Free-running cardiac magnetic resonance fingerprinting: Joint T1/T2 map and Cine imaging.

PURPOSE: To develop and evaluate a novel non-ECG triggered 2D magnetic resonance fingerprinting (MRF) sequence allowing for simultaneous myocardial T1 and T2 mapping and cardiac Cine imaging. METHODS: Cardiac MRF (cMRF) has been recently proposed to provide joint T1/T2 myocardial mapping by triggering the acquisition to mid-diastole and relying on a subject-dependent dictionary of MR signal evolutions to generate the maps. In this work, we propose a novel "free-running" (non-ECG triggered) cMRF framework for simultaneous myocardial T1 and T2 mapping and cardiac Cine imaging in a single scan. Free-running cMRF is based on a transient state bSSFP acquisition with tiny golden angle radial readouts, varying flip angle and multiple adiabatic inversion pulses. The acquired data is retrospectively gated into several cardiac phases, which are reconstructed with an approach that combines parallel imaging, low rank modelling and patch-based high-order tensor regularization. Free-running cMRF was evaluated in a standardized phantom and ten healthy subjects. Comparison with reference spin-echo, MOLLI, SASHA, T2-GRASE and Cine was performed. RESULTS: T1 and T2 values obtained with the proposed approach were in good agreement with reference phantom values (ICC(A,1) > 0.99). Reported values for myocardium septum T1 were 1043 ± 48 ms, 1150 ± 100 ms and 1160 ± 79 ms for MOLLI, SASHA and free-running cMRF respectively and for T2 of 51.7 ± 4.1 ms and 44.6 ± 4.1 ms for T2-GRASE and free-running cMRF respectively. Good agreement was observed between free-running cMRF and conventional Cine 2D ejection fraction (bias = -0.83%). CONCLUSION: The proposed free-running cardiac MRF approach allows for simultaneous assessment of myocardial T1 and T2 and Cine imaging in a single scan.

Machine learning in whole-body MRI: experiences and challenges from an applied study using multicentre data.

Machine learning is now being increasingly employed in radiology to assist with tasks such as automatic lesion detection, segmentation, and characterisation. We are currently involved in an National Institute of Health Research (NIHR)-funded project, which aims to develop machine learning methods to improve the diagnostic performance and reduce the radiology reading time of whole-body magnetic resonance imaging (MRI) scans, in patients being staged for cancer (MALIBO study). We describe here the main challenges we have encountered during the course of this project. Data quality and uniformity are the two most important data traits to be considered in clinical trials incorporating machine learning. Robust data pre-processing and machine learning pipelines have been employed in MALIBO, a task facilitated by the now freely available machine learning libraries and toolboxes. Another important consideration for achieving the desired clinical outcome in MALIBO, was to effectively host the resulting machine learning output, along with the clinical images, for reading in a clinical environment. Finally, a range of legal, ethical, and clinical acceptance issues should be considered when attempting to incorporate computer-assisting tools into clinical practice. The road from translating computational methods into potentially useful clinical tools involves an analytical, stepwise adaptation approach, as well as engagement of a multidisciplinary team.

Identifying the optimal regional predictor of right ventricular global function: a high-resolution three-dimensional cardiac magnetic resonance study.

Right ventricular (RV) function has prognostic value in acute, chronic and peri-operative disease, although the complex RV contractile pattern makes rapid assessment difficult. Several two-dimensional (2D) regional measures estimate RV function, however the optimal measure is not known. High-resolution three-dimensional (3D) cardiac magnetic resonance cine imaging was acquired in 300 healthy volunteers and a computational model of RV motion created. Points where regional function was significantly associated with global function were identified and a 2D, optimised single-point marker (SPM-O) of global function developed. This marker was prospectively compared with tricuspid annular plane systolic excursion (TAPSE), septum-freewall displacement (SFD) and their fractional change (TAPSE-F, SFD-F) in a test cohort of 300 patients in the prediction of RV ejection fraction. RV ejection fraction was significantly associated with systolic function in a contiguous 7.3 cm2 patch of the basal RV freewall combining transverse (38%), longitudinal (35%) and circumferential (27%) contraction and coinciding with the four-chamber view. In the test cohort, all single-point surrogates correlated with RV ejection fraction (p < 0.010), but correlation (R) was higher for SPM-O (R = 0.44, p < 0.001) than TAPSE (R = 0.24, p < 0.001) and SFD (R = 0.22, p < 0.001), and non-significantly higher than TAPSE-F (R = 0.40, p < 0.001) and SFD-F (R = 0.43, p < 0.001). SPM-O explained more of the observed variance in RV ejection fraction (19%) and predicted it more accurately than any other 2D marker (median error 2.8 ml vs 3.6 ml, p < 0.001). We conclude that systolic motion of the basal RV freewall predicts global function more accurately than other 2D estimators. However, no markers summarise 3D contractile patterns, limiting their predictive accuracy.

White matter hyperintensity and stroke lesion segmentation and differentiation using convolutional neural networks.

White matter hyperintensities (WMH) are a feature of sporadic small vessel disease also frequently observed in magnetic resonance images (MRI) of healthy elderly subjects. The accurate assessment of WMH burden is of crucial importance for epidemiological studies to determine association between WMHs, cognitive and clinical data; their causes, and the effects of new treatments in randomized trials. The manual delineation of WMHs is a very tedious, costly and time consuming process, that needs to be carried out by an expert annotator (e.g. a trained image analyst or radiologist). The problem of WMH delineation is further complicated by the fact that other pathological features (i.e. stroke lesions) often also appear as hyperintense regions. Recently, several automated methods aiming to tackle the challenges of WMH segmentation have been proposed. Most of these methods have been specifically developed to segment WMH in MRI but cannot differentiate between WMHs and strokes. Other methods, capable of distinguishing between different pathologies in brain MRI, are not designed with simultaneous WMH and stroke segmentation in mind. Therefore, a task specific, reliable, fully automated method that can segment and differentiate between these two pathological manifestations on MRI has not yet been fully identified. In this work we propose to use a convolutional neural network (CNN) that is able to segment hyperintensities and differentiate between WMHs and stroke lesions. Specifically, we aim to distinguish between WMH pathologies from those caused by stroke lesions due to either cortical, large or small subcortical infarcts. The proposed fully convolutional CNN architecture, called uResNet, that comprised an analysis path, that gradually learns low and high level features, followed by a synthesis path, that gradually combines and up-samples the low and high level features into a class likelihood semantic segmentation. Quantitatively, the proposed CNN architecture is shown to outperform other well established and state-of-the-art algorithms in terms of overlap with manual expert annotations. Clinically, the extracted WMH volumes were found to correlate better with the Fazekas visual rating score than competing methods or the expert-annotated volumes. Additionally, a comparison of the associations found between clinical risk-factors and the WMH volumes generated by the proposed method, was found to be in line with the associations found with the expert-annotated volumes.

A spatio-temporal reference model of the aging brain.

Both normal aging and neurodegenerative disorders such as Alzheimer's disease (AD) cause morphological changes of the brain. It is generally difficult to distinguish these two causes of morphological change by visual inspection of magnetic resonance (MR) images. To facilitate making this distinction and thus aid the diagnosis of neurodegenerative disorders, we propose a method for developing a spatio-temporal model of morphological differences in the brain due to normal aging. The method utilizes groupwise image registration to characterize morphological variation across brain scans of people with different ages. To extract the deformations that are due to normal aging we use partial least squares regression, which yields modes of deformations highly correlated with age, and corresponding scores for each input subject. Subsequently, we determine a distribution of morphologies as a function of age by fitting smooth percentile curves to these scores. This distribution is used as a reference to which a person's morphology score can be compared. We validate our method on two different datasets, using images from both cognitively normal subjects and patients with Alzheimer disease (AD). Results show that the proposed framework extracts the expected atrophy patterns. Moreover, the morphology scores of cognitively normal subjects are on average lower than the scores of AD subjects, indicating that morphology differences between AD subjects and healthy subjects can be partly explained by accelerated aging. With our methods we are able to assess accelerated brain aging on both population and individual level. A spatio-temporal aging brain model derived from 988 T1-weighted MR brain scans from a large population imaging study (age range 45.9-91.7y, mean age 68.3y) is made publicly available at www.agingbrain.nl.

Instantiated mixed effects modeling of Alzheimer's disease markers.

The assessment and prediction of a subject's current and future risk of developing neurodegenerative diseases like Alzheimer's disease are of great interest in both the design of clinical trials as well as in clinical decision making. Exploring the longitudinal trajectory of markers related to neurodegeneration is an important task when selecting subjects for treatment in trials and the clinic, in the evaluation of early disease indicators and the monitoring of disease progression. Given that there is substantial intersubject variability, models that attempt to describe marker trajectories for a whole population will likely lack specificity for the representation of individual patients. Therefore, we argue here that individualized models provide a more accurate alternative that can be used for tasks such as population stratification and a subject-specific prognosis. In the work presented here, mixed effects modeling is used to derive global and individual marker trajectories for a training population. Test subject (new patient) specific models are then instantiated using a stratified "marker signature" that defines a subpopulation of similar cases within the training database. From this subpopulation, personalized models of the expected trajectory of several markers are subsequently estimated for unseen patients. These patient specific models of markers are shown to provide better predictions of time-to-conversion to Alzheimer's disease than population based models.

Construction of a fetal spatio-temporal cortical surface atlas from in utero MRI: Application of spectral surface matching.

In this study, we construct a spatio-temporal surface atlas of the developing cerebral cortex, which is an important tool for analysing and understanding normal and abnormal cortical development. In utero Magnetic Resonance Imaging (MRI) of 80 healthy fetuses was performed, with a gestational age range of 21.7 to 38.9 weeks. Topologically correct cortical surface models were extracted from reconstructed 3D MRI volumes. Accurate correspondences were obtained by applying a joint spectral analysis to cortices for sets of subjects close to a specific age. Sulcal alignment was found to be accurate in comparison to spherical demons, a state of the art registration technique for aligning 2D cortical representations (average Fréchet distance≈0.4 mm at 30 weeks). We construct consistent, unbiased average cortical surface templates, for each week of gestation, from age-matched groups of surfaces by applying kernel regression in the spectral domain. These were found to accurately capture the average cortical shape of individuals within the cohort, suggesting a good alignment of cortical geometry. Each spectral embedding and its corresponding cortical surface template provide a dual reference space where cortical geometry is aligned and a vertex-wise morphometric analysis can be undertaken.

Automated fetal brain segmentation from 2D MRI slices for motion correction.

Motion correction is a key element for imaging the fetal brain in-utero using Magnetic Resonance Imaging (MRI). Maternal breathing can introduce motion, but a larger effect is frequently due to fetal movement within the womb. Consequently, imaging is frequently performed slice-by-slice using single shot techniques, which are then combined into volumetric images using slice-to-volume reconstruction methods (SVR). For successful SVR, a key preprocessing step is to isolate fetal brain tissues from maternal anatomy before correcting for the motion of the fetal head. This has hitherto been a manual or semi-automatic procedure. We propose an automatic method to localize and segment the brain of the fetus when the image data is acquired as stacks of 2D slices with anatomy misaligned due to fetal motion. We combine this segmentation process with a robust motion correction method, enabling the segmentation to be refined as the reconstruction proceeds. The fetal brain localization process uses Maximally Stable Extremal Regions (MSER), which are classified using a Bag-of-Words model with Scale-Invariant Feature Transform (SIFT) features. The segmentation process is a patch-based propagation of the MSER regions selected during detection, combined with a Conditional Random Field (CRF). The gestational age (GA) is used to incorporate prior knowledge about the size and volume of the fetal brain into the detection and segmentation process. The method was tested in a ten-fold cross-validation experiment on 66 datasets of healthy fetuses whose GA ranged from 22 to 39 weeks. In 85% of the tested cases, our proposed method produced a motion corrected volume of a relevant quality for clinical diagnosis, thus removing the need for manually delineating the contours of the brain before motion correction. Our method automatically generated as a side-product a segmentation of the reconstructed fetal brain with a mean Dice score of 93%, which can be used for further processing.

Manifold population modeling as a neuro-imaging biomarker: application to ADNI and ADNI-GO.

We propose a framework for feature extraction from learned low-dimensional subspaces that represent inter-subject variability. The manifold subspace is built from data-driven regions of interest (ROI). The regions are learned via sparse regression using the mini-mental state examination (MMSE) score as an independent variable which correlates better with the actual disease stage than a discrete class label. The sparse regression is used to perform variable selection along with a re-sampling scheme to reduce sampling bias. We then use the learned manifold coordinates to perform visualization and classification of the subjects. Results of the proposed approach are shown using the ADNI and ADNI-GO datasets. Three types of classification techniques, including a new MRI Disease-State-Score (MRI-DSS) classifier, are tested in conjunction with two learning strategies. In the first case Alzheimer's Disease (AD) and progressive mild cognitive impairment (pMCI) subjects were grouped together, while cognitive normal (CN) and stable mild cognitive impaired (sMCI) subjects were also grouped together. In the second approach, the classifiers are learned using the original class labels (with no grouping). We show results that are comparable to other state-of-the-art methods. A classification rate of 71%, of arguably the most clinically relevant subjects, sMCI and pMCI, is shown. Additionally, we present classification accuracies between CN and early MCI (eMCI) subjects, from the ADNI-GO dataset, of 65%. To our knowledge this is the first time classification accuracies for eMCI patients have been reported.

Automatic quantification of normal cortical folding patterns from fetal brain MRI.

We automatically quantify patterns of normal cortical folding in the developing fetus from in utero MR images (N=80) over a wide gestational age (GA) range (21.7 to 38.9weeks). This work on data from healthy subjects represents a first step towards characterising abnormal folding that may be related to pathology, facilitating earlier diagnosis and intervention. The cortical boundary was delineated by automatically segmenting the brain MR image into a number of key structures. This utilised a spatio-temporal atlas as tissue priors in an expectation-maximization approach with second order Markov random field (MRF) regularization to improve the accuracy of the cortical boundary estimate. An implicit high resolution surface was then used to compute cortical folding measures. We validated the automated segmentations with manual delineations and the average surface discrepancy was of the order of 1mm. Eight curvature-based folding measures were computed for each fetal cortex and used to give summary shape descriptors. These were strongly correlated with GA (R(2)=0.99) confirming the close link between neurological development and cortical convolution. This allowed an age-dependent non-linear model to be accurately fitted to the folding measures. The model supports visual observations that, after a slow initial start, cortical folding increases rapidly between 25 and 30weeks and subsequently slows near birth. The model allows the accurate prediction of fetal age from an observed folding measure with a smaller error where growth is fastest. We also analysed regional patterns in folding by parcellating each fetal cortex using a nine-region anatomical atlas and found that Gompertz models fitted the change in lobar regions. Regional differences in growth rate were detected, with the parietal and posterior temporal lobes exhibiting the fastest growth, while the cingulate, frontal and medial temporal lobes developed more slowly.

Whole-brain mapping of structural connectivity in infants reveals altered connection strength associated with growth and preterm birth.

Cerebral white-matter injury is common in preterm-born infants and is associated with neurocognitive impairments. Identifying the pattern of connectivity changes in the brain following premature birth may provide a more comprehensive understanding of the neurobiology underlying these impairments. Here, we characterize whole-brain, macrostructural connectivity following preterm delivery and explore the influence of age and prematurity using a data-driven, nonsubjective analysis of diffusion magnetic resonance imaging data. T1- and T2-weighted and -diffusion MRI were obtained between 11 and 31 months postconceptional age in 49 infants, born between 25 and 35 weeks postconception. An optimized processing pipeline, combining anatomical, and tissue segmentations with probabilistic diffusion tractography, was used to map mean tract anisotropy. White-matter tracts where connection strength was related to age of delivery or imaging were identified using sparse-penalized regression and stability selection. Older children had stronger connections in tracts predominantly involving frontal lobe structures. Increasing prematurity at birth was related to widespread reductions in connection strength in tracts involving all cortical lobes and several subcortical structures. This nonsubjective approach to mapping whole-brain connectivity detected hypothesized changes in the strength of intracerebral connections during development and widespread reductions in connectivity strength associated with premature birth.

Automated analysis of atrial late gadolinium enhancement imaging that correlates with endocardial voltage and clinical outcomes: a 2-center study.

BACKGROUND: For late gadolinium enhancement (LGE) cardiovascular magnetic resonance (CMR) assessment of atrial scar to guide management and targeting of ablation in atrial fibrillation (AF), an objective, reproducible method of identifying atrial scar is required. OBJECTIVE: To describe an automated method for operator-independent quantification of LGE that correlates with colocated endocardial voltage and clinical outcomes. METHODS: LGE CMR imaging was performed at 2 centers, before and 3 months after pulmonary vein isolation for paroxysmal AF (n = 50). A left atrial (LA) surface scar map was constructed by using automated software, expressing intensity as multiples of standard deviation (SD) above blood pool mean. Twenty-one patients underwent endocardial voltage mapping at the time of pulmonary vein isolation (11 were redo procedures). Scar maps and voltage maps were spatially registered to the same magnetic resonance angiography (MRA) segmentation. RESULTS: The LGE levels of 3, 4, and 5SDs above blood pool mean were associated with progressively lower bipolar voltages compared to the preceding enhancement level (0.85 ± 0.33, 0.50 ± 0.22, and 0.38 ± 0.28 mV; P = .002, P < .001, and P = .048, respectively). The proportion of atrial surface area classified as scar (ie, >3 SD above blood pool mean) on preablation scans was greater in patients with postablation AF recurrence than those without recurrence (6.6% ± 6.7% vs 3.5% ± 3.0%, P = .032). The LA volume >102 mL was associated with a significantly greater proportion of LA scar (6.4% ± 5.9% vs 3.4% ± 2.2%; P = .007). CONCLUSIONS: LA scar quantified automatically by a simple objective method correlates with colocated endocardial voltage. Greater preablation scar is associated with LA dilatation and AF recurrence.

Benchmarking framework for myocardial tracking and deformation algorithms: an open access database.

In this paper we present a benchmarking framework for the validation of cardiac motion analysis algorithms. The reported methods are the response to an open challenge that was issued to the medical imaging community through a MICCAI workshop. The database included magnetic resonance (MR) and 3D ultrasound (3DUS) datasets from a dynamic phantom and 15 healthy volunteers. Participants processed 3D tagged MR datasets (3DTAG), cine steady state free precession MR datasets (SSFP) and 3DUS datasets, amounting to 1158 image volumes. Ground-truth for motion tracking was based on 12 landmarks (4 walls at 3 ventricular levels). They were manually tracked by two observers in the 3DTAG data over the whole cardiac cycle, using an in-house application with 4D visualization capabilities. The median of the inter-observer variability was computed for the phantom dataset (0.77 mm) and for the volunteer datasets (0.84 mm). The ground-truth was registered to 3DUS coordinates using a point based similarity transform. Four institutions responded to the challenge by providing motion estimates for the data: Fraunhofer MEVIS (MEVIS), Bremen, Germany; Imperial College London - University College London (IUCL), UK; Universitat Pompeu Fabra (UPF), Barcelona, Spain; Inria-Asclepios project (INRIA), France. Details on the implementation and evaluation of the four methodologies are presented in this manuscript. The manually tracked landmarks were used to evaluate tracking accuracy of all methodologies. For 3DTAG, median values were computed over all time frames for the phantom dataset (MEVIS=1.20mm, IUCL=0.73 mm, UPF=1.10mm, INRIA=1.09 mm) and for the volunteer datasets (MEVIS=1.33 mm, IUCL=1.52 mm, UPF=1.09 mm, INRIA=1.32 mm). For 3DUS, median values were computed at end diastole and end systole for the phantom dataset (MEVIS=4.40 mm, UPF=3.48 mm, INRIA=4.78 mm) and for the volunteer datasets (MEVIS=3.51 mm, UPF=3.71 mm, INRIA=4.07 mm). For SSFP, median values were computed at end diastole and end systole for the phantom dataset(UPF=6.18 mm, INRIA=3.93 mm) and for the volunteer datasets (UPF=3.09 mm, INRIA=4.78 mm). Finally, strain curves were generated and qualitatively compared. Good agreement was found between the different modalities and methodologies, except for radial strain that showed a high variability in cases of lower image quality.

A combined manifold learning analysis of shape and appearance to characterize neonatal brain development.

Large medical image datasets form a rich source of anatomical descriptions for research into pathology and clinical biomarkers. Many features may be extracted from data such as MR images to provide, through manifold learning methods, new representations of the population's anatomy. However, the ability of any individual feature to fully capture all aspects morphology is limited. We propose a framework for deriving a representation from multiple features or measures which can be chosen to suit the application and are processed using separate manifold-learning steps. The results are then combined to give a single set of embedding coordinates for the data. We illustrate the framework in a population study of neonatal brain MR images and show how consistent representations, correlating well with clinical data, are given by measures of shape and of appearance. These particular measures were chosen as the developing neonatal brain undergoes rapid changes in shape and MR appearance and were derived from extracted cortical surfaces, nonrigid deformations, and image similarities. Combined single embeddings show improved correlations demonstrating their benefit for further studies such as identifying patterns in the trajectories of brain development. The results also suggest a lasting effect of age at birth on brain morphology, coinciding with previous clinical studies.

Combining morphological information in a manifold learning framework: application to neonatal MRI.

MR image data can provide many features or measures although any single measure is unlikely to comprehensively characterize the underlying morphology. We present a framework in which multiple measures are used in manifold learning steps to generate coordinate embeddings which are then combined to give an improved single representation of the population. An application to neonatal brain MRI data shows that the use of shape and appearance measures in particular leads to biologically plausible and consistent representations correlating well with clinical data. Orthogonality among the correlations suggests the embedding components relate to comparatively independent morphological features. The rapid changes that occur in brain shape and in MR image appearance during neonatal brain development justify the use of shape measures (obtained from a deformation metric) and appearance measures (obtained from image similarity). The benefit of combining separate embeddings is demonstrated by improved correlations with clinical data and we illustrate the potential of the proposed framework in characterizing trajectories of brain development.

A common neonatal image phenotype predicts adverse neurodevelopmental outcome in children born preterm.

Diffuse white matter injury is common in preterm infants and is a candidate substrate for later cognitive impairment. This injury pattern is associated with morphological changes in deep grey nuclei, the localization of which is uncertain. We test the hypotheses that diffuse white matter injury is associated with discrete focal tissue loss, and that this image phenotype is associated with impairment at 2years. We acquired magnetic resonance images from 80 preterm infants at term equivalent (mean gestational age 29(+6)weeks) and 20 control infants (mean GA 39(+2)weeks). Diffuse white matter injury was defined by abnormal apparent diffusion coefficient values in one or more white matter region (frontal, central or posterior white matter at the level of the centrum semiovale), and morphological difference between groups was calculated from 3D images using deformation based morphometry. Neurodevelopmental assessments were obtained from preterm infants at a mean chronological age of 27.5months, and from controls at a mean age of 31.1months. We identified a common image phenotype in 66 of 80 preterm infants at term equivalent comprising: diffuse white matter injury; and tissue volume reduction in the dorsomedial nucleus of the thalamus, the globus pallidus, periventricular white matter, the corona radiata and within the central region of the centrum semiovale (t=4.42 p<0.001 false discovery rate corrected). The abnormal image phenotype is associated with reduced median developmental quotient (DQ) at 2years (DQ=92) compared with control infants (DQ=112), p<0.001. These findings indicate that specific neural systems are susceptible to maldevelopment after preterm birth, and suggest that neonatal image phenotype may serve as a useful biomarker for studying mechanisms of injury and the effect of putative therapeutic interventions.

Registration of serial SPECT/CT images for three-dimensional dosimetry in radionuclide therapy.

For radionuclide therapy, individual patient pharmacokinetics can be measured in three dimensions by sequential SPECT imaging. Accurate registration of the time series of images is central for voxel-based calculations of the residence time and absorbed dose. In this work, rigid and non-rigid methods are evaluated for registration of 6-7 SPECT/CT images acquired over a week, in anatomical regions from the head-and-neck region down to the pelvis. A method for calculation of the absorbed dose, including a voxel mass determination from the CT images, is also described. Registration of the SPECT/CT images is based on a CT-derived spatial transformation. Evaluation is focused on the CT registration accuracy, and on its impact on values of residence time and absorbed dose. According to the CT evaluation, the non-rigid method produces a more accurate registration than the rigid one. For images of the residence time and absorbed dose, registration produces a sharpening of the images. For volumes-of-interest, the differences between rigid and non-rigid results are generally small. However, the non-rigid method is more consistent for regions where non-rigid patient movements are likely, such as in the head-neck-shoulder region.

Multi-atlas based segmentation of brain images: atlas selection and its effect on accuracy.

Quantitative research in neuroimaging often relies on anatomical segmentation of human brain MR images. Recent multi-atlas based approaches provide highly accurate structural segmentations of the brain by propagating manual delineations from multiple atlases in a database to a query subject and combining them. The atlas databases which can be used for these purposes are growing steadily. We present a framework to address the consequent problems of scale in multi-atlas segmentation. We show that selecting a custom subset of atlases for each query subject provides more accurate subcortical segmentations than those given by non-selective combination of random atlas subsets. Using a database of 275 atlases, we tested an image-based similarity criterion as well as a demographic criterion (age) in a leave-one-out cross-validation study. Using a custom ranking of the database for each subject, we combined a varying number n of atlases from the top of the ranked list. The resulting segmentations were compared with manual reference segmentations using Dice overlap. Image-based selection provided better segmentations than random subsets (mean Dice overlap 0.854 vs. 0.811 for the estimated optimal subset size, n=20). Age-based selection resulted in a similar marked improvement. We conclude that selecting atlases from large databases for atlas-based brain image segmentation improves the accuracy of the segmentations achieved. We show that image similarity is a suitable selection criterion and give results based on selecting atlases by age that demonstrate the value of meta-information for selection.

Comparison and evaluation of segmentation techniques for subcortical structures in brain MRI.

The automation of segmentation of medical images is an active research area. However, there has been criticism of the standard of evaluation of methods. We have comprehensively evaluated four novel methods of automatically segmenting subcortical structures using volumetric, spatial overlap and distance-based measures. Two of the methods are atlas-based - classifier fusion and labelling (CFL) and expectation-maximisation segmentation using a dynamic brain atlas (EMS), and two model-based - profile active appearance models (PAM) and Bayesian appearance models (BAM). Each method was applied to the segmentation of 18 subcortical structures in 270 subjects from a diverse pool varying in age, disease, sex and image acquisition parameters. Our results showed that all four methods perform on par with recently published methods. CFL performed significantly better than the other three methods according to all three classes of metrics.

Automated morphological analysis of magnetic resonance brain imaging using spectral analysis.

Analysis of structural neuroimaging studies often relies on volume or shape comparisons of labeled neuroanatomical structures in two or more clinical groups. Such studies have common elements involving segmentation, morphological feature extraction for comparison, and subject and group discrimination. We combine two state-of-the-art analysis approaches, namely automated segmentation using label fusion and classification via spectral analysis to explore the relationship between the morphology of neuroanatomical structures and clinical diagnosis in dementia. We apply this framework to a cohort of normal controls and patients with mild dementia where accurate diagnosis is notoriously difficult. We compare and contrast our ability to discriminate normal and abnormal groups on the basis of structural morphology with (supervised) and without (unsupervised) knowledge of each individual's diagnosis. We test the hypothesis that morphological features resulting from Alzheimer disease processes are the strongest discriminator between groups.