The link between diffusion MRI and tumor heterogeneity: Mapping cell eccentricity and density by diffusional variance decomposition (DIVIDE)

• DOI: 10.1016/j.neuroimage.2016.07.038
• Publication Date: 2016-07-21T13:21:40Z
• Authors: Filip Szczepankie...; Danielle van Westen; Elisabet Englund; Carl-Fredrik Westin; Freddy Ståhlberg; Jimmy Lätt; Pia C. Sundgren; Markus Nilsson
• Keywords: Kurtosis

The structural heterogeneity of tumor tissue can be probed by diffusion MRI (dMRI) in terms the variance apparent diffusivities within a voxel. However, link between diffusional and is not well-established. To investigate this we test hypothesis that variance, caused microscopic anisotropy isotropic heterogeneity, associated with variable cell eccentricity density brain tumors. We performed dMRI using novel encoding scheme for decomposition (DIVIDE) 7 meningiomas 8 gliomas prior to surgery. was quantified from total mean kurtosis (MKT), DIVIDE used decompose MKT into components (MKA) (MKI). Diffusion evaluated fractional (FA) (μFA). Quantitative microscopy on excised tissue, where were structure tensor analysis nuclei segmentation, respectively. In order validate parameters they correlated corresponding derived microscopy. found an excellent agreement parameters; MKA (r=0.95, p<10-7) MKI (r=0.83, p<10-3). voxel-scale (FA, r=0.80, p<10-3) scale (μFA, r=0.93, p<10-6). A multiple regression showed conventional parameter reflects both density, therefore lacks specificity microstructure characteristics. obtained decomposing two contributions; only eccentricity, variance. primarily (mean±s.d.) MKA=1.11±0.33 vs MKI=0.44±0.20 (p<10-3), whereas gliomas, it mostly MKI=0.57±0.30 MKA=0.26±0.11 (p<0.05). conclusion, allows non-invasive mapping reflect density. These results constitute convincing evidence exists specific aspects dMRI. Decomposing effects facilitates improved interpretation as well macroscopic scale.