Abstract:Quantitative susceptibility mapping (QSM) quantifies the distribution of magnetic susceptibility and shows great potential in assessing tissue contents such as iron, myelin, and calcium in numerous brain diseases. The accuracy of QSM reconstruction was challenged by an ill-posed field-to-susceptibility inversion problem, which is related to the impaired information near the zero-frequency response of the dipole kernel. Recently, deep learning methods demonstrated great capability in improving the accuracy and efficiency of QSM reconstruction. However, the construction of neural networks in most deep learning-based QSM methods did not take the intrinsic nature of the dipole kernel into account. In this study, we propose a dipole kernel-adaptive multi-channel convolutional neural network (DIAM-CNN) method for the dipole inversion problem in QSM. DIAM-CNN first divided the original tissue field into high-fidelity and low-fidelity components by thresholding the dipole kernel in the frequency domain, and it then inputs the two components as additional channels into a multichannel 3D Unet. QSM maps from the calculation of susceptibility through multiple orientation sampling (COSMOS) were used as training labels and evaluation reference. DIAM-CNN was compared with two conventional model-based methods [morphology enabled dipole inversion (MEDI) and improved sparse linear equation and least squares (iLSQR) and one deep learning method (QSMnet)]. High-frequency error norm (HFEN), peak signal-to-noise-ratio (PSNR), normalized root mean squared error (NRMSE), and the structural similarity index (SSIM) were reported for quantitative comparisons. Experiments on healthy volunteers demonstrated that the DIAM-CNN results had superior image quality to those of the MEDI, iLSQR, or QSMnet results. Experiments on data with simulated hemorrhagic lesions demonstrated that DIAM-CNN produced fewer shadow artifacts around the bleeding lesion than the compared methods. This study demonstrates that the incorporation of dipole-related knowledge into the network construction has a potential to improve deep learning-based QSM reconstruction.

NeXtQSM -- A complete deep learning pipeline for data-consistent quantitative susceptibility mapping trained with hybrid data

MoDL-QSM: Model-based Deep Learning for Quantitative Susceptibility Mapping

MoG-QSM: Model-based Generative Adversarial Deep Learning Network for Quantitative Susceptibility Mapping.

Nonlinear Formulation of the Magnetic Field to Source Relationship for Robust Quantitative Susceptibility Mapping

DF‐QSM: Data Fidelity based Hybrid Approach for Improved Quantitative Susceptibility Mapping of the Brain

A Self-Supervised Deep Learning Approach for Quantitative Susceptibility Mapping Without Ground Truth Labels

Overview of quantitative susceptibility mapping using deep learning -- Current status, challenges and opportunities

Model-Based Learning for Quantitative Susceptibility Mapping.

Deep Quantitative Susceptibility Mapping for Background Field Removal and Total Field Inversion

Clinical Quantitative Susceptibility Mapping (QSM): Biometal Imaging and Its Emerging Roles in Patient Care

msQSM: Morphology-based self-supervised deep learning for quantitative susceptibility mapping

IR2QSM: Quantitative Susceptibility Mapping via Deep Neural Networks with Iterative Reverse Concatenations and Recurrent Modules

xQSM: Quantitative Susceptibility Mapping with Octave Convolutional and Noise Regularized Neural Networks

Spinet-QSM: model-based deep learning with schatten p-norm regularization for improved quantitative susceptibility mapping

Quantitative susceptibility mapping using multi-channel convolutional neural networks with dipole-adaptive multi-frequency inputs

Learn Less, Infer More: Learning in the Fourier Domain for Quantitative Susceptibility Mapping.

A subject-specific unsupervised deep learning method for quantitative susceptibility mapping using implicit neural representation

Plug-and-Play Latent Feature Editing for Orientation-Adaptive Quantitative Susceptibility Mapping Neural Networks

QSMGAN: Improved Quantitative Susceptibility Mapping using 3D Generative Adversarial Networks with Increased Receptive Field

Quantitative Susceptibility Inversion Through Parcellated Multiresolution Neural Networks and K-Space Substitution