This study used DSC perfusion imaging to measure CBF, CBV, and MTT in NAWM and sub-cortical NAGM of patients with CIS in comparison to those with early RR-MS and to healthy controls. While CBF was decreased in the periventricular NAWM in both CIS and RR-MS patients in comparison to controls, sub-cortical GM CBF was decreased in patients with early RR-MS but not in those with CIS.
Our findings in the early RR-MS group of patients are quite consistent with those of prior MR perfusion studies in patients with the same clinical course [13
]. One of the early studies using DSC MRI demonstrated that acute gadolinium-enhancing lesions in MS patients have higher relative CBV than contra-lateral NAWM [29
]. However, since NAWM may be inappropriate as reference due to pathology of NAWM in its own right, more recent studies employed absolute measurements of CBF, CBV and MTT. The overall perfusion of MS patients in chronic lesions, and several regions of NAWM and NAGM is decreased compared to healthy control subjects [13
A novel finding of our study is the presence of reduced perfusion in the NAWM of patients with CIS, albeit at a lesser degree than in patients with RR-MS. This is not surprising since patients with CIS already show evidence of “hidden” WM damage detected by non-conventional MRI techniques [31
]. For example, magnetization transfer ratios in NAWM of CIS patients have been shown to be decreased [32
] suggesting demyelination and axonal injury. Likewise, mean diffusivity was increased and fractional anisotropy was decreased in the NAWM of CIS patients [34
]. Furthermore, spectroscopy studies in patients with CIS have shown decreases in whole-brain N-acetylaspartate reflecting axonal pathology [35
] and increased myo-inositol in the NAWM, reflecting glial activity [36
The interpretation of NAWM hypoperfusion in MS is not only based on the close vicinity of evolving plaques to deep WM venules, and on the intrinsic vulnerability of WM to hypoxia, but also on several histological and pathophisyological observations. Microvascular damage such as small venous thrombosis, vein wall hyalinization and intravascular fibrin deposition has been observed in MS lesions at histological examination [9
]. Toxic inflammatory mediators might contribute to tissue hypoxia through mitochondrial injury [37
]. Recent studies provide evidence for mithocondrial damage in MS lesions, which could be mediated by reactive oxygen and nitric oxide (NO) species [37
]. In addition to NO and cytokine mediated changes in cerebral perfusion, altered glial cell homeostasis may contribute to vascular changes in MS [40
]. Recently, it was reported that a rise of calcium levels within astrocytes induces constriction of blood vessels and consequently reduces CBF [41
]. Finally, recent studies on gene expression in the NAWM of patients with MS showed that mRNAs of multiple genes involved in hypoxic preconditioning are significantly unregulated in comparison with control WM [42
]. This notion is also supported by the observation that one of the four patterns of pathology described in MS lesions, distal dying-back oligodendrogliopathy, is the same pattern found in acute WM stroke, likely mediated in part by hypoxia inducible factor (HIF 1 alpha) [43
]. Interestingly, this particular type of lesion has also been described in a study on early stages of MS lesions [45
Unlike in patients with RR-MS, tissue perfusion was not decreased in the sub-cortical GM of patients with CIS suggesting a continuum of brain perfusion decreases in MS beginning in periventricular white matter, and spreading to other WM as well as GM regions during disease progression. Several direct and indirect mechanisms might have contributed to the deep GM hemodynamic impairment in RR-MS patients. First of all, MS lesions are frequently found in the thalamus and basal ganglia at post-portem examination. However, lesions are often missed on T2-W images due to their small size and poor contrast with the surrounding GM. In addition, diminished blood supply has been described in MS plaques [29
]. Although macroscopic deep GM lesions were not detected in our MS patients, theoretically, small deep GM lesions, not detected on MRI, could have contributed to the decrease of perfusion. Second, axonal transection in active WM lesions could lead indirectly to anterograde and retrograde degeneration of axons running within the thalamus and basal ganglia. Finally, axonal loss itself secondary to hypoperfusion, demyelination or Wallerian degeneration might contribute to CBF decrease via reduction in local metabolic activity. All these mechanisms, already active since the earliest stages of the disease, may lead to GM hypoperfusion as the disease progresses. Furthermore, the secondary nature of GM hypoperfusion seems to be supported by the association between WM and GM hypoperfusion in our patient's group.
To our knowledge, this is the first study to suggest heretofore that perfusion in the NAWM of CIS patients is decreased compared to healthy controls and that sub-cortical GM might be affected as the disease progresses secondarily to WM injury. There are several caveats to our study. First of all our sample size was small and, although we suggest that GM hypoperfusion develops as disease progresses, caution must be exercised before drawing firm conclusions since the patients with RR-MS were not the same ones who had CIS some years earlier. Second, our study is cross-sectional and we are essentially examining snapshots in time, which may have overlooked the dynamics of tissue perfusion changes. For instance, a previous longitudinal perfusion MRI study examining RR-MS patients found that increases in NAWM perfusion could be detected as much as 3 weeks before the development of a subsequent lesion [14
], however these increases were over the background of overall decreased perfusion. Finally, the relative low resolution of DSC MRI and the high density of vessels in the cortical GM precluded us from measuring tissue perfusion in the cortex by means of region of interest analysis.
Examining perfusion patterns in patients with CIS and following them over time will help confirm our findings regarding the spread of perfusion decreases if and when those patients convert to clinically definite MS.