Efficacy and safety of ABP 980 compared with reference trastuzumab in women with HER2-positive early breast cancer (LILAC study): a randomised, double-blind, phase 3 trial.

BACKGROUND: ABP 980 (Amgen Inc, Thousand Oaks, CA, USA) is a biosimilar of trastuzumab, with analytical, functional, and pharmacokinetic similarities. We compared the clinical safety and efficacy of ABP 980 with that of trastuzumab in women with HER2-positive early breast cancer.

METHODS: We did a randomised, multicentre, double-blind, active-controlled equivalence trial at 97 study centres in 20 countries, mainly in Europe and South America. Eligible women were aged 18 years or older, had histologically confirmed HER2-positive invasive early breast cancer, an Eastern Cooperative Oncology Group performance status score of 0 or 1, and were planning to have surgical resection of the breast tumour with sentinel or axillary lymph node dissection and neoadjuvant chemotherapy. After four cycles of run-in anthracycline-based chemotherapy, patients were assigned 1:1 to receive ABP 980 or trastuzumab with a permuted block design (blocks of four) computer-generated randomisation schedule. Patients received neoadjuvant therapy with a loading dose (8 mg/kg) of ABP 980 or trastuzumab plus paclitaxel 175 mg/m2 in a 90 min intravenous infusion, followed by three cycles of 6 mg/kg intravenous ABP 980 or trastuzumab plus paclitaxel 175 mg/m2 every 3 weeks in 30 min intravenous infusions (or 80 mg/m2 paclitaxel once per week for 12 cycles if that was the local standard of care). Randomisation was stratified by T stage, node status, hormone receptor status, planned paclitaxel dosing schedule, and geographical region. Surgery was completed 3-7 weeks after the last dose of neoadjuvant treatment, after which adjuvant treatment with ABP 980 or trastuzumab was given every 3 weeks for up to 1 year after the first dose in the study. Patients had been randomly assigned at baseline to continue APB 980, continue trastuzumab, or switch from trastuzumab to APB 980 as their adjuvant treatment. The co-primary efficacy endpoints were risk difference and risk ratio (RR) of pathological complete response in breast tissue and axillary lymph nodes assessed at a local laboratory in all patients who were randomly assigned and received any amount of neoadjuvant investigational product and underwent surgery. We assessed safety in all patients who were randomly assigned and received any amount of investigational product. This trial is registered with ClinicalTrials.gov, number NCT01901146 and Eudra, number CT 2012-004319-29.

FINDINGS: Of 827 patients enrolled, 725 were randomly assigned to receive ABP 980 (n=364) or trastuzumab (n=361). The primary endpoint was assessable in 696 patients (358 who received ABP 980 and 338 who received trastuzumab). Pathological complete response was recorded in 172 (48%, 95% CI 43-53) of 358 patients in the ABP 980 group and 137 (41%, 35-46) of 338 in the trastuzumab group (risk difference 7·3%, 90% CI 1·2-13·4; RR 1·188, 90% CI 1·033-1·366), with the upper bounds of the CIs exceeding the predefined equivalence margins of 13% and 1·318, respectively. Pathological complete response in the central laboratory assessment was seen in 162 (48%) of 339 patients assigned to ABP 980 at baseline and 138 (42%) of 330 assigned to trastuzumab at baseline (risk difference 5·8%, 90% CI -0·5 to 12·0, and RR 1·142, 90% CI 0·993 to 1·312). Grade 3 or worse adverse events during the neoadjuvant phase occurred in 54 (15%) of 364 patients in the ABP 980 group and 51 (14%) of 361 patients in the trastuzumab group, of which the most frequent grade 3 or worse event of interest was neutropenia, occurring in 21 (6%) patients in both groups. In the adjuvant phase, grade 3 or worse adverse events occurred in 30 (9%) of 349 patients continuing ABP 980, 11 (6%) of 171 continuing trastuzumab, and 13 (8%) of 171 who switched from trastuzumab to ABP 980, the most frequent grade 3 or worse events of interest were infections and infestations (four [1%], two [1%], and two [1%]), neutropenia (three [1%], two [1%], and one [1%]), and infusion reactions (two [1%], two [1%], and three [2%]). Two patients died from adverse events judged to be unrelated to the investigational products: one died from pneumonia while receiving neoadjuvant ABP 980 and one died from septic shock while receiving adjuvant ABP 980 after trastuzumab.

INTERPRETATION: Although the lower bounds of the 90% CIs for RR and risk difference showed non-inferiority, the upper bounds exceeded the predefined equivalence margins when based on local laboratory review of tumour samples, meaning that non-superiority was non-conclusive. In our sensitivity analyses based on central laboratory evaluation of tumour samples, estimates for the two drugs were contained within the predefined equivalence margins, indicating similar efficacy. ABP 980 and trastuzumab had similar safety outcomes in both the neoadjuvant and adjuvant phases of the study.

FUNDING: Amgen.