Spinal cord artifacts from truncation errors during MR imaging.

The significance of linear regions of altered signal intensity that appear in sagittal magnetic resonance (MR) images along the length of the spinal cord was investigated. Examinations were performed on ten healthy volunteers and one patient with spinal cord edema. A 0.5-T or a 1.5-T MR system was used. Sampling-related effects (Gibbs phenomenon) at spinal cord edges and cerebrospinal fluid interfaces can lead to different signal patterns within the spinal cord and canal. These artifacts cause problems in interpretation, especially with the use of small object-to-pixel size ratios, by obscuring anatomy and simulating pathologic conditions such as pseudosyringes. Analysis of these intensity variations and of their dependence on sampling may improve the clinical accuracy of MR imaging.