Therapeutic options after first-line treatment in patients with RAS mutant metastatic colorectal cancer (mCRC) are limited by poor clinical outcomes and lack of activity of anti-epidermal growth factor receptor drugs. Maintaining angiogenesis inhibition and switching the chemotherapy backbone represents the current therapeutic strategy of second-line mCRC therapy. Regorafenib, an oral broad-spectrum multikinase inhibitor targeting angiogenesis, has been shown to prolong overall survival (OS) and is approved after progression on chemotherapy. STREAM trial was an academic, multicentric, phase 2, single-arm, Simon’s two-stage design, aimed to establish the activity of regorafenib, defined as the rate of patients alive and progression free after 6months from study entry (6mo-PFS) in patients with RAS mutant mCRC, after progression on fluoropyrimidine, oxaliplatin and bevacizumab.Setting α and β errors at 0.10, defining a 6mo-PFS rate of 30% as p0 and 50% as p1, to consider the study positive ≧8 patients without progression at 6mo, were needed in the first stage (N=22) and ≧18 in the overall population (N=46). Regorafenib was administered orally at the approved schedule of 160mg day1-21 q28. Secondary endpoints were the evaluation of toxicity, objective response rate (ORR), progression-free survival (PFS), OS. Early metabolic response assessed by PET-CT scan after 2 weeks of treatment was an exploratory endpoint. Translational analyses are ongoing and will be presented at a later stage. Between November 2015, and December 2020, 46 patients were enrolled. median age was 67ys; 80.4% had ECOG performance status 0, 32.6% single organ involvement, 26% lung-limited disease. The study did not meet its predefined primary endpoint. Eight of the first 22 patients and 14 in the overall population were 6mo-PFS, as compared to the 18 required by the study protocol. At a median follow-up of 50.2 months (95%CI=24.2-56.3), ORR was 10.9%, Disease Control Rate (DCR) was 54.6%, median (m)PFS was 3.6mo (95%CI=1.9-6.7).Despite the short mPFS, regorafenib did not preclude a subsequent treatment: mPFS2 (from study entry to progression to subsequent treatment line) was 13.3mo (95%CI=8.4-19.7) and mOS was 18.9mo (95%CI=10.3-35.3). No unexpected toxicity was reported. Grade≥3 AEs occurred in 18 patients(39.1%), mostly hand foot syndrome(13%), fatigue and bilirubin increase(6.5%). Early metabolic response with PET/CT was not associated with ORR, mPFS and mOS. Baseline metabolic assessment, not predictive of ORR and mPFS, was significantly associated with mOS. A subgroup of patients, with single organ involvement, mostly lung-limited disease, low baseline PET-CT parameters, a low rate of early progression in first line (<6months=21.4%) and a long mPFS2 (23.3mo 95%CI=13.3-39.9), reported a long disease control with regorafenib: ORR=35.7%; mPFS=10.2mo (95%CI =8.6-13.5), mOS=38.8mo (95%CI=23.9-NR). Conversely, patients with higher radiological and metabolic burden of disease, higher rate of early progression in first-line (<6months=56.3%) and a shorter mPFS2 (8.4mo CI95%=6.1-13.3), reported unfavorable outcomes. Despite the study did not meet its primary endpoint, treatment with regorafenib had no unexpected toxicity and did not preclude efficacy of subsequent treatments. A subgroup of patients characterized by good prognostic features (limited disease, lung metastasis and low metabolic burden) had clinical benefit with regorafenib. Appropriate patients selection might guide the clinical development of regorafenib in an earlier setting, ensuring a chemotherapy-free interval in the treatment sequence.