Background: Regorafenib is a treatment option for refractory mCRC pts. Considering its limited clinical benefit and the palliative setting, patients’ selection is essential to optimize the cost-effectiveness ratio. We aimed to describe the modulation of selected circulating angiogenic factors by regorafenib and to investigate their correlation with clinical outcome. Methods: IL-8, Ang-2, PDGF, bFGF, Tie-2, sVEGFR2, sVEGFR3, PlGF, VEGF-A, VEGF-B were assessed by ELISA on plasma samples collected at baseline (d1), after 15 days of treatment (d15), at the best RECIST response, at disease progression (PD) in mCRC pts treated with regorafenib, as per indication. Comparisons among concentrations of each marker at different timepoints were performed by Wilcoxon test. Markers showing significant changes were analyzed for correlation with outcome. Results: 105 pts were included. Median PFS and OS were 2.1 and 7.0 months (mos), respectively. As compared to d1, IL-8 and Ang-2 levels increased at PD. An early decrease at d15 was observed for PDGF, Tie-2, sVEGFR2 and sVEGFR3 levels, followed by an increase at PD. Conversely, PlGF levels increased at d15 and then decreased at PD. Baseline levels of Ang-2 and Tie-2 below the median value were associated with longer PFS (HR:0.61 [0.37-0.83] p = 0.006; HR:0.69 [0.43-0.95] p = 0.04, respectively) and OS (HR:0.45 [0.26-0.60] p < 0.0001; HR:0.68 [0.44-0.98] p = 0.04, respectively). With regard to Ang-2, 40 (45%) out of 89 pts with available plasma samples at d15 showed increased levels at d15 as compared to d1. Among them, 21 (53%) achieved disease control, as compared to 14 out of 49 (29%) pts with Ang-2 decreased levels (p = 0.03). Median PFS of pts with increased and decreased Ang-2 levels were 3.1 and 1.8 mos, respectively (HR:0.57 [0.33-0.78] p = 0.004). Conclusions: A dynamic modulation of plasma angiogenic factors occurs during the treatment with regorafenib. Low baseline Ang-2 and Tie-2 levels seem to be associated with good prognosis. The early modulation of Ang-2 levels may predict benefit from regorafenib. These results need validation in independent series.