Investigating the accuracy of FatNav‐derived estimates of temporal B₀ changes and their application to retrospective correction of high‐resolution 3D GRE of the human brain at 7T

Purpose To investigate the precision of estimates of temporal variations of magnetic field achievable by double‐echo fat image navigators (FatNavs), and their potential application to retrospective correction of 3‐dimensional gradient echo–based sequences. Methods Both head motion and temporal changes of B 0 were tracked using double‐echo highly accelerated 3‐dimensional FatNavs as navigators, allowing estimation of the temporal changes in low spatial‐order field coefficients. The accuracy of the method was determined by direct comparison to controlled offsets in the linear imaging gradients. Double‐echo FatNavs were also incorporated into a high‐resolution, 3‐dimensional gradient echo–based sequence to retrospectively correct for both motion and temporal changes in B 0 during natural and deep breathing. The additional scan time was 5 min (a 40% increase). Correction was also investigated using only the first echo of the FatNav to explore the trade‐off in accuracy versus scan time. Results Excellent accuracy (0.27 Hz, 1.57–2.75 Hz/m) was achieved for tracking field changes, and no significant bias could be observed. Artifacts in the 3‐dimensional gradient echo–based images induced by temporal field changes, if present, were effectively reduced using either the field estimates from the double echo or the first echo only from the FatNavs. Conclusion The FatNavs were shown to be an excellent candidate for accurate, fast, and precise estimation of global field variations for the tested patterns of respiration. Future work will investigate ways to increase the temporal sampling to increase robustness to variations in breathing patterns. Magn Reson Med 80:585–597, 2018.