Vessel radius mapping for realistic vascular architecture in a U87-glioblastoma xenograft model

Purpose: To produce cerebral vessel radius maps that incorporate both susceptibility and diffusion effects in random vessel geometries to monitor and quantify microvascular changes in glioblastoma multiforme. Methods: The cerebral microvascular arrangement is codified in the free induction decay through local differences in magnetic susceptibility between blood-filled capillaries and the surrounding tissue that generates local magnetic field inhomogeneities. We have developed a model that considers both diffusion and susceptibility effects around randomly positioned and oriented arrangements of vessels that provides a connection between relaxation rate R2* and voxel-specific capillary radius R. The model is an extension of the strong-collision approximation (SCA) model (1). Results: Relaxation rates are shown versus capillary radii for a set of typical model parameters in brain tissue (Fig. 1): the model is validated against a numerically simulated signal decay on randomly positioned and oriented vessels (2). Radius maps are produced in a U87-glioblastoma xenograft mouse model (3) based on a multi-gradient-echo sequence (Fig. 2) that show significantly larger radii in glioblastoma tissue compared to contralateral healthy tissue of 6.42 [+ or -] 0.55 pm vs. 2.81 [+ or -] 0.44 pm (n = 7, p = 0.011). Conclusion: The obtained vessel radius maps can be used to evaluate capillary networks in healthy and pathological brain tissue.