The most common method for creating perfusion-weighted images with MRI is to use a T2*-weighted, echo planar pulse sequence.

While we typically think of gadolinium contrast as resulting in changes perceptible on T1-weighted images, Gadolinium also results in T2* shortening, producing a loss of signal on T2* images.

The amount of decrease is related to the amount of gadolinium, which, in turn, is related to the amount of blood passing through a voxel of tissue.

PWI Source Images

Below is a movie loop during the injection of gadolinium. Note the decrease in signal, particularly in gray matter, as the bolus passes through the brain.

Signal Profile

If the signal intensity of a certain voxel is graphed over time, the decrease in signal can be appreciated (Note the high intensity at timepoint 1, when the tissue has not yet been RF saturated):

Post-Processing Steps

Unlike MRA or diffusion images where inspection of the source or raw images may be helpful, viewing the source images from a perfusion sequence is rarely useful, other than to confirm possible artifacts. Rather, the images are processed to produce perfusion-weighted images, including relative cerebral blood volume (rCBV), mean transit time (MTT), or cerebral blood flow (CBF).

This poster will focus on rCBV maps, though since all are derived from the same source images, artifacts and correction methods are similar.

rCBV Images

When gadolinium passes through tissue, it causes a decrease in signal on T2* images, formally known as “delta-R2” :

DR₂(t)=-1/TE In[ S(t) / S₀ ]

The total volume of blood in a unit of tissue (CBV) is proportional to the reduction of signal. Because the gadolinium does not arrive at one instant, it is necessary to integrate the total reduction of signal.