1. Abstract 8504: E4697: Phase III Cooperative Group Study of Yeast Derived GM-CSF vs Placebo as Adjuvant Treatment of Patients with Completely Resected Stage III-IV Melanoma DH Lawson, SJ Lee, AA Tarhini, KA Margolin, MS Ernstoff, JM Kirkwood Winship Cancer Institute of Emory University, Dana-Farber Cancer Institute, University of Pittsburgh Cancer Institute, University of Washington, Dartmouth Hitchcock Medical Center

2. Objectives Primary objective: –To compare OS of patients given GM-CSF vs placebo Secondary objectives –To compare DFS of patients with GM-CSF vs placebo –To compare DFS of HLA-A2+ patients given peptide vaccine vs. placebo –To compare OS of HLA-A2+ patients given peptide vaccine vs. placebo

3. GM-CSF+ Peptide Vaccine Placebo+ Peptide Vaccine GM-CSF+ Peptide Placebo Placebo+ Peptide Placebo RANDOMIZE GM-CSF GM-CSF Placebo RANDOMIZE HLA-A2 Positive HLA-A2 Negative Stratify: HLA-A2 Status 1. Positive 2. Negative Site of Metastases 1. Visceral 2. Non-visceral 3. Visceral/Non-visceral Number of Metastases 1. 1 2. 2 - 3 3. 4 or more

4. Eligibility for E4697: Stage III Intransit metastases including local recurrence Gross extracapsular extension Recurrence in previously resected nodal basin Four or more involved nodes Ulcerated primary with any clinically involved nodes Locoregional recurrence after IFN or S0008

5. Eligibility for E4697 Resected locoregional mucosal melanoma Completely resected stage IV melanoma, –Cutaneous, mucosal, ocular, and unknown 1º

6. Treatment Plan GM-CSF 250 mcg SC 14 days of 28 for one year (13 cycles) Patients with resectable recurrences encouraged to continue 6 months past recurrence or for one year, whichever is longer Disease assessments every 3 mos

7. Treatment Plan Vaccinations on days 1 and 15 of first cycle, then day 1 of subsequent cycles. Peptides used: –Tyrosinase: 368-376 (370D) –gp-100: 209-217 (210M) –MART-1: 27-35

8. Statistical Considerations 800 patients gives 80% power to detect a 33% increase in median survival from 40 to 53 months with 2-sided type 1 error rate of 0.05 80% power to detect 24% increase in DFS from 11 to 13.6 months Placebo controlled to enhance validity of the DFS endpoint

9. E4697: Accrual 815 pts accrued between December 29, 1999 and October 31, 2006 –53.5% HLA-A2+ –46.5% HLA-A2- 72 ineligible 743 included in this analysis –368 received GMCSF –375 received GMCSF-placebo

10. Data Analyses Unblinding specified in protocol at 30 months from completion of treatment of the last patient (April, 2009) even if full information not reached by then.

11. Data Analyses September 2009 analysis Positive effect of GM-CSF on DFS (p=.03) with Hazard Ratio (HR) 0.82, 95% CI(.69,.98) No significant effect on OS (p=.55) with HR 0.94, 95% CI (.75, 1.16)

12. Data Analyses Updated analysis for ASCO planned for April 2010

13. Current Status 345 deaths 86% information time Full information anticipated in 2014 Results of April, 2010 analysis presented today

14. DFS by GM-CSF (n=743) P = 0.14 HR = 0.88, 95% CI(.73,1.04) Median DFS Placebo: 9.2 mos, 95% CI (7.68, 12.48) GM-CSF:11.5 mos, 95% CI (9.72, 15.48)

15. OS by GM-CSF Treatment Status P= 0.78 HR= 0.96, 95% CI(.78,1.04) Median OS: Placebo : 62.4 mos, 95% CI (45.6, -) GM-CSF: 69.6 mos, 95% CI (50.4, -)

16. Effect of Peptide Vaccination on Disease Free and Overall Survival in HLA-A2+ Patients

17. DFS by PEP in HLA-A2+ (n=398) P=.709 HR=.93, 95%CI(.73,1.27)

18. OS by PEP in HLA-A2+ (n=398) P =.670 HR=.94, 95%CI(.70,1.26)

19. Subset Analysis of Effect of GMCSF on OS and DFS by Stage Stage IIIB and IIIC: 374 pts Stage IV: 258 pts

20. Stage III Melanoma p = 0.52 HR=.92, 95%CI(.72,1.17) DFSOS p = 0.80 HR=.97, 95%CI(.71,1.31)

21. Stage IV Melanoma DFS OS P= 0.04 HR=.74, 95%CI(.56,.99) P= 0.07 HR=.72, 95%CI(.50,1.02)

22. Conclusions Neither GMCSF nor peptide vaccination achieved OS and DFS objectives There is a suggestion of favorable effect of GMCSF on DFS Subset analysis suggests effects of GMCSF on DFS and OS are largest among Stage IV subjects

23. Conclusions Further study of adjuvant GMCSF is warranted (patients, dose, duration) Use of GMCSF in the adjuvant setting is worthy of discussion (IFN failures, resected stage IV) Laboratory immunologic responses in relation to clinical outcome are under investigation

24. Acknowledgements Coordinating Centers of ECOG, SWOG, and CALGB PI’s, Research Nurses and Coordinators at the sites Carol Hill, RN, PIN nurse As always, our patients and their families