION ON ANGIOGENESIS

SLIDE 43

NCIC-CTG Intergr

Completely resected NSCLC

N=239

Stages IB-II (Excl T3N0) Stratify N0 vs N1

RAS pos vs neg

N=243

vs neg vs unk

roup Trial BR.10

Observation Vinorelbine 25 mg/m2 g/ weekly x 16 weeks Cisplatin 50 mg/m2 days 1 & 8 x 4 months days 1 & 8 x 4 months

W inton et al. N Engl J Med. 2 0 0 5 ;3 5 2 :2 5 8 9 -9 7 .

SLIDE 44

BR.10: Over

100 80

ge

40 60

ercentag

20

P

HR 0.7, p=0.012

0.0 2.0 4

, p

Tim

rall Survival

69% Vin/Cis 54% / 2 Observation

.0 6.0 8.0

mes

W inton et al. N Engl J Med. 2 0 0 5 ;3 5 2 :2 5 8 9 -9 7 .

SLIDE 45

BR.10 Overall

Mutation Pos Mutation Pos

100 60 80 40 60

Observa

20

*HR 0.96, p=0.8

0.0 2.0

In the Cox model, significant inte and RAS mutation was not dete

Survival in RAS

sitive patients sitive patients

Vin/Cis ation 8865

4.0 6.0

W inton et al. N Engl J Med. 2 0 0 5 ;3 5 2 :2 5 8 9 -9 7 .

eraction between chemotherapy ected, Interaction p value 0.29

SLIDE 46

KRAS KRAS an

KRAS mutations see smokers smokers KRAS mutations see th ll t

ther cell types

KRAS mutations alm patients with EGFR

d EGFR nd EGFR

en more frequently in en in both adeno and most never seen in mutations

SLIDE 47

KRAS Gene Mutat

Progression-Free Survival

1 0 0.8 1.0 0.4 0.6 0.2 5 10 15 20 Months

Chemotherapy, wild type (n=103) Chemotherapy, mutant (n=30)

tions in TRIBUTE

Overall Survival

1 0 0.8 1.0 0.4 0.6 0.2

KRAS mutant

5 10 15 20 Months

Erlotinib + chemotherapy, wild type (n=104) Eberhard D, et al. J Clin Oncol 2 0 0 5 ;2 5 :5 9 0 0 – 9 . Erlotinib + chemotherapy, mutant (n=25)

SLIDE 48

BR.21 Survival Acc KRAS Muta KRAS Muta

100

KRAS Wild Type

80 60

ge

Median: 7.5 (5.4, 10.7) 3.4 (3.0, 7.1) HR=0.69 (0.49-0.97) P=0.0311 40 20

Percenta

Erlotinib

20 6 12 18 24

Time (months)

Placebo

Interaction P

Time (months)

Despite the large numerical diff interaction could be demons

Interaction P

interaction could be demons

cording to Updated ation Status ation Status

100

KRAS Mutation

80 60

ge

Median: 3.7 (1.9, 7.9) 7.0 (1.7, 19.5) HR=1.67 (0.62-4.50) P=0.3096 40 20

Percenta

Placebo

20 6 12 18 24

Time (months)

Erlotinib

P value = 0 09

Time (months)

ferences in the HRs, no significant strated Interaction p value 0.09

Shepherd et al. Proc ASCO 2007. Abstract 7571.

P value = 0.09

strated Interaction p value 0.09

SLIDE 49

RAS Su RAS Su

It appears that KRAS m
It appears that KRAS m

predict survival benef chemotherapy chemotherapy

Analyses are ongoing
Adequate power to co
Adequate power to co

achieved in the pooled

KRAS mutation status
KRAS mutation status

the trials of adjuvant E currently on-going cu e t y o go g

ummary ummary

mutation status may mutation status may fit from adjuvant g in other studies

nfirm this may be
nfirm this may be

d LACE-Bio analyses s may affect outcome in s may affect outcome in EGFR inhibitors

SLIDE 50

BRCA1

SLIDE 51

BRCA1 in BRCA1 in

Ind ces esista Induces resista cisplatin-mediat Increased levels poorer OS in co poorer OS in co chemotherapy-n

BRCA1 > 5 vs ≤

p=0.02

n NSCLC n NSCLC

nce to nce to ted apoptosis s associated with mpletely resected mpletely resected, naïve NSCLC

≤ 5: HR 1.98,

Rosell et al. ASCO 2 0 0 7 . Abst. 7 5 5 1 .

SLIDE 52

Spanish Customized

Completely Resect Completely Resect (SCAT

CONTROL

D

Resected NSCLC

CONTROL

D

pN1 / pN2

EXPERIMENTAL

EGFR mutations in adenoca

P ti t t tifi d b i i

Patients stratified by invasi

(CSF-1, EGFR & CA IX

d Adjuvant Therapy in ted N1 & N2 NSCLC ted N1 & N2 NSCLC T Trial)

Docetaxel/Cis Gem/Cis

Q 1 BRCA1

Docetaxel/Cis

Q 2 & 3 BRCA1

Docetaxel/Cis

BRCA1 Q 4 BRCA1

Docetaxel

Q 4 BRCA1

Docetaxel

arcinomas i i t ive gene signature ) & tumor size

SLIDE 53

Micro-Array Profiling y g

SLIDE 54

Micro Micro-

Array Expres

Array Expres P di O P di O

D d f

Predict O Predict O

Alive 5 years Dead of Genes Genes Tumor Sample (Pa Tumor Sample (Pa

Potti et al. N Engl J Med 355: 570-80, 2006

ssion Profiles That ssion Profiles That O

b 2 5

Outcome Outcome

cancer by 2.5 years atients) atients)

SLIDE 55

Prognostic Gene Si by Microarray by Microarray

Tu

Raponi/Beer (2006) Squa Potti (2006) Lu (2006) Larsen (2007) L (2007) S Larsen (2007) S Chen (2007) Bianchi (2007) Bianchi (2007) Lau (2007) Director’s Challenge 08

N

ignatures in NSCLC y or RT QPCR y or RT-QPCR

umor Type Gene Set

amous/Adeno 50/47 NSCLC Metagene sets NSCLC 64 Adeno 54 S 111 Squamous 111 NSCLC 5 NSCLC 10 NSCLC 10 NSCLC 3 Adeno Multiple sets

None of these was tested as predictive marker

SLIDE 56

BR.10 15-gene Signa Stage I and Stage II Stage I and Stage II

Stage IB (n=34)

100

entage

60 80

Perce

20 40

Low Risk

20 20 3 19 6

Time (Years)

13 High risk Low risk

No. at Risk

HR 13.32 (95% CI 2.86-62.11) p<0.0001

Low Risk High Risk

20 14 19 6 13 2

ature is Prognostic in Observation Patients Observation Patients

Stage II (n=28)

100

centage

60 80

Perc

20 40 11 3 9 6

Time (Years)

7 9 High risk Low risk 9 1 11 17 9 3 7 1

HR 13.47 (95% CI 3.00-60.43) p<0.0001

1

Tsao et al. Proc ASCO, 2008

SLIDE 57

CALGB 30506 Sch

Resection T (1.75 to 4. Resection T (1.75 to 4. LM Score <0.55; 850 LM Score <0.55; 850 Observation Observation Observation Observation

LM Scores Blinde LM Scores Blinde

hema (Stage IA/IB)

.0) N0 Patients + Array .0) N0 Patients + Array N=1296 N=1296 LM Score LM Score > > 0.55; 446 0.55; 446 Randomize Randomize Randomize Randomize

Observation Observation Adjuvant Adjuvant Chemotherapy Chemotherapy ed to Investigators ed to Investigators

SLIDE 58

CALGB 30506 S

Resection T (1.75 to 4. Resection T (1.75 to 4. LM Score <0.55; 850 LM Score <0.55; 850 Randomize Randomize Randomize Randomize

Adjuvant Adjuvant Chemotherapy Chemotherapy N=425 N=425 Observation Observation N=425 N=425 LM Scores Blinde LM Scores Blinde

Schema Updated

.0) N0 Patients + Array .0) N0 Patients + Array N=1296 N=1296 LM Score LM Score > > 0.55; 446 0.55; 446 Randomize Randomize Randomize Randomize

Observation Observation N=223 N=223 Adjuvant Adjuvant Chemotherapy Chemotherapy N=223 N=223 ed to Investigators ed to Investigators

SLIDE 59

Primum Non Nocere

SLIDE 60

Reasons For My C C t CALG Current CALG

The study is limited to

ti t ith t patients with tumors <

With the exception of J

UFT, no study of platin chemotherapy has eve advantage in these pat

UFT never studied in W
Meta-gene study propo

risk Stage IA and smal risk Stage IA and smal chemotherapy or obse

Concern With The B T i l D i B Trial Design

Stage IA and Stage IB 4 <4cm Japanese studies of num-based er shown a survival tients Western patients p

sed to randomize low

ll IB patients to ll IB patients to ervation

SLIDE 61

Reasons For My C Reasons For My C Current CALG

LACE analysis sugg

HARM from platinum HARM from platinum chemotherapy in Sta

Concern With The Concern With The GB Trial Design

gested potential m-based adjuvant m based adjuvant age IA

SLIDE 62

LACE: Effect Chemothera Chemothera

CT may be detrimental for stage generally not given the potentiall vinorelbine (13% of stage IA patie vinorelbine (13% of stage IA patie

t of Adjuvant apy by Stage apy by Stage

rd ratio

py / Control)

HR [95% CI] rd ratio

py / Control)

HR [95% CI] 1.41 [0.96;2.09] 0.92 [0.78;1.10]

py )

1.41 [0.96;2.09] 0.92 [0.78;1.10]

py )

[ ; ] 0.83 [0.73;0.95] 0.83 [0.73;0.95] [ ; ] 0.83 [0.73;0.95] 0.83 [0.73;0.95]

ntrol better

1.5 2.0 2.5

ntrol better

1.5 2.0 2.5

IA, but stage IA patients were ly best combination cisplatin + ents vs ~43% for other stages) ents vs ~43% for other stages)

SLIDE 63

Reasons For My C Reasons For My C Current CALG

LACE analysis suggested pot

based adjuvant chemotherapy

CALGB d BR 10

CALGB and BR.10 r

subgroup analyses and the potential for patients with tumors p

Concern With The Concern With The B Trial Design

ential HARM from platinum- y in Stage IA

t ti retrospective showed NO benefit r harm in Stage IB s <4cm

SLIDE 64

CALGB 9633: S IB Patients by

Survival: Patients with Tumours ? 4 0 cm

>

Survival: Patients with Tumours ? 4.0 cm

0.8 1.0

Observation Chem o

>

0.4 0.6

Probability

N=97 N=99

2 4 6 8

S i l Ti (Y )

0.0 0.2 1 2 3 4 5 6 7 8 9 1 2 3 4 5 6 7 8 9

HR=0.66; 90% CI: 0.45-0.97; p=0.04 Survival Tim e (Years)

urvival of Stage g y Tumour Size

Survival: Patients With Tumour < 4 cm

1.0 0.6 0.8

bability

Observation Chem o

N=74 N=74

0.2 0.4

Prob HR=1.02; 90% CI: 0.67-1.55; p=0.51

2 4 6 8

Survival Tim e (Years)

0.0 1 2 3 4 5 6 7 8 9

SLIDE 65

BR.10: Surviva by Tumo

BR.10: Survival Patients with BR.10: Survival Patients with Tumour > 4.0 cm

BR.10 OVERALL SURVIVAL

Stage IB, Tumour Size > 4 cm

Observation Vin/Plat

100 Percen tage 40 60 80

# at Risk

Observation

P 20

41

2

29

4

18

6 Time(Years)

8

8 10

Observation Vin/Plat Hazard Ratio (95% C.I.): Vin/Plat/Observation: 0.48 (0.23,1.00) Test for equality of groups: Log Rank: p=0.0453

Summary Statistics:

41 50 29 46 18 32 8 14 2

Hazard Ratio 0.48 (CI 0.23-1.00), LR p=0.048

al of IB Patients

ur Size

BR.10: Survival Patients with Tumour < 4.0 cm

BR.10 OVERALL SURVIVAL

Stage IB, Tumour Size <= 4 cm

Observation Vin/Plat

100 Percentage 40 60 80

# at Risk

Observation Vi /Pl t

20

67 61

2

62 47

4

39 33

6 Time(Years)

21 13

8

8 4

10

Vin/Plat Hazard Ratio (95% C.I.): Vin/Plat/Observation: 1.46 (0.78,2.75) Test for equality of groups: Log Rank: p=0.2349

Summary Statistics:

61 47 33 13 4

Hazard Ratio 1.46 (CI 0.78-2.75), LR p=0.23

SLIDE 66

Reasons For My C Reasons For My C Current CALG

LACE analysis suggested pot

based adjuvant chemotherapy

CALGB and BR 10 retrospecti
CALGB and BR.10 retrospecti

NO benefit in Stage IB patient

BR 10 micro-array p
BR.10 micro-array p

harm to low risk pat

Concern With The Concern With The B Trial Design

ential HARM from platinum- y in Stage IA ive subgroup analyses showed ive subgroup analyses showed s with tumors <4cm

profiling suggests profiling suggests tients

SLIDE 67

Chemotherapy Bene but Not Low

100

BR.10, high risk (n=67)

but Not Low

rcentag

60 80

Per e

20 40 Observ. 31 9 3 3 6

Time (Years)

9 Chemo Observation

No. at Risk

HR 0.33 (95% CI 0.17-0.63) p=0.0005

Chemo 36 25 15 1

In

efits BR.10 High Risk Risk Patients Risk Patients

100

BR.10, low risk (n=66)

rcentage

60 80

Per

20 40 Chemo Observation 31 3 28 6

Time (Years)

20 9 1 Chemo Observation

HR 3.67 (95% CI 1.22-11.06) p=0.0133

35 28 19 3

nteraction p = 0.0001

SLIDE 68

What About Plain

Higher response r
Higher response r

in adenocarcinom H

However, no conv

a differential effec

Pemetrexed inferio

(docetaxel and ge ( g squamous cell can

Is this due to high

Is this due to high

n Old Histology?

rates to EGFR TKIs rates to EGFR TKIs ma i i id f vincing evidence of ct on survival

r to other agents

mcitabine) in ) ncer her TS levels?? her TS levels??

SLIDE 69

Cis/Pem vs. Cis/Gem Overall Survival in A Overall Survival in A

0.9 1.0 Probability 0.6 0.7 0.8 l Survival P 0 3 0.4 0.5 Overal 0 0 0.1 0.2 0.3

* non-squamous = patients with adenocarcinoma or larg Overall Survival Time (months) in

6 12 0.0

Scagliotti et al. J Clin Oncol doi:10.1200/JCO.2007.15.0375

Cis/Pem, cisplatin/pemetrexed; Cis/Gem, cisplatin/gem mos, months

m in First-Line NSCLC: Adeno or Large Cell Adeno or Large Cell

Ci /P (N 512) 11 8 (10 4 13 2) OS Median (95% CI) Cis/Pem (N=512) 11.8 mos (10.4, 13.2) Cis/Gem (N=488) 10.4 mos (9.6, 11.2) OS Adjusted HR (95% CI) Cis/Pem vs Cis/Gem 0 81(0 70-0 94)

Cis/Pem statistically superior OS vs. Cis/Gem

Cis/Pem vs. Cis/Gem 0.81(0.70-0.94)

ge cell carcinoma n Non-Squamous* Patients

18 24 30

citabine; CI, confidence interval; HR, hazard ratio;

SLIDE 70

Cis/Pem vs. Cis/Gem Overall Survival

1.0 0 9

bability

0.7 0.8 0.9

Survival Pro

0.4 0.5 0.6

Overall S

0.1 0.2 0.3

Cis/Pem cisplatin/pemetrexed; Cis/Gem cisplatin/gemcita Overall Survival Time (months)

6 12 0.0

Scagliotti et al. J Clin Oncol doi:10.1200/JCO.2007.15.0375

Cis/Pem, cisplatin/pemetrexed; Cis/Gem, cisplatin/gemcita cisplatin+pemetrexed; CG, patients treated with cisplatin+

m in First-line NSCLC: in Squamous Cell

OS Median (95% CI) Cis/Pem (N=244) 9.4 mos (8.4, 10.2) Cis/Gem (N=229) 10.8 mos (9.5, 12.1) OS Adjusted HR (95% CI) j ( % ) Cis/Pem vs. Cis/Gem 1.23 (1.00-1.51) OS effect with Cis/Pem less than with Cis/Gem

abine; CI confidence interval; CP patients treated with ) in Squamous Patients

18 24 30

abine; CI, confidence interval; CP, patients treated with gemcitabine; HR, hazard ratio; mos, months

SLIDE 71

JMEI 2nd li Survival by

JM E I 2 n d lin e N S C L C : E

1 00

P a tie n ts R a n d o m ize d to P e m e tre e d rviving P e m e trex e d

N on- squam ous: M edian = 9 2

50

Percent Su

9.2

O ve ra ll S u rviva l (m o n th s )

S quam ous: M edian = 6.2

ne NSCLC y Histology

E fficacy b y H isto lo g y

10 0 P a tie n ts R a n d o m ize d to D o c e taxe l

urviving

N on-squam ous: M edian = 8.2

50

Percent Su

S quam ous: M edian = 7.4

O ve ra ll S u rviva l (m o n th s )

SLIDE 72

JMEI 2nd lin Efficacy by

S OS Non-Squamous OS

(n = 399; 279 events) Adj HR = 0.778 (95%CI:0.607-

0 997) p=0 0475 Alimta Med = 9.3 mos. 0.997) p=0.0475 Med 9.3 mos. Docetaxel Median = 8.0 mos.

ne NSCLC y Histology

S OS Squamous OS

(n = 172; 130 events) Adj HR = 1.563

(95%CI:1 079-2 264) (95%CI:1.079 2.264) p=0.0182 Docetaxel Median = 7.4 mos. Alimta Alimta Median = 6.2 mos.

SLIDE 73

JMEI: Treatment -by-

Treatment by Histology

Non-squa

Histology Interaction: Survival Adjusted for Cofactors

Pemetrexe (n=205)

j (p=0.001)

Median OS, months

9.3

Adjusted OS HR (95% CI)

0.778 (0

Median PFS, months

3.1

Adjusted PFS HR (95% CI) 0.823 (0

Histology Interaction

amous group Squamous group ed Docetaxel (n=194) Pemetrexed (n=78) Docetaxel (n=94)

8.0 6.2 7.4 .607, 0.997) 1.563 (1.079, 2.264) 3.0 2.3 2.7 .664, 1.020) 1.403 (1.006, 1.957)

SLIDE 74

My Favourite Mo My Favourite Mo

(Or what I would like to see

KRAS
EGFR

EGFR

ERCC1

EGFR

EGFR
P53 (IHC)
Proper histologic
Micro-array not re

Micro array not re

lecular Markers
lecular Markers
n every pathology report)

subtyping ady for prime time ady for prime time

SLIDE 75