Lesion Recognition in Multiple Sclerosis: A Sequence Comparison and Quantification Study at 3T

Introduction Lesion detection in multiple sclerosis (MS) is an essential part of its clinical diagnosis. In addition, radiological characterisation of MS lesions is an important research field that aims at distinguishing different MS types, monitoring drug response and prognosis. To date, various MR protocols have been proposed to obtain optimal lesion contrast for early and comprehensive diagnosis of the MS disease. In this study, we compare the sensitivity of five different MR contrasts for lesion detection: (i) the DIR sequence (Double Inversion Recovery, [4]), (ii) the Dark-fluid SPACE acquisition schemes, a 3D variant of a 2D FLAIR sequence [1], (iii) the MP2RAGE [2], an MP-RAGE variant that provides homogeneous T1 contrast and quantitative T1-values, and the sequences currently used for clinical MS diagnosis (2D FLAIR, MP-RAGE). Furthermore, we investigate the T1 relaxation times of cortical and sub-cortical regions in the brain hemispheres and the cerebellum at 3T. Methods 10 early-stage female MS patients (age: 31.64.7y; disease duration: 3.81.9y; disability score, EDSS: 1.80.4) and 10 healthy controls (age and gender-matched: 31.25.8y) were included in the study after obtaining informed written consent according to the local ethic protocol. All experiments were performed at 3T (Magnetom Trio a Tim System, Siemens, Germany) using a 32-channel head coil [5]. The imaging protocol included the following sequences, (all except for axial FLAIR 2D with 1x1x1.2 mm3 voxel and 256x256x160 matrix): DIR (TI1/TI2/TR XX/3652/10000 ms, iPAT=2, TA 12:02 min), MP-RAGE (TI/TR 900/2300 ms, iPAT=3, TA 3:47 min); MP2RAGE (TI1/TI2/TR 700/2500/5000 ms, iPAT=3, TA 8:22 min, cf. [2]); 3D FLAIR SPACE (only for patient 4-6, TI/TR 1800/5000 ms, iPAT=2, TA=5;52 min, cf. [1]); Axial FLAIR (0.9x0.9x2.5 mm3, 256x256x44 matrix, TI/TR 2500/9000 ms, iPAT=2, TA 4:05 min). Lesions were identified by two experienced neurologist and radiologist, manually contoured and assigned to regional locations (s. table 1). Regional lesion masks (RLM) from each contrast were compared for number and volumes of lesions. In addition, RLM were merged in a single "master" mask, which represented the sum of the lesions of all contrasts. T1 values were derived for each location from this mask for patients 5-10 (3D FLAIR contrast was missing for patient 1-4). Results & Discussion The DIR sequence appears the most sensitive for total lesions count, followed by the MP2RAGE (table 1). The 3D FLAIR SPACE sequence turns out to be more sensitive than the 2D FLAIR, presumably due to reduced partial volume effects. Looking for sub-cortical hemispheric lesions, the DIR contrast appears to be equally sensitive to the MP2RAGE and SPACE, but most sensitive for cerebellar MS plaques. The DIR sequence is also the one that reveals cortical hemispheric lesions best. T1 relaxation times at 3T in the WM and GM of the hemispheres and the cerebellum, as obtained with the MP2RAGE sequence, are shown in table 2. Extending previous studies, we confirm overall longer T1-values in lesion tissue and higher standard deviations compared to the non-lesion tissue and control tissue in healthy controls. We hypothesize a biological (different degree of axonal loss and demyelination) rather than technical origin. Conclusion In this study, we applied 5 MR contrasts including two novel sequences to investigate the contrast of highest sensitivity for early MS diagnosis. In addition, we characterized for the first time the T1 relaxation time in cortical and sub-cortical regions of the hemispheres and the cerebellum. Results are in agreement with previous publications and meaningful biological interpretation of the data.

Machine learning algorithms for spatial data. Case studies: Environmental pollution, natural hazards, renewable resources

This paper presents general problems and approaches for the spatial data analysis using machine learning algorithms. Machine learning is a very powerful approach to adaptive data analysis, modelling and visualisation. The key feature of the machine learning algorithms is that they learn from empirical data and can be used in cases when the modelled environmental phenomena are hidden, nonlinear, noisy and highly variable in space and in time. Most of the machines learning algorithms are universal and adaptive modelling tools developed to solve basic problems of learning from data: classification/pattern recognition, regression/mapping and probability density modelling. In the present report some of the widely used machine learning algorithms, namely artificial neural networks (ANN) of different architectures and Support Vector Machines (SVM), are adapted to the problems of the analysis and modelling of geo-spatial data. Machine learning algorithms have an important advantage over traditional models of spatial statistics when problems are considered in a high dimensional geo-feature spaces, when the dimension of space exceeds 5. Such features are usually generated, for example, from digital elevation models, remote sensing images, etc. An important extension of models concerns considering of real space constrains like geomorphology, networks, and other natural structures. Recent developments in semi-supervised learning can improve modelling of environmental phenomena taking into account on geo-manifolds. An important part of the study deals with the analysis of relevant variables and models' inputs. This problem is approached by using different feature selection/feature extraction nonlinear tools. To demonstrate the application of machine learning algorithms several interesting case studies are considered: digital soil mapping using SVM, automatic mapping of soil and water system pollution using ANN; natural hazards risk analysis (avalanches, landslides), assessments of renewable resources (wind fields) with SVM and ANN models, etc. The dimensionality of spaces considered varies from 2 to more than 30. Figures 1, 2, 3 demonstrate some results of the studies and their outputs. Finally, the results of environmental mapping are discussed and compared with traditional models of geostatistics.

Cre recombinase-mediated inactivation of H-2Dd transgene expression: evidence for partial missing self-recognition by Ly49A NK cells.

We have established H-2D(d)-transgenic (Tg) mice, in which H-2D(d) expression can be extinguished by Cre recombinase-mediated deletion of an essential portion of the transgene (Tg). NK cells adapted to the expression of the H-2D(d) Tg in H-2(b) mice and acquired reactivity to cells lacking H-2D(d), both in vivo and in vitro. H-2D(d)-Tg mice crossed to mice harboring an Mx-Cre Tg resulted in mosaic H-2D(d) expression. That abrogated NK cell reactivity to cells lacking D(d). In D(d) single Tg mice it is the Ly49A+ NK cell subset that reacts to cells lacking D(d), because the inhibitory Ly49A receptor is no longer engaged by its D(d) ligand. In contrast, Ly49A+ NK cells from D(d) x MxCre double Tg mice were unable to react to D(d)-negative cells. These Ly49A+ NK cells retained reactivity to target cells that were completely devoid of MHC class I molecules, suggesting that they were not anergic. Variegated D(d) expression thus impacts specifically missing D(d) but not globally missing class I reactivity by Ly49A+ NK cells. We propose that the absence of D(d) from some host cells results in the acquisition of only partial missing self-reactivity.

Education of Murine NK Cells Requires Both cis and trans Recognition of MHC Class I Molecules.

Although NK cells use invariant receptors to identify diseased cells, they nevertheless adapt to their environment, including the presence of certain MHC class I (MHC-I) molecules. This NK cell education, which is mediated by inhibitory receptors specific for MHC-I molecules, changes the responsiveness of activating NK cell receptors (licensing) and modifies the repertoire of MHC-I receptors used by NK cells. The fact that certain MHC-I receptors have the unusual capacity to recognize MHC-I molecules expressed by other cells (trans) and by the NK cell itself (cis) has raised the question regarding possible contributions of the two types of interactions to NK cell education. Although the analysis of an MHC-I receptor variant suggested a role for cis interaction for NK cell licensing, adoptive NK cell transfer experiments supported a key role for trans recognition. To reconcile some of these findings, we have analyzed the impact of cell type-specific deletion of an MHC-I molecule and of a novel MHC-I receptor variant on the education of murine NK cells when these mature under steady-state conditions in vivo. We find that MHC-I expression by NK cells (cis) and by T cells (trans), and MHC-I recognition in cis and in trans, are both needed for NK cell licensing. Unexpectedly, modifications of the MHC-I receptor repertoire are chiefly dependent on cis binding, which provides additional support for an essential role for this unconventional type of interaction for NK cell education. These data suggest that two separate functions of MHC-I receptors are needed to adapt NK cells to self-MHC-I.

Intracellular pattern-recognition receptors

The last ten years of research in the field of innate immunity have been incredibly fertile: the transmembrane Toll-like receptors (TLRs) were discovered as guardians protecting the host against microbial attacks and the emerging pathways characterized in detail. More recently, cytoplasmic sensors were identified, which are capable of detecting not only microbial, but also self molecules. Importantly, while such receptors trigger crucial host responses to microbial insult, over-activity of some of them has been linked to autoinflammatory disorders, hence demonstrating the importance of tightly regulating their actions over time and space. Here, we provide an overview of recent findings covering this area of innate and inflammatory responses that originate from the cytoplasm

Comparison of fundus autofluorescence images acquired by the confocal scanning laser ophthalmoscope (488 nm excitation) and the modified Topcon fundus camera (580 nm excitation).

To compare autofluorescence (AF) images obtained with the confocal scanning laser ophthalmoscope (using the Heidelberg retina angiograph; HRA) and the modified Topcon fundus camera, in a routine clinical setting. A prospective comparative study conducted at the Jules-Gonin Eye Hospital. Fifty-six patients from the medical retina clinic. All patients had complete ophthalmic slit-lamp and fundus examinations, colour and red-free fundus photography, AF imaging with both instruments, and fluorescein angiography. Cataract and fixation were graded clinically. AF patterns were analyzed for healthy and pathological features. Differences of image noise were analyzed by cataract grading and fixation. A total of 105 eyes were included. AF patterns discovered by the retina angiograph and the fundus camera images, respectively, were a dark optic disc in 72 % versus 15 %, a dark fovea in 92 % versus 4 %, sub- and intraretinal fluid visible as hyperautofluorescence on HRA images only, lipid exudates visible as hypoautofluorescence on HRA images only. The same autofluorescent pattern was found on both images for geographic atrophy, retinal pigment changes, drusen and haemorrhage. Image noise was significantly associated with the degree of cataract and/or poor fixation, favouring the fundus camera. Images acquired by the fundus camera before and after fluorescein angiography were identical. Fundus AF images differ according to the technical differences of the instruments used. Knowledge of these differences is important not only for correctly interpreting images, but also for selecting the most appropriate instrument for the clinical situation.

Extent of hypoattenuation on CT angiography source images in basilar artery occlusion: prognostic value in the Basilar Artery International Cooperation Study.

BACKGROUND AND PURPOSE: The posterior circulation Acute Stroke Prognosis Early CT Score (pc-ASPECTS) quantifies the extent of early ischemic changes in the posterior circulation with a 10-point grading system. We hypothesized that pc-ASPECTS applied to CT angiography source images predicts functional outcome of patients in the Basilar Artery International Cooperation Study (BASICS). METHODS: BASICS was a prospective, observational registry of consecutive patients with acute symptomatic basilar artery occlusion. Functional outcome was assessed at 1 month. We applied pc-ASPECTS to CT angiography source images of patients with CT angiography for confirmation of basilar artery occlusion. We calculated unadjusted and adjusted risk ratios (RRs) of pc-ASPECTS dichotomized at ≥8 versus <8. Primary outcome measure was favorable outcome (modified Rankin Scale scores 0-3). Secondary outcome measures were mortality and functional independence (modified Rankin Scale scores 0-2). RESULTS: Of 158 patients included, 78 patients had a CT angiography source images pc-ASPECTS≥8. Patients with a pc-ASPECTS≥8 more often had a favorable outcome than patients with a pc-ASPECTS<8 (crude RR, 1.7; 95% CI, 0.98-3.0). After adjustment for age, baseline National Institutes of Health Stroke Scale score, and thrombolysis, pc-ASPECTS≥8 was not related to favorable outcome (RR, 1.3; 95% CI, 0.8-2.2), but it was related to reduced mortality (RR, 0.7; 95% CI, 0.5-0.98) and functional independence (RR, 2.0; 95% CI, 1.1-3.8). In post hoc analysis, pc-ASPECTS dichotomized at ≥6 versus <6 predicted a favorable outcome (adjusted RR, 3.1; 95% CI, 1.2-7.5). CONCLUSIONS: pc-ASPECTS on CT angiography source images independently predicted death and functional independence at 1 month in the CT angiography subgroup of patients in the BASICS registry.