Lack of animal model system that accurately predicts therapy responses in humans is blocking the development of new and more effective anti-cancer therapies. The objective of this application is to demonstrate that patient-tumor derived orthotopic xenograft mouse models of malignant brain tumors can accurately predict drug responses in humans, and that these models can be prospectively developed within a short time frame (3-4 months) concurrent with global genomic analysis of the primary patient tumors in real time in a clinical setting. Our central hypothesis is that patient tumor-derived orthotopic xenograft tumors will respond to anti-cancer therapies similarly to the corresponding human primary tumors and can reliably predict drug responses in the patients. Additionally, rapid advancement of clinical whole exome sequencing and other genomic methods has created an urgent need for personalized animal models for timely and customized preclinical drug screenings. Our second hypothesis is that personalized orthotopic xenograft mouse models can be prospectively developed within 3-4 months for patients enrolled in a clinical whole exome sequencing study and utilized to test targeted therapies selected based on each patient's genomic data. To test these hypotheses, we propose two Specific Aims. In Aim 1, we will treat each of the existing xenograft mouse models in vivo with the same chemotherapeutic agents and/or ionizing radiation that were administered to the corresponding patients and to compare xenograft responses (tumor size shrinkage and survival time extension) directly with that observed in the matched original patient tumors. The doses and schedule of the patient treatment will be modified so that the drug concentrations in mouse serum are similar to those in human beings. We will also examine the underlying mechanisms of tumor resistance by analyzing the inter- and intra-tumoral differences of therapy responses both in vitro and in vivo, and by examining the efficiency of drug delivery into xenograft tumors in vivo in mouse brains. In Aim 2, we will determine the impact of greater cell viability and increased tumor cell number on shortening time frame of orthotopic xenograft tumor formation from the same brain tumor patients enrolled in the NHGR/NCI funded U01 Clinical Exploratory Sequencing Research (CSER) project, which is aimed to sequence all the newly diagnosed pediatric solid tumors at Texas Children's Hospital. We will confirm the preservation of patient primary tumor gene mutations in the xenograft tumors; and conduct customized preclinical drug screenings by targeting the tumor-specific (and druggable) genomic aberration(s) with or without combination with the standard therapies. Since all our models are patient-specific with treatment history and clinical outcome data from the matched original patient tumors, and we have optimized surgical procedures that allows for rapid and safe injection of tumor cells into various locations in mouse brains (>160 mice/day), we are well positioned to address the PQD-5 in a clinically relevant setting to provide accurate and objective evaluation of the predicative power of this model system.