Author Of 1 Presentation
P0569 - Distribution profile of T1 relaxation time in white matter lesions on 7-Tesla MRI in multiple sclerosis reflects disease severity and phenotype (ID 1208)
Abstract
Background
Recent advancements in quantitative neuroimaging have revealed signal heterogeneity in multiple sclerosis (MS) brains. Although most work focuses on central tendency measures (i.e., mean or median), distribution features (i.e., density profile, skewness, kurtosis) of quantitative metrics from magnetic resonance image (MRI) may also provide insightful information about disease severity and progression.
Objectives
We aimed to characterize white matter lesion (WML), normal-appearing white matter (NAWM), and cortical gray matter (GM) in MS brains utilizing distribution features of T1 values from 7-Tesla (7T) MRI to demonstrate their potential as biomarkers of MS disease phenotype and disability.
Methods
Forty-eight participants with MS underwent brain MRI on a 7T scanner. Magnetization prepared 2 rapid gradient echo (MP2RAGE) image was acquired, and quantitative T1 relaxation times were calculated from two inversion images from the acquisition. T1-weighted image reconstructed from MP2RAGE was used for the segmentation of the brain into WML, NAWM, and GM tissue types. T1 values of all participants were concatenated and subgrouped by either disability or disease subtype. T1 distributions in three tissue segments were compared using cumulative distribution function and Two-sample Kolmogorov-Smirnov test (D-statistic). Associations of various T1 features with clinical measures were assessed by Spearman or Pearson methods with controlling for age, as appropriate.
Results
Concatenated T1 distribution of participants’ WML in groups with a higher disability or more progressive MS phenotype appeared wider and shifted toward a higher T1 value. For example, the higher disability group had a higher IQR of T1 (p = 0.040) and a higher median T1 (p = 0.018). The distribution profile of WML was distinctively different between low and high EDSS groups and relapsing versus progressive MS (D = 0.323, p = < 0.001; D = 0.314, p = < 0.001 respectively). Distribution profiles of NAWM and GM were also significantly different between groups, but the magnitude of the difference was smaller (D = 0.058 and D = 0.024, respectively). Despite the difference in the appearance of distribution profiles in WML between groups, skewness and kurtosis were not significantly different. Disability (measured as Expanded Disability Status Scale: EDSS) was significantly correlated with median T1 (rho = 0.405, p = 0.005) and skewness of T1 (rho = -0.301, p = 0.040) in WML, and median T1 (rho = 0.435, p = 0.002) and IQR of T1 (rho = 0.452, p = 0.001) in NAWM. Neither central tendency nor distribution measures in GM significantly correlated with EDSS.
Conclusions
Our study suggests that differences in T1 distribution between groups reflect increasing T1 values of WML as disease advances toward more severe status (i.e., higher disability or progressive MS). This finding supports T1 measures as a potential in vivo biomarker in the diagnosis and prognosis of MS brains.