Phase III, open-label, randomized, comparative study of tremelimumab (CP-675,206) and chemotherapy (temozolomide [TMZ] or dacarbazine [DTIC]) in patients with advanced melanoma

A. Ribas, A. Hauschild, R. Kefford, C. J. Punt, J. B. Haanen, M. Marmol, C. Garbe, J. Gomez-Navarro, D. Pavlov, M. Marshall

Research output: Contribution to journalMeeting abstract

Abstract

Background: Tremelimumab, a fully human anti-cytotoxic T lymphocyte-associated antigen 4 (CTLA4) monoclonal antibody was tested in a phase III study to compare the overall survival (OS) with standard, single-agent chemotherapy. Methods: Pts with unresectable stage IIIc-IV melanoma without brain metastasis, LDH below 2x ULN, and no prior systemic treatment for advanced melanoma were randomized 1:1 to either tremelimumab 15 mg/kg IV q90d, or physician's choice of TMZ 200 mg/m² p.o. d1–5 q28d or DTIC 1,000 mg/m² IV q21d (chemotherapy arm). Primary endpoint was OS, secondary endpoints included response, durable tumor response, 6-mo PFS, and safety. 537 deaths would provide 90% power to detect a 33% improvement in true median OS with an unstratified log-rank test at overall 2-sided significance level of 0.045. Two equally spaced interim analyses were planned based on the group sequential design using the Lan-DeMets alpha and beta spending approach to an O'Brien-Fleming boundary. Results: 655 pts were enrolled between March 2006 and July 2007, 328 pts randomized to tremelimumab (324 treated), 327 to chemotherapy (319 treated). There were no significant imbalances in age, sex, LDH, or stage (5% stage IIIc, 15% M1a, 22% M1b, 58% M1c). Diarrhea (43% overall, 14% grade 3/4), pruritus (25%) and rash (23%) were the most common treatment-related adverse events in the tremelimumab arm. 3% of pts had pituitary or adrenal gland toxicities, and 4% thyroid. There were 3 treatment-related deaths in the tremelimumab arm and none in the chemotherapy arm. Based on a protocol-specified second interim analysis with 340 deaths, the independent DSMC advised to stop the study on 3/28/08 since the Logrank test-statistic (P-value of 0.729) had crossed the O'Brien-Fleming futility boundary. Median OS by intent-to-treat was 11.8 mo (95% CI 10.4, 13.9) in the tremelimumab arm, and 10.7 mo (95% CI 9.3, 12.0) in the chemotherapy arm, with a HR (chemotherapy over tremelimumab) of 1.04 (95% CI 0.84, 1.28). Conclusions: Tremelimumab as a single agent failed to demonstrate an improvement in OS as a first line treatment in patients with metastatic melanoma when compared with standard chemotherapy. Analysis of secondary endpoints may provide further information on the role of tremelimumab in the management of metastatic melanoma.
Original languageEnglish
Pages (from-to)LBA9011-LBA9011
Number of pages1
JournalJournal of Clinical Oncology
Volume26
Issue number15 supplement
DOIs
Publication statusPublished - 20 May 2008

Cite this