Our purpose was to develop an imaging method which could quickly estimate the homogeneity of the magnetic field using readily available, echo planar imaging techniques. This method is valuable both clinically and for routine quality assurance of the imager even when double echo gradient phase map imaging is not available.On a 1.5T Siemens Aera MR imager we acquired echo planar and gradient echo images of the ACR quality control phantom. We deliberately adjusted the linear and quadratic shim gradients to create known magnetic field inhomogeneities. For each condition, we acquired images with the phase encode gradient oriented in both positive and negative directions. We combined these two images by integrating the image intensity along each column in the phase encoding direction to produce one combined, distortion-free image and a second image corresponding to the amount of shift required to remove the distortion.3 Knowing the bandwidth in the phase encoding direction of the echo planar imaging technique we easily converted this image into a measurement of magnetic field inhomogeneity to within approximately 5%.Inhomogeneous magnetic fields leads to several problems in MRI including image geometric distortion, non-uniformity in image intensity and inhomogeneous fat suppression.² As a result the American College of Radiology asks for regular measurements of field homogeneity. Additionally, measuring field homogeneity is valuable in clinical situations to optimize image performance. When available the best method to measure field homogeneity is a phase contrast technique where the phase in a gradient echo sequence is measured at two echo times. In many cases, however, a phase contrast technique is not readily available except to the Field Service Engineer. We developed a method using echo planar images whose image fidelity is very sensitive to field homogeneity. We also developed image analysis methods to automatically estimate the distortion and from that infer the field inhomogeneity.Figure 1 shows the measured inhomogeneity derived from the measurement of the dimension of the phantom in the phase encode direction versus the nominal field inhomogeneity derived from the deliberate misadjustment of the Gy linear gradient. Figure 2 shows in each row, the composite collection of echo planar images acquired with the phase encode gradient oriented down or up, as well as an image derived from the distorted images and, a an image representing the displacement necessary to transform the distorted image into the undistorted-combined image. This set of four images is repeated for the three settings of the quadratic shim gradient, Gxy=X²-Y² of -50 µT/m², 0, or +50 µT/m². Figure 3 is a plot of the field inhomogeneity expressed as frequency offset in Hz. Note the expected quadratic shape of the field inhomogeneity arising from the X²-Y² gradient. Figure 4 plots the vertical profiles taken through the displacement images shown in figure 2 taken through the phantom. Plots are given for the conditions of Gxy= -50 µT/m², 0 , or +50 µT/m².The EPI method was successful in estimating linear field inhomogeneities accurately. By swapping the direction of the phase encoding gradient and acquiring two EPI images in rapid succession one can estimate the in-plane distortion in the phase encoding direction and hence, the field homogeneity over a large volume. The total acquisition time is on the order of 2 TR which can be as low as 5 sec. The method can also be applied clinically where it may be used to estimate the magnetic field homogeneity where movement from cardiac or respiratory motion would normally corrupt the measurement. The echo planar method for measuring field homogeneity can produce accurate estimates of in-plane magnetic field homogeneity throughout a large volume. The measurement time is short and the measurements can be performed on most MR imagers without special pulse sequences.