Background: HER2+ breast cancer is driven by HER2/HER3/PI3K signaling. Single agent targeted therapy (tx) is limited by redundant mechanisms of pathway activation and feedback loops; preclinical models support combinatorial tx. This phase I trial examined HER2, HER3 and PI3K inhibition in pts with HER2+ MBC (NCT02167854). Methods: Pts received weekly intravenous (IV) trastuzumab (T) at 2 mg/kg, LJM716 (L, HER3 inhibitor) at 20 mg/kg, and escalating dose cohorts of daily oral BYL719 (B, PI3Kα inhibitor) starting at 250 mg. Eligible pts had HER2+ MBC with a PIK3CA mutation and prior ado-trastuzumab emtansine and pertuzumab. Endpoints were: to define MTD (toxicity based on CTCAE 4.0) using the Continual Reassessment Method, to assess efficacy (RECIST v1.1), to evaluate genomics and proteomics of pre- and on-tx tumor biopsies, and to quantify cell-free DNA and PIK3CA mutant allele fraction (MAF). Results: 8 pts have been treated with a median (M) age 56y (range (R): 46-68), M ECOG of 1 (R: 0-1), and M of 6 (R: 3-10) prior tx. M duration on study was 10 weeks (R: 2-21.9). 7 pts were treated with B 250 mg. 1 pt was treated at 300 mg and had a DLT of supraventricular tachycardia in the setting of hypokalemia. A second pt had a DLT of G3 transaminitis at B 250 mg. Significant toxicities (and worst grades) included diarrhea (G3: 5 pts; G1/2: 2 pts); hyperglycemia (G3: 2 pts; G1/2: 6 pts); hypokalemia (G3: 2 pts; G1: 4 pts); mucositis: (G3: 1 patient; G1/2: 4 pts); and transaminitis: (G3: 2 pts), with no G4 toxicity. Despite prophylactic antidiarrheals, diarrhea led to dose reduction in 1 (13%) and tx interruptions in 5 (63%) pts. Overall, toxicities limited drug delivery, with only 72%/83% of total planned B/L doses given. In addition, B was dose-reduced in 2 (25%) and L in 4 (50%) of pts. Best response was SD in 5 of 6 evaluable pts. Pre-tx genomics and PIK3CA MAF will be correlated with response. Conclusions: The combination of T, L, and B has antitumor activity in these pre-treated HER2+ pts with MBC with PIK3CA mutations. Clinically significant gastrointestinal and metabolic toxicities limit drug delivery and dose escalation of B/L. Based on preclinical modeling, exploration of intermittent dosing schedules is warranted. Clinical trial information: NCT02167854