Roger C. Tam

Deep 3D Convolutional Encoder Networks With Shortcuts for Multiscale Feature Integration Applied to Multiple Sclerosis Lesion Segmentation

We propose a novel segmentation approach based on deep 3D convolutional encoder networks with shortcut connections and apply it to the segmentation of multiple sclerosis (MS) lesions in magnetic resonance images. Our model is a neural network that consists of two interconnected pathways, a convolutional pathway, which learns increasingly more abstract and higher-level image features, and a deconvolutional pathway, which predicts the final segmentation at the voxel level. The joint training of the feature extraction and prediction pathways allows for the automatic learning of features at different scales that are optimized for accuracy for any given combination of image types and segmentation task. In addition, shortcut connections between the two pathways allow high- and low-level features to be integrated, which enables the segmentation of lesions across a wide range of sizes. We have evaluated our method on two publicly available data sets (MICCAI 2008 and ISBI 2015 challenges) with the results showing that our method performs comparably to the top-ranked state-of-the-art methods, even when only relatively small data sets are available for training. In addition, we have compared our method with five freely available and widely used MS lesion segmentation methods (EMS, LST-LPA, LST-LGA, Lesion-TOADS, and SLS) on a large data set from an MS clinical trial. The results show that our method consistently outperforms these other methods across a wide range of lesion sizes.

The association between cognitive function and white matter lesion location in older adults: a systematic review

BackgroundMaintaining cognitive function is essential for healthy aging and to function autonomously within society. White matter lesions (WMLs) are associated with reduced cognitive function in older adults. However, whether their anatomical location moderates these associations is not well-established. This review systematically evaluates peer-reviewed evidence on the role of anatomical location in the association between WMLs and cognitive function.MethodsIn accordance with the preferred reporting items for systematic reviews and meta-analysis (PRISMA) statement, databases of EMBASE, PUBMED, MEDLINE, and CINAHL, and reference lists of selected papers were searched. We limited our search results to adults aged 60 years and older, and studies published in the English language from 2000 to 2011. Studies that investigated the association between cognitive function and WML location were included. Two independent reviewers extracted: 1) study characteristics including sample size, sample characteristic, and study design; 2) WML outcomes including WML location, WML quantification method (scoring or volume measurement), strength of the MRI magnet in Tesla, and MRI sequence used for WML detection; and 3) cognitive function outcomes including cognitive tests for two cognitive domains of memory and executive function/processing speed.ResultsOf the 14 studies included, seven compared the association of subcortical versus periventricular WMLs with cognitive function. Seven other studies investigated the association between WMLs in specific brain regions (e.g., frontal, parietal lobes) and cognitive function. Overall, the results show that a greater number of studies have found an association between periventricular WMLs and executive function/processing speed, than subcortical WMLs. However, whether WMLs in different brain regions have a differential effect on cognitive function remains unclear.ConclusionsEvidence suggests that periventricular WMLs may have a significant negative impact on cognitive abilities of older adults. This finding may be influenced by study heterogeneity in: 1) MRI sequences, WML quantification methods, and neuropsychological batteries; 2) modifying effect of cardiovascular risk factors; and 3) quality of studies and lack of sample size calculation.

Shape simplification based on the medial axis transform

We present a new algorithm for simplifying the shape of 3D objects by manipulating their medial axis transform (MAT). From an unorganized set of boundary points, our algorithm computes the MAT, decomposes the axis into parts, then selectively removes a subset of these parts in order to reduce the complexity of the overall shape. The result is simplified MAT that can be used for a variety of shape operations. In addition, a polygonal surface of the resulting shape can be directly generated from the filtered MAT using a robust surface reconstruction method. The algorithm presented is shown to have a number of advantages over other existing approaches.

Manifold Learning of Brain MRIs by Deep Learning

Manifold learning of medical images plays a potentially important role for modeling anatomical variability within a population with pplications that include segmentation, registration, and prediction of clinical parameters. This paper describes a novel method for learning the manifold of 3D brain images that, unlike most existing manifold learning methods, does not require the manifold space to be locally linear, and does not require a predefined similarity measure or a prebuilt proximity graph. Our manifold learning method is based on deep learning, a machine learning approach that uses layered networks (called deep belief networks, or DBNs) and has received much attention recently in the computer vision field due to their success in object recognition tasks. DBNs have traditionally been too computationally expensive for application to 3D images due to the large number of trainable parameters. Our primary contributions are (1) a much more computationally efficient training method for DBNs that makes training on 3D medical images with a resolution of up to 128 x 128 x 128 practical, and (2) the demonstration that DBNs can learn a low-dimensional manifold of brain volumes that detects modes of variations that correlate to demographic and disease parameters.

Reproducibility of myelin water fraction analysis: a comparison of region of interest and voxel-based analysis methods

This study compared region of interest (ROI) and voxel-based analysis (VBA) methods to determine the optimal method of myelin water fraction (MWF) analysis. Twenty healthy controls were scanned twice using a multi-echo T(2) relaxation sequence and ROIs were drawn in white and grey matter. MWF was defined as the fractional signal from 15 to 40 ms in the T(2) distribution. For ROI analysis, the mean intensity of voxels within an ROI was fit using non-negative least squares. For VBA, MWF was obtained for each voxel and the mean and median values within an ROI were calculated. There was a slightly higher correlation between Scan 1 and 2 for the VBA method (R(2)=0.98) relative to the ROI method (R(2)=0.95), and the VBA mean square difference between scans was 300% lower, indicating VBA was the most consistent between scans. For the VBA method, mean MWF was found to be more reproducible than median MWF. As the VBA method is more reproducible and gives more options for visualization and analysis of MWF, it is recommended over the ROI method of MWF analysis.

A Hybrid Geometric–Statistical Deformable Model for Automated 3-D Segmentation in Brain MRI

We present a novel 3-D deformable model-based approach for accurate, robust, and automated tissue segmentation of brain MRI data of single as well as multiple magnetic resonance sequences. The main contribution of this study is that we employ an edge-based geodesic active contour for the segmentation task by integrating both image edge geometry and voxel statistical homogeneity into a novel hybrid geometric-statistical feature to regularize contour convergence and extract complex anatomical structures. We validate the accuracy of the segmentation results on simulated brain MRI scans of both single T1-weighted and multiple T1/T2/PD-weighted sequences. We also demonstrate the robustness of the proposed method when applied to clinical brain MRI scans. When compared to a current state-of-the-art region-based level-set segmentation formulation, our white matter and gray matter segmentation resulted in significantly higher accuracy levels with a mean improvement in Dice similarity indexes of 8.55% (p<0.0001) and 10.18% (p<0.0001), respectively.

Deep Convolutional Encoder Networks for Multiple Sclerosis Lesion Segmentation

We propose a novel segmentation approach based on deep convolutional encoder networks and apply it to the segmentation of multiple sclerosis (MS) lesions in magnetic resonance images. Our model is a neural network that has both convolutional and deconvolutional layers, and combines feature extraction and segmentation prediction in a single model. The joint training of the feature extraction and prediction layers allows the model to automatically learn features that are optimized for accuracy for any given combination of image types. In contrast to existing automatic feature learning approaches, which are typically patch-based, our model learns features from entire images, which eliminates patch selection and redundant calculations at the overlap of neighboring patches and thereby speeds up the training. Our network also uses a novel objective function that works well for segmenting underrepresented classes, such as MS lesions. We have evaluated our method on the publicly available labeled cases from the MS lesion segmentation challenge 2008 data set, showing that our method performs comparably to the state-of-theart. In addition, we have evaluated our method on the images of 500 subjects from an MS clinical trial and varied the number of training samples from 5 to 250 to show that the segmentation performance can be greatly improved by having a representative data set.

MRI Brain Extraction with Combined Expectation Maximization and Geodesic Active Contours

This paper presents a new fully automated method for the extraction of brain cortex from Tl-weighted magnetic resonance imaging (MRI) head scans. Combined with the expectation maximization (EM) algorithm, and a hybrid of pre- and post-processing techniques, incorporating mathematical morphology and connected component analysis, geodesic active contours are evolved in 3D space to segment the cortex. The robustness and accuracy of our proposed method are validated with both synthetic and real MRI data. Our method outperforms standard techniques including the brain extraction tool (BET) and statistical parametric mapping (SPM) by lowering the misclassification rate, especially when analyzing real MRI data

The impact of intensity variations in T1-hypointense lesions on clinical correlations in multiple sclerosis.

Background: The correlations between T1-hypointense lesion (‘black hole’) volume and clinical measures have varied widely across previous studies. The degree of hypointensity in black holes is associated with the severity of tissue damage, but the impact on the correlation with disability is unknown. Objectives: To determine how variations in the intensity level used for lesion classification can impact clinical correlation, specifically with the Expanded Disability Status Scale (EDSS), and whether using a restricted range can improve correlation. Methods: A highly automated image analysis procedure was applied to the scans of 24 multiple sclerosis (MS) patients with well-distributed EDSS scores to compute their black hole volumes at nine different levels of intensity relative to the reference intensities sampled in normal-appearing white matter (NAWM) and cerebrospinal fluid (CSF). Two methods of volume computation were used. Results: The black hole volume–EDSS Spearman correlations ranged between 0.49–0.73 (first method) and 0.54–0.74 (second method). The strongest correlations were observed by only including the voxels with maximum intensities at 30–40% of the CSF to NAWM range. Conclusions: Intensity variations can have a large impact on black hole–EDSS correlation. Restricting the measurement to a subset of the darkest voxels may yield stronger correlations.

Modeling the Variability in Brain Morphology and Lesion Distribution in Multiple Sclerosis by Deep Learning

Changes in brain morphology and white matter lesions are two hallmarks of multiple sclerosis (MS) pathology, but their variability beyond volumetrics is poorly characterized. To further our understanding of complex MS pathology, we aim to build a statistical model of brain images that can automatically discover spatial patterns of variability in brain morphology and lesion distribution. We propose building such a model using a deep belief network (DBN), a layered network whose parameters can be learned from training images. In contrast to other manifold learning algorithms, the DBN approach does not require a prebuilt proximity graph, which is particularly advantageous for modeling lesions, because their sparse and random nature makes defining a suitable distance measure between lesion images challenging. Our model consists of a morphology DBN, a lesion DBN, and a joint DBN that models concurring morphological and lesion patterns. Our results show that this model can automatically discover the classic patterns of MS pathology, as well as more subtle ones, and that the parameters computed have strong relationships to MS clinical scores.