Spiral MR Imaging

Figure 13. Signal loss due to turbulent flow in a stenosis phantom (A) using a conventional technique and (B) recovery of that signal using spiral imaging. Flow is downward at an average velocity of 50 cm/sec.

Conventional methods of data collection, along with the EPI and FSE methods described above, collect data with only the frequency-encoding field gradient on. Consequently, the spatial waves in a straight line through k -space (i.e., phase twists changing in one direction only) can be measured. The most notable reason is that image reconstruction with the Fourier transform is far more efficient when adding spatial waves measured in this fashion. However, modern computational power combined with unconventional thinking has led to a myriad of data collection `pathways' through k -space. One of the better known is a path that follows an Archimedean spiral. By appropriately changing the direction and magnitude of the field gradient, data can be measured along a spiral path in k -space, which can be more efficient than conventional measurements on `linear' paths.

Importantly, the slowly varying waves—at the center of k-space—are measured first. These waves have the greatest effect on the final appearance of the image. Motion during the data measurement period, which can corrupt the image, only affects the measurement of rapidly varying waves collected later in the measurement. Spiral MR images are then much less prone to image artifacts from motion than conventional scans (Fig. 13).

Figure 14. Echo-planar image from a stroke patient (A) without and (B) with the application of diffusion-sensitizing gradients. The cerebrospinal fluid and vitreous signal reduce the most because of the high mobility of water in these regions. Image analysis reveals that the signal in the small area of stroke did not reduce as much as that of surrounding brain tissue, indicating reduced water diffusion.