31
7

Uncertainty-Aware Self-supervised Neural Network for Liver T1ρT_{1ρ} Mapping with Relaxation Constraint

Abstract

T1ρT_{1\rho} mapping is a promising quantitative MRI technique for the non-invasive assessment of tissue properties. Learning-based approaches can map T1ρT_{1\rho} from a reduced number of T1ρT_{1\rho} weighted images, but requires significant amounts of high quality training data. Moreover, existing methods do not provide the confidence level of the T1ρT_{1\rho} estimation. To address these problems, we proposed a self-supervised learning neural network that learns a T1ρT_{1\rho} mapping using the relaxation constraint in the learning process. Epistemic uncertainty and aleatoric uncertainty are modelled for the T1ρT_{1\rho} quantification network to provide a Bayesian confidence estimation of the T1ρT_{1\rho} mapping. The uncertainty estimation can also regularize the model to prevent it from learning imperfect data. We conducted experiments on T1ρT_{1\rho} data collected from 52 patients with non-alcoholic fatty liver disease. The results showed that our method outperformed the existing methods for T1ρT_{1\rho} quantification of the liver using as few as two T1ρT_{1\rho}-weighted images. Our uncertainty estimation provided a feasible way of modelling the confidence of the self-supervised learning based T1ρT_{1\rho} estimation, which is consistent with the reality in liver T1ρT_{1\rho} imaging.

View on arXiv
Comments on this paper