Imaging Poster Presentation

P0581 - Gradient echo magnetic resonance imaging to detect central vein sign in patients with Multiple Sclerosis (ID 1528)

Speakers
  • V. Levasseur
Authors
  • V. Levasseur
  • B. Xiang
  • A. Salter
  • S. Lancia
  • D. Yablonskiy
  • A. Cross
Presentation Number
P0581
Presentation Topic
Imaging

Abstract

Background

Diagnostic criteria for multiple sclerosis (MS) have undergone several iterations in the past 20 years, resulting in increased sensitivity and earlier diagnosis, but also increased misdiagnoses due to reduced specificity. Biomarkers are needed to distinguish MS from its mimics and between MS subtypes. MS lesions in white matter (WM) typically form around a central vein, which can be visualized with FLAIR* imaging (Sati, et al, 2012). Central vein sign (CVS) presence on MRI can help differentiate MS from other diseases with WM T2-weighted hyperintensities and may increase the sensitivity and specificity of MS diagnosis.

Objectives

To apply MRI-based gradient echo plural contrast imaging (GEPCI) approach (Luo, et al, 2012) to generate FLAIR*-like images and detect CVS in patients with progressive MS (PMS) and relapsing remitting MS (RRMS). To quantitatively evaluate tissue damage in lesions with and without CVS using tissue-specific GEPCI R2t* metric (Xiang, et al, 2019).

Methods

MRI scans of PMS (n=39) and RRMS subjects (n=30) were analyzed for the CVS within WM lesions. Presence of CVS, lesion volume, and anatomic location were determined. To quantitatively evaluate the severity of brain-tissue damage in MS lesions, mean R2t* was calculated. R2t* is a quantitative measure that correlates with brain tissue cellular density and is decreased in areas of reduced tissue integrity. The proportion of total lesions with CVS was calculated, excluding confluent lesions and lesions with more than 1 central vein. Median proportions in PMS and RRMS subjects were compared using the Wilcoxon sign rank test. Individual lesion CVS status was examined using generalized linear models. Linear mixed models adjusted for age were used to evaluate predictors of mean R2t* in nonconfluent lesions with only one central vein of sufficient size. Associations of CVS and clinical data from time of MRI scan, including Expanded Disability Status Scale (EDSS), symbol digit modality test (SDMT), and multiple sclerosis functional composite (MSFC) were made using Spearman correlations.

Results

The PMS group had significantly higher EDSS scores and poorer performance on SDMT and MSFC than the RRMS group. There were no significant differences in total CVS and percentage of CVS per lesion between MS subtypes. Controlling for age, EDSS, and lesion volume did not increase the odds of either group having CVS. Lesions with CVS had lower R2t* (greater tissue damage). However, accounting for MS type, age, and lesion count reduced significance (p = 0.09). Mean R2t* was significantly lower in PMS than RRMS (p = 0.027) and declined with age (p = 0.023).

Conclusions

This study did not find that the presence of CVS could distinguish between patients with RRMS and PMS. Our data suggest that lesions with CVS have more tissue damage. Due to reduced significance after accounting for additional variables, further studies with more patients are needed.

Collapse