Effect of echo time and T2-weighting on GRASE-based T1w/T2w ratio measurements at 3T

Tissue contrast can be enhanced by dividing T1-weighted (T1w) images by T2-weighted (T2w) images to map the so-called T1w/T2w ratio, which has become an increasingly popular technique for quantifying brain tissue changes associated with neurodevelopment, aging, and a variety of neurodegenerative diseases. However, although it is self-evident that T1w/T2w ratios increase with the amount of T2-weighting in the T2w image - which is determined by the echo time (TE), all else being equal - longer TEs also reduce the signal-to-noise ratio (SNR) of the T2w images, and it is not clear how these SNR characteristics affect the reliability of T1w/T2w measurements. Therefore, the current study systematically investigated how different amounts of T2-weighting affected T1w/T2w measurements in order to determine whether there is an optimal amount of T2-weighting. T1w images were acquired from 10 neurologically healthy adults using a 3D turbo field echo (TFE) sequence, and a series of T2-weighted images were extracted from a multi-echo 3D combined gradient- and spin-echo (GRASE) sequence. Analyses of 12 anatomically defined brain regions revealed that both the mean and standard deviation of the T1w/T2w measurements increased exponentially with TE of the T2w images, and that T2w images with TE ≈ 120-160 ms yielded the most consistent/reproducible contrast between white matter ROIs and the whole-brain T1w/T2w signal. Furthermore, comparisons between T1w/T2w measurements and multi-component T2-relaxation myelin water fractions (MWFs) in the same brain regions revealed that T2w images with TE ≥ 160 ms drastically reduced the degree of correlation between T1w/T2w measurements and MWF. Overall, these findings suggest that: 1) there is a substantial trade-off between increased T1w/T2w contrast (based on longer TEs for the T2w images) and the reliability of quantitative T1w/T2w signals; and 2) the optimal TE for T2w GRASE scans is between 120 ms and 160 ms for calculating T1w/T2w ratios.