Final Efficacy Results of A3671009, a Phase III Study of Tremelimumab vs Chemotherapy (Dacarbazine or Temozolomide) in First-line Patients With Unresectable Antoni Ribas,1 Axel Hauschild,2 Richard Kefford,3 Dmitri Pavlov,4 Bo Huang,4 Margaret A. Marshall4

1University of California Los Angeles, Los Angeles, CA; 2Department of Dermatology, University of Kiel, Kiel, Germany; 3Westmead Institute for Cancer Research and Sydney Melanoma Unit, University of Sydney, NSW, Australia; 4Pfizer Oncology Global Research & Development, New London, CT

Presented at: 25th Annual Meeting of the International Society for Biological Therapy of Cancer (iSBTc); October 2-4, 2010; Washington, DC. Late-breaking abstract 174. Tremelimumab

• Fully human IgG2 developed by that is specific for CTLA4 (CD152) with a plasma half-life of 22.1 days1,2 • Promising activity was observed in phase I and II trials in patients with melanoma

• 1001: first-in-human, single-dose escalation, Patients with phase I clinical trial2 Study Phase melanoma, n – 4 (14%) objective responses A36710012 I29 – All responses lasted ≥ 18 months • 1002: multidose phase I/II clinical trial3 A36710023 I/II 117 – 8 objective responses (9.5%) among 84 evaluable in phase II A36710084 II 246 – 6 patients had responses lasting 15+ months • 1008: multidose phase II clinical trial4 – 16 (6.6%) objective responses – Response duration 8.9 to 29.8 months

Abbreviations: CTLA4, cytotoxic T-lymphocyte 4; Ig, immunoglobulin.

1. Ribas A, et al. Oncologist. 2007;12(7):873-883; 2. Ribas A, et al. J Clin Oncol. 2005;23(35):8968-8977; 3. Camacho LH, et al. J 2 Clin Oncol. 2009;27(7):1075-1081; 4. Kirkwood JM, et al. Clin Cancer Res. 2010;16(3):1042-1048. Tremelimumab Phase III Study A3671009

Objective

• This phase III study was conducted to test the hypothesis that tremelimumab can improve survival in patients with surgically incurable metastatic melanoma

Primary Analysis of Survival

• 537 events would provide 90% power for 2-sided log-rank test at .045 significance when true HR ≥ 1.33 (chemotherapy over tremelimumab)

• 2 equally spaced interim analyses based on O’Brien-Fleming–type boundary were planned when ~1/3 and ~2/3 of events had been observed to stop the clinical trial for futility or to claim efficacy

3 Abbreviation: HR, hazard ratio. Tremelimumab Phase III Study A3671009 Schema and Endpoints

• Primary endpoint was overall survival • Secondary endpoints included best overall response, durable response, duration of tumor response, PFS (at 6 months postrandomization) safety

• First-line therapy R • Melanoma stage IIIc/IV A • LDH < 2× ULN N Tremelimumab (15 mg/kg) IV on day 1 Q90D • No brain metastasis D • No ocular melanoma O • No history of autoimmune M Dacarbazine (1,000 mg/m2) IV on day 1 Q3W disease I • No history of inflammatory OR Z 2 bowel disease E Temozolomide (200 mg/m ) PO on days 1 - 5 Q4W

• Accrual period: March 2006 - July 2007 • Crossed futility boundary at second interim analysis: March 2008

Abbreviations: IV, intravenous; LDH, lactate dehydrogenase; PFS, progression-free survival; PO, orally; 4 ULN, upper limit of normal. Tremelimumab Phase III Study A3671009 Patient Characteristics

Tremelimumab Chemotherapy Randomized patients, n 328 327 Male, % 58 56 White, % 93 93 Mean age, years (range) 57 (22-90) 56 (22-90) Age  65 years, % 34 28 ECOG = 0, % 68 70 Disease stage, % IIIC 6 4 M1a 14 15 M1b 23 21 M1c 57 59 LDH  ULN, % 30 35 Nonmeasurable disease, % 6 6

5 Abbreviation: ECOG, Eastern Cooperative Oncology Group. Tremelimumab Phase III Study A3671009 Kaplan-Meier Estimate of Overall Survivala

1.0 Tremelimumab: mOS = 12.58 Chemotherapy: mOS = 10.71 0.8 HR: 1.14 (> 1 favors treme) P value: 0.127 0.6

0.4 Survival rate

0.2

0 0 10 20 30 40 Survival time, months Patients at risk Tremelimumab 328 189 105 71 20 Chemotherapy 327 167 82 59 12 Abbreviation: mOS, median overall survival. 6 aData from September 2010. Tremelimumab Phase III Study A3671009 Secondary Endpoint: Responses to Therapy and Progression-free Survivala

Tremelimumab Chemotherapy Randomized patients, n 328 327

Complete response (CR),b n (%) 11(3.4) 8(2.4)

Partial response (PR),b n (%) 25 (7.6) 24 (7.3)

Objective response (CR + PR),b n (%) 36 (11.0) 32 (9.8)

95% CI for objective response rateb (%) (7.8, 14.9) (6.8, 13.5)

6-Month progression-free survival (PFS),c % 20.1 18.1

Abbreviation: CI, confidence interval. aBest overall response as confirmed by sponsor. bData from September 2010. 7 cData from May 2010. Tremelimumab Phase III Study A3671009 Secondary Endpoint: Duration of Objective Responsea,b

1.00

0.75 Tremelimumab

0.50

Chemotherapy 0.25 Survival distribution function Survival distribution

0 0 5 10 15 20 25 30 35 40 Survival time, months aDuration of response from time of randomization. 8 bData from September 2010. Exploratory Analysis of Factors Potentially Associated With In favor of chemotherapy In favor of tremelimumab Overall Survival Patients, N Arm A Arm B P value Overall 328 327 1.14 .1267 Disease stage IIIc 19 14 0.77 .5112 1.13 IV M1a, IV M1b 121 119 .4257 IV M1c 188 194 1.16 .1759 Baseline LDH 1.09 ≤ ULN 220 211 .3968 1.27 1 - 2 × ULN 83 87 .1497 Number of disease sites 0.97 18185 .8848 1.21 > 1 244 231 .0523 Patient characteristics Age 1.14 < 65 years 218 237 .2022 1.13 ≥ 65 years 110 90 .4481 Baseline ECOG 1.1 0 222 228 1.26 .3806 1 101 90 .1497 Prior adjuvant therapy 1.08 76 83 1.19 .6665 None 241 236 .0918 Region 0.90 US 62 58 1.20 .6230 Rest of world 266 269 .0568 Lab assays HLA 1.13 A2 145 131 1.14 .3553 Other 159 157 .3107 CRP 1.49 ≤ 1.5 × ULN 164 164 0.86 .0017 > 1.5 × ULN 96 105 .381 Lymphocytes 0.69 < 0.9 × LLN 52 55 1.22 .053 ≥ 0.9 × LLN 217 215 .053

9 Data from May 2010. Tremelimumab Phase III Study A3671009 Survival by Baseline CRP

Subset of patients with CRP ≤ 1.5 × ULN Subset of patients with CRP > 1.5 × ULN

1.0 Median overall survival 1.0 Median overall survival Tremelimumab: 19.1 months Tremelimumab: 6.1 months Chemotherapy: 12.7 months Chemotherapy: 6.1 months 0.8 0.8 Hazard ratio: 1.49 Hazard ratio: 0.88 P value: 0.0017 P value: 0.381 0.6 0.6

0.4 0.4 Survival rate Survival rate

0.2 0.2

0 0 0 10 20 30 40 0 10 20 30 40 Months Months

Abbreviation: CRP, C-reactive protein.

Marshall MA, Ribas A, Huang B. Poster presented at: 46th Annual Meeting of the American Society of Clinical 10 Oncology (ASCO); June 4-8, 2010; Chicago, IL. Abstract 2609. Tremelimumab Phase III Study A3671009 Distribution of Baseline Lymphocyte Count and Objective Tumor Response, by Treatment Arm

11 Data from May 2010. Tremelimumab Phase III Study A3671009 Survival by Baseline Lymphocytes Subset

Patients with Patients with A B 1.00 C1D1 lymphocytes 1.00 C1D1 lymphocytes < 0.9 × ULN ≥ 0.9 × ULN

0.75 0.75

0.50 0.50

Tremelimumab 0.25 Chemotherapy 0.25 Survival distribution function Survival distribution function Survival distribution Chemotherapy Tremelimumab 0 0 0 5 1015202530354045 0 5 1015202530354045 Survival time, months Survival time, months

12 Data from May 2010. Tremelimumab Phase III Study A3671009 Treatment-Related Adverse Events (AEs)a

Tremelimumab, Chemotherapy, Patients n (%) n (%) Evaluable for AEs 325 319 With grade 3 or 4 AEs 110 (33.8) 74 (23.2) With serious adverse events 80 (24.6) 16 (5.0) With grade 5 AEs 6 (1.8) 1 (0.3) Discontinued because of AEs 39 (12.0) 8 (2.5)

13 aData from September 2010. Balancing Inflammation and Immunity in the Tumor Microenvironment

14 Tremelimumab Study A3671009 Conclusions

• Tremelimumab compared with chemotherapy resulted in a nonsignificant (P = 0.127) improvement in survival of patients with metastatic melanoma treated at first line

• The duration of first objective tumor responses to tremelimumab was significantly longer than responses to chemotherapy

• A low baseline CRP and a baseline absolute lymphocyte count in the normal range selected for a patient population with higher tumor response rate and better survival outcome with tremelimumab compared with chemotherapy – This may reflect an interaction between the tumor microenvironment, tumor inflammation, and an adaptive immune response 15 Acknowledgments Pfizer GRD New London J Gomez-Navarro (clinical lead) U Tuebingen Christie Hosp U St. Radbound M Marshall (study lead clinician) C Garbe P Lorigan D Pavlov (statistician) CJA Punt U Kiel Bo Huang (statistician) Mount Vernon Hosp Netherlands Nat’l Cancer Inst A Hauschild P Carlson (study manager) P Nathan JBAG Haanen Charité-U Medicine Berlin M Song (clinical protocol manager) Newcastle Ge’rl Hosp U Trefzer E Plummer

Cross Cancer Inst Maria Sklodowska-Curie Mem Inst M Smyle J Rolski Tom Baker Cancer Centre Great Poland Cancer Center T Cheng A Mackiewicz

Nantes-Hotel Dieu B Dreno Petrov Res Inst V Moiseenko U Barcelona M Marmol Athens U Huntsman Cancer Inst H Gogas W Samlowski Ohio State U K Kendra Inst Toscano Tumori Washington U M Maio Sheba Medical Center G Linette J Schachter U Pittsburgh J Kirkwood Mary Potter Oncology Centre GL Cohen Peter Maccallum Cancer Centre G McArthur Patients, families, and caregivers from 24 countries Steering Committee Members A Ribas (Chair), UCLA Listed are the 24 highest enrolling sites out R Kefford, Westmead Hosp 16 of 114 sites open to patient accrual A Hauschild, U Kiel