Correction of Rf Inhomogeneities in Flash-based T1 Mapping Using Unified Segmentation

Introduction: Quantitative T1 mapping based on 3D FLASH acquisitions with variable excitation flip angles (VFA) is fast and robust (1, 2). However, high accuracy can only be achieved when the local flip angle is known precisely (1). At higher static magnetic fields local flip angles may deviate considerably from the nominal flip angle due to inhomogeneities of the RF transmit/B1 field (3, 4). T1 maps can be corrected using maps of the local RF transmit field (1, 5), requiring the availability of precise and accurate whole-brain B1 mapping methods (3) and adding experimental time. Here we propose a method that corrects the bias in T1 maps due to RF inhomogeneities using post-processing only and does not require knowledge of local flip angles which must be provided by additional data. The method exploits the facts that 1) the local flip angle simply scales the local T1 value (1), 2) B1 distribution varies smoothly across the brain (3), and the assumption that 3) T1 values in gray/white matter and CSF are normally distributed and do not vary significantly across the brain. We compare the performance of the proposed model-based correction with the established correction based on measured RF transmit maps (1, 3).