Iron stores and Cerebral Veins in MS Studied by Susceptibility Weighted Imaging (SWI)
E. Mark Haackea, James Garberna, Yanwei Miaob, Charbel Habiba, Manju Liua.
aDepartment of Radiology, Wayne State University, Detroit, MI 48202, USA
bDepartment of Radiology, the First Affiliated Hospital, Dalian Medical University, Dalian,
Liaoning 116011, China
E. Mark Haacke, PhD
Wayne State University
MR Research Facility
Department of Radiology
HUH—MR Research G030/Radiology
3990 John R Road
Detroit, MI 48201
Tel. (313) 745-1395
Fax (313) 745-9182
Aim: Multiple sclerosis (MS) is a disease whose etiology until recently has remained a mystery.
A possible explanation for MS has been put forward by Zamboni et al 1 that it is caused by a
chronic cerebrospinal venous insufficiency (CCSVI). In this paper, we show that the iron
deposition as seen by susceptibility weighted imaging (SWI) in the basal ganglia and thalamus is
consistent with this interpretation.
Methods: 14 MS patients were recruited for this study with a mean age of 38 ranging from 19
to 66 years old. A velocity compensated 3D gradient echo sequence was used to generate
susceptibility weighted images (SWI) with a high sensitivity to iron content. We evaluated iron
in the following structures: substantia nigra, red nucleus, globus pallidus, putamen, caudate
nucleus, thalamus and pulvinar thalamus. Each structure was broken into two parts, a high iron
content region and a low iron content region. The measured values were compared to previously
established baseline iron content in these structures as a function of age.
Results: Of the 14 cases, 12 revealed an increase in iron above normal levels and with a
particular pattern of iron deposition in the medial venous drainage system.
Conclusion: The backward iron accumulation pattern seen in the basal ganglia, thalamus and
midbrain of most MS patients is consistent with the hypothesis of venous hypertension.
Key words: brain iron deposition, susceptibility weighted imaging, iron changes in multiple
It has been suggested recently that multiple sclerosis (MS) may develop as a neurodegenerative
disease because of a venous vascular insufficiency 1 and the resulting venous hypertension
related directly to abnormalities (narrowing and restricted flow) in the jugular and azygous veins.
Although evidence of venous involvement has been known for a long time 2-4, pioneering efforts
in this direction have basically been ignored until Zamboni et al 1 demonstrated the presence of
stenoses in MS patients. This is truly a paradigm shift, where one must look outside the brain, in
this case the cardiovascular system, to understand the source of the brain’s problem in MS.
Evidently, the venous hypertension that leads to multiple sclerosis can have a damaging effect on
the vessel wall 5,6. This is best understood from research related to chronic venous disease where
it has been shown that changes in venous hemodynamics leads to a disruption of flow, a
reduction of the shear stress in the vessel wall, an inflammatory degenerative response and,
subsequently, a deterioration of the venous wall 7-9. Similar effects are known in atherosclerosis
when normal flow is disrupted. And even in multiple sclerosis research dating back to the work
of Adams 2, and later Trojano 10, venous wall breakdown in multiple sclerosis has been well
Following this argument a bit further, one might expect then local iron increases from the
damaged vessel wall, loss of vessel function, reduced transit times and reduced local blood flow
in the chronic stages of the disease. One might also expect that the vascular pathways most
closely associated with the hypertension and local reflux (especially in the areas of the internal
cerebral veins and vein of Galen) associated with obstructed venous flow would in turn lead to
local venous damage in multiple sclerosis 11-13. The areas most affected by this might well be the
thalamus and basal ganglia since they are associated with the medial venous drainage system 14.
With this philosophy in mind, we investigated the local iron content in the basal ganglia and
thalamus using susceptibility weighted imaging (SWI) to see if there were any evidence of such
microbleeding or vascular breakdown that followed a pathway reverse to that of the draining
veins as noted by Fog in 1964 3 and later espoused by Schelling 4.
SWI is a gradient echo MR imaging technique that uses phase as an alternate source of
information to enhance contrast 15. It uses the susceptibility differences between tissues to
generate this contrast. Susceptibility is very different than the usual spin density, T1 and T2
properties. It is represented indirectly by the phase images and in this case by the SWI filtered
phase images that are used to remove background field inhomogeneities. These phase images
produce the same image, whether acquired at 1.5T, 3T or any other field strength as long as the
product of BoTE is a constant. Thus, while other tissue properties change from field strength to
field strength, phase properly adjusted for TE will not. That is because susceptibility does not
change from one field strength to another. The early applications of SWI stemmed from the
signal changes that took place in the SWI images for veins 16. The presence of deoxygenated
blood leads to both a T2* effect and a change in phase in the veins since deoxyhemoglobin is
paramagnetic. Combining these two characteristics into SWI data led to a beautiful perspective
of the venous system in the brain. Similar to how the blood oxygenation level dependent
(BOLD) effect is used in functional brain imaging, SWI shows veins with less deoxyhemoglobin
as less suppressed and those with more deoxyhemoglobin as more suppressed. Several examples
of the venograms that are possible with SWI at 3T are shown in Figure 1. In this paper, our
interest is to combine the venous vascular information in SWI processed data with the putative
iron content seen in SWI high pass filtered phase data. These two pieces of information can then
be used to assess the state of the venous vasculature in multiple sclerosis patients. Some example
SWI filtered phase images good for visualizing iron are shown in Figure 2.
To this end, we compared the iron in a series of 14 multiple sclerosis patients with that in a set of
100 normal subjects using SWI. Our goal was to look for iron abnormalities and evaluate
whether these could be integrated into this new theory of cerebral spinal venous insufficiency.
Materials and Methods
Fourteen (14) clinically defined MS patients were imaged with mean age of 38 ranging from 19
to 66 years old. All patients signed a consent form approved by the Internal Review Board (IRB-
approved). A velocity compensated 3D gradient echo sequence was used to generate
susceptibility weighted images (SWI) with a high sensitivity to iron content. All SWI images
were acquired at 1.5T Sonata [Siemens, Erlangen, Germany] with a resolution of 0.5x0.5x2mm3.
Imaging parameters were TR=57ms, TE=40ms; FA=20°; and BW=80 Hz/pixel. A 64x64 low
spatial frequency kernel was used to complex-divide into the original data to create an effective
high pass filtered phase image. The resulting SWI filtered phase images were used as a means to
quantify iron content in vivo. The following 7 deep gray matter structures were studied for iron
content including: the globus pallidus (GP), the caudate nucleus (CN), the putamen (PUT), the
thalamus (THA), the substantia nigra (SN), the red nucleus (RN) and the pulvinar thalamus (PT).
A two region analysis concept was used where each structure was separated into two regions-of-
interest (ROI): the normal iron content region (RI) and the high iron content region (RII). These
regions were differentiated using threshold values cited in previous work by Haacke et al. in
2007 17. The measured values were compared to a previously established baseline of iron content