Abstract
PURPOSE: To determine the feasibility of tagged MRI at 0.55 T and tag persistence for cardiac and speech production applications. This is particularly challenging due to the short muscle T(1) at 0.55 T (701 ms) and requires advanced techniques for maximizing initial tag contrast and signal-to-noise ratio (SNR)-efficient imaging. METHODS: Tagged MRI was performed using phase-sensitive tagging (REALTAG) two-dimensional grid-based preparation and spiral balanced steady-state free precession (bSSFP) imaging. Experiments were performed in healthy adult volunteers using a 0.55 T MRI system with high-performance gradients on two functional anatomies. Cardiac scans had 7.5-mm tag spacing, 1.3-mm(2) in-plane resolution, and were acquired during about 18s breath-holds. Speech scans had 10-mm tag spacing, 2-mm(2) spatial resolution, and were acquired in real time. Resulting tagged images were compared with conventional gradient-recalled echo (GRE). Tag persistence and tag contrast-to-noise ratio were measured and compared. RESULTS: REALTAG preparation and phase-sensitive reconstruction resulted in taglines that recover quickly when compared to 1.5 T, for both bSSFP and GRE imaging approaches. The bSSFP approach, at optimal flip angles, provided a 1.35-fold improved SNR and 1.6-fold improved tag persistence. For a tag contrast-to-noise ratio threshold of 6, the tag persistence was about 506 ms for GRE and about 851 ms for bSSFP. CONCLUSION: Tagged MRI is feasible at 0.55 T and can be used for functional analysis of muscle motion such as cardiac and speech applications, despite the short T(1). Spiral bSSFP along with REALTAG preparation pulses provide adequate SNR and contrast to visualize cardiac systolic mechanics and to visualize tongue deformation during short speech utterances.