Validation of retroactively derived T1 relaxation values from 3D T1-weighted images with clinical and MRI measures of disability in multiple sclerosis

TL;DR

This study validates a new method to estimate T1 relaxation values from MRI images and links them to disability measures in multiple sclerosis patients.

Contribution

The novel contribution is the clinical validation of retroactive T1 mapping as a tool for monitoring MS disability progression.

Findings

Lesion T1 values correlated strongly with MTR and disability metrics like EDSS and BPF.

Lesion volumes stratified by T1 thresholds showed significant associations with disease duration and brain atrophy.

NAWM T1 values were significantly associated with EDSS scores.

Abstract

Quantitative T1 mapping is a valuable technique for assessing tissue injury in multiple sclerosis (MS) lesions. We previously introduced a novel methodology for converting high-resolution anatomical 3D T1-weighted (T1W) images into parametric T1 relaxometry maps. Herein, we correlate MS lesion pathology as quantified by retroactive T1 mapping with clinical and MRI metrics of disability and with magnetization transfer ratio (MTR).

38 subjects with relapsing-remitting MS (RRMS) were examined, contributing to 587 unique lesions for analysis. T1 and MTR values were compared using correlation statistics. Univariate correlations between lesional T1 or MTR and Expanded Disability Status Scale (EDSS) were examined using Spearman’s rho (ρ), and for disease duration and brain parenchymal fraction (BPF), Pearson’s r. Mean T1 values of lesions were compared across different categories of EDSS severity using Kruskal-Wallis test. Ordinal regression model was used to assess the association between EDSS and T1 values of select brain regions.

The mean T1 of lesions showed a high correlation with - MTR, r = 0.68. T2 lesion volumes stratified based on different T1 thresholds showed a significant correlation with MS disease metrics: lesion volume threshold at 700 < T1 < 900 was correlated with disease duration (r = 0.34, p = 0.04) and BPF (r = -0.47, p = 0.003); lesion volume threshold at 900 < T1 < 1100 was correlated with EDSS (ρ = 0.42, p = 0.01), disease duration (r = 0.45, p = 0.01), and BPF (r = -0.56, p < 0.001); lesion volume threshold at T1 > 1100 was correlated with EDSS (ρ = 0.41, p = 0.01), disease duration (r = 0.45, p = 0.001), and BPF (r = -0.51, p < 0.001). T1 values at the 25th, 50th, and 75th percentiles significantly correlated with BPF (r = -0.41, p = 0.01; r = -0.41, p = 0.01; r = -0.38, p = 0.02). MTR showed a significant correlation with EDSS but not with disease duration or BPF. Mean T1 values in the NAWM showed a significant association with EDSS (coefficient = 0.03, pseudo R2 = 0.07, p = 0.03).

We provide clinical validation of retroactive T1 mapping as a complementary post-processing technique for monitoring disease activity and disability progression in MS.

Figures