As in spin echo imaging, gradient echo techniques use an RF excitation pulse to create transverse magnetization. However, instead of applying a second RF pulse to refocus the magnetization and form a spin echo, the dephasing transverse magnetization (FID) following the excitation pulse is collected and used to form an image*. The FID is frequency-encoded using a readout gradient pulse, just as the echo is encoded in spin echo imaging. However, because there is no 180° pulse, the readout gradient must be preceded by a negative dephasing lobe to create a gradient echo at the center of the acquisition window (rather than two gradient pulses with the same polarity on either side of a 180° pulse as in spin echo). The gradient echo is encoded with both kx and ky information, and stored into k-space. (Fig. 1.23)

Figure 1.23 Pulse sequence timing diagram for a typical gradient echo sequence.

The main advantage of gradient echo sequences is that the TR can be made very short without exceeding SAR limits. These sequences lend themselves well to imaging applications that require short acquisition times. Because gradient echo sequences do not reverse the order of the spins by flipping them 180° around a transverse axis, the static dephasing effects discussed in Section 1.3.2 above are not refocused. The size of the signal collected is therefore determined by the T2* envelope of the FID, and images obtained using gradient echo sequences will therefore always have some intrinsic T2* weighting. Gradient echo sequences are not appropriate for use in regions of high magnetic susceptibility variation, such as around metal, or interfaces with air because of the high degree of signal loss associated with the T2* decay in these regions. However, it is often useful to perform a GRE sequence to look for small focal areas of increased susceptibility such as regions with increased hemosiderin content (e.g., pigmented villonodular synovitis, hematoma), or calcium deposition (e.g., calcifications around tendons or bursae).

The main determinants of contrast in a gradient echo sequence are the user's choices of flip angle, TE and TR. In addition, several pathways for magnetization exist by which contributions to the acquired echo can be formed. For long TR, the gradient echo sequence behaves very much like a spin echo sequence in that only transverse magnetization created by the latest excitation pulse contributes to the acquired signal. For short TR (less than the tissue T2) as is used in fast imaging applications, the next excitation pulse is played out before all of the transverse magnetization has decayed away through T2 processes. The gradient echo sequence plays out a train of excitation pulses ( pulses) separated by TR, where each pulse can create an echo from previously excited transverse magnetization. (Fig. 1.24)

Figure 1.24 Following each alpha pulse in a gradient echo sequence, a FID is formed as a result of that pulse. In addition, a Hahn echo can also be formed at the time of the alpha pulse, resulting from the train of alpha pulses that have already been played out. This echo can either be included in the collected signal and used to form an image (GRASS), or can be "spoiled" by offsetting the phase of the alpha pulses so that no coherent echo can be formed (SPGR).

These echoes will form at the time of the alpha pulses, and can be collected along with the most recently created FID to add significantly to the strength of the total MR signal. In general, flip angles for the pulses of much less than 90° are preferred to prevent saturation and to ensure that the next pulse has ample longitudinal magnetization available to excite. However, in the case when TR << T2 (and, hence T1), a large flip angle of 90° can be used to selectively enhance fluids. For very fast acquisition times, short TRs and small flip angles as low as 5° may also be used to avoid saturation effects. Gradient echo sequences fall into three main categories, each of which are tailored to provide more or less signal from each of these "coherence pathways", i.e. the echo pathways, or the FID. These three categories are referred to in GE terminology as GRASS, SPOILED GRASS (SPGR), and STEADY STATE FREE PRECESSION (SSFP). SSFP is not used routinely for orthopedic applications and therefore will not be discussed here.

---

* In any imaging application where gradients are necessarily used to provide spatial encoding of the signal, the gradients must be balanced to ensure that the magnetization is refocussed at exactly TE. The echo collected in a spin echo sequence is actually a combined spin and gradient echo.

Return to Table of Contents Advance to Next Section