Purpose: A uniform static magnetic field (B₀) is imperative in clinical MRI applications to minimize artifacts such as geometric distortion, signal loss, banding artifacts, and failures in frequency-selective techniques. Several methods to measure B₀ homogeneity (ΔB₀) have been developed. The purpose of this work is to assess the variability in measurements from different methods and potential contributing factors.
Methods: ΔB₀ was measured by multiple users and scanner types. Four methods to measure ΔB₀ were analyzed: Spectral Peak (N=41), Bandwidth (N=45), Phase Difference Map (N=55), and Field Map (N=35). This dataset was retrospectively analyzed: average values of ΔB₀ and coefficients of variation (CV) were calculated for each method. These data were further grouped by field strength and phantom size and tested for significant differences to determine any contributing factors of ΔB₀ variability.
Results: Average ΔB₀ in ppm and CV of the four methods were as follows: Spectral Peak (0.93, CV=0.68), Bandwidth (1.23, CV=0.82), Phase Difference Map (1.52, CV=0.43), Field Map (1.45, CV=0.43). Spectral peak was reported as root-mean-square, while the other methods were reported as peak-to-peak. The Bandwidth difference method had the highest CV while the two mapping methods had the lowest. Measurements for each method grouped by field strength did not differ significantly. ΔB₀ was significantly lower (p=0.0464) when measured in smaller phantoms (16-29 cm) compared with larger phantoms (30-45 cm) using the Spectral Peak method but did not differ significantly when measured with the other methods.
Conclusion: This analysis of ΔB₀ measurements using four common methods serves as an initial reference for those conducting their own homogeneity tests. These initial data are highly variable, which may be explained by differences in scanner technology and reflect variability in methods across physicists. However, as more data points are added to the study, these reference values can be reported more precisely.