Microstructural differentiation of cerebral metastases, glioblastoma, meningioma and primary CNS lymphoma using advanced diffusion imaging techniques

Microstructural tumor characteristics discriminate metastases, glioblastoma, meningioma and primary CNS lymphoma. We aimed to assess these intracranial neoplasms utilizing multiparametric diffusion imaging as a translational measure of morphology.

We investigated 101 newly diagnosed intracranial tumors (35 metastases, 34 glioblastomas(GB), 21 meningiomas, 11 primary CNS lymphomas(PCNSL)) with advanced diffusion MRI including Diffusion Tensor Imaging (DTI), Neurite Orientation and Dispersion Imaging (NODDI), and Diffusion Microstructure Imaging (DMI). Beyond DTI-derived metrics (aD, FA, MD, rD), we extracted the NODDI and DMI intra-axonal (NODDI ICVF, DMI V-intra), extra-axonal cellular (DMI V-extra), and free water (NODDI ISO-VF, DMI V-CSF) fractions using a multi-compartment model. These metrics were read from contrast-enhancing tumor portions and compared across the entities

Various microstructural parameters served as effective discriminators in pair-wise comparisons: ISO-VF demonstrated high accuracy in distinguishing metastases from PCNSL (accuracy 90.13%) and meningiomas (accuracy 80.69%). aD was most accurate in discriminating GB from PCNSL (accuracy 89.57%) and meningioma from PCNSL (accuracy 74.03%), similar to MD which distinguished GB from meningiomas (accuracy 77.73%). FA performed best in discriminating GB from metastases (accuracy 83.11%). Discrimination on two axes of directionality and compartmentalization illustrate the comprehensive approach of tumor assessment.

Advanced microstructural imaging facilitates discrimination of four common intracranial neoplasms. Features such as cell density, extent of free water, and directional cellular elements are reflected in the diffusion metrics to varying degrees. As part of a first non-invasive assessment, they may direct early diagnostic and therapeutic procedures.