The most difficult challenge in treatment management of brain tumor patients is the need to accurately identify if a suspicious lesion on a post-treatment MRI scan is a benign treatment-effect or a “true” cancer recurrence. Both radiation effects and tumor recurrence have similar clinical symptoms and appearances on routine MRI scans. Currently, a highly invasive brain biopsy is the only option for confirmation of disease presence. Each biopsy procedure costs $20,000-$50,000/patient. Further, over 15% of patients who undergo biopsy will get an incorrect diagnosis due to difficulty in sampling of reliable locations of the tumor. There is hence a need for non-invasive image techniques to reliably differentiate benign treatment effects from tumor in brain tumor patients. Our team has developed new image-based biomarkers that use routine MRI scans to differentiate between these two conditions with an accuracy of 92% on n>200 studies. We propose to validate our image-based biomarkers in a limited clinical trial to reliably sample locations of tumor recurrence from benign radiation effects. The clinical trial will be based on creation of a “GPS” map of the locations of tumor and benign radiation necrosis in the tumor using MRI scans. This GPS map will assist neurosurgeons in reliably identifying locations to biopsy from during surgery. The proposed project, when successful, will thus have significant implications in personalizing treatment decisions in brain tumors.