OIVKV S. Kargcr AG. Bawl
0014-3072/89/029$ 0294S2.7S/0

Eur Neurol 1989:29:294-297

Positron Emission Tomography Studies of Changes in Cerebral
Blood Flow and Oxygen Metabolism in Arteriovenous Malformation
of the Brain
(With I color plate)

J De Reuck*. J. Van Akenh, I f Van I.undegema, A. Vakael*
Departments of “Neurology and ‘’Anaesthesiology. University Hospital; Cyclotron Unit, laboratory of Analytical Chemistry.
Institute for Nuclear Sciences; Laboratory for Electronics and Metrology. Gent. Belgium

Key Words. Arteriovenous malformation

PET • Seizure • Transient focal neurological deficit

Introduction
Arteriovenous malformations consist of a tangle of
dilated vessels, which form an abnormal communication
between the arterial and venous systems. I lalf of the
patients present with a cerebral haemorrhage. 30% with
a seizure and 20% with migrainous headache or a tran­
sient neurological deficit (1J. The aim of the study is to
analyse by means of positron emission tomography
(PET) the possible mechanisms that had caused tempo­
rary disturbances of brain function in 2 patients with
arteriovenous malformations.
Patients and Methods
Patients
Patient /. A 48-year-old man had a history of two episodes of
auditory hallucinations with subsequent loss of consciousness and
generalized convulsions.
Neurological examination on admission was normal. Computed
tomography (CT) scan of the brain and angiography demonstrated a
small arteriovenous malformation in the left insular region The

mallormation was already filled in the early artenul phase and
drained by a large vcnc to the lateral sinus.
Patient 2. A 65-year-old right-handed woman had a history of
recurrent convulsions, starting at the right leg with secondary generali/alion. from the age of 35 years on. A large arteriovenous malfor­
mation. fed by the end branches of the left anterior cerebral artery',
in the left parietal region, was demonstrated by angiography at the
age of 42 years. The malformation was considered not to be surgi­
cally removable. Although high doses of anti-epileptic drugs were
given, this patient still had recurrent epileptic fits, but also transient
episodes of right hemiparesis.
In 1984. a partial embolization of the malformation was per­
formed. Afterwards the patient had no more seizures, but the epi­
sodes of transient hemiparesis recurred approximately every 2
months. CT scan of the brain demonstrated only the malformation
at the medial side of the left cerebral hemisphere. No infarction was
observed. Left carotid angiography showed a dilated left anterior
cerebral artery, feeding a large medial parietal arteriovenous malfor­
mation Venous drainage appeared in the early arterial phase. The
branches of the left middle artery were visible at the same time as
the malformation. The right carotid angiography showed poor opa­
cification of the right anterior cerebral artery, by draining of the
contrast media to the left anterior cerebral artery by the anterior
communicating artery. The branches of the right middle cerebral
artery were well tilled.

Downloaded by:
Nagoya University
133.6.82.173 - 1/13/2019 6:47:45 AM

Abstract. Two patients with arteriovenous malformation were studied by position emission tomography, using
the steady-state technique with l502. In the first patient, who had seizures, preceded by auditory hallucinations, an
area of decreased cerebral blood How and oxygen consumption was shown just behind the malformation. In the
second patient, who had repeated attacks of right hemiparesis. an area of decreased blood flow and oxygen metab­
olism was shown at distance of the malformation. These areas of low flow and metabolism are most probably the
result of a vascular steal phenomenon and the origin of the transient neurological symptoms in both patients.

Plate I

De Kcuck/Van Aken/Van 1andegem/Vakaet

d

Fig. 1. Comparison of the CT scan (a. c) with the ^CO; scan
(h, d) in two slides, respectively, 3 mm (a. b) and 35 mm (c. d) above
the orbilomealul line in patient I. The left arteriovenous malforma­
tion (arrow) can be picked up on the CT scan as well as on the CBF
scan (a. b). In addition, increased flow is demonstrated in the upper
cortical regions and the basal ganglia of the left hemisphere.

Eur Neurol. Vol. 29

c

d

Fig. 2. Comparison of the CT scan (a, c) with the ,5C 0 2 scan
(b. d) in two slides, respectively. 3 mm (a. b) and 35 mm (c. d) above
the orbitomeatal line in patient 2. The arteriovenous malformation,
drained by the left anterior cerebral artery, and a zone of increased
rCnF in the territory' of the feeding vessels arc show n. In addition, a
restricted area of decreased flow is demonstrated at the convex side
o f the left cerebral hemisphere (d. arrow). No corresponding struc­
tural change can be shown on the CT scan

S. Karger. Basel

Downloaded by:
Nagoya University
133.6.82.173 - 1/13/2019 6:47:45 AM

c

295

PET and Arteriovenous Malformation

PET Scan
For the measurements of the regional blood flow (rCBF), oxygen
extraction rate (rOER) and oxygen consumption (rCMROz), the
procedures were essentially similar to those described by Frackowiak et al. [2].
A Neuro-Ecat PET scanner (EG&Ortec, Oak Ridge, Tenn.,
USA), equipped with 2 rings of detectors and with lateral and axial
resolution of 8.1 and 14.0 mm, respectively, were used. The distance
between the centres of the 2 detector rings was 32 mm.
After a bianco scan had been performed with an external ger­
manium-68 ring source, the head of the patient was positioned to
allow the accumulation of data from 2 coronal slices, with their
centres 3 and 35 mm above the orbitomeatal line. Thereafter, a
transmission scan was performed with the external germanium-68
ring source to measure photon attenuation and to correct the subse­
quent emission scans.
For measuring rCBF, ISC02 was inhaled by the patient at a rate
of 20|xCi/min. After 10 min of equilibration, the radioactivity in
the head was scanned during 300 s. Two arterial blood samples were
obtained anaerobically at the beginning and at the end of the
scan.
For measuring rOER, lsCC>2 was inhaled at a rate of 40 pCi/min,
and, after an equilibration period of 10 min, the scan time was
extended to 400 s and also two arterial blood samples were taken
anaerobically at the beginning and at the end of the procedure.
The measurement of the radioactivity in the blood samples and
the rCBF, rOER and rCMR02 determinations were calculated
according to the methods described by Weyne et al. [3].
Several cortical regions, basal ganglia and white matter of the
cerebral hemisphere with the arteriovenous malformation were ex­
amined separately and the values obtained were compared with
those of the corresponding regions of the contralateral hemi­
sphere.
Statistical significance was determined by test for paired values
and p values of less than 0.05 were considered to be significant.

Results
Patient 1
The rCBF, rOER and rCMR02 values obtained 14
days after the last epileptic fit are shown in table 1.
In the left frontal and anterior temporal cortical areas,
the values for rCBF and rCMR02 are increased by more
than +100%, while the changes in rOER are similar to
those of the corresponding regions in the non-affected
hemisphere.
Also in the left basal ganglia, a deviation of + 65% is
found for rCBF and rCMR02, with comparable values
for rOER between the left and right side. On the other
hand, the rCBF is lowered to 70% and the rCMR02 to

62% in the left parietotemporal cortex, lying behind the
arteriovenous malformation.
The CT scan of the corresponding sections shows only
a small contrast in the left insular region. No other struc­
tural lesions are observed (fig. 1).
Patient 2

PET scan was performed in this patient 7 days after
the last episode of transient right hemiparesis. The val­
ues of rCBF, rOER and rCMR02 are shown in table 2. In
the medial cortical part of the left cerebral hemisphere,
the increase of rCBF is approximately + 115% and that
of rCMR02, +96%, with no significant deviation of the
rOER. In the area where the large malformation is

Table I. Comparison of the rCBF, rOER and rCMRÛ2 between
different regions of the left (L) cerebral hemisphere with the arterio­
venous malformation and the right (R) cerebral hemisphere in
patient 1 (p < 0.05)
Hemisphere
region

rCBF
rOER
ml/min/lOO g %

rCMROj
ml/min/IOOg

Anterior
insular cortex

L
R

104.6
50.3

61.2
59.5

9.9
4.6

Posterior
insular cortex

L
R

36.2
51.7

61.1
65.5

3.2
5.2

Frontal cortex

L
R

68.2
38.3

53.8
50.2

6.2
3.6

Basal ganglia

L
R

48.9
29.4

49.1
53.3

4.8
3.0

Table 2. Comparison of rCBF, rOER and rCMRÛ2 between dif­
ferent regions of the left (L) cerebral hemisphere with the arteriove­
nous malformation and the right (R) cerebral hemisphere in pa­
tient 2 (p < 0.05)
Hemisphere
region

rCBF
ml/min/100g

rOER
%

rCM R02
ml/min/100g

Medial
frontal cortex

L
R

121.3
55.9

59.9
66.7

12.2
6.6

Medial
parietal cortex

L
R

126.8
51.9

57.8
68.8

11.6
5.8

Lateral
frontal cortex

L
R

18.4
37.1

53.7
54.4

1.7
3.9
Downloaded by:
Nagoya University
133.6.82.173 - 1/13/2019 6:47:45 AM

On neurological examination between the attacks, only brisk
tendon reflexes and a Babinski sign on the right side, without loss of
strength, were found.

296

De Reuck/Van Aken/Van Landegem/Vakaet

explain the periods of transient hemiparesis without
speech disturbances. The values found for those regions
suggest ischaemic infarction [14], not seen on the corre­
sponding slides of the CT scan.
Concerning the cause of the cerebral hypoperfusion of
some regions adjacent or at distance from the arteriove­
nous malformation several hypotheses can be discussed.
Vasospasm occurs rarely in arteriovenous malforma­
tions [15, 16], and is associated with the presence of a
subarachnoid bleeding, which is not the case in our 2
patients. Diaschisis has been described as a phenomenon
of reduction of rCBF in uninvolved areas distant from
Discussion
an acute stroke [17]. This has never been described to
In the 2 patients examined, an increased CBF and occur in arteriovenous malformations. So the most prob­
CMRO 2 are shown in the regions lying in the supply area ably cause of the low rCBF and ÆMRO 2 is an intracere­
of the feeding vessels of the arteriovenous malforma­ bral arterial steal phenomenon, caused by the shunting
of blood through the arteriovenous malformation.
tion.
This phenomenon has already been demonstrated to
Even in the first case with only a small malformation
in the insular region, high values of rCBF and rCMRC>2 occur frequently in these conditions, as well in regions
are found in an extensive zone of the left cerebral hemi­ adjacent to the malformation as in remoted areas [6, 1820 ],
sphere.
As we have no normal controls one cannot exclude
that flow and oxygen metabolism in the contralateral
hemisphere are diminished. Neither have we determined
Acknowledgments
the regional blood volume, which can be important for a
This study was supported by a grant of the Scientific Research
more precise determination of the rCMRO?. Omitting
Committee
of the Departments of the Prime Minister (no.
such corrections leads to a systematic overestimation of
O.O.A.I2053 B85).
OER, that increases with increasing cerebral blood vol­
We would like to thank Prof. J. Wcyne. Dr. G. De Ley, Prof. G.
ume [4], However, in both cases, the rOER values are De Meester, Dr. K. Stryckmans, Dr. P. Goethals and Mr. De Kesel
comparable in the corresponding regions of both cere­ for their technical assistance and help.
bral hemispheres.
This allows to postulate that the impact of the malfor­
mation as a 'blood lake’ does not interfere to much in the
References
determination of the CMRO2. In spite of these limita­
1 Adams R. Victor M: Principles of Neurology, New York,
tions, we can put forward that the regions of increased
McGraw-Hill, 1977, pp 546-547.
CBF and CMRO 2 extend far beyond the anatomical lim­
2 Frackowiak RSJ, Lenzi GL, Jones T, et al: Quantitative mea­
its of the arteriovenous malformation and that even in
surement of regional cerebral blood flow and oxygen metabolism
case of a small malformation the disturbances of flow
in man using lsOi and positron emission: Theory, procedure and
and metabolism are extensive [5-10]. Also in both cases
normal values. J Comput Assist Tomogr 1980;4:727-736.
3 Weyne J, De Ley G, De Meester G, et al: PET studies of changes
other regions have lowered rCBF and rCMR02- Their
in cerebral blood flow and oxygen metabolism after unilateral
decrease is probably responsible for the temporary brain
microembolization of the brain in anaesthetized dogs. Stroke
dysfunction.
1987;18:128-137.
In the first case, the auditory hallucinations as aura of
4 Lammerstma AA, Wise RJS, Heather JD, et al: Correction for
the epileptic insults can have their origin in the left audi­
the presence of intravascular oxygen-15 in the steady-state tech­
nique for measuring regional oxygen extraction ratio in the
tory cortex of the post-insular region with low rCBF and
brain:
2. Results in normal subjects and brain tumour and stroke
rCMRCK Epileptic foci interictally are frequently recog­
patients. J Ccreb Blood Flow Metab 1983;3:425-431.
nized as areas of low metabolism and blood flow [11 —
5 Feindel W. Yasamoto Y, Hodge C: Red cerebral veins and the
13].
cerebral steal syndrome: Evidence from fluorescein angiography
In the second case, the finding of a zone of low rCBF
and microregional blood flow by radio-isotopes during excision
of an angioma. J Neurosurg 1971;35:167-179.
and rCMRC>2 in the left upper frontoparietal region can

Downloaded by:
Nagoya University
133.6.82.173 - 1/13/2019 6:47:45 AM

present, the rCBF increases up to more than 400% and
the rCMRC>2 up to more than 250%. In the frontoparie­
tal region at the convex side of the left cerebral hemi­
sphere, a significant decrease of rCBF of 50% and of
rCMRC>2 of 58% is observed, without deviation of the
values of rOER.
The CT scan of comparable sections can only demon­
strate the large arteriovenous malformation at the me­
dial side of the left cerebral hemisphere (fig. 2).

297

PET and Arteriovenous Malformation

222 .

15 Mohr JP, Kase CS: Spasme artériel cérébral dans les malforma­
tions vasculaires cérébrales. Rev Neurol 1983;139:99-113.
16 Sasaki T. Mayanagi Y, Yano T, et al: Cerebral vasospasm with
subarachnoid hemorrhage from cerebral arteriovenous malfor­
mations. Surg Neurol 1983:16:183-187.
17 Reivich M: Pathophysiology of cerebral ischemia; in Harrison
MJG, Dyken ML (eds): Cerebral Vascular Disease. London, But­
terworths, 1983, pp 86-136.
18 Deutsch G: Blood flow changes in arteriovenous malformation
during behavioral activation. Ann Neurol 1983;13:38-43.
19 Solomon RA, Michelson WJ: Defective cerebrovascular autoreg­
ulation in regions proximal to arteriovenous malformations of
the brain: A case report and topic review. Neurosurgery 1981; 14:
78-82.
20 Homan RW, Devous MD, Stokely EM, et al: Quantification of
intracerebral steal in patients with arteriovenous malformation.
Arch Neurol 1986:43:779-785.

Received: December 2, 1988
Accepted: March 6, 1989
Jacques Dc Reuck. MD, PhD
Department of Neurology
University Hospital
De Pintelaan 185
B-9000 Gent (Belgium)

Downloaded by:
Nagoya University
133.6.82.173 - 1/13/2019 6:47:45 AM

6 Prosenz P, Heiss WD, Kvicola V, et al: Contribution to the
hemodynamics of arteriovenous malformations. Stroke 1971 ;2:
279-289.
7 Kusske J, Kelly W: Embolization and reduction of the ‘steal'
syndrome in cerebral arteriovenous malformations. J Neurosurg
1974:40:313 -3 2 1.
8 Nomes H, Grip A: Hemodynamic aspects of cerebral arteriove­
nous malformations. J Neurosurg 1980;53:456-464.
9 Stein BW, Wolpert SM: Arteriovenous malformations of the
brain: Current concepts and treatment. Arch Neurol I980;37:
1-5.
10 Takeuchi S, Kikuchi H, Karasawa J. et al: Cerebral hemodynam­
ics in arteriovenous malformations: Evaluation by single photon
emission CT. Am J Neuroradiol 1987;8:193-197.
11 Kuhl DE. Engel JJ, Phelps ME. et al: Epileptic patterns of local
cerebral metabolism and perfusion in humans, determined by
emission computed tomography of l8FDG and IJNHj. Ann Neu­
rol 1980:8:342-360.
12 Mazziotta JC, Engel JJ: The use and impact of positron com­
puted tomography scanning in epilepsy. Epilepsia 1984;25(suppl
21:86-104.
13 Dasheiff RM, Rosenbek J. Mattheus C, et al: Epilepsy surgery
improves glucose metabolism on PET scan: A case report. J Neu­
rol 1987;234:283-288.
14 Wise JS, Bernardi S. Frackowiak RSJ, et al: Serial observations
on the pathophysiology of acute stroke. Brain 1983; 106:197—