Brain Segmentation From Computed Tomography of Healthy Aging and Geriatric Concussion at Variable Spatial Resolutions

Abstract

When properly implemented and processed, anatomic T1-weighted magnetic resonance imaging (MRI) can be ideal for the noninvasive quantification of white matter (WM) and gray matter (GM) in the living human brain. Although MRI is more suitable for distinguishing GM from WM than computed tomography (CT), the growing clinical use of the latter technique has renewed interest in head CT segmentation. Such interest is particularly strong in settings where MRI is unavailable, logistically unfeasible or prohibitively expensive. Nevertheless, whereas MRI segmentation is a sophisticated and technically-mature research field, the task of automatically classifying soft brain tissues from CT remains largely unexplored. Furthermore, brain segmentation methods for MRI hold considerable potential for adaptation and application to CT image processing. Here we demonstrate this by combining probabilistic, atlas-based classification with topologically-constrained tissue boundary refinement to delineate WM, GM and cerebrospinal fluid (CSF) from head CT images. The feasibility and utility of this approach are revealed by comparison of MRI-only vs. CT-only segmentations in geriatric concussion victims with both MRI and CT scans. Comparison of the two segmentations yields mean Sørensen-Dice coefficients of 85.5 ± 4.6% (WM), 86.7 ± 5.6% (GM) and 91.3 ± 2.8% (CSF), as well as average Hausdorff distances of 3.76 ± 1.85 mm (WM), 3.43 ± 1.53 mm (GM) and 2.46 ± 1.27 mm (CSF). Bootstrapping results suggest that the segmentation approach is sensitive enough to yield WM, GM and CSF volume estimates within ~5%, ~4%, and ~3% of their MRI-based estimates, respectively. To our knowledge, this is the first 3D segmentation approach for CT to undergo rigorous within-subject comparison with high-resolution MRI. Results suggest that (1) standard-quality CT allows WM/GM/CSF segmentation with reasonable accuracy, and that (2) the task of soft brain tissue classification from CT merits further attention from neuroimaging researchers.

A Method to Estimate Brain Volume from Head CT Images and Application to Detect Brain Atrophy in Alzheimer Disease

“…4 However, their method cannot be applied to singletime-point CT images with no image enhancement. Irimia et al 16 adapted SPM12 (http://www.fil.ion.ucl.ac.uk/spm/), 17 a widely used MR imaging-based segmentation method, for CT segmentation and validated it by comparing it with MR images segmented using FreeSurfer (http://surfer.nmr.mgh.harvard.edu). However, they validated only the accuracy of ventricular CSF and not TIV or TBV.…”

Creation of an anthropomorphic CT head phantom for verification of image segmentation

“…These are lower than values obtained in a recent study comparing an enhanced segmentation approach based on SPM12 but incorporating neuroanatomy-constrained correction of tissue boundaries based on the local topological properties of the GM/WM interface. 31 Their approach obtains higher DSC for WM and CSF (0.85 and 0.91, respectively), indicating the success of their method, with a similar DSC measured for GM of 0.87. It is worth noting, however, that in order to demonstrate the utility of their method the study authors had to collect data from consented patients and had to exclude poor quality scans.…”

“…The objective function assists this process by weighting the probability maps of MNI space according to Bayesian inference principles and then deforming them so that they match the volumes being segmented. 10,31 For this work two sets of a priori tissue probability maps were employed; first, the default maps provided with the SPM12 software were used. Second, compatible maps from a study that created a database of adult, age-specific MRI brain, and head templates were used.…”

“…30 Another study uses a segmentation protocol based on SPM12 combined with topologically constrained tissue boundary refinement to delineate WM, GM, and cerebrospinal fluid (CSF) and directly assesses accuracy by comparing CT segmentation for MRI segmentation in the same subjects. 31 Before clinical use can be contemplated a method is required to test the accuracy and reproducibility of the entire imaging chain, including the scanner, reconstruction, and the software for segmentation and quantitative analysis of the brain.…”

Multimodal imaging of cerebral microhemorrhages and white matter degradation in geriatric patients with mild traumatic brain injury

“…Intensity normalization is implemented to account for the potentially confounding effect of intersubject intensity brain differences. 4. The additional rotation of the B matrix in DWI image processing is necessary because resulting orientational information may be affected by the increased head motion which is often associated with DWI, given this modality's relatively long acquisition times [21].…”