Abstract: Background: We tested the oral mammalian target of rapamycin (mTOR) inhibitor everolimus in addition to paclitaxel in patients with HER2-negative tumours not responding to initial neoadjuvant cytotoxic and anti-angiogenic treatment. Methods: Patients with primary HER2-negative tumours received four neoadjuvant cycles of epirubicin/cyclophosphamide (EC) with or without bevacizumab. Patients without clinical response were randomised to receive weekly paclitaxel (80mg/m2) with or without everolimus (5mg p.o. daily, after a step-wise dose–escalation starting from 2.5mg bid) for 12weeks before surgery. To detect an increase in pathological complete response (pCR; ypT0 ypN0) from 5% to 12.1% (odds ratio 2.62) 566 patients had to be recruited. The trial was stopped prematurely due to completion of accrual in the main study. Findings: Of 1948 patients initially starting neoadjuvant treatment 403 were randomised. A total of 18 (4.6%) patients, 7 (3.6%) treated with paclitaxel and everolimus and 11 (5.6%) treated with paclitaxel alone had a pCR (odds ratio 0.36 (OR) (95% confidence interval (CI), 0.24–1.6) p =0.34). Overall response rate in breast and lymph nodes at surgery was 52.2% after paclitaxel plus everolimus and 61.7% after paclitaxel alone (p =0.063). Breast conserving treatment was performed in 54.4% of patients with the combination treatment and 61.9% with paclitaxel alone (p =0.20). Mucosal inflammation, thrombocytopenia, neutropenia, infection, and skin rash were more frequent when everolimus was added to paclitaxel. Interpretation: Neoadjuvant therapy with everolimus and paclitaxel for patients with HER2-negative disease unresponsive to EC with or without bevacizumab did not improve the pCR rate. Long-term outcome is awaited. Funding: Novartis, Roche, and Sanofi-Aventis. [Copyright &y& Elsevier]