Background: FOLFOXIRI and bevacizumab achieves high objective response rate (ORR)/R0 resections in pts with unresectable CLM. Bevacizumab increases pathological response, a surrogate endpoint of survival after CLM resection. COI-B regimen is feasible in pts with advanced colorectal cancer. In this phase II study, we aimed at assessing safety and efficacy of perioperative COI-B for pts with borderline resectable or high recurrence risk CLM. Methods: Inclusion criteria: borderline resectability with portal embolization/2-stage hepatectomy, involvement of > 1 hepatic vein or > 4 segments; and/or at least one poor prognostic factor: > 4 metastases; CEA > 200; synchronicity. Limited resectable extraepatic disease and in situ primary allowed. Primary endpoint: TRG1-3 (Rubbia-Brandt et al) according to Simon 2-stage design (first step 22; target 46 pts); secondary endpoints: ORR; R0 resection; safety; progression-free and overall survival. Pts received bi-weekly irinotecan (180 mg/m²) and bevacizumab (5 mg/Kg) day 1, oxaliplatin (85 mg/m²) day 2 and capecitabine (1000 mg/m²/day b.i.d.) days 2–6; 5 cycles pre-operatively (the last without bevacizumab) and 4 post-operatively. Results: We present preliminary data on the first 25 pts. M/F: 13/12, median age 61 years, synchronous disease 19 (76%), multiple nodules 15 (60%), N+ primary tumour 14 (56%), CEA > 200 4 (16%), extrahepatic disease 1, RAS mutation 15 (60%)/BRAF mutation 0. Involvement ≥ 4 hepatic segments 9 (36%) and involvement of > 1 hepatic veins 4 (16%). ORR was 91% (21 out of 23 evaluable, 2 too early), while 1 SD and 1 PD. Surgery performed in 21 pts (4 awaiting) and R0 resection in 18 (86%). TRG1-2 was observed in 6 (29%), TRG3 in 9 (43%), i.e. pathological response 72% (first step reached). Grade ≥ 3 toxicities: neutropenia 2, diarrhea 2, thrombosis 2. Median follow up 13 months, with only 4 relapses and 1 death. Conclusions: COI-B regimen is a feasible neoadjuvant strategy for borderline resectable or high recurrence risk CLM. This is the first trial to select pathological response as primary surrogate endpoint with encouraging results. Clinical trial information: NCT02086656