Improved calculation of the equilibrium magnetization of arterial blood in arterial spin labeling

Purpose: To propose and assess an improved method for calculating the equilibrium magnetization of arterial blood (M-0a), used for calibration of perfusion estimates in arterial spin labeling.Methods: Whereas standard M-0a calculation is based on dividing a proton density-weighted image by an average brain-blood partition coefficient, the proposed method exploits partial-volume data to adjust this ratio. The nominator is redefined as the magnetization of perfused tissue, and the denominator is redefined as a weighted sum of tissue-specific partition coefficients. Perfusion data were acquired with a pseudo-continuous arterial spin labeling sequence, and partial-volume data were acquired using a rapid saturation recovery sequence with the same readout module. Results from 7 healthy volunteers were analyzed and compared with the conventional method.Results: The proposed method produced improved M-0a homogeneity throughout the brain in all subjects. The mean gray matter perfusion was significantly higher with the proposed method compared with the conventional method: 61.2 versus 56.3 mL/100 g/minute (+8.7%). Although to a lesser degree, the corresponding white matter values were also significantly different: 20.8 versus 22.0 mL/100 g/minute (-5.4%). The spatial and quantitative differences between the 2 methods were similar in all subjects.Conclusion: Compared with the conventional approach, the proposed method produced more homogenous M-0a maps, corresponding to a more accurate calibration. The proposed method also yielded significantly different perfusion values across the whole brain, and performed consistently in all subjects. The new M-0a method improves quantitative perfusion estimation with arterial spin labeling, and can therefore be of considerable value in perfusion imaging applications.