High Dynamic Range Turbo‐ FLASH B1 Mapping for Characterizing Multi‐Transmit Multi‐Receive RF Arrays, Validated at 7 T and 11.7 T
Magnetic Resonance in Medicine · 2027-01 · 97 (1) · 240-249
DOI: 10.1002/mrm.70569 ↗Abstract
ABSTRACT Purpose Transmit (B 1 + ) and receive (B 1 − ) sensitivities of parallel‐transmit/receive RF array are a prerequisite for modern ultra‐high‐field MRI. Their accurate mapping remains challenging for coils with high transmit dynamic range, particularly in regions with low combined‐mode B 1 + magnitude, where standard reconstruction is impaired by noise. This work presents a new B 1 + and B 1 − model‐based fitting approach for interferometric presaturated TurboFLASH (satTFL) data. With an optimized acquisition scheme, this method improves B 1 + /B 1 − mapping accuracy. Methods Voxel‐wise B 1 + and B 1 − are expressed as the product of a complex unit‐norm vector (phasor) and a scalar (efficiency). Phasors are estimated through singular value decomposition of satTFL signals, and efficiencies are recovered with fitting to the signal equation. This method was evaluated in phantom simulations for 7 T and 11.7 T 8Tx/32Rx RF arrays and compared with standard approaches. As it supports a variable number of pre‐saturated scans, different interferometric schemes were optimized for B 1 + accuracy. Finally, schemes yielding acceptable B 1 + error with minimal acquisition time were tested in vitro at 7 T and 11.7 T and in vivo at 7 T. Results The proposed method provided lower median B 1 + and B 1 − errors than conventional satTFL methods and fewer outliers; the number of voxels with error > 30% is reduced from several dozens to zero. In vitro results were consistent with the simulations, and in vivo maps demonstrated good quality. Conclusion The proposed method improves B 1 + and B 1 − mapping robustness to noise and enables more flexible satTFL acquisition scheme, allowing better trade‐offs between accuracy and scan time.
EUVIMED