Abstract
PURPOSE: To evaluate the effect of RF saturation parameters and different amide proton transfer (APT) quantification methods on human brain tumor contrast. METHODS: Chemical exchange saturation transfer (CEST) images were acquired with varied RF saturation strengths, saturation times, and duty cycles (DC) at 3 T from brain tumor patients with glioblastoma. Different quantification methods, including magnetization transfer ratio asymmetry (MTR(asym)), apparent exchange-dependent relaxation (AREX), and extrapolated semisolid MT reference (EMR), were compared for measuring APT. The effects of different MT contrast (MTC) lineshape assumptions, such as Gaussian vs. Lorentzian or super-Lorentzian, and symmetric vs. asymmetric, were also evaluated with respect to APT contrast in human brain tumors. RESULTS: Overall, higher RF saturation strengths (> 1.5 μT) and longer saturation times with 100 % DC resulted in better enhancement of APT tumor contrast. The asymmetric super-Lorentzian MTC lineshape assumption was found to be the most accurate for in vivo brain tissue and provided the best fit for the acquired data. The EMR technique provided highly positive contrast for brain tumors, although still influenced by water relaxation effects. CONCLUSIONS: RF saturation parameters and APT quantification methods can greatly influence brain tumor contrast. The choice of an APT quantification metric must be carefully considered according to RF saturation parameter settings. The EMR-based CEST calculation metric demonstrated high contrast for brain tumors, making it a powerful clinical biomarker for tumor detection.