Kaikai Shen

Detecting global and local hippocampal shape changes in Alzheimer's disease using statistical shape models.

The hippocampus is affected at an early stage in the development of Alzheimers disease (AD). With the use of structural magnetic resonance (MR) imaging, we can investigate the effect of AD on the morphology of the hippocampus. The hippocampal shape variations among a population can be usually described using statistical shape models (SSMs). Conventional SSMs model the modes of variations among the population via principal component analysis (PCA). Although these modes are representative of variations within the training data, they are not necessarily discriminative on labeled data or relevant to the differences between the subpopulations. We use the shape descriptors from SSM as features to classify AD from normal control (NC) cases. In this study, a Hotellings T2 test is performed to select a subset of landmarks which are used in PCA. The resulting variation modes are used as predictors of AD from NC. The discrimination ability of these predictors is evaluated in terms of their classification performances with bagged support vector machines (SVMs). Restricting the model to landmarks with better separation between AD and NC increases the discrimination power of SSM. The predictors extracted on the subregions also showed stronger correlation with the memory-related measurements such as Logical Memory, Auditory Verbal Learning Test (AVLT) and the memory subscores of Alzheimer Disease Assessment Scale (ADAS).

Patient Specific Prostate Segmentation in 3-D Magnetic Resonance Images

Accurate localization of the prostate and its surrounding tissue is essential in the treatment of prostate cancer. This paper presents a novel approach to fully automatically segment the prostate, including its seminal vesicles, within a few minutes of a magnetic resonance (MR) scan acquired without an endorectal coil. Such MR images are important in external beam radiation therapy, where using an endorectal coil is highly undesirable. The segmentation is obtained using a deformable model that is trained on-the-fly so that it is specific to the patients scan. This case specific deformable model consists of a patient specific initialized triangulated surface and image feature model that are trained during its initialization. The image feature model is used to deform the initialized surface by template matching image features (via normalized cross-correlation) to the features of the scan. The resulting deformations are regularized over the surface via well established simple surface smoothing algorithms, which is then made anatomically valid via an optimized shape model. Mean and median Dices similarity coefficients (DSCs) of 0.85 and 0.87 were achieved when segmenting 3T MR clinical scans of 50 patients. The median DSC result was equal to the inter-rater DSC and had a mean absolute surface error of 1.85 mm. The approach is showed to perform well near the apex and seminal vesicles of the prostate.

Investigating brain connectivity heritability in a twin study using diffusion imaging data

Heritability of brain anatomical connectivity has been studied with diffusion-weighted imaging (DWI) mainly by modeling each voxels diffusion pattern as a tensor (e.g., to compute fractional anisotropy), but this method cannot accurately represent the many crossing connections present in the brain. We hypothesized that different brain networks (i.e., their component fibers) might have different heritability and we investigated brain connectivity using High Angular Resolution Diffusion Imaging (HARDI) in a cohort of twins comprising 328 subjects that included 70 pairs of monozygotic and 91 pairs of dizygotic twins. Water diffusion was modeled in each voxel with a Fiber Orientation Distribution (FOD) function to study heritability for multiple fiber orientations in each voxel. Precision was estimated in a test-retest experiment on a sub-cohort of 39 subjects. This was taken into account when computing heritability of FOD peaks using an ACE model on the monozygotic and dizygotic twins. Our results confirmed the overall heritability of the major white matter tracts but also identified differences in heritability between connectivity networks. Inter-hemispheric connections tended to be more heritable than intra-hemispheric and cortico-spinal connections. The highly heritable tracts were found to connect particular cortical regions, such as medial frontal cortices, postcentral, paracentral gyri, and the right hippocampus.

Automated 3D Segmentation of Vertebral Bodies and Intervertebral Discs from MRI

Recent developments in high resolution MRI scanning of the human spine are providing increasing opportunities for the development of accurate automated approaches for pathoanatomical assessment of intervertebral discs and vertebrae. We are developing a fully automated 3D segmentation approach for MRI scans of the human spine based on statistical shape analysis and template matching of grey level intensity profiles. The algorithm reported in the present study was validated on a dataset of high resolution volumetric scans of lower thoracic and lumbar spine obtained on a 3T scanner using the relatively new 3D SPACE (T2-weighted) pulse sequence, and on a dataset of axial T1-weighted scans of lumbar spine obtained on a 1.5T system. A 3D spine curve is initially extracted and used to position the statistical shape models for final segmentation. Initial validating experiments show promising results on both MRI datasets.

Increasing power to predict mild cognitive impairment conversion to alzheimer's disease using hippocampal atrophy rate and statistical shape models

Identifying mild cognitive impairment (MCI) subjects who will convert to clinical Alzheimers disease (AD) is important for therapeutic decisions, patient counselling and clinical trials. Hippocampal volume and rate of atrophy predict clinical decline at the MCI stage and progression to AD. In this paper, we create p-maps from the differences in the shape of the hippocampus between 60 normal controls and 60 AD subjects using statistical shape models, and generate different regions of interest (ROI) by thresholding the p-maps at different significance levels. We demonstrate increased statistical power to classify 86 MCI converters and 128 MCI stable subjects using the hippocampal atrophy rates calculated by the boundary shift integral within these ROIs.

Statistical machine learning to identify traumatic brain injury (TBI) from structural disconnections of white matter networks

Identifying diffuse axonal injury (DAI) in patients with traumatic brain injury (TBI) presenting with normal appearing radiological MRI presents a significant challenge. Neuroimaging methods such as diffusion MRI and probabilistic tractography, which probe the connectivity of neural networks, show significant promise. We present a machine learning approach to classify TBI participants primarily with mild traumatic brain injury (mTBI) based on altered structural connectivity patterns derived through the network based statistical analysis of structural connectomes generated from TBI and age-matched control groups. In this approach, higher order diffusion models were used to map white matter connections between 116 cortical and subcortical regions. Tracts between these regions were generated using probabilistic tracking and mean fractional anisotropy (FA) measures along these connections were encoded in the connectivity matrices. Network-based statistical analysis of the connectivity matrices was performed to identify the network differences between a representative subset of the two groups. The affected network connections provided the feature vectors for principal component analysis and subsequent classification by random forest. The validity of the approach was tested using data acquired from a total of 179 TBI patients and 146 controls participants. The analysis revealed altered connectivity within a number of intra- and inter-hemispheric white matter pathways associated with DAI, in consensus with existing literature. A mean classification accuracy of 68.16%±1.81% and mean sensitivity of 80.0%±2.36% were achieved in correctly classifying the TBI patients evaluated on the subset of the participants that was not used for the statistical analysis, in a 10-fold cross-validation framework. These results highlight the potential for statistical machine learning approaches applied to structural connectomes to identify patients with diffusive axonal injury.

Alterations in erythrocyte fatty acid composition in preclinical Alzheimer's disease

Brain and blood fatty acids (FA) are altered in Alzheimer’s disease and cognitively impaired individuals, however, FA alterations in the preclinical phase, prior to cognitive impairment have not been investigated previously. The current study therefore evaluated erythrocyte FA in cognitively normal elderly participants aged 65–90 years via trans-methylation followed by gas chromatography. The neocortical beta-amyloid load (NAL) measured via positron emission tomography (PET) using ligand 18F-Florbetaben, was employed to categorise participants as low NAL (standard uptake value ratio; SUVR < 1.35, N = 65) and high NAL or preclinical AD (SUVR ≥ 1.35, N = 35) wherein, linear models were employed to compare FA compositions between the two groups. Increased arachidonic acid (AA, p < 0.05) and decreased docosapentaenoic acid (DPA, p < 0.05) were observed in high NAL. To differentiate low from high NAL, the area under the curve (AUC) generated from a ‘base model’ comprising age, gender, APOEε4 and education (AUC = 0.794) was outperformed by base model + AA:DPA (AUC = 0.836). Our findings suggest that specific alterations in erythrocyte FA composition occur very early in the disease pathogenic trajectory, prior to cognitive impairment. As erythrocyte FA levels are reflective of tissue FA, these alterations may provide insight into the pathogenic mechanism(s) of the disease and may highlight potential early diagnostic markers and therapeutic targets.

Elevated plasma ferritin in elderly individuals with high neocortical amyloid-β load

Ferritin, an iron storage and regulation protein, has been associated with Alzheimer’s disease (AD); however, it has not been investigated in preclinical AD, detected by neocortical amyloid-β load (NAL), before cognitive impairment. Cross-sectional analyses were carried out for plasma and serum ferritin in participants in the Kerr Anglican Retirement Village Initiative in Aging Health cohort. Subjects were aged 65–90 years and were categorized into high and low NAL groups via positron emission tomography using a standard uptake value ratio cutoff=1.35. Ferritin was significantly elevated in participants with high NAL compared with those with low NAL, adjusted for covariates age, sex, apolipoprotein E ɛ4 carriage and levels of C-reactive protein (an inflammation marker). Ferritin was also observed to correlate positively with NAL. A receiver operating characteristic curve based on a logistic regression of the same covariates, the base model, distinguished high from low NAL (area under the curve (AUC)=0.766), but was outperformed when plasma ferritin was added to the base model (AUC=0.810), such that at 75% sensitivity, the specificity increased from 62 to 71% on adding ferritin to the base model, indicating that ferritin is a statistically significant additional predictor of NAL over and above the base model. However, ferritin’s contribution alone is relatively minor compared with the base model. The current findings suggest that impaired iron mobilization is an early event in AD pathogenesis. Observations from the present study highlight ferritin’s potential to contribute to a blood biomarker panel for preclinical AD.

Heritability and genetic correlation between the cerebral cortex and associated white matter connections.

The aim of this study is to investigate the genetic influence on the cerebral cortex, based on the analyses of heritability and genetic correlation between grey matter (GM) thickness, derived from structural MR images (sMRI), and associated white matter (WM) connections obtained from diffusion MRI (dMRI). We measured on sMRI the cortical thickness (CT) from a large twin imaging cohort using a surface‐based approach (N = 308, average age 22.8 ± 2.3 SD). An ACE model was employed to compute the heritability of CT. WM connections were estimated based on probabilistic tractography using fiber orientation distributions (FOD) from dMRI. We then fitted the ACE model to estimate the heritability of CT and FOD peak measures along WM fiber tracts. The WM fiber tracts where genetic influence was detected were mapped onto the cortical surface. Bivariate genetic modeling was performed to estimate the cross‐trait genetic correlation between the CT and the FOD‐based connectivity of the tracts associated with the cortical regions. We found some cortical regions displaying heritable and genetically correlated GM thickness and WM connectivity, forming networks under stronger genetic influence. Significant heritability and genetic correlations between the CT and WM connectivity were found in regions including the right postcentral gyrus, left posterior cingulate gyrus, right middle temporal gyri, suggesting common genetic factors influencing both GM and WM. Hum Brain Mapp 37:2331–2347, 2016.

Surface-Base Approach Using a Multi-scale EM-ICP Registration for Statistical Population Analysis

The human cortex is a folded ribbon of neurons with a high inter-individual variability. It is a challenging structure to study especially when measuring small changes resulting from normal aging and neurodegenerative disorders such as Alzheimers Disease (AD). Recent studies have proposed surface based approaches for statistical population comparison of cortical changes since such approaches better cope with the surfacic nature of the cortex. In this paper we present a new multi-scale EM-ICP registration that is embedded into a surface-based approach. We compare this new registration algorithm with the shape context in the context of statistical population analysis. When comparing the cortical thickness between healthy elderly subjects to Alzheimers disease patients, the new pipeline reduces the intra class variability while increasing the statistical power of the T-tests between both groups.