Motivation: In voxel-wise, whole-brain studies of quantitative susceptibility mapping(QSM), variations in background field removal(BFR) methods, inversion algorithms, reference choices, and statistical indicators(signed or absolute) can lead to divergent findings and affect the reliability of results. Goal(s): To investigate the impacts of various BFR methods, inversion algorithms, reference choices, and statistical indicators on the outcomes of voxel-wise analysis. Approach: Voxel-wise analysis compared three BFR methods, two inversion algorithms, two reference choices, and signed/absolute values to assess susceptibility differences between young and elderly healthy subjects. Results: These combinations of BFR methods, inversion methods, reference choices, and statistical indicators showed varying sensitivity to age-related susceptibility changes. Impact: This study revealed inconsistencies in results arising from factors such as reconstruction methods, reference choices, and statistical indicators in voxel-wise QSM analysis. Careful consideration of these factors is essential to improve accuracy and reliability in future voxel-wise QSM research.
Tang et al. (Tue,) studied this question.