Kilian M. Pohl

GLISTR: Glioma Image Segmentation and Registration

We present a generative approach for simultaneously registering a probabilistic atlas of a healthy population to brain magnetic resonance (MR) scans showing glioma and segmenting the scans into tumor as well as healthy tissue labels. The proposed method is based on the expectation maximization (EM) algorithm that incorporates a glioma growth model for atlas seeding, a process which modifies the original atlas into one with tumor and edema adapted to best match a given set of patients images. The modified atlas is registered into the patient space and utilized for estimating the posterior probabilities of various tissue labels. EM iteratively refines the estimates of the posterior probabilities of tissue labels, the deformation field and the tumor growth model parameters. Hence, in addition to segmentation, the proposed method results in atlas registration and a low-dimensional description of the patient scans through estimation of tumor model parameters. We validate the method by automatically segmenting 10 MR scans and comparing the results to those produced by clinical experts and two state-of-the-art methods. The resulting segmentations of tumor and edema outperform the results of the reference methods, and achieve a similar accuracy from a second human rater. We additionally apply the method to 122 patients scans and report the estimated tumor model parameters and their relations with segmentation and registration results. Based on the results from this patient population, we construct a statistical atlas of the glioma by inverting the estimated deformation fields to warp the tumor segmentations of patients scans into a common space.

Adolescent Development of Cortical and White Matter Structure in the NCANDA Sample: Role of Sex, Ethnicity, Puberty, and Alcohol Drinking.

Brain structural development continues throughout adolescence, when experimentation with alcohol is often initiated. To parse contributions from biological and environmental factors on neurodevelopment, this study used baseline National Consortium on Alcohol and NeuroDevelopment in Adolescence (NCANDA) magnetic resonance imaging (MRI) data, acquired in 674 adolescents meeting no/low alcohol or drug use criteria and 134 adolescents exceeding criteria. Spatial integrity of images across the 5 recruitment sites was assured by morphological scaling using Alzheimers disease neuroimaging initiative phantom-derived volume scalar metrics. Clinical MRI readings identified structural anomalies in 11.4%. Cortical volume and thickness were smaller and white matter volumes were larger in older than in younger adolescents. Effects of sex (male > female) and ethnicity (majority > minority) were significant for volume and surface but minimal for cortical thickness. Adjusting volume and area for supratentorial volume attenuated or removed sex and ethnicity effects. That cortical thickness showed age-related decline and was unrelated to supratentorial volume is consistent with the radial unit hypothesis, suggesting a universal neural development characteristic robust to sex and ethnicity. Comparison of NCANDA with PING data revealed similar but flatter, age-related declines in cortical volumes and thickness. Smaller, thinner frontal, and temporal cortices in the exceeds-criteria than no/low-drinking group suggested untoward effects of excessive alcohol consumption on brain structural development.

WESD--Weighted Spectral Distance for Measuring Shape Dissimilarity

This paper presents a new distance for measuring shape dissimilarity between objects. Recent publications introduced the use of eigenvalues of the Laplace operator as compact shape descriptors. Here, we revisit the eigenvalues to define a proper distance, called Weighted Spectral Distance (WESD), for quantifying shape dissimilarity. The definition of WESD is derived through analyzing the heat trace. This analysis provides the proposed distance with an intuitive meaning and mathematically links it to the intrinsic geometry of objects. We analyze the resulting distance definition, present and prove its important theoretical properties. Some of these properties include: 1) WESD is defined over the entire sequence of eigenvalues yet it is guaranteed to converge, 2) it is a pseudometric, 3) it is accurately approximated with a finite number of eigenvalues, and 4) it can be mapped to the ([0,1)) interval. Last, experiments conducted on synthetic and real objects are presented. These experiments highlight the practical benefits of WESD for applications in vision and medical image analysis.

Harmonizing DTI measurements across scanners to examine the development of white matter microstructure in 803 adolescents of the NCANDA study.

Neurodevelopment continues through adolescence, with notable maturation of white matter tracts comprising regional fiber systems progressing at different rates. To identify factors that could contribute to regional differences in white matter microstructure development, large samples of youth spanning adolescence to young adulthood are essential to parse these factors. Recruitment of adequate samples generally relies on multi-site consortia but comes with the challenge of merging data acquired on different platforms. In the current study, diffusion tensor imaging (DTI) data were acquired on GE and Siemens systems through the National Consortium on Alcohol and NeuroDevelopment in Adolescence (NCANDA), a multi-site study designed to track the trajectories of regional brain development during a time of high risk for initiating alcohol consumption. This cross-sectional analysis reports baseline Tract-Based Spatial Statistic (TBSS) of regional fractional anisotropy (FA), mean diffusivity (MD), axial diffusivity (L1), and radial diffusivity (LT) from the five consortium sites on 671 adolescents who met no/low alcohol or drug consumption criteria and 132 adolescents with a history of exceeding consumption criteria. Harmonization of DTI metrics across manufacturers entailed the use of human-phantom data, acquired multiple times on each of three non-NCANDA participants at each sites MR system, to determine a manufacturer-specific correction factor. Application of the correction factor derived from human phantom data measured on MR systems from different manufacturers reduced the standard deviation of the DTI metrics for FA by almost a half, enabling harmonization of data that would have otherwise carried systematic error. Permutation testing supported the hypothesis of higher FA and lower diffusivity measures in older adolescents and indicated that, overall, the FA, MD, and L1 of the boys were higher than those of the girls, suggesting continued microstructural development notable in the boys. The contribution of demographic and clinical differences to DTI metrics was assessed with General Additive Models (GAM) testing for age, sex, and ethnicity differences in regional skeleton mean values. The results supported the primary study hypothesis that FA skeleton mean values in the no/low-drinking group were highest at different ages. When differences in intracranial volume were covaried, FA skeleton mean reached a maximum at younger ages in girls than boys and varied in magnitude with ethnicity. Our results, however, did not support the hypothesis that youth who exceeded exposure criteria would have lower FA or higher diffusivity measures than the no/low-drinking group; detecting the effects of excessive alcohol consumption during adolescence on DTI metrics may require longitudinal study.

PORTR: Pre-Operative and Post-Recurrence Brain Tumor Registration

We propose a new method for deformable registration of pre-operative and post-recurrence brain MR scans of glioma patients. Performing this type of intra-subject registration is challenging as tumor, resection, recurrence, and edema cause large deformations, missing correspondences, and inconsistent intensity profiles between the scans. To address this challenging task, our method, called PORTR, explicitly accounts for pathological information. It segments tumor, resection cavity, and recurrence based on models specific to each scan. PORTR then uses the resulting maps to exclude pathological regions from the image-based correspondence term while simultaneously measuring the overlap between the aligned tumor and resection cavity. Embedded into a symmetric registration framework, we determine the optimal solution by taking advantage of both discrete and continuous search methods. We apply our method to scans of 24 glioma patients. Both quantitative and qualitative analysis of the results clearly show that our method is superior to other state-of-the-art approaches.

Segmentation of myocardium using deformable regions and graph cuts

Deformable models and graph cuts are two standard image segmentation techniques. Combining some of their benefits, we introduce a new segmentation system for (semi-) automatic delineation of epicardium and endocardium of Left Ventricle of the heart in Magnetic Resonance Images (MRI). Specifically, a temporal information among consecutive phases is exploited via a coupling between deformable models and graph cuts which provides automated accurate cues for graph cuts and also good initialization scheme for de-formable model that ultimately leads to more accurate and smooth segmentation results with lower interaction costs than using only graph cut segmentation. In addition, we define deformable model as a region defined by two nested contours and segment epicardium and endocardium in an unified way by optimizing single energy functional. This approach provides inherent coherency among the two contours thus leads to more accurate results than deforming separate contours for each target. We show promising results on the challenging problems of left ventricle segmentation.

Semi-supervised Pattern Classification: Application to Structural MRI of Alzheimer's Disease

This paper presents an image-based classification method, and applies it to classification of brain MRI scans of individuals with Mild Cognitive Impairment (MCI). The high dimensionality of the image data is reduced using nonlinear manifold learning techniques, thereby yielding a low-dimensional embedding. Features of the embedding are used in conjunction with a semi-supervised classifier, which utilizes both labeled and unlabeled images to boost performance. The method is applied to 237 scans of MCI patients in order to predict conversion from MCI to Alzheimers Disease. Experimental results demonstrate better prediction accuracy compared to a state-of-the-art method.

Cognitive, emotion control, and motor performance of adolescents in the NCANDA study: Contributions from alcohol consumption, age, sex, ethnicity, and family history of addiction.

OBJECTIVE To investigate development of cognitive and motor functions in healthy adolescents and to explore whether hazardous drinking affects the normal developmental course of those functions. METHOD Participants were 831 adolescents recruited across 5 United States sites of the National Consortium on Alcohol and NeuroDevelopment in Adolescence 692 met criteria for no/low alcohol exposure, and 139 exceeded drinking thresholds. Cross-sectional, baseline data were collected with computerized and traditional neuropsychological tests assessing 8 functional domains expressed as composite scores. General additive modeling evaluated factors potentially modulating performance (age, sex, ethnicity, socioeconomic status, and pubertal developmental stage). RESULTS Older no/low-drinking participants achieved better scores than younger ones on 5 accuracy composites (general ability, abstraction, attention, emotion, and balance). Speeded responses for attention, motor speed, and general ability were sensitive to age and pubertal development. The exceeds-threshold group (accounting for age, sex, and other demographic factors) performed significantly below the no/low-drinking group on balance accuracy and on general ability, attention, episodic memory, emotion, and motor speed scores and showed evidence for faster speed at the expense of accuracy. Delay Discounting performance was consistent with poor impulse control in the younger no/low drinkers and in exceeds-threshold drinkers regardless of age. CONCLUSIONS Higher achievement with older age and pubertal stage in general ability, abstraction, attention, emotion, and balance suggests continued functional development through adolescence, possibly supported by concurrently maturing frontal, limbic, and cerebellar brain systems. Determination of whether low scores by the exceeds-threshold group resulted from drinking or from other preexisting factors requires longitudinal study. (PsycINFO Database Record

Regional Manifold Learning for Deformable Registration of Brain MR Images

We propose a method for deformable registration based on learning the manifolds of individual brain regions. Recent publications on registration of medical images advocate the use of manifold learning in order to confine the search space to anatomically plausible deformations. Existing methods construct manifolds based on a single metric over the entire image domain thus frequently miss regional brain variations. We address this issue by first learning manifolds for specific regions and then computing region-specific deformations from these manifolds. We then determine deformations for the entire image domain by learning the global manifold in such a way that it preserves the region-specific deformations. We evaluate the accuracy of our method by applying it to the LPBA40 dataset and measuring the overlap of the deformed segmentations. The result shows significant improvement in registration accuracy on cortex regions compared to other state of the art methods.

Regional Manifold Learning for Disease Classification

While manifold learning from images itself has become widely used in medical image analysis, the accuracy of existing implementations suffers from viewing each image as a single data point. To address this issue, we parcellate images into regions and then separately learn the manifold for each region. We use the regional manifolds as low-dimensional descriptors of high-dimensional morphological image features, which are then fed into a classifier to identify regions affected by disease. We produce a single ensemble decision for each scan by the weighted combination of these regional classification results. Each weight is determined by the regional accuracy of detecting the disease. When applied to cardiac magnetic resonance imaging of 50 normal controls and 50 patients with reconstructive surgery of Tetralogy of Fallot, our method achieves significantly better classification accuracy than approaches learning a single manifold across the entire image domain.