14542 Background: A multi-center phase II trial has been designed to evaluate toxicity, anti-tumour and immune-biological activity of the bio-chemotherapy GOLFIG regimen in advanced colorectal cancer (A-CRC). Methods: The trial involved 46 patients (34/46 2nd line or more). The biweekly GOLFIG regimen consisted of gemcitabine (1,000 mg/m2, day 1), oxaliplatin (85 mg/m2, day 2), levofolinic acid (100 mg /m2, day 1,2) and 5-FU (400 mg/m2 as a bolus, and 800 mg/m2 as 24 hour infusion, days 1,2) followed by sc GM-CSF (100 μg, days 3 to 8) and sc IL-2 (0.5 X 106 IUs twice a day, days 9 to 14). Results: GOLFIG regimen was well tolerated and resulted very active (ORR = 56.5%; disease control = 91.3%), fulfilling the pre-selected (40–60%) target of activity, with a promising TTP [12.26 months, 95% CI; 9.2–15.2 mo.] and OS [18.76 months, 95% CI; 15.2–22.3 mo.]. We detected a significant rise in lymphocyte number and colon cancer-specific cytotoxic T cells (CTLs) paralleled by a reduction in suppressive T-regulatory CD4+CD25+ FoxP3+ lymphocytes -Treg. Self-limiting autoimmunity (AI) occurred in 6 (14%) patients. 1 patient developed cutaneous lesions, with pathological diagnosis of Discoid Lupus Erythematosus (DLE), while 5 other presented a mono / oligoarticular arthritis mainly localized at the knees, elbows, shoulders and fingers with signs of synovitis (pain and swelling) paralleled by a significant increase of inflammatory markers (ESR, C-RP and rheumatoid factor). Subgroup analysis disclosed a mean time to progression of 23.8 months (95% CI; 12.1–35.56 mo.) and an overall survival of 31.83 months (95% CI; 19.9–43.7 mo.) in the 6 AI patients, significantly longer than in patients who did not develop AI [TTP and OS respectively 10.52 (95% CI; 7.7–13.3) and 16.8 months (95% CI; 13.3–20.2)], P ≤ 0.0039 and 0.0080. 5/6 AI patients are still alive and 3 have so far achieved a survival of 37, 41 and 50 months. Conclusions: The GOLFIG regimen exerts immunological and anti-tumour activity and has manageable toxicity in A-CRC. The occurrence of AI is predictive of a favourable outcome. These results provided the rationale for a presently ongoing phase III trial aimed to compare the efficacy of GOLFIG vs. FOLFOX-4 regimen as first line treatment in A- CRC patients. No significant financial relationships to disclose.