The optimal chemotherapy backbone for dual HER2 blockade in the neoadjuvant setting for early breast cancer is unknown. We investigated whether the addition of anthracyclines would improve pathological complete response compared with a carboplatin–taxane regimen, when given in combination with the HER2-targeted agents trastuzumab and pertuzumab.

The TRAIN-2 study is an open-label, randomised, controlled, phase 3 trial being done in 37 hospitals in the Netherlands. We recruited patients aged 18 years or older with previously untreated, histologically confirmed stage II–III HER2-positive breast cancer. Patients were randomly allocated using central randomisation software (1:1 ratio) with minimisation without a random component, stratified by tumour stage, nodal stage, oestrogen receptor status, and age, to receive 5-fluorouracil (500 mg/m²), epirubicin (90 mg/m²), and cyclophosphamide (500 mg/m²) every 3 weeks for three cycles followed by paclitaxel (80 mg/m² on days 1 and 8) and carboplatin (area under the concentration–time curve [AUC] 6 mg/mL per min on day 1 or optionally, as per hospital preference, AUC 3 mg/mL per min on days 1 and 8) every 3 weeks for six cycles, or to receive nine cycles of paclitaxel and carboplatin at the same dose and schedule as in the anthracycline group. Patients in both study groups received trastuzumab (6 mg/kg, loading dose 8 mg/kg) and pertuzumab (420 mg, loading dose 840 mg) concurrently with all chemotherapy cycles. The primary endpoint was the proportion of patients who achieved a pathological complete response in breast and axilla (ypT0/is ypN0) in the intention-to-treat population. Safety was analysed in patients who received at least one treatment cycle according to actual treatment received. This trial is registered with ClinicalTrials.gov, number NCT01996267, and follow-up for long-term outcome is ongoing.

Between Dec 9, 2013, and Jan 14, 2016, 438 patients were enrolled and randomly assigned to the two treatment groups (219 patients to each group), of whom 418 were evaluable for the primary endpoint (212 in the anthracycline group and 206 in the non-anthracycline group). The median follow-up for all patients was 19 months (IQR 16–23 months). A pathological complete response was recorded in 141 (67%, 95% CI 60–73) of 212 patients in the anthracycline group and in 140 (68%, 61–74) of 206 in the non-anthracycline group (p=0·95). One patient randomly allocated to the non-anthracycline group did receive anthracyclines and was thus included in the anthracycline group for safety analyses; therefore, for the safety analyses there were 220 patients in the anthracycline group and 218 in the non-anthracycline group. Serious adverse events were reported in 61 (28%) of 220 patients in the anthracycline group and in 49 (22%) of 218 in the non-anthracycline group. The most common adverse events of any cause were grade 3 or worse neutropenia (in 131 [60%] of 220 patients in the anthracycline group vs 118 [54%] of 218 in the non-anthracycline group), grade 3 or worse diarrhoea (26 [12%] vs 37 [18%]), and grade 2 or worse peripheral neuropathy (66 [30%] vs 68 [31%]), with no substantial differences between the groups. Grade 3 or worse febrile neutropenia was more common in the anthracycline group than in the non-anthracycline group (23 [10%] vs three [1%], p<0·0001). Symptomatic left ventricular systolic dysfunction was rare in both groups (two [1%] of 220 vs 0 of 218). One patient in the anthracycline group died because of a pulmonary embolism, which was possibly treatment related.

In view of the high proportion of pathological complete responses recorded in both groups and the fact that febrile neutropenia was more frequent in the anthracycline group, omitting anthracyclines from neoadjuvant treatment regimens might be a preferred approach in the presence of dual HER2 blockade in patients with early HER2-positive breast cancer. Long-term follow-up is required to confirm these results.

Roche Netherlands.