Effects of magnetic susceptibility artifacts and motion in evaluating the cervical neural foramina on 3DFT gradient-echo MR imaging.

The purpose of this study was to evaluate in vitro the effects of magnetic susceptibility and motion on the estimation of neural foraminal diameter with three-dimensional Fourier transformation (3DFT) gradient-recalled MR imaging as compared with CT. A cervical spine phantom was constructed from desiccated human cervical vertebral bodies embedded in a water-based proteinaceous gel. The phantom was imaged with thin-section 1.5-mm axial CT and 1.5-mm axial 3DFT gradient-recalled MR using a constant TR (35 msec) and flip angle (5 degrees), while the TE was varied from 11 to 22 msec. During imaging, the phantom either was kept stationary or underwent subtle, intermittent motion. Compared with CT, MR consistently underestimated the diameters of the neural foramina, leading to overestimation of neural foraminal stenosis. The degree of over-estimation varied directly with increasing TE values, from 8% (TE = 11 msec) to 27% (TE = 22 msec). Motion artifacts also increased foraminal overestimation and mimicked osseous hypertrophy. The effect of image degradation due to motion was noted to increase with longer TE values. Image degradation caused by magnetic susceptibility and motion artifacts can be minimized by using the shortest TE possible. We do not recommend the use of 3DFT gradient-recalled MR imaging for the evaluation of cervical radiculopathy if patient motion is anticipated.