Ultra-low-field paediatric MRI in low- and middle-income countries: super-resolution using a multi-orientation U-Net

Abstract

AbstractSuper-resolution has been applied to ultra-low-field MRI, however it is yet to be applied specifically to paediatric ultra-low-field imaging. Owing to the high cost of modern MRI systems, their use in clinical care and neurodevelopmental research is limited to hospitals and universities in high income countries. Ultra-low-field systems with significantly lower scanning costs bear the potential for global adoption, however their reduced SNR compared to 1.5 or 3T systems limits their applicability for research and clinical use. In this paper, we describe a deep learning-based super-resolution approach to generate high-resolution isotropic T2-weighted scans from low-resolution inputs. We train a ‘multi-orientation U-Net’, which uses multiple low-resolution anisotropic images acquired in orthogonal orientations to construct a super-resolved output. Our approach exhibits improved quality of outputs compared to current state-of-the-art methods for super-resolution of ultra-low-field scans in paediatric populations. Crucially for paediatric development, our approach improves reconstruction on deep brain structures for all measured regions with the greatest improvement in the caudate, where Spearman’s correlation coefficient, Md [Q1, Q3], between model outputs and high-field targets increases from 0.75 [0.64, 0.81] (current state-of-the-art) to 0.90, [0.86, 0.93] (U-Net). Our research serves as proof-of-principle of the viability of training deep-learning based super-resolution models for use in neurodevelopmental research and presents the first U-Net trained exclusively on paired ultra-low-field and high-field data from infants.