Correlation of Multiple Sclerosis Measures Derived from T2-Weighted, T1-Weighted, Magnetization Transfer, and Diffusion Tensor MR Imaging

Multiple sclerosis (MS) is an immune-mediated disorder selectively affecting the central nervous system. The pathologic hallmark of MS is inflammatory demyelination, which can be limited by reparative mechanisms (including remyelination) or can become irreversible and ultimately lead to tissue loss (1). In MS, the progressive accumulation of tissue damage is likely to be one of the major factors contributing to disabling neurologic deficits (2). As a consequence, achieving an accurate in vivo assessment of pathologic signs of MS might be a rewarding exercise for increasing our understanding of the mechanisms leading to irreversible disability and our ability to monitor the efficacy of experimental treatments. In this context, measures derived from conventional MR imaging, such as the load of T2 hyperintense lesions, have several limitations, including the lack of specificity to the heterogeneous pathologic substrates of MS lesions (3, 4) and the inability to detect subtle abnormalities in the normal-appearing white matter (NAWM) (5–8). This can explain why, in patients with MS, the correlation between clinical and conventional MR findings is, at best, moderate (4). Magnetization transfer (MT) and diffusion tensor (DT) MR imaging, may, however, go some way toward overcoming these limitations.

MT is based on the interactions between protons in a relatively free environment and those wherein motion is restricted. In the brain, these two states correspond to the protons in tissue water, and in the macromolecules of myelin and cell membranes. Off-resonance irradiation is applied that saturates the magnetization of the less mobile protons and is then transferred to the mobile protons, thus reducing the observable signal intensity. Low MT ratio (MTR) indicates a reduced capacity of the macromolecules in brain tissue to exchange magnetization with the surrounding water molecules, reflecting damage to myelin or to the axonal membrane (9). A postmortem study found a correlation between low MTR and the percentage of residual axons in MS lesions (10). Animal studies have also shown that low MTR correlates with histopathologic findings of myelin loss and axonal destruction (11), whereas edematous lesions result in slightly increased MTR values (12). Dramatically reduced MTR is also seen in the “pure” demyelinating lesions of patients with progressive multifocal leukoencephalopathy (13) or central pontine myelinolysis (14).

Diffusion is the microscopic random translational motion of molecules, and water molecular diffusion can be measured in vivo by use of diffusion-weighted MR imaging (15). Because diffusion is affected by the properties of the medium wherein molecular motion occurs (16), the measurement of diffusion inside biological tissues provides information about tissue structure at a microscopic level (17). The motion of water molecules can be hindered by the presence of structural barriers at a cellular or subcellular level. Pathologic processes that alter tissue organization by decreasing or increasing the number of barriers to water molecular motion or that alter the permeability of the barriers cause abnormal water diffusivity. In addition, diffusion is inherently a three-dimensional process, and in some tissues with an oriented microstructure, such as brain white matter, the molecular mobility is not the same in all directions. This property is called anisotropy, and results in a variation in the measured diffusivity with tissue measurement direction (18, 19). White matter fiber tracts consist of aligned myelinated axons and, therefore, hindrance of water diffusion is much greater across rather than along the major axis of axonal fibers (20). Under these conditions, a full characterization of diffusion can only be found in terms of a tensor (21), a 3 × 3 matrix where the on-diagonal elements represent the diffusion coefficients along the axes of the reference frame, whereas the off-diagonal elements account for the correlations between molecular displacement along orthogonal directions. From the tensor, it is possible to derive some scalar indices, invariant to the changes in the frame of reference, which reflect the diffusion characteristics of the tissue. These measures include 1) the mean diffusivity (D̄) (equal to one third of the trace of the diffusion tensor), which is a measure of the average molecular motion independent of any tissue directionality and is affected by cellular size and integrity (22, 23); and 2) the fractional anisotropy (FA), which is one of the most commonly used measures of deviation from isotropy (22) and reflects the degree of alignment of cellular structures within fiber tracts, as well as their structural integrity. The pathologic elements of MS have the potential to alter the permeability or geometry of structural barriers to water molecular diffusion in the brain and, consistent with this, water diffusivity is higher and FA lower in MS lesions than in NAWM and in white matter of healthy volunteers (24–26).

MT and DT characteristics can be analyzed on a region of interest (ROI) basis or on a more global basis by using histogram analysis (8, 27–30), an approach that allows evaluation of all the brain tissue, thus providing an assessment of both macroscopic and microscopic disease burden in MS. Because the relative contributions of conventional, MT, and DT MR imaging to study MS in vivo have never been investigated, aims of this study were to investigate the magnitude of the correlations between various MT and DT MR imaging–derived metrics and between these quantities and measures derived from conventional MR imaging.

Methods

Results

In Table 1, the values of the different conventional, MT, and DT MR quantities studied are reported for the whole patient sample. No T2-hyperintense or T1-hypointense lesions were found on the scans of healthy control subjects. Average D̄ was significantly lower (0.90 × 10⁻³ mm² s⁻¹, P = .0001) and average MTR significantly higher (40.3%, P = 0.0001) in control subjects than in patients. All the other MTR and D̄ histogram–derived metrics also significantly differed between control subjects and MS patients (P values ranging from .003 to .0001).

No significant correlation was found between any of the MTR and FA histogram–derived metrics and conventional MR measures (T2 and T1 LV and brain volume). Modest, but statistically insignificant correlations were found between T2 and T1 LV and the peak heights of the MTR histograms (r values ranging from −0.31 to −0.45). All the BT and NABT D̄ histogram–derived metrics were significantly correlated with T2 and T1 LV (Table 2), whereas they were not significantly correlated with brain volume.

No significant correlations were found among the corresponding quantities of the MTR, D̄, and FA histograms. Modest, but statistically insignificant correlations were found among the average BT D̄ and FA (r = −0.39), the peak positions of the MTR histograms, and the average D̄ values from both BT and NABT (r values ranging from −0.34 to −0.43) and among the average FA and the peak heights of BT or NABT D̄ histograms (r = 0.32). Average lesion MTR, D̄, and FA were neither significantly correlated nor showed a trend toward significance.

None of the MR quantities we measured were significantly correlated with patient EDSS score. EDSS was modestly, but not significantly, correlated only with T1 LV (r = 0.36).

Discussion

Although the assessment of lesion burden from T2-weighted MR scans is widely used as a surrogate marker of disease evolution in MS studies (4), abnormalities seen on T2-weighted images do not provide specific information regarding the heterogeneous pathologic substrates of MS lesions (3, 4), which can range from inflammation and vasogenic edema to irreversible demyelination and axonal loss. Another major limitation of T2-weighted MR imaging is its inability to detect and quantify the extent and severity of the microscopic damage known to occur in NAWM (5–8). The nature and extent of lesions and NAWM pathologic abnormalities are likely to influence MS manifestations and evolution (8, 37, 38). For instance, evidence of tissue loss (1) and widespread NAWM changes (8) can be found in MS patients with severe and irreversible neurologic disability, even when T2 and enhancing lesion loads are relatively low. For all these reasons, quantitative MR techniques, with the potential to provide more pathologically specific and accurate information about MS, have recently been introduced in the assessment of MS patients (2–4). MT and DT MR imaging are two of the most promising techniques for at least two reasons. First, they provide quantitative and objective measures. Second, they allow a large brain coverage, which is important in multifocal and widespread diseases, such as MS. Previous MS studies used MT or DT MR imaging in isolation and, therefore, did not have the opportunity to investigate the correlation among quantities derived from these techniques. Because different MR techniques necessarily provide overlapped information, defining the magnitude of such a correlation is an important prerequisite to avoid the acquisition of redundant MR data.

In this study, we found that quantities derived from MTR histogram analysis of the BT and NABT are not significantly correlated with the corresponding quantities from D̄ and FA histograms. This confirms and extends the results of a previous study that used diffusion-weighted MR imaging and showed a lack of correlation between average MTR and D̄ of BT (27). We also did not find significant correlations between lesion MTR, D̄, and FA. These observations suggest that MT and DT MR imaging provide relatively independent measures of MS pathologic abnormalities and that their combined use might result in a gain of relevant information leading to an understanding of the mechanisms underlying the clinical manifestations of the disease. Although correlative studies with histopathology are needed, we believe that the lack of correlation between MTR and DT MR metrics in the brain tissue is the result of the complex relationship between destructive (inflammation, demyelination, and axonal loss) and reparative (remyelination and gliosis) mechanisms occurring within and outside T2-visible lesions and their variable effects on MTR, D̄, and FA values. That MTR and DT MR metrics provide complementary and partially independent information is supported by another finding that DT histogram metrics are strongly correlated with T2 and T1 LV, whereas quantities derived from MTR histograms are modestly, but not significantly, correlated.

We did not find any correlation between brain volume and quantities derived from MT and DT MR imaging. The measurement of brain volume has recently been suggested as a marker of MS severity with the potential to monitor the disease evolution accurately (39). Several studies found that MR imaging measures of brain atrophy were correlated with the level of disability or the course of the disease (40–44) and found moderate correlation between brain volume and MTR histogram–derived metrics (45–47). The discrepancy between our results and those of previous studies (45–47) might be due to the clinical characteristics of the patients we studied, who were mildly disabled by RRMS. Although brain volumes vary markedly in healthy individuals and, as a consequence, a normalization of brain volumes would have been desirable (43), this observation suggests that brain atrophy measurement might not be sensitive to the most early and subtle aspects of the MS pathologic abnormalities. It is indeed conceivable that brain atrophy in MS is a late phenomenon, which follows the appearance of other evidences of tissue loss and disorganization, detectable by MT and DT MR imaging. This is consistent with recent MR spectroscopy findings, indicating that the presence of axonal damage can be detected in RRMS patients in the absence of concomitant decreases of brain volume (48).

Conclusion

Although preliminary and warranting further investigation, our results suggest that the application of different MR techniques with variable sensitivity to the heterogeneous pathologic aspects of MS might contribute to a better understanding of MS pathophysiology and call for a multiparametric MR approach in the study of patients with MS.