Learning white matter subject-specific segmentation from structural MRI

An iterative ROC procedure identifies white matter tracts diagnostic for traumatic brain injury: an exploratory analysis in U.S. Veterans

OBJECTIVE

Understanding the pathophysiology of traumatic brain injury (TBI) is crucial for effectively managing care. Diffusion tensor imaging (DTI) is an MRI technology that evaluates TBI pathology in brain white matter. However, DTI analysis generates numerous measures. Choosing between them remains an obstacle to clinical translation. In this study, we leveraged an iterative receiver operating characteristic (ROC) analysis to examine white matter tracts in a group of 380 Veterans, consisting of TBI (n = 243) and non-TBI patients (n = 137).

METHOD

For each participant, we obtained a whole brain tractography and extracted DTI measures from 50 tracts. The ROC analyzed these variables and produced decision trees of tracts diagnostic for TBI. We expanded our findings by applying jackknife resampling. This procedure removed potential outliers and yielded tracts not observed in the initial ROCs. Finally, we used logistic regression to confirm the tracts predicted TBI status.

RESULTS

Our results indicate ROC can identify tracts diagnostic for TBI. We also found that groups of tracts are more predictive than any single one.

CONCLUSIONS

These analyses show that ROC is a useful tool for exploring large, multivariate datasets and may inform the design of clinical algorithms.

Tractography with T1-weighted MRI and associated anatomical constraints on clinical quality diffusion MRI

Diffusion MRI (dMRI) streamline tractography, the gold-standard for in vivo estimation of white matter (WM) pathways in the brain, has long been considered as a product of WM microstructure. However, recent advances in tractography demonstrated that convolutional recurrent neural networks (CoRNN) trained with a teacher-student framework have the ability to learn to propagate streamlines directly from T1 and anatomical context. Training for this network has previously relied on high resolution dMRI. In this paper, we generalize the training mechanism to traditional clinical resolution data, which allows generalizability across sensitive and susceptible study populations. We train CoRNN on a small subset of the Baltimore Longitudinal Study of Aging (BLSA), which better resembles clinical scans. We define a metric, termed the epsilon ball seeding method, to compare T1 tractography and traditional diffusion tractography at the streamline level. We show that under this metric T1 tractography generated by CoRNN reproduces diffusion tractography with approximately three millimeters of error.

Convolutional-recurrent neural networks approximate diffusion tractography from T1-weighted MRI and associated anatomical context

Diffusion MRI (dMRI) streamline tractography is the gold-standard for in vivo estimation of white matter (WM) pathways in the brain. However, the high angular resolution dMRI acquisitions capable of fitting the microstructural models needed for tractography are often time-consuming and not routinely collected clinically, restricting the scope of tractography analyses. To address this limitation, we build on recent advances in deep learning which have demonstrated that streamline propagation can be learned from dMRI directly without traditional model fitting. Specifically, we propose learning the streamline propagator from T1w MRI to facilitate arbitrary tractography analyses when dMRI is unavailable. To do so, we present a novel convolutional-recurrent neural network (CoRNN) trained in a teacher-student framework that leverages T1w MRI, associated anatomical context, and streamline memory from data acquired for the Human Connectome Project. We characterize our approach under two common tractography paradigms, WM bundle analysis and structural connectomics, and find approximately a 5-15% difference between measures computed from streamlines generated with our approach and those generated using traditional dMRI tractography. When placed in the literature, these results suggest that the accuracy of WM measures computed from T1w MRI with our method is on the level of scan-rescan dMRI variability and raise an important question: is tractography truly a microstructural phenomenon, or has dMRI merely facilitated its discovery and implementation?

Prediction of the Topography of the Corticospinal Tract on T1-Weighted MR Images Using Deep-Learning-Based Segmentation

INTRODUCTION

Tractography is an invaluable tool in the planning of tumor surgery in the vicinity of functionally eloquent areas of the brain as well as in the research of normal development or of various diseases. The aim of our study was to compare the performance of a deep-learning-based image segmentation for the prediction of the topography of white matter tracts on T1-weighted MR images to the performance of a manual segmentation.

METHODS

T1-weighted MR images of 190 healthy subjects from 6 different datasets were utilized in this study. Using deterministic diffusion tensor imaging, we first reconstructed the corticospinal tract on both sides. After training a segmentation model on 90 subjects of the PIOP2 dataset using the nnU-Net in a cloud-based environment with graphical processing unit (Google Colab), we evaluated its performance using 100 subjects from 6 different datasets.

RESULTS

Our algorithm created a segmentation model that predicted the topography of the corticospinal pathway on T1-weighted images in healthy subjects. The average dice score was 0.5479 (0.3513-0.7184) on the validation dataset.

CONCLUSIONS

Deep-learning-based segmentation could be applicable in the future to predict the location of white matter pathways in T1-weighted scans.

Automatic segmentation of the core of the acoustic radiation in humans

Introduction

Acoustic radiation is one of the most important white matter fiber bundles of the human auditory system. However, segmenting the acoustic radiation is challenging due to its small size and proximity to several larger fiber bundles. TractSeg is a method that uses a neural network to segment some of the major fiber bundles in the brain. This study aims to train TractSeg to segment the core of acoustic radiation.

Methods

We propose a methodology to automatically extract the acoustic radiation from human connectome data, which is both of high quality and high resolution. The segmentation masks generated by TractSeg of nearby fiber bundles are used to steer the generation of valid streamlines through tractography. Only streamlines connecting the Heschl's gyrus and the medial geniculate nucleus were considered. These streamlines are then used to create masks of the core of the acoustic radiation that is used to train the neural network of TractSeg. The trained network is used to automatically segment the acoustic radiation from unseen images.

Results

The trained neural network successfully extracted anatomically plausible masks of the core of the acoustic radiation in human connectome data. We also applied the method to a dataset of 17 patients with unilateral congenital ear canal atresia and 17 age- and gender-paired controls acquired in a clinical setting. The method was able to extract 53/68 acoustic radiation in the dataset acquired with clinical settings. In 14/68 cases, the method generated fragments of the acoustic radiation and completely failed in a single case. The performance of the method on patients and controls was similar.

Discussion

In most cases, it is possible to segment the core of the acoustic radiations even in images acquired with clinical settings in a few seconds using a pre-trained neural network.