SUMMARY CORE 2 This Program Project comprises four individual projects, which will: implement evidence-based sequential multiple treatment assignment strategies for patients predicted to have insufficient response to their initial neoadjuvant targeted and/or chemotherapy (Project 1); qualify non-invasive imaging methods as early markers of non-response (Project 2); characterize the biology of non-responders to inform treatment selection (Project 3); and develop a portfolio of agents and decision tools for treatment re-assignment matched to biology of non- responding tumors (Project 4). The Bioinformatics and Statistics Core will act as a centralized resource where the analytical goals of these projects converge - where we work closely with each of the project groups not only to provide project-specific analytical support, but also to build predictive models across multiple modalities (imaging, molecular profiles and their combination) and facilitate cross-project interactions towards the common goal of building robust decision algorithms to enable adaptation of treatment for individual women with poor response to their initial neoadjuvant targeted and/or chemotherapy. This undertaking will utilize substantial the archived and newly generated datasets from the I-SPY 1 and I-SPY 2 trials to develop and validate algorithms that will enable the transition to I-SPY 2+, where patients predicted as insufficient responders by an optimized, subtype-specific MRI-based predictor “Virtual RCB” can be identified during the course of their initial therapy (after completion of their taxol +/- experimental agent and 2 cycles of AC) and offered alternative treatment strategies based on their tumor biology in order to mitigate recurrence and improve long term outcomes. The primary goal of the Biostatistics and Bioinformatics Core is to provide biostatistics and bioinformatics support to individual projects and facilitate cross-project analyses and results sharing within the Program Project Framework. The specific aims are listed as follows: Specific Aim 1: To provide innovative bioinformatics and statistical modeling and analytical approaches needed by the projects to achieve their Specific Aims. Specific Aim 2: To develop SMART (sequential, multiple assignment, randomized trial) methods for adaptive treatment of predicted non-responders within the I-SPY 2+ Program Project framework. Specific Aim 3: To synthesize biomarker data within and across projects into actionable clinical information.