The Effect of Axial Reorientation on Deep Learning-Based Susceptibility Mapping

Motivation: The rotation of oblique MRI acquisitions to axial orientation may degrade susceptibility maps' quality depending on the interpolation method. Goal(s): This study aimed to determine the optimal interpolation method and stage within the QSM pipeline to apply rotation, improving susceptibility outcomes and supporting the use of deep learning QSM in clinical applications. Approach: On an obliquely acquired post-mortem brain scan, we applied trilinear, spline, and sinc interpolation at different stages in the QSM pipeline and compared outcomes against its axially acquired scan. Results: Spline interpolation before background correction best preserved image clarity, achieving high SSIM scores and maintaining sharpness compared to trilinear. Impact: Using spline-based interpolation before background correction in QSM improves the visibility of clinical features, aiding the accuracy and effectiveness of QSM applications in diagnosing brain disorders.