Background: Neoadjuvant endocrine therapy with an aromatase inhibitor has shown efficacy comparable to that of neoadjuvant chemotherapy in postmenopausal breast cancer. Pre-/Clinical data have shown that metformin, a widely used anti-diabetic drug also one of mTOR inhibitor have shown anti-tumor activity. We report the result of prospective, multicenter, phase II randomized, placebo controlled trial aiming to evaluate the direct anti-tumor effect of metformin in non-diabetic postmenopausal women with hormone-receptor (HR) positive breast cancer. Methods: 203 postmenopausal women diagnosed with hormone receptor positive, T1-3/N0-2 invasive breast cancer were randomized to 24 weeks of neoadjuvant letrozole (2.5 mg/day) and either metformin (2000 mg/day) or placebo. Women with history of diabetes were excluded. Primary endpoint was clinical response rate (complete, partial response by caliper). Secondary endpoint was pathologic complete response rate, breast conservation rate, percent mammographic density change. PEPI score and toxicity profile were compared between two groups. Results: 153 intention-to treat population were analyzed (72 metformin, 75 placebo group). Overall clinical response rate was 61.4% (94/153) by caliper and did not reach statistical significance between metformin versus placebo groups (66.7% versus 56.4%, p = 0.193). Breast conservation rate was 68.0% (100/147) (66.7% versus 69.3%). Overall, 87.3% (103/118) displayed Ki67 < 10% at surgical specimen and 16.7% (21/126) had zero PEPI score. Neither Ki67% nor PEPI score was different between two groups. However, among the 20 patients with core-needle biopsy after 4 weeks of medication, greater number of patients displayed Ki67 < 10% in metformin group than in placebo group (87.5% versus 33.3%, p = 0.017). Patients with 4week Ki67 < 10% had higher clinical response rate (100% versus 57.1%, p = 0.038). Grade 3 side effects were reported in three patients (vomiting, high blood pressure, weight loss) and no hypoglycemia event was observed. Conclusions: 61.7% overall clinical response was achieved with 24-weeks of neoadjuvant letrozole, with numerically > 10% higher response rate in letrozole+metformin group (66.7% versus 56.4%). 4-weeks Ki67 < 10% level was predictive of clinical response. With < 2% grade 3 side effects, preoperative letrozole (with/without metformin) followed by 4-week Ki67 evaluation may indeed serve as primary choice to postmenopausal hormone receptor positive breast cancers. Clinical trial information: NCT01589367