Download ASCO_2009_files/Tabernero Discussant PD ASCO 2009

K-Ras and Beyond
Josep Tabernero, MD
Vall d’Hebron University Hospital
Barcelona, Spain
Disclosures
• Participated in Advisory Boards of Merck, Amgen, Imclone,
Sanofi-Aventis, Onyx, and Roche
K-Ras and B-Raf in CRC
•
•
Constitutive mutations of K-Ras
predict resistance to anti-EGFR
MoAbs in CRC:
– refractory1 setting
– first-line2-3 setting
– basis for regulatory approval
(EMEA) & national guidelines
(NCCN)
Role of mutations of other signal
transducer proteins is being
evaluated:
– i.e. B-Raf: refractory setting4
EGFR
RAS
RAF
PI3K
Akt
MEK
MAPK
Cell
Cell
Survival Proliferation
1Lièvre,
A. Cancer Res; 66:3992-3995, 2006
Cutsem, E. et al. N Engl J Med; 360:1408-1417, 2009
3Bokemeyer, C. et al. J Clin Oncol; 27:663-671, 2009
4Di Nicolantonio, F. et al. J Clin Oncol; 26:5702-5712, 2008
2Van
K-Ras, B-Raf, N-Ras and PIK3CA
mutations and cetuximab efficacy
A4020 – Poster Board #: 11; Diether Lambrechts et al.
The role of KRAS, BRAF, NRAS, and PIK3CA mutations as markers of
resistance to cetuximab in chemorefractory metastatic colorectal cancer.
Lambrechts: Patients and Methods
Endpoint
Utility
Performance of 4 tumor based tests: K-Ras, BRaf, N-Ras and PIK3CA mutation status
Predictive biomarker
Specimen
Tumor specimens (paraffin-embedded)
Patients
Sample size
Refractory mCRC treated with Irinotecan +
Cetuximab
276 tumors  580 tumors (European
consortium)
Sequenom MALDI TOF MassArray system
Assay
Lambrechts: Results (1)
KRAS
BRAF
Mutations included
% coverage of
potential mutations
(Cosmic)
Mutation rate
detected
G12S, G12R , G12C, G12D , G12A , G12V ,
G13D, A146T, G13A, G13V, A59T, Q61K , Q61E,
Q61P, Q61R, Q61L, Q61H
99.2%
36.5%
V600E ,K601E, D594G ,V600M
97%
( 622 samples)
5%
(589 samples)
NRAS
PI3K
Q61P,Q61L,Q61H,Q61H,Q61Q,Q61E,G13S,G13C
,G13R,Q61K,Q61R, G12D,G12S ,G12C
97%
H1047R, H1047L , K179T, P539R,Q546K,Q546E,
E81K, R88Q,G106V,N345K, R93W, S158L,
H160N,R38H,E542K, E542Q,E545K,E545Q,
G118D, G12D,K567R,H1047Y, P134S, R108H,
C420R,H701P,K184E, C901F,M1004I, G1049R,
G1007R, G1049S
86%
6%
(261 samples worked)
13%
(578 samples)
Lambrechts: Results (1)
• K-Ras, B-Raf and N-Ras mutually exclusive
• 17.7% K-Ras mt and 10.4% K-Ras wt had a
PIK3CA mutation
(p= 0.009 Pearson Chi square)
• 6% B-Raf mutants and 13% B-Raf wt had a
PIK3CA mutation
(p= 0.412 Fisher’s Exact test)
• Representative series: outcomes in accordance
with the literature
– mPFS 18 wks, mOS 38 wks (≈BOND)
Lambrechts: Results (2) - RR
KRAS
CR + PR
SD + PD
total
p
WT
130 (36%)
226 (64%)
356
p<.001
Mut
11 (5%)
192 (95%)
203
BRAF
CR + PR
SD + PD
total
p
WT
141 (26%)
399 (74%)
540
p=.035
Mut
2 (8%)
24 (92%)
26
NRAS
CR + PR
SD + PD
total
p
WT
50 (21%)
179 (79%)
239
p=.317
Mut
1 ( 6%)
14 (94%)
15
PI3K
CR + PR SD + PD
total
PI3K
128 (27%) CR357
(73%) SD + PD
WT
485
+ PR
In KRAS wt
Mut
WT
10 (14%)
Mut
60 (86%)
117 (38%)
8 (24%)
p
p=.028
total
70
195 (62%)
312
26 (76%)
34
p
p=0.107
Lambrechts: Results (3) – PFS & OS
KRAS
Median PFS
HR (95% CI)
P value (log rank)
WT
MUT
N=369
N=212
24 weeks
12 weeks
0.542 (0.452-0.650)
<0.001
Lambrechts: Results (3) – PFS & OS
KRAS
BRAF
Median PFS
Median PFS
HR (95% CI)
HR (95% CI)
P value (log rank)
P value (log rank)
WT
MUT
WT
MUT
N=369
N=212
N= 561
N= 28
24 weeks
12 weeks
19 weeks
7.8 weeks
0.542 (0.452-0.650)
0.410 (0.275-0.610)
<0.001
<0.001
Lambrechts: Results (3) – PFS & OS
KRAS
WT
MUT
N=369
N=212
WT
MUT
WT
MUT
24
weeks
12
N= 561
N= 28weeks
505
73
HR (95%
0.542 (0.452-0.650)
Median
PFS CI)
19 weeks
7.8 weeks
Median PFS
19 weeks 12.5
valueCI)
(log rank)
<0.001
HR P(95%
0.410 (0.275-0.610)
HR (95% CI)
0.772 (0.601-0.991)
P value (log rank)
<0.001
P value (log rank)
0.036
BRAF
PI3K
Median PFS
Lambrechts: Results (3) – PFS & OS
KRAS
WT
MUT
N=369
N=212
MUT
All KRAS wt WT
Median
PFS
24 weeks
weeks MUT
PI3K PI3K
N=
561
WT WTN=12
28
MUT
HR (95%
0.542323
(0.452-0.650)
Median
PFS CI)
19 weeks
505
7.8 weeks
73
36
BRAF
valueCI)
(log
HR P(95%
Median
PFSrank)
Median
PFS
P value
HR (log
(95%
CI) CI)
HRrank)
(95%
P value
(log rank)
P value
(log rank)
<0.00112.5
19
0.410
weeks
(0.275-0.610)
24
weeks
23 weeks
0.772
<0.001
(0.601-0.991)
0.848(0.599-1.201)
0.036
0.338
Lambrechts: Results (4) – Multivariate
OR Logistic regresion OR
95%CI
P value
KRAS
0.093
0.048 – 0.177
p<.001
BRAF
0.140
0.032 – 0.604
p=.008
Pi3K
Not retained
PFS Cox regresion
HR
KRAS
p=.136
95%CI
P value
0.523
0.434 – 0.631
p<.001
BRAF
0.328
0.217 – 0.497
p<.001
Pi3K
0.798
0.620 – 1.027
p=.079
OS Cox regresion
HR
95%CI
P value
KRAS
0.549
0.452 – 0.667
p<.001
BRAF
0.378
0.250 – 0.572
p<.001
Pi3K
Not retained
p=.187
Lambrechts: Conclusions
• K-Ras impact ≈ literature1
• N-Ras impact: not mature, full series to be
analyzed, currently mt incidence 6%
• B-Raf ≈ literature2. Most powerful negative
predictor
• PIK3CA: little effect, no effect if restricted to K-Ras
wt, not retained in multivariate analysis
 Discrepancy with the literature3,4 (although
limited number of patients)
1Lièvre,
2Di
A. Cancer Res; 66:3992-3995, 2006
Nicolantonio, F. et al. J Clin Oncol; 26:5702-5712, 2008
3 Sartore-Bianchi, A et al Cancer Res; 69:1851-7, 2009
4Ann Oncol ;20:84-90, 2008
Lambrechts: Implications
• Strengths:
– Unique and consistent population
– Large database
– Not influenced by other treatments
Lambrechts: Implications
• Weakness:
– Not all the mutations have the same addictive role
– Other possible deregulations not considered so
far: PTEN mutations, PTEN loss of function, Src
mutations, p53 mutations, …
– Other potential predictors:
• Role of the ligands
• Polymorphisms1-3:
– EGFR, EGF, …
– Fc receptors (ADCC): FcgammaRIIa-H131R and
FcgammaRIIIa-V158F
1Lurge,
J et al. Clin Cancer Res 1;14:7884-95,2008
W wt al. J Clin Oncol 20;25:3712-8,2007
3Bibeau, F et al. J Clin Oncol; 27:1122-9,2009
2Zhang,
Amphiregulin/Epiregulin
A4016 – Poster Board #: 7; Derek J Jonker et al.
High epiregulin (EREG) gene expression plus K-ras wild-type (WT) status as
predictors of cetuximab benefit in the treatment of advanced colorectal
cancer (ACRC): Results from NCIC CTG CO.17—A phase III trial of
cetuximab versus best supportive care (BSC).
A4019 – Poster Board #: 10; Hans Prenen et al.
Use of amphiregulin and epiregulin mRNA expression in primary tumors to
predict outcome in metastatic colorectal cancer treated with cetuximab.
A4021 – Poster Board #: 12; Fotios Loupakis et al.
Amphiregulin (AR) expression in the prediction of benefit from cetuximab
plus irinotecan in KRAS wild-type metastatic colorectal cancer (mCRC)
patients.
Amphiregulin/Epiregulin
• EGFR ligands:
– 1 in C. Elegans
– 4 in Drosophila
– 7 in mammals: EGF, TGF-α,
HB-EGF, amphiregulin
(AREG), betacellulin,
epiregulin (EREG) and
epigen1
– EREG and AREG bind more
weakly to EGFR than EGF but
much more potently and
prolonged
– EREG preferentially activates
heterodimers2
• High gene expression levels of
EREG and AREG predict
response to cetuximab3
1Singh,
AB et al. Cell Signal; 17:1183-1193,2005
M et al. J Biol Chem; 273:10496-10505,1998
3Khambata-Ford, S. et al. J Clin Oncol; 25:3230-3237, 2007
2Shelly,
Jonker: Patients and Methods
Endpoint
Utility
Three tumor based tests: K-Ras mutation status
and EREG & AREG (not shown) expression
Predictive biomarker
Tumor specimens (paraffin-embedded material) –
Study NCIC CTG CO.17
Patients
Refractory mCRC treated with Cetuximab or BSC
Sample size K-Ras 394/572 (69%); EREG 385/572 (67%)
Assay
EREG gene expression by quantitative RT-PCR
Specimen
Jonker: Background
NCIC CTG CO.17: mCRC Cetuximab vs BSC
HR OS: ITT 0.7  K-Ras wt 0.55
1Jonker,
2Karapetis,
DJ et al. NEJM; 357:2040-8,2007
CS. et al. NEJM;359:1757-65,2008
Jonker: Results (1)
• EREG in K-Ras wt as a continuous variable:
prognostic and predictive
EREG and OS in patients with K-Ras wild-type
Adjusted HR (95% CI) for 1 unit
increase in EREG normCT (toward
normal)
P
value
CET/BSC
1.17 (1.04-1.32)
0.01
BSC
1.13(1.01-1.27)
0.04
Study arm
Test for treatment / biomarker
interaction (adjusted p value)
HR 1.03 (0.88-1.20) p=0.75
EREG and PFS in patients with K-Ras wild-type
Adjusted HR (95% CI) for 1 unit
increase in EREG normCT (toward
normal)
P
value
CET/BSC
1.13 (1.01-1.26)
0.03
BSC
0.96 (0.87-1.07)
0.48
Study arm
Test for treatment / biomarker
interaction (adjusted p value)
HR 1.14 (0.98-1.33) p=0.08
Jonker: Results (2)
• EREG in K-Ras wt as a categorical variable
(high vs low): predictive but not prognostic
– In K-Ras wt patients on BSC, high EREG
expression did not correlate with OS using:
• pre-specified threshold: adjusted HR 0.82
[0.58-1.15], p=0.24
• minimum p threshold: adjusted HR 0.85 [0.591.22], p=0.38
Jonker: Results (3)
• Combimarker: K-Ras wt and high EREG
– Pre-especified threshold1
– Minimum threshold: 169/384 (44%)
• response rate was 15.5 vs 0% for cetuximab vs BSC
• median PFS was 5.1 vs 1.9 months for cetuximab vs
BSC (HR, 0.33; p<0.0001)
• median OS was 9.9 vs 5.0 months for cetuximab vs
BSC (HR, 0.46; p<0.001)
• Implications in patients to be treated:
– All comers
 394 (100%)
HR: 0.7
– K-Ras wt
 230 (58%)
HR: 0.55
– Combimarker  169 (44%)
HR: 0.46
1Khambata-Ford,
S. et al. J Clin Oncol; 25:3230-3237, 2007
Jonker: Results (4)
• Combimarker: K-Ras wt and high EREG
– Minimum threshold: 169/384 (44%)
High EREG by minimum-p threshold
100
Cetuximab + BSC
80
Proportion alive
Proportion alive
100
Low EREG by minimum-p threshold
60
40
BSC alone
20
Cetuximab + BSC
80
60
40
BSC alone
20
HR 0.46 [0.32-0.65], p<0.0001
HR 0.93 [0.51-1.71], p=0.81
0
0
0
2
4
6
8
10
12
14
0
2
4
6
8
10
84
85
80
73
76
54
66
26
43
19
28
14
18
10
8
5
30
26
25
18
16
15
13
10
8
5
5
3
Time from randomization (months)
Time from randomization (months)
Prenen: Patients and Methods
Endpoint
Utility
Three tumor based tests: K-Ras mutation
status and EREG and AREG expression
Predictive biomarker
Specimen
Tumor specimens (paraffin-embedded)
Patients
Sample size
Irinotecan refractory mCRC treated with
Irinotecan + Cetuximab
220 tumors + 67 tumors (external validation)
EREG and AREG gene expression by
quantitative RT-PCR
Assay
Prenen: Results (1)
• EREG expression is higher in K-Ras wt than
in K-Ras mut tumors (p=0.0002)
Prenen: Results (2)
• EREG and AREG expression as a continuous variable
is predictive of response in K-Ras wt but not in mut
tumors
EREG
p=.0005
AREG
p=.0017
Prenen: Results (3)
• EREG and AREG expression as a categorical variable
is predictive of RR, DCR, PFS, OS in K-Ras wt tumors
Odd ratio
EREG
AREG
RR
5.04
DCR
20.7
OR
5.46
DCR
6.86
• However, the cut-offs points are different by ROCanalysis for each end-point
Prenen: Results (3)
• Combination of K-Ras wt and EREG or AREG and OS
EREG HR OS: 0.42 (95% CI 0.28 – 0.63)
p<.001
Jonker & Prenen: Implications
• Strengths:
– Large series
• One randomized study: 394 pts.
• One multicentric cohort series: 287 pts.
– Not influenced by other treatments
– Proof of concept of AREG & EREG well
established, beyond K-Ras
Jonker & Prenen: Implications
• Weakness:
– Do not discriminate between AREG & EREG
– Underestimate other relevant mutations
– Reproducibility: magnitude and cut-off
– Variability in the categorization and loss of
power
Loupakis: Patients and Methods
Endpoint
Utility
Three tumor based tests: K-Ras and B-Raf
mutation status and AREG expression
Predictive biomarker
Specimen
Tumor specimens (paraffin-embedded)
Patients
Sample size
Refractory mCRC treated with Irinotecan +
Cetuximab
87 tumors (4 centers in Italy)
AREG expression by IHC (Mo Ab cl 31221,
RD) H-Score (0-300)
Mutations K-Ras & B-Raf (not described)
Assay
Loupakis: Results - Implications
• RR:
•
•
•
•
- ITT: 16%
- K-Ras wt: 25%; K-Ras wt + B-Raf wt: 30%
AREG: High expression associated with B-Raf wt
(p=.0005) but not with K-Ras wt
AREG in K-Ras wt and B-Raf wt: no relation with RR, PFS
and OS
In the multivariate analysis only B-Raf status keep the
prognostic value
 Difficult to conciliate with the literature due the low
frequency of B-Raf mut (5-10%)
AREG by IHC not standardized
Polymorphisms
A4022 – Poster Board #: 13; Dongyun Yang et al.
Pharmacogenetic analysis in metastatic colorectal cancer (mCRC) patients
(pts) treated with second-line irinotecan (IR)+/- cetuximab (CB): The EPIC
experience.
Yang: Methods and Results
Endpoint
Two tumor based tests: K-Ras mutation status and
EGFR-CA repeats in Intron 1
Utility
Predictive biomarker
Specimen
Tumor specimens (paraffin-embedded)
Patients
Sample size
Oxaliplatin-refractory mCRC treated with Irinotecan +
Cetuximab
84 pts treated in the US (Ir/Cmab) - EPIC study
Assay
EGFR-CA repeats in Intron 1 (PCR)
Mutations (method not defined)
Yang: Results - Methods
• K-Ras mutation status was not significantly associated with PFS or response
• EGFR-CA- repeat in intron 1 in arm be associated with PFS (p=0.031)
• Results difficult to interpret: few patients in variant 20/20
Yang: Implications
US patients
Ir/Cmab (n=84)
Ir (n=102)
p
RR
13.1
5.9
-
mTTP (m)
3.0
2.7
-
Ir/Cmab (n=84)
Ir (n=102)
p
RR
16.4
4.2
<.05
mTTP (m)
4.0
2.6
<.05
Total
• Behavior of homozygous variants (20/20 & <20/<20) is different to the
heterozygous (20/<20)
• Biology?
• Sample size
Methodology in K-Ras mutations
determination
A4018 – Poster Board #: 9; Andreas Jung et al.
The German quality assurance system for the molecular-pathological
detection of KRAS-mutations in colorectal cancer.
Jung: Patients and Methods
Endpoint
Quality audit of K-Ras mutation status test
Utility
Predictive biomarker
Specimen
Tumor specimens (paraffin-embedded)
Patients
mCRC patients from German-speaking countries
Sample size (Austria, Germany and Switzerland)
10 patients; 50 institutions
Assay
K-Ras mutation analysis by DDS (disesoxy
sequencing – Sanger), ARMS (amplification
refractory mutation sequencing) and MPA
(melting point analysis, pyrosequencing)
Jung: Results - Implications
• 10 patients: 74 different K-Ras determinations
• Limited number of patients and analysis
• The authors raise concerns on the difficulties to
establish quality assurance systems
• The authors state there is no technique/method
superior to another?
• Delay in the result higher than expected (>14 days)
• 15% conflicting results: not disclosed
• SOP: critical step
IGF-1/IGF1R axis in the treatment
with anti-EGFR MoAbs
A4017 – Poster Board #: 8; Mario Scartozzi et al.
Correlation of insulin-like growth factor 1 (IGF-1) expression and clinical
outcome in K-RAS wild-type colorectal cancer patients treated with
cetuximab-irinotecan.
Scartozzi: Methods and Results
Endpoint
Two tumor based tests: K-Ras mutation status and
Insulin-like growth factor (IGF-1) expression
Utility
Predictive biomarker
Specimen
Tumor specimens (paraffin-embedded)
Patients
Sample size
Refractory mCRC treated with Irinotecan + Cetuximab
62 tumors (4 centers in Italy)
Assay
IGF-1 expression by IHC (Cell Signaling)
Mutations (method not defined)
Total 62 pts.
IGF-1 -
IGF-1 +
p
RR PR
7 (50%)
1 (5%)
.004
11
3.2
.03
mTTP (m)
Scartozzi: Results - Implications
• Combined IGF-1 IHC expression and K-Ras mutation analysis
may represent an effective strategy for a better selection of
responding colorectal tumors for cetuximab treatment
• Caveats:
• IHC considered positive if 2
• These results should be externally validated
• Reproducibility of IHC for IGF-1, IGFBPs and IGF-1R is
cumbersome
• Potential role for anti-EGFR and anti-IGF1R combinations:
• Activation of IGF-1/IGF1R reduces sensitivity to EGFR TKI in
cancer cells. IGF-1R inhibition restores sensitivity to EGFR
TKIs1,2
1Jones,
2Guix,
HE et al. Br J Cancer 95;172-180, 2006
M et al. J Clin Invest. 118:2609–2619, 2008
Conclusions
• Each of these studies constitute and
Academic effort to personalize the treatment
in patients with mCRC by tuning the target
population beyond the standard of care (KRas status)
• In order to completely define the ultimate
role of the different predictive factors an
international collaboration is needed
Conclusions
• Predictive factors accepted:
– K-Ras status
• Far advanced:
– B-Raf status
• To be defined:
–
–
–
–
N-Ras, PIK3CA status
Loss of PTEN
Ligands: AREG, EREG
Polymorphisms: EGFR, EGF, Fc receptors (ADCC):
FcgammaRIIa-H131R and FcgammaRIIIa-V158F
– Others
Acknowledgements
• ASCO Program Committee
• Poster presenters for providing their
presentations in a timely fashion
• Eduardo Vilar, MD and Javier Hernández,
PhD for their thoughtful comments
• Audience