Accuracy in the quantification of chemical exchange saturation transfer (CEST) and relayed nuclear Overhauser enhancement (rNOE) saturation transfer effects.
AUTHORS
- PMID: 28272761 [PubMed].
- PMCID: PMC5490367.
- NIHMSID: NIHMS868776
ABSTRACT
Accurate quantification of chemical exchange saturation transfer (CEST) effects, including dipole-dipole mediated relayed nuclear Overhauser enhancement (rNOE) saturation transfer, is important for applications and studies of molecular concentration and transfer rate (and thereby pH or temperature). Although several quantification methods, such as Lorentzian difference (LD) analysis, multiple-pool Lorentzian fits, and the three-point method, have been extensively used in several preclinical and clinical applications, the accuracy of these methods has not been evaluated. Here we simulated multiple-pool Z spectra containing the pools that contribute to the main CEST and rNOE saturation transfer signals in the brain, numerically fit them using the different methods, and then compared their derived CEST metrics with the known solute concentrations and exchange rates. Our results show that the LD analysis overestimates contributions from amide proton transfer (APT) and intermediate exchanging amine protons; the three-point method significantly underestimates both APT and rNOE saturation transfer at -3.5 ppm (NOE(-3.5)). The multiple-pool Lorentzian fit is more accurate than the other two methods, but only at lower irradiation powers (≤1 μT at 9.4 T) within the range of our simulations. At higher irradiation powers, this method is also inaccurate because of the presence of a fast exchanging CEST signal that has a non-Lorentzian lineshape. Quantitative parameters derived from in vivo images of rodent brain tumor obtained using an irradiation power of 1 μT were also compared. Our results demonstrate that all three quantification methods show similar contrasts between tumor and contralateral normal tissue for both APT and the NOE(-3.5). However, the quantified values of the three methods are significantly different. Our work provides insight into the fitting accuracy obtainable in a complex tissue model and provides guidelines for evaluating other newly developed quantification methods.
Accurate quantification of chemical exchange saturation transfer (CEST) effects, including dipole-dipole mediated relayed nuclear Overhauser enhancement (rNOE) saturation transfer, is important for applications and studies of molecular concentration and transfer rate (and thereby pH or temperature). Although several quantification methods, such as Lorentzian difference (LD) analysis, multiple-pool Lorentzian fits, and the three-point method, have been extensively used in several preclinical and clinical applications, the accuracy of these methods has not been evaluated. Here we simulated multiple-pool Z spectra containing the pools that contribute to the main CEST and rNOE saturation transfer signals in the brain, numerically fit them using the different methods, and then compared their derived CEST metrics with the known solute concentrations and exchange rates. Our results show that the LD analysis overestimates contributions from amide proton transfer (APT) and intermediate exchanging amine protons; the three-point method significantly underestimates both APT and rNOE saturation transfer at -3.5 ppm (NOE(-3.5)). The multiple-pool Lorentzian fit is more accurate than the other two methods, but only at lower irradiation powers (≤1 μT at 9.4 T) within the range of our simulations. At higher irradiation powers, this method is also inaccurate because of the presence of a fast exchanging CEST signal that has a non-Lorentzian lineshape. Quantitative parameters derived from in vivo images of rodent brain tumor obtained using an irradiation power of 1 μT were also compared. Our results demonstrate that all three quantification methods show similar contrasts between tumor and contralateral normal tissue for both APT and the NOE(-3.5). However, the quantified values of the three methods are significantly different. Our work provides insight into the fitting accuracy obtainable in a complex tissue model and provides guidelines for evaluating other newly developed quantification methods.
Tags: 2017