Recurrent Inference Machines as inverse problem solvers for MR relaxometry

TL;DR

This paper introduces Recurrent Inference Machines (RIMs) as a novel neural network approach for high-precision, fast, and robust T1 and T2 relaxometry mapping in MRI, outperforming traditional methods.

Contribution

It demonstrates that RIMs can effectively solve non-linear inverse problems in MRI, providing a flexible, accurate, and significantly faster alternative to existing techniques like MLE.

Findings

RIM achieves higher accuracy and precision than MLE and ResNet.

Inference with RIM is 150 times faster than MLE.

RIM shows robustness to variations in scanning parameters.

Abstract

In this paper, we propose the use of Recurrent Inference Machines (RIMs) to perform T1 and T2 mapping. The RIM is a neural network framework that learns an iterative inference process based on the signal model, similar to conventional statistical methods for quantitative MRI (QMRI), such as the Maximum Likelihood Estimator (MLE). This framework combines the advantages of both data-driven and model-based methods, and, we hypothesize, is a promising tool for QMRI. Previously, RIMs were used to solve linear inverse reconstruction problems. Here, we show that they can also be used to optimize non-linear problems and estimate relaxometry maps with high precision and accuracy. The developed RIM framework is evaluated in terms of accuracy and precision and compared to an MLE method and an implementation of the ResNet. The results show that the RIM improves the quality of estimates compared to the other techniques in Monte Carlo experiments with simulated data, test-retest analysis of a system phantom, and in-vivo scans. Additionally, inference with the RIM is 150 times faster than the MLE, and robustness to (slight) variations of scanning parameters is demonstrated. Hence, the RIM is a promising and flexible method for QMRI. Coupled with an open-source training data generation tool, it presents a compelling alternative to previous methods.