Comparison of in vivo and ex vivo diffusion tensor imaging in rhesus macaques at short and long diffusion times

Abstract

Diffusion tensor imaging (DTI) is widely used to non-invasively study neural tissue micro-structure. While DTI tractography of large nerve fibers is well accepted, visualization of smaller fibers and resolution of branching fibers is challenging. Sensitivity of DTI to diffusion anisotropy can be further enhanced using long diffusion time that can provide a more accurate representation of the tissue micro-structure. We previously reported that ex vivo fixed brain DTI at long t(diff) (192 ms) showed improved sensitivity to fiber tracking compared to short t(diff) (48 ms) in 4% formalin-fixed non-human primate (NHP) brains. This study further tested the hypothesis that DTI at longer diffusion time improves DTI fiber tracking in the in vivo NHP brains on a clinical 3 Tesla MRI scanner. Compared to fixed brains, the in vivo ADC was larger by a factor of 5. Also, the white-matter FA was 28% higher in the in vivo study as compared to our ex vivo experiments. Compared to short t(diff), long t(diff) increased white-matter FA by 6.0±0.5%, diffusion was more anisotropic, tensor orientations along major fiber tracts were more coherent, and tracked fibers were about 10.1±2.9% longer in the corpus callosum and 7.3±2.8% longer along the cortico-spinal tract. The overall improvements in tractography were, however, less pronounced in the in vivo brain than in fixed brains. Nonetheless, these in vivo findings reinforce that DTI tractography at long diffusion time improves tracking of smaller fibers in regions of low fractional anisotropy.

Keywords: DTI; Fiber tracking; Fixed brain.; Fractional anisotropy; MRI; Non-human primate.

Fig. (1)

Stimulated Echo Acquisition Mode Sequence with diffusion gradients. The mixing time, TM, can be increased, thereby increasing Δ (duration for which diffusion gradients are applied), and thereby achieve longer diffusion times without increase in TE.

Fig. (2)

(a) ADC in the GM and WM in the in vivo and ex vivo experiments. Compared to the in vivo values, ADC in the WM was reduced by almost by 61 ± 5% and ADC in the GM decreased by 39 ± 8% in the ex vivo samples. (b) Comparison of FA in the in vivo and ex vivo samples. FA reduced by almost 28% in the WM in the ex vivo study while in the GM the decrease was less prominent at 10% compared to the in vivo FA. (Ex vivo data obtained from [11])

Fig. (3)

Geometric measures of diffusion: CL, CP and CS are plotted as three phase plot. (a) Comparison of short and long t_diff by STEAM from the same animal. (b) Comparison of DSE and long t_diff by STEAM from the same animal. Light red dots indicate diffusion in voxels at short t_diff = 50 ms, Green dots represent diffusion in the same voxels at t_diff = 200 ms. At long t_diff, diffusion tends to be significantly (p < 0.05) more linearly anisotropic. The red and green squares represent the mean diffusivity at t_diff = 50 ms and 200 ms respectively.

Fig. (4)

DTI tractography: For the ex vivo study, fiber tracking using (a) DSE, (b) STEAM48 and (c) STEAM192 is shown in the same sample. White arrows and a yellow bracket mark the regions of improvement (longer fibers) using long t_diff. In vivo (d) short and (e) long t_diff tractography results by STEAM sequence were obtained from the same animal. In vivo (f) DSE and (g) long t_diff by STEAM sequence were obtained from the same animal. Arrows mark the regions of improvement (longer fibers) using long diffusion time. (Ex vivo tractography obtained from experimental data in [11])