Model-based Learning for Quantitative Susceptibility Mapping

TL;DR

This paper introduces uQSM, a model-based deep learning method for quantitative susceptibility mapping in MRI that improves accuracy and quality without relying on ground-truth labels, overcoming limitations of previous techniques.

Contribution

uQSM is the first deep learning approach for QSM that is trained solely on physical models, eliminating the need for ground-truth susceptibility maps or synthetic data.

Findings

uQSM outperforms existing methods in quantitative accuracy on multi-orientation datasets.

uQSM produces higher quality QSM images qualitatively on single-orientation datasets.

The method demonstrates robustness without dependence on labeled training data.

Abstract

Quantitative susceptibility mapping (QSM) is a magnetic resonance imaging (MRI) technique that estimates magnetic susceptibility of tissue from Larmor frequency offset measurements. The generation of QSM requires solving a challenging ill-posed field-to-source inversion problem. Inaccurate field-to-source inversion often causes large susceptibility estimation errors that appear as streaking artifacts in the QSM, especially in massive hemorrhagic regions. Recently, several deep learning (DL) QSM techniques have been proposed and demonstrated impressive performance. Due to the inherent non-existent ground-truth QSM references, these DL techniques used either calculation of susceptibility through multiple orientation sampling (COSMOS) maps or synthetic data for network training. Therefore, they were constrained by the availability and accuracy of COSMOS maps, or suffered from performance drop when the training and testing domains were different. To address these limitations, we present a model-based DL method, denoted as uQSM. Without accessing to QSM labels, uQSM is trained using the well-established physical model. When evaluating on multi-orientation QSM datasets, uQSM achieves higher levels of quantitative accuracy compared to TKD, TV-FANSI, MEDI, and DIP approaches. When qualitatively evaluated on single-orientation datasets, uQSM outperforms other methods and reconstructed high quality QSM.