Deep learning estimation of multi-tissue constrained spherical deconvolution with limited single shell DW-MRI

Abstract

Diffusion-weighted magnetic resonance imaging (DW-MRI) is the only non-invasive approach for estimation of intravoxel tissue microarchitecture and reconstruction of in vivo neural pathways for the human brain. With improvement in accelerated MRI acquisition technologies, DW-MRI protocols that make use of multiple levels of diffusion sensitization have gained popularity. A well-known advanced method for reconstruction of white matter microstructure that uses multishell data is multi-tissue constrained spherical deconvolution (MT-CSD). MT-CSD substantially improves the resolution of intra-voxel structure over the traditional single shell version, constrained spherical deconvolution (CSD). Herein, we explore the possibility of using deep learning on single shell data (using the b=1000 s/mm 2 from the Human Connectome Project (HCP)) to estimate the information content captured by 8 th order MT-CSD using the full three shell data (b=1000, 2000, and 3000 s/mm 2 from HCP). Briefly, we examine two network architectures: 1.) Sequential network of fully connected dense layers with a residual block in the middle (ResDNN), 2.) Patch based convolutional neural network with a residual block (ResCNN). For both networks an additional output block for estimation of voxel fraction was used with a modified loss function. Each approach was compared against the baseline of using MT-CSD on all data on 15 subjects from the HCP divided into 5 training, 2 validation, and 8 testing subjects with a total of 6.7 million voxels. The fiber orientation distribution function (fODF) can be recovered with high correlation (0.77 vs 0.74 and 0.65) and low root mean squared error ResCNN:0.0124, ResDNN:0.0168 and sCSD:0.0323 as compared to the ground truth of MT-CST, which was derived from the multi-shell DW-MRI acquisitions. The mean squared error between the MT-CSD estimates for white matter tissue fraction and for the predictions are ResCNN:0.0249 vs ResDNN:0.0264. We illustrate the applicability of high definition fiber tractography on a single testing subject with arcuate and corpus callosum Tractography. In summary, the proposed approach provides a promising framework to estimate MT-CSD with limited single shell data. Source code and models have been made publicly available.

Deep Diffusion MRI Registration (DDMReg): A Deep Learning Method for Diffusion MRI Registration

“…Third, currently the computation of a TOM is relatively time-consuming because of the computation of CSD and FOD peaks. A future investigation could include deep learning to perform CSD computation to further improve computational speed [81]. Fourth, currently, we demonstrated our method on HCP data; future investigation could include data from different sources for training and testing.…”

Deep Learning Improves Pre-Surgical White Matter Visualization in Glioma Patients

“…Using our technique, any institution possessing a DTI-capable MR device can theoretically obtain high-resolution tractography without acquisition of research-grade hardware or personnel. Furthermore, scanning time may be substantially reduced making high-quality tractographic reconstructions clinically feasible: For example, a single shell-shell, 32 direction scan may take approximately 5 minutes to acquire, whereas a 3-shell, 90 direction scan may take approximately 40 minutes 14 . Our approach demonstrates that high-quality tractography can be conducted in short, clinically-feasible timeframes.…”

“…We adapted the deep learning model proposed by Nath et al (2020). The input consists of 3 x 3 x 3 x 45 cubic patches of voxels.…”

“…Deep learning is a powerful tool that permits learning of non-linear mappings between a set of inputs and outputs. Nath et al (2020) presented a deep learning-based pipeline for recovery of MT-CSD from single-shell diffusion data using a residual convolutional neural network (ResCNN) 14 . In this study, we adapted the deep learning model, trained using normal subjects from the Human Connectome Project (HCP) 15 , and utilized it to process single-shell DTI data from 2 real tumor cases.…”

“…At present, the majority of these non-tensor sequences are utilized for research purposes, rather than in clinical settings 6,13 . Relative to DTI, however, acquisition of multi-shell, non-tensor sequences is more expensive and time-consuming 14 , making them impractical to implement or utilize clinically.…”