White Matter Abnormalities in Patients with Multiple Sclerosis: A Systematic Review and Network-Meta-Analysis of Diffusion Kurtosis Imaging Studies

Abstract

Introduction: Multiple sclerosis (MS) is a chronic demyelinating disease characterised by oedema, inflammation, demyelination, and axonal loss in the central nervous system. Magnetic resonance imaging (MRI) can provide information and track disease progression to improve MS diagnosis and evaluation in clinical practice. However, conventional MRI cannot capture microstructural damage information. Diffusion tensor imaging (DTI) is widely used in clinical practice and research to evaluate microstructural changes in white matter lesions (WMLs) and normal-appearing white matter (NAWM), which provides more sensitive measures for clinically relevant brain abnormalities. However, the diffusion of water molecules in brain tissue follows a non-Gaussian distribution, which DTI does not entirely characterise. Diffusion kurtosis imaging (DKI) is an extension of DTI that can quantify the non-Gaussian diffusion properties of water molecules in tissues and provides more accurate tissue microstructure information than DTI.

Objectives: The primary objective is to assess the primary DKI parameter known as Mean Kurtosis (MK) that has been used to quantify microstructural abnormalities in neuroaxonal pathology in WMLs and NAWM in MS subjects compared with the WM in healthy controls. It was measured as standardised mean difference (SMD), where MK's lower values will equal more advanced axonal pathology.

Materials and methods: Three reviewers conducted the literature search of four electronic databases (Medline, Embase, Scopus, and PubMed) according to the updated PRISMA guidelines. We performed a random-effect network meta-analysis. Pairwise comparison estimates for each category were in tabular format, and rankings represented the probability of each node producing the best outcome. The rankings were presented with mean ranks, 95% confidence intervals, and the surface under the cumulative ranking curve. Analyses were conducted using MetaInsight Software and an open-source web interface to perform analysis in an intuitive ‘point and click’ manner.

Result: We included 6 studies and a total of 239 participants with three comparison groups in our network meta-analysis. 3 studies compared MK in NAWM and WM among MS subjects; 4 studies compared MK in WM among control subjects and patients with MS, 5 studies compared MK in WM and NAWM among control and MS subjects. We identified more reduced MK values in MS WMLs compared to NAWM and healthy WM. Also, less reduced MK values were seen in NAWM compared with MS WMLs. There was a statistically significant difference in standardised mean difference among groups.

Conclusion: This network meta-analysis confirmed that the MK is reduced in MS WMLs and NAWM from WM in healthy participants, which may correspond to axonal loss or myelin sheath damage.