CT-P6 is a proposed biosimilar to reference trastuzumab. In this study, we aimed to establish equivalence of CT-P6 to reference trastuzumab in neoadjuvant treatment of HER2-positive early-stage breast cancer. In this randomised, double-blind, active-controlled, phase 3 equivalence trial, we recruited women aged 18 years or older with stage I-IIIa operable HER2-positive breast cancer from 112 centres in 23 countries. Inclusion criteria were an Eastern Cooperative Oncology Group performance status score of 0 or 1; a normal left ventricular ejection fraction of at least 55%; adequate bone marrow, hepatic, and renal function; at least one measureable lesion; and known oestrogen and progesterone receptor status. Exclusion criteria included bilateral breast cancer, previous breast cancer treatment, previous anthracycline treatment, and pregnancy or lactation. We randomly allocated patients 1:1 to receive neoadjuvant CT-P6 or reference trastuzumab intravenously (eight cycles, each lasting 3 weeks, for 24 weeks; 8 mg/kg on day 1 of cycle 1 and 6 mg/kg on day 1 of cycles 2-8) in conjunction with neoadjuvant docetaxel (75 mg/m2 on day 1 of cycles 1-4) and FEC (fluorouracil [500 mg/m2], epirubicin [75 mg/m2], and cyclophosphamide [500 mg/m2]; day 1 of cycles 5-8) therapy. We stratified randomisation by clinical stage, receptor status, and country and used permuted blocks. We did surgery within 3-6 weeks of the final neoadjuvant study drug dose, followed by an adjuvant treatment period of up to 1 year. We monitored long-term safety and efficacy for 3 years after the last patient was enrolled. Participants and investigators were masked to treatment until study completion. The primary efficacy endpoint, analysed in the per-protocol population, was pathological complete response, assessed via specimens obtained during surgery, analysed by masked central review of local histopathology reports. The equivalence margin was -0·15 to 0·15. This trial is registered with ClinicalTrials.gov, number NCT02162667, and is ongoing, but no longer recruiting. Between Aug 7, 2014, and May 6, 2016, we randomly allocated 549 patients (271 [49%] to CT-P6 vs 278 [51%] to reference trastuzumab). A similar proportion of patients achieved pathological complete response with CT-P6 (116 [46·8%; 95% CI 40·4-53·2] of 248 patients) and reference trastuzumab (129 [50·4%; 44·1-56·7] of 256 patients). The 95% CI of the estimated treatment outcome difference (-0·04% [95% CI -0·12 to 0·05]) was within the equivalence margin. 19 (7%) of 271 patients in the CT-P6 group reported serious treatment-emergent adverse events versus 22 (8%) of 278 in the reference trastuzumab group; frequent (occurring in more than one patient) serious adverse events were febrile neutropenia (four [1%] vs one [<1%]) and neutropenia (one [<1%] vs two [1%]). Grade 3 or worse treatment-related adverse events occurred in 17 (6%) of 271 patients in the CT-P6 group versus 23 (8%) of 278 in the reference trastuzumab group; the most frequently reported adverse event was neutropenia in ten (4%) versus 14 (5%). CT-P6 showed equivalent efficacy to reference trastuzumab and adverse events were similar. Availability of trastuzumab biosimilars could increase access to this targeted therapy for HER2-positive early-stage cancer. Celltrion Inc.