Location of brain lesions predicts conversion of clinically isolated syndromes to multiple sclerosis.
ABSTRACT OBJECTIVES: To assess in a large population of patients with clinically isolated syndrome (CIS) the relevance of brain lesion location and frequency in predicting 1-year conversion to multiple sclerosis (MS). METHODS: In this multicenter, retrospective study, clinical and MRI data at onset and clinical follow-up at 1 year were collected for 1,165 patients with CIS. On T2-weighted MRI, we generated lesion probability maps of white matter (WM) lesion location and frequency. Voxelwise analyses were performed with a nonparametric permutation-based approach (p < 0.05, cluster-corrected). RESULTS: In CIS patients with hemispheric, multifocal, and brainstem/cerebellar onset, lesion probability map clusters were seen in clinically eloquent brain regions. Significant lesion clusters were not found in CIS patients with optic nerve and spinal cord onset. At 1 year, clinically definite MS developed in 26% of patients. The converting group, despite a greater baseline lesion load compared with the nonconverting group (7 ± 8.1 cm(3) vs 4.6 ± 6.7 cm(3), p < 0.001), showed less widespread lesion distribution (18% vs 25% of brain voxels occupied by lesions). High lesion frequency was found in the converting group in projection, association, and commissural WM tracts, with larger clusters being in the corpus callosum, corona radiata, and cingulum. CONCLUSIONS: Higher frequency of lesion occurrence in clinically eloquent WM tracts can characterize CIS subjects with different types of onset. The involvement of specific WM tracts, in particular those traversed by fibers involved in motor function and near the corpus callosum, seems to be associated with a higher risk of clinical conversion to MS in the short term.
SourceAvailable from: Mary Beth Nebel[Show abstract] [Hide abstract]
ABSTRACT: Brain lesion localization in multiple sclerosis (MS) is thought to be associated with the type and severity of adverse health effects. However, several factors hinder statistical analyses of such associations using large MRI datasets: 1) spatial registration algorithms developed for healthy individuals may be less effective on diseased brains and lead to different spatial distributions of lesions; 2) interpretation of results requires the careful selection of confounders; and 3) most approaches have focused on voxel-wise regression approaches. In this paper, we evaluated the performance of five registration algorithms and observed that conclusions regarding lesion localization can vary substantially with the choice of registration algorithm. Methods for dealing with confounding factors due to differences in disease duration and local lesion volume are introduced. Voxel-wise regression is then extended by the introduction of a metric that measures the distance between a patient-specific lesion mask and the population prevalence map.PLoS ONE 09/2014; 9(9):e107263. DOI:10.1371/journal.pone.0107263 · 3.53 Impact Factor
[Show abstract] [Hide abstract]
ABSTRACT: We hypothesized that appraisal of brain connectivity may shed light on the substrate of the radiologically isolated syndrome (RIS), a term applied to asymptomatic subjects with brain MRI abnormalities highly suggestive of multiple sclerosis. We thus used a multimodal MRI approach on the human brain by modeling measures of microstructural integrity of white matter (WM) tracts with those of functional connectivity (FC) at the level of resting state networks in RIS subjects, demographically matched normal controls (NC), and relapsing-remitting (RR) MS patients, also matched with RIS for brain macrostructural damage (i.e., lesions and atrophy). Compared with NC, in both RIS subjects and MS patients altered integrity of WM tracts was present. However, RIS subjects showed, at a less conservative threshold, lower diffusivities than RRMS patients in distinct cerebral associative, commissural, projection, and cerebellar WM tracts, suggesting a relatively better anatomical connectivity. FC was similar in NC and RIS subjects, even in the presence of important risk factors for MS (spinal cord lesions, oligoclonal bands, and dissemination in time on MRI) and increased in RRMS patients in two clinically relevant networks subserving "processing" (sensorimotor) and "control" (working memory) functions. In RIS, the lack of functional reorganization in key brain networks may represent a model of "functional reserve," which may become upregulated, with an adaptive or maladaptive role, only at a later stage in case of occurrence of clinical deficit. Copyright © 2015 the authors 0270-6474/15/350550-09$15.00/0.The Journal of Neuroscience : The Official Journal of the Society for Neuroscience 01/2015; 35(2):550-8. DOI:10.1523/JNEUROSCI.2557-14.2015 · 6.75 Impact Factor
[Show abstract] [Hide abstract]
ABSTRACT: We aim to determine if machine learning techniques, such as support vector machines (SVMs), can predict the occurrence of a second clinical attack, which leads to the diagnosis of clinically-definite Multiple Sclerosis (CDMS) in patients with a clinically isolated syndrome (CIS), on the basis of single patient's lesion features and clinical/demographic characteristics. Seventy-four patients at onset of CIS were scanned and clinically reviewed after one and three years. CDMS was used as the gold standard against which SVM classification accuracy was tested. Radiological features related to lesional characteristics on conventional MRI were defined a priori and used in combination with clinical/demographic features in an SVM. Forward recursive feature elimination with 100 bootstraps and a leave-one-out cross-validation was used to find the most predictive feature combinations. 30 % and 44 % of patients developed CDMS within one and three years, respectively. The SVMs correctly predicted the presence (or the absence) of CDMS in 71.4 % of patients (sensitivity/specificity: 77 %/66 %) at 1 year, and in 68 % (60 %/76 %) at 3 years on average over all bootstraps. Combinations of features consistently gave a higher accuracy in predicting outcome than any single feature. Machine-learning-based classifications can be used to provide an “individualised” prediction of conversion to MS from subjects' baseline scans and clinical characteristics, with potential to be incorporated into routine clinical practice.01/2015; 7:281-7. DOI:10.1016/j.nicl.2014.11.021