A Correlation Study Of Clinical Findings In Multiple Sclerosis With Radiological Findings

AimTo explore the relationship between spatial distributions of multiple sclerosis (MS) lesions in brain and spinal cord magnetic resonance (MR) imaging with clinical manifestation. Hypothesizing that symptomatology may partly be determined by lesion location, this retrospective study explored relations between lesion location and disability using voxel wise analyses in standard space. Determining spinal cord atrophy among subtypes in MS and whether this offers diagnostic and clinical correlative information beyond that provided by other MRI markers and lastly to analyze the clinical and laboratory features of each subtype of MS and correlate it with radiological findings.MethodWe retrospectively analyzed the medical records of41consecutive patients with a clinical diagnosis of MS, who had been admitted to our university hospital from2006to2013. They all met criteria for definite MS, by clinical or laboratory standards. We reviewed the following items:1) clinical manifestations;2) spinal and brain MRI findings. Using nonparametric permutation based statistics, voxel based lesions probability on T2lesions masks was related to expanded disabiltiy status scale (EDSS) and MS functional composite (MSFC) sub domain scores and demographic characteristics of MS patients compared. To indentify statistically significant locations, a cluster-forming threshold of3.1was used.Upper cervical cord cross-sectional area (UCCA), brain and spinal cord lesion loads and brain atrophy were measured. Expanded Disability Status Scale (EDSS), the timed25-foot walk test (TWT), and the nine-hole peg tests were used. UCCA was compared between groups with the Mann-Whitney U test. Correlations were assessed with the Spearman r test. Multivariate associations between UCCA and clinical and other MR imaging parameters, including number of hypo intense brain lesions on T1-weighted MR images, presence of diffuse abnormalities, and number of involved segments in the spinal cord, were assessed by using multiple linear regressions.ResultsPatients with secondary progressive disease tended to have longer disease duration, a higher lesion load (LL) and worse disability scores than patients with RRMS (all P<0.001; Mann-Whitney U test). When patient characteristics were analyzed for interrelations, age (Spearman’s rho0.16; P=0.006), disease duration (Spearman’s rho0.28; P=0.001), and disability measurements (EDSS [N=41], Spearman’s rho0.27;9HPT [N=21], Spearman’s rho0.35; PASAT (N=29), Spearman’s rho-0.28; Spearman’s rho0.32; all P<0.001) correlated moderately with LL. Patients with progressive disease had a higher LL than patients with relapsing disease (N=41; P＜0.001, Mann-Whitney U test), but onset type was not related to LL (N=41; P=0.97, Mann-Whitney U test). Patients with more pronounced disability according to the "worse" quartile of disability scores had higher LL than patients in the best performing quartile (P<0.001, Mann-Whitney U test). Most lesions occurred in the periventricular region, with a peak in voxel wise lesion frequency of29%.In voxel wise analyses, lesion probability around the ventricles correlated significantly with EDSS and MSFC sub domain scores, with age, and with disease duration (all voxel wise P<0.001).UCCA, number of hypointense brain lesions on T1-weighted MR images, presence of diffuse abnormalities, and number of involved segments in the spinal cord were found to be significant explanatory factors for clinical disability (R2=0.564).The UCCA and the number of hypointense brain lesions on T1-weighted images was the strongest MR imaging parameters for explaining physical disability, as measured with the EDSS. In the whole group, UCCA had weak to barely moderate correlation with all investigated clinical parameters:UCCA was inversely correlated with disease duration (p=-0.33, P<.001), age (p=-0.23, P<.001), and EDSS score (p=-0.39, P<.001).Spinal cord lesions were present in31of the41patients with spinal cord images (74.3%), whereas almost all patients (39of41,99.5%) had abnormalities on the available T2-weighted brain MR images. Diffuse spinal cord abnormalities were present in5of the41patients with spinal cord images (14.5%). In patients with spinal cord images, a median of two spinal cord lesions (interquartile range,0-4) was found, and the median number of involved segments was1.75(interquartile range,0-3.5ConclusionOf all possible locations throughout the brain, periventricular lesions seem most important in determining disability as measured by EDSS and MSFC in MS. The lack of more specific localizations may be explained by the effect of LL on disability acting through a disconnection syndrome, and adaptive changes. As this method proved feasible in relating lesion distribution to several markers of disease, future studies may focus on the use of lesion probability mapping for studying aspects of MS less determined by LL, for instance relations between lesion location and local brain atrophy, or longitudinal lesion distribution changes.Spinal cord abnormalities have a strong effect on clinical disability in MS. MRI derived UCCA was found to be the most significant spinal cord parameter for explaining EDSS score.