Correction for gradient nonlinearity in continuously moving table MR imaging

Abstract

Recently, a number of methods have been demonstrated for large field of view MR imaging using continuous table motion. As with conventional, fixed-table MRI, the spatial encoding is performed using magnetic field gradients. However, it is demonstrated in this work that as a consequence of every measurement being made at a slightly different displacement between the object and the gradient field, gradient nonlinearities are manifest as blurring in addition to spatial distortion. Moreover, the blurring is spatially dependent. It is also shown that correcting all phase-encoding steps individually or in groups can reduce these effects. (1)(2)(3)(4)(5)(6)(7)(8)(9). With these techniques the data must be spatially registered based on the table position. Because the pixel value at each location in the reconstructed image is formed from data acquired at multiple positions of the object within the gradient field, one might assume that artifacts from gradient nonlinearities manifest themselves differently for these types of acquisitions than for those with a fixed table position. To our knowledge, these gradientrelated effects have not been directly addressed in moving table imaging. Rather, the severity of these effects has been reduced by minimizing the FOV of the data acquisition.In MR image reconstruction it is assumed that there is a direct linear relationship between position and precessional frequency based on the gradient field strength. Any deviation from linearity results in a spatial distortion of the reconstructed image. Correction for this spatial distortion in imaging a fixed FOV is well understood and involves both a spatial remapping as well as an intensity correction (10,11). The technique used in this work is from Ref. 10 and is commonly referred to as "GradWarp."In one method of continuous moving table imaging in which frequency-encoding is along the direction of motion (4), 3D imaging volumes with a limited FOV are repeated multiple times while the table is moved continuously to extend the total imaging FOV. In this case, each phaseencoding step is first Fourier-transformed along the frequency-encoding direction and then registered along the same direction based on the distance the table has moved. This step is repeated for all phase-encoding steps in each acquisition. The resulting hybrid data that has been Fourier-transformed along the frequency-encoding direction can then be transformed along the other two dimensions to generate an imaging volume with a reconstructed FOV larger than the frequency-encoding FOV.In this work we describe the impact of gradient nonlinearity on continuous table motion acquisitions. First, we provide a qualitative description of the effects, followed by a more quantitative derivation of the point-spread function. We then present a technique for correction of such effects and demonstrate the performance of the correction in the specific application of table motion along the frequency-encoding direction. DESCRIPTIONThe effect of gradient nonlinearity in continuous table mot...

Continuous table acquisition MRI for radiotherapy treatment planning: Distortion assessment with a new extended 3D volumetric phantom

“…The technique is designed to reduce the distortion components from the z axis with no additional benefit in the axial plane. One angiography study using SMS 18 demonstrated additional distortion effects and blurring in all 3 imaging planes when utilising this acquisition technique. 105…”

Whole‐body magnetic resonance imaging featuring moving table continuous data acquisition with high‐precision position feedback

“…Here, averaging was performed in a weighted manner in hybrid k-space, reducing contrast variations due to gradient nonlinearity. In addition to averaging, the image quality can further be optimized by accurately correcting gradient nonlinearity (14). By implementing a gradient nonlinearity correction, a larger portion of the imaging FOV in the z-direction can be included in the data for reconstruction, thus decreasing necessary overlap of consecutive measurements and simultaneously increasing efficiency.…”

Extended field‐of‐view imaging with table translation and frequency sweeping

“…However, for incremental-FOV imaging with readout along the table direction, a significant cropping (i.e., a reduction in effective regional FOV length) may be necessary to make up for the inadequacy of existing methods for gradient nonlinearity effect correction. A recent study demonstrated improvements provided by the use of a new correction algorithm (12). Simulation studies and algorithm development are among our priorities for maximizing effective regional FOV (and hence scan efficiency) through advanced nonlinearity effect correction.…”