Geodesic information flows

Generating multi-pathological and multi-modal images and labels for brain MRI

The last few years have seen a boom in using generative models to augment real datasets, as synthetic data can effectively model real data distributions and provide privacy-preserving, shareable datasets that can be used to train deep learning models. However, most of these methods are 2D and provide synthetic datasets that come, at most, with categorical annotations. The generation of paired images and segmentation samples that can be used in downstream, supervised segmentation tasks remains fairly uncharted territory. This work proposes a two-stage generative model capable of producing 2D and 3D semantic label maps and corresponding multi-modal images. We use a latent diffusion model for label synthesis and a VAE-GAN for semantic image synthesis. Synthetic datasets provided by this model are shown to work in a wide variety of segmentation tasks, supporting small, real datasets or fully replacing them while maintaining good performance. We also demonstrate its ability to improve downstream performance on out-of-distribution data.

Technical and clinical validation of commercial automated volumetric MRI tools for dementia diagnosis-a systematic review

Developments in neuroradiological MRI analysis offer promise in enhancing objectivity and consistency in dementia diagnosis through the use of quantitative volumetric reporting tools (QReports). Translation into clinical settings should follow a structured framework of development, including technical and clinical validation steps. However, published technical and clinical validation of the available commercial/proprietary tools is not always easy to find and pathways for successful integration into the clinical workflow are varied. The quantitative neuroradiology initiative (QNI) framework highlights six necessary steps for the development, validation and integration of quantitative tools in the clinic. In this paper, we reviewed the published evidence regarding regulatory-approved QReports for use in the memory clinic and to what extent this evidence fulfils the steps of the QNI framework. We summarize unbiased technical details of available products in order to increase the transparency of evidence and present the range of reporting tools on the market. Our intention is to assist neuroradiologists in making informed decisions regarding the adoption of these methods in the clinic. For the 17 products identified, 11 companies have published some form of technical validation on their methods, but only 4 have published clinical validation of their QReports in a dementia population. Upon systematically reviewing the published evidence for regulatory-approved QReports in dementia, we concluded that there is a significant evidence gap in the literature regarding clinical validation, workflow integration and in-use evaluation of these tools in dementia MRI diagnosis.

Characterizing the Clinical Features and Atrophy Patterns of MAPT-Related Frontotemporal Dementia With Disease Progression Modeling

BACKGROUND AND OBJECTIVE

Mutations in the MAPT gene cause frontotemporal dementia (FTD). Most previous studies investigating the neuroanatomical signature of MAPT mutations have grouped all different mutations together and shown an association with focal atrophy of the temporal lobe. The variability in atrophy patterns between each particular MAPT mutation is less well-characterized. We aimed to investigate whether there were distinct groups of MAPT mutation carriers based on their neuroanatomical signature.

METHODS

We applied Subtype and Stage Inference (SuStaIn), an unsupervised machine learning technique that identifies groups of individuals with distinct progression patterns, to characterize patterns of regional atrophy in MAPT-associated FTD within the Genetic FTD Initiative (GENFI) cohort study.

RESULTS

Eighty-two MAPT mutation carriers were analyzed, the majority of whom had P301L, IVS10+16, or R406W mutations, along with 48 healthy noncarriers. SuStaIn identified 2 groups of MAPT mutation carriers with distinct atrophy patterns: a temporal subtype, in which atrophy was most prominent in the hippocampus, amygdala, temporal cortex, and insula; and a frontotemporal subtype, in which atrophy was more localized to the lateral temporal lobe and anterior insula, as well as the orbitofrontal and ventromedial prefrontal cortex and anterior cingulate. There was one-to-one mapping between IVS10+16 and R406W mutations and the temporal subtype and near one-to-one mapping between P301L mutations and the frontotemporal subtype. There were differences in clinical symptoms and neuropsychological test scores between subtypes: the temporal subtype was associated with amnestic symptoms, whereas the frontotemporal subtype was associated with executive dysfunction.

CONCLUSION

Our results demonstrate that different MAPT mutations give rise to distinct atrophy patterns and clinical phenotype, providing insights into the underlying disease biology and potential utility for patient stratification in therapeutic trials.

SVM recursive feature elimination analyses of structural brain MRI predicts near-term relapses in patients with clinically isolated syndromes suggestive of multiple sclerosis

•

RFE-SVMs predict future outcome of CIS patients with conservative accuracy estimates between 64.9% and 88.1%.

•

Recursive feature selection improves classification performance compared to using all information.

•

Relevant features include regional WM lesion load and GM density, as well as the type of CIS onset.

•

Cross-validation introduces positive bias on accuracy estimate.

Machine learning classification is an attractive approach to automatically differentiate patients from healthy subjects, and to predict future disease outcomes. A clinically isolated syndrome (CIS) is often the first presentation of multiple sclerosis (MS), but it is difficult at onset to predict who will have a second relapse and hence convert to clinically definite MS. In this study, we thus aimed to distinguish CIS converters from non-converters at onset of a CIS, using recursive feature elimination and weight averaging with support vector machines. We also sought to assess the influence of cohort size and cross-validation methods on the accuracy estimate of the classification.

We retrospectively collected 400 patients with CIS from six European MAGNIMS MS centres. Patients underwent brain MRI at onset of a CIS according to local standard-of-care protocols. The diagnosis of clinically definite MS at one-year follow-up was the standard against which the accuracy of the model was tested. For each patient, we derived MRI-based features, such as grey matter probability, white matter lesion load, cortical thickness, and volume of specific cortical and white matter regions. Features with little contribution to the classification model were removed iteratively through an interleaved sample bootstrapping and feature averaging approach. Classification of CIS outcome at one-year follow-up was performed with 2-fold, 5-fold, 10-fold and leave-one-out cross-validation for each centre cohort independently and in all patients together.

The estimated classification accuracy across centres ranged from 64.9% to 88.1% using 2-fold cross-validation and from 73% to 92.9% using leave-one-out cross-validation. The classification accuracy estimate was higher in single-centre, smaller data sets than in combinations of data sets, being the lowest when all patients were merged together.

Regional MRI features such as WM lesions, grey matter probability in the thalamus and the precuneus or cortical thickness in the cuneus and inferior temporal gyrus predicted the occurrence of a second relapse in patients at onset of a CIS using support vector machines. The increased accuracy estimate of the classification achieved with smaller and single-centre samples may indicate a model bias (overfitting) when data points were limited, but also more homogeneous. We provide an overview of classifier performance from a range of cross-validation schemes to give insight into the variability across schemes. The proposed recursive feature elimination approach with weight averaging can be used both in single- and multi-centre data sets in order to bridge the gap between group-level comparisons and making predictions for individual patients.

Longitudinal neuroanatomical and cognitive progression of posterior cortical atrophy

Posterior cortical atrophy is a clinico-radiological syndrome characterized by progressive decline in visual processing and atrophy of posterior brain regions. With the majority of cases attributable to Alzheimer's disease and recent evidence for genetic risk factors specifically related to posterior cortical atrophy, the syndrome can provide important insights into selective vulnerability and phenotypic diversity. The present study describes the first major longitudinal investigation of posterior cortical atrophy disease progression. Three hundred and sixty-one individuals (117 posterior cortical atrophy, 106 typical Alzheimer's disease, 138 controls) fulfilling consensus criteria for posterior cortical atrophy-pure and typical Alzheimer's disease were recruited from three centres in the UK, Spain and USA. Participants underwent up to six annual assessments involving MRI scans and neuropsychological testing. We constructed longitudinal trajectories of regional brain volumes within posterior cortical atrophy and typical Alzheimer's disease using differential equation models. We compared and contrasted the order in which regional brain volumes become abnormal within posterior cortical atrophy and typical Alzheimer's disease using event-based models. We also examined trajectories of cognitive decline and the order in which different cognitive tests show abnormality using the same models. Temporally aligned trajectories for eight regions of interest revealed distinct (P < 0.002) patterns of progression in posterior cortical atrophy and typical Alzheimer's disease. Patients with posterior cortical atrophy showed early occipital and parietal atrophy, with subsequent higher rates of temporal atrophy and ventricular expansion leading to tissue loss of comparable extent later. Hippocampal, entorhinal and frontal regions underwent a lower rate of change and never approached the extent of posterior cortical involvement. Patients with typical Alzheimer's disease showed early hippocampal atrophy, with subsequent higher rates of temporal atrophy and ventricular expansion. Cognitive models showed tests sensitive to visuospatial dysfunction declined earlier in posterior cortical atrophy than typical Alzheimer's disease whilst tests sensitive to working memory impairment declined earlier in typical Alzheimer's disease than posterior cortical atrophy. These findings indicate that posterior cortical atrophy and typical Alzheimer's disease have distinct sites of onset and different profiles of spatial and temporal progression. The ordering of disease events both motivates investigation of biological factors underpinning phenotypic heterogeneity, and informs the selection of measures for clinical trials in posterior cortical atrophy.

Uncovering the heterogeneity and temporal complexity of neurodegenerative diseases with Subtype and Stage Inference

The heterogeneity of neurodegenerative diseases is a key confound to disease understanding and treatment development, as study cohorts typically include multiple phenotypes on distinct disease trajectories. Here we introduce a machine-learning technique—Subtype and Stage Inference (SuStaIn)—able to uncover data-driven disease phenotypes with distinct temporal progression patterns, from widely available cross-sectional patient studies. Results from imaging studies in two neurodegenerative diseases reveal subgroups and their distinct trajectories of regional neurodegeneration. In genetic frontotemporal dementia, SuStaIn identifies genotypes from imaging alone, validating its ability to identify subtypes; further the technique reveals within-genotype heterogeneity. In Alzheimer’s disease, SuStaIn uncovers three subtypes, uniquely characterising their temporal complexity. SuStaIn provides fine-grained patient stratification, which substantially enhances the ability to predict conversion between diagnostic categories over standard models that ignore subtype (p = 7.18 × 10⁻⁴) or temporal stage (p = 3.96 × 10⁻⁵). SuStaIn offers new promise for enabling disease subtype discovery and precision medicine.

Progressive diseases tend to be heterogeneous in their underlying aetiology mechanism, disease manifestation, and disease time course. Here, Young and colleagues devise a computational method to account for both phenotypic heterogeneity and temporal heterogeneity, and demonstrate it using two neurodegenerative disease cohorts.

Presymptomatic white matter integrity loss in familial frontotemporal dementia in the GENFI cohort: A cross-sectional diffusion tensor imaging study

Objective

We aimed to investigate mutation-specific white matter (WM) integrity changes in presymptomatic and symptomatic mutation carriers of the C9orf72,MAPT, and GRN mutations by use of diffusion-weighted imaging within the Genetic Frontotemporal dementia Initiative (GENFI) study.

Methods

One hundred and forty mutation carriers (54 C9orf72, 30 MAPT, 56 GRN), 104 presymptomatic and 36 symptomatic, and 115 noncarriers underwent 3T diffusion tensor imaging. Linear mixed effects models were used to examine the association between diffusion parameters and years from estimated symptom onset in C9orf72,MAPT, and GRN mutation carriers versus noncarriers. Post hoc analyses were performed on presymptomatic mutation carriers only, as well as left-right asymmetry analyses on GRN mutation carriers versus noncarriers.

Results

Diffusion changes in C9orf72 mutation carriers are present significantly earlier than both MAPT and GRN mutation carriers - characteristically in the posterior thalamic radiation and more posteriorly located tracts (e.g., splenium of the corpus callosum, posterior corona radiata), as early as 30 years before estimated symptom onset. MAPT mutation carriers showed early involvement of the uncinate fasciculus and cingulum, sparing the internal capsule, whereas involvement of the anterior and posterior internal capsule was found in GRN. Restricting analyses to presymptomatic mutation carriers only, similar - albeit less extensive - patterns were found: posteriorly located WM tracts (e.g., posterior thalamic radiation, splenium of the corpus callosum, posterior corona radiata) in presymptomatic C9orf72, the uncinate fasciculus in presymptomatic MAPT, and the internal capsule (anterior and posterior limbs) in presymptomatic GRN mutation carriers. In GRN, most tracts showed significant left-right differences in one or more diffusion parameter, with the most consistent results being found in the UF, EC, RPIC, and ALIC.

Interpretation

This study demonstrates the presence of early and widespread WM integrity loss in presymptomatic FTD, and suggests a clear genotypic "fingerprint." Our findings corroborate the notion of FTD as a network-based disease, where changes in connectivity are some of the earliest detectable features, and identify diffusion tensor imaging as a potential neuroimaging biomarker for disease-tracking and -staging in presymptomatic to early-stage familial FTD.

Selection bias in the reported performances of AD classification pipelines

The last decade has seen a great proliferation of supervised learning pipelines for individual diagnosis and prognosis in Alzheimer's disease. As more pipelines are developed and evaluated in the search for greater performance, only those results that are relatively impressive will be selected for publication. We present an empirical study to evaluate the potential for optimistic bias in classification performance results as a result of this selection. This is achieved using a novel, resampling-based experiment design that effectively simulates the optimisation of pipeline specifications by individuals or collectives of researchers using cross validation with limited data. Our findings indicate that bias can plausibly account for an appreciable fraction (often greater than half) of the apparent performance improvement associated with the pipeline optimisation, particularly in small samples. We discuss the consistency of our findings with patterns observed in the literature and consider strategies for bias reduction and mitigation.

•

Demonstration and measurement of selection bias in AD classification experiments

•

Bias accounts for much of the performance improvement seen with pipeline optimisation.

•

Assessment of key risk factors and guidance on best research practices

•

Evidence of selection bias in results collected by a recent literature review

A multi-time-point modality-agnostic patch-based method for lesion filling in multiple sclerosis

Multiple sclerosis lesions influence the process of image analysis, leading to tissue segmentation problems and biased morphometric estimates. Existing techniques try to reduce this bias by filling all lesions as normal-appearing white matter on T1-weighted images, considering each time-point separately. However, due to lesion segmentation errors and the presence of structures adjacent to the lesions, such as the ventricles and deep grey matter nuclei, filling all lesions with white matter-like intensities introduces errors and artefacts. In this paper, we present a novel lesion filling strategy inspired by in-painting techniques used in computer graphics applications for image completion. The proposed technique uses a five-dimensional (5D), patch-based (multi-modality and multi-time-point), Non-Local Means algorithm that fills lesions with the most plausible texture. We demonstrate that this strategy introduces less bias, fewer artefacts and spurious edges than the current, publicly available techniques. The proposed method is modality-agnostic and can be applied to multiple time-points simultaneously. In addition, it preserves anatomical structures and signal-to-noise characteristics even when the lesions are neighbouring grey matter or cerebrospinal fluid, and avoids excess of blurring or rasterisation due to the choice of the segmentation plane, shape of the lesions, and their size and/or location.

Using florbetapir positron emission tomography to explore cerebrospinal fluid cut points and gray zones in small sample sizes

INTRODUCTION

We aimed to assess the feasibility of determining Alzheimer's disease cerebrospinal fluid (CSF) cut points in small samples through comparison with amyloid positron emission tomography (PET).

METHODS

Twenty-three individuals (19 patients, four controls) had CSF measures of amyloid beta (Aβ)1-42 and total tau/Aβ1-42 ratio, and florbetapir PET. We compared CSF measures with visual and quantitative (standardized uptake value ratio [SUVR]) PET measures of amyloid.

RESULTS

Seventeen of 23 were amyloid-positive on visual reads, and 14 of 23 at an SUVR of ≥1.1. There was concordance (positive/negative on both measures) in 20 of 23, of whom 19 of 20 were correctly classified at an Aβ1-42 of 630 ng/L, and 20 of 20 on tau/Aβ1-42 ratio (positive ≥0.88; negative ≤0.34). Three discordant cases had Aβ1-42 levels between 403 and 729 ng/L and tau/Aβ1-42 ratios of 0.54-0.58.

DISCUSSION

Comparing amyloid PET and CSF biomarkers provides a means of assessing CSF cut points in vivo, and can be applied to small sample sizes. CSF tau/Aβ1-42 ratio appears robust at predicting amyloid status, although there are gray zones where there remains diagnostic uncertainty.

Multiple Orderings of Events in Disease Progression

The event-based model constructs a discrete picture of disease progression from cross-sectional data sets, with each event corresponding to a new biomarker becoming abnormal. However, it relies on the assumption that all subjects follow a single event sequence. This is a major simplification for sporadic disease data sets, which are highly heterogeneous, include distinct subgroups, and contain significant proportions of outliers. In this work we relax this assumption by considering two extensions to the event-based model: a generalised Mallows model, which allows subjects to deviate from the main event sequence, and a Dirichlet process mixture of generalised Mallows models, which models clusters of subjects that follow different event sequences, each of which has a corresponding variance. We develop a Gibbs sampling technique to infer the parameters of the two models from multi-modal biomarker data sets. We apply our technique to data from the Alzheimer's Disease Neuroimaging Initiative to determine the sequence in which brain regions become abnormal in sporadic Alzheimer's disease, as well as the heterogeneity of that sequence in the cohort. We find that the generalised Mallows model estimates a larger variation in the event sequence across subjects than the original event-based model. Fitting a Dirichlet process model detects three subgroups of the population with different event sequences. The Gibbs sampler additionally provides an estimate of the uncertainty in each of the model parameters, for example an individual's latent disease stage and cluster assignment. The distributions and mixtures of sequences that this new family of models introduces offer better characterisation of disease progression of heterogeneous populations, new insight into disease mechanisms, and have the potential for enhanced disease stratification and differential diagnosis.

Multi-contrast attenuation map synthesis for PET/MR scanners: assessment on FDG and Florbetapir PET tracers

Positron Emission Tomography/Magnetic Resonance Imaging (PET/MR) scanners are expected to offer a new range of clinical applications. Attenuation correction is an essential requirement for quantification of PET data but MRI images do not directly provide a patient-specific attenuation map.

Methods We further validate and extend a Computed Tomography (CT) and attenuation map (μ-map) synthesis method based on pre-acquired MRI-CT image pairs. The validation consists of comparing the CT images synthesised with the proposed method to the original CT images. PET images were acquired using two different tracers (¹⁸F-FDG and ¹⁸F-florbetapir). They were then reconstructed and corrected for attenuation using the synthetic μ-maps and compared to the reference PET images corrected with the CT-based μ-maps. During the validation, we observed that the CT synthesis was inaccurate in areas such as the neck and the cerebellum, and propose a refinement to mitigate these problems, as well as an extension of the method to multi-contrast MRI data.

Results With the improvements proposed, a significant enhancement in CT synthesis, which results in a reduced absolute error and a decrease in the bias when reconstructing PET images, was observed. For both tracers, on average, the absolute difference between the reference PET images and the PET images corrected with the proposed method was less than 2%, with a bias inferior to 1%.

Conclusion With the proposed method, attenuation information can be accurately derived from MRI images by synthesising CT using routine anatomical sequences. MRI sequences, or combination of sequences, can be used to synthesise CT images, as long as they provide sufficient anatomical information.

Presymptomatic cognitive and neuroanatomical changes in genetic frontotemporal dementia in the Genetic Frontotemporal dementia Initiative (GENFI) study: a cross-sectional analysis

BACKGROUND

Frontotemporal dementia is a highly heritable neurodegenerative disorder. In about a third of patients, the disease is caused by autosomal dominant genetic mutations usually in one of three genes: progranulin (GRN), microtubule-associated protein tau (MAPT), or chromosome 9 open reading frame 72 (C9orf72). Findings from studies of other genetic dementias have shown neuroimaging and cognitive changes before symptoms onset, and we aimed to identify whether such changes could be shown in frontotemporal dementia.

METHODS

We recruited participants to this multicentre study who either were known carriers of a pathogenic mutation in GRN, MAPT, or C9orf72, or were at risk of carrying a mutation because a first-degree relative was a known symptomatic carrier. We calculated time to expected onset as the difference between age at assessment and mean age at onset within the family. Participants underwent a standardised clinical assessment and neuropsychological battery. We did MRI and generated cortical and subcortical volumes using a parcellation of the volumetric T1-weighted scan. We used linear mixed-effects models to examine whether the association of neuropsychology and imaging measures with time to expected onset of symptoms differed between mutation carriers and non-carriers.

FINDINGS

Between Jan 30, 2012, and Sept 15, 2013, we recruited participants from 11 research sites in the UK, Italy, the Netherlands, Sweden, and Canada. We analysed data from 220 participants: 118 mutation carriers (40 symptomatic and 78 asymptomatic) and 102 non-carriers. For neuropsychology measures, we noted the earliest significant differences between mutation carriers and non-carriers 5 years before expected onset, when differences were significant for all measures except for tests of immediate recall and verbal fluency. We noted the largest Z score differences between carriers and non-carriers 5 years before expected onset in tests of naming (Boston Naming Test -0·7; SE 0·3) and executive function (Trail Making Test Part B, Digit Span backwards, and Digit Symbol Task, all -0·5, SE 0·2). For imaging measures, we noted differences earliest for the insula (at 10 years before expected symptom onset, mean volume as a percentage of total intracranial volume was 0·80% in mutation carriers and 0·84% in non-carriers; difference -0·04, SE 0·02) followed by the temporal lobe (at 10 years before expected symptom onset, mean volume as a percentage of total intracranial volume 8·1% in mutation carriers and 8·3% in non-carriers; difference -0·2, SE 0·1).

INTERPRETATION

Structural imaging and cognitive changes can be identified 5-10 years before expected onset of symptoms in asymptomatic adults at risk of genetic frontotemporal dementia. These findings could help to define biomarkers that can stage presymptomatic disease and track disease progression, which will be important for future therapeutic trials.

FUNDING

Centres of Excellence in Neurodegeneration.

Longitudinal measurement of the developing thalamus in the preterm brain using multi-modal MRI

Preterm birth is a significant public health concern. For infants born very preterm (≤ 32 weeks completed gestation), there is a high instance of developmental disability. Due to the heterogeneity of patient outcomes, it is important to investigate early markers of future ability to provide effective and targeted intervention. As a neuronal relay centre, the thalamus is critical for effective cognitive function and, thus, development of white matter connections between the thalamus and cortex is vital. By non-invasively examining the state of the thalamus we can monitor development in the preterm period. To track the development we develop a novel registration technique to combine data from multiple modalities, in order to derive the transformation from a preterm scan, to a scan of the same infant at term-equivalent age. By measuring the changes in diffusion parameters over this period on a per-voxel basis, we hope to provide unique insight into neurodevelopment.

Attenuation correction synthesis for hybrid PET-MR scanners: application to brain studies

Attenuation correction is an essential requirement for quantification of positron emission tomography (PET) data. In PET/CT acquisition systems, attenuation maps are derived from computed tomography (CT) images. However, in hybrid PET/MR scanners, magnetic resonance imaging (MRI) images do not directly provide a patient-specific attenuation map. The aim of the proposed work is to improve attenuation correction for PET/MR scanners by generating synthetic CTs and attenuation maps. The synthetic images are generated through a multi-atlas information propagation scheme, locally matching the MRI-derived patient's morphology to a database of MRI/CT pairs, using a local image similarity measure. Results show significant improvements in CT synthesis and PET reconstruction accuracy when compared to a segmentation method using an ultrashort-echo-time MRI sequence and to a simplified atlas-based method.

Hierarchical manifold learning for regional image analysis

We present a novel method of hierarchical manifold learning which aims to automatically discover regional properties of image datasets. While traditional manifold learning methods have become widely used for dimensionality reduction in medical imaging, they suffer from only being able to consider whole images as single data points. We extend conventional techniques by additionally examining local variations, in order to produce spatially-varying manifold embeddings that characterize a given dataset. This involves constructing manifolds in a hierarchy of image patches of increasing granularity, while ensuring consistency between hierarchy levels. We demonstrate the utility of our method in two very different settings: 1) to learn the regional correlations in motion within a sequence of time-resolved MR images of the thoracic cavity; 2) to find discriminative regions of 3-D brain MR images associated with neurodegenerative disease.

Hierarchical unbiased graph shrinkage (HUGS): a novel groupwise registration for large data set

Normalizing all images in a large data set into a common space is a key step in many clinical and research studies, e.g., for brain development, maturation, and aging. Recently, groupwise registration has been developed for simultaneous alignment of all images without selecting a particular image as template, thus potentially avoiding bias in the registration. However, most conventional groupwise registration methods do not explore the data distribution during the image registration. Thus, their performance could be affected by large inter-subject variations in the data set under registration. To solve this potential issue, we propose to use a graph to model the distribution of all image data sitting on the image manifold, with each node representing an image and each edge representing the geodesic pathway between two nodes (or images). Then, the procedure of warping all images to their population center turns to the dynamic shrinking of the graph nodes along their graph edges until all graph nodes become close to each other. Thus, the topology of image distribution on the image manifold is always preserved during the groupwise registration. More importantly, by modeling the distribution of all images via a graph, we can potentially reduce registration error since every time each image is warped only according to its nearby images with similar structures in the graph. We have evaluated our proposed groupwise registration method on both infant and adult data sets, by also comparing with the conventional group-mean based registration and the ABSORB methods. All experimental results show that our proposed method can achieve better performance in terms of registration accuracy and robustness.

Attenuation correction synthesis for hybrid PET-MR scanners

The combination of functional and anatomical imaging technologies such as Positron Emission Tomography (PET) and Computed Tomography (CT) has shown its value in the preclinical and clinical fields. In PET/CT hybrid acquisition systems, CT-derived attenuation maps enable a more accurate PET reconstruction. However, CT provides only very limited soft-tissue contrast and exposes the patient to an additional radiation dose. In comparison, Magnetic Resonance Imaging (MRI) provides good soft-tissue contrast and the ability to study functional activation and tissue microstructures, but does not directly provide patient-specific electron density maps for PET reconstruction. The aim of the proposed work is to improve PET/MR reconstruction by generating synthetic CTs and attenuation-maps. The synthetic images are generated through a multi-atlas information propagation scheme, locally matching the MRI-derived patient's morphology to a database of pre-acquired MRI/CT pairs. Results show improvements in CT synthesis and PET reconstruction accuracy when compared to a segmentation method using an Ultrashort-Echo-Time MRI sequence.

Multi-atlas segmentation with robust label transfer and label fusion

Multi-atlas segmentation has been widely applied in medical image analysis. This technique relies on image registration to transfer segmentation labels from pre-labeled atlases to a novel target image and applies label fusion to reduce errors produced by registration-based label transfer. To improve the performance of registration-based label transfer against registration errors, our first contribution is to propose a label transfer scheme that generates multiple warped versions of each atlas to one target image through registration paths obtained by composing inter-atlas registrations and atlas-target registrations. The problem of decreasing quality of warped atlases caused by accumulative errors in composing multiple registrations is properly addressed by an atlas selection method that is guided by atlas segmentations. To improve the performance of label fusion against registration errors, our second contribution is to integrate the probabilistic correspondence model employed by the non-local mean approach with the joint label fusion technique, both of which have shown excellent performance for label fusion. Experiments on mitral-valve segmentation in 3D transesophageal echocardiography (TEE) show the effectiveness of the proposed techniques.