Uncertainty-Aware Self-supervised Neural Network for Liver T_1ρ Mapping with Relaxation Constraint

Abstract

T_1 mapping is a promising quantitative MRI technique for the non-invasive assessment of tissue properties. Learning-based approaches can map T_1 from a reduced number of T_1 weighted images, but requires significant amounts of high quality training data. Moreover, existing methods do not provide the confidence level of the T_1 estimation. To address these problems, we proposed a self-supervised learning neural network that learns a T_1 mapping using the relaxation constraint in the learning process. Epistemic uncertainty and aleatoric uncertainty are modelled for the T_1 quantification network to provide a Bayesian confidence estimation of the T_1 mapping. The uncertainty estimation can also regularize the model to prevent it from learning imperfect data. We conducted experiments on T_1 data collected from 52 patients with non-alcoholic fatty liver disease. The results showed that our method outperformed the existing methods for T_1 quantification of the liver using as few as two T_1-weighted images. Our uncertainty estimation provided a feasible way of modelling the confidence of the self-supervised learning based T_1 estimation, which is consistent with the reality in liver T_1 imaging.

Tasks

Reproductions