3D sodium (23 Na) magnetic resonance fingerprinting for time-efficient relaxometric mapping

Abstract

Purpose: To develop a framework for 3D sodium (²³ Na) MR fingerprinting (MRF), based on irreducible spherical tensor operators with tailored flip angle (FA) pattern and time-efficient data acquisition for simultaneous quantification of T₁ , $T_{2l}^{\ast }$ , $T_{2s}^{\ast }$ , and $T_2^{\ast }$ in addition to ΔB₀ .

Methods: ²³ Na-MRF was implemented in a 3D sequence and irreducible spherical tensor operators were exploited in the simulations. Furthermore, the Cramér Rao lower bound was used to optimize the flip angle pattern. A combination of single and double echo readouts was implemented to increase the readout efficiency. A study was conducted to compare results in a multicompartment phantom acquired with MRF and reference methods. Finally, the relaxation times in the human brain were measured in four healthy volunteers.

Results: Phantom experiments revealed a mean difference of 1.0% between relaxation times acquired with MRF and results determined with the reference methods. Simultaneous quantification of the longitudinal and transverse relaxation times in the human brain was possible within 32 min using 3D ²³ Na-MRF with a nominal resolution of (5 mm)³ . In vivo measurements in four volunteers yielded average relaxation times of: T_1,brain = (35.0 ± 3.2) ms, $T_{2l,\text{brain}}^{\ast }$ = (29.3 ± 3.8) ms and $T_{2s,\text{brain}}^{\ast }$ = (5.5 ± 1.3) ms in brain tissue, whereas T_1,CSF = (61.9 ± 2.8) ms and $T_{2,\text{CSF}}^{\ast }$ = (46.3 ± 4.5) ms was found in cerebrospinal fluid.

Conclusion: The feasibility of in vivo 3D relaxometric sodium mapping within roughly ½ h is demonstrated using MRF in the human brain, moving sodium relaxometric mapping toward clinically relevant measurement times.

Keywords: 7 Tesla; Cramér Rao lower bound; X-nuclei; magnetic resonance fingerprinting; relaxometry; sodium.