Eur Radiol (2002) 12:2717–2722
DOI 10.1007/s00330-001-1293-y

D. Ducreux
M.C. Petit-Lacour
K. Marsot-Dupuch
J. Bittoun
P. Lasjaunias

Received: 16 August 2001
Revised: 8 November 2001
Accepted: 21 November 2001
Published online: 19 March 2002
© Springer-Verlag 2002
D. Ducreux (✉) · M.C. Petit-Lacour
K. Marsot-Dupuch · P. Lasjaunias
Department of Neuroradiology,
CHU de Bicêtre, Paris XI University,
78 rue du Général Leclerc,
94270 Le Kremlin-Bicêtre, France
e-mail: denis.ducreux@bct.ap-hop-paris.fr
Tel.: +33-1-45212601
Fax: +33-1-45212317
J. Bittoun
CIERM (U2R2M-CNRS),
CHU de Bicêtre, Paris XI University,
78 rue du Général Leclerc,
94270 Le Kremlin-Bicêtre, France

NEURO

MR perfusion imaging in a case of cerebral
proliferative angiopathy

Abstract Proliferative angiopathy is
an uncommon type of cerebral arteriovenous shunt characterized by an
extensive capillary network with
normal brain intermingled and few
clinical symptoms (mostly seizures
and headaches). This case report depicts an extensive proliferative angiopathy located in the right hemisphere. Its hemodynamic disturbances were studied with MR perfusion imaging (local or remote areas
of increased time to peak, decrease
ratio of signal, and relative regional
cerebral blood volume values).

Keywords Magnetic resonance
imaging · Intracranial arteriovenous
malformations · Perfusion

Introduction

Case report

Proliferative angiopathy (PA) is an uncommon pial arteriovenous shunt (AVS) that combines diffuse cortical
network with moderately enlarged veins compared with
what is seen in a true brain arteriovenous malformation
[1]. This very particular kind of microcapillary network
may be associated with seizure [1], but there is a lack of
physiological evidence depicting the way it acts on adjacent cortex. Many hypotheses have been proposed [2, 3,
4]; one of them was about the local or regional hemodynamic disturbances induced by the arteriovenous shunt
[2]. This hypothesis may be enlightened by MR gadolinium-perfusion imaging. We describe a case of primitive
proliferative angiopathy without embolization, imaged
1 year after the onset of symptoms with qualitative and
semi-quantitative MR perfusion parameters: time to peak
(TTP), decrease ratio of signal (DR), and regional relative cerebral blood volume (rrCBV).

A 36-year-old woman without clinical background nor actual therapy was admitted to the neurological care unit for a first partial
then generalized seizure onset. After seizure, the patient had no
headaches, and the neurological examination was normal. Magnetic resonance imaging was performed, which demonstrated an extensive capillary network located in the right frontal lobe, with
moderately enlarged veins (Fig. 1). This typical AVS pattern was
further analyzed with angiography (Fig. 2). The middle-sized
AVS, extended in the right subcortical area, was fed by cortical,
lenticulostriate, and the Heubner arteries. The post-bifurcation M1
right segment was occluded, and the right hemisphere was supplied by this proliferative angiopathic network. The angiostructure
was drained by an enlarged inferior ventricular vein flowing in the
Galen vein. The patient was rapidly discharged with heparin and
Depakin (antiepileptic) therapy (no endovascular treatment). As
annual follow-up, the patient underwent both clinical and imaging
examinations. The MR imaging was performed using a 1.5-T
scanner, with two sequences: a 5-mm slice thickness with no interslice gap, 24×24 cm2 field of view (FOV), and 28 axial slices on
T1-weighted spin-echo sequence (256×256 matrix, TR=500 ms,
TE=15 ms) was performed, followed by a 5-mm slice thickness,

2718

Fig. 1a–c T1-weighted spin-echo MR multiplanar reconstructions
centered on the proliferative angiopathy (red dot). a Axial and
b coronal reconstruction showing a dense vascular network (red
dot) centered on the anterior basal ganglia, with ventricles deformation. c Sagittal reconstruction showing the enlarged inferior
ventricular vein (arrowhead) compressing the right ventricle and
the right thalamus area. The size of the draining vein is here not a
discriminative feature of the pathology

0.5-mm interslice gap, 24×24 cm2 FOV, and 18 axial slices per
volume single-shot gradient-echo echo-planar imaging sequence
(96×64 matrix, TR=2.4 s, TE=38 ms) to perform the MR perfusion
bolus tracking, with the sequence started 3 s after the Gd-DTPA injection. Twenty milliliters of Gd-DTPA (0.2 mmol/kg) was manually injected for 3 s in a left antecubital vein, immediately followed by a 20-cc saline flush at the same rate. The entire brain
volume was imaged every 2.4 s for 47 s after injection of GdDPTA (20 volumes in total). No smoothing was performed on the
images. The first acquired bolus tracking volume was used only
for display purposes since its magnetization contrast was not yet
saturated by the pulse repetition. Time zero was assigned to the
second acquired volume. All images were transferred to an offline
workstation for processing using programs written in Delphi 5
(Borland Inprise Inc.). The MR signal intensity was converted to
gadolinium concentration using the equation Cm(t)=–K.ln[S(t)/S0],
where Cm(t) is the measured concentration of Gd-DTPA with respect to time. K is a proportionality constant depending on the

magnetic field, type of tissue, type of contrast agent, and echo
time of imaging sequence which can be further simplified for
computation. S(t) is the MRI signal intensity with respect of time,
and S0 is the baseline MRI signal before the presence of Gd-DTPA
and after a steady-state magnetization has been achieved. Three
images were averaged to calculate S0 (second, third, and fourth acquired bolus tracking volumes). Cm(t), maps, and regions of interest of TTP, DR, and rrCBV in brain tissue were calculated on a
voxel-by-voxel basis. The TTP was established by calculating the
time between the first T2*-weighted measurement and the bolus
peak (the highest Cm(t) curves value) [5]. The signal DR was calculated with the equation: DR=S(TTP)/S0, where S(TTP) is the
signal value of the time to peak, and S0 is the baseline signal value
averaged on second, third, and fourth acquired bolus tracking volumes. The rrCBV was calculated with the indicator dilution theory
[6] by integrating the Cm(t) curves with the following equation:

where κ is a constant which corrects for the fact that hematocrit in
large vessels is larger than the hematocrit of the small vessels [7],
and ρ is the density of the brain tissue. These constants were not
considered for further computation because only relative values
(in arbitrary units) were recorded. Four mirror regions of interest
of 49 voxels (172 mm2) were drawn in the right basifrontal AVS
area and in the normal-appearing contralateral hemisphere, and remotely from the AVS in the coronae radiatae.

2719

Fig. 2a–d Bilateral carotid angiography. a Right internal carotid selective angiography, coronal view, showing occlusion
of its distal segments from the
post-bifurcation M1 segment
(arrow). Slightly opacified nidus (arrowhead). b Left internal carotid selective angiography, coronal view, showing the
contralateral right lenticulostriate and Heubner arteries (arrow) feeding the nidus (arrowhead). c Left internal carotid
selective angiography, coronal
view, showing late opacification of the nidal area (arrow)
and persistent right middle cerebral artery distal occlusion.
d Right internal carotid selective angiography, sagittal view,
showing slightly opacified right
inferior ventricular draining
vein (arrow). This enlarged
vein is here quite unusual in
proliferative angiopathy

Table 1 Results of measurements. ROI region of interest
Location

ROI

Time to peak
(s, mean±SD)

Decrease signal ratio
(%, mean±SD)

Relative regional
cerebral blood volume
(AU, mean±SD)

Right basifrontal hemisphere
Left basifrontal hemisphere
Right corona radiata
Left corona radiata

1
2
3
4

14.7912±0.896
16.7016±0.843
21.552±1.341
19.248±0.343

98.492±0.949
24.227±9.583
40.992±9.816
31.123±14.942

50.55±11.841
4.642±2.417
12.694±4.389
6.248±4.614

Results
The results of measurements are reported in Table 1.
In the right basifrontal AVS area, brain parenchyma
appeared as edematous areas of T1-weighted spin-echo
hypointensities and T2-weighted echo-planar imaging
hyperintensities. Small vessels appeared hypointense on
both T1- and T2-weighted sequences. In this large right
region decreased TTP values, increased DR, and increased rrCBV covered the whole AVS area compared
with those of contralateral left hemisphere which was
taken as reference (Fig. 3).

Apart from AVS (coronae radiatae), no significant
changes were observed on both conventional T1- and
T2-weighted MR imaging. Conversely, we observed a
wide right parietal area of increased TTP, DR, and
rrCBV values compared with those of contralateral left
hemisphere which was taken as reference (Fig. 4).

Discussion
Proliferative angiopathy was individualized from the cerebral AVM group in 1989 [1]. Proliferative angiopathy is

2720

Fig. 3a–d Magnetic resonance
perfusion imaging in the PA nidus area. a T2*-weighted echoplanar image showing right basifrontal hydric hyperintensities
mixed with vascular hypointensities (arrow). Enlarged draining vein (arrowhead), which is
quite unusual in PA. b Time to
peak (TTP) mapping showing
in the nidus region a wide anterior area of slight decreased
(green), and few posterior areas
of slight increased (red) TTP
values. c Decrease ratio of signal (DR) mapping showing the
wide area of increased DR values (red). d Regional relative
cerebral blood volume (rrCBV)
mapping showing the matching
DR area of increased rrCBV
values (green and red). Severe
increase in rrCBV values in the
internal parietal area suggesting
the presence of the enlarged
draining vein (arrow)

thought to be an uncommon singular AVS with an extensive capillary network, occurring in young adults. Clinical onset of symptoms are the most common with seizures (72.2%), but headaches (27.8%), neurological deficit (44.4%), or bleeding (5.6% stand alone, with recurrent
bleeding of 27.8%) may also occur (R. Stendel, unpublished data). Diagnosis is suspected on enhanced cerebral
CT Scan or MR study which showed a dense parenchymal vascular network and was confirmed on angiography.
The angiostructure combines angiogenesis and angioectasias, and there is a discrepancy between the “nidus-like”
network of vessels and the normal or slightly enlarged
appearance of the draining veins [1]. The natural history
is poorly known but often associated with progressive occlusion of cerebral arteries. In that case, angiogenesis (angioectasia and transdural angiogenesis) supply the local
brain vascularization (Fig. 2). The hemodynamic behavior of these PAs was demonstrated on angiography and
shows a rapid venous filling usually due to faster capillary transit time and slower-than-true AVM. What seems
peculiar in that case are the perfusion abnormalities

which extended far beyond the boundaries of the morphological abnormalities seen on conventional MR sequences. In the nidal and the para-nidal right frontal area, the
abnormalities in perfusion parameters were of similar
features and were quite expected because of the shunt effect (decreased TTP and increased rrCBV values) and the
angioectasia (increased rrCBV values). As previously reported, microcapillaries (Fig. 2b) affect the Cm(t) contrast vs time course variation [8, 9]. Remote from the nidus was a right parietal area which was undetectable on
both conventional MR imaging and angiography. This area was highlighted by the slightly increased TTP values
probably due to the M1 arterial occlusion. These local
and remote perfusion abnormalities may explain the main
clinical presentation (seizures) as cortical or subcortical
hemodynamic dysregulation which could act as an irritative spike [10], without reported evidence of local cortical ischemia [11, 12]. The minor clinical expression of
PA over time illustrates the tremendous plasticity of the
brain hemodynamic [13, 14]; thus, embolization of such a
fragile system may have disastrous consequences and en-

2721

Fig. 4a–d Magnetic resonance
perfusion imaging on the corona radiata level (remotely from
the nidus). a T2-weighted
echo-planar image without any
signal abnormality. b The TTP
mapping showing a wide right
parietal area of slightly increased TTP values (red) probably due to the right M1 occlusion, and prolonging the abnormality seen on Fig. 3b. The c
DR and d rrCBV maps showing a tiny right frontal area of
increased DR and rrCBV values (green and red) which does
not match the TTP mapping of
right parietal area increased
TTP values. Strong increase in
rrCBV values may be due to
marginal abnormal microvessels which are invisible in a

dovascular treatment should only be used with targeted
objectives. Moreover, our timely restricted vision of PA is
unable to describe all the processes implied in brain hemodynamic regulation which might quickly change from
one state to another. Performing MR perfusion imaging in
a larger cohort of patients with similar conditions may
highlight a threshold beyond which a major risk of seizure could be documented.

Conclusion
There are many structural and hemodynamic cerebral abnormalities in proliferative angiopathy. In this case of
clearly depicted PA disturbed perfusion parameters (especially TTP) showed local and remote dysregulations with
seizure as the main clinical symptom. Larger studies are
necessary to highlight a possible focal cortical ischemia.

References
1. Lasjaunias P, Berenstein L, Ter Brugge
KG (eds) (1997) Surgical neuroangiography, vol 1. Springer, Berlin Heidelberg New York, pp 14–22
2. Gao E, Young WL, Ornstein E, PileSpellman J, Ma Q (1997) A theoretical
model of cerebral hemodynamics: application to the study of arteriovenous
malformations. J Cereb Blood Flow
Metab 17:905–918

3. Kader A, Young WL (1996) The effects of intracranial arteriovenous malformations on cerebral hemodynamics.
Neurosurg Clin North Am 7:767–781
4. Hacein-Bey L, Nour R, Pile-Spellman
J, Van Heertum R, Esser PD, Young
WL (1995) Adaptive changes of autoregulation in chronic cerebral hypotension with arteriovenous malformations:
an acetazolamide-enhanced single-photon emission CT study. Am J Neuroradiol 16:1865–1874

5. Neumann-Haefelin T, Wittsack HJ,
Wenserski F et al. (1999) Diffusionand perfusion-weighted MRI. The
DWI/PWI mismatch region in acute
stroke. Stroke 30:1591–1597
6. Meier P, Zierler LL (1954) On the theory of the indicator-dilution method for
measurement of blood flow and volume. J Appl Physiol 6:731–744

2722

7. Axel L (1980) Cerebral blood flow determination by rapid-sequence computed tomography. Radiology
137:679–686
8. Fukuda Y, Murata Y, Umehara I,
Yamashita T, Ono C, Iwai T, Shibuya
H (1999) Perfusion and blood pool
scintigraphy for diagnosing soft-tissue
arteriovenous malformations. Clin
Nucl Med 24:232–234
9. Yamada S, Thio S, Iacono RP, Yamada
BS, Brauer FS, Hayward W, Morgese
VJ, Moghtader M (1993) Total blood
flow to arteriovenous malformations.
Neurol Res 15:379–383

10. Weinand ME (2000) Vascular stent
model of human temporal lobe epileptogenicity: the relationship between
electrocorticographic interhemispheric
propagation time and cerebral blood
flow. Med Hypotheses 54:717–720
11. Meyer B, Schaller C, Frenkel C,
Schramm J (1998) Physiological steal
around AVMs of the brain is not equivalent to cortical ischemia. Neurol Res
20 (Suppl 1):S13–S17
12. Meyer B, Schaller C, Frenkel C,
Ebeling B, Schramm J (1999) Distributions of local oxygen saturation and its
response to changes of mean arterial
blood pressure in the cerebral cortex
adjacent to arteriovenous malformations. Stroke 30:2623–2630

13. Takeshita G, Toyama H, Nakane K,
Nomura M, Osawa H, Ogura Y, Katada
K, Takeuchi A, Koga S, Kato Y et al.
(1994) Evaluation of regional cerebral
blood flow changes on perifocal brain
tissue SPECT before and after removal
of arteriovenous malformations. Nucl
Med Commun 15:461–468
14. Young WL, Pile-Spellman J,
Prohovnik I, Kader A, Stein BM
(1994) Evidence for adaptive autoregulatory displacement in hypotensive cortical territories adjacent to arteriovenous malformations. Columbia University AVM Study Project. Neurosurgery
34:601–611