Abstract BACKGROUND: Approximately 50% of TNBC pts treated with standard taxane/anthracycline-based NACT will have chemo-insensitive disease (CID) manifested as extensive residual disease (RCB-II or III) at the time of surgery. 40-80% of these pts will develop recurrence within 3 years of initial diagnosis. Recent advances in molecular profiling have identified subsets of TNBC with distinct, targetable molecular features. We developed a clinical trial to identify and characterize CID (ARTEMIS: A Randomized, TNBC Enrolling trial to confirm Molecular profiling Improves Survival). In the ARTEMIS trial, treatment naïve pts with localized TNBC undergo a pretreatment biopsy and then immediately start their initial phase of anthracycline-based chemotherapy so that the results of the molecular characterization are used in combination with response assessment (clinical exam/diagnostic imaging) to identify CID and inform the second phase of NACT, thus using a 'second hit' strategy in the middle of NACT to overcome drug resistance. The mesenchymal subtypes of TNBC have a high incidence of PI3K pathway activation. Preclinical models demonstrated response to PI3K inhibitors in this subtype. Metaplastic breast cancers make up ∼30% of tumors characterized as 'claudin-low/mesenchymal' by gene signature and are also associated with a high rate of PI3K activating molecular aberrations. A combination regimen of liposomal doxorubicin, bevacizumab and the mTOR inhibitors temsirolimus or everolimus (DAT or DAE) demonstrated response (including durable complete responses) in metastatic metaplastic breast cancer. PRIMARY OBJECTIVE: Determine the rate of pathologic complete response (pCR/RCB-0) or minimal residual disease (RCB-I) after 4 cycles of DAE for treatment of mesenchymal TNBC deemed to be CID through the ARTEMIS trial TRIAL DESIGN AND STATISTICAL METHODS: Only pts deemed to have mesenchymal CID on the ARTEMIS trial can enter this non-randomized phase II study. Realizing that pts without response to their initial cycles of chemotherapy have very low chance (5%) of achieving pCR with additional cycles of chemotherapy, it would be clinically meaningful to see pCR in this pt population improved to 20%. Counting pCR (RCB-0) or RCB-I as response, a two-stage Gehan-type design will be employed with 14 pts in the first stage. If at least one pt responds, 23 more pts will be added for a total of 37 pts. This design has a 49% chance of terminating after the first stage if the true response rate is 0.05, 23% chance if the true rate is 0.10, 10% if the true rate is 0.15 and 4% if the true rate is 0.20. If accrual continues to the second stage and a total of 37 pts are enrolled, the 95% confidence interval for a 0.20 response rate will extend from 0.10 to 0.35. BRIEF ELIGIBILITY CRITERIA: Inclusion: localized TNBC enrolled onto ARTEMIS trial, adequate organ, bone marrow and cardiac parameters Exclusion: metastatic disease, pregnant or lactating pts, medical illness that increases chance of moderate to severe toxicity CORRELATIVE SCIENCE: Correlate vimentin expression by IHC, mesenchymal signatures and PI3K pathway aberrations with response. Citation Format: Moulder S, Hess K, Rauch M, Astrada B, Litton J, Mittendorf E, Ueno N, Tripathy D, Lim B, Piwnica-Worms H, Thompson A, Symmans WF. NCT02456857: A phase II trial of liposomal doxorubicin, bevacizumab and everolimus (DAE) in patients (pts) with localized triple-negative breast cancer (TNBC) with tumors predicted insensitive to standard neoadjuvant chemotherapy (NACT) [abstract]. In: Proceedings of the 2016 San Antonio Breast Cancer Symposium; 2016 Dec 6-10; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2017;77(4 Suppl):Abstract nr OT2-01-22.