Background: This large randomized open-label phase 2 trial compared the efficacy and safety of Neratinib (N), the irreversible pan-HER tyrosine kinase inhibitor, + Paclitaxel (P) vs Trastuzumab (T) + P as first-line treatment in HER2+ metastatic breast cancer (MBC). Methods: 479 pts were enrolled between 08/2009 and 08/2011. Women ≥18y with locally recurrent or MBC, HER2 gene amplification/HER2 overexpression, no progression within 12m (neo)adjuvant therapy, and no prior treatment for advanced disease were eligible. Pts were randomized to oral N (240mg od) + P or T+P. Dosing: T 4mg/kg iv then 2mg/kg iv weekly, P 80mg/m² iv weekly 3/4w. Primary endpoint: progression-free survival (PFS). Secondary endpoints: overall survival (OS); overall response rate (ORR); duration of response; clinical benefit rate; CNS progression; safety. PFS analysis was ITT by stratified log-rank test. Randomization was stratified by region, prior T, prior lapatinib, ER/PR status. 304 PFS events were required to detect 30% improvement in median PFS with 80% power (2-sided α=0.15). Results: 479 pts made up the ITT population (6 pts in NP arm and 12 pts in TP arm had known brain metastases at baseline; other baseline characteristics were balanced). Main efficacy findings are shown below. Most common adverse events (AEs) with NP vs TP were: diarrhea 93 vs 33% (G3 30 vs 4%; no G4; G3 median duration of 4.5d); nausea 44 vs 30% (G≥3 2 vs 1%); vomiting 36 vs 16% (G≥3 3 vs 1%); fatigue 32 vs 27% (G≥3 3 vs 3%); rash 31 vs 24% (G≥3 1 vs < 1%); cardiac disorders 2.5 vs 5.1% (G≥3 < 1 vs < 1%). Conclusions: NP has similar efficacy to TP; of interest, NP may be more effective than TP in reducing CNS progression. With no primary loperamide prophylaxis, diarrhea was significantly higher with NP than TP; non-GI AEs occurred at similar rates in both arms. Clinical trial information: NCT00915018

^a49 events were CNS only.