Background: The immunomodulator MGN1703, a TLR-9 agonist, has shown good safety profile in patients with metastatic solid tumors. The IMPACT trial was conducted to assess clinical efficacy, safety, and immunological effects of MGN1703 as maintenance therapy twice weekly. Methods: The international randomized (2:1) double-blind placebo-controlled phase 2 IMPACT trial in mCRC patients with disease control after 1st-line chemotherapy by FOLFOX/XELOX or FOLFIRI +/- bevacizumab. After randomization of 59 out of 129 planned patients (43 MGN1703, 16 placebo) the trial was prematurely closed due to slow recruitment. Results: A superior effect of MGN1703 compared to placebo was shown: HR for the primary endpoint PFS on maintenance was 0.55 (p=0.041) by local and 0.56 (p=0.070) by independent assessment. For PFS from start of induction therapy the HR was 0.50 (p=0.022) and 0.49 (p=0.030), respectively. Notably, at time of study closure 4 MGN1703 patients were still free of PD and continued in compassionate use. A possible predictive effect was identified by Cox regression analyses: HR of 0.07 (p<0.0001) for patients with normal CEA and 0.39 (p=0.005) for patients with an objective response. Evaluation of biological activity via cell populations and chemokine serum levels confirmed immune system activation in MGN1703 patients. Cox regression and ROC analyses identified of activated NKT-cells (CD3⁺/CD56⁺/CD69⁺) at baseline as potentially predictive for a benefit from MGN1703 treatment (HR: 0.27, p=0.007). Study treatment was well tolerated: 32.6% vs. 18.8% of patients (MGN1703 vs. placebo) had drug-related AE and in only 1 patient per arm it was of grade 3 (MGN1703: sensory polyneuropathy, placebo: papular exanthema). Seven SAEs were reported of which only 1 was possible drug-related (moderate atypical pneumonia). Conclusions: Following induction chemotherapy, MGN1703 maintenance in mCRC patients is associated with improved PFS and an activation of the innate and adaptive immune system. Treatment was safe and well tolerated. There is preliminary evidence that CEA levels, tumor response, and activated NKT cells before start of MGN1703 therapy allow for identification of benefitting patients. Clinical trial information: NCT01208194.