Purpose: To enable clinical applications of quantitative magnetization transfer (qMT) imaging by developing a fast method to map one of its fundamental model parameters, the bound pool fraction (BPF), in the human brain.
Theory and methods: The theory of steady-state MT in the fast-exchange approximation is used to provide measurements of BPF, and bound pool transverse relaxation time ( ). A sequence that allows sampling of the signal during steady-state MT saturation is used to perform BPF mapping with a 10-min-long fully echo planar imaging-based MRI protocol, including inversion recovery T1 mapping and B1 error mapping. The approach is applied in 6 healthy subjects and 1 multiple sclerosis patient, and validated against a single-slice full qMT reference acquisition.
Results: BPF measurements are in agreement with literature values using off-resonance MT, with average BPF of 0.114(0.100-0.128) in white matter and 0.068(0.054-0.085) in gray matter. Median voxel-wise percentage error compared with standard single slice qMT is 4.6%. Slope and intercept of linear regression between new and reference BPF are 0.83(0.81-0.85) and 0.013(0.11-0.16). Bland-Altman plot mean bias is 0.005. In the multiple sclerosis case, the BPF is sensitive to pathological changes in lesions.
Conclusion: The method developed provides accurate BPF estimates and enables shorter scan time compared with currently available approaches, demonstrating the potential of bringing myelin sensitive measurement closer to the clinic.
Keywords: 2-pool model; EPI; bound pool fraction; magnetization transfer; myelin; quantitative steady-state.
© 2019 International Society for Magnetic Resonance in Medicine.