Multicenter, randomized phase II trial of GM-CSF (GM) plus ipilimumab (Ipi) versus Ipi alone in metastatic melanoma: E1608.

Melanoma/Skin Cancers
Session Type and Session Title: 
Oral Abstract Session, Melanoma/Skin Cancers
Abstract Number: 
J Clin Oncol 31, 2013 (suppl; abstr CRA9007)
F. Stephen Hodi, Sandra J. Lee, David F. McDermott, Uma N. M. Rao, Lisa H. Butterfield, Ahmad A. Tarhini, Philip D. Leming, Igor Puzanov, John M. Kirkwood, Eastern Cooperative Oncology Group; Dana-Farber Cancer Institute, Boston, MA; Beth Israel Deaconess Medical Center, Boston, MA; University of Pittsburgh Physicians, Pittsburgh, PA; University of Pittsburgh Cancer Institute, Pittsburgh, PA; University of Pittsburgh Medical Center, Pittsburgh, PA; Cincinnati Hematology Oncology, Inc., Cincinnati, OH; Vanderbilt University Medical Center, Nashville, TN

Abstracts that were granted an exception in accordance with ASCO's Conflict of Interest Policy are designated with a caret symbol (^).

Abstract Disclosures


Background: CTLA-4 blockade and GM secreting tumor vaccine combinations demonstrate therapeutic synergy in multiple preclinical models. GM has activity in prostate and ovarian carcinoma and is being evaluated in phase III adjuvant trials for melanoma and lymphoma. GM enhances dendritic cell activation and potentiates antitumor T and B cell responses. GM may induce regulatory immune responses. A key issue is whether systemic GM might synergize with CTLA-4 blockade. Methods: Eligibility: measureable disease, ≤1 prior therapy, no CNS mets, ECOG PS 0-1, > 4 wks prior therapy, adequate end organ function, no autoimmune disease, no prior CTLA-4 blockade/CD137 agonist. OS was primary endpoint. Pts randomized to Arm A Ipi 10 mg/kg q3 wks IV x 4 then q12 wks plus GM 250 μg SC days 1-14 of 21 day cycles vs. Arm B Ipi 10 mg/kg as in Arm A alone. Due to known inflammatory effects of treatments, pts were permitted to continue up to 100% increase in SPD and four new lesions in absence of declining performance status and discretion of treating physician. Drug supply, funding from Sanofi/Bristol-Myers Squibb. Results: 245 pts were enrolled. Arms were balanced for demographics. Median follow up 13.3 mos. RR Arm A 11.3 % (6.1, 18.6), Arm B 14.7% (8.6,22.8) (not significant;NS). PFS Arm A 3 mos (2.9,4.3), Arm B 3.2 mos (3,4) (NS). Median OS Arm A not reached, Arm B 12.6 mos (9.2,-). One year OS for Arm A was 67.9% (59%,76%), Arm B 51.2% (42.6%, 61.3%) (stratified log rank p1=0.016, p2=0.033). HR for mortality on Ipi + GM/Ipi=0.65. Per study design (overall one-sided type I error 0.10), OS interim analysis was conducted at 69% info time. O’Brien-Fleming boundary was crossed for OS. Toxicity assessed for all cases regardless of eligibility. Gr 3-5 AEs 45% Arm A, 57% Arm B (p2=0.078). Gr 5 AEs: Arm A colonic perforation (1), cardiac arrest (1); Arm B multiorgan failure (2), colonic perforation (2), hepatic failure (1), respiratory failure (2). Conclusions: Ipi plus GM significantly improves OS over Ipi alone. No significant differences in toxicity were observed. A trend toward improved tolerability is noted in the GM arm. Clinical trial information: NCT01134614.