Quantification and reproducibility assessment of the regional brain T relaxation in naïve rats at 7T.

Purpose: To measure the reproducibility of T relaxation and to determine the statistical power of T mapping in the rat brain as a characteristic of the baseline performance of the T relaxation as a potential biomarker of neurotoxicity.

Materials And Methods: Multislice multiecho spin-echo imaging was utilized to obtain the quantitative T maps in 138 naïve rats at 7T. Images were skull-stripped and coregistered to the common anatomical reference. A full anatomical segmentation mask, which included all major brain structures, was created using the same reference T map. The overall variability map was also calculated from all T maps and the areas with arbitrarily high variability (coefficient of variation >25%) were excluded from the full segmentation mask to produce a trimmed mask. T maps were segmented using both these masks and statistical power analysis was conducted in all segmented areas.

Results: The coefficient of variation of T relaxation in different brain areas varied from 5.4% (cerebrospinal fluid) to 1.2% (cortex) when using a full segmentation mask. The use of a trimmed segmentation mask decreased the coefficient of variation in many areas, which ranged between 3.2% (inferior colliculi) and 1.2% (cortex) in this case. As revealed by statistical power analysis to detect 5% change with power of 0.8, the minimum number of observations needed for different areas ranged from 3 (cortex) to 8 (inferior colliculi) in the case of use of a trimmed segmentation mask.

Conclusion: T relaxation is a very reproducible MRI parameter of the rat brain with high statistical power, which allows detecting very small changes in groups consisting of a minimal number of experimental animals.

Level Of Evidence: 1 J. Magn. Reson. Imaging 2017;45:700-709.