Normal appearing and diffusely abnormal white matter in patients with multiple sclerosis assessed with quantitative MR

Abstract

Introduction: Magnetic Resonance Imaging is a sensitive technique for detecting white matter (WM) MS lesions, but the relation with clinical disability is low. Because of this, changes in both 'normal appearing white matter' (NAWM) and 'diffusely abnormal white matter' (DAWM) have been of interest in recent years. MR techniques, including quantitative magnetic resonance imaging (qMRI) and quantitative magnetic resonance spectroscopy (qMRS), have been developed in order to detect and quantify such changes. In this study, qMRI and qMRS were used to investigate NAWM and DAWM in typical MS patients and in MS patients with low number of WM lesions. Patient data were compared to 'normal white matter' (NWM) in healthy controls.

Methods: QMRI and qMRS measurements were performed on a 1.5 T Philips MR-scanner. 35 patients with clinically definite MS and 20 healthy controls were included. Twenty of the patients fulfilled the 'Barkhof-Tintoré criteria' for MS, ('MRIpos'), whereas 15 showed radiologically atypical findings with few WM lesions ('MRIneg'). QMRI properties were determined in ROIs of NAWM, DAWM and lesions in the MS groups and of NWM in controls. Descriptive statistical analysis and comparisons were performed. Correlations were calculated between qMRI measurements and (1) clinical parameters and (2) WM metabolite concentrations. Regression analyses were performed with brain parenchyma fraction and MSSS.

Results: NAWM in the MRIneg group was significantly different from NAWM in the MRIpos group and NWM. In addition, R1 and R2 of NAWM in the MRIpos group correlated negatively with EDSS and MSSS. DAWM was significantly different from NWM, but similar in the MS groups. N-acetyl aspartate correlated negatively with R1 and R2 in MRIneg. R2 of DAWM was associated with BPF.

Conclusions: Changes in NAWM and DAWM are independent pathological entities in the disease. The correlation between qMRI and clinical status may shed new light on the clinicoradiological paradox.

Figure 1. Sample qMRI data maps from a 45-years-old female MRI_pos MS patient.

a) R1 map, b) R2 map, c) PD map, d) synthetic T2-weighted image is shown for reference. Several partially liquefied lesions are visible (red arrows in d).

Figure 2. Typical VOI placement in MS patients.

Figure 3. Typical ROI placement in MS patients and healthy controls.

a) DAWM in MRI_neg, b) NAWM in MRI_neg, c) DAWM in MRI_pos, d) NAWM in MRI_pos, e) WM lesion in MRI_pos, f) NWM in a healthy control. MRI_neg subject was a 43 years old female, MRI_pos subject was a 66 years old female and healthy subject was a 42 years old female.

Figure 4. Tissue clusters and in vivo data, in the 3D feature space formed by R1-R2-PD, projected on a,d) R1-R2, b,e) R1-PD and c,f) R2-PD planes.

Figures a-c shows NAWM, DAWM and NAWM, figures d-f shows NWM and WM lesions. Cluster ellipses indicates the calculated covariance's of each tissue type separately, red ellipses correspond to the MRI_neg group, blue ellipses correspond to the MRI_pos group, and the dashed black ellipses correspond to the NWM tissue cluster in healthy controls. The indicated ellipses enclose 95% of the in vivo data.