Optimizing pTILT perfusion imaging in the presence of off-resonance frequency

To optimize the perfusion measurements and to recover loss of tagging efficiency from the blood flow measurements in the presence of static field inhomogeneity with the pseudo-continuous arterial spin labeling technique of pTILT, which is sensitive to off-resonance effects due to the employment of concatenated radiofrequency (RF) pulses for labeling.

Numerical Bloch simulations were performed to explore the labeling responses of concatenated RF pulses (45°, ±45°) in pTILT as a function of off-resonance frequency. A correction method was proposed by curve-fitting the measured blood flow signal to the signal model obtained from the simulations. The performance of the proposed correction method was examined for three healthy subjects on a 3 Tesla magnet after good shimming as well as under a "worst-case" scenario with deliberate mis-shimming.

In cases after good shimming, in which the off-resonance frequency was less than 50 Hz, the perfusion measurements by the optimized pTILT sequence were improved by 12.7% compared with the original pTILT. With a "worst-case" mis-shimming and frequency offset by 80 Hz, the blood flow signal was increased by 68.9% and 64.8% for optimized global and localized pTILT, respectively.

Addressing the impact of off-resonance frequency on concatenated RF pulses, tagging efficiency can be effectively recovered in pTILT perfusion measurements. This strategy may be extended to other applications of concatenated RF pulses where sensitivity to magnetic field inhomogeneity prevents accurate quantification.