Chapter 1: MRI Techniques for Orthopedic Imaging

Figure 19   Figure 20
     
Figure 1.19
Pulse sequence timing diagram for a typical FSE sequence. All 90 degree and 180 degree pulses are slice-selective. Each acquired echo is phase encoded into a separate line in k-space.		   Figure 1.20
Example of a FSE acquisition where the TEeff corresponds to the third collected echo (the initial FID is not collected in FSE imaging). This echo is used to fill the center of k-space, and has the strongest effect on image contrast. TEeff will be adjusted from the user? choice to a value that corresponds to the closest echo.
     
     
     

1.4.3 - Fast Spin Echo

Fast Spin Echo is a variation of spin echo that uses a train of 180? refocusing pulses to produce a series of Hahn echoes that are each phase-encoded into different ky lines of the same k-space (See Figure 1.19). This series of echoes is called an echo train. For an echo train length (ETL) equal to eight, eight lines of k-space are collected following a single 90? excitation pulse. The time to completely fill k-space is reduced from conventional spin echo by the ETL. Following the acquisition of an echo train, it is necessary to wait before exciting the same slice again to allow the magnetization to recover along +z through T1 relaxation. As in conventional spin echo, it is possible to interleave acquisitions from different slices within each TR interval. In FSE, the extra time spent acquiring data from each slice requires a longer TR to be selected, resulting in improved tissue contrast for T2-wtd images while still effectively using the entire TR interval to acquire data. The FSE acquisition scheme therefore allows much greater SNR per unit acquisition time than conventional spin echo. For FSE, the actual TE of the image is not precisely determined because the image is reconstructed from many lines of k-space that are acquired at multiple different TEs. An effective TE (TEeff) is defined by the TE of the echoes that are used to fill the center of k-space (Fig. 1.20).

These low spatial frequency (ky) lines dominate the contrast in the image. The acquired echoes from each echo train are ordered in k-space according to the user's choice of timing parameters, in particular, the TEeff and the ETL. The user inputs values for TEeff, Xres, ETL and RBW. The Xres and RBW together are the main determinants of the minimum ESP that can be achieved (the width of the 180? pulses is constant). Choosing a higher RBW, and/or decreasing Xres will result in shorter echo spacing. The scanner will use this minimum echo spacing as the actual spacing for echoes in the echo train. The echo that occurs closest to the user's choice of TEeff will be used to fill the center lines of k-space. Note that because the echo spacing has already been determined, this echo may not occur at the user's exact choice of TEeff. TEeff must be an integer multiple of the echo spacing. The actual TEeff for the image will be displayed correctly in the image annotation.

The contrast in a FSE image is mainly determined by the TEeff, but is inherently different from a conventional spin echo image even with TE set equal to TEeff and equal TR. This difference in contrast is due to the way k-space is filled for FSE. There are several different important factors that affect contrast in an FSE sequence. Magnetization transfer contrast is increased in FSE relative to conventional SE because of the many 180? pulses that are played out during each TR interval. Each slice feels increased off-resonance effects from the many refocusing pulses that are applied to its neighboring slices. The increased MTC results in reduced signal from articular cartilage and other tissues with effective magnetization transfer. In general, the MTC in FSE tends to increase the apparent T2-weighting in the image. In addition, the efficiency of magnetization transfer in fat is negligible, leading to an apparently higher signal from fat in the image relative to other tissues whose signal intensities are suppressed due to their more efficient magnetization transfer properties. A second contribution to increased signal from fat on FSE is that the T2 of fat is actually increased for short ESP due to the reduction of J-coupling effects for multiple 180? pulses with short ESP. In a conventional spin echo image, J-coupling in fat molecules decreases the apparent T2 of the fat, making the signal from fat artificially darker. In an FSE sequence, the train of 180? refocusing pulses disturbs the J-coupling mechanism, and the apparent T2 of fat is increased3,4.

FSE-XL is the most recent version of FSE on GE Signa scanners, and is able to achieve shorter echo spacing than FSE through use of narrower RF pulses, and more efficient use of the higher gradient strengths and faster slew rates that are available on newer systems. One of the most important advantages of using FSE-XL is that image blurring from T2 decay during the echo train can be greatly reduced. For the same ETL, the image blurring will be much smaller using FSE-XL compared to FSE. Alternatively, longer ETLs may be used without image degradation, because the total length of the echo train is comparable in the two cases. Blurring due to T2 decay is most pronounced in short-TE images because the center sections of k-space are acquired when the T2 decay curve is the steepest. With FSE-XL, the short echo-spacing allows PD-weighted images to be acquired with reduced blurring in the image by using a shorter echo train. PD-weighted FSE images with an intermediate weighting (TEeff between 30-36ms) have increased in use for the evaluation of articular cartilage pathology. With this TEeff, the articular cartilage is of intermediate signal intensity, in contrast with the high signal intensity seen in the adjacent fluid. Fluid is relatively bright on FSE despite a modest echo time due to the magnetization transfer contrast inherent to the FSE technique.


3 Henkelman RM, Hardy PA, Bishop JE et al. Why fat is bright in RARE and FSE imaging. JMRI 2:533-540, 1992.

4 Coherence transfer by isotropic mixing in Carr-Purcell-Meiboom-Gill imaging: Implications for the bright fat phenomenon in FSE imaging. MRM 35:506-513, 1996.