Phase Contrast MRI 2: The Revenge

Last week, I talked briefly about how we use successive slices in a cine MRI sequence to evaluate heart function. Using this type of image analysis, we are able to find information such as cardiac output and ejection fraction, as well as volume and mass information during various stages of the heart. In addition, we can use a technique called Phase Contrast MRI to obtain flow information. While the data obtained from Phase Contrast MRI is slightly different than the data obtained from segmenting a cine MRI, it is possible to correlate the two and from this comparison, we can determine the agreement between our two measurements.

Phase contrast works by exploiting the nature of MR imaging. Two gradients, with equal magnitude but opposite orientation, are applied to the area of interest in rapid succession. Objects that remain completely stationary undergo no net phase shift, as the opposing gradients “cancel out” the effects of one another. However, objects that are moving undergo a change in phase proportional to their velocity. This concept is illustrated in the diagram below.

Phase Contrast gradient fields on stationary and moving objects
(image obtained from radiographics.rsnajnls.org)

The top panels demonstrate the gradients on a stationary object. The first gradient “tips” the atoms by varying amounts, depending on their location within the gradient field. These tip angles are eliminated upon the application of the second gradient field. The end result gives the effect as if no gradient was applied at all, and all magnetic moments are uniform and in a single direction.

On the other hand, a moving object will undergo a predictable change in phase. The bottom panels illustrate what happens when a single atom travels linearly through the same region. The end result is a phase distinctly different than that of the stationary atoms.

This information can be visualized by viewing the phase information of a particular MR image sequence. Typically, a neutral gray color is assigned to pixels in which there is no net phase difference. Black indicates that the objects within the pixel are moving towards the viewer, while a white color indicates that objects are moving away. Varying velocities are indicated by the intensity of these shapes. From this, we can quickly determine the direction and amount of flow. Further, by using computer automation and manual segmentation, we can determine the amount of flow in specific areas.

In this study, we will be performing phase contrast analyses of the aorta. By quantifying the amount of blood flowing through the aorta, we can indirectly determine the stroke volume of the heart, which will provide an additional form of “ground truth” by which we can further evaluate our automatic segmentation algorithm.

Magnitude MR image ("Regular")
(image obtained from radiographics.rsnajnls.org)

Phase Contrast MR image (new and improved!)
(image obtained from radiographics.rsnajnls.org)