B -sensitivity analysis of quantitative magnetization transfer imaging.

PURPOSE

To evaluate the sensitivity of quantitative magnetization transfer (qMT) fitted parameters to B inaccuracies, focusing on the difference between two categories of T mapping techniques: B -independent and B -dependent.

METHODS

The B -sensitivity of qMT was investigated and compared using two T measurement methods: inversion recovery (IR) (B -independent) and variable flip angle (VFA), B -dependent). The study was separated into four stages: 1) numerical simulations, 2) sensitivity analysis of the Z-spectra, 3) healthy subjects at 3T, and 4) comparison using three different B imaging techniques.

RESULTS

For typical B variations in the brain at 3T (±30%), the simulations resulted in errors of the pool-size ratio (F) ranging from -3% to 7% for VFA, and -40% to > 100% for IR, agreeing with the Z-spectra sensitivity analysis. In healthy subjects, pooled whole-brain Pearson correlation coefficients for F (comparing measured double angle and nominal flip angle B maps) were ρ = 0.97/0.81 for VFA/IR.

CONCLUSION

This work describes the B -sensitivity characteristics of qMT, demonstrating that it varies substantially on the B -dependency of the T mapping method. Particularly, the pool-size ratio is more robust against B inaccuracies if VFA T mapping is used, so much so that B mapping could be omitted without substantially biasing F. Magn Reson Med 79:276-285, 2018. © 2017 International Society for Magnetic Resonance in Medicine.