Q-space-coordinate-guided neural networks for high-fidelity diffusion tensor estimation from minimal diffusion-weighted images

Abstract

Diffusion tensor imaging (DTI) is a widely used imaging technique for mapping living human brain tissue's microstructure and structural connectivity. Recently, deep learning methods have been proposed to rapidly estimate diffusion tensors (DTs) using only a small quantity of diffusion-weighted (DW) images. However, these methods typically use the DW images obtained with fixed q-space sampling schemes as the training data, limiting the application scenarios of such methods. To address this issue, we develop a new deep neural network called q-space-coordinate-guided diffusion tensor imaging (QCG-DTI), which can efficiently and correctly estimate DTs under flexible q-space sampling schemes. First, we propose a q-space-coordinate-embedded feature consistency strategy to ensure the correspondence between q-space-coordinates and their respective DW images. Second, a q-space-coordinate fusion (QCF) module is introduced which efficiently embeds q-space-coordinates into multiscale features of the corresponding DW images by linearly adjusting the feature maps along the channel dimension, thus eliminating the dependence on fixed diffusion sampling schemes. Finally, a multiscale feature residual dense (MRD) module is proposed which enhances the network's feature extraction and image reconstruction capabilities by using dual-branch convolutions with different kernel sizes to extract features at different scales. Compared to state-of-the-art methods that rely on a fixed sampling scheme, the proposed network can obtain high-quality diffusion tensors and derived parameters even using DW images acquired with flexible q-space sampling schemes. Compared to state-of-the-art deep learning methods, QCG-DTI reduces the mean absolute error by approximately 15% on fractional anisotropy and around 25% on mean diffusivity.

1. Introduction

Magnetic resonance imaging (MRI) is currently the mainstream and commonly used clinical detection and diagnostic method in medical imaging. Diffusion magnetic resonance imaging (dMRI) is a new type of MRI technology that has developed in recent years (Alexander et al., 2007). For imaging, dMRI uses mainly the different diffusion motion characteristics of water molecules in various organ tissues under the influence of magnetic fields with different diffusion gradient directions. The difference between dMRI and traditional MRI images is that the features of each voxel not only are described by grayscale values but also contain a high-dimensional tensor to describe the diffusion characteristics of water molecules in organs.

In clinical applications, it is necessary to collect diffusion-weighted imaging (DWI) images through dMRI to calculate the diffusion tensor (DT) (Le Bihan et al., 2001; Alexander et al., 2007) and use DT to derive a series of parameters, such as mean diffusivity (MD), fractional anisotropy (FA), axial diffusivity (AD), and radial diffusivity (RD). These tensor-derived parameters have been widely used in studies of brain development (Barnea-Goraly et al., 2005; Lebel et al., 2008) and various diseases (Eriksson et al., 2001; Barnea-Goraly et al., 2004; Roosendaal et al., 2009). DT imaging (DTI) is an essential tool for neuroscience research and has many clinical applications.

In traditional algorithms, only six diffusion-weighted (DW) images and one non-DW image are theoretically needed to estimate the DT. However, DW images obtained in clinical practice usually have low signal-to-noise ratios (SNRs). The traditional DT estimation algorithms cannot accurately estimate the DT for DW images with low SNR. The ordinary least-squares (OLS) method was initially proposed to estimate DTs (Pierpaoli et al., 1996), but it is susceptible to noise and variations in local magnetic susceptibility. To enhance the reliability of the estimation results, the weighted least-squares (WLS) method was proposed (Basser et al., 2000). WLS adjusts the accuracy of the data and the noise level through a weighting matrix. However, these traditional methods still have certain limitations in DT estimation. Consequently, it is...