Purpose: To introduce a novel time-efficient method, termed true constructive interference in the steady state (trueCISS), that not only solves the problem of banding artifacts for balanced steady-state free precession (bSSFP) but also provides its genuine, that is, true, on-resonant signal.
Methods: After a compressed sensing reconstruction from a set of highly undersampled phase-cycled bSSFP scans, the local off-resonance, relaxation time ratio, and equilibrium magnetization are voxel-wise estimated using a dictionary-based fitting routine. Subsequently, on-resonant bSSFP images are generated using the previously estimated parameters. Due to the high undersampling factors used, the acquisition time is not prolonged with respect to a standard CISS acquisition.
Results: From a set of 16 phase-cycled SSFP scans in combination with an eightfold undersampling, both phantom and in vivo whole-brain experiments demonstrate that banding successfully can be removed. Additionally, trueCISS allows the derivation of synthetic bSSFP images with arbitrary flip angles, which enables image contrasts that may not be possible to acquire in practice due to safety constraints.
Conclusion: TrueCISS offers banding-free bSSFP images with on-resonant signal intensity and without requiring additional acquisition time compared to conventional methods. Magn Reson Med 79:1901-1910, 2018. © 2017 International Society for Magnetic Resonance in Medicine.
Keywords: SSFP; balanced steady-state free precession; compressed sensing; iterative reconstruction; phase cycling; sparse sampling.
© 2017 International Society for Magnetic Resonance in Medicine.