Regional phase correction of inversion-recovery MR images.

Many MR imaging systems are limited in their ability to successfully display inversion-recovery images. The reason is that part of the contrast is encoded as phase differences between pixels, whereas in the more commonly used spin-echo pulse sequence all the information is contained in the pixel magnitude. Inversion-recovery images are often displayed in magnitude form, resulting in loss of potentially useful phase information contained in the data. Before this phase information can be used, phase errors which result from scanner imperfections must be removed. While most of the necessary correction can be accomplished using data obtained by scanning a uniform phantom, this approach has several disadvantages. An alternative method by which phase errors can be readily removed without phantom data is described. This method has been applied to images of the head, knee, and liver with good results. It is concluded that this technique is useful for producing phase corrected inversion-recovery MR images.