Abstract
Diffusion tensor imaging (DTI) and its metrics, such as mean diffusivity (MD) and fractional anisotropy (FA), have been used to detect the extent of brain tumors and understand tumor growth and its influence on the surrounding tissue. However, there are conflicting reports on how DTI metrics can be used for tumor diagnosis. The physiological interpretation of these metrics in terms of tumor growth is also not clear. The objective of this study was to investigate the DTI parameters in two rat brain tumor models (9L and F98) with different patterns of aggressiveness by longitudinal monitoring of tumor growth and comparing the DTI parameters of these two tumor models. In addition to the standard DTI metrics, MD and FA, we measured other metrics representing diffusion tensor shape, such as linear and planar anisotropy coefficients (CL and CP), and orientational coherence measured by lattice index (LI), to characterize the two tumor models. The 9L tumor had higher FA, CL, and LI than the F98 tumor. F98 had a larger difference in anisotropies between tumor and peritumor regions than 9L. From the eigenvalues, it was found that the increase in CL and trace of the 9L tumor was due to an increase in the primary eigenvalue, whereas the increase in CP in the peritumor region was due to an increase in both primary and secondary eigenvalues and a decrease in tertiary eigenvalue. Our results indicate that shapeβoriented anisotropy measures, such as CL and CP, and orientational coherence measures, such as LI, can provide useful information in differentiating these two tumor models and also differentiating tumor from peritumoral regions. Copyright Β© 2007 John Wiley & Sons, Ltd.