Focus on angiogenesis: new learning points from clinical trials

1
(No Transcript)
2
Chairs introduction

• Nick Thatcher
• Christie Hospital NHS Trust,Manchester, UK

3
Lung cancer the burden of disease

• Global incidence 1.3 million cases/year
• Global deaths 1.2 million/year

Approximately 70 of NSCLC patients are diagnosed
with advanced disease
Ferlay J, et al. GLOBOCAN 2002 http//www-dep.iarc
.fr/
4
What is the ideal situation?

• Being able to assess a patients disease profile
  and use this to determine the best possible
  treatment to optimise clinical benefit
• How do we do this?
• By identifying quantifiable molecular
  characteristics that predict for benefit with a
  particular therapeutic option
• a biomarker

5
Biomarker testing the Herceptin example

• Human epidermal growth factor receptor 2 (HER2)
  overexpressed in 30 breast cancers
• Herceptin (trastuzumab) monoclonal antibody
  directly targets HER2 and is highly effective in
  early-stage, HER2 patients1
• Diagnostic test developed to identify patients
  with high levels of HER2 expression (IHC 3)
• Test is used to identify patients who will have
  greatest benefit from Herceptin,2 but has still
  not been fully optimised or standardised after 5
  years

1. Romond EH, et al. N Engl J Med
2005353167384 2. Smith IE. Anticancer Drugs
200112S310
6
Objectives of the meeting

• Review the potential biomarkers being assessed
  for Tarceva and the challenges we face in
  identifying a viable biomarker
• Examine some possible biomarkers for Avastin and
  look at how these will be investigated
• Consider the impact of the introduction of
  testing procedures on clinical practice and the
  implications for future lung cancer therapy

7
(No Transcript)
8
Question cards and evaluation forms

• Question cards are available from the hostesses
  if you have a question for one of the speakers,
  please complete the card and hand back to a
  hostess
• Please also complete the evaluation form inside
  your abstract book and hand to a hostess as you
  leave the meeting

9
(No Transcript)
10
Testing times the challenge of biomarker
research in lung cancer

• Marc van de Vijver
• Academic Medical CentreAmsterdam, The Netherlands

11
What is a biomarker?

• A characteristic that is objectively measured and
  evaluated as an indicator of normal biological
  processes, pathogenic processes or
  pharmacological responses to therapeutic
  intervention1

1Biomarker Definitions Working Group, Biomarkers
and Surrogate Endpoints Preferred Definitions
and Conceptual Framework. Clin Pharmacol Ther
2001698995
12
Prognostic versus predictive markersan
important distinction
Prognostic marker Indicates the likelihood of
outcome (tumour recurrence or patient survival)
regardless of the specific treatment the patient
receives
Predictive marker Indicates the likelihood of
response to a specific therapy
www.cancerdiagnosis.nci.nih.gov
13
Many steps involved in establishing a clinically
viable and useful biomarker
Generation of a biomarker hypothesis
Identification of candidate markers
Develop the assay
Clinical validation of candidate markers using
assay
Refinement of companion diagnostic assay
Approval of test and distribution to laboratories
Auditing to ensure consistency and validity of
results
14
Testing methods for molecular biomarkers

• Protein expression
• immunohistochemistry (IHC)
• enzyme-linked immunosorbent assay (ELISA)
• Gene expression
• assessed by microarray technology or reverse
  transcription-polymerase chain reaction (RT-PCR)
• Gene copy number
• fluorescent/chromogenic in-situ hybridisation
  (FISH/CISH)
• Gene sequence
• DNA sequencing (other methods possible for known
  mutations)

Only IHC, FISH and ELISA are established in
diagnostic laboratories other methods are
limited to research laboratories
15
The main challenges in biomarker testing
Variations in testing methods used at different
sites results cannot be directly compared
Validation of biomarkers and testing methods is a
complex and lengthy process
Availability of tissue in sufficient quantities
and of high enough quality
Need to have consensus of opinion and cooperation
between industry and academia
Future biomarkers, e.g. proteomics and gene chips
may require new technology and methods
Assays need to be quick, reliable, inexpensive
and easy to establish in laboratories
16
EGFR activates multiple pathwaysleading to
invasive growth
Extracellular
Intracellular
P
P
JNK
PI3K Akt
MAPK

• Proliferation
• Invasion
• Metastasis
• Angiogenesis
• Inhibition of apoptosis

Woodburn J. Pharmacol Ther 19998224150 Lynch
TJ, et al. N Engl J Med 2004350212939Knowlden
JM, et al. Endocrinology 2003144103244
Chakravarti A, et al. Cancer Res 200262430715
17
Tarceva targets EGFR-mediated cellular processes
Extracellular
Intracellular
Tarceva
Etessami A, Bourhis J. Drugs Future
2000258959 Moyer J, et al. Cancer Res
199757483848Harari PM, Huang SM. Semin Radiat
Oncol 200212(Suppl. 2)216
18
Potential markers in the EGFR pathway
Proliferation
Survival
PIP2
PIP3
PTEN
GTP-RAS
GDP-RAS
RAF
P
P
Adaptor
Adaptor
PI3K
MEK
AKT
MAPK
Transcription factors
Apoptosis regulators
Nucleus
19
EGFR a logical biomarker for Tarceva?

• EGFR is expressed at high levels in up to 80 of
  NSCLC cases1
• Tarceva specifically targets the tyrosine kinase
  domain of EGFR and directly affects the cellular
  processes mediated by this receptor
• The EGFR pathway appears to be the logical choice
  for investigation, but any biomarker for clinical
  benefit with Tarceva needs to be validated via
  prospective, controlled trials

1Rusch V, et al. Clin Cancer Res 1997351522
20
Potential biomarkers predicting level of benefit
with Tarceva

• EGFR protein or mRNA expression
• EGFR gene copy number
• EGFR gene mutations
• KRAS gene mutations
• Markers of downstream signalling, such as pMAPK
  and pAKT
• EGFR ligands

21
Assessing EGFR protein expression by IHC
membranous staining in ?10 of tumour cells
Incomplete membrane and cytoplasmic staining
Complete membrane staining
Tsao M-S, et al. N Engl J Med 200535313344
Of any intensity
22
IHC testing

• IHC is widespread in pathology laboratories
• However, IHC is semi-quantitative
• Laboratories must use correct testing procedures
  and undergo adequate quality control
• No tests validated for NSCLC (although one is
  available for colorectal cancer)

23
Assessing EGFR gene amplification by FISH
scoring categories (Cappuzzo et al.)
Disomy
Trisomy
High polysomy
Gene amplification
Cappuzzo F, et al. J Natl Cancer Inst
20059764355
24
FISH testing

• More widespread availability in pathology
  laboratories since advent of HER2 testing in
  breast cancer
• technicians will not have as much experience as
  with IHC
• HER2 FISH analyses are usually only performed in
  HER2 IHC 2 cases (approximately 10 of all
  cases)
• Scoring is objective but may vary between centres
• Currently, no validated test is available for
  NSCLC

25
EGFR gene mutations what are we looking for?

• The most commonly observed EGFR mutations are
  in-frame deletions in exon 19 and the L858R
  mutation in exon 21, although other mutations
  have also been identified
• Finding the appropriate technique is a challenge
• sequencing is time-consuming and not easily
  automated
• better alternatives are needed

Pao W, Miller VA. J Clin Oncol 200523255668
26
Learning from HER2 identifying the biomarker
HER2 mRNA
Normal levelsof HER2 protein
Over-replication selection of cells with growth
advantage
HER2 mRNA
Elevated levelsof HER2 protein
27
Learning from HER2 the rationale for selecting
patients to receive Herceptin
28
Learning from HER2 breast cancer testing
algorithm
Patient tumour sample
FISH / CISH
IHC


1
2
3
0
Herceptin therapy
Herceptin therapy
FISH / CISH



Herceptin therapy
Bilous M, et al. Breast 200312928Hanna WM,
Kwok K. Mod Pathol 2006194817
29
Summary

• Identifying and validating a biomarker is a
  complex process that can take many years
• Numerous difficulties need to be overcome to
  determine a biomarker for NSCLC
• When the process works (as with HER2 testing for
  breast cancer), it can lead to more
  individualised therapy and improved outcomes for
  patients

30
(No Transcript)
31
Leading change Tarceva at the forefront of
biomarker research

• Martin Reck
• Hospital GrosshansdorfGrosshansdorf, Germany

32
An unmet need in second-line NSCLC

• Many patients are diagnosed with advanced disease
  and are unsuitable for surgery
• first-line chemotherapy is usual course of
  treatment
• For patients who fail on first-line therapy,
  there are limited options available in second
  line
• patients may have reduced performance status due
  to previous treatment or progression
• need for new and different therapeutic options to
  offer more choice for clinicians

33
BR.21 significant improvement in overall
survival with Tarceva
1.00 0.75 0.50 0.25 0
Tarceva Placebo
Hazard ratio (HR)0.73, p0.001
Survival distribution function

• 42.5 improvement in median survival

0 5 10 15 20 25 30
Survival time (months)
HR and p (log-rank test) adjusted for
stratification factors at randomisation and
epidermal growth factor receptor (EGFR) status
Shepherd F, et al. N Engl J Med
200535312332Tarceva Summary of Product
Characteristics, F. Hoffmann-La Roche Ltd
34
All patient subgroups derived a survival benefit
in BR.21
Tarcevaplacebo PS 01 0.73 (0.60.9) PS 23
0.77 (0.61.0) Male 0.76 (0.60.9) Female
0.80 (0.61.1) lt65 years 0.75 (0.60.9) ?65
years 0.79 (0.61.0) Adenocarcinoma 0.71
(0.60.9) Squamous-cell carcinoma 0.67
(0.50.9) Other histology 1.04 (0.71.5)
Tarcevaplacebo Never-smoker 0.42
(0.30.6) Current/ex-smoker 0.87 (0.71.1) 1
prior regimen 0.76 (0.61.0) 2 prior regimens
0.75 (0.61.0) Best prior response CR/PR 0.67
(0.50.9) Best prior response SD 0.83
(0.61.1) Best prior response PD 0.85
(0.61.2) Asian 0.61 (0.41.0) Other 0.79
(0.70.9)
Factors
0 1 2
0 1 2
HR
HR
Shepherd F, et al. N Engl J Med 200535312332
HR lt1 improved survival with Tarceva
35
BR.21 response to Tarceva according to EGFR
biomarker status
However, RR does not take into account survival
benefit associated with stable
disease Progression-free survival (PFS) and
overall survival (OS) are the relevant endpoints
to consider
Includes indeterminate variantsExon 19
deletions and L858RIHC immunohistochemistry
Tsao M-S, et al. N Engl J Med 200535313344 Tsao
M-S, et al. N Engl J Med 20063545278
Shepherd FA, et al. J Clin Oncol 200725(Suppl.
18 Pt I)402s (Abs. 7571)
36
BR.21 survival according to EGFR gene copy
number using FISH
EGFR low copy
EGFR high copy
100 80 60 40 20 0
100 80 60 40 20 0
HR0.80 (CI 0.491.29)p0.03525
HR0.43 (CI 0.230.78)p0.0042
Percentage
Percentage
0 6 12 18 24
0 6 12 18 24
Time (months)
Time (months)
Interaction p0.12 (not significant)
Shepherd FA, et al. J Clin Oncol 200725(Suppl.
18 Pt I)402s (Abs. 7571)
37
BR.21 survival according to EGFR mutation status
EGFR wild-type
Exon 19 or 21 mutations
100 80 60 40 20 0
100 80 60 40 20 0
HR0.74 (CI 0.521.05)p0.0924
HR0.55 (CI 0.251.19)p0.1217
Percentage
Percentage
0 6 12 18 24
0 6 12 18 24
Time (months)
Time (months)
Interaction p0.47 (not significant)
Shepherd FA, et al. J Clin Oncol 200725(Suppl.
18 Pt I)402s (Abs. 7571)
38
BR.21 EGFR biomarker status does not predict for
survival benefit with Tarceva
0.25
0.12
0.47
p value for subgroup compared with placebop
value for interactionIncludes indeterminate
variantsExon 19 deletions and L858RFISH
fluorescence in-situ hybridisation
Tsao M-S, et al. N Engl J Med 200535313344 Tsao
M-S, et al. N Engl J Med 20063545278
Shepherd FA, et al. J Clin Oncol 200725(Suppl.
18 Pt I)402s (Abs. 7571)
39
Survival according to EGFR IHC and EGFR FISH
status in TRUST open-label phase IV study
1.00 0.75 0.50 0.25 0
1.00 0.75 0.50 0.25 0
EGFR IHC (n229) EGFR IHC (n55)
EGFR FISH (n49) EGFR FISH (n156)
Survival distribution function
Survival distribution function
HR0.662p0.0341
HR0.766p0.1056
0 100 200 300 400 500 600 700 800
0 100 200 300 400 500 600 700 800
Overall survival (days)
Overall survival (days)

• TRUST is a single-arm study and cannot therefore
  distinguish between prognostic and predictive
  value

Based on staining in ?10 tumour cellsLog-rank
test
Schneider CP, et al.J Clin Oncol 200725(Suppl.
18 Pt I)427s (Abs. 7674)
40
The presence of KRAS mutations any predictive
value?
100 80 60 40 20 0
100 80 60 40 20 0
KRAS wild-type
KRAS mutation
Median (months) Tarceva 3.7 Placebo 7.0
HR1.67 (CI 0.624.50)p0.3096
Percentage
Percentage
0 6 12 18 24
0 6 12 18 24
Time (months)
Time (months)

• Small patient numbers in BR.21 analysis1
  difficult to distinguish between prognostic and
  predictive value
• data from PA.3 study in pancreatic cancer found
  no clear relationship between KRAS mutations and
  lack of survival benefit2

1Shepherd FA, et al. J Clin Oncol 200725(Suppl.
18 Pt I)402s (Abs. 7571) 2OSI Pharmaceuticals,
Ltd. Data on file.
41
Erlotinib first line in patients with EGFR
mutations response rate
Overall RR 82 (95 CI 6692) Exon 19 RR 95
(95 CI 75100) Exon 21 RR 67 (95 CI 4187)
Paz-Ares L, et al J Clin Oncol 200624(Suppl. 18
Pt 1) (Abs. 7020)
42
Tarceva first line in patients with EGFR
mutations time to progression

• Median TTP 13.3 months

1.0 0.5 0
1.0 0.5 0
Exon 19
Cumulative survival
Cumulative survival
Exon 21
Log-rank p0.06 Breslow p0.04
0 2 4 6 8 10 12
0 2 4 6 8 10 12
Time (months)
Time (months)
Paz-Ares L, et al J Clin Oncol 200624(Suppl. 18
Pt 1) (Abs. 7020)
Censored
43
Spanish Lung Cancer Group phase III trial in
EGFR-mutated NSCLC
Stage IIIB/IV NSCLC Mutations of exon 19 or 21 of
EGFR No previous treatment for metastatic
disease Measurable or evaluable disease gt18
years of age ECOG PS 2
Arm A Tarceva 150mg/day po
R A N D O M I S E D
Crossover at PD
Planned n147
Arm B CDDP 100mg/m2 i.v./gemcitabine
1,250mg/m2 CDDP 75mg/m2 i.v./docetaxel 75mg/m2
i.v. CBDCA AUC5/gemcitabine 1,000mg/m2 CBDCA
AUC5/docetaxel 75mg/m2 i.v. (up to four cycles)
ECOG PS Eastern Cooperative Oncology Group
performance status
44
Summary of the current biomarker situation for
Tarceva

• We now have hypotheses of potentially predictive
  biomarkers for Tarceva
• No current data from randomised, prospective,
  controlled trials to support patient selection,
  or exclusion from receiving Tarceva
• Patient selection based on clinical
  characteristics can not be justified, given that
  all patient subgroups in BR.21 had a survival
  benefit

45
Ongoing biomarker research programme
46
Marker identification trial (MERIT) results
available at ECCO 2007
n264 recruited, second line after standard
first-line treatment
Mandatorysamples Tumour biopsy Tumour block(if
available) RNA blood sample
Biopsy
Tarceva 150mg/day
PD
Data from the MERIT study will be presented
during the Basic science/Medics programme Lung
cancer (2) session on Wednesday 26 September at
10.00
SecondRNA bloodsample
Day 1
Day 28 Day 1
Week 6
Follow-up every 12 weeks
Clinical assessment at screening and every 6
weeks until PD
Primary endpoint gene expression profiling
(prediction of clinical benefit) Secondary
endpoints EGFR mutation analysis molecular
(explorative) assessment of putative
alterations of downstream targets of EGFR
47
SATURN sequential Tarceva in unresectable NSCLC
Stratify by EGFR protein expression (IHC) results
Tumour samples (mandatory)
Tarceva 150mg/day
PD
Off study
Chemotherapy naïvestage IIIB/IV NSCLC Planned
n1,700
Four cycles of first-line standard platinum-based
doublet
Non-PD
11
(n850)
Placebo
PD
Off study

• The largest randomised trial ever to
  prospectively address relationships between EGFR
  markers and clinical outcomes due to report in
  2008/2009

48
Other Tarceva trials performing biomarker analyses
49
Challenges in identifying and implementing a
biomarker for Tarceva

• Potential biomarkers need to be validated
• Need to differentiate between prognostic and
  predictive value
• can only be confirmed in large prospective,
  randomised trials
• Assays need to be standardised
• Obtaining adequate samples
• Working in a multi-disciplinary approach

50
Summary

• No evidence to support patient selection by
  baseline characteristics, as all patient groups
  derive a survival benefit from Tarceva
• Molecular markers are promising but more data are
  needed
• No tests have been validated so far
• Cooperation from academics, clinicians and
  industry is essential
• A large-scale clinical trial programme is
  underway to investigate numerous potential
  biomarkers for Tarceva

51
(No Transcript)
52
The role of erythropoietinreceptors in oncology

• Robert Pirker
• Medical University of Vienna, Austria

53
Erythropoietin a well established and effective
treatment for anaemia in patients with lung cancer

• Epoetin beta 30 000 IU Once Weekly improves
  haemoglobin (Hb) levels, reduces blood
  transfusions and improves Quality of Life1,2

Patients remaining transfusion-free ()1
Mean Hb level (g/dL)1
n40
1. Pirker et al. Lung Cancer. 20075589942. de
Castro et al. Lung Cancer. 2007 (In press)
54
The erythropoietin receptor
Erythropoietin

• Homodimerisation
• JAK 2 phosphorylation
• Tyrosine phosphorylation of EPO-R and cytosolic
  signal transducers
• Activation of transcription factors (STAT 5) and
  enzymes (PI-3K, MAPK)
• Survival, proliferation, differentiation

S
S
SOS
JAK 2
JAK 2
STAT 5
Ras
P
P
GRB2
RAF
P
SHC
STAT 5
P
MEK
P
P
PI-3K
MAPK
Transcription
Jelkmann. Intern Med. 200443649659
55
EPO-R in cancer cells

• Recent reports claim that patients with tumours
  expressing EPO-R are at risk of disease
  progression if treated with erythropoietin1
• However, current data suggest no survival risk
  when erythropoietin is used within current
  guidelines2
• Current research aims at identifying EPO-R in
  tumour biopsies and cancer cell lines

1. Henke et al. J Clin Oncol. 20062447084713 2.
Bokemeyer et al. Eur J Cancer. 200743258270
56
Detection and measurement of EPO-R expression

• Transcription level
• Protein level (cell surface or intracellularly)
• Functionality of EPO-R
• in vitro tests
• animal models

Sinclair et al. Cancer. 2007110477488
57
Transcription level

• RT-PCR (reverse transcriptase-polymerase chain
  reactions)
• EPO-R transcripts have been detected in
• cancer tissues and tumour cell lines
• normal tissues (e.g. bone marrow, kidney, heart,
  brain, endothelium and smooth muscle)
• High expression levels only in normal bone marrow
• Only transcripts detected and not functional EPO-R

Sinclair et al. Cancer. 2007110477488
58
Protein level

• EPO-R expression is detected at the protein level
• Commercially available EPO-R antibodies include1
• C20
• M-20
• H-194
• 07-311
• Major limitation1,2
• lack of specificity

1. Elliott et al. Blood. 200610718921895 2.
Österborg et al. Eur J Cancer. 200743510519
59
Commercially available antibodies lack
specificity for detecting EPO-R
Santa Cruz C20
FL AG-EPO-R?40
UT-7 (EPO-R ve)
FL AG-EPO-R
Empty vector
SH-SY5Y
Caki-2
MCF-7
769-P
HeLa
kDa
100
Anti-EPO-R antibodies are non-specific and bind
to multiple targets
80
EPO-R
60
50
40
30

• Antibody C20 recognised at least three additional
  non-EPO-R proteins in cancer cells
• All other antibodies tested (M-20, H-194, 07-311)
  also detected non-EPO-R proteins (up to 20
  different types)

Elliott et al. Blood. 200610718921895
60
Functional level

• In vitro studies
• conflicting results stimulation vs no
  stimulation
• experimental doses which showed an effect were
  much higher than pharmacological (therapeutic)
  doses
• Animal models show no evidence of tumour cell
  proliferation

Sinclair et al. Cancer. 2007110477488
61
Clinical evidence meta-analysis data suggest no
negative impact of erythropoietin on survival
CI, confidence interval HR, hazard ratio OS,
overall survival RR, relative risk
1. Bohlius et al. J Natl Cancer Inst.
200597489498 2. Bohlius et al. J Natl Cancer
Inst. 200698708714 3. Hedenus et al. J Clin
Oncol. 20052369416948 4. Aapro et al. Br J
Cancer. 20069514671473
62
Future requirements

• Reliable and clinically relevant assays for EPO-R
  
• Prospective studies on large, homogenous
  populations with homogenous treatments

63
Summary

• The role of EPO-R in cancer needs further
  evaluation
• Results from previous experimental studies on
  EPO-Rs should not be over-interpreted in terms
  of their clinical relevance
• Clinical decisions on the use of erythropoietin
  shouldbe based on evidence from clinical studies

64
(No Transcript)
65
Focus on angiogenesis new learning points from
clinical trials

• Nick Thatcher
• Christie Hospital NHS TrustManchester, UK

66
Chemotherapy in advanced NSCLCreached a plateau
in efficacy
14 12 10 8 6 4 2 0

• Single-agent platinum compounds significantly
  improved overall survival versus BSC1
• Platinum-based doublet chemotherapy provided
  significantly superiorsurvival versus single
  agent1,2
• Modern platinum-based chemotherapy doubletsyield
  similar efficacy3

Therapeutic plateau
Platinum-based doublets810 months
Single-agent platinum68 months
Median survival (months)
BSC 25 months
1970s
1980s
1990s
1Bunn. JCO 2002 2Delbaldo, et al. JAMA
2004 3Schiller, et al. NEJM 2002
BSC best supportive care
67
Phase III trial of Avastin plus CP in NSCLC
(E4599) trial design
CP (n444)
PD
Previously untreated stage IIIB/IV non-squamous
NSCLC (n878)
Avastin (15mg/kg) every 3 weeks CP (n434)
Avastin every 3 weeks until progression
PD
No crossover permitted

• Primary endpoint overall survival
• Avastin 15mg/kg i.v. administered every 3 weeks
• Carboplatin i.v. to AUC 6mg/mL and paclitaxel
  200mg/m2 i.v. every 3 weeks

CP carboplatin/paclitaxel PD progression of
disease i.v. intravenous AUC area under the
curve
Sandler, et al. NEJM 2006
68
E4599 overall survival
Avastin CP CP
Median overall survival gt12 months
HR0.79 (0.670.92) p0.003
15 10 5 0
1.0 0.8 0.6 0.4 0.2 0
Median overall survival (months)
Probability of survival
Avastin CP
CP
10.3
12.3
0 6 12 18 24 30 36 42
Time (months)
Sandler, et al. NEJM 2006
HR hazard ratio
69
Breaking through the therapeutic plateau
14 12 10 8 6 4 2 0

• Single-agent platinum compounds significantly
  improved overall survival versus BSC1
• Platinum-based doublet chemotherapy provided
  significantly superior survival versus single
  agent1,2
• Avastin plus platinum doublet was the first
  regimen in a large phase III trial to extend
  overall survival beyond historical benchmark of1
  year3

Therapeutic plateau
E4599 Avastin platinum-based doublet 12.3
months
Platinum-based doublets810 months
Single-agent platinum68 months
Median survival (months)
BSC 25 months
1970s
1980s
1990s
2006
1Bunn. JCO 2002 2Delbaldo, et al. JAMA
2004 3Sandler, et al. NEJM 2006
70
E4599 Avastin has awell-characterised safety
profile
20 18 16 14 12 10 8 6 4 2 0
CP Avastin 15mg/kg CP
Percentage
Hypertension
Proteinuria
Headache
Rash/ desquamation
Arterial thrombosis
Venous thrombosis
Grade 3 adverse events (AEs)
Sandler, et al. NEJM 2006
71
Patient selection significantly reduced risk of
pulmonary bleeding in E45991 compared with the
earlier phase II trial of Avastin plus CP2
CNS central nervous system GI gastrointestinal
1Sandler, et al. NEJM 2006 2Johnson, et al. JCO
2004
72
Phase III trial of Avastin plus CG in NSCLC
(AVAiL) trial design
Avastin 7.5mg/kg CG (n345)
R A N D O M I S E
Avastin
PD
Previously untreated, stage IIIB, IV or recurrent
non-squamous NSCLC (n1,043)
Placebo CG (n347)
Placebo (no crossover allowed)
PD
Avastin 15mg/kg CG (n351)
Avastin
PD

• Primary endpoint PFS
• Initiated to complement the E4599 trial and
  evaluate Avastin in combination with a
  platinum-based chemotherapy regimen commonly used
  in Europe

Manegold, et al. ECCO 2007
CG cisplatin/gemcitabine PFS
progression-free survival
73
AVAiL improvement in PFS
1.0 0.8 0.6 0.4 0.2 0
33 improvement
Possibility of PFS
22 improvement
0 3 6 9 12 15 18
Time (months)
Manegold, et al. ECCO 2007
74
Clinical relevance of HR

• The HR
• is an indication of the risk over time
• takes the entire survival curve into account
• does not focus on a single time point

75
Avastin-based therapy improves duration of
response and response rate

• Avastin therapy should be continued until disease
  progression
• Avastin inhibits new/recurrent vessel growth,
  delaying disease progression

7 6 5 4 3 2 1 0
Response rate 34
Response rate 30
Response rate 20
Median responseduration (months)
Avastin 7.5mg/kg CG (n323)
Avastin 15mg/kg CG (n332)
Placebo CG (n324)
Manegold, et al. ECCO 2007
76
AVAiL overall survival

• Overall survival data are immature due to short
  duration of follow up
• Overall survival data are likely to be available
  in 2008

77
No unexpected safety signalswere observed in
AVAiL
VTE venous thromboembolism
Manegold, et al. ECCO 2007
78
AVAiL incidence of pulmonary haemorrhage

• 38 of patients had central lesions
• 4/10 patients with severe pulmonary haemorrhage
  had central lesions
• Therapeutic anticoagulation was initiated in 9
  of patients during the trial
• of these, none had a severe pulmonary haemorrhage

Manegold, et al. ECCO 2007
79
AVAiL and E4599have comparable PFS benefit
PROGRESSION OR DEATH
Avastin or placebo chemotherapy
Second-line antineoplastic therapy
1Sandler, et al. NEJM 2006 2Manegold, et al.
ECCO 2007
80
AVAiL had slightly more stringent exclusion
criteria than E4599
81
Similar safety profile was observed in E4599 and
AVAiL
50 45 40 35 30 25 20 15 10 5 0
Grade ?3 AEs
E4599 Avastin 15mg/kg CP AVAiL Avastin 7.5mg/kg
CG AVAiL Avastin 15mg/kg CG
Percentage
Bleeding
Hypertension
Proteinuria
Neutropenia
Febrile neutropenia
Sandler, et al. NEJM 2006 Manegold, et al. ECCO
2007
E4599 reports only grade 4/5 haematological
events
82
Appropriate patient selection improves the
riskbenefit ratio
AVF0757g Avastin 7.5 or 15mg/kg CP E4599
Avastin 15mg/kg CP AVAiL Avastin 7.5mg/kg
CG AVAiL Avastin 15mg/kg CG
10 8 6 4 2 0
Restricting eligibility to patients with
non-squamous histology and minimal baseline
haemoptysis
Percentage
Grade ?3 pulmonary haemorrhage
Johnson, et al. JCO 2004 Sandler, et al. NEJM
2006 Manegold, et al. ECCO 2007
Phase II trial including patients with squamous
cell histology
83
24 August 2007 EU approvalof Avastin for NSCLC
The European Union has approved the use of
Avastin at a dose of 7.5mg/kg or 15mg/kg, in
combination with platinum-based chemotherapy, for
the first-line treatment of patients with
unresectable advanced, metastatic or recurrent
NSCLC other than predominantly squamous-cell
histology

• The approval is based on data from the pivotal US
  phase III trial (E4599) and the Avastin in Lung
  (AVAiL) phase III trial, which both demonstrate
  that Avastin is effective in combination with
  platinum-based chemotherapy

84
Conclusions Avastin consistently improves
outcomes in advanced NSCLC

• Avastin administered until progression with
  platinum-based chemotherapy
• extends overall survival beyond 12 months
• significantly delays disease progression
• has a well-characterised safety profile
• E4599 and AVAiL demonstrate that Avastin can be
  successfully combined with different chemotherapy
  backbones in NSCLC
• Avastin plus platinum-based chemotherapy
  represents the standard of care for
  Avastin-eligible patients with advanced NSCLC

85
(No Transcript)
86
Focus on angiogenesisthe quest for a biomarker

• Peter Carmeliet
• Katholieke Universiteit LeuvenBelgium

87
Mechanism of action of Avastin
Regression
Inhibition (key in adjuvant setting)
Normalisation
88
Considerable interest in patient selection based
on molecular markers

• Currently no validated biomarkers in lung cancer
• main focus has been on biomarkers for epidermal
  growth factor receptor (EGFR)-targeted therapies
• biomarkers for chemotherapy are being
  investigated (ERCC1 and RRM1)
• no validated data to indicate that any biomarker
  is predictive of clinical benefit with
  anti-angiogenic agents
• Avastin has proven activity in non-squamous
  non-small cell lung cancer (NSCLC)
• identification of a biomarker for Avastin could
  potentially lead to further efficacy gains

89
Hurdles/caveats

• Data from randomised trials required to validate
  potential biomarkers
• Development of standardised assays required
• Avastin is used in combination with chemotherapy
• may influence the effect identified for a
  particular biomarker

90
Biomarkers for anti-angiogenesis agents what do
we currently know?
Due to small sample sizes, most investigations
are hypothesis-generating only
Most analyses are retrospective and therefore
exploratory
?
No biomarker that reliably predicts patient
outcome on Avastin-based therapy has been
identified to date
Identifying a biomarker for Avastin in
combination with chemotherapy may not be clear-cut
91
Phase III trial E4599 demonstrated a survival
benefit for Avastin plus chemotherapy
CP (n444)
PD
Previously untreated stage IIIB/IV non-squamous
NSCLC (n878)
Avastin (15mg/kg) every 3 weeks CP (n434)
Avastin every 3 weeks until progression
PD
No crossover permitted
CP carboplatin/paclitaxel PD progression of
disease
Sandler, et al. NEJM 2006
92
Measurement of biomarkers in E4599

• Plasma VEGF, bFGF, ICAM-1 and E-selectin levels
  measured by ELISA pretreatment
• bFGF, ICAM-1 and E-selectin levels also assessed
  after 7 weeks of therapy
• Pretreatment levels were correlated with response
  to therapy and changes were correlated with
  response to endothelial cell apoptosis
• Pre- and post-treatment measurements were
  available for 150166 and 113 patients,
  respectively

VEGF vascular endothelial growth factorbFGF
basic fibroblast growth factor ICAM-1
intercellular adhesion molecule-1 ELISA
enzyme-linked immunosorbent assay
Dowlati, et al. ASCO 2006
93
Rationale for selection of biomarkers

• VEGF and bFGF well-known angiogenic factors
• ICAM-1 E-selectin
• stimulate angiogenesis, activate endothelial and
  epithelial cells
• mediate extravasation of inflammatory cells
  (immunity)
• mediate extravasation of tumour cells
  (metastasis)
• elevated levels seen in a variety of malignancies
• soluble form shed from dysfunctional apoptotic
  endothelium (anti-angiogenic agents)
• soluble form can be measured in plasma

Dowlati, et al. ASCO 2006
94
Median pretreatment biomarker levelsin E4599
were comparable between arms
Dowlati, et al. ASCO 2006
95
Mean biomarker levels pretreatmentand at 7 weeks
in E4599
Dowlati, et al. ASCO 2006
96
Interpretation of changes in biomarker levels
from baseline to week 7 in E4599

• bFGF levels increased significantly from baseline
  to week 7
• No association between baseline measurements and
  clinical factors including
• gender
• age
• race
• sites of metastasis

Dowlati, et al. ASCO 2006
97
Patients with lower than median baseline ICAM-1
had higher response rates than those with higher
than median baseline ICAM-1
Dowlati, et al. ASCO 2006
98
Patients with lower than median baseline ICAM-1
had better overall and 1-year survival than those
with higher than median baseline ICAM-1
Survival by baseline ICAM-1
lt260.5ng/mL (62 deaths/75 cases)gt260.5ng/mL (70
deaths/75 cases) p0.00005
1.0 0.8 0.6 0.4 0.2 0
1 year 60
Probability
1 year 25
0 10 20 30 40 50
Months

• Baseline plasma ICAM-1 levels may have prognostic
  value in advanced NSCLC
• Trials are needed to determine if these will be
  better markers of clinical outcome than
  conventional radiographical response assessment

Dowlati, et al. ASCO 2006
99
Mechanism of ICAM-1?

• ICAM-1
• stimulates angiogenesis, activates endothelial
  and epithelial cells
• mediates extravasation of inflammatory cells
  (immunity)
• mediates extravasation of tumor cells
  (metastasis)
• elevated levels seen in a variety of malignancies
• soluble form shed from dysfunctional apoptotic
  endothelium (anti-angiogenic agents)
• soluble form can be measured in plasma

Dowlati, et al. ASCO 2006
100
Biomarkers in E4599 conclusions and issues

• ICAM-1 levels appear to be associated with
  outcomes in patients with advanced NSCLC treated
  with Avastin CP. However
• no diagnostic test for ICAM-1 has been validated
  for clinical use, limiting the clinical utility
  of these data
• the current analysis does not show that high
  ICAM-1 levels preclude a response to therapy
• further prospective analysis of ICAM-1 levels in
  clinical trials is required
• grouping the values into high and low levels
  in this analysis results in a loss of power
• use of a data-derived median cut-off leads to
  serious bias

101
The future developing assays for biomarkers

• Markers under investigation include

D-dimers PAI-1
Imaging (DCE-MRI, FDG-PET, DC-US)
102
The future the potential of biomarkers for
Avastin safety

• Plasma samples are collected in clinical trials
  of Avastin to
• investigate biomarkers for Avastin-associated
  adverse events
• Thrombosis
• decreases anti-thrombotic protection (less NO and
  PGI2)
• increases endothelial damage dysfunction
• upregulates plasminogen activator inhibitor-1
  (PAI-1), tissue factor, etc.
• effects on plaque rupture due to destabilisation
  (?)
• Bleeding
• decreases renewal capacity of endothelial cells
• causes endothelial dysfunction and defects of the
  interior vascular lining (pruning regression)

103
The future the potential of biomarkers for
Avastin safety (contd)

• Proteinuria
• Anti-VEGF therapy may distort the glomerular and
  peritubular capillary network and impair the
  glomerular filtration unit
• Hypertension
• Anti-VEGF inhibits the vasodilative effect of
  VEGF (nitric oxide)

104
The future the potential of biomarkers for
Avastin safety (contd)

• Anti-VEGF may aggravate heart disease, in
  particular in patients receiving a combination
  treatment (cardiotoxic chemotherapy) or in
  patients at risk for (ischaemic) heart disease
• Mechanism pruning of the coronary
  microvasculature (?)
• Plasma samples for cardiac biomarker analysis are
  collected from consenting patients at the same
  time as assessment of left ventricular systolic
  function
• The cardiac biomarker analysis will include
  markers for myocardial dysfunction resulting from
  apoptosis and necrosis, such as troponin, proBNP,
  etc.

105
The future preclinical rationale for anti-VEGF
therapy in the adjuvant setting
Control Control MAb (200µg) Anti-VEGF MAb
(10µg) Anti-VEGF MAb (50µg) Anti-VEGF MAb
(100µg) Anti-VEGF MAb (200µg)
1,600 1,200 800 400 0

• In preclinical models, inhibition of VEGF has
  been shown to
• suppress new vascular sprouting within24
  hours1,2
• inhibit growth of colorectal tumours3

Tumour volume (mm3)
0 7 14 21
Time (days)
1Baluk, et al. Curr Opin Genet Dev 2005 2Inai, et
al. Am J Pathol 2004 3Warren, et al. J Clin
Invest 1995
Anti-VEGF agents VEGF-Trap and AG013736
106
The future Avastin in the adjuvant setting
107
The future an extensive biomarker program for
Avastin
Advanced NSCLC
Ovarian cancer
Adjuvant BC
BO20571
ICON-7
INNOVATIONS
BEATRICE
Avastin
Pancreatic cancer
AVITA
NO16966
AVOREN
AVADO
AVEREL
AVANT
CRC
mRCC
HER2ve HER2ve
CRC colorectal cancer BC breast cancer mRCC
metastatic renal cell carcinoma HER2 human
epidermal growth factor receptor 2
108
Conclusions

• To date, no biomarker can be used to select
  patients for, or predict outcome of,
  Avastin-based therapy in patients with advanced
  NSCLC
• Retrospective data indicate a role for low ICAM-1
  as a potential prognostic marker
• will need to be clinically validated
• Three areas of focus for Avastin are biomarkers
  that
• predict decreased benefit and rapid progression
  after initial response
• predict clinical benefit
• identify patients at risk of developing
  particular adverse effects

109
(No Transcript)
110
A continued commitment to improving patient
outcomes

• Stefan Scherer
• Biomarker Programme Leader, Roche Basel

111
Complexity in combination with cytotoxic
chemotherapy

• Pharmacokinetics
• Drug perfusion
• Intrinsic drug resistance

Anti-angiogenesis is a complex, multistep process
and thus unlikely to be reflected by a single
biomarker
112
Biomarkers for Avastin what do we know?

• Relatively few prospective biomarker
  investigations have been undertaken
• To date no biomarkers are available to suggest
  which patients derive most benefit from Avastin
• those that have been undertaken included only
  small sample numbers and are hypotheses-generating
  only

113
Many biomarkers are under investigation
tVEGF (VEGF-expression)
MVD
pVEGF levels (VEGF-A/-B/-C/-D/-E)
VEGFR
VEGFR phosphorylation
THBS-2
sFGF
D-dimers
sVEGFR-1/-2/-3
CEC
CEP
ICAM
VCAM
E-selectin
114
Which specimens will be collected

• Additional methodologies may be applied for
  corroboration, including RT-PCR, in-situ
  hybridisation and gene expression profiling. The
  collected tumour tissue and blood samples may be
  used to develop and validate diagnostic assays
  and might allow the generation of statistically
  meaningful biomarker data

optional, separate signed ICF
115
Which biomarker will be tested?
optional, separate signed ICF
116
Biomarkers for Avastin safety

• Plasma samples are collected in clinical trials
  of Avastin to investigate biomarkers for
  Avastin-associated adverse events

Proteinuria
Bleeding
Hypertension
Thrombosis
Cardiac safety
117
Extensive biomarker programme for Avastin in all
tumour types
Please note, this is not a comprehensive list of
all trials
118
Summary

• A biomarker must be reproducible, applicable and
  validated
• Biomarker research is at an early stage
• angiogenesis and anti-angiogenesis are complex
  and unlikely to be reflected by a single marker
• Investigation of biomarkers for Avastin is active
  with three main areas of focus
• biomarkers that identify patients who may benefit
  from an alternative Avastin-based treatment
  strategy to those being currently used
• biomarkers that predict clinical benefit
• biomarkers that will allow physicians to monitor
  more closely patients at risk for adverse events

119
(No Transcript)
120
Chairs close

• Nick Thatcher
• Christie Hospital NHS Trust,Manchester, UK

121
Summary and conclusions

• An extensive clinical trial programme is
  investigating potential biomarkers for Tarceva
• candidate biomarkers include EGFR protein
  expression, EGFR gene amplification and EGFR gene
  mutations/polymorphisms
• MERIT is reporting at ECCO SATURN will report in
  2008/2009
• many challenges to overcome in implementing
  biomarker testing in practice
• commitment to identifying those patients who will
  obtain greatest benefit from Tarceva

122
Summary and conclusions

• Latest trial data demonstrate that Avastin, when
  combined with platinum-based chemotherapy,
  improved outcomes in first-line NSCLC
• a consistent benefit across subgroups was
  observed
• Research is underway to identify potential
  biomarkers for Avastin and this is reflected in
  the biomarker analyses included in many ongoing
  Avastin trials
• complicated by unique mechanism of action of
  Avastin and its use in combination rather than as
  single agent

123
Thank you for attending

• Please complete your evaluation forms and hand
  them in

124
(No Transcript)