Quantitative muscle water T₂ mapping using RF phase-modulated 3D gradient echo imaging

To propose a motion robust 3D sequence for water T₂ ( $T{2}_{H_{2}O}$ ) estimation in skeletal muscle tissues.

A $T{2}_{H_{2}O}$ estimation method is proposed, using 10 image volumes acquired with a partially spoiled gradient echo (pSPGR) sequence, varying the RF phase-cycling increment and prescribed flip angle. The complex signal evolution is fit with a bi-component water/fat model to extract $T{2}_{H_{2}O}$ and account for B₁ and fat fraction confounders. Accuracy and precision were evaluated using numerical simulations. Cartesian and radial implementations of the sequence were tested. In phantoms, results were compared with reference spectroscopic and multi-spin echo imaging techniques. Several in vivo experiments evaluated robustness to B₁ field inhomogeneities, sensitivity to physiological and pathological variations in $T{2}_{H_{2}O}$ on the thigh muscles.

In phantoms, $T{2}_{H_{2}O}$ values were highly correlated with reference spectroscopy and multi spin echo values (R² > 0.8). In vivo, $T{2}_{H_{2}O}$ values were correlated with reference values in healthy controls (R² = 0.69) and pathological muscles (R² = 0.87) and were not affected by B₁ inhomogeneities (R² = 0.06). In the tongue muscle, a significant reduction in the SD of $T{2}_{H_{2}O}$ values was observed using the radial compared to the Cartesian pSPGR sequence (-28%).

The proposed approach provides efficient 3D $T{2}_{H_{2}O}$ estimation in skeletal muscle, including small moving organs like the tongue. This broadens the range of accessible targets for characterizing heterogeneous impairment of muscle tissue, while retaining durations compatible with clinical research.