J.
Kevin
De Marco,
MD
#{149} William
Dural Arterlovenous Evaluation with The preangiographic diagnosis of cerebral dural arteriovenous fistula (DAVF) can be difficult. The magnetic resonance (MR) images of 12 patients with angiographically proved DAVF were evaluated to characterize
the
appearance
of these
lesions and to identify those patients at increased risk for complications. Patients with DAVF demonstrating venous occlusive disease are at higher risk for complications from the arterialized collateral yenous system. This venous occlusive disease is demonstrated best at anteniography.
The
MR
imaging
appear-
ance of dilated cortical veins without a parenchymal nidus is suggestive of a DAVF with veno-occlusive disease. Eight of the 12 patients in our series demonstrated this finding at angiography. Complications, including infarction and hemorrhage, were identified at MR imaging in eight patients with MR imaging evidence of veno-occiusive disease. At angiography 42% of these complications were not apparent. In one patient with a DAVF draining into an unobstructed
right
sigmoid
sinus,
results of MR study were normal. Although patients with DAVF without veno-occlusive disease may have normal findings at MR imaging, DAVF associated with veno-occlusive disease and dilated pial yenous drainage can be documented on MR images. This subset of DAVF patients, many of whom were identified only at MR imaging, is at higher risk for complications due to veno-occlusive disease. These patients are believed to require more urgent therapy. MR imaging is useful in the pretherapeutic planning for patients with DAVF. Index terms: Arteniovenous malformations, cranial, 17.75 #{149}Fistula, arteriovenous #{149} Fistula, dural, 17.75 #{149}Sinuses, dural, 17.75 Radiology
1990; 175:193-199
P. Dillon,
MD
V. Halbach,
MD
arteniovenous fistulas (DAVFs) represent 10%-15% of all intracranial vascular malformations (1). The majority of these involve the venous sinuses along the base of the brain. The clinical symptoms of DAVFs involving the sagittal, lateral, sigmoid, and straight sinuses are often related to the route of URAL
venous
drainage
(2). Unfortunately,
the initial complaints are frequently nonspecific. The preangiographic diagnosis of a DAVF is difficult and mequires a high degree of clinical suspicion. This differs from diagnosis in patients with DAVFs involving the cavernous sinus, who often present with classic symptoms such as proptosis, chemosis, and bruit. These were therefore excluded from this study. The magnetic resonance (MR) imaging appearance of DAVF has not been covered in detail previously. Our goal was to evaluate the MR imaging appearance of angiographically proved sagittal, lateral, sigmoid, and straight sinus DAVFs and attempt to identify features that place patients at high risk for development of complications.
PATIENTS
AND
METHODS
Twelve patients with angiographically proved DAVFs of the sagittal, lateral, sigmoid, or straight sinus were evaluated with MR imaging. There were nine men and three women ranging in age from 35 to 70 years, with a mean age of 52 years. Patient information is summarized in the Table. Angiograms obtained in two orthogo-
1
From
the
Neuroradiology ogy, University Parnassus
#{149} Jay
S. Tsuruda,
MD
Fistulas: Imaging’
MR
D
#{149} Van
Ave.
Diagnostic
and
Interventional
nab planes Spin-echo
were available (SE) MR images
on a i.5-T
unit
(Signa;
for all patients. were obtained
GE Medical
Sys-
tcms, Milwaukee) in all but one patient. In the remaining patient, MR images were obtained on a 0.5-T unit (Diasonics, South San Francisco). Images were obtamed in two orthogonal planes in all patients. In eight cases images in all three orthogonal planes were provided. Tiweighted images were obtained with a repetition time (TR) of 600-800 msec and
an echo
time (TE) of 20-30
msec (TR/TE
600-800/20-30). Proton-density were obtained at 2,000-2,800/20-35. weighted images were acquired
images
T2at 2,0002,800/70. The matrix size was i92 X 256 in four patients and 256 X 256 in eight patients. Section thickness was 5 mm. In selected patients, flow-sensitive gradient recalled echo (GRE) images (150/15; flip angle, 50#{176}) were obtained. In patient 8, axial cine MR imaging (21/9; flip angle, 30#{176}) was performed. The appearance of the arterial supply,
nidus,
venous
drainage,
patency
of the
involved sinus, and associated parenchymal abnormalities were correlated with the angiographic findings (Table). The MR images and angiograms were obtamed within 7 days in all but two cases. In these two cases there was no significant change in the patient’s clinical status between the two examinations. Three patients also underwent contrast materialenhanced computed tomography (CT) (GE 9800; GE Medical Systems) within i week of MR imaging and angiography.
RESULTS As outlined in the Table, the arterial supply to the DAVF was not demonstrated on MR images, except for a possible tentonial arterial feeder seen in patient 2 (Fig 1). The site of the anteniovenous (AV) fistula was always demonstrated at angiognaphy but never seen at MR imaging. No panen-
Section, Department of Radio!of California San Francisco, 505 San
Francisco,
CA
94143.
From
the 1989 RSNA
annual meeting. Received August 29, 1989; revision requested October 26; revision received November 29; accepted Dccember 11. Address reprint requests to J.K.D. © RSNA, 1990
Abbreviations: dural
called =
AV
arteniovenous
echo,
repetition
SE
arteriovenous, fistula, GRE
spin echo,
TE
DAVF gradient reecho time, TR
time.
193
chymab nidus was demonstrated at MR imaging. Abnormal dilated draining veins were identified at MR imaging and angiognaphy in eight of the 12 paticnts (Table). One additional patient demonstrated an occluded sinus without dilated cortical veins. This correlated with angiographicabby proved vcno-occlusive disease. Seven of these nine patients (78%) demonstratcd complications from the DAVF. In one patient with no abnommal draining veins identified on MR images, an unexplained subdurab he-
: a
Ca
6 05
was
correctly
identified
2
.-.
h
6
-E
a .c
.
t-
a
>-
Ca
)
a-
5SC
00.
GJ
Ca U
U
Z(n
>
0
-a
a
a
0.
Ca
a
-g
a a)
-.
a a 0 z
a a
-u .-. x
Zu
a) >.
a) >.
Ca
Ca
Ca
6
6
66
a
.
a)
;s.
bO Ca
6
a 0
05 U
arterial
the
major
sinus
194
#{149} Radiology
network may
in the
0
0
0
z
zz
(IC
a
0.
0
-0 a) -0
-
0 -
-
a)
Ca
bO 0
U
U
0
Ca
6
6
0
0
-
0.
Ca
-
QJ
.
;.
cn
j
Ca
Ca
c0
c a)
a
a
a
a U) a
(IC
(IC a)
a
a
0
0
0
0
0
0
Ca
a) 0
0
U U
0
0
0
0
aj Cl) a a
ZZZZZ
a 5) a
a
a) a) a 0
z
000
Z>
>.ZZ>.>.Z
Ca
0
0
0
0
0
0
ZZZZZZ
ZZ
0 a)
-
Ca
Ca
66 U U
0
0
0
ZZ
-
-
-
Ca
Ca
Ca
EE6E
U U
‘
0
Ca
.6 0
OZ
U
‘
U
0
0
OZZZOZ
0 a)
:
ii
6, 0.
0.
a
6
a
a
.
a)
a)
;.
a)
a)
Z
o.
o.
Z
o.
a)
00
a
-0 a)
-
0
a bO Ca
Ca
ZZ
0.
-
Ca
a a
0
a
0
0.,
(1)
Ca
a
0
a
a) 0.
0
a)
b0
a
a)
0.
(I)
-:;i
ILl
.bo 1)
a
2aaa
I-
a)
(I)
u(IC
0_
a
u(IC
a)
a)
;..
o.
o.
Z
;.
Ca
ZO
a 0.
a)
a)
.
.
g
Z
(IC
(IC
(I)
a
a
a
-
a)
.
a
._
i.
.-.
bO
bO
(IC
(IC
aCa
aCa
-g C
OJ
r: a) ,._
: C)
(3
5)
(1)
0 0.
.0-u
-
‘
0
a
>a
,
a
so
.,
g
a) U Ca
a
4-
a
-0
-.
.
Ca
Ca
ILl
CU
CU
(IC
a
a
-;;i
-
0
.
ba’...
n.
bo
Ca
(IC
a
a 0
Ca _
-j
..
I..
-:ao
a
_ ‘
-g6
a
n
a
Ca
05
.-
‘.
U)
-.
a a Ca a 5j
U
.
bO
deep
walls
Ca
0
z
a
attempt
a
ZZ>n.ZZ
(I>
U
0
DISCUSSION
dunab
a
a
bC
a a
0>
(5), the
SJ
:-.
a)
a
venous system (6,7), and the cavernous sinus (8). Several authors have postulated that the development of DAVF may begin with thrombosis or occlusion of the involved sinus (2,9). The rich
5
6
a
a
of durab arteries and a vein or venous sinus (3). Many of the DAVFs involve the venous sinuses along the base of the brain, and here the transverse sinus is the most commonly affcctcd (2). DAVFs can also involve veins in the anterior cranial fossa (4),
a
:a
a
bO
‘.-
sinus
-
sJ
aJ
6 a’
-
bO 0
.
sagittab
OO
.
aJ
6
In one case, results of MR imaging were falsely negative. No dilated contical veins, sinus occlusion, on complications were seen (Fig 4). The angiogram demonstrated a right sigmoid DAVF without veno-occlusive disease.
superior
.ua n
LI
(1)
ab-
Ca
--
0.
dunal
DAVFs are infrequent vascular normalities consisting of numerous tiny connections between branches
2Ca 6
Ca
.
-
u_u a o.
0
in sev-
involved
-u
.,
a
en patients. Two patients demonstrated occlusion of the involved sinus at both MR imaging and angiography. Partial on complete sinus thrombosis was underestimated on routine SE images in three patients. When used, GRE images were helpful in evaluating flow in draining
veins as well as in the sinuses (Fig 3).
ca)
(IC
matoma was noted (Fig 2). Overall, 12 complications were seen in eight patients including infarction (n 3), intraparcnchymal hematoma (n = 3), subdumal hcmatoma (n = 2), diffuse hemosidenin deposit possibly secondary to venous infarct (n 1), vasogenic edema (n = 1), hydmocephabus (n 1), and proptosis (n 1) (Table). Five of the 12 complications (42%) were not readily apparent at angiography. The patency of the involved dural
sinus
n.
E
bo
.9
a)
b0
: Lfl ‘
Ca’ ‘
‘
m
n
(0
0
r
00
0
Lt)
0’
L.
.
0 N.
C’4 LO
C1
Z
to necanApril
1990
b. Figure 1. DAVF involving lated vessel coursing along
dural
arterial
supply
can
left transverse the left tentorium be seen
against
the
(a)
sinus.
(arrows). tentorium
Axial
MR
image
(2,000/35)
The signal
void
but
be impossible
would
of this
reveals
a di-
presumed to separate artery
from cortical bone on standard SE images. (b) Lateral view of the left internal carotid after injection of contrast material confirms the presence of a dilated tentomial branch arrows) supplying the fistula located in the left transverse sinus (open arrows).
(solid
synchronous
tinnitus.
unobstructed
venous
In DAVFs
drainage,
with
corti-
cal venous pathways usually do not develop. Hemorrhagic complications with these DAVFs are infrequent (5). Patients with carotid cavernous DAVFs, unlike those with DAVFs seen in other locations, often present with classic clinical findings of proptosis, chemosis, and bruit. They rarely demonstrate complications such as stroke or hemorrhage. Such patients were excluded from this study. The most severe symptoms associated with DAVF are due to cortical venous drainage secondary to venoocclusive disease (10). In a review of DAVF by Castaigne et al (11), 42% of DAVFs draining into cortical veins were associated with hemorrhage. However, Houser et al (2) point out that the sinus need not even be thrombosed to cause retrograde yenous drainage into cortical veins.
Partial
filling
of sinuses
by the
high-
pressure arterial blood can impede antegrade venous drainage. Because of their location away from major dumal sinuses, DAVFs involving the anterior cranial fossa and tentonium almost always drain into intradural veins and are associated with a high frequency of hemorrhage (12). In addition to hemorrhage, retrograde yenous flow accompanied by increased venous pressure can lead to chronic passive congestion and venous infarcts. This is a major cause of focal neurologic deficits. Hydrocephalus is another complication of DAVF. This can be related to decreased absorption of cerebrospinal fluid due to elevated venous pressure or mechanical compression of the aqueduct by mesencephalic veins. Proper
important
b. Figure
(a)
2.
Axial
sagittal
tion
DAVF MR
of superior
image
Ti-weighted
disorder,
of contrast terial
sagittal
(2,000/70)
sinus
seen
demonstrates
The
hematoma. sinus
is patent
on
(not shown). Since there was no history of trauma or coagulawas suspected. (b) Image of left common carotid artery after injection in the late arterial phase reveals a DAVF with left middle meningeal amarrows) entering in the inferior/lateral sinus wall (solid arrow). images
material
supply
(open
seen
at angiogmaphy.
The
volved dunal sinus receives arterialized blood flow that can lead to mechanical obstruction of the sinus and result in retrograde drainage of blood away from the sinus and
Volume
subdural
hematoma.
a DAVF
alize the sinus by developing direct artery-to-sinus communications. This may lead to the angiomatous network of multiple feeding arteries and numerous AV shunts within a pantially mecanalized sinus that is fre-
quently
as a spontaneous
a subdural
175
#{149} Number
1
in-
into
cortical
veins.
A better understanding of the natumal course of DAVF has bed to an improvement in the diagnosis and treatment of this disease. Bruit and headache arc the common symptoms of a DAVF and are usually unrelated to cortical venous drainage. When the DAVF involves a dural sinus in direct contact with the petrous pyramid, the resulting higher turbulent flow is often manifested as pulse-
MR
imaging
when
technique
assessing
is
patients
with DAVF, as both parenchymal abnormalities and venous thrombosis must be assessed. Acquisition of Tiand T2-weighted images is mandatory. At least two orthogonal planes should be used. If flow compensation techniques are used with T2-weighted images, refocusing of laminar flow will result in increased signal
intensity
and
may
obscure
or simu-
late thrombosis. This technique may also reduce the conspicuity of dilated cortical veins. Saturation bands are helpful on both Ti- and T2-weighted images to reduce intraluminal signal intensity caused by slow flow or the entry-section phenomenon (13). GRE images can be helpful in evaluating the vascular anatomy of DAVF and the patency of the durab sinuses.
Radiology
#{149} 195
Figure 3. DAVF of left sigmoid sinus with occlusion of left transverse sinus. (a) Lateral view after selective injection of contrast material in the left occipital artery demonstrates
a DAVF
involving
the
patent
but
nar-
rowed left sigmoid sinus (open arrows). Proximally the left transverse sinus is occluded (solid arrow). (b) Axial MR image (2,800/80) ty in the
suggests left jugular
increased bulb but
signal is not
intensidiagnos-
tic of partial occlusion. Occlusion of the left transverse sinus is not seen. (c) Coronal GRE MR image (150/15; flip angle, 50#{176}) demonstrates an occluded distal left transverse sinus.
Irregular
narrowing
of the
left
sigmoid
sinus (not shown) secondary to the DAVF was also noted. (d) Coronal GRE MR image (150/15; flip angle, 50#{176}) obtained posterior to c reveals
that
left transverse
the
sinus
proximal
portion
of the
is patent.
a.
b.
C.
d.
Although the actual fistula beartery and vein was depicted in all patients at angiography, it was never seen directly with MR imaging. This is probably related to the small size of the fistula as well as the lack of contrast between the signal void in dural vessels and the lack of signal from cortical bone. Indeed, a patient with a DAVF without venoocclusive disease or dilated cortical veins may have completely normal findings at MR examination (Fig 4). Thus, we do not believe MR imaging tween
is suitable
as the
sole
screening
cx-
amination for DAVF. Dilated cortical veins in patients with DAVF arc usually, but not a!ways, associated with veno-occlusive disease. Angiography had demonstrated that veno-occlusive disease can occur at any point in the draining venous system. Thus, the nanrowed vessel may be at the junction of the AV fistula and the dural sinus, within the dural sinus, or in a draining vein away from the fistula site (14). MR imaging was able to document all cases of dilated cortical veins associated with veno-occlusive disease seen at angiognaphy. Further, the MR imaging appearance of dilated cortical veins in the absence of a panenchymal AV malformation nidus suggests the presence of a DAVF. A parenchymal AV fistula, pial AV ma!formation, or anomalous cortical yenous drainage could have a similar appearance at MR imaging. This constellation of findings at MR imaging should prompt subselective angiography. An isolated dunal thrombosis could have this appearance, although the thrombus is usually seen at MR imaging. In many instances dumal sinus thrombosis may be obvious with routine SE sequences (15,16). In our se196
Radiology
#{149}
a.
b.
Figure
4.
(2,800/80) injection
DAVF
DAVF
sigmoid
at the level of the of contrast material
(black
arrows)
of veno-occlusive
nies
of right
one
patient
left transverse hypointensity
and
disease
with
sinus within
Ti- and T2-wcighted flow-sensitive GRE
sinus
sigmoid in the
without
opacification was
of the right
subsequently
occlusion
images. coronal
disease.
(b) Lateral demonstrates
jugular
vein
(a) Axial view a right
(white
MR image
after selective sigmoid sinus
arrows).
No evidence
demonstrated.
of the
demonstrated this sinus
veno-occlusive
sinus is unremarkable. right occipital artery
correctly
nus on
Only images
demonstrated
thrombosis
in a dunal
sinus
nized
fibrous
strates
bow
the
(Fig may
tissue, signal
dural
si-
3). A thrombus contain
orga-
which
demon-
intensity
on
both
April
1990
a.
b.
C.
Figure 5. subfrontal
along artery
DAVF of anterior cranial fossa with parenchymal hematoma. (a) Non-contrast-enhanced axial CT scan demonstrating a left hematoma. (b) Coronal MR image (600/20) confirms the parenchymal hematoma and demonstrates a dilated draining vein the inferior aspect of the hematoma (arrow). (c) Lateral view after selective injection of contrast material in the left internal carotid demonstrates a DAVF supplied by the anterior ethmoidal branches of the ophthalmic artery (solid black arrows) and the draining
intracranial
vein
(open
arrows)
with
a focal
aneurysmal
dilatation
at the
site
of parenchymal
hemorrhage
I
(white
arrow).
Figure 6. DAVF of superior sagittal sinus demonstrating greater specificity of MR imaging than CT. (a) Non-contrast-enhanced axial CT scan demonstrating a well-defined high-attenuation lesion in the left parietal lobe with surrounding area of low attenuation mimicking a parenchymal hematoma. (b) Axial MR image (600/20) after intravenous administration of gadolinium reveals phase misregistration indicating that the CT finding is an aneurysmally dilated vessel. (C, d) After injection of contrast material in the right internal carotid artery, anteropostenor (C) and lateral view (d) confirm the DAVF of the superior sagittal sinus supplied by right superficial temporal and middle
meningeal
branches
occlusive disease the venous varix.
Ti- and simulating a.
b.
(17)
associated
(black
T2-weighted a flow
recently
arrows)
void.
with
veno-
including
SE images, Strother
described
a similar
thus et al MR
imaging appearance of spontaneous thrombus occurring in giant aneurysms. He noted that the MR imaging appearance of thrombus is diverse and influenced by many factons, including the physical state of the thrombus as well as the varieties of hemoglobin products and proportion of white blood cells, platelet products, and fibnin within the thrombus. Thus, flow-sensitive GRE images may be helpful in evaluating the patency of the dunal sinuses. In our series, 78% of the patients with MR imaging evidence of abnormal venous drainage had complications
d.
C.
Volume
175
#{149} Number
1
seen
on
and parenchymal features in five
MR
images.
Subdumab
hematomas were of the 12 complicaRadiology
197
#{149}
tions seen. These hemorrhagic complications were directly related to vcno-occlusivc disease demonstrated at angiognaphy. In two of the three patients studied with both MR imaging and CT, MR imaging was supenion to CT in delineating the cause of the hemorrhage (Figs 5, 6). In the third patient a spontaneous subdunal hematoma, unrelated to trauma or coagulopathy, was seen on CT and MR images (Fig 2). It should be emphasized that an unexplained subdunal hematoma in a young patient without a history of trauma on coagulation disorder should suggest an underlying DAVF. Venous infarction was the next most common complication seen at MR
imaging.
This
is probably
related
to venous congestion secondary to arterialization of cortical veins. Often, venous infarcts are not found in the same anatomic regions as arterial infancts (Fig 7). MR imaging is supenor to angiography in demonstrating venous infarction (Table). One patient also demonstrated diffuse hypointensity of the left cerebellar hemisphere seen on the predominantly T2-weighted images. At sungery a markedly abnormal left cerebellan hemisphere with diffuse hemosidenin staining was seen. An arterialized vein was cannulated for embolization. No parenchyma was resected. The T2 shortening was thought to represent hemosidenin deposition from chronic venous congestion and/or repeated hemorrhage (Fig 8). There is a high degree of association between dilated cortical veins and the complications depicted at MR imaging. Since both were present at the initial examination, a cause and effect relationship cannot be shown by this review. Previous angiognaphic studies have supported this theory, however (7). The mcidence of complications is probably underestimated at angiography. Thus, MR imaging can suggest venoocclusive disease associated with DAVF by depicting dilated cortical veins and can readily demonstrate complications not seen at angiography.
In our
experience,
this
subgroup
of patients with complicated require more urgent therapy.
DAVF U
2.
198
Newton
TH,
Cronqvist
dural
arteries
in
nous
malformations.
S.
Involvement
intracranial
Radiology
arteriove-
1969;
7. DAVF of vein of Galen with a pontine infarct. (a) Axial demonstrates a focal pontine infarct (arrow). (b) Sagittal MR image cephalus secondary to aqueductal stenosis from the dilated central
a.
(2,700/20) reveals hydrovein (arrows).
d.
matoma
DAVF
of straight
MR image in the inferior
MR image
of
MR image (600/20) draining
b.
C.
(600/20)
sinus
(600/20) vermis.
shows
with
cerebellar
demonstrates This was
dilated
cortical
not
hematoma
and
hemosiderin
methemoglobin (arrow) apparent at angiography.
veins
within
deposition.
in a parenchymal (b) Left parasagittal
the cerebellum.
he-
No parenchymal
AVM nidus is seen. The left cerebellar hemisphere is of diffusely decreased signal intensity. (c) Axial MR image (2,800/80). The more heavily 12-weighted image accentuates the low signal intensity in the left cerebellar hemisphere. The 12 shortening was believed to represent diffuse hemosiderin deposition (confirmed at surgery) from chronic venous congestion. Subacute hematoma in the vermis is high in signal intensity (arrow). (d) Lateral view after
93:1071-1078.
right
Houser 0, Campbell J, Campbell R. Sundt I. Arteriovenous malformation affecting the transverse dural venous sinus: an ac-
straight
#{149} Radiology
b.
Figure
Figure 8. (a) Sagittal
References 1.
a.
vertebral sinus
injection supplied
sion occurred between draining cortical veins
of contrast by
right
material posterior
the superior vermian (white arrows).
shows
a DAVF
meningeal
vein
artery
and
the
(solid
black
arrows)
(open
arrows).
The
straight
sinus,
resulting
involving venous
occ!u-
in dilated
April
1990
quired
3.
nous
4.
5.
6.
lesion.
Clin
embolization
Proc
1979;
8.
of dural
fistulas
in-
M, Zimmerman
aneurysm
and straight surg
1988;
Halbach Wilson
C.
associated
sinus
with
thrombosis.
9.
CB, Hardin
RI, Hieshima
CW,
Kwan
ment of dural fistulas involving cerebral venous system. AJNR
E. the
Huang
C-I,
Wu
C-
VP, Cho SH,
14.
an acquired
Houser 0, Baker H Jr. Rhoton Intracranial dural arteriovenous
11.
Castaigne
mations. fistules
Radiology
P. Bones
1972;
art#{233}rioveineuses
veineux cortical. 1976; 132:169-181. Ito J, Imamura H, Kobayashi
Dural arteriovenous base of the anterior 13.
Neurol
1983;
Felmlee
JP, Ehman
tion:
a method
P. et al. Rev
Neurol
using
Dural
sinus
intermediate
16.
17.
put Assist Tomogr 1986; 10:10-15. Strother CM, Eldevik P. Kikuchi Y, Graves V, Partington C, Merlis A. Thrombus for-
mation
K, et al. malformations of the cranial fossa. Rev
G. Berg BO. study
field strength MR imaging. Radiology 1986; 161:83-86. Macchi P. Grossman R, Gomoni J, Zimmerman R, Bilianiuk L. High field MR imaging of cerebral venous thrombosis. J Com-
Les
m#{233}ning#{233}es pures
Higashida RI, Hieshima GB, D, Newton TH. Dural fistulas involving the transverse and sigmoid sinuses: results of treatment in 28 patients. Radiology 1987; 163:443-447. McMurdo SK, Brant-Zawadzki M, Bradley thrombosis:
AJNR
A Jr. et al. malfor-
VV,
WG Jr. Chang
163:55-64.
J, Brunet
a drainage
GB,
15.
YP. Dural of the major
lesion.
Halbach
Goto K, Norman
3:13-19.
10.
12.
Treatdeep
1989;
W-Y,
MY, Sachdev
sinuses:
1982;
AVM
J Neuro-
Chaudhary
venous
68:288-291.
VV, Higashida
Guo
Weitzner I Jr. Puljic S, Huang arteriovenous malformations
Vein of Gadural
MMH,
9:539-546.
nous malformation in the anterior fossa. No Shinkei Geka 1987; 15:1025-1030. Vi#{241}uela F, Fox AJ, Pelz DM, Drake CG. Unusual clinical manifestations of dural arteriovenous malformations. J Neurosurg 1986; 64:554-558.
Mayberg
leng
C, Chang T. Occlusion of arteriovenous malformations of the cavernous sinus via the superior ophthalmic vein. AJNR 1988;
RI, Hieshima GB, CW. Transve-
volving the transverse and sigmoid sinuses. AJNR 1989; 10:385-392. Takeda N, Fujita 5, Kondo I, Yasuda M, Nakamura M. A case of dural arteriove-
len
7.
Mayo
54:651-661. Halbach VV, Higashida Mehninger CM, Hardin
mass
and structure
and the evolution
of
effect
in intracranial aneurysms treated by balloon embo!ization: emphasis on MR findings. AJNR 1989; 10:787-796.
24:149-154.
RL.
Spatial
for suppressing
presatura-
flow
anti-
facts and improving depiction of vascular anatomy in MR imaging. Radiology 1987;
10:393-399.
164:559-564.
Volume
175
Number
#{149}
1
Radiology
199
#{149}
Kevin
De Marco,
MD
#{149} William
Dural Arterlovenous Evaluation with The preangiographic diagnosis of cerebral dural arteriovenous fistula (DAVF) can be difficult. The magnetic resonance (MR) images of 12 patients with angiographically proved DAVF were evaluated to characterize
the
appearance
of these
lesions and to identify those patients at increased risk for complications. Patients with DAVF demonstrating venous occlusive disease are at higher risk for complications from the arterialized collateral yenous system. This venous occlusive disease is demonstrated best at anteniography.
The
MR
imaging
appear-
ance of dilated cortical veins without a parenchymal nidus is suggestive of a DAVF with veno-occlusive disease. Eight of the 12 patients in our series demonstrated this finding at angiography. Complications, including infarction and hemorrhage, were identified at MR imaging in eight patients with MR imaging evidence of veno-occiusive disease. At angiography 42% of these complications were not apparent. In one patient with a DAVF draining into an unobstructed
right
sigmoid
sinus,
results of MR study were normal. Although patients with DAVF without veno-occlusive disease may have normal findings at MR imaging, DAVF associated with veno-occlusive disease and dilated pial yenous drainage can be documented on MR images. This subset of DAVF patients, many of whom were identified only at MR imaging, is at higher risk for complications due to veno-occlusive disease. These patients are believed to require more urgent therapy. MR imaging is useful in the pretherapeutic planning for patients with DAVF. Index terms: Arteniovenous malformations, cranial, 17.75 #{149}Fistula, arteriovenous #{149} Fistula, dural, 17.75 #{149}Sinuses, dural, 17.75 Radiology
1990; 175:193-199
P. Dillon,
MD
V. Halbach,
MD
arteniovenous fistulas (DAVFs) represent 10%-15% of all intracranial vascular malformations (1). The majority of these involve the venous sinuses along the base of the brain. The clinical symptoms of DAVFs involving the sagittal, lateral, sigmoid, and straight sinuses are often related to the route of URAL
venous
drainage
(2). Unfortunately,
the initial complaints are frequently nonspecific. The preangiographic diagnosis of a DAVF is difficult and mequires a high degree of clinical suspicion. This differs from diagnosis in patients with DAVFs involving the cavernous sinus, who often present with classic symptoms such as proptosis, chemosis, and bruit. These were therefore excluded from this study. The magnetic resonance (MR) imaging appearance of DAVF has not been covered in detail previously. Our goal was to evaluate the MR imaging appearance of angiographically proved sagittal, lateral, sigmoid, and straight sinus DAVFs and attempt to identify features that place patients at high risk for development of complications.
PATIENTS
AND
METHODS
Twelve patients with angiographically proved DAVFs of the sagittal, lateral, sigmoid, or straight sinus were evaluated with MR imaging. There were nine men and three women ranging in age from 35 to 70 years, with a mean age of 52 years. Patient information is summarized in the Table. Angiograms obtained in two orthogo-
1
From
the
Neuroradiology ogy, University Parnassus
#{149} Jay
S. Tsuruda,
MD
Fistulas: Imaging’
MR
D
#{149} Van
Ave.
Diagnostic
and
Interventional
nab planes Spin-echo
were available (SE) MR images
on a i.5-T
unit
(Signa;
for all patients. were obtained
GE Medical
Sys-
tcms, Milwaukee) in all but one patient. In the remaining patient, MR images were obtained on a 0.5-T unit (Diasonics, South San Francisco). Images were obtamed in two orthogonal planes in all patients. In eight cases images in all three orthogonal planes were provided. Tiweighted images were obtained with a repetition time (TR) of 600-800 msec and
an echo
time (TE) of 20-30
msec (TR/TE
600-800/20-30). Proton-density were obtained at 2,000-2,800/20-35. weighted images were acquired
images
T2at 2,0002,800/70. The matrix size was i92 X 256 in four patients and 256 X 256 in eight patients. Section thickness was 5 mm. In selected patients, flow-sensitive gradient recalled echo (GRE) images (150/15; flip angle, 50#{176}) were obtained. In patient 8, axial cine MR imaging (21/9; flip angle, 30#{176}) was performed. The appearance of the arterial supply,
nidus,
venous
drainage,
patency
of the
involved sinus, and associated parenchymal abnormalities were correlated with the angiographic findings (Table). The MR images and angiograms were obtamed within 7 days in all but two cases. In these two cases there was no significant change in the patient’s clinical status between the two examinations. Three patients also underwent contrast materialenhanced computed tomography (CT) (GE 9800; GE Medical Systems) within i week of MR imaging and angiography.
RESULTS As outlined in the Table, the arterial supply to the DAVF was not demonstrated on MR images, except for a possible tentonial arterial feeder seen in patient 2 (Fig 1). The site of the anteniovenous (AV) fistula was always demonstrated at angiognaphy but never seen at MR imaging. No panen-
Section, Department of Radio!of California San Francisco, 505 San
Francisco,
CA
94143.
From
the 1989 RSNA
annual meeting. Received August 29, 1989; revision requested October 26; revision received November 29; accepted Dccember 11. Address reprint requests to J.K.D. © RSNA, 1990
Abbreviations: dural
called =
AV
arteniovenous
echo,
repetition
SE
arteriovenous, fistula, GRE
spin echo,
TE
DAVF gradient reecho time, TR
time.
193
chymab nidus was demonstrated at MR imaging. Abnormal dilated draining veins were identified at MR imaging and angiognaphy in eight of the 12 paticnts (Table). One additional patient demonstrated an occluded sinus without dilated cortical veins. This correlated with angiographicabby proved vcno-occlusive disease. Seven of these nine patients (78%) demonstratcd complications from the DAVF. In one patient with no abnommal draining veins identified on MR images, an unexplained subdurab he-
: a
Ca
6 05
was
correctly
identified
2
.-.
h
6
-E
a .c
.
t-
a
>-
Ca
)
a-
5SC
00.
GJ
Ca U
U
Z(n
>
0
-a
a
a
0.
Ca
a
-g
a a)
-.
a a 0 z
a a
-u .-. x
Zu
a) >.
a) >.
Ca
Ca
Ca
6
6
66
a
.
a)
;s.
bO Ca
6
a 0
05 U
arterial
the
major
sinus
194
#{149} Radiology
network may
in the
0
0
0
z
zz
(IC
a
0.
0
-0 a) -0
-
0 -
-
a)
Ca
bO 0
U
U
0
Ca
6
6
0
0
-
0.
Ca
-
QJ
.
;.
cn
j
Ca
Ca
c0
c a)
a
a
a
a U) a
(IC
(IC a)
a
a
0
0
0
0
0
0
Ca
a) 0
0
U U
0
0
0
0
aj Cl) a a
ZZZZZ
a 5) a
a
a) a) a 0
z
000
Z>
>.ZZ>.>.Z
Ca
0
0
0
0
0
0
ZZZZZZ
ZZ
0 a)
-
Ca
Ca
66 U U
0
0
0
ZZ
-
-
-
Ca
Ca
Ca
EE6E
U U
‘
0
Ca
.6 0
OZ
U
‘
U
0
0
OZZZOZ
0 a)
:
ii
6, 0.
0.
a
6
a
a
.
a)
a)
;.
a)
a)
Z
o.
o.
Z
o.
a)
00
a
-0 a)
-
0
a bO Ca
Ca
ZZ
0.
-
Ca
a a
0
a
0
0.,
(1)
Ca
a
0
a
a) 0.
0
a)
b0
a
a)
0.
(I)
-:;i
ILl
.bo 1)
a
2aaa
I-
a)
(I)
u(IC
0_
a
u(IC
a)
a)
;..
o.
o.
Z
;.
Ca
ZO
a 0.
a)
a)
.
.
g
Z
(IC
(IC
(I)
a
a
a
-
a)
.
a
._
i.
.-.
bO
bO
(IC
(IC
aCa
aCa
-g C
OJ
r: a) ,._
: C)
(3
5)
(1)
0 0.
.0-u
-
‘
0
a
>a
,
a
so
.,
g
a) U Ca
a
4-
a
-0
-.
.
Ca
Ca
ILl
CU
CU
(IC
a
a
-;;i
-
0
.
ba’...
n.
bo
Ca
(IC
a
a 0
Ca _
-j
..
I..
-:ao
a
_ ‘
-g6
a
n
a
Ca
05
.-
‘.
U)
-.
a a Ca a 5j
U
.
bO
deep
walls
Ca
0
z
a
attempt
a
ZZ>n.ZZ
(I>
U
0
DISCUSSION
dunab
a
a
bC
a a
0>
(5), the
SJ
:-.
a)
a
venous system (6,7), and the cavernous sinus (8). Several authors have postulated that the development of DAVF may begin with thrombosis or occlusion of the involved sinus (2,9). The rich
5
6
a
a
of durab arteries and a vein or venous sinus (3). Many of the DAVFs involve the venous sinuses along the base of the brain, and here the transverse sinus is the most commonly affcctcd (2). DAVFs can also involve veins in the anterior cranial fossa (4),
a
:a
a
bO
‘.-
sinus
-
sJ
aJ
6 a’
-
bO 0
.
sagittab
OO
.
aJ
6
In one case, results of MR imaging were falsely negative. No dilated contical veins, sinus occlusion, on complications were seen (Fig 4). The angiogram demonstrated a right sigmoid DAVF without veno-occlusive disease.
superior
.ua n
LI
(1)
ab-
Ca
--
0.
dunal
DAVFs are infrequent vascular normalities consisting of numerous tiny connections between branches
2Ca 6
Ca
.
-
u_u a o.
0
in sev-
involved
-u
.,
a
en patients. Two patients demonstrated occlusion of the involved sinus at both MR imaging and angiography. Partial on complete sinus thrombosis was underestimated on routine SE images in three patients. When used, GRE images were helpful in evaluating flow in draining
veins as well as in the sinuses (Fig 3).
ca)
(IC
matoma was noted (Fig 2). Overall, 12 complications were seen in eight patients including infarction (n 3), intraparcnchymal hematoma (n = 3), subdumal hcmatoma (n = 2), diffuse hemosidenin deposit possibly secondary to venous infarct (n 1), vasogenic edema (n = 1), hydmocephabus (n 1), and proptosis (n 1) (Table). Five of the 12 complications (42%) were not readily apparent at angiography. The patency of the involved dural
sinus
n.
E
bo
.9
a)
b0
: Lfl ‘
Ca’ ‘
‘
m
n
(0
0
r
00
0
Lt)
0’
L.
.
0 N.
C’4 LO
C1
Z
to necanApril
1990
b. Figure 1. DAVF involving lated vessel coursing along
dural
arterial
supply
can
left transverse the left tentorium be seen
against
the
(a)
sinus.
(arrows). tentorium
Axial
MR
image
(2,000/35)
The signal
void
but
be impossible
would
of this
reveals
a di-
presumed to separate artery
from cortical bone on standard SE images. (b) Lateral view of the left internal carotid after injection of contrast material confirms the presence of a dilated tentomial branch arrows) supplying the fistula located in the left transverse sinus (open arrows).
(solid
synchronous
tinnitus.
unobstructed
venous
In DAVFs
drainage,
with
corti-
cal venous pathways usually do not develop. Hemorrhagic complications with these DAVFs are infrequent (5). Patients with carotid cavernous DAVFs, unlike those with DAVFs seen in other locations, often present with classic clinical findings of proptosis, chemosis, and bruit. They rarely demonstrate complications such as stroke or hemorrhage. Such patients were excluded from this study. The most severe symptoms associated with DAVF are due to cortical venous drainage secondary to venoocclusive disease (10). In a review of DAVF by Castaigne et al (11), 42% of DAVFs draining into cortical veins were associated with hemorrhage. However, Houser et al (2) point out that the sinus need not even be thrombosed to cause retrograde yenous drainage into cortical veins.
Partial
filling
of sinuses
by the
high-
pressure arterial blood can impede antegrade venous drainage. Because of their location away from major dumal sinuses, DAVFs involving the anterior cranial fossa and tentonium almost always drain into intradural veins and are associated with a high frequency of hemorrhage (12). In addition to hemorrhage, retrograde yenous flow accompanied by increased venous pressure can lead to chronic passive congestion and venous infarcts. This is a major cause of focal neurologic deficits. Hydrocephalus is another complication of DAVF. This can be related to decreased absorption of cerebrospinal fluid due to elevated venous pressure or mechanical compression of the aqueduct by mesencephalic veins. Proper
important
b. Figure
(a)
2.
Axial
sagittal
tion
DAVF MR
of superior
image
Ti-weighted
disorder,
of contrast terial
sagittal
(2,000/70)
sinus
seen
demonstrates
The
hematoma. sinus
is patent
on
(not shown). Since there was no history of trauma or coagulawas suspected. (b) Image of left common carotid artery after injection in the late arterial phase reveals a DAVF with left middle meningeal amarrows) entering in the inferior/lateral sinus wall (solid arrow). images
material
supply
(open
seen
at angiogmaphy.
The
volved dunal sinus receives arterialized blood flow that can lead to mechanical obstruction of the sinus and result in retrograde drainage of blood away from the sinus and
Volume
subdural
hematoma.
a DAVF
alize the sinus by developing direct artery-to-sinus communications. This may lead to the angiomatous network of multiple feeding arteries and numerous AV shunts within a pantially mecanalized sinus that is fre-
quently
as a spontaneous
a subdural
175
#{149} Number
1
in-
into
cortical
veins.
A better understanding of the natumal course of DAVF has bed to an improvement in the diagnosis and treatment of this disease. Bruit and headache arc the common symptoms of a DAVF and are usually unrelated to cortical venous drainage. When the DAVF involves a dural sinus in direct contact with the petrous pyramid, the resulting higher turbulent flow is often manifested as pulse-
MR
imaging
when
technique
assessing
is
patients
with DAVF, as both parenchymal abnormalities and venous thrombosis must be assessed. Acquisition of Tiand T2-weighted images is mandatory. At least two orthogonal planes should be used. If flow compensation techniques are used with T2-weighted images, refocusing of laminar flow will result in increased signal
intensity
and
may
obscure
or simu-
late thrombosis. This technique may also reduce the conspicuity of dilated cortical veins. Saturation bands are helpful on both Ti- and T2-weighted images to reduce intraluminal signal intensity caused by slow flow or the entry-section phenomenon (13). GRE images can be helpful in evaluating the vascular anatomy of DAVF and the patency of the durab sinuses.
Radiology
#{149} 195
Figure 3. DAVF of left sigmoid sinus with occlusion of left transverse sinus. (a) Lateral view after selective injection of contrast material in the left occipital artery demonstrates
a DAVF
involving
the
patent
but
nar-
rowed left sigmoid sinus (open arrows). Proximally the left transverse sinus is occluded (solid arrow). (b) Axial MR image (2,800/80) ty in the
suggests left jugular
increased bulb but
signal is not
intensidiagnos-
tic of partial occlusion. Occlusion of the left transverse sinus is not seen. (c) Coronal GRE MR image (150/15; flip angle, 50#{176}) demonstrates an occluded distal left transverse sinus.
Irregular
narrowing
of the
left
sigmoid
sinus (not shown) secondary to the DAVF was also noted. (d) Coronal GRE MR image (150/15; flip angle, 50#{176}) obtained posterior to c reveals
that
left transverse
the
sinus
proximal
portion
of the
is patent.
a.
b.
C.
d.
Although the actual fistula beartery and vein was depicted in all patients at angiography, it was never seen directly with MR imaging. This is probably related to the small size of the fistula as well as the lack of contrast between the signal void in dural vessels and the lack of signal from cortical bone. Indeed, a patient with a DAVF without venoocclusive disease or dilated cortical veins may have completely normal findings at MR examination (Fig 4). Thus, we do not believe MR imaging tween
is suitable
as the
sole
screening
cx-
amination for DAVF. Dilated cortical veins in patients with DAVF arc usually, but not a!ways, associated with veno-occlusive disease. Angiography had demonstrated that veno-occlusive disease can occur at any point in the draining venous system. Thus, the nanrowed vessel may be at the junction of the AV fistula and the dural sinus, within the dural sinus, or in a draining vein away from the fistula site (14). MR imaging was able to document all cases of dilated cortical veins associated with veno-occlusive disease seen at angiognaphy. Further, the MR imaging appearance of dilated cortical veins in the absence of a panenchymal AV malformation nidus suggests the presence of a DAVF. A parenchymal AV fistula, pial AV ma!formation, or anomalous cortical yenous drainage could have a similar appearance at MR imaging. This constellation of findings at MR imaging should prompt subselective angiography. An isolated dunal thrombosis could have this appearance, although the thrombus is usually seen at MR imaging. In many instances dumal sinus thrombosis may be obvious with routine SE sequences (15,16). In our se196
Radiology
#{149}
a.
b.
Figure
4.
(2,800/80) injection
DAVF
DAVF
sigmoid
at the level of the of contrast material
(black
arrows)
of veno-occlusive
nies
of right
one
patient
left transverse hypointensity
and
disease
with
sinus within
Ti- and T2-wcighted flow-sensitive GRE
sinus
sigmoid in the
without
opacification was
of the right
subsequently
occlusion
images. coronal
disease.
(b) Lateral demonstrates
jugular
vein
(a) Axial view a right
(white
MR image
after selective sigmoid sinus
arrows).
No evidence
demonstrated.
of the
demonstrated this sinus
veno-occlusive
sinus is unremarkable. right occipital artery
correctly
nus on
Only images
demonstrated
thrombosis
in a dunal
sinus
nized
fibrous
strates
bow
the
(Fig may
tissue, signal
dural
si-
3). A thrombus contain
orga-
which
demon-
intensity
on
both
April
1990
a.
b.
C.
Figure 5. subfrontal
along artery
DAVF of anterior cranial fossa with parenchymal hematoma. (a) Non-contrast-enhanced axial CT scan demonstrating a left hematoma. (b) Coronal MR image (600/20) confirms the parenchymal hematoma and demonstrates a dilated draining vein the inferior aspect of the hematoma (arrow). (c) Lateral view after selective injection of contrast material in the left internal carotid demonstrates a DAVF supplied by the anterior ethmoidal branches of the ophthalmic artery (solid black arrows) and the draining
intracranial
vein
(open
arrows)
with
a focal
aneurysmal
dilatation
at the
site
of parenchymal
hemorrhage
I
(white
arrow).
Figure 6. DAVF of superior sagittal sinus demonstrating greater specificity of MR imaging than CT. (a) Non-contrast-enhanced axial CT scan demonstrating a well-defined high-attenuation lesion in the left parietal lobe with surrounding area of low attenuation mimicking a parenchymal hematoma. (b) Axial MR image (600/20) after intravenous administration of gadolinium reveals phase misregistration indicating that the CT finding is an aneurysmally dilated vessel. (C, d) After injection of contrast material in the right internal carotid artery, anteropostenor (C) and lateral view (d) confirm the DAVF of the superior sagittal sinus supplied by right superficial temporal and middle
meningeal
branches
occlusive disease the venous varix.
Ti- and simulating a.
b.
(17)
associated
(black
T2-weighted a flow
recently
arrows)
void.
with
veno-
including
SE images, Strother
described
a similar
thus et al MR
imaging appearance of spontaneous thrombus occurring in giant aneurysms. He noted that the MR imaging appearance of thrombus is diverse and influenced by many factons, including the physical state of the thrombus as well as the varieties of hemoglobin products and proportion of white blood cells, platelet products, and fibnin within the thrombus. Thus, flow-sensitive GRE images may be helpful in evaluating the patency of the dunal sinuses. In our series, 78% of the patients with MR imaging evidence of abnormal venous drainage had complications
d.
C.
Volume
175
#{149} Number
1
seen
on
and parenchymal features in five
MR
images.
Subdumab
hematomas were of the 12 complicaRadiology
197
#{149}
tions seen. These hemorrhagic complications were directly related to vcno-occlusivc disease demonstrated at angiognaphy. In two of the three patients studied with both MR imaging and CT, MR imaging was supenion to CT in delineating the cause of the hemorrhage (Figs 5, 6). In the third patient a spontaneous subdunal hematoma, unrelated to trauma or coagulopathy, was seen on CT and MR images (Fig 2). It should be emphasized that an unexplained subdunal hematoma in a young patient without a history of trauma on coagulation disorder should suggest an underlying DAVF. Venous infarction was the next most common complication seen at MR
imaging.
This
is probably
related
to venous congestion secondary to arterialization of cortical veins. Often, venous infarcts are not found in the same anatomic regions as arterial infancts (Fig 7). MR imaging is supenor to angiography in demonstrating venous infarction (Table). One patient also demonstrated diffuse hypointensity of the left cerebellar hemisphere seen on the predominantly T2-weighted images. At sungery a markedly abnormal left cerebellan hemisphere with diffuse hemosidenin staining was seen. An arterialized vein was cannulated for embolization. No parenchyma was resected. The T2 shortening was thought to represent hemosidenin deposition from chronic venous congestion and/or repeated hemorrhage (Fig 8). There is a high degree of association between dilated cortical veins and the complications depicted at MR imaging. Since both were present at the initial examination, a cause and effect relationship cannot be shown by this review. Previous angiognaphic studies have supported this theory, however (7). The mcidence of complications is probably underestimated at angiography. Thus, MR imaging can suggest venoocclusive disease associated with DAVF by depicting dilated cortical veins and can readily demonstrate complications not seen at angiography.
In our
experience,
this
subgroup
of patients with complicated require more urgent therapy.
DAVF U
2.
198
Newton
TH,
Cronqvist
dural
arteries
in
nous
malformations.
S.
Involvement
intracranial
Radiology
arteriove-
1969;
7. DAVF of vein of Galen with a pontine infarct. (a) Axial demonstrates a focal pontine infarct (arrow). (b) Sagittal MR image cephalus secondary to aqueductal stenosis from the dilated central
a.
(2,700/20) reveals hydrovein (arrows).
d.
matoma
DAVF
of straight
MR image in the inferior
MR image
of
MR image (600/20) draining
b.
C.
(600/20)
sinus
(600/20) vermis.
shows
with
cerebellar
demonstrates This was
dilated
cortical
not
hematoma
and
hemosiderin
methemoglobin (arrow) apparent at angiography.
veins
within
deposition.
in a parenchymal (b) Left parasagittal
the cerebellum.
he-
No parenchymal
AVM nidus is seen. The left cerebellar hemisphere is of diffusely decreased signal intensity. (c) Axial MR image (2,800/80). The more heavily 12-weighted image accentuates the low signal intensity in the left cerebellar hemisphere. The 12 shortening was believed to represent diffuse hemosiderin deposition (confirmed at surgery) from chronic venous congestion. Subacute hematoma in the vermis is high in signal intensity (arrow). (d) Lateral view after
93:1071-1078.
right
Houser 0, Campbell J, Campbell R. Sundt I. Arteriovenous malformation affecting the transverse dural venous sinus: an ac-
straight
#{149} Radiology
b.
Figure
Figure 8. (a) Sagittal
References 1.
a.
vertebral sinus
injection supplied
sion occurred between draining cortical veins
of contrast by
right
material posterior
the superior vermian (white arrows).
shows
a DAVF
meningeal
vein
artery
and
the
(solid
black
arrows)
(open
arrows).
The
straight
sinus,
resulting
involving venous
occ!u-
in dilated
April
1990
quired
3.
nous
4.
5.
6.
lesion.
Clin
embolization
Proc
1979;
8.
of dural
fistulas
in-
M, Zimmerman
aneurysm
and straight surg
1988;
Halbach Wilson
C.
associated
sinus
with
thrombosis.
9.
CB, Hardin
RI, Hieshima
CW,
Kwan
ment of dural fistulas involving cerebral venous system. AJNR
E. the
Huang
C-I,
Wu
C-
VP, Cho SH,
14.
an acquired
Houser 0, Baker H Jr. Rhoton Intracranial dural arteriovenous
11.
Castaigne
mations. fistules
Radiology
P. Bones
1972;
art#{233}rioveineuses
veineux cortical. 1976; 132:169-181. Ito J, Imamura H, Kobayashi
Dural arteriovenous base of the anterior 13.
Neurol
1983;
Felmlee
JP, Ehman
tion:
a method
P. et al. Rev
Neurol
using
Dural
sinus
intermediate
16.
17.
put Assist Tomogr 1986; 10:10-15. Strother CM, Eldevik P. Kikuchi Y, Graves V, Partington C, Merlis A. Thrombus for-
mation
K, et al. malformations of the cranial fossa. Rev
G. Berg BO. study
field strength MR imaging. Radiology 1986; 161:83-86. Macchi P. Grossman R, Gomoni J, Zimmerman R, Bilianiuk L. High field MR imaging of cerebral venous thrombosis. J Com-
Les
m#{233}ning#{233}es pures
Higashida RI, Hieshima GB, D, Newton TH. Dural fistulas involving the transverse and sigmoid sinuses: results of treatment in 28 patients. Radiology 1987; 163:443-447. McMurdo SK, Brant-Zawadzki M, Bradley thrombosis:
AJNR
A Jr. et al. malfor-
VV,
WG Jr. Chang
163:55-64.
J, Brunet
a drainage
GB,
15.
YP. Dural of the major
lesion.
Halbach
Goto K, Norman
3:13-19.
10.
12.
Treatdeep
1989;
W-Y,
MY, Sachdev
sinuses:
1982;
AVM
J Neuro-
Chaudhary
venous
68:288-291.
VV, Higashida
Guo
Weitzner I Jr. Puljic S, Huang arteriovenous malformations
Vein of Gadural
MMH,
9:539-546.
nous malformation in the anterior fossa. No Shinkei Geka 1987; 15:1025-1030. Vi#{241}uela F, Fox AJ, Pelz DM, Drake CG. Unusual clinical manifestations of dural arteriovenous malformations. J Neurosurg 1986; 64:554-558.
Mayberg
leng
C, Chang T. Occlusion of arteriovenous malformations of the cavernous sinus via the superior ophthalmic vein. AJNR 1988;
RI, Hieshima GB, CW. Transve-
volving the transverse and sigmoid sinuses. AJNR 1989; 10:385-392. Takeda N, Fujita 5, Kondo I, Yasuda M, Nakamura M. A case of dural arteriove-
len
7.
Mayo
54:651-661. Halbach VV, Higashida Mehninger CM, Hardin
mass
and structure
and the evolution
of
effect
in intracranial aneurysms treated by balloon embo!ization: emphasis on MR findings. AJNR 1989; 10:787-796.
24:149-154.
RL.
Spatial
for suppressing
presatura-
flow
anti-
facts and improving depiction of vascular anatomy in MR imaging. Radiology 1987;
10:393-399.
164:559-564.
Volume
175
Number
#{149}
1
Radiology
199
#{149}
