449P Randomized phase II study comparing pathological responses of resected colorectal cancer metastases (CRCM) after bevacizumab (BEV) with FOLFOX or FOLFIRI (BEV-ONCO trial)

Abstract

Pathological response (PR) of resected CRCM after preop treatment is a recognized prognostic factor. Retrospective studies reported that BEV + oxaliplatin-based chemotherapy increased PR compared to irinotecan-based chemotherapy. In this trial, we aim to demonstrate that preop BEV + FOLFOX would increase PR. BEV-ONCO (NCT01858649) is a multicenter prospective randomized (1/1) phase II trial evaluating PR on resected CRCM after 3 to max 6 cycles of mFOLFOX (ARM A) or FOLFIRI (ARM B) + BEV (5mg/kg/2 weeks). Primary endpoint is the major pathological response rate (MPRR) defined as the % of patients presenting CRCMs with a mean tumor regression grade (TRG) <3. Secondary endpoints include DFS, OS, safety, complete PR, R0 resection rate and liver toxicity comprising sinusoidal obstruction syndrome (SOS) and nodular regenerative hyperplasia (NRH). 54 pts (27 per arm) are needed to detect a difference (alpha=0.05; beta=0.2) of MPRR proportion of 0.40 between treatment arms (two-sided Fisher's Exact test). Among 65 pts included between 06/2013 and 09/2018, 57 pts (28 ARM A / 29 ARM B) have had CRCM resection. Clinical and treatment characteristics were similar in both treatment arms (median age 60 y-old, 51% male, 33% RAS wt, 98% liver CRCM, 75% synchronous, median 2 CRCM/pt, median of 4 chemo cycles and 3 BEV cycles). 11/28 pts presented 1-month postop surgical complications in ARM A (39%, grade 3-4: 17.9%) and 9/29 pts in ARM B (31%, grade 3-4: 6.9%, p=0.58). MPRR was 32% in ARM A and 21% in ARM B (p=0.38). 4 pts presented complete PR (ARM A/B: 14%/0%, p=0.05). No difference between treatment arms was observed for R0 resection (ARM A/B: 89%/93%, p=0.80), SOS (ARM A/B: 54%/38%, p=0.50), NRH (ARM A/B: 21%/17%, p=0.75), DFS (ARM A/B: HR=1.14, 95%CI:0.58-2.21, p=0.71) and OS (ARM A/B: HR=1.38, 95%CI:0.48-4.00, p=0.55). Pts with PR among all CRCM (Max TRG≤3; 44% of pts) had a lower risk of relapse/death (DFS: HR=0.41, 95%CI=0.20-0.82, p=0.01) and death (OS: HR=0.34, 95%CI=0.10-1.11, p=0.07). This trial fails to demonstrate any significant difference of PR between BEV with FOLFOX or FOLFIRI but confirms PR as a prognostic factor.