biomarker defined subgroups in gastric cancer, a case study events... · *results based on the...
TRANSCRIPT
Biomarker defined subgroups in gastric cancer, a case study
Nigel Baker
EFSPI, November 30, 2012
Overview
• Background• Gastric cancer• Biomarkers
• Phase 2 trial• Overall• Biomarker subgroup
• Conclusions
• Lessons
2
Gastric cancer
• Second leading cause of cancer-related death
• Poor 5-year survival (~25%) all stages combined
• ~10 month survival for metastatic stage
• Late diagnosis with no standard screening in North America: ~60% regional or metastatic disease at diagnosis
• No single global standard of care chemotherapy: epirubicin, cisplatin, fluoropyrimidine, based regimens most commonly used
3
Stomach - AnatomyStomach - Anatomy
Possible markers to define subgroups
© www.Abcam.com
PIPPIP33PIPPIP22
Gab1Gab1PI3KPI3K
ERKERK
MEKMEKRasRas
Shp2Shp2Gab1Gab1
METMET
ProliferationProliferationMigrationMigrationInvasionInvasion
cdc42cdc42Rac1Rac1
RafRafAKTAKT
SurvivalSurvival
HGF/SFHGF/SF
Rilotumumab (AMG 102)Rilotumumab (AMG 102)
X XX
Possible markers to define subgroups
• Pathway defined potential biomarkers
5
HGF/MET Pathway in Gastric Cancer
• The MET receptor and its ligand, hepatocyte growth factor (HGF), also known as scatter factor (SF), regulate multiple cellular functions, including proliferation, survival, motility, and morphogenesis1
• MET and HGF are important in gastric and esophagogastric junction (G/EGJ) cancer
• MET is expressed in 26% to 74% of gastric cancer (GC) cases, and MET is amplified in 2% to 23% of GC cases1-8
• MET expression is associated with tumor depth of invasion, metastasis, stage, and poor prognosis in patients with GC2,3,6,7,9
• MET expression occurs in 80% to 100% of EGJ cases and is associated with poor outcome10,11
• Biomarkers can potentially be used to identify patients likely to benefit from agents targeting the HGF/MET pathway
• Rilotumumab is an investigational, fully human monoclonal antibody to HGF that blocks the binding of HGF to MET
Key Inclusion Criteria•Pathologically confirmed unresectable locally advanced or metastatic G/EGJ cancer•No prior systemic therapy for locally advanced or metastatic disease
Endpoints•Primary: PFS•Others: OS, ORR, PK, biomarkers
Patient Enrollment/Sites • 121 patients were randomized from 42
sites (from 13 countries)
Data Analyses• Estimation study to evaluate efficacy endpoints
– Each dose of rilotumumab separately– Combined rilotumumab group
• Primary analysis: 79 PFS events, November 30, 2010• Updated analysis:102 PFS events and 74 OS events (~ 60%), April 1, 2011
ARM A (n=40)Rilotumumab (15 mg/kg) + ECX Q3W
ARM B (n=42)Rilotumumab (7.5 mg/kg) + ECX Q3W
ARM C (n=39)Placebo+ ECX Q3W
E: Epirubicin: 50 mg/m2 IV, day 1C: Cisplatin: 60 mg/m2 IV, day 1
X: Capecitabine: 625 mg/m2 BID orally, days 1-21
Stratification factors:ECOG PS 0 vs 1LA vs metastatic
Phase 2 Study of Rilotumumab + Epirubicin, Cisplatin, and Capecitabine (ECX) in G/EGJ cancer
0
20
40
60
80
100
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 Pro
gres
sion
-Fre
e Su
rviv
al (%
)
Time (months)
0
20
40
60
80
100
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17Time (months)
Ove
rall
Surv
ival
(%)
Clinical Efficacy in the Intent-to-Treat Population*
*Results based on the updated analysis with data cutoff of April 1, 2011. †Adjusted for baseline randomization stratification variables (ECOG status [0 or 1] and disease extent [locally advanced or metastatic]).Iveson T, et al. European Multidisciplinary Cancer Congress, September 23-27, 2011, Stockholm, Sweden; abstract #6504.
Median Months(80% CI)
HR†
(80% CI) P Value
Rilotumumab + ECX (n = 82) 5.6 (4.9–6.9) 0.64 (0.48–0.85) 0.045
Placebo + ECX (n = 39) 4.2 (3.7–4.6)
Progression-Free Survival
Overall Survival Median Months(80% CI)
HR†
(80% CI) P Value
Rilotumumab + ECX (n = 82) 11.1 (9.5–12.1) 0.73 (0.53–1.01) 0.215
Placebo + ECX (n = 39) 8.9 (5.7–10.6)
Hypothesis testing requirements:Patient stratification biomarkers• Effective drug or evidence of biologic activity
• Well-defined biologic hypothesis• Knowledge of pathway and target marker• Meaningful biomarker variation that is prevalent in indication of interest
(e.g., KRAS in colorectal vs head and neck cancer)
• Appropriate setting to perform robust test• Adequate sample size with high rate of marker ascertainment
• Depends on both biomarker prevalence and level of efficacy• Control arm to dissect prognostic vs. predictive value of marker
• Validated assay that addresses testable hypothesis• KRAS: limited known mutations, DNA assay• PTEN: huge gene, cytoplasmic and nuclear protein
challenges in assay interpretation and standardization
3
Biomarker Enrichment?
• Primary analysis – encouraging
• Strong historical evidence for a predictive and prognostic biomarker
• Pre-specified analysis plan to find biomarker that defines enriched population
• Aims:• Continue Amgen’s focus on personalised medicine• Strengthen chance of surviving portfolio planning• Strengthen chance of Regulatory success
10
Sample Ascertainment
Patients,* n (%)Patients, n 118
MET immunohistochemistryAcceptable tumor sample available, n (%) 90 (76%)
†METHigh, n (%) 38 (42%)†METLow, n (%) 52 (58%)
Evaluable patients in each treatment arm Rilotumumab + ECX 62 (78%)
Placebo + ECX 28 (74%)
*Per protocol analysis set.
†Criteria chosen for clinical trial assay:METHigh:> 50% of tumor cells with cytoplasmic stainingMETLow: ≤ 50% of tumor cells with cytoplasmic staining
Choosing Cut Point - OS by MET Expression With Increasing Threshold of Percent Positive
• Cox PH modeling suggests increasingly
favorable HR up to a 50% cutpoint.
– Above 50%, no apparent gain is observed.
BiomarkerSubgroup
(% cell staining) HR (95% CI) P valueInteraction
P value
FavorsRilotumumab + ECX
FavorsPlacebo + ECX
0.001 0.01 0.1 1.0 10.0
High (> 50) 0.290 (0.111–0.760) 0.012 0.007Low (≤ 50) 1.838 (0.778–4.343) 0.166High (> 40) 0.501 (0.228–1.098) 0.084 0.051Low (≤ 40) 1.699 (0.669–4.312) 0.265High (> 30) 0.710 (0.327–1.541) 0.387 0.332Low (≤ 30) 1.286 (0.498–3.323) 0.604High (> 20) 0.855 (0.414–1.765) 0.672 0.769Low (≤ 20) 1.027 (0.360–2.933) 0.960High (> 10) 0.847 (0.429–1.672) 0.633 0.563Low (≤ 10) 1.469 (0.406–5.313) 0.557
High (> 90) 0.035 (0.002–0.678) 0.027 0.028Low (≤ 90) 1.209 (0.625–2.337) 0.573High (> 80) 0.166 (0.033–0.823) 0.028 0.010Low (≤ 80) 1.473 (0.700–3.102) 0.308High (> 70) 0.292 (0.099–0.858) 0.025 0.012Low (≤ 70) 1.641 (0.743–3.628) 0.221High (> 60) 0.262 (0.095–0.725) 0.010 0.004Low (≤ 60) 1.879 (0.806–4.381) 0.144
0
20
40
60
80
100
Time (Months)0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17
Ove
rall
Surv
ival
(%)
PFS and OS May be Improved in METHigh
PatientsProgression-Free Survival
Overall Survival
*HR adjusted for baseline disease extent and ECOG PS.
Time (Months)0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
100
80
60
40
20
0
Prog
ress
ion-
Free
Surv
ival
(%) Median Months
(80% CI)HR*
(95% CI) P Value
Rilotumumab + ECX (n = 27) 6.9 (5.1–7.5) 0.51 (0.24–1.10) 0.085
Placebo + ECX (n = 11) 4.6 (3.7–5.2)
Median Months(80% CI)
HR+
(95% CI) P Value
Rilotumumab + ECX (n = 27) 11.1 (9.2–13.3) 0.29 (0.11–0.76) 0.012
Placebo + ECX (n = 11) 5.7 (4.5–10.4)
Distribution of Bootstrap Hazard Ratios
14
0.64
Percentile HR
100 2.43
95 0.64
50 0.26
5 0.07
0 0
Ove
rall
Surv
ival
(%)
0102030405060708090
100
Time (Months)0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17
High Levels of Tumor MET May Be Prognostic and Predictive
• Patients with gastric tumors with high MET expression may have a poorer prognosis but may receive more benefit from rilotumumab
Median Months(80% CI)
Logrank P value
Placebo + ECX (METLow, n = 17) NE (8.5–NE)0.023Placebo + ECX (METHigh, n =
11) 5.7 (4.5–10.4)
0.01 0.1 1.0 10.0 100.0
Favors METHigh Favors METLow
Rilotumumab + ECX 0.677 (0.350, 1.310) 0.2470.007
Placebo + ECX 3.218 (1.076, 9.630) 0.037
TreatmentGroup HR (95% CI) P value
InteractionP valueTreatment Group HR (High vs Low)
(95% CI) P-value Interaction p-value
Rilotumumab + ECX 0.68 (0.35, 1.31) 0.2470.023
Placebo + ECX 3.22 (1.08, 9.63) 0.037
Rilotumumab Safety Profile in the METHigh
Patient Subgroup
METHigh METLow Unselected†
Rilotumumab(n = 27)
Placebo(n = 11)
Rilotumumab(n = 35)
Placebo(n = 17)
Rilotumumab(n = 81)
Placebo(n =39)
Any AE, n (%) 27 (100) 11 (100) 34 (97) 17 (100) 80 (99) 39 (100)
Grade ≥ 3 AE, n (%) 24 (89) 9 (82) 31 (89) 13 (76) 70 (86) 29 (74)
Serious AE, n (%) 16 (59) 7 (64) 22 (63) 6 (35) 47 (58) 19 (49)
Fatal AE, n (%) 0 (0) 2 (18) 8 (23) 2 (12) 9 (11) 6 (15)
†Safety analysis set.
Conclusions
• High tumor MET expression may predict clinical benefit for the addition of rilotumumab to ECX in patients with advanced G/EGJ cancer
• The outcome by MET expression in the placebo arm suggests that high tumor levels of MET is a marker of poor prognosis that may be reversed by rilotumumab
• Rilotumumab has a manageable safety profile among patients with high tumor MET expression
• Based on these results, a phase 3 trial is planned to confirm the efficacy of rilotumumab added to ECX in patients with MET-positive G/EGJ tumors as detected by a MET in vitro diagnostic assay
Lessons
• May increase probability of success• Internal decision making• Predicted Regulatory success
• May make Regulatory buy-in to a Phase 3 design easier• Exclude patients who are less likely to benefit
• Needs careful assessment• Biological justification• Pre-specified analysis plan
18
Acknowledgements
• Kelly Oliner, Amgen for her ASCO 2012 Slides
• Elwyn Loh, Rui (Sammi)Tang,Joe Jiang and Chrissie Fletcher for helpful review comments
19
References
• Slide 3:Jemal et al, CA Cancer J Clin 2011;61: 60-90Jemal et al, CA Cancer J Clin 2010;60;277-300Kang et al, Curr Treat Options Oncol 2011;12:96-106
• Slide 6:
• Slide 9:2008 ODAC briefing book
Altar CA, Amakye D, Bounos D, et al. A prototypical process for creating evidentiary standards for biomarkers and diagnostics.
Clin Pharmacol Ther. 2008;83:368-71.
Hayes DF, Bast RC, Desch CE, et al. Tumor marker utility grading system: a framework to evaluate clinical utility of tumor markers.
J Natl Cancer Inst. 1996;88:1456- 66.
Woolsey R. Drug-diagnostic co-development: a new paradigm. 2008; http://www.cpath.org/Portals/0/PersonMedv2.pdf
Food and Drug Administration. Drug-diagnostic co-development concept paper. 2005; http://www.fda.gov/Cder/genomics/pharmacoconceptfn.pdf.
20
Nakajima M, et al. Cancer. 1999;85:1894-1902.Taniguchi T, et al. Br J Cancer. 1997;75:673-677.Wu C, et al. Oncol Rep. 1998;5:817-822.Graziano F, et al. J Clin Oncol. 2011;29:4789-4795.Tuynaman JB, et al. Br J Cancer. 2008;6:1102-1108.
Birchmeier C, et al. Nat Rev Mol Cell Biol. 2003;4:915-925.Janjigian YY, et al. Cancer Epidemiol Biomarkers Prev. 011;20:1021-1027.Lennerz JK, et al. J Clin Oncol. 2011;29:1-8.Drebber UTA, et al. Oncol Rep. 2008;19:1477-1483.Amemiya H, et al. Oncol. 2002;63:286-296.Kubicka S, et al. Dig Dis Sci. 2002;47:114-121.