Sainte-Foy-lès-Lyon, France
Sainte-Foy-lès-Lyon, France

Time filter

Source Type

Merida I.,University of Lyon | Merida I.,Siemens AG | Reilhac A.,CERMEP Imagerie du Vivant | Redoute J.,CERMEP Imagerie du Vivant | And 5 more authors.
Physics in Medicine and Biology | Year: 2017

In simultaneous PET-MR, attenuation maps are not directly available. Essential for absolute radioactivity quantification, they need to be derived from MR or PET data to correct for gamma photon attenuation by the imaged object. We evaluate a multi-atlas attenuation correction method for brain imaging (MaxProb) on static [18F]FDG PET and, for the first time, on dynamic PET, using the serotoninergic tracer [18F]MPPF. A database of 40 MR/CT image pairs (atlases) was used. The MaxProb method synthesises subject-specific pseudo-CTs by registering each atlas to the target subject space. Atlas CT intensities are then fused via label propagation and majority voting. Here, we compared these pseudo-CTs with the real CTs in a leave-one-out design, contrasting the MaxProb approach with a simplified single-atlas method (SingleAtlas). We evaluated the impact of pseudo-CT accuracy on reconstructed PET images, compared to PET data reconstructed with real CT, at the regional and voxel levels for the following: radioactivity images; time-activity curves; and kinetic parameters (non-displaceable binding potential, BPND). On static [18F]FDG, the mean bias for MaxProb ranged between 0 and 1% for 73 out of 84 regions assessed, and exceptionally peaked at 2.5% for only one region. Statistical parametric map analysis of MaxProb-corrected PET data showed significant differences in less than 0.02% of the brain volume, whereas SingleAtlas-corrected data showed significant differences in 20% of the brain volume. On dynamic [18F]MPPF, most regional errors on BPND ranged from -1 to +3% (maximum bias 5%) for the MaxProb method. With SingleAtlas, errors were larger and had higher variability in most regions. PET quantification bias increased over the duration of the dynamic scan for SingleAtlas, but not for MaxProb. We show that this effect is due to the interaction of the spatial tracer-distribution heterogeneity variation over time with the degree of accuracy of the attenuation maps. This work demonstrates that inaccuracies in attenuation maps can induce bias in dynamic brain PET studies. Multi-atlas attenuation correction with MaxProb enables quantification on hybrid PET-MR scanners, eschewing the need for CT. © 2017 Institute of Physics and Engineering in Medicine.

Yankam Njiwa J.,Neurodis Foundation | Bouvard S.,CERMEP Imagerie du Vivant | Mauguiere F.,University of Lyon | Ryvlin P.,University of Lyon | Hammers A.,Neurodis Foundation
NeuroImage: Clinical | Year: 2013

A third of patients with intractable temporal lobe epilepsy and hippocampal sclerosis (HS) are not seizure free (NSF) after surgery. Increased periventricular [11Cflumazenil (FMZ) binding, reflecting heterotopic neuron concentration, has been described as one predictor of NSF outcome at the group level. We aimed to replicate this finding in an independent larger cohort and investigated whether NSF outcome can be predicted in individuals. Preoperative [11CFMZ summed radioactivity images were available for 16 patients with HS and 41 controls. Images were analyzed using SPM8, explicitly including the white matter, and correction for global radioactivity via group-specific ANCOVA. Periventricular increases were assessed with a mask and different cutoffs for distinguishing NSF and seizure free (SF) patients. NSF patients had increased [11CFMZ binding around the posterior horn of the ventricles ipsilaterally (z = 2.53) and contralaterally (z = 4.44) to the seizure focus compared with SF patients. Compared with controls, SF patients had fewer periventricular increases (two clusters, total volume 0.87 cm 3, zmax = 3.8) than NSF patients (two ipsilateral and three contralateral clusters, 6.15 cm3, zmax = 4.8). In individuals and at optimized cutoffs, five (63%) of eight NSF patients and one (13%) of eight SF patients showed periventricular increases compared with controls (accuracy 75%). Only one (2%) of the 41 controls had increases at the same cutoff. The association between periventricular [11CFMZ increases and NSF outcome after temporal lobe resection for HS has been confirmed in an independent cohort on simple summed activity images. [ 11CFMZ-PET may be useful for individual preoperative counseling with clinically relevant accuracy. © 2013. Published by The Author. All rights reserved.

Azami M.E.,University of Lyon | Hammers A.,Neurodis Foundation | Hammers A.,King's College London | Jung J.,French Institute of Health and Medical Research | And 3 more authors.
PLoS ONE | Year: 2016

Pattern recognition methods, such as computer aided diagnosis (CAD) systems, can help clinicians in their diagnosis by marking abnormal regions in an image. We propose a machine learning system based on a one-class support vector machine (OC-SVM) classifier for the detection of abnormalities in magnetic resonance images (MRI) applied to patients with intractable epilepsy. The system learns the features associated with healthy control subjects, allowing a voxelwise assessment of the deviation of a test subject pattern from the learned patterns. While any number of various features can be chosen and learned, here we focus on two texture parameters capturing image patterns associated with epileptogenic lesions on T1-weighted brain MRI e.g. heterotopia and blurred junction between the grey and white matter. The CAD output consists of patient specific 3D maps locating clusters of suspicious voxels ranked by size and degree of deviation from control patterns. System performance was evaluated using realistic simulations of challenging detection tasks as well as clinical data of 77 healthy control subjects and of eleven patients (13 lesions). It was compared to that of a mass univariate statistical parametric mapping (SPM) single subject analysis based on the same set of features. For all simulations, OCSVM yielded significantly higher values of the area under the ROC curve (AUC) and higher sensitivity at low false positive rate. For the clinical data, both OC-SVM and SPM successfully detected 100% of the lesions in the MRI positive cases (3/13). For the MRI negative cases (10/13), OC-SVM detected 7/10 lesions and SPM analysis detected 5/10 lesions. In all experiments, OC-SVM produced fewer false positive detections than SPM. OC-SVM may be a versatile system for unbiased lesion detection. © 2016 El Azami et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.

Rizzo G.,University of Padua | Veronese M.,University of Padua | Veronese M.,King's College London | Heckemann R.A.,Gothenburg University | And 5 more authors.
Journal of Cerebral Blood Flow and Metabolism | Year: 2014

Substantial efforts are being spent on postmortem mRNA transcription mapping on the assumption that in vivo protein distribution can be predicted from such data. We tested this assumption by comparing mRNA transcription maps from the Allen Human Brain Atlas with reference protein concentration maps acquired with positron emission tomography (PET) in two representative systems of neurotransmission (opioid and serotoninergic). We found a tight correlation between mRNA expression and specific binding with 5-HT1A receptors measured with PET, but for opioid receptors, the correlation was weak. The discrepancy can be explained by differences in expression regulation between the two systems: transcriptional mechanisms dominate the regulation in the serotoninergic system, whereas in the opioid system proteins are further modulated after transcription. We conclude that mRNA information can be exploited for systems where translational mechanisms predominantly regulate expression. Where posttranscriptional mechanisms are important, mRNA data have to be interpreted with caution. The methodology developed here can be used for probing assumptions about the relationship of mRNA and protein in multiple neurotransmission systems. © 2014 ISCBFM.

Yakushev I.,University Medical Center Mainz | Gerhard A.,University Medical Center Mainz | Gerhard A.,University of Manchester | Mu ller M.J.,University Medical Center Mainz | And 9 more authors.
Brain Structure and Function | Year: 2011

Abnormal microstructural integrity and glucose metabolism of the hippocampus are common in subjects with Alzheimer's disease (AD) that typically manifest as episodic memory impairment. The above-tissue alterations can be captured in vivo using diffusion tensor imaging (DTI) and positron emission tomography with [18F] fluorodeoxyglucose (FDG-PET). Here, we explored relationships between the above neuroimaging and cognitive markers of early AD-specific hippocampal damage. Twenty patients with early AD (MMSE 25.7 ± 1.7) were studied using DTI and FDG-PET. Episodic memory performance was assessed using the free delayed verbal recall task (DVR). In the between-modality correlation analysis, FDG uptake was strongly associated with diffusivity in the left anterior hippocampus only (r = -0.81, p<0.05 Bonferroni's corrected for multiple tests). Performance on DVR significantly correlated with left anterior (r = -0.80, p<0.05) and left mean (r = -0.72, p<0.05) hippocampal diffusivity, while the correlation with left anterior FDG uptake did not reach statistical significance (r = 0.52, n.s.). DTI-derived diffusivity of the anterior hippocampus might be a sensitive early marker of hippocampal dysfunction as reflected at the synaptic and cognitive levels. This neurobiological distinction of the anterior hippocampus might be related to the disruption of the perforant pathway that is known to occur early in the course of AD. © Springer-Verlag 2011.

Keihaninejad S.,Imperial College London | Heckemann R.A.,Imperial College London | Heckemann R.A.,Neurodis Foundation | Gousias I.S.,Imperial College London | And 7 more authors.
PLoS ONE | Year: 2012

Brain images contain information suitable for automatically sorting subjects into categories such as healthy controls and patients. We sought to identify morphometric criteria for distinguishing controls (n = 28) from patients with unilateral temporal lobe epilepsy (TLE), 60 with and 20 without hippocampal atrophy (TLE-HA and TLE-N, respectively), and for determining the presumed side of seizure onset. The framework employs multi-atlas segmentation to estimate the volumes of 83 brain structures. A kernel-based separability criterion was then used to identify structures whose volumes discriminate between the groups. Next, we applied support vector machines (SVM) to the selected set for classification on the basis of volumes. We also computed pairwise similarities between all subjects and used spectral analysis to convert these into persubject features. SVM was again applied to these feature data. After training on a subgroup, all TLE-HA patients were correctly distinguished from controls, achieving an accuracy of 96 ± 2% in both classification schemes. For TLE-N patients, the accuracy was 86 ± 2% based on structural volumes and 91 ± 3% using spectral analysis. Structures discriminating between patients and controls were mainly localized ipsilaterally to the presumed seizure focus. For the TLE-HA group, they were mainly in the temporal lobe; for the TLE-N group they included orbitofrontal regions, as well as the ipsilateral substantia nigra. Correct lateralization of the presumed seizure onset zone was achieved using hippocampi and parahippocampal gyri in all TLE-HA patients using either classification scheme; in the TLE-N patients, lateralization was accurate based on structural volumes in 86 ± 4%, and in 94 ± 4% with the spectral analysis approach. Unilateral TLE has imaging features that can be identified automatically, even when they are invisible to human experts. Such morphometric image features may serve as classification and lateralization criteria. The technique also detects unsuspected distinguishing features like the substantia nigra, warranting further study. © 2012 Keihaninejad et al.

Malik S.J.,Imperial College London | Keihaninejad S.,Imperial College London | Hammers A.,Neurodis Foundation | Hajnal J.V.,Imperial College London
Magnetic Resonance in Medicine | Year: 2012

Brain images acquired at 3T often display central brightening with spatially varying tissue contrast, caused by inhomogeneity in the transmit radiofrequency fields used for excitation. Tailored radiofrequency pulses can provide mitigation of radiofrequency field inhomogeneity, but previous designs have been unsuitable for 3D imaging in rapid pulse sequences. This article presents a nonselective pulse design based on a short (1 ms) 3D spiral k-space trajectory that covers low spatial frequencies. The resulting excitations are optimized to produce a uniform excitation within a specified volume of interest covering the whole brain. B1 mapping and pulse calculation times were reduced by optimizing in only five slices within the brain. The method has been tested with both single and parallel transmission: in phantom experiments, normalized root-mean-square error in excitation was 0.022 for single and 0.020 for parallel transmission. The corresponding results in vivo were 0.066 and 0.055 respectively. A pilot brain imaging study using the proposed pulses for excitation within the Alzheimer's disease neuroimaging initiative magnetization prepared rapid gradient echo (MP-RAGE) protocol, yielded excellent image quality with improved signal to noise ratio in peripheral brain regions and enhanced uniformity of contrast compared with standard excitation. Greatest performance enhancement was achieved using parallel transmission, but single channel transmission offers significant improvement over standard excitation pulses. Copyright © 2011 Wiley Periodicals, Inc.

Gousias I.S.,Imperial College London | Hammers A.,Imperial College London | Hammers A.,Neurodis Foundation | Counsell S.J.,Imperial College London | And 2 more authors.
IST 2012 - 2012 IEEE International Conference on Imaging Systems and Techniques, Proceedings | Year: 2012

Automatic anatomical segmentation of pediatric brain MR data sets can be pursued with the use of registration algorithms when segmentation priors (atlases) are in hand. We investigated the performance of a maximum probability pediatric atlas (MPPA), template based registration and label propagation. The MPPA was created from the 33 pediatric data sets, available through We evaluated the performance of the MPPA comparing with manual segmentations by means of the Dice overlap coefficient. Dice values, averaged across representative regions, were 0.90 ± 0.03 for the hippocampus, 0.92 ± 0.01 for the caudate nucleus and 0.92 ± 0.02 for the pre-central gyrus. Segmentations of 36 further unsegmented target 3T images (1-year-olds and 2-year-olds) yielded visibly high-quality results. This registration approach allows the rapid construction of automatically labeled pediatric brain atlases in a single registration step. © 2012 IEEE.

Yankam Njiwa J.,Neurodis Foundation | Gray K.R.,Imperial College London | Costes N.,Cermep Imagerie du Vivant | Mauguiere F.,University of Lyon | And 2 more authors.
NeuroImage: Clinical | Year: 2015

Purpose We have previously shown that an imaging marker, increased periventricular [11C]flumazenil ([11C]FMZ) binding, is associated with failure to become seizure free (SF) after surgery for temporal lobe epilepsy (TLE) with hippocampal sclerosis (HS). Here, we investigated whether increased preoperative periventricular white matter (WM) signal can be detected on clinical [18F]FDG-PET images. We then explored the potential of periventricular FDG WM increases, as well as whole-brain [11C]FMZ and [18F]FDG images analysed with random forest classifiers, for predicting surgery outcome. Methods Sixteen patients with MRI-defined HS had preoperative [18F]FDG and [11C]FMZ-PET. Fifty controls had [18F]FDG-PET (30), [11C]FMZ-PET (41), or both (21). Periventricular WM signal was analysed using Statistical Parametric Mapping (SPM8), and whole-brain image classification was performed using random forests implemented in R ( Surgery outcome was predicted at the group and individual levels. Results At the group level, non-seizure free (NSF) versus SF patients had periventricular increases with both tracers. Against controls, NSF patients showed more prominent periventricular [11C]FMZ and [18F]FDG signal increases than SF patients. All differences were more marked for [11C]FMZ. For individuals, periventricular WM signal increases were seen at optimized thresholds in 5/8 NSF patients for both tracers. For SF patients, 1/8 showed periventricular signal increases for [11C]FMZ, and 4/8 for [18F]FDG. Hence, [18F]FDG had relatively poor sensitivity and specificity. Random forest classification accurately identified 7/8 SF and 7/8 NSF patients using [11C]FMZ images, but only 4/8 SF and 6/8 NSF patients with [18F]FDG. Conclusion This study extends the association between periventricular WM increases and NSF outcome to clinical [18F]FDG-PET, but only at the group level. Whole-brain random forest classification increases [11C]FMZ-PET's performance for predicting surgery outcome.

Ledig C.,Imperial College London | Wolz R.,Imperial College London | Aljabar P.,Imperial College London | Lotjonen J.,VTT Technical Research Center of Finland | And 3 more authors.
Proceedings - International Symposium on Biomedical Imaging | Year: 2012

In recent years, multi-atlas segmentation has emerged as one of the most accurate techniques for the segmentation of brain magnetic resonance (MR) images, especially when combined with intensity-based refinement techniques such as graph-cut or expectation-maximization (EM) optimization. However, most of the work so far has focused on intensity-based refinement strategies for individual anatomical structures such as the hippocampus. In this work we extend a previously proposed framework for labeling whole brain scans by incorporating a global and stationary Markov random field that ensures the consistency of the neighbourhood relations between structures with an a-priori defined model. In particular we improve the segmentation result of a locally weighted multi-atlas fusion method for 41 different structures simultaneously by applying a subsequent EM optimization step. We evaluate the proposed approach on 30 manually annotated brain MR images and observe an improvement of label overlaps to a manual reference by up to 6%. We also achieved a considerably improved group separation when the proposed segmentation framework is applied to a volumetric analysis of 404 subjects from the Alzheimer's Disease Neuroimaging Initiative (ADNI) cohort. © 2012 IEEE.

Loading Neurodis Foundation collaborators
Loading Neurodis Foundation collaborators