Effects of MR Image Noise on Estimation of Short T2 Values from T2-weighted Image Series

Abstract

We examined how the noise from magnetic resonance (MR) imaging affects the calculation of T(2) in skeletal muscle, a tissue with short T(2) values. The measured pixel intensity of the MR image (: the magnitude image) was the superimposed signal which was composed of the MR signal and the noise, and we demonstrated that noise from a magnitude image matches the DC component of the T(2) decay curve. In materials with long T(2) values, the noise has no influence on the selective echo time (TE) in calculating T(2). However, in materials with short T(2) values, noise clearly influences the selective TE. In this study, we proposed a T(2) effective signal-ratio, T(2)SR, as an index for determining whether the noise of the magnitude image can be ignored in calculating T(2). When T(2)SR and the signal-to-noise ratio (SNR), an index of image quality, were compared as indices to evaluate the influence of noise in the calculation of T(2), T(2)SR was useful and SNR was not. The use of multiple spin echo (MSE) technique shortened imaging time, but required detailed understanding of the MSE. Our results indicated that T(2) can be calculated correctly for skeletal muscle and other tissues with short T(2) even when the receiver coil has a low SNR and few measurement points are available.