INTRODUCTION: Cisplatin plus pemetrexed followed by pemetrexed is an efficacious platinum combination regimen for advanced non-squamous, non-small cell lung cancer (NSCLC). Data regarding the addition of bevacizumab, especially in maintenance treatment, are insufficient. METHODS: Eligibility criteria included: no prior chemotherapy; advanced, non-squamous, NSCLC; performance status ≤1; and epidermal growth factor receptor mutation-negative. Patients (N = 108) received induction chemotherapy with cisplatin, pemetrexed, and bevacizumab every 3 weeks for four cycles, and tumor response was needed to confirm four-week response duration. Patients with at least stable disease were randomized to pemetrexed/bevacizumab or pemetrexed alone. The primary endpoint was progression-free survival (PFS) after induction chemotherapy. Myeloid-derived suppressor cell (MDSC) counts of peripheral blood samples were also analyzed. RESULTS: Thirty-five patients each were randomized to the pemetrexed/bevacizumab group and the pemetrexed alone group. PFS was significantly better in the pemetrexed/bevacizumab group than in the pemetrexed alone group (7.0 vs. 5.4 months, hazard ratio: 0.56 [0.34-0.93], log-rank p = 0.023). In patients with partial response to induction therapy, median overall survival was 23.3 months in the pemetrexed alone group and 29.6 months in the pemetrexed/bevacizumab group (log-rank p = 0.077). Pretreatment monocytic (M)-MDSC counts tended to be greater in the pemetrexed/bevacizumab group with poor PFS than in those with good PFS (p = 0.0724). CONCLUSIONS: Addition of bevacizumab to pemetrexed as maintenance therapy prolonged PFS in patients with untreated, advanced, non-squamous NSCLC. Furthermore, an early response to induction therapy and pretreatment M-MDSC counts may be related to the survival benefit of the addition of bevacizumab to the combination of cisplatin and pemetrexed.