Abstract

Akio Ogura*, Yui Taguch and Toru Takemura

ABSTRACT

Purpose: The subtraction method is highly accurate for measuring the SNR in MRI, and it is frequently used for this reason. However, in clinical settings, this technique is not used very frequently due to its requirement of obtaining the same image twice. The pixel shift method was reported, which we believe can overcome this issue adequately. The pixel shift method can be used to calculate the SNR with just one round of imaging, but the effect of the number of pixels involved in the pixel shift and its direction have not yet been investigated. Materials & Methods: The cephalic images of the participants and phantom images were obtained for the number of different additions. Differential images of the various pixel shift and their directions were created for each image and were considered for the optimal pixel shift method by comparison with the theoretical value of the number of additions. Results In the pixel shift method, the edge of the structure was emphasized in a differential image. The noise was overestimated for the ROI setting, including at the edge of the structure. However, we calculated the SNR with high precision, despite using a theoretical value for comparison without depending on the number of pixels involved in the shift or its direction, when we set the ROI at the position without any surrounding structures or tissues. Conclusion: We estimated the position of the ROI where there are no surrounding structures or tissue, and we assessed the effect of the number of pixels involved in the pixel shift and its direction on the efficacy of the method for the pixel shift method.