Breast tumors expressing estrogen receptor-alpha (ER) tend to be less responsive to cytotoxic chemotherapy than ER-negative breast cancers; advanced ER+ breast cancer remains largely incurable. Moreover, this breast cancer subtype frequently exhibits marked dormancy, conferring upon patients a high risk of experiencing a late recurrence (emergence from dormancy) that persists for decades after their initial diagnosis and treatment. Why ER+ breast is the most strongly associated with dormancy is unknown and represents a critical barrier to progress in the field. Since 70% of newly diagnosed breast cancers are ER+ but ~50% of these eventually recur, many doing so years after cessation of an otherwise apparently successful initial intervention, significance of the dormant phenotype is clear. Using transcriptome data from pretreatment tumors from women with ER+ breast cancer who were subsequently treated with tamoxifen (TAM) as their only systemic therapy, we derived an initial molecular classifier that robustly separates early (d 3 yrs) from late (e 5 yrs; emergence from dormancy) distant recurrences in both training and independent datasets. We then identified novel features of the "early" (E) vs. "late" (L) signaling network's topology in these datasets. We also developed an adaptation of the DMBA-induced rat mammary tumor model and show that it represents E, L, and ""not"" (N) recurring tumors, molecular features of which are present in human breast tumors and cell lines. Mechanistically, our data implicate a "rewiring" (including epigenetic events) of the signaling from the unfolded protein response (UPR) and autophagy, and adaptations in cellular metabolism, as the molecular events that may enable cells to maintain, and later escape from, dormancy. Our Central hypothesis is that dormancy reflects the growth arrest induced by endocrine therapies in both residual tumors (in-breast recurrences) and micro-metastases (distant recurrences), and that the signaling maintaining dormancy after systemic therapy ends is epigenetically regulated. Indeed, prolonged hormone therapy may well work by extending dormancy. Individual cells can emerge from dormancy as their metabolic capacity becomes sufficient to support both survival and replication, and the integrated balance between autophagy (prosurvival) and apoptosis (prodeath) signaling enables more cells in a tumor to proliferate and the tumor population size to grow. We propose an integrated, multidisciplinary research program that, if successful, will generate innovative insights into the molecular drivers of dormancy in ER+ breast cancers, develop new in silico signaling models, identify predictors of patient risk to experience a late recurrence, and explore mechanisms of dormancy with the longer term goal of discovering potentially novel therapeutic interventions.