Background: Preoperative treatment of resectable liver metastases from CRC is an hot topic.The aims of this study were to test the feasibility and activity of Bevacizumab+FOLFIRI in this setting and to explore the role of PET/CT in predicting the efficacy of treatment and to compare it to the standard RECIST response. Methods: A single-arm phase 2 study design enrolling 39 patients was applied with 1-year PFS as primary end point. PET/CT was performed before and after 1 cycle of treatment. For each lesion a ≤ -50% change from baseline was used as threshold for significant metabolic response for maximum SUV.RECIST response was assessed with CT after 3 months of treatment.The association between metabolic and CT/RECIST and pathologic response(PR) was tested as well as the ability to predict PFS and OS. Results: Response rate was 66.7%. 37 patients (94.9%) were operated with a R0 rate of 84.6%. 5 patients had a complete PR(14%). At 1 year 24 patients were alive and free from disease(61.6%).Median PFS and OS were 14 and 38 months. Early metabolic PET/CT response had a stronger, independent and statistically significant predictive value for PFS and OS than both CT/RECIST and PR at multivariate analysis. Conclusion: Preoperative treatment of patients with resectable liver metastases from CRC with bevacizumab plus FOLFIRI is feasible. PET/CT response was significantly predictive of long-term outcomes during preoperative treatment and its predictive ability was higher than that of CT/RECIST response after 3 months of treatment.