Melanoma in 2009: therapeutic p rogress on two fronts tumor and - - PowerPoint PPT Presentation

Melanoma in 2009: therapeutic progress on two fronts—tumor and

host

John M. Kirkwood, MD

Director, Melanoma and Skin Cancer Program University of Pittsburgh Cancer Institute Professor and Vice Chairman for Clinical Research Departments of Medicine and Dermatology University of Pittsburgh School of Medicine Chair, Melanoma Committee Eastern Cooperative Oncology Group kirkwood@pitt.edu

Systemic therapies for melanoma are needed

• One FDA-approved cytotoxic agent in use

– Dacarbazine (Temozolomide): 6.8-15% response rate in modern trials, rarely durable

• Never compared to observation or supportive care
• Temozolomide no better than dacarbazine (Patel 2008)
• One therapy approved in modern times

– High-dose IL-2: 16% response rate and ~6% durable as tested in collected phase II trials

• No phase III trial previously conducted
• Phase III trial building on IL-2 first reported by

Schwartzentruber et al., ASCO 2009

Atkins et al., JCO 1999 Luikart, Kirkwood, JCO 1984; Hill et al Cancer 1984 (n=580); Middleton et al., JCO 2001;Avril et al., JCO 2004; Patel et al ESMO 2008

What have we learned in the past 35 years of phase II clinical trials?

• Meta-analysis of 42 Phase II trials of SWOG,

ECOG, CALGB, NCCTG, and NCIC-CTG between 1975 and 2005

• Benchmarks for overall survival (OS) , and

progression-free survival (PFS) endpoints sought for future phase II trials

– 1-year OS rate – 6-month PFS rate

New benchmarks for Phase II trials Survival at 1 year; Progression-free Survival at 6 mos

Numbers of patients Numbers of patients One year OS 25% Six month PFS 15% Survival at 1 year = 25% Progression Free Survival at 1 year = 15% Meta-analysis of 70 Phase II Trials in 2100 Patients over past 35 Years

Korn…Kirkwood, J. Clin. Onc, 2008

Current Therapy of Metastatic Melanoma

• Systemic therapy does not alter natural history of

advanced melanoma

– Dacarbazine (Temozolomide) palliative in a minority of patients – High-dose IL-2 cures a small ~6% fraction of patients

• No established mechanistic basis or predictive factors established
• VEGF a suggested potential biomarker of resistance (Sabatini, JCO

2009)

• Gp100 vaccine adds to IL-2 (Schwartzentruber, Proc. ASCO 2009)

– Phase III trials negative to date

• Chemotherapy (dacarbazine) plus antisense Bcl-2 (Genasense) 2006
• Chemobiotherapy (Cisplatin, Vinblastine, DTIC, IL-2, IFN 2008
• Chemotherapy (Paclitaxel) plus elesclomol 2009 Chemotherapy

(Paclitaxel and Carboplatin) +/- Sorafenib E2603

– Pending Phase III results

• Ab-Paclitaxel
• CTLA4 blocking antibodies
• Allovectin B7

Sabatini et al., J Clin Onc 2009 Kirkwood and Tarhini, JCO 2009 Bedekian et al JCO 2007

A Randomized, Double-blind, Phase 3 Trial of STA-4783 (elesclomol) in Combination with Paclitaxel versus Paclitaxel Alone for Treatment of Patients with Stage IV Metastatic Melanoma (SYMMETRY) Review of Preliminary Data

• A. Hauschild, A.M. Eggermont, E. Jacobson, S. O’Day.

University of Kiel, Kiel, Germany; Erasmus University Medical Center, Rotterdam, The Netherlands; Synta Pharmaceuticals, Lexington, MA; The Angeles Clinic and Research Institute, Santa Monica, CA On behalf of the SYMMETRY Trial Investigators

Elesclomol: Mechanism of Action

• Elesclomol is an investigational drug candidate that induces
• xidative stress (reactive oxygen species, ROS)1
• Oxidative stress induction represents a potential novel way of

selectively targeting and killing cancer cells

• Cancer cells produce higher levels of reactive oxygen species

(ROS) than normal cells, making them potentially more susceptible to further oxidative stress and ROS mediated apoptosis2

• Elevation of ROS may facilitate the ability of taxanes to induce

apoptosis through the intrinsic mitochondrial pathway3

• Preclinical in vivo studies demonstrated synergistic efficacy of

paclitaxel and elesclomol in a variety of solid tumor models, including melanoma

1. Kirshner et al. (2008) Molecular Cancer Therapy 7:2319-2327 2. Kong et al. (2000) Medical Hypothesis 55:29-35; Pelicano et al. (2004) Drug Resistance Updates 7:97-110 3. Ramanathan et al. (2005) Cancer Research 65:8455-8460

Cyt C

Elesclomol

Elesclomol-ROS- apoptosis pathway

 Pro-apoptotic factors (Bax) Opening of mitochondrial membrane pores  Anti-apoptotic factors (Bcl-2; XIAP)

Cyt C (1) Direct apoptosis induction

ROS Increase

H2O2, OH, O-

Apoptosis

Caspase cascade

Cyt C

Elesclomol

Elesclomol-ROS- apoptosis pathway

Chemotherapy agents

acting through intrinsic mitochondria apoptosis pathway (taxanes)

(2) Enhancement

• f chemo-driven

apoptosis

Opening of mitochondrial membrane pores

Cyt C

ROS Increase

H2O2, OH, O-

Caspase cascade

Apoptosis

(1) Direct apoptosis induction

 Pro-apoptotic factors (Bax)  Anti-apoptotic factors (Bcl-2; XIAP)

A multi-center, randomized Phase 2 trial of elesclomol in combination with paclitaxel for the treatment of patients with metastatic melanoma

• Double-blind, randomized, controlled; 21 centers in U.S.
• Cross-over for paclitaxel alone arm after PD

Study Population

• Stage IV

metastatic melanoma

• 0-1 prior

chemo for metastatic disease

• ECOG PS 0-2
• No brain mets

paclitaxel 80 mg/m2 + elesclomol 213 mg/m2 (N=53) Randomization 2:1 (N=81) paclitaxel 80 mg/m2 (N=28)

Principal Investigator: Steven O’Day, M.D., The Angeles Clinic and Research Institute (Los Angeles, CA) Ref: O’Day, et al. JCO in press

Primary endpoint: Progression- free survival

qw for 3 weeks; 1 week off

Kaplan-Meier Plot of Progression-Free Survival ITT population

ELPAC (N=53) PAC (N=28) Median (months) 95% CI 3.7 (2.5 – 5.5) 1.8 (1.6 – 3.4) 6 Month PFS Rate 35% 15% HR (ELPAC vs. PAC) 95% CI 0.58 (0.35 – 0.97) p-value 0.035

*Chemo-Naïve subgroup had a significantly higher median PFS which led to the chemo-naïve patient population for the SYMMETRY study

Ref: O’Day, et al. JCO in press

A Randomized, Double-blind, Phase 3 Trial of STA-4783 (elesclomol) in Combination with Paclitaxel versus Paclitaxel Alone for Treatment of Patients with Stage IV Metastatic Melanoma (SYMMETRY) Review of Preliminary Data

SYMMETRY Study design

• 160 centers in 15 countries
• Tumor Assessment: at baseline and every 8 weeks from time of

randomization (RECIST)

• No patient cross-over

Study Population

• Stage IV

metastatic melanoma

• Chemo-naïve
• LDH ≤ 2x ULN
• ECOG PS 0-2
• Absence of

CNS mets paclitaxel 80mg/m2 + elesclomol 213 mg/m2 (N=315) Randomization 1:1 (N=630) Stratification Factors

• LDH status
• M1 subclass
• Prior permitted

(non-chemo) therapy* paclitaxel 80mg/m2 (N=315)

Primary endpoint: Progression- free survival

qw for 3 weeks; 1 week off

* Kinase inhibitor, immunotherapy, biologic therapy, vaccine, or investigational non-chemo Study Steering Committee: Steven O’Day, Axel Hauschild, Alexander Eggermont

Subject Accrual, Randomization, and Treatment

• Accrual period: Sep 2007- Feb

2009

• Peak accrual: June 08-Jan09

Enrolled (N = 651) * N=621 1:1 Randomization ELPAC (n = 309)

Adverse Event (n = 15) PD (n = 157) Withdrew consent (n = 9)

PAC alone (n = 312)

*Data cut April 2009 ITT

ELPAC (n = 304) PAC alone (n = 311) Safety

Withdrew consent (n = 5) PD (n = 186)

Died (n = 1)

Symptomatic Deterioration (n=10)

Died (n=5)

Sponsor Decision (n=105) Sponsor Decision (n=97) Adverse Event (n = 13) Symptomatic Deterioration (n=6)

Other (n=8) Other (n=4) ITT- all randomized Safety- received at least 1 dose

ELPAC PAC Events/Number at risk 170/309 192/312 Median (months) 95% CI 3.4 (2.6 - 3.6) 1.9 (1.9 - 2.9) 6 Month PFS Rate 23% 17% HR (ELPAC vs. PAC) 95% CI 0.84 (0.68 – 1.04) p-value (Stratified Log-Rank) 0.1107

Progression-Free Survival - ITT Population

PAC 312 200 87 42 20 8 2 2 1 ELPAC 309 187 107 53 21 10 6 3 2 2 Subjects at Risk

Overall Survival: Not Yet Mature Currently Favors PAC Arm

ITT Population (N=651) Data Cut % censored HR (CI) p-value Feb 2009 80% 1.62* (1.14 - 2.31) 0.0068 April 2009 72% 1.31 (0.98 - 1.76) 0.0719

* There were 80 OS events in the ELPAC arm vs. 53 OS events in the PAC arm

Pts enrolled as of Sep 1, 2008 (N=300) Data Cut % censored HR (CI) p-value Feb 2009 63% 1.28 (0.88 - 1.87) 0.1930 April 2009 54% 1.22 (0.87 - 1.71) 0.2552

Conclusions

• Despite a trend in improvement of PFS, elesclomol in

combination with paclitaxel (ELPAC) failed to demonstrate a statistically significant improvement when compared with paclitaxel alone in chemo-naive patients with metastatic melanoma

• There was a statistically significant increase in PFS with

ELPAC in the subgroup of patients with normal LDH (68% of the ITT, pre-specified exploratory analysis)

• An imbalance in deaths favoring the paclitaxel arm was
• bserved, leading to early study termination
• No organ-specific toxicities have been identified that explain

the observed imbalance in deaths at this time; safety data is continuing to be evaluated

• At this point OS data is not mature; mature data will be

presented at a future scientific meeting

Prediction of response to alkylator-based chemotherapy in metastatic melanoma (MM) using gene expression and promoter methylation signatures (ASCO #9009)

Hussein Tawbi MD, MS

• H. Tawbi, S. Buch, P. Pancoska, Y. Lin, M. Saul, M. Romkes, R. Sobol, J. Kirkwood

University of Pittsburgh Melanoma and Skin Cancer Program Tawbi et al., Proc ASCO 2009 #9009

Individualized therapy

• Epigenomic/ Genomic/proteomic profiling
• Identify the ―5-7%‖ that will respond
• Identify patients that will not respond
• Identify a proportion of patients with

modifiable epigenomic signature in which epigenetic agents will reverse resistance

Pilot Study

• Archived tumor tissue obtained from

metastatic melanoma patients treated with alkylator-based therapy

• IRB-approved
• Pre-treatment samples
• Response classified:

– Responder: PR and CR and SD>4 m – Non-Responder: PD

Patient Characteristics

Age at diagnosis (years) Mean (Range) 48.3 (30-90) Gender Males 14 Females 7 Chemotherapy Response Non responder (NR) 12 Responder (R) 9 Mean follow up time (months) 20.7 Median survival time (months) Non-Responder 9 Responder 21

Experimental methods

• DNA was extracted using ArchivePure DNA Cell/Tissue Kit

and RNA using PerfectPure RNA Tissue Kit from 5 Prime

• Whole genome gene expression analysis was performed

using the HumanHT-12 Expression BeadChip from Illumina, Inc.

• Whole genome promoter methylation analysis was

performed using the HumanMethylation27 DNA Analysis BeadChip from Illumina, Inc.

• Statistical analyses were performed using BRB array tools,

Maple 12, Weka (Waikato Environment for Knowledge Analysis) software

Classical Analyses

• Predictive analysis of microarray (PAM) analysis of

gene expression data revealed that a classifier set consisting of 82 genes was able to predict NR from R with 83% sensitivity and 89% specificity

• Differential promoter methylation analysis revealed

that 63.6% of promoter sites were hypomethylated in tumors obtained from Responder patients (p<0.0001)

• However, promoter methylation profiling did not

independently correlate with Response phenotype

Methylation-Gene Expression Matrix: Response Phenotype

Methylation Expression

Log10(Expression)

Data Reduction:

Differential Methylation-Expression R vs NR

Methylation

1 = non-responder 0 = responder 1 = responder 0 = non-responder

Neural Network Analysis:

Pattern Recognition for Prediction of Response

Neural Network Analysis:

Weighted impact on the prediction algorithm

Data Reduction: Refining the Model

Correctly classified instances 21 100% Incorrectly classified instances 0%

Final Predictive Model

Final Predictive Model

True positive rate False positive rate Precision ROC area Class 1 1 1 NR 1 1 1 R

Observed Phenotype Responder Non-Responder Predicted Phenotype 9 Responder 12 Non-Responder

Final Predictive Model: Performance

Pathways

Gene Top functions CCDC89 Hematological Disease, Organismal Injury and Abnormalities, Cancer CCDC94 Cellular Development, Lipid Metabolism, Molecular Transport DYSF Genetic Disorder, Skeletal and Muscular Disorders, Cellular Assembly and Organization RASSF4 Cell Signaling, Cell Death, Dermatological Diseases and Conditions SLC8A2 Inflammatory Disease, Genetic Disorder, Skeletal and Muscular Disorders CPN1 Cell Signaling, Molecular Transport, Small Molecule Biochemistry NKG7 Natural killer cell group 7 sequence IRX2 Iroquois homeobox 2

Summary

• Epigenetic regulation is an important element for

molecular profiling

• Integrative analysis requires novel methodology
• Neural network analysis accounts for interaction

between genes and their regulatory regions

• Neural network analysis allows the identification of

molecular patterns highly predictive of phenotype

Conclusions

• Epigenetic regulation of gene expression adds a new

dimension to our understanding of biology and allows for better prediction of a therapeutic response phenotype

• A step towards personalized medicine and patient

selection to maximize benefit and minimize risk

• Identifying a modifiable epigenetic signature allows for

novel combination therapies that have the potential to modulate the response phenotype (UPCI 07-008)

Rational Therapy of Metastatic Melanoma according to Molecular and Immune Lesion(s) of Tumor and Host

• Promise of new agents builds on understanding of

molecular pathogenesis and immunobiology of tumor

– Understand and overcome drug resistance – Target fundamental elements of tumor progression pathway

• BRAF
• C-Kit
• bFGF/FGFR1
• VEGF

– Targeting immunopathology of progression and tolerance

• DC
• T cell (CD8, CD4)
• T regulatory cells
• STAT 3, TGFβ, VEGF, IL-10, IDO and other molecular processes of

local immunosuppression mediated by tumor…

―The Melanomas‖

AFBR PTEN CCND1, KIT KIT, CCND1 KIT, CDK4

Curtin J et al. NEJM 2005

RAS BRAF MEK ERK CRAF

15% mutated 60-70% mutated

PI3K AKT mTor

45% deleted 33% amplified

PTEN

70% deleted

CDK4 CDK2

p16

Cyclin D

10% amplified 5% mutated

MITF

20% amplified c-kit 3% mutated

Pathways of Progression in Melanoma

After KT Flaherty 2009

RAS BRAF MEK ERK

Targeted agents in development

PI3K AKT mTor

PTE N

CDK4 CDK2

p16

Cyclin D MITF

PD0325901 AZD6244 Sorafenib RAF-265 PLX4032 R115777 SCH66336 PD032991 CYC202 temsirolimus everolimus AP23573 GSK690693 VQD-002 BMS-387032 SF1126 XL147 Imatinib, dasatinib TKI-258

After KT Flaherty 2009

Phase I study of PLX4032: Proof of concept for V600E BRAF mutation as a therapeutic target in human cancer ASCO #9000

• K. Flaherty,1 I. Puzanov2, J. Sosman2, K. Kim3, A. Ribas4,
• G. McArthur5, R. Lee6, J. Grippo6, K. Nolop7,
• P. Chapman8

1University of Pennsylvania 5Peter MacCallum Cancer Centre 2Vanderbilt University 6Hoffmann-La Roche Inc 3MD Anderson Cancer Center 7Plexxikon Inc 4University of Califorinia, Los Angeles 8Memorial Sloan-Kettering Cancer

Center

NRAS CRAF BRAF PI3K MEK AKT mTOR CDK2 CDK4 Sorafenib RAF-265 XL281 PLX4032 PD0325901 ARRY-142886 PTEN p16 CyclinD

PLX 4032: new potent selective B- Raf kinase inhibitor

• Activating mutations of B-Raf observed in

multiple solid tumor types

• B-Raf mutations found in 55-68% of

melanoma (with intermittant sun-damage)

• Oncogenic mutations of B-Raf correlate

with poorer prognosis

• PLX 4032 more selectively inhibits

mutated B-Raf kinase

PLX4032: novel, small molecule inhibitor

Selectivity for BRAFV600E in vitro and in vivo

Phospho-ERK IC50 (nM) A375 COLO829 COLO205 SW620 SKMEL2 20 10 30 >40,000 14,000

Selective in cellular assays

V600E WT

Selective regression of V600E tumors Selective for BRAFV600E kinase among 70 kinases screened 1205Lu V600E

+ PLX4032

C8161 WT

+ PLX4032 Tumor Size 100 mm3 10–100 IC50 (nM) 100–1000 1000–10000

PLX4720 co-structure with kinase domain of BRAFV600E (Tsai J et al. 2008 PNAS)

Kinase domain binding

2006 EORTC-NCI-AACR Molecular Targets & Cancer Therapeutics BRAF

PLX4032 safety summary (n=55 patients)

• Nearly all AEs mild and transient up to and including 720 mg BID
• Most frequent drug attributable AEs through 1120 mg BID:

– Gr 1 and 2 rash (n=17); fatigue (n=9); photosensitivity (n=7) – Cutaneous SCC following chronic dosing (n=6)

• Dose-limiting toxicities:

– 1 DLT out of 5 patients at 720 mg BID

• Gr 4 pancytopenia resolved with dose reduction

– 4 DLTs out of 6 patients at 1120 mg BID

• Gr 3 rash ± grade 3 fatigue (n= 3); Gr 3 arthralgia (n= 1)
• 960 mg BID is under evaluation as MTD

BRAFV600E melanoma patient demographics enrolled at > 240 mg BID

Melanoma patients at > 240 mg BID N= 21

• V600E present

16

• V600E absent

5

BRAFV600E melanoma patients at > 240 mg BID (N) AJCC stage Prior treatmentsECOG PS

M1a 4 3 7 M1b 1 1 6 1 9 M1c 11 > 2 7

BRAFV600E melanoma patient PET scan at baseline and day +15 after PLX4032 treatment at 320 mg BID

Pt 45 – MD Anderson

Day 0 Day 15

BRAFV600E melanoma patients treated with PLX4032 > 240 mg BID

% change from baseline (sum of lesion size)

*

100 75 50 25

• 25
• 50
• 75
• 100

(RECIST cutoff for PR, 30%)

Patients (n=15)*

All melanoma patients

• One M1c patient had 55% reduction in target lesions, but PD in
• non-target lesions; died before end C2 (not included above)

Interim PLX4032 phase I Kaplan-Meier plot:

BRAFV600E melanoma patients (≥ 240mg BID)

% progression-free survival Time since first dose (days)

100 90 80 70 60 50 40 30 20 10 60 120 180 240 300 360 420 480 540 600

V600E (n=16)

Data as of 5.13.09

Interim PLX4032 phase I Kaplan-Meier plot:

non- BRAFV600E melanoma patients (≥ 240mg BID)

% progression-free survival Time since first dose (days)

100 90 80 70 60 50 40 30 20 10 60 120 180 240 300 360 420 480 540 600

BRAFV600E absent (n=5)

Conclusions PLX4032 Phase I dose escalation study

• Twice daily administration tolerable up to 720 mg
• 960 mg BID dose under evaluation as MTD
• Pharmacokinetics show limited variability and exposure adequate

to inhibit target in animal models

• Responses observed in V600E+ melanoma patients > 240 mg

BID:

• 9 PRs (7 confirmed 49% - 100%, 2 unconfirmed 31% - 62%)
• Regression of liver, lung and bone lesions
• Symptom improvement in many patients
• Interim PFS ~ 6 months, with many patients still on therapy
• In 3 V600E+ thyroid patients, 1 confirmed PR and 2 SD

Phase II Study of Aflibercept (VEGF Trap) in Recurrent Inoperable Stage III or Stage IV Melanoma of Cutaneous or Ocular

• rigin (ASCO #9028)

Ahmad A. Tarhini, Scott Christensen, Kim Margolin, Paul Frankel, Christopher Ruel, Stergios Moschos, Hussein Tawbi, Janice Shipe- Spotloe, John M. Kirkwood

Support: NCI, Sanofi-aventis, Regeneron, UPCI CTRC University of Pittsburgh, Pittsburgh, PA City of Hope National Medical Center, Duarte CA USC/Norris Comprehensive Cancer Center, Los Angeles University of California, Davis Cancer Center, Sacramento

Introduction

• Aflibercept (VEGF Trap) is a unique fusion protein combining

the Fc portion of human IgG1 with the principal extracellular ligand-binding domains of human VEGFR1 & VEGFR2

• Acts as a high-affinity soluble decoy VEGF receptor and potent

angiogenesis inhibitor

• Aflibercept has highest binding affinity for VEGF described to date.

Dissociation constant 0.5 pM

Study Eligibility

• Adults with recurrent inoperable stage III or stage IV melanoma

– Cutaneous, ocular or mucosal primaries allowed

• No prior systemic chemotherapy or hormonal therapy
• ECOG PS < 2 (Karnofsky > 60%)
• No brain metastases
• Normal organ and marrow function
• If UPCR > 1, a 24-hour urine protein should be obtained and

should be <500 mg

• No clinically significant CVD (MI, CABG or UA within 6 mo), CVA

within 6 mo, uncontrolled HTN, NYHA ≥grade III CHF, serious cardiac arrhythmia requiring medication, clinically significant PVD within 6 mo, VTE within 6 mo

• No evidence of bleeding diathesis or coagulopathy
• Informed consent

Endpoints

• Primary endpoints:

– Overall response rate – 4-month PFS rate

• Secondary endpoints:

– Overall survival – Toxicity – Laboratory correlates

Treatment Plan

• Schedule of administration

– Aflibercept 4 mg/kg I.V. every 2 wks for 8 wks (4 cycles; 1 cycle=14 days) – Response assessment every 8 wks

• Correlative Studies

– Pharmacokinetics – Anti-VEGF Trap Antibody

Trial Design - Statistics

• A 2-stage design was adopted focusing upon RR (RECIST) and

4-month PFS rate

• First stage accrual of 21 patients

– Either promising objective response rate (OR: CR, PR)

• ≥1 OR in the first 21

– Or promising 4-mo PFS rate

• ≥ 8 instances of 4-mo PFS
• Final accrual of 41 is planned

– 87% power to detect OR rate of 15% – 85% power to detect 4-month PFS rate of 50% (median PFS of 4 mo)

• If at least 4 OR (at least 10%), or at least 17 instances of 4-

month PFS (at least 41%), were observed among the 41 enrolled patients, this agent would be considered worthy of further testing in this disease

Trial Design – PFS Rate

• Three recent phase III RCT

– Atkins: E3695, BCT vs. CVD (JCO 2008) – Middleton: TMZ vs. Dacarbazine (JCO 2000) – Avril: Fotemustine vs. Dacarbazine (JCO 2000)

• Average of the median PFS in the control and

treatment arms: 2.3 months

– Goal to improve median PFS from 2.3 mos (external standard) to 4 mos – Corresponds to improvement of the 4-month PFS rate from 30% to 50% based on a constant hazard model

RCT: Randomized Controlled Trial; CVD: Cisplatin/Vinblastine/Dacarbazine; TMZ: Temozolomide; DTIC: Dacarbazine

Demographics & Baseline Characteristics (December, 2008)

Variable

• No. of Patients (N=41)

Age, years 57 (23-83) Sex 15 F, 26 M PS (Karnofsky) 19 (100), 14 (90), 8 (80) Primary cutaneous 28 Primary ocular 10 Unknown primary 3 Prior therapies Prior drug therapy IFN (adjuvant) IFN, anti-CTLA4 Melphalan (Isolated Limb Perfusion) CRO11-VCMMAE 12 9 1 1 1 Radiation 11 Classification M1a 3 M1b 8 M1c 28 IIIC (N3) 2

Summary of Adverse Events by Severity (Possible, Probable and Definite; N = 41 patients)

Hematologic Type All Grades Grade 3/4

• No. Pts.

%

• No. Pts.

% Anemia 3 7 Lymphopenia 5 12 Thrombocytopenia 1 2 1 2 Non-Hematologic Type Cardiovascular Hypertension 26 63 8 20 Hypotension 1 2 1 2 LV Diastolic Dysfunction 1 2 1 2 Constitutional Insomnia 2 5 Fatigue 16 39 1 2 Weight Loss 4 10

Efficacy Summary: Patients with 4 Months PFS (N=41)

First 21 Patients** Primary Classificatio n PFS (Days) 1.Ocular M1c 382 2.Cutaneous M1b 414+ 3.Ocular M1c 280 4.Ocular M1c 151+ 5.Cutaneous M1c 186 6.Ocular M1c 185 7.Cutaneous M1c 177 8.Ocular M1c 174 9.Cutaneous M1b 343+ 10.Unknown M1a 148 Next 20 Patients*** Primary Classificat ion PFS (Days) 11.Cutaneous M1c 293+ 12.Cutaneous M1a 329+ 13.Cutaneous M1c 231 14.Cutaneous* M1c 125+ 15.Cutaneous M1b 126+ 16.Cutaneous* N3 113+

*Imaging studies confirming SD done few days short of 4 months time point ** Excluding one patient who is NED, but considered non-evaluable for efficacy ***8 additional patients continue on treatment, but have not reached the 4 month PFS milestone

Efficacy Summary: Responses (N=41)**

No. Pts (%) Primary

• No. Pts (%)

Classification

• No. Pts (%)

Cutaneous Ocular Unknown M1a, N3 M1b M1c RR* 1 (2) 1 (100) 1 (100) SD 20 (49) 13 (65) 5 (25) 2 (10) 4 (20) 5 (25) 11 (55) PD 13 (32) 9 (69) 3 (23) 1 (8) 1 (8) 2 (15) 10 (77) Not Eval 5 (12) Too Early 2 (5)

Conclusions

• Aflibercept has promising clinical activity in metastatic

melanoma of cutaneous or ocular origin

• It is highly likely that we will meet the 4-month

median PFS milestone (17 instances of 4-month PFS with longer follow up)

• Serious adverse events noted with this class of drugs

such as hypertension require close follow up and monitoring of patients

• Oral AEs observed and should prompt oral/dental

evaluation prior to and during therapy with particular vigilance in patients with a history of radiation therapy to the head and neck, and those with poor dentition

Avenues of Immunotherapy for Melanoma

• IFNs augment

effector cell numbers and repolarize host response to tumor

• IFNs inhibit tumor

proliferation, angiogenesis, and STAT3

Th0 Th1 Th2 Tr/Th3 IL-12 IL-18 IFN-γ IL-4 IL-5 IL-10 IL-10 TGF-β

Cell-mediated antitumor immunity Humoral antitumor immunity

DC1 DC2

iDC

GM-CSF IFN- Anti- CTLA4

Vaccines IL-2 Adoptive T cell, DC Therapy 65

0.0 0.2 0.4 0.6 0.8 1.0 2 4 6 8 10 12 14 16 Time (Years) Proportion Alive and Relapse Free

E1684: IFN vs Observation**

0.0 0.2 0.4 0.6 0.8 1.0 1 2 3 4 5 6 7 8 9 10 Time (Years) Proportion Alive and Relapse Free

0-2 2-4 4-6 6-8 8-10 Observ. 105/212 16/94 5/72 2/44 0/13 IFN 98/215 15/108 5/85 2/53 0/20

E1690: IFN vs Observation*

0.0 0.2 0.4 0.6 0.8 1.0 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 Time (Years) Proportion Alive and Relapse Free

0-1 1-2 2-3 3-4 4-5 IFN 118/436 28/257 8/123 3/47 0/3 GMK 153/439 40/240 6/113 3/40 0/0

E1694: IFN vs GMK**

E1684, E1690, and E1694: Durable and Significant Impact upon Relapse-free * and Overall Survival**

Time Interval (Years) 0-2 2-4 4-6 6-8 8-10 10-12 12-14 14-16 Observ. 89/140 12/51 3/39 0/35 1/32 1/29 0/15 0/3 IFN 73/146 14/68 3/53 1/50 2/48 2/44 0/31 0/10 (No. events/No. at risk)

• Kirkwood. Clin Cancer Res. 2004;10:1670; Ives et al., 2008

Time Interval (Years) (No. events/No. at risk) Time Interval (Years) (No. events/No. at risk)

HR=1.38 P2=0.02 HR=1.24 P2=0.09 HR=1.33 P2=0.006

Meta-analysis of all trials of IFN also confirms RFS, OS impact

Autoimmunity is Predictive of IFN Adjuvant Therapy Benefit

• Subset analysis of phase III He13/97 trial
• 200 Patients with stage IIB/IIIC melanoma

Stage IIB/IIIC melanoma

Arm 1: IFN-α2b 15 MU/m2 5 x weekly for 4 wks, then observation (N=96) Arm 2: Same as Arm 1 + IFN-2b 10 MU 3 x weekly for 48 wks (N=104)

Manifestation of Autoimmunity

N % Antinuclear Antibodies 12 6 Anticardiolipin Antibodies 10 5 Vitiligo 11 6 Clinical Manifestations 19 10 Multiple Manifestations 16 8

Gogas et al., N Engl J Med. 2006;354:709.

Multivariate Analysis for OS in High-Risk Melanoma Patients Receiving HDI

Positive Autoimmunity Status RFS OS HR (95% CI) P HR (95% CI) P At 3 mo 0.15 (0.06-0.37) <0.001 0.07 (0.02-0.28) <0.001 At 12 mo 0.08 (0.03-0.22) <0.001 0.02 (<0.01-0.15) <0.001

• Gogas. N Engl J Med. 2006;354:709.

TTP

Months

100 80 60 40 20 1,0 ,5 0,0

OS

Months

100 80 60 40 20 1,0 ,5 0,0

Patients with autoimmunity (N=52) Patients with autoimmunity (N=52) Patients without autoimmunity (N=148) Patients without autoimmunity (N=148)

Time to Progression OS

Baseline Pro-Inflammatory Cytokine Levels Predict 5-Year Relapse-Free Survival with High-Dose IFNα but not GMK Vaccine

• High-throughput xMAP multiplex immunobead assay

– Tested 29 analytes: cytokines, chemokines, angiogenic factors, growth factors, and soluble receptors – Serum of 378 matched healthy subjects vs 179 patients with melanoma from ECOG E1694

• Phase III trial of HDI vs ganglioside vaccine in resected, high-risk,

cutaneous melanoma

• Serum concentrations of many markers were found to

be higher in patients with resected, high-risk melanoma than in healthy individuals

• Yurkovetsky. Clin Cancer Res. 2007;13:2422; Kirkwood. J Clin Oncol. 2001;19:2370.

E1694

High-dose IFN-2b x 52 wks GMK x 96 wks Randomized N=774 Surgery

Baseline Pro-Inflammatory Cytokine Levels Predict 5-Year Relapse-Free Survival with High-Dose IFNα but not GMK Vaccine

500 1000 1500

*

IL-1b, pg/mL

1000 2000 3000 4000

*

IL-6, pg/mL

1000 2000 3000

*

TNF-α pg/mL

Patients Receiving HDI

250 500 750 1000

IL-1 b , pg/mL

500 1000 1500 2000 2500

IL-6, pg/mL

500 1000 1500 2000 2500 3000 3500

Patients receiving GMK Vaccine

TNF-α pg/mL

Yurkovetsky, Kirkwood, Edington, Clin Cancer Res. 2007 Years <1 1-5 >5 <1 1-5 >5 <1 1-5 >5 Years <1 1-5 >5 <1 1-5 >5 <1 1-5 >5

Dissection of the role of IFN induction Dose, route, and duration of IFN therapy

• All trials of IFN with durable RFS and

OS impact utilized IV induction at 20MU/m2

(Cmax >10,000u/ml)

• Is one month of IV IFN2b necessary

and sufficient?

– Intergroup E1697 – Neoadjuvant Trial UPCI 00-008 – Hellenic Oncology Group Trial 13A/97 – Italian Melanoma Intergroup and DeCOG trials

E1697 - A randomized study of four weeks of high-dose interferon alpha-2b in stage T3-T4 or N1 (microscopic) melanoma

STRATIFICATION Pathologic Lymph Node Status Known Unkown Lymph Node Staging Procedure Sentinel Lymph Node Procedure Elective Lymph Node Dissection No Lymphadenectomy Breslow Depth 1.5 - 3 mm 3.1 - 4 mm > 4 mm Ulceration of Primary Lesion Yes No Disease Stage Lymph Node Positive Lymph Node Negative

R A N D O M I Z E

Arm A: Observation Arm B: 4 week high-dose IFN alfa-2b (Intron A) 20 MU/m2/d qd IV for 5 consecutive days out of 7 (M-F) every week times 4 weeks

Hypothesis: Induction IV IFN is necessary and sufficient to achieve durable adjuvant benefit in intermediate-risk melanoma patients

Pittsburgh Skin Cancer SPORE Project 1: Analysis of genetic and proteomic predictors of IFN benefit

Current Controversies

• Is One Month ~ One Year HD IFN?

– Gogas J. Clin. Oncology 2009

• Nonstandard month at 15Mu/M2/dose (n=182)
• Nonstandard year at 60% of FDA Approved

dosage (n=182)

• Study would have termed non-inferior a

treatment with 15% lower RFS/OS

• No control—and no evidence that either arm

was active – US Cooperative Group Proposal to test equivalence of 1 month and 1 year (5% threshold) required 3000 pts (1991) and never commenced

Current Controversies (2)

• Is Longer Therapy Needed?

– DeCOG 5 year vs. 18 month IFN Trial

• No difference in RFS or OS with 5 years compared to 18 months of

lower-dose IFN regimen (DeCOG)

– EORTC 18952 trial of intermediate dosages over 2 year vs 1 year show non-significant difference in favor of 2 year regimen – Nordic IFN trial of 2 years vs 1 year show nonsignificant difference in favor of 1 year regimen – EORTC 18991 PEG-IFN neither improves OS nor DMFS overall

• RFS difference of 16% at 5 years is significant in

microscopic nodal disease subset

Eggermont et al Lancet Oncology 2008 Eggermont et al JCO 2007 Kirkwood and Tarhini: Nature Clinical Practice Oncology 2009

Controversies (3): What is the Meaning

• f Ulceration in Primary Melanoma?

Ulceration and Responsiveness to Adjuvant Interferon Therapy:

analysis of EORTC18952 and EORTC18991 in 2644 patients

Alexander M. M. Eggermont, Stefan Suciu, Sandro Testori, Poulam Patel, Alain Spatz

Proc ASCO 2009: EORTC Melanoma Group Study

1 2 3 4 5 6

10 20 30 40 50 60 70 80 90 100

O N 124 437 55 233

No Ulceration

HR = 1.11 95% CI (0.80 , 1.56) Stratified for trial: p=0.53 (yrs) 1 2 3 4 5 6

10 20 30 40 50 60 70 80 90 100

O N 128 333 80 151

Ulceration

HR = 0.58 95% CI (0.43 , 0.78) Stratified for trial: p=0.0003

Survival by treatment according to ulceration in Stage IIb & III (N1)

(PEG)-IFN Obs (PEG)-IFN Obs

• Conclusions:

– Ulceration of primary appears a major predictor of clinical benefit from IFN – IPD Meta-analysis of Ives, Wheatley (2005) first suggested this – May increase impetus for application of IFN in ulcerated lesions – Not yet ready to be adopted for standard clinical decision-making

• No central review of EORTC pathology
• Incomplete samples and selection bias possible
• Not supported by ECOG-US Intergroup experience with HDI
• Limitations:

– Unclear whether applicable to FDA-approved dose/ schedule – Retrospective analysis of ECOG adjuvant IFN datasets show no

• bvious association where there is 100% central path review

– Validation is required in prospective study with rigorous Pathology evaluation

Antitumor Vaccines—The Holy Grail Multiple vaccine approaches are currently under study that target molecular mechanisms in immune function and tumor-host interactions

Stimulate antitumor immune response Tumor-derived proteins DNA-encoding tumor antigens Whole cells (autologous or allogeneic) Tumor-derived peptides

A phase III multi-institutional randomized study of immunization with the gp100: 209-217(210M) peptide followed by high dose IL-2 compared with high dose IL-2 alone in patients with metastatic melanoma

• Prior clinical data:

– NCI phase II study: OR in 13/31 (42%); Rosenberg et al. Nature Med 1998 – CWG phase II trials: OR in 20/121 (16%); Sosman et al. JCO 2008 – Single institution experience: OR in >30%; Sullivan et al., ASCO 2009

• Multicenter phase III trial of IL-2 +/- gp100

– focus immune response? – Patient Selection: HLA A2 positive only

• 185 patients

Endpoint IL-2 IL-2+peptides Significance 1 ORR 9.7% 22.1% P = 0.022 (favoring IL-2+pep) 2 PFS 1.6 mo 2.9 mo P = 0.01 (favoring IL-2+pep) OS 12.8 mo 17.6 mo P = 0.0968 (favoring IL-2+pep) Schwartzentruber et al., Proc ASCO 2009 # 9011

Vaccination with the Gp 100 peptide improves the outcome of IL-2

PFS

Hallmark of immunotherapy: Very few relapses beyond 2 years

Schwartzentruber et al., Proc ASCO 2009 # 9011

Peptide vaccine & IL-2 therapy

• Despite reaching primary endpoint of improved response

rate, impact on standard practice unclear

• Selection of patients for IL-2 therapy according to VEGF

levels, IL-6, CRP have the potential to improve therapeutic index of IL-2 by avoiding treatment of nonresponsive patients

– Sabatini JCO 2009

• Vaccine results support concept that immune response

can be focused and that combinations of vaccines with

• ther immunotherapies are reasonable

– IFN – IL-2 – Anti-CTLA 4 blocking antibodies

Schwartzentruber et al., Proc ASCO 2009 # 9011

Overcoming immunoregulation in melanoma: CTLA4 immune checkpoint as target of therapy

Cytotoxic T lymphocyte-associated antigen 4 (CTLA4) blockade prevents down-regulation of T cell activity

CTLA-4 Blocking mAbs in Phase I-III Study

Antibody Name Former Names Type of Antibody Ig Subtype Plasma Half-life Ipilimumab MDX010 BMS-734,016 Fully human IgG1 12-14 days Tremelimumab CP-675,206 ticilimumab Fully human IgG2 22 days

• Ribas. J Clin Oncol. 2005; Benjamini. Immunology: A Short Course. 3rd ed. New York,

NY: Wiley-Liss, Inc. 1996; Paul, ed. Fundamental Immunology. 3rd ed. New York, NY: Raven Press, Ltd. 1993; Korman. Adv Immunol. 2006;90:297.

IgG1 IgG2 IgG3 IgG4 Antibody-dependent Cellular Cytotoxicity +++ ± +++ + Complement Fixation ++ + +++ – Plasma Half-life 23 days 23 days 9 days 23 days

Baseline And On-Study Tumor Biopsy Markers Are Associated With Clinical Activity In Patients With Advanced Melanoma Treated With Ipilimumab (CA184-004 ASCO # 9008)

Omid Hamid,1 Scott D. Chasalow,2 Zenta Tsuchihashi,2 Suresh Alaparthy, 2 Susan Galbraith,2 and David Berman2

1 The Angeles Clinic and Research Institute, Santa Monica

CA, USA

2 Bristol-Myers Squibb, Princeton, NJ, USA

Email: ohamid@theangelesclinic.org

Study Design: CA184-004

Q3W = every 3 weeks, Q12W = every 12 weeks; W= week

• Between the two cohorts, there were no meaningful differences in patient

demographics, efficacy, or safety

10 mg/kg Q12W 10 mg/kg (n=42) Q3W x4

W12 W24

Induction Maintenance

W48+

Screening Follow-up Advanced melanoma

W1 n=82

Treatment naïve (n= 20) Pretreated (n= 62) 3 mg/kg Q12W 3 mg/kg (n=40) Q3W x4

Ipilimumab Ipilimumab

Ipilimumab Dosing:

W1 W4 W7 W10 W24 W36 W48+

Sampling For Biomarker Analyses

Assessment Baseline W4 W7 W10 W12 Tumor biopsy*   Peripheral blood ALC      Influ/Pneumo vaccine booster†  Viral/Bacterial antibody titers  

ALC= absolute lymphocyte count; W = week; Influ/Pneumo= Influenza/Pneumococcus * Tumor biopsy was performed at baseline and 24-72 hours after the 2nd dose of ipilimumab

† Influenza/pneumococcal booster given 5 days after 1st dose of ipilimumab

Tumor Biopsy Evaluation

• Tumor biopsies were obtained pretreatment and

at W4

• 14 candidate biomarkers were evaluated by

central, independent pathologist

– Expression of 7 proteins by IHC using a 9-point scale – 7 tumor characteristics, including TILs, by H&E staining, using a 3-point scale

• Clinical activity (based on mWHO criteria)

defined as:

– CR or PR or – SD lasting at least 24 W from 1st dose of ipilimumab

W= week; IHC= immunohistochemistry; TILs= tumor infiltrating lymphocytes; H&E= hematoxylin and eosin

Representative Photomicrograph Of T-cell FoxP3 Staining In Tumor At Baseline

• Marker of regulatory T-cells and an immunosuppressive environment
• Presence in a tumor may indicate increased sensitivity to immunotherapy

Gajewski TF, et al. J Immunother 2006;29:233-240 Ribas A, et al. Clin Can Res 2009;15:390-399

Representative Photomicrograph Of Tumor IDO Staining At Baseline

• IDO (indoleamine 2,3-dioxygenase) is produced by tumor cells
• Contributes to an immunosuppressive tumor microenvironment

Brandacher G, et al. Clin Cancer Res 2006;12:1144-1151 Okamoto A, et al. Clin Cancer Res 2005;11:6030-6039

High Levels Of Tumor-infiltrating Lymphocytes (TILs) 4 Weeks After 1st Dose of Ipilimumab

• TILs represent an immune response to tumors
• Presence of TILs is a favorable prognostic indicator in many cancers

Dunn GP, et al. Annu Rev Immunol 2004;22:329-360 Rosenberg SA, et al. Nat Rev Cancer 2008;8:299-308

Association Of Clinical Activity With Tumor Biomarkers

Biomarkera Clinical Activity

(CR, PR or SD ≥ 24 Weeks)

Without Clinical Activity W4 TILs change from baseline (N=27) (P=0.005)b 4/7 (57%) increase 2/20 (10%) increase FoxP3 expression at baseline (N=33) (P=0.014)b 6/8 (75%) positive 9/25 (36%) positive IDO expression at baseline (N=35) (P=0.012)b 3/8 (38%) positive 3/27 (11%) positive

a Not all samples were evaluable for every parameter, and not all patients provided data for

both time points

b P values uncorrected for multiple testing

N=total number of patients with data suitable for the given analysis

mRNA Expression In Tumor Biopsy Had Significant Changes From Baseline*

• Significantly increased expression from baseline in various immune-

response genes: – immunoglobulins – granzyme B (GZMB), perforin 1 (PRF1), granulysin (GNLY) – CD8 alpha subunit (CD8A), T cell receptor alpha and beta subunits – 2 of the interferon-gamma-inducible chemokines (CXCL9 and CXCL10)

• Significantly decreased gene expression from baseline in known

melanoma antigens: – tyrosinase-related protein 2 (DCT) – gp100 (SILV) – tyrosinase-related protein 1 (TYRP1) – tyrosinase (TYR) – Melan-A (MLANA)

* 466 mRNA probe sets; corrected for multiple testing; p-value <0.05

No Association With Clinical Activity For Other Candidate Biomarkers

• By H&E analysis of tumor biopsy

– Normal tissue, viable tumor, necrotic tumor, fibrotic regression, total infiltrate, and peri-tumoral immune cells

• By IHC analyses of tumor biopsy

– Granzyme B, Perforin, CD4, CD8, CD45RO

• Genotyping of patients at baseline

– 22 polymorphisms in immune-related genes – Presence of allele HLA-2A*201 at locus HLA-A – Medium-resolution HLA genotypes

H&E= hematoxylin and eosin; IHC= immunohistochemistry

Association Between Change In Absolute Lymphocyte Count (ALC) During Induction And Clinical Activity

• Greater mean increase in ALC over time in patients with clinical

activity than in those without

Dose (mg/kg) Response Evaluable Patients, n Mean change in ALC (1000 cells/μL/week) 3 Clinical activity No clinical activity 6 21 0.030 −0.019 10 Clinical activity No clinical activity 6 23 0.153 0.030

Association Between Change In Absolute Lymphocyte Count (ALC) During Induction And Clinical Activity (Cont.)

In an independent cross-study analysis (n=379) from 3 Phase II ipilimumab studies in melanoma NPV= 100% (Berman et al. ASCO 2009, Abstract 3020) Negative predictive value: 95% Positive predictive value: 30%

Number of Patients (%), N=56 Clinical Activity No Yes Total ALC Decrease 18 (95) 1 (5) 19 Increase 26 (70) 11 (30) 37

Conclusions

• Patients with an immunosuppressive tumor microenvironment at

baseline may be more sensitive to immunotherapy – Expression of FoxP3 and IDO at baseline and increase from baseline of TILs at W4 may predict ipilimumab clinical activity

• mRNA analysis of tumor biopsies consistent with increased TILS
• Change in ALC may be a biomarker of clinical activity

– Decrease in ALC during induction may predict lack of clinical activity

• Ipilimumab treatment led to increased humoral immune response in

an antigen-specific fashion

• Current data support the need for further exploration of the identified

biomarkers as predictors of clinical activity with immunotherapy

Conclusions

• Induction of de novo immune responses to tumor

antigens and autoimmunity are common features of effective IL-2, IFN, CTLA4 blocking antibody therapy

– Genetic basis and novel antigens to be identified

• Combinations of IL-2 and vaccines of interest

(Schwartzentruber Proc. ASCO 2009)

• Phase III results of ipilimumab study awaited; Adjuvant

trials in progress vs placebo (EORTC 18971) and pending in US (E1609)

• New costimulatory options (anti-CD137, anti-CD40L

and TLR agonists [CpG]), PD-1 source of hope

• More precisely targeted agents for BRAF (PLX4032) are

active and proof of concept for this class of agents; more mature data are needed, and these agents may need to be combined with other agents for durable impact