Parameter Lexicon

Indirectly, it cuts down on the scanning time, because a higher bandwidth increases the maximum number of slices available / acquisition. Consequently, the TR necessary for slice coverage will decrease.

Increasing the bandwidth is important to compensate for the presence of metallic hardware.

An additional advantage of increasing the Bw within the context of multi-echo sequences stems from the fact that the blurring effect decreases, partially offsetting the often-debated effect that blur may have in the interpretation of meniscal pathology.

The drawback to increasing the bandwidth is that there is a secondary cut in SNR. This reduction in SNR is not as dramatic as it may be expected, because it is a function of the square root of the change. This is due to a consequent decrease in the TEeff.

Decreasing the Bw results in higher SNR. Since the MSK system does not place the same challenges as body imaging (respiration, cardiac motion), the Bw can be kept low to maximize the SNR.

GE has a 'variable bandwidth' imaging option. When selected, the bandwidth will be optimized according to the matrix and the coil with +/- 15.6 KHz / field for 256 matrix (read-out direction), and +/- 32 KHz for 512 matrix (read-out direction). In older platforms, selection 'variable bandwidth' automatically would gray-out the Bandwidth box; this is no longer true: even when 'variable bandwidth' is selected, a specific bandwidth can be typed into the bandwidth box. Any typed bandwidth will supersede the 'variable bandwidth' default values.

Concatenation:

Whenever the number of slices exceeds the TR-related maximum, the number of acquisitions will be doubled, with a consequent marked increased in the scanning time. However, the two acquisitions will be scanned as part of the same series without need for pre-scanning of the second stack of slices.

Interpolation:

Zero-filling of K-space presents the information gathered as though it had been acquired with a higher matrix. The advantage is that it enhances the definition of anatomical structures; the drawback lies in its decreased SNR. However, the drop in SNR is not nearly as prominent as if the K-space information had been acquired with an equivalent matrix to that provided by the zero-filling feature.

The second point to remember is that this feature does not work with conventional spin echo.

For a zip 512, the frequency matrix will be first affected. For example, if the original matrix is 256 x 192, and zip 512 is turned on, the final apparent matrix will be 512 x 192. However, if the selected matrix is 512 x 256, and the zip 512 feature is turned on, the zero-filling will work on the phase matrix, bringing up the apparent matrix to 512 x 512.

GE systems are designed to turn on automatically the zip 512 feature whenever the matrix selected is higher than 256. So, if the matrix is 312 x 312, the apparent matrix will be 512 x 512, regardless of whether the zip 512 is turned on or off.

Interleaving:

Slice selection is a function of the gradients and the profile of the radiofrequency pulses; this process is less than mathematical perfect, leading to excitation of adjacent slices, and eventual saturation of spins outside the selected slice.

This problem can be addresses by maintaining a safety gap between scans. Alternatively, the gap between slices can be completely eliminated by acquiring interleaved slices. The obvious benefit is the elimination of volume-averaging - or the maximization of out-of-plane resolution. However, the differences in signal intensity between slices are marked, forcing constant readjustments in window levels between scans.

T2-weighting:

The tenet is that a TR of > 3000 msec. is necessary to obtain a T2 weighted sequence with FSE. The reason is that the slowest recovery time is for water (TR = 1 sec. at 1.5 T). Since a 95% recovery is to be expected at 3 x T1, it will take 3 seconds (or 3000 msec.) to get rid of all T1 contrast in the image.

This is theoretical, though. Fluid sensitivity is very apparent on intermediate sequences with TR values <3000 msec.

When creating protocols, TR values of >3000 msec. can be achieved without paying a time penalty by increasing the ETL. An increased ETL will cut down the acquisition time, and also the maximum number of images per acquisition. Therefore, the TR will naturally have to be brought up to compensate for the decreased number of slices.

Asymmetric FoV (rectangular FoV):

1. Variable FOV: The default phase-to-frequency FOV ratio is 1:1. However, a lower ratio can be chosen when the anatomy in the phase direction is smaller than the FOV, an asymmetric FOV can be chosen, and the amount of data acquired in the phase direction will be reduced. E.g. with a 48 cm FOV, a phase FOV selection of 1:2 will result in the acquisition of only the middle 24 cm of data in the phase direction. "No Phase Wrap" is not available with asymmetric FOV; hence it can only be used when wrap-around is not an issue: make sure anatomy doesn't extend beyond FOV. The scan time does not depend on the FOV, but on the actual matrix.

In the example above, with a ratio of 1:2, a selected matrix of 256 x 256 would result in an actual matrix of 256 x 128, and square pixels. The actual resolution for this case is 256 x 256. With the same ratio of 1:2, a selected matrix of 256 x 128 results in an actual matrix of 256 x 64, and rectangular pixels. The actual resolution would be 256 x128. Here, the ratio dominates, and is going to slice the number of phase steps accordingly, but won't vary the relative size of the pixels.

2. Square pixels: The same effect can be achieved by selecting the "square pixel" function. With this function selected, a 256 x 128 would result in an actual matrix of 256 x 128, but with a resolution of 256 x 256. The ratio of frequency: phase would be 1:2. Just imagine square cobblestones with only half the number of rows in one direction. The final ratio is determined by the number of phase steps and the selection of "square pixel."

Fast Recovery FSE:

Applies a -90 degree pulse at the end of the echo-train to eliminate residual transverse magnetization, resulting in an increase in the number of slices/TR.

T1 FLAIR:

Improves the T1 contrast of the image. Results in an increased imaging time of 30'' to 1', and has the potential to interfere with signal from gadolinium.