Abstract
The most common methods to determine the signal-to-noise ratio (SNR) of magnetic resonance imaging (MRI) are based on the signal statistics in regions of interest (ROIs) in a magnitude image. For this, methods to calculate the ROI have still several discussions; we assumed SNR of a magnitude image could be estimated from standard deviation of a phase image (the phase method). The purpose of this study is to evaluate the validity of the phase method to determine the SNR. The simulation using digital phantom was carried out for evaluation of the effect to measure SNR; fluctuation in the phase image and SNR of the magnitude image. The phantom study was also performed for evaluation of the validity of estimation using the phase method in comparison with the conventional method (the identical ROI method and the subtraction method). The result of the simulation showed that SNR of magnitude image is larger than 4 for the SNR measurement using the phase method and this results reliable. The influence of fluctuation of the phase image should be eliminated for practical purposes. In the phantom study, phase method showed similar results compared to conventional methods in condition with elimination of the fluctuation of the phase image. Though there was a difference in the results of the phase method and the subtraction method according to the position of the ROI, the error was less than 4%. In this study, the method using the phase image to determine the SNR was identified as valuable.
