Longitudinal Radio-pathomic Mapping of Non-Enhancing Tumor Growth in Glioblastoma

About This Grant

Project Summary Glioblastoma (GBM) remains the most aggressive primary brain tumor with median survival of only 12-15 months despite intensive multimodal therapy. Current treatment paradigms focus primarily on contrast-enhancing tumor regions visible on MRI, failing to adequately address non-enhancing tumor components that harbor resistant cells and serve as reservoirs for recurrence. These non-enhancing regions, which can extend up to 10 cm beyond visible enhancement, significantly impact patient prognosis but remain poorly understood and inadequately targeted by current therapeutic approaches. The proposed research addresses this critical gap by leveraging the world's largest neuro-oncology brain bank to develop and validate advanced radio-pathomic mapping techniques that predict pathological tumor characteristics using standard clinical MRI. Aim 1 conducts retrospective analysis of tumor progression patterns, examining how clinical, demographic, and treatment factors impact enhancing and non-enhancing tumor growth dynamics. This includes identifying growth patterns predictive of survival outcomes, characterizing clinical modulators such as age, sex, and MGMT status, and assessing differential treatment response and efficacy on distinct tumor compartments. Aim 2 develops second- order radio-pathomic maps incorporating cellular morphometry and molecular markers including Ki-67, CD68, and SOX2 to improve non-enhancing tumor detection. This project then integrates first and second-order radio- pathomic feature maps to create comprehensive tumor habitat maps that characterize pathologically distinct microenvironmental regions. The expected outcomes include novel prognostic tools for personalized GBM management that substantially improve upon the pathological resolution of our previously developed models, allowing for the identification of previously unrecognized therapeutic targets in non-enhancing regions, and mechanistic insights into treatment resistance patterns. The radio-pathomic mapping approach will enable non- invasive monitoring of complete tumor burden throughout disease progression, potentially transforming clinical decision-making in neuro-oncology. By improving detection and characterization of non-enhancing tumor regions, this work addresses the fundamental limitation in current GBM imaging and treatment management that contributes to treatment failure and poor outcomes. The development of MRI-based tumor probability maps of second order features using standard clinical sequences ensures immediate translational potential and broad applicability across institutions without requiring additional scan time or specialized protocols, facilitating rapid clinical implementation and widespread adoption.