Echo planar imaging is very sensitive to field heterogeneity. This is the basis for detecting the gadolinium bolus. But it can also create problems, as seen on the following slide where the sphenoid sinus results in field distortion and signal loss. Gradient echo echo planar images are more susceptible to this problem.

While the susceptibility effect is desired to show where gadolinium is flowing, it also distorts the images in areas of field distortion, such as in this case, near the sphenoid sinus.

Correction of Susceptibility

Since the phenomenon producing perfusion images is based on susceptibility effects, one does not want to eliminate this effect. However, in cases where you may need to reduce this weighting, spin-echo echo-planar imaging can be used. In this case, it may be necessary to double the gadolinium dose to get a perceptible signal decrease.

GRE-EPI	DD SE-EPI

Ghosting Effects

In cases where there is inadequate sampling, one can see ghosting. As with other undersampling problems, this can be addressed by increasing the sampling, increasing the field of view, or sometimes, exchanging the phase and frequency directions. The GE-EP image on the previous page is an example of ghosting due to inadequate sampling.

Noise

The images produced with DCE-PWI have reasonable signal-to-noise ratio. However, it is not just the image itself that determines noise, but also the magnitude of the change that the bolus produces. This is affected by the amount injected, the rate it is injected, as well as the more standard imaging-determined factors.

Correction for Noise

A low pass filter can be used, but degrades edges. The "salt-and-pepper" nature of noise in these images is well addressed by median filters.

Original	Low Pass Filter	Median Filter