LINEAR: Learning Implicit Neural Representation With Explicit Physical Priors for Accelerated Quantitative T1rho Mapping

TL;DR

This paper introduces LINEAR, a novel unsupervised neural method that accelerates T1rho mapping by integrating physical priors, achieving high acceleration and superior image quality from undersampled data.

Contribution

The study presents a subject-specific implicit neural approach that incorporates physical priors for accelerated T1rho mapping without requiring fully-sampled training data.

Findings

Achieves up to 14x acceleration in T1rho mapping.

Outperforms state-of-the-art methods in artifact suppression.

Demonstrates robustness on both retrospective and prospective data.

Abstract

Quantitative T1rho mapping has shown promise in clinical and research studies. However, it suffers from long scan times. Deep learning-based techniques have been successfully applied in accelerated quantitative MR parameter mapping. However, most methods require fully-sampled training dataset, which is impractical in the clinic. In this study, a novel subject-specific unsupervised method based on the implicit neural representation is proposed to reconstruct T1rho-weighted images from highly undersampled k-space data, which only takes spatiotemporal coordinates as the input. Specifically, the proposed method learned a implicit neural representation of the MR images driven by two explicit priors from the physical model of T1rho mapping, including the signal relaxation prior and self-consistency of k-t space data prior. The proposed method was verified using both retrospective and prospective undersampled k-space data. Experiment results demonstrate that LINEAR achieves a high acceleration factor up to 14, and outperforms the state-of-the-art methods in terms of suppressing artifacts and achieving the lowest error.