A Deep Learning based Harmonic Field Extension in SMWI with Reduced Spatial Coverage

Motivation: SMWI enhances contrast for Parkinson's biomarkers in the substantia nigra but requires long scan times. Reducing the FOV to the substantia nigra can shorten scan time but leads to significant QSM underestimation, weakening SMWI contrast.
Goal(s): To address contrast loss in limited FOV SMWI by correcting underestimation issue in limited FOV QSM.
Approach: Applying a deep learning-based harmonic field extension method corrects underestimation in limited FOV QSM, thereby enhancing contrast in limited FOV SMWI.
Results: The proposed method reduced scan time by approximately 40% compared to the original, while maintaining SMWI contrast comparable to full FOV.
Impact: This study offers a practical approach to enhance SMWI contrast in limited FOV scans, enabling faster, more effective imaging for Parkinson's patients. The method holds potential for clinical application with reduced scan time and robust contrast maintenance.