Generalized Q-space MRI reveals macroscopic patterns of tumor architecture in vivo

Current approaches for studying tumor activity in patients involve molecular characterization in excised tissue or biopsied samples. Recognizing that tumors are composed of heterogeneous arrays of cells and their environment, there is a compelling rationale to explore the macroscopic organization of tumor tissue. We present a novel methodology for probing the micro-structural constituents of tumors in vivo utilizing generalized Q-space MRI. This approach employs varying magnetic field gradients and diffusion sensitivities to yield voxel-scale probability distribution functions of proton diffusivity, and then maps multi-voxel cellular alignment with tractography. Using this methodology, we describe the presence of macroscopic organizational features in patients with head and neck cancers, specifically depicting regional differences between the geometrically coherent periphery and incoherent core region. Such methods may comprise a method for assessing attributes of tumor biology in vivo and for predicting the response of such tumors to various drugs and interventions.