Survival of patients considered for participation to contemporary dose-seeking phase trial: Matter of tumour burden, nature of treatment or of dose-levels?

Juliette Bouchet,Carlos Alberto Gomez-Roca,Stephanie Clisant,Alice Levart,Philippe Alexandre Cassier,Nicolas Penel,Isabelle Desmoulins,Veronique Favre,Jean-Pierre Delord,Nicolas Isambert,Peggy Philippe,Fatima-Zhora Meniai

doi:10.1200/jco.2013.31.15_suppl.2549

Abstract

2549 Background: We have analyzed the survival of pts considered for participation to contemporary phase 1 trial. Methods: All consecutive pts having signed the PIS/IC have been analyzed. OS have been measured using Kalan-Meier method. RMS had been calculated, RMS (0 to 3) is sum of the following prognostic factors: LDH>ULN, met. sites>2 and albumin <35 g/L. Comparisons have been done with Log-rank tests and Cox model. Results: OS of the entire cohort was 448 days. 73.4% of pts having been enrolled. Among not enrolled pts, 74.1% of pts received another treatment. The OS was 497, 247 and 110 days, in pts enrolled in phase I trial, in pts not enrolled but receiving another treatment and in non-treated pts (p=0.001). After adjustment to RMS and with pts not enrolled but receiving other treatment as reference, the HR was 0.47 (95-CI:0.34-0.66; p=0.0001) in pts enrolled in phase 1 compared and 3.54 (1.92-6.52; p=0.0001) in non-treated pts. We have then more specifically analyzed the pts enrolled in single-agent dose-escalating phase I. The OS was 894, 272 and 395 days in pts receiving the 2 first dose-levels, in those receiving intermediate dose-levels and those receiving the phase 2-recommended dose, respectively (p=0.001). The OS was 328 in pts receiving molecular targeted agent and 539 in those receiving cytotoxic agents (p=0.004). In a multivariate analysis, the nature of investigational agent and the dose-level were not associated with better outcome. The sole prognostic factor for OS in multivariate analysis was the RMS (0+1 vs 2+3: HR=3.80 [1.76-8.20], p=0.01). Conclusions: Inclusion in phase 1 trial was associated with better outcome in both crude analysis and after adjustment to RMS. Among enrolled pts, in multivariate analysis RMS reflecting the tumor burden was the sole prognostic factor, the nature of the drug and the dose-level were not associated with the outcome.

Full Text