Elsevier

NeuroImage

Volume 66, 1 February 2013, Pages 103-109
NeuroImage

Segmentation of magnetization transfer ratio lesions for longitudinal analysis of demyelination and remyelination in multiple sclerosis

https://doi.org/10.1016/j.neuroimage.2012.10.059Get rights and content

Abstract

We demonstrate a new technique to quantify longitudinal changes in magnetization transfer ratio (MTR) magnetic resonance imaging (MRI). These changes are indicative of demyelination and remyelination. This technique comprises a definition of ΔMTR lesions, which are identified directly from the MTR images, and an automatic procedure for segmenting these lesions. We used this technique to analyze MTR changes in lesions of subjects with rapidly progressing multiple sclerosis before and after treatment with immunoablation and autologous stem cell transplant. Subjects who experienced clinical improvement after treatment showed significantly improved MTR recovery in lesions that were recovering during treatment (p < 0.0001) while those who were clinically stable after treatment showed significantly poorer MTR recovery (p = 0.002). The statistical power of this technique to detect treatment effects on MTR recovery was shown to be considerably better than previous methods. These results suggest that longitudinal measurements of MTR in ΔMTR lesions may be an important technique for the assessment of treatment effects on remyelination in clinical trials.

Introduction

Multiple sclerosis (MS) is an inflammatory demyelinating disease of the central nervous system. Although the disease is known to have various diffuse effects on the brain, the hallmark feature of the disease is the formation of focal demyelinating lesions in the white matter (WM). Due to its noninvasive nature and excellent soft tissue contrast, magnetic resonance (MR) imaging is often employed to diagnose and monitor MS, primarily through identification of these lesions.

The formation of WM lesions in MS is associated with a breakdown in the blood–brain barrier (BBB) and the invasion of peripheral immune cells into brain tissue. This stage of lesion formation is characterized by inflammatory demyelination and is associated with enhancement on T1-weighted (T1w) images acquired after administration of a gadolinium (Gd) contrast agent, as well as hyperintense signal on T2-weighted (T2w) images (T2 lesion). Gd enhancement lasts for approximately 1 month (Giovannoni et al., 2000). Remyelination usually occurs, to a variable extent, over several months after lesion formation. Gd enhancement is primarily a measure of BBB integrity, and T2w hyper-intensity is a non-specific, indirect marker of tissue abnormality (Chen et al., 2008).

Closer study of demyelination and remyelination in MS, as well as development of new drugs that modify these processes, requires more myelin-specific imaging information. Although myelin concentration cannot be directly imaged with MR, two imaging techniques have shown particular promise for improving myelin specificity: magnetization transfer ratio (MTR) and myelin water fraction (MWF) (Mackay et al., 1994). MTR imaging is sensitive to changes in myelin density (Pike et al., 2000). MTR relies on the acquisition of a T1- or PD-weighted (PDw) reference scan (MTOFF), and a second scan that differs only by the addition of a magnetization transfer preparation pulse (MTON). An MTR image is calculated by taking the ratio of these two scans. MTR sequences are available on most commercial MR scanners, reconstruction is simple, and the technique can provide full-brain coverage quickly enough to be compatible with clinical trial imaging protocols. For these reasons, this paper focuses on analysis of MTR images. However, the techniques we describe should be easily generalizable to any MR technique with myelin specificity.

MTR information has been analyzed in a variety of ways. Promising results have been obtained by examining the MTR mean, median or histogram in the brain tissue of MS patients (Filippi et al., 2000, Richert et al., 2001, Van Buchem et al., 1999). Other approaches have examined MTR in a voxel-wise manner (Chen et al., 2007, Chen et al., 2008). However, these techniques are generally limited to inspecting MTR changes within T1, T2 or Gd lesions, or within normal appearing tissue. Chen et al., 2007, Chen et al., 2008 for example, examined changes in Gd lesions and thresholded voxels into three groups – those showing MTR increase, decrease or stability. This approach, while decreasing noise, also discards information about the magnitude of the changes.

In this study, we describe a straightforward method for identifying focal areas of MTR decrease (MTR lesions) directly on paired MTR images. We quantify MTR changes in both MTR lesions and Gd lesions, and demonstrate the ability of these methods to assess the extent of demyelination and remyelination in vivo. The proposed methods are compared to previously described techniques, and an example is provided from an ongoing clinical trial.

Section snippets

Patients

Data from the Canadian study of immunoablation and autologous hematopoietic stem cell transplantation (ASCT) for the treatment of multiple sclerosis (Kurtzke, 1983a) was used to test the methods proposed in this study. In this trial, patients with aggressive MS underwent (1) stem cell mobilization using cyclophosphamide and granulocyte-colony stimulating factor (GCSF), (2) stem cell harvest from peripheral blood, (3) immunoablation, then (4) autologous hematopoietic stem cell transplantation.

Correspondence between Gd lesions and ΔMTR lesions

Fig. 3 shows an example slice with Gd lesions indicated in (A) and ΔMTR lesions indicated in (B). On this slice, two large ΔMTR lesions were identified. However, while Gd lesions were identified partially overlapping the ΔMTR lesions, the Gd lesions are mostly adjacent to the region of MTR change. Most of the Gd-enhancing tissue does not show abnormal MTR changes. This result is consistent with the rest of the dataset: 35% ± 12% of Gd-enhancing voxels did not correspond to a ΔMTR lesion voxel,

Discussion

In general, analysis of MTR imaging in MS consists of two steps: (1) identifying a tissue mask of interest, such as T2 or Gd lesions, and (2) calculating the value of a chosen metric within that tissue. In this paper we propose both a novel lesion type (ΔMTR lesions) as well as two novel metrics, both based on semi-quantitative MTR values (MTRRecovered and MTRDrop). Both metrics may be calculated using any lesion type and we have demonstrated their use with Gd and ΔMTR lesions.

Numerous

Conclusions

We demonstrate several new techniques for quantifying MTR recovery in multiple sclerosis lesions using magnetization transfer ratio MR imaging, and also a method for automatically identifying newly formed ΔMTR lesions. As a novel feature, analysis of ΔMTR lesions does not require contrast enhanced scans, or the presence of gadolinium enhancing lesions. ΔMTR lesions may better indicate areas of acute demyelination and are more prevalent than actively enhancing lesions, resulting in fewer

Acknowledgments

The Canadian Study of Immunoablation and Autologous Hematopoetic Stem Cell Transplantation for the Treatment of Multiple Sclerosis is funded by The Multiple Sclerosis Scientific Research Foundation. The current study was funded by the Canadian Multiple Sclerosis Society and the Canadian Institutes of Health Research.

References (28)

  • J. Chen

    Voxel-based analysis of the evolution of magnetization transfer ratio to quantify remyelination and demyelination with histopathological validation in a multiple sclerosis lesion

    NeuroImage

    (2007)
  • M. Deloire-Grassin

    In vivo evaluation of remyelination in rat brain by magnetization transfer imaging

    J. Neurol. Sci.

    (2000)
  • F. Barkhof

    Remyelinated lesions in multiple sclerosis: magnetic resonance image appearance

    Arch. Neurol.

    (2003)
  • R.A. Brown et al.

    Normalization of MTR imaging for multi-centre clinical trials

    International Society for Magnetic Resonance in Medicine White Matter Study Group Meeting

    (2011)
  • R.A. Brown et al.

    Normalization of magnetization transfer ratio Mri for multicentre clinical trials

  • J.T. Chen et al.

    Magnetization transfer ratio evolution with demyelination and remyelination in multiple sclerosis lesions

    Ann. Neurol.

    (2008)
  • D.L. Collins et al.

    Automated 3d nonlinear deformation procedure for determination of gross morphometric variability in human brain

    Proc. SPIE

    (1994)
  • V. Dousset

    Experimental allergic encephalomyelitis and multiple sclerosis: lesion characterization with magnetization transfer imaging

    Radiology

    (1992)
  • A.N. Dula et al.

    Multiexponential T2, magnetization transfer, and quantitative histology in white matter tracts of rat spinal cord

    Magn. Reson. Med.

    (2010)
  • F. Faul et al.

    G* Power 3: a flexible statistical power analysis program for the social, behavioral, and biomedical sciences

    Behav. Res. Methods

    (2007)
  • M. Filippi

    Magnetization transfer imaging to monitor the evolution of MS

    Neurology

    (2000)
  • Francis, S. J. Automatic lesion identification in MRI of multiple sclerosis patients. Thesis, Department of Neurology...
  • M.I. Gaitán

    Evolution of the blood–brain barrier in newly forming multiple sclerosis lesions

    Ann. Neurol.

    (2011)
  • P.J. Gareau et al.

    Magnetization transfer and multicomponent T2 relaxation measurements with histopathologic correlation in an experimental model of MS

    J. Magn. Reson. Imaging

    (2000)
  • Cited by (0)

    1

    Fax: + 1 514 398 2975.

    View full text