Brain tumors are a heterogeneous group of neoplasm with a diverse histological, molecular, and genetic spectrum and a widely variable clinical course and prognosis. Although most brain tumors, especially the malignant variety, remain difficult to cure, there are promising novel therapies and drug delivery systems that are under active investigation. One of the greatest challenges in developing effective therapy for brain tumors is the lack of specific markers to directly and accurately assess antitumor effect early and noninvasively. Further challenge lies in the fact that early treatment response can be transient and may not necessarily translate into long-term response or a favorable clinical outcome. In addition, there may be a small window of opportunity to assess therapeutic efficacy so that ineffective toxic therapy can be switched over to more effective therapy before there is widespread damage to the normal brain. The search for reliable and accurate predictors of treatment outcome that can be used to guide therapy and to improve survival in patients in malignant brain tumor has continued over several decades with modest success. This article will provide a general overview and current status of using quantitative maps derived from physiology-based magnetic resonance imaging to assess therapy response and to predict clinical outcome early during the course of therapy.