Randomized phase III trial of intravenous (IV) versus hepatic intra-arterial (HIA) fotemustine in patients with liver metastases from uveal melanoma: Final results of the EORTC 18021 study.

Abstract

8532 Background: HIA fotemustine has shown promising results in Phase II studies that led to the EORTC randomized phase III trial (18021) in unpretreated patients (pts) with liver metastases from uveal melanoma. Methods: The treatment consisted in an induction cycle of either HIA (fotemustine 100 mg/m² over 4 hours, day 1, 8, 15, 22) vs IV control arm (fotemustine 100 mg/m² over 1 hour, day 1, 8, 15). After a 5-week break, maintenance cycles were given every 3 weeks. Randomization was stratified by PS (0 vs 1), LDH (normal vs abnormal) and center. Main endpoint was overall survival (OS). Required accrual per protocol was set to 262 pts, with final analysis planned after 220 deaths (hazard ratio (HR) =0.67, power=85%, 1-sided α=2.5%). Due to poor accrual an interim analysis was done after 134 deaths, in order to test futility (power=79%). Results: Between Feb-2005- Feb-2011, 171 pts were randomized (HIA: 86, IV: 85). Characteristics: PS 1: 20%, abnormal LDH: 42%, male: 50%, median age: 59 y.; balanced between arms. In the HIA arm 20 (23%) pts never started treatment mainly due to catheter problems and 2 pts in the IV arm. In those who started the treatment, leucopenia grade 3-4 was 18% and thrombopenia grade 3-4: 21% in the HIA arm compared to 32% and 42% in the IV arm. Non-hematological grade 3-4 toxicities were minimal (GI toxicity, catheter complications). In May 2011, as the OS HR=1.097 was > critical value 0.87, the IDMC recommended stopping accrual for futility. The final results from Jan-2012 are presented in the table below. Treatment comparison adjusted by PS and LDH provided similar results. Conclusions: Even if HIA fotemustine administration could not start in 23% of pts, it led to a higher ORR and longer PFS compared to IV administration. HIA did not translate into an improvement in OS. [Table: see text]