The effect of residual Nyquist ghost in quantitative echo-planar diffusion imaging

Single-shot diffusion-weighted echo-planar imaging (EPI) is typically used for most clinical diffusion studies due to its low sensitivity to patient motion. Although the Nyquist ghost artifact in EPI can be substantially reduced, there is frequently a residual ghost with low signal intensity. As reported in this study, this residual ghost can produce severe artifacts when maps of the apparent diffusion coefficient (ADC) are calculated from single-shot echo-planar images. The artifacts presented in this paper appear as regions of apparently low ADC which simulate regions of reduced diffusion, but are in fact generated by b-value dependent Nyquist ghosts of the orbits. Data acquired in vivo were used to demonstrate that these artifacts can be avoided by including standard methods of spatial presaturation or fluid-suppression in the diffusion-weighted EPI protocol. In addition, phantom studies were used to illustrate how phase and amplitude variations in the ghost generate the artifacts and theoretical expressions, derived elsewhere, were used to provide a detailed understanding of the artifacts observed in vivo. The level of Nyquist ghost reported for the current generation of commercial scanners suggests that this is a general phenomenon which should be a consideration in all EPI-based diffusion studies. Magn Reson Med 42:385-392, 1999.