Abstract:Multi-parametric mapping of MRI relaxations in liver has the potential of revealing pathological information of the liver. A self-supervised learning based multi-parametric mapping method is proposed to map T$T_{1\rho}$ and T$_2$ simultaneously, by utilising the relaxation constraint in the learning process. Data noise of different mapping tasks is utilised to make the model uncertainty-aware, which adaptively weight different mapping tasks during learning. The method was examined on a dataset of 51 patients with non-alcoholic fatter liver disease. Results showed that the proposed method can produce comparable parametric maps to the traditional multi-contrast pixel wise fitting method, with a reduced number of images and less computation time. The uncertainty weighting also improves the model performance. It has the potential of accelerating MRI quantitative imaging.

What problem does this paper attempt to address?

The paper primarily focuses on addressing the issue of multiparametric mapping in liver quantitative magnetic resonance imaging (qMRI), with the specific goal of improving the mapping accuracy of T1ρ and T2 relaxation times while reducing the number of images and computation time. The research team proposed a self-supervised learning-based multitask mapping method that can map T1ρ and T2 simultaneously, utilizing relaxation constraints to optimize the learning process, and by considering the data noise of different mapping tasks, the model is endowed with uncertainty awareness. This allows for dynamic adjustment of the learning weights for different tasks. The key innovations of the paper are: 1. **Self-supervised learning framework**: This approach does not require high-quality annotated data and leverages the intrinsic physical relationship between T1ρ and T2 weighted images (i.e., relaxation constraints) during training to estimate both parameters simultaneously. 2. **Uncertainty-weighted loss function**: By introducing uncertainty weights, the method can automatically adjust the relative importance of each mapping task in the overall learning process based on the data noise of each task. This helps to achieve a more balanced learning outcome in the presence of varying task difficulties and data qualities. 3. **Experimental validation**: The research was validated using a dataset from 51 patients with non-alcoholic fatty liver disease, showing that the proposed method can generate comparable parameter maps with reduced image numbers and computation time compared to traditional pixel-wise fitting methods. The clinical significance of this research lies in providing a new avenue for accelerating liver multiparametric mapping and promoting the application of quantitative magnetic resonance imaging in clinical settings. By reducing scanning time and improving quantification accuracy, this method is expected to enhance efficiency in large-scale qMRI studies, especially in scenarios where rapid and precise assessment of liver pathology is required. Additionally, the study discusses two forms of uncertainty weighting—heteroscedastic and homoscedastic uncertainty—and their performance differences in dealing with signal scaling variations and data corruption. Future research directions will include exploring the correlation between T1ρ and T2 mapping, as well as applying multivariate multitask learning techniques to further optimize multiparametric mapping.

Uncertainty-weighted Multi-tasking for $T_{1ρ}$ and T$_2$ Mapping in the Liver with Self-supervised Learning

Uncertainty-Aware Self-supervised Neural Network for Liver $T_{1ρ}$ Mapping with Relaxation Constraint

An Uncertainty Aided Framework for Learning based Liver $T_1ρ$ Mapping and Analysis

Three-dimensional Simultaneous Brain Mapping of T1, T2, T2*and Magnetic Susceptibility with MR Multitasking

Three-dimensional simultaneous brain T 1 , T 2 , and ADC mapping with MR Multitasking.

Free‐breathing Multitasking Multi‐echo MRI for Whole‐liver Water‐specific T1, Proton Density Fat Fraction, and Quantification

Free-breathing, Non-Ecg, Simultaneous Myocardial T1 , T2 , T2 *, and Fat-Fraction Mapping with Motion-Resolved Cardiovascular MR Multitasking.

Multimodal Priors Guided Segmentation of Liver Lesions in MRI Using Mutual Information Based Graph Co-Attention Networks.

Three-dimensional simultaneous brain T<sub>1</sub>, T<sub>2</sub>, and ADC mapping with MR Multitasking

Liver Tumor Detection Via A Multi-Scale Intermediate Multi-Modal Fusion Network on MRI Images.

Liver-Buda-Sage: Simultaneous Whole Liver T2 and T*2 Mapping in One Breath-Hold.

Deep Learning-Assisted Diagnosis of Liver Tumors Using Non-Contrast Magnetic Resonance Imaging: A Multi-Center Study

Multi-phase Liver Tumor Segmentation with Spatial Aggregation and Uncertain Region Inpainting

United adversarial learning for liver tumor segmentation and detection of multi-modality non-contrast MRI

Free-breathing simultaneous native myocardial T1, T2 and T1ρ mapping with Cartesian acquisition and dictionary matching

Alternating low‐rank tensor reconstruction for improved multiparametric mapping with cardiovascular MR Multitasking

Uncertainty in multitask learning: joint representations for probabilistic MR-only radiotherapy planning

Free‐breathing, fat‐corrected T1 mapping of the liver with stack‐of‐stars MRI, and joint estimation of T1, PDFF, R2*, and B1+

Simultaneous Multiparameter Mapping of the Liver in a Single Breath‐Hold or Respiratory‐Triggered Acquisition Using Multi‐Inversion Spin and Gradient Echo MRI

Task relevance driven adversarial learning for simultaneous detection, size grading, and quantification of hepatocellular carcinoma via integrating multi-modality MRI

Edge-aware Multi-task Network for Integrating Quantification Segmentation and Uncertainty Prediction of Liver Tumor on Multi-modality Non-contrast MRI