Patients and study design
This phase II trial, which opened for accrual in the US and Canadian NSABP Foundation sites, was designed originally as a two-arm study with 2:1 randomization to evaluate trastuzumab or neratinib with paclitaxel followed by AC. In December 2011, after 30 patients were enrolled, accrual was placed on hold. The decision to hold accrual was based on reports that dual anti-HER2 inhibition with trastuzumab plus lapatinib or trastuzumab plus pertuzumab increased the pCR rate in neoadjuvant breast cancer [11, 17]. From May 2011 to July 2012, the NSABP Foundation conducted a phase I dose-escalation study in patients with HER2+ metastatic disease evaluating the combination of trastuzumab, neratinib, and paclitaxel. This study established the recommended phase II dose of neratinib as 200 mg/day in combination with trastuzumab and paclitaxel . In August 2012, FB-7 reopened as a randomized (1:1:1) study adding the third arm of trastuzumab, neratinib, and paclitaxel followed by AC. The randomization was adjusted so that at the completion of the study, the three arms, would have an equal number of evaluable patients (total study N = 126). On September 30, 2013, the US FDA granted accelerated approval to pertuzumab in combination with trastuzumab and docetaxel for use in the neoadjuvant setting in women with HER2+ breast cancer based on improved pCR in the NeoSphere study (pertuzumab, trastuzumab, docetaxel, pCR rate 45.8%)  and TRYPHAENA study (pertuzumab, trastuzumab, carboplatin, docetaxel, pCR rate 66.2%) .
On October 22, 2013, the NSABP closed arms 1 and 2 of FB-7 to additional accrual in the USA. From October 22, 2013, to April 2014, arm 3 remained open as a non-randomized arm to accrue an additional 12 patients in order to obtain more robust safety information on US patients. These patients are included only in the biomarker analysis with the exception of the MammaPrint® analysis, which included only randomized patients. European sites, in which dual anti-HER2 therapy was not the standard of care, were opened to complete the accrual to the randomized three-arm trial (total study N = 126). Canadian sites continued to accrue to the three-arm trial. The last patient enrolled on November 12, 2014.
Eligible patients included women ≥ 18 years with ECOG PS of 0–1, invasive adenocarcinoma of the breast, HR+ or HR− disease, and HER2 positivity defined as a score of 3+ by immunohistochemistry (IHC), or evidence of gene amplification by fluorescence in situ hybridization (FISH) or chromogenic in situ hybridization (CISH). HER2 testing was performed locally without central confirmation. We are not able to provide HER2 status based upon the guidelines published in 2018 .
Patients with AJCC stage IIB, IIIA, IIIB, or IIIC were eligible. Patients were required to have the following baseline laboratory studies: an absolute neutrophil count of ≥ 1000/mm3, platelet count of ≥ 100,000/mm3, hemoglobin of ≥ 9 g/dL, serum creatinine ≤ 1.5× the upper limits of normal (ULN), total bilirubin of ≤ 1.5× the ULN, AST, and ALT of ≤ 1.5× ULN. Left ventricular ejection fraction (LVEF) ≥ 50% assessed by either 2D echocardiogram or MUGA scan was required. The first 60 patients were required to have a research biopsy before therapy initiation; this was subsequently dropped as a requirement.
Patients were excluded if they had any evidence of metastatic disease, active hepatitis B or C with abnormal liver function tests, intrinsic lung disease causing dyspnea, persistent ≥ grade 2 diarrhea, sensory-motor neuropathy ≥ grade 2, conditions that would prohibit intermittent administration of corticosteroids for paclitaxel premedication, or active cardiac disease including recent myocardial infarction, symptomatic arrhythmia, or angina pectoris.
The study protocol was approved by the Institutional Review Boards of each participating institution, and all patients provided written informed consent. The study was conducted according to the Good Clinical Practice and the Declaration of Helsinki and its amendments. The authors had full control of all primary data. The datasets during and/or analyzed during the current study will be available from the corresponding author on reasonable request and with permission from PUMA Biotechnology.
Patients in arm 1 (control) received 4 cycles of paclitaxel 80 mg/m2 administered on days 1, 8, and 15 of a 28-day cycle with trastuzumab 4 mg/kg loading dose, then 2 mg/kg weekly for a total of 16 doses. Following paclitaxel and trastuzumab, doxorubicin (A) 60 mg/m2 and cyclophosphamide (C) 600 mg/m2 were given every 3 weeks for 4 cycles. In arm 2 (experimental), in place of trastuzumab, patients received neratinib 240 mg taken orally once daily beginning on day 1 of paclitaxel and continuing through day 28 of the final cycle of paclitaxel. In arm 3 (experimental), both trastuzumab and neratinib were given with paclitaxel as described above; however, neratinib was administered at 200 mg/day. After recovery from all chemotherapy, patients had their definitive surgery and completion of 1 year of trastuzumab. The decisions regarding hormonal therapy and radiotherapy were at the discretion of the treating physician.
Standard pre-medications were given before each paclitaxel administration and during AC. Because diarrhea is expected with neratinib, early in the study, diarrhea management was initiated after the first diarrheal stool with loperamide 4 mg and then 2 mg after each loose stool thereafter. Despite this management plan, diarrhea was still consistently noted within 2 weeks of study therapy. Subsequently, diarrheal guidelines were amended to mandate primary prophylaxis with loperamide beginning with the first dose of therapy. At all treating sites, patients were contacted at 24, 48, and 72 h during the first week of treatment on the neratinib-containing arms to reenforce anti-diarrheal management.
The primary endpoint was pCR, defined as no residual invasive disease in the breast and lymph nodes (ypT0/N0). Randomly assigned patients who received any protocol therapy were included in the analysis. Secondary endpoints included clinical complete response, pCR in the breast, and rate of adverse events (AEs). Due to the small sample size and a limited number of tissue samples available, all molecular correlates are exploratory.
Safety was assessed by physical examination, interim history, and laboratory assessment. AE assessment occurred on days 1 and 15 of cycle 1 and on day 1 of each subsequent cycle, and 2 to 4 weeks after the last cycle of AC. AE reporting was assessed according to the National Cancer Institute Common Terminology Criteria for Adverse Events (NCI CTCAE) version 3.0. AEs occurring at the US or Canadian sites were continuously monitored and reviewed by the NSABP medical review team. European sites were monitored by MedSIR every 6 to 10 weeks.
Tumor RNA extraction
RNA was isolated from pre-treatment tumor biopsies and residual disease utilizing one to five 5-μm tissue sections depending on the size of the tumor area within a slide. Separate tissue sections were required for RNA and DNA. Extractions were prepared using the AllPrep DNA/RNA FFPE kit from Qiagen and following the manufacturer’s recommendations.
Whole blood was collected in ACD tubes from patients (n = 80) before treatment. Peripheral blood monocytes (PBMC) were isolated and stored at − 80 °C. Genotyping of rs1801274 (FCGR2A-131R/H) and rs396991 (FCGR3A-158V/F) was performed as previously described .
Details of the preparation, performance, and analysis of the RNA-Seq libraries are included in Additional file 1: Methods.
The proprietary MammaPrint 70-gene Breast Cancer Recurrence Assay was performed at Agendia, using 50 ng of RNA for all available patients with pre-treatment biopsies with sufficient RNA (n = 45); one case failed RNA and hybridization quality control, and in five cases, pCR information was missing. MammaPrint was performed retrospectively to determine the number of low-risk patients enrolled in the study.
This is a non-comparative, randomized phase II study in which patients with HER2+, locally advanced disease were randomized to one of the three arms. The study was designed originally as a two-arm trial with 2:1 randomization. However, when it became apparent that dual anti-HER2 regimens showed early favorable results, the trastuzumab plus neratinib arm was added and the randomization balanced among arms. The primary endpoint was pCR, defined as no evidence of invasive disease in the breast and nodes (ypT0/N0). We determined that a sample size per arm of 42 patients would offer 80% power to test the null hypothesis that the response rate would be 30% or less at a one-sided alpha of 0.05. pCR was analyzed descriptively by treatment arm and, for exploratory purposes, was compared across the three arms (two-by-two comparisons) using the stratified Cochran-Manel-Haenszel statistic (two-sided P values). The analyses were stratified in such a way that patients randomly assigned to the study prior to the addition of arm 3 were considered in a separate stratum from those randomly assigned after the addition of arm 3.
Secondary aims and exploratory molecular analyses were performed for exploratory purposes with Fisher’s exact or chi-square tests, not adjusted for multiplicity. Any comparison that reached nominal statistical significance should be interpreted with due regard to the multiple comparisons performed, the small number of patients in the subgroups, and the fact that some analyses were performed retrospectively.