Probing tumor microenvironment with time-dependent diffusion MRS and machine learning based modeling in C6 Glioma

Motivation: Diffusion-weighted MR Spectroscopy(DW-MRS) enables access to diffusion properties of intracellular metabolites to characterize brain cell-specific microstructure, potentially reflecting cellular changes in tumor microenvironment(TME). Goal(s): To measure time-dependent diffusion kurtosis of intracellular metabolites and to compute biophysical model of cellular changes in TME. Approach: In-vivo DW-MRS and machine learning(ML)-based model were applied to measure tumor microstructural changes with different diffusion-times(Td) in 11 normal and 11 C6-glioma rat models. Results: Changes in neuronal(tNAA) and glial(tCho) metabolites' time-dependent diffusion indicate neuronal atrophy and glial activation within tumor core. Microstructural and metabolic alterations, including specific cell size changes during tumor growth, suggest increased invasiveness in malignant tumors. Impact: This work provides a unique insight into the Td-dependency of intracellular metabolites and water to probe the microstructural changes during the pathological tumor progression, which helps revealing spatial structural and metabolite profile to address the complexity of the TME.