The GOLFIG-2 phase III trial was designed to compare the immunobiological activity and antitumor efficacy of GOLFIG chemoimmunotherapy regimen with standard FOLFOX-4 chemotherapy in frontline treatment of metastatic colorectal cancer (mCRC) patients. This trial was conceived on the basis of previous evidence of antitumor and immunomodulating activity of the GOLFIG regimen in mCRC. GOLFIG-2 is a multicentric open/label phase III trial (EUDRACT: 2005-003458-81). Chemo-naive mCRC patients were randomized in a 1:1 ratio to receive biweekly standard FOLFOX-4 or GOLFIG [gemcitabine (1000 mg/m2, day 1); oxaliplatin (85 mg/m2, day 2); levofolinate (100 mg/m2, days 1–2), 5-fluorouracil (5-FU) (400 mg/m2 in bolus followed by 24 h infusion at 800 mg/m2,days 1–2), sc. GM-CSF (100 μg, days 3–7); sc. aldesleukin (0·5 MIU bi-daily, days 8–14 and 17–30)] treatments. The study underwent early termination because of poor recruitment in the control arm. After a median follow-up of 43.83 months, GOLFIG regimen showed superiority over FOLFOX in terms of progression-free survival [median 9·23 (95% confidence interval (CI), 6·9–11.5) vs. median 5.70 (95% CI, 3.38–8.02) months; hazard ratio (HR): 0.52 (95% CI, 0.35–0.77), P=0·002] and response rate [66.1% (95% CI, 0.41–0.73) vs. 37·0% (95% CI, 0.28–0.59), P=0.002], with a trend to longer survival [median 21.63 (95% CI, 18.09–25.18) vs. 14.57 mo (95% CI, 9.07–20.07); HR: 0·79 (95% CI, 0.52–1.21); P=0.28]. Patients in the experimental arm showed higher incidence of non-neutropenic fever (18.5%), autoimmunity signs (18.5%), an increase in the number of monocytes, eosinophils, CD4+ T lymphocytes, natural killer cells, and a decrease in immunoregulatory (CD3+CD4+CD25+FoxP3+) T cells. Taken together, these findings provide proof-of-principle that GOLFIG chemoimmunotherapy may represent a novel reliable option for first-line treatment of mCRC.