J Neurosurg 80:422-432, 1994
The natural history of familial cavernous malformations:
results of an ongoing study
JOSEPH M. ZABRAMSKI, M.D., THOMAS M. WASCHER, M.D., ROBERT F. SPETZLER, M.D.,
BLAKE JOhnSON, M.D., JOaN GOLrINOS, M.D., BURTON P. DRAYER, M.D.,
BEN BROWN, M.D., DANIEL RIGAMONTI, M.D., AND GERALDINE BROWN, R.N., M.S.
Divisions of Neurological Surgery and Neuroradiolo83; Barrow Neurological Institute, St. Joseph's
Hospital, Phoenix, Arizona; Division of Neuroradiology, Kaiser Permanente Hospital, Portland,
Oregon; and Division of Neurological Surgery, University of Maryland Hospital, Baltimore, Maryland
~" Cavernous malformations are congenital abnormalities of the cerebral vessels that affect 0.5% to 0.7% of
the population. They occur in two forms: a sporadic form characterized by isolated lesions, and a familial
form characterized by multiple lesions with an autosomal dominant mode of inheritance. The management
of patients with cavernous malformations, particularly those with the familial form of the disease, remains
a challenge because little is known regarding the natural history.
The authors report the results of an ongoing study in which six families afflicted by familial cavernous
malformations have been prospectively followed with serial interviews, physical examinations, and magnetic
resonance (MR) imaging at 6- to 12-month intervals. A total of 59 members of these six families were screened
for protocol enrollment; 31 (53%) had MR evidence of familial cavernous malformations. Nineteen (61%)
of these 31 patients were symptomatic, with seizures in 12 (39%), recurrent headaches in 16 (52%), focal
sensory/motor deficits in three (10%), and visual field deficits in two (6%). Twenty-one of these 31 patients
underwent at least two serial clinical and MR imaging examinations. A total of 128 individual cavernous
malformations (mean 6.5 -+ 3.8 lesions/patient) were identified and followed radiographically. During a mean
follow-up period of 2.2 years (range 1 to 5.5 years), serial MR images demonstrated 17 new lesions in six
(29%) of the 21 patients; 13 lesions (10%) showed changes in signal characteristics, and five lesions (3.9%)
changed significantly in size. The incidence of symptomatic hemorrhage was 1.1% per lesion per year.
The results of this study demonstrate that the familial form of cavernous malformations is a dynamic disease;
serial MR images revealed changes in the number, size, and imaging characteristics of lesions consistent with
acute or resolving hemorrhage. It is believed that the de novo development of new lesions in this disease has
not been previously reported. These findings suggest that patients with familial cavernous malformations require
careful follow-up monitoring, and that significant changes in neurological symptoms warrant repeat MR imaging. Surgery should be considered only for lesions that produce repetitive or progressive symptoms. Prophylactic resection of asymptomatic lesions does not appear to be indicated.
KEY WORDS
cavernous malformation 9 angioma 9 cerebrovascular malformation
familial malformation 9 hereditary disorder
magnetic resonance imaging
EREBROVASCULAR malformations are developmental abnormalities that affect the blood vessels supplying the brain. They include venous
malformations, arteriovenous malformations (AVM's),
cavernous malformations, and telangiectases. 43 The
results of postmortem studies suggest that approximately 4% of the population harbor such lesions. 27,z9,3~
Although cavernous malformations make up only 8%
to 15% of cerebrovascular malformations reported
in postmortem studies, they are responsible for a disproportionately larger percentage of the clinical
symptomatology caused by this group of vascular 1esions. 12,23,27,30,43 Symptomatic lesions usually present
c
422
9
with evidence of recent hemorrhage in association with
seizures when located above the tentorium, and with
focal neurological deficits (or much more rarely, death)
when located in the posterior f o s s a . 10,15,24,26,4~,49,50 They
are the second most common vascular malformation
identified in reviews of surgical pathology, being outnumbered only by A V M ' s ? 3
Cavernous malformations occur in two forms: a sporadic or nonhereditary form, in which patients tend to
have a single isolated lesion, and a familial form characterized by the presence of multiple lesions and an
autosomal dominant pattern of inheritance. Prior to the
introduction of magnetic resonance (MR) imaging, fa-
3. Neurosurg. / Volume80 / March, 1994
Natural history of familial cavernous malformations
milial cavernous malformations were thought to be
rare. 2.5,1~
In a review of the clinical data from our
institution, Rigamonti, et aL, 38 found that 54% of patients with cavernous malformation have a strong family history consistent with a hereditary pattern of development. The familial form appears to occur more
frequently in the Hispanic population2 ,18,25,36,38
When an isolated cavernous malformation is identified in a symptomatic patient and the lesion is readily
accessible, surgical excision is recommended by most
authors. When lesions are located in deep or eloquent
structures or are discovered incidentally, management
is more controversial. The problem is complicated by
the lack of data on the natural history of cavernous
malformations. We have previously identified a series
of families with the inherited form of cavernous malformations. 38 This group of patients, many with multiple incidental lesions, provides a unique opportunity
to study the natural history of these lesions.
In this report we describe the results of a prospective
study of the members of six unrelated Arizona families
with the diagnosis of familial cavernous malformations. The patients were followed with serial clinical
interviews, physical examinations, and MR images.
Information regarding genetic transmission, growth
rate, incidence of hemorrhage/rehemorrhage, MR signal changes, and frequency of neurological symptoms
related to familial cavernous malformations was obtained and analyzed.
TABLE 1
Magnetic resonance (MR) imaging classification for
cavernous malformation
Lesion
Type
Type I
Type 1I
MR Signal Characteristic*
T~:hyperintense core
T~:hyper- or hypointense
core with surrounding
hypointense rim
T6 reticulated mixed signal core
T2:reticulated mixed signal core with surrounding hypointenserim
Pathological
Characteristics
subacute hemorrhage, surrounded by a rim of hemosiderin-stained macrophages & gliotic brain
loculated areas of hemorrhage & thrombosisof
varying age, surrounded
by gliotic, hemosiderinstained brain; in large lesions, areas of calcification may be seen
chronic resolved hemorrhage, with hemosiderin
staining within & around
the lesion
q~pe III Tt: iso- or hypointense
T2:hypointensewith a
hypointense rim that
magnifies the size of
the lesion
GE: hypointensewith
greater magnification
than T:
Type IV Tj: poorly seen or not
two lesions in the category
visualized at all
have been pathologically
1"2:poorly seen or not
documented to be telangivisualized at all
ectasias
GE: punctate hypointense
lesions
* T~ and T2 denote T~- and Tz-weightedMR images, respectively;
GE = gradient-echosequences.
Clinical Material and Methods
During the 5~ years from July 1, 1986, to January
1, 1992, six unrelated families with familial cavernous
malformation were followed prospectively with serial
interviews, comprehensive neurological examinations,
and MR imaging studies of the brain. These six families
were originally identified at the Barrow Neurological
Institute as a result of investigation of at least one
symptomatic propositus in each family. A family was
considered to harbor familial cavernous malformations
if at least two members of the same family had an MR
image that demonstrated the characteristic appearance
of cavernous malformation(s) (Table 1). All family
members in an identified family were eligible for the
protocol and were invited to participate. Asymptomatic
individuals with an initial negative MR imaging study
were excluded from further analysis. Any patient with
an initial MR image that demonstrated at least one cavernous malformation underwent follow-up comprehensive interviews and physical examinations at 6- to 12month intervals. Inquiries were specifically directed at
the new onset of or changes in existing seizure disorder,
headache, visual difficulties, weakness or sensory abnormalities, or other neurological complaints. Any neurological sign or symptom was documented and followed prospectively. Follow-up data were defined in
terms of "lesion-years," determined by multiplying the
number of cavernous malformations by the follow-up
period in years for each patient.
Serial MR images of the brain were obtained at 6J. Neurosurg. / Volume 80 / March, 1994
to 12-month intervals coincident with interviews and
physical examinations. All MR images were obtained
on one of two identical machines (1.5-tesla MR system). Two spin-echo pulse sequences were routinely
obtained in the axial and sagittal planes; Tt-weighted,
spin-echo, sagittal (TR 500 to 700 msec, TE 16 to 20
msec) and axial (TR 700 to 800 msec, TE 20 msec)
images, as well as intermediate- and T2-weighted (TR
2500 to 2800 msec, TE 30 or 90 msec) images were
obtained. Gradient-echo sequences (TR 750 to 800
msec, TE 15 or 30 msec, 0 = 20 ~ completed the MR
studies. Imaging was routinely performed with a field
view of 20 or 24 cm, an acquisition matrix of 256 x
192, an image thickness of 5 mm, and two excitations
with a total imaging time of approximately 15 minutes.
The MR images were reviewed by two neuroradiologists (B.P.D. and B.J.) who were unaware of the
patients' clinical history. To permit accurate follow-up
monitoring and to eliminate discrepancies in descriptive terminology, each lesion seen on an MR imaging
study was characterized according to its size, location,
and signal characteristics. A change in size greater than
0.3 cm on subsequent studies was defined as significant. The appearance of new lesions was also recorded
and characterized according to signal characteristics.
For the purposes of this study, individual cavernous
malformations were divided into four categories on the
423
J. M. Zabramski, et al.
FIG. 1. Magnetic resonance (MR) Tl-weighted (A and B) and T2-weighted (C and D) imaging sections
obtained 1 year after surgery in a 45-year-old Hispanic woman (Patient B3 in Fig. 4), who presented with
the sudden onset of double vision and severe headache. The MR images at the time of her initial presentation
revealed an approximately 1.5-cm acute hemorrhage in the rectum of the left midbrain, as well as multiple
other hemorrhagic lesions of varying size and age. The MR findings and a strong family history of seizures
were believed to be consistent with a diagnosis of familial cavernous malformations. Gross total resection
of the midbrain lesion was carried out using a supracerebellar-infratentorial approach. Pathological evaluation
confirmed the diagnosis of cavernous malformation. Sections from this follow-up MR imaging again reveal
multiple cavernous malformations including three Type I lesions (straight arrows) and three Type III lesions
(curved arrows). In addition, the cystic postoperative remnant of the resected brain-stem lesion is readily apparent (arrowhead). Note that the areas of subacute hemorrhage that characterized Type I lesions are hyperintense on Tcweighted images, but may be hyper- or hypointense on T2-weighted images depending on the
age of the hemorrhage (see the Clinical Material and Methods section). Type III lesions (chronic hemorrhage)
are iso- or hypointense on Tt-weighted images, but are always hypointense on T2-weighted spin-echo sequences
because of the presence of hemosiderin and ferritin. In contrast, note that the MR imaging signal of the cystic
postoperative defect in the midbrain follows cerebrospinal fluid signal intensity: the small halo of hypointensity
that surrounds the cyst on the T2-weighted images betrays its hemorrhagic past. This pattern of cystic change
may be seen in areas of chronic resolved hemorrhage of other etiologies; however, it is not characteristic of
lesions in patients with intact cavernous malformations.
basis of pathological correlations and MR signal characteristics (Table 1), as described more fully below.
Type I Lesions
Type I lesions are characterized by the presence of
subacute hemorrhage, and are therefore frequently well
visualized on computerized tomography (CT) scans.
Type I lesions (Fig. 1) exhibit a hyperintense core on
Tl-weighted MR images due to the presence of methemoglobin? ,16,17 Signal intensity on T2-weighted images is initially hyperintense; however, as the hematoma ages, methemoglobin is rapidly broken down and
converted to hemosiderin and ferritin. The longer echo
times used to generate T2-weighted spin-echo images
make them more suspectible to the paramagnetic effects of these breakdown products. This process occurs
first along the margins of the hematoma, decreasing
the size of the hyperintense core and producing a hypointense halo around the lesion on T2-weighted scans.
This effect is well illustrated in the left medial temporal
lobe cavernous malformation seen in Fig. 1, where the
lesion appears hypointense on the T2-weighted scan,
despite a hyperintense core on the Tl-weighted image.
In this grading system, lesions are considered to be subacute, or Type I, until the MR signal on Tl-weighted
images becomes iso- or hypointense with the surrounding brain.
424
Type H Lesions
Type II lesions are characterized pathologically
by loculated areas of thrombosis and hemorrhage of
varying age, surrounded by gliosis and hemosiderin
staining. Large lesions may contain areas of organized
thrombus, with calcification. On CT scans, these lesions are poorly visualized or are not seen at all. Occasionally, CT may reveal an area of stippled calcification. Magnetic resonance images of Type II lesions
(Fig. 2) demonstrate a mixed, or reticulated, core of
high and low signal intensity, surrounded by a hypointense ring that has become almost pathognomonic of
these lesions in the literatureY 7,37 The reticulated MR
appearance of these lesions combined with the relatively common finding of calcium suggests that thrombosis and hemorrhage within the lesion are an ongoing,
repetitive process.
Type 111 Lesions
Pathologically, Type III lesions (Figs. I and 2) correspond to chronic areas of hemorrhage in which the
presence of residual hemosiderin in and around the
cavernous malformation produces a markedly hypointense lesion on T2-weighted and gradient-echo images. 16 On Tl-weighted images, Type III lesions may
be iso- or hypointense to the surrounding brain.
J. Neurosurg. / Volume 80 / March, 1994
Natural history of familial cavernous malformations
FIc. 2. Magnetic resonance (MR) Tl-weighted (A and B) and T2-weighted (C and D) follow-up images
in a 10-year-old Caucasian boy (Case A16 in Fig. 4). This child was asymptomatic when initially screened
in June, 1989, at 8 years of age. An MR image at that time demonstrated four foci consistent with cavernous
malformation, including the two lesions demonstrated here in November, 1991. The small hypointense Type
III lesion (straight arrow, A and C) in the medial right temporal lobe and the much larger variegated Type
II lesion (curved arrow, B and D) in the left temporal-occipital region have remained essentially unchanged.
Failure of the left temporo-occipital lesion to resolve to a more chronic form suggests continued (re)hemorrhage
and/or thrombosis within the cavernous malformation. Complex partial seizures developed at 9 years of age
and have been clearly localized to this left-sided lesion.
Results
Patient Characteristics
F-he. 3. Gradient-echo magnetic resonance images in a 51year-old Hispanic man (Case C5, Fig. 4) revealing multiple
hypointense lesions "too numerous to count" consistent with
cavernous malformations. The Type IV lesions are the small
punctate areas of low signal intensity, several of which have
been arbitrarily indicated by arrows. Two Type IV lesions at
the authors' institution have been confirmed pathologically to
be capillary telangiectasias, 39 suggesting that these lesions
form a continuum in the development of the more typical cavernous malformation.
Type IV Lesions
Type IV lesions are poorly visualized on both T~- and
T2-weighted MR images, and are seen best with a
gradient-echo sequence as small, punctate hypointense
foci (Fig. 3). The MR signal characteristics suggest that
minute deposits of hemosiderin are present within the
lesions. Whether these lesions represent minute cavernous malformations or a histologically distinct precursor remains controversial. At Barrow Neurological
Institute, two lesions in this category have been verified
pathologically as capillary telangiectasias. 39
J. Neurosurg. / Volume 80 / March, 1994
The series consisted of 118 surviving members (64
males and 54 females) of six unrelatedArizona families
known to harbor familial cavernous malformations
(Fig. 4). Five of these six families were of Hispanic
descent. Of these patients, 59 were located and were
willing to undergo a screening evaluation. In this group,
MR imaging studies revealed evidence of at least one
cavernous malformation in 31 patients (53%), including 13 (48%) of the 27 males and 18 (56%) of the 32
females. Multiple lesions were present in 26 (84%) of
these 31 patients, while five patients harbored only one
lesion.
Of these original 31 patients, 2 t agreed to participate
in the long-term natural history study. These included
10 males and 11 females, with a mean age of 25 years
(range 7 to 51 years). A total of 128 individual cavernous malformations were identified in this group,
ranging in number from one lesion in five patients to
"too numerous to count" in two patients (Fig. 3), who
were excluded from the numerical analysis because
their lesions could not be reliably enumerated and differentiated on follow-up studies. In the remaining patients, there were an average of 6.5 _+ 3.8 lesions per
patient (Table 2). The follow-up period for this group
averaged 2.2 -+ 1.2 years (range 1.0 to 5.5 years), which
corresponded to a mean follow-up time of 46 patientyears, or 282 lesion-years.
Initial Clinical Findings
Of the original 31 patients with positive MR imaging
studies, 19 (61%) had a history of seizure activity, severe headache, hemiparesis, hemisensory deficits, or
425
J. M. Zabramski, et al.
Fl6. 4. Pedigrees of six families with familial cavernous malformations. In the six pedigrees shown, circles
denote females and squares denote males. Open symbols denote normal subjects, and solid symbols denote
patients with at least one cavernous malformation on magnetic resonance (MR) imaging. Symbols with diagonal
lines represent deceased patients. Arrows designate the propositus in each family. A horizontal bar over a symbol
indicates that pathological confirmation of the disorder was obtained. Symbols with solid centers indicate that
the patient described a history of seizures, severe headaches, hemiparesis or hemisensory deficits, or visual
difficulties; symbols with horizontal bars indicate that the patient exhibited an abnormal neurological examination. The 21 patients with MR imaging and clinical follow-up monitoring described in detail in the text
are designated by a letter-number combination, such as A3.
426
J. Neurosurg. / Volume 80 / March, 1994
Natural history of familial cavernous malformations
TABLE 2
TABLE 3
Magnetic resonance imaging findings in 31 cases
of familial cavernous malformations*
Summary of neurological symptoms and signs in 31
patients with familial cavernous malformations*
]No. of
Lesions
No.
1
5
Cases
Percent
2
2
3
5
4
4
5-10
11
> 10
4
* Mean 6.5 + 3.8 lesions per patient.
16
6
16
I3
35
13
visual obscuration (Table 3). Eight (26%) of these patients had a distinctly abnormal neurological examination. Twelve patients (39%) were completely asymptomatic with normal neurological examinations. Five
(16%) of the 31 had undergone previous surgery (four
intracranial, one intraspinal) for the removal of histopathologically verified cavernous malformations.
Of the 21 patients with complete follow-up data, 15
Cases
Symptoms & Signs
No.
asymptomatic
12
symptomatic
19
recurrent headaches
16
seizures
12
hemisensory/motordeficits
3
visual field deficits
2
no neurologicalsigns
23
neurologicalsigns
8
hyperreflexia/Babinskisigns
5
ataxic dysmetria
3
nystagmus
2
hemiparesis
2
hemisensory deficit
2
intellectualimpairment
2
optic atrophy,Marcus-Gunnpupil
1
* A patient may have more than one sign or symptom.
Percent
39
61
52
39
10
6
74
26
16
10
6
6
6
6
3
(71%) were symptomatic and seven (33%) exhibited
neurological abnormalities. Most commonly observed
were seizures, seen in 48%. The remainder of this report is focused on these 21 patients.
Clinical Course, Rate of Hemorrhage, Surgical
Results
Two patients experienced exacerbation of their seizure disorder and one previously asymptomatic patient
developed seizures during the follow-up period. These
three patients were found to have rebled (Fig. 5) into
a supratentorial cavernous malformation. All three patients underwent craniotomy for resection of the symptomatic lesion without complication, at which time the
diagnosis of cavernous malformation was confirmed
by pathological examination.
The rate of symptomatic hemorrhage in this series
of familial cavernous malformations was 6.5 % per patient per year, or 1.1% per lesion-year. Four additional
patients who were asymptomatic or minimally symptomatic (complaint of headaches) had M R evidence of
recurrent hemorrhage during the follow-up period, corresponding to an overall hemorrhage rate of 13% per
patient per year, or 2% per lesion-year.
Findings on MR Imaging
Of the 128 distinct cavernous malformations identified in the 21 patients, 118 (92%) were supratentorial
and 10 (8%) were infratentorial (Table 4). Gradient refocused images were more sensitive than spin-echo images in detecting cavernous malformations, especially
those with hypointense signal characteristics. Familial
cavernous malformations most commonly involved the
frontal lobes, where 38% were located at the gray-white
J. Neurosurg. / Volume 80 / March, 1994
Fro. 5. Follow-up magnetic resonance T 1- (left) and T2(righO weighted images in a 31-year-old Caucasian woman
(Case A3 in Fig. 4). The patient first experienced generalized
tonic-clonic seizures at 10 years old; her family history was
remarkable for seizures in two of her five siblings. Serial MR
images demonstrated a large cavernous malformation in the
left insular cortex with the now almost classic MR appearance
of a mixed signal intensity core surrounded by a ring of hypointensity (Type II lesion). Her seizures were relatively well
controlled until the spring of 1991, when there was an acute
exacerbation. Repeat MR imaging, shown here, demonstrated
findings consistent with an area of subacute hemorrhage
within the insular mass, with a region of hyperintense signal
(arrows) on both the %- and Tz-weighted images. Violent almost daily seizures persisted despite maximum medical
therapy, and the patient underwent craniotomy with complete
resection of a multiloculated cavernous malformation. Postoperative mild right hemiparesis and expressive aphasia have
resolved completely, and her seizures are presently well controlled on a single medication.
matter junction. Signs and symptoms were seen almost
exclusively in patients with evidence of subacute or
chronic resolving hemorrhage on MR images (Type I
or Type II lesions). In fact, 14 (93%) of the 15 patients
with Type I or Type II lesions were symptomatic.
427
J. M. Zabramski, et al.
TABLE 4
Location and (vpe qf cavernous ma!fi)rmations ident!~ied in 2I study patients with complete data*
Lesion
Location
frontal
temporal
parietal
occipital
basal ganglia/internal capsule
thalamus
parieto-occipital
cerebellum
midbrain
ports
medulla
corpus callosum
totals (%)
* MR
MR Type I
MR Type 11
MR Type IV
Right
Left
Right
Left
Right
Left
Right
Left
3
0
2
0
0
0
0
0
6
2
0
1
1
0
0
0
4
0
0
I
0
0
0
0
2
3
1
I
1
0
0
0
7
9
I
8
0
0
1
I
7
10
2
2
0
1
0
4
13
1
7
1
O
2
0
O
6
2
3
0
0
1
1
0
0
3
0
3
21 (16%)
0
0
0
0
13 (10%)
0
0
0
2
55 (43%)
0
2
0
0
39 (31%)
Total
No.
Percent
48
37
27
21
16
]2
14
t1
2
2
4
3
2
2
5
4
0
0
5
4
0
0
5
4
128 (100%)
magneticresonanceimaging. For definitionof lesion type see Table I.
MR Imaging Appearance of New Lesions
During the follow-up period, six (29%) of the 21
patients developed a total of 17 new lesions not visualized on earlier MR imaging studies. These "new"
lesions occurred at an overall frequency of 0.4 lesions
per patient-year. Of the 17 "new" cavernous malformations, 11 (65%) identified by MR imaging were
small, hypointense, chronic (Type III) lesions, while
the remaining six lesions contained varying amounts
of subacute hemorrhage. All but one of these new lesions were clinically silent.
Changes in Lesion Size on MR Imaging
A change in size was observed in five lesions in four
patients. An average 0.8-cm increase in size was seen
in four lesions due to recurrent hemorrhage, and was
symptomatic in three of the four lesions. One cavernous malformation that initially presented with MR evidence of subacute hemorrhage decreased 1.0 cm in size
over 2~- years without evidence of recurrent hemorrhage.
MR Imaging Signal Intensity Changes
Signal intensity changes were observed on MR imaging in 13 lesions (10%) in eight (38%) of the 21 patients. Overall, six (4.7%) of the 128 lesions demonstrated repeat hemorrhage on MR imaging during the
follow-up period. In three patients, MR evidence of
recurrent hemorrhage was associated clinically with
the exacerbation of seizures (in two) or new seizure
occurrence (in one). There were no instances of a hyperintense malformation becoming hypointense over
the follow-up period.
Associated lntracranial Abnormalities
Four of the original 31 patients with MR imaging
positive for cavernous malformations had evidence of
4:98
MR Type III
associated benign intracranial lesions. Signal characteristics were consistent with a pericallosal lipoma in
one patient, a small lipoma of the quadrigeminal plate
in one, a venous malformation in the frontal region in
one, and a left cerebellar venous angioma in one.
Discussion
Cavernous malformations represent one of four
common cerebrovascular malformations. In postmortem studies they have been found to affect 0.5% to 0.7%
of the population5 3.29 On gross inspection, cavernous
malformations appear as a discrete, lobulated, wellcircumscribed, mulberry-like lesion that may vary in
diameter from several millimeters to several centimeters. 23,29,43 The histopathological features of these
lesions have been well described. 23,28,29,43Microscopically they are characterized by the appearance of
abnormally dilated, sinusoidal vascular channels without intervening brain parenehyma. The wails of the dilated sinusoidal spaces are composed of a single layer
of flattened endothelial cells, without smooth muscle,
and are separated from each other by collagenous hyalinized or fibrous tissue. Elastic fibers in the walls of
the vascular spaces are characteristically absent.
Thrombosis, organization and inflammatory changes,
and occasional calcification may be seen in larger lesions.2.28A2.43
Evidence of prior hemorrhage is a nearly constant
feature of cavernous malformations. These lesions are
thought to grow and produce symptoms by recurrent
episodes of hemorrhage. Hemorrhage is characteristically confined within the lesion, and produces neurological deficits secondary to local mass effect rather
than direct parenchymal injury. Smaller, nonsymptomatic episodes of hemorrhage are thought to contribute to the development of seizures. These smaller
hemorrhages result in the progressive deposition of
hemosiderin in the cerebral parenchyma surrounding
the cavernous malformation. Iron, which is present
J. Neurosurg. / Volume80 / March, 1994
Natural history of familial cavernous malformations
in hemosiderin, is a well-known epileptogenic mate
rial that is used to induce seizures in laboratory models
of epilepsy.21,~5-53Indeed, seizures are the most common symptom in patients with cavernous malformations, accounting for 40% to 60% of the presenting
complaintsJ~.3s.474,~51 (39% in the present series). Other
common presenting symptoms include recurrent headaches, acute hemorrhage, and focal neurological deft
cits from m a s s e f f e c t . 152~,26,47,4'),51 Rarer presentations
include hydrocephalus, cranial neuropathies such as
trigeminal neuralgia, papilledema, hypothalamic disturbances, and progressive or transient neurological
deficits, j5,33,43,45 Symptomatic lesions have been described in all age groups, with most studies indicating
an onset of symptoms between the third and fifth
decades without a sex predominance. 24,2(''3~',5tClinically, cavernous malformations are thought to compose a
significant percentage of "angiographic occult cerebral
vascular malformation. '4,~',7,'~,~1,123722,31,32,46,48,52
The hallmarks of familial cavernous malformations
are the characteristic findings on MR imaging of multiple cavernous malformations combined with a positive family history of seizures. These vascular lesions
seem to be more prevalent among families of Hispanic
(Mexican-American) heritage. 1,1~-~-5,3~',3sThe first report
suggesting a familial occurrence for these lesions was
by Kufs in 1928. 2o Subsequent reports indicated that
the familial form of cavernous malformations was
likely inherited as an autosomal dominant trait. 2,~~
The findings in the present study are consistent with
this mode of inheritance; approximately 50% of all offspring had MR evidence of the disease, males and females were equally affected, and the disease was transferred maternally or fraternally to male offspring with
no skipped generations.
With the introduction of MR imaging, it has become
apparent that familial cavernous malformations are
more common than previously postulated. In our institution, approximately 50% of patients with symptomatic cavernous malformation are found to have a
family history consistent with familial inheritance. 3~
Clinical suspicion of the familial disease should be
raised by the finding of multiple lesions, which were
present in 84% of patients in this study compared to
the 10% to 15% incidence of multiplicity reported in
sporadic c a s e s . 8,4f~'zl,43
Previously, the natural history of familial cavernous
malformations has been obscure. This study indicated
the dynamic nature of familial cavernous malformation
with regard to clinical manifestation, number of lesions, and size and signal characteristics on MR imaging. The majority of affected patients (approximately
60%) are symptomatic. As with the sporadic form of
this disease, the most common presenting symptoms
of patients with familial cavernous malformations in
this series were recurrent headaches (52%) and seizures (39%).
Change in lesion size and signal intensity (seen in
19% and 38% of patients, respectively) indicates that
familial cavernous malformations are not static lesions.
They tend to increase in size secondary to episodes of
J. Neurosurg. / Volume 80 / March, 1994
hemorrhage into the malformation, and may decrease
in size during periods of quiescence as resolution of
the hemorrhage occurs. In the present series the incidence of symptomatic hemorrhage averaged 1.1% per
lesion-year for patients with familial cavernous malformation. This compares favorably to the hemorrhage
rate of 0.7% reported by Robinson, et al. 4c~but is considerably higher than the 0.1% rate reported by Curling,
e l al.; s this latter group estimated the rate of symptomatic hemorrhage in their study population on the
basis of a retrospective historical review of symptoms,
a method that is likely to lead to underestimation. Age
may also play a factor as hemorrhage has been reported
to be more common in younger patients? 4~,4~,5~,51The
average age of patients in the present study was 25
years versus 35 and 38 years, respectively, for the patients in the reports by the two groups above,s,4~ Alternatively, the risk of hemorrhage may be inherently
higher in the familial form of this disease.
An unexpected finding was the appearance of "new"
lesions on serial MR images. New lesions consistent
with cavernous malformation appeared in six (29%) of
21 patients, at a rate of 0.4 new lesions per patient per
year. To our knowledge, the development of new lesions has not been reported previously. New lesions
may represent the growth of a very small nidus due
to capillary proliferation or focal hemorrhage, or some
combination of these two f a c t o r s . 34'56 Another possibility is that these "new" lesions were in fact present,
but not identified on earlier studies. This might occur
in attempts to image small lesions, as a result of partial
volume-averaging or interslice spacing. If this hypothesis were true, one would expect by the same reasoning
that lesions identified during the protocol would have
an equal likelihood of disappearing on follow-up MR
imaging; however, none of the initial 128 lesions disappeared on follow-up images.
Determinants of clinical severity appear related to
size and location as well as appearance on MR imaging.
Cavernous malformations located within the brain stem
can pursue an especially aggressive c o u r s e . 13'54'56 In the
present study, patients harboring hyperintense lesions
on spin-echo MR images (lesions with evidence of subacute or mixed-age hemorrhage: Type I or Type II lesions) were almost uniformly symptomatic (14 (93%)
of 15 patients), while only two (33%) of the six patients
with exclusively hypointense lesions on MR images
(Type III or Type IV lesions) demonstrated symptomatology. The hyperintense familial cavernous malformations also showed a tendency to symptomatic recurrent hemorrhage, resulting in exacerbation of
seizures and/or neurological deficits related to mass effect. In addition to primary genetic abnormality, other
factors such as blood pressure, metabolites (especially
hemoglobin degradation products), or hormonal or enzymatic levels may have some role in lesion growth
and in determining the clinical severity of cavernous
malformations.~8,44
Although it might be expected that lesions demonstrating hyperintense MR signal characteristics would
429
J. M. Z a b r a m s k i , et al.
exhibit a hypointense signal over time as a result of
the natural evolution and resorption of blood to hemosiderin, this has not been the case thus far. None of the
34 Type I or Type II hyperintense (subacute or mixedage) lesions observed in this study converted to a Type
III hypointense (chronic) lesion during a mean follow-up period of 2.2 years. This finding provides additional evidence that, once hemorrhagic, the majority of
cavernous malformations continue to bleed intermittently.
Cavernous malformations should be considered in
the differential diagnosis of a patient presenting with
new onset of seizures, intracranial hemorrhage, or focal
neurological deficit. This consideration is especially
important if the patient is of Hispanic descent and has
a family history of seizures. When more than one firstdegree relative has a cavernous malformation, genetic
counseling and serial follow-up monitoring consisting
of physical examinations and MR imaging (which must
include gradient-echo imaging), should be offered to
all family members. 1~
Establishing a diagnosis in
asymptomatic family members permits delineation of
potential risks, identifies the requirement for close
follow-up review, and indicates the need for appropriate investigation in the event of onset of neurological
symptoms.
Previous recommendations to "excise all surgically
accessible lesions" are not practical in cases of familial
cavernous malformation with multiple lesions. Surgical intervention in patients with multiple lesions is
recommended for clinically significant hemorrhage,
uncontrollable seizures, or progressive neurological
deterioration. It appears that asymptomatic patients
with familial cavernous malformations do not require
prophylactic surgery, since the morbidity and mortality
associated with hemorrhage from these lesions (except
those in the brain stem) are relatively low compared
to that from an aneurysm or AVM. Surgical extirpation
of symptomatic brain-stem cavernous malformations
appears to be the treatment of choice when the lesion
is located superficially and an operative approach can
spare eloquent tissueJ 3,55,56
Conclusions
Although this cohort of patients with familial cavernous malformations has been followed for only a
relatively short period of time, several important conclusions can be drawn. They are as follows.
1. Familial cavernous malformations are dynamic lesions that change with regard to number, size, and signal characteristics on MR imaging. Signal changes correlate with episodes of hemorrhage and clinical signs
and symptoms.
2. Symptoms are most commonly associated with
hyperintense lesions (those that show evidence of recent sizable hemorrhage). The hyperintense appearance may indicate a predisposition to a more aggressive
clinical course. These lesions tend to rehemorrhage,
both clinically and subclinically. Although rarely cata430
strophic, significant hemorrhage was usually symptomatic (seizures and/or focal neurological deficits).
Likewise, exacerbation of symptoms was typically associated with MR imaging evidence of recent (subacute) hemorrhage.
3. The importance of high-quality MR imaging utilizing gradient-refocused imaging to diagnose and follow patients at risk cannot be overstated. History and
.physical examination alone are inadequate for screenmg afflicted family members because more than onethird of patients harboring lesions are asymptomatic.
4. Due to the dynamic nature of these lesions, until
the natural history of cavernous malformations is better
defined, repeat MR imaging for symptomatic individuals at 12-month intervals is recommended for routine
follow-up monitoring. Magnetic resonance imaging
should be offered to all family members at risk. Until
more is known, no specific recommendations can be
made regarding follow-up evaluation of asymptomatic
patients.
These findings represent the results of an ongoing
prospective study of the natural history of familial cavernous malformations. Further follow-up study will
better define the changes that occur in this disease process. In this fashion, the subset of patients at greatest
risk of morbidity and mortality can be identified, thus
delineating a rational comprehensive plan for appropriate management.
Acknowledgment
The authorsgratefullyacknowledgethe assistance of the Barrow Neurological Institute Editorial Office in preparing this
manuscript.
References
l. BicknellJM: Familialcavernousangiomaof the brainstem
dominantly inherited in Hispanics. Neurosurgery 24:
102-105, 1989
2. Bicknell JM, Carlow TJ, Kornfeld M, et al: Familial cavernous angiomas. Arch Neurol 35"746-749, 1978
3. BradleyWG Jr, Schmidt PG: Effect of methemoglobinformation on the MR appearanceof subarachnoidhemorrhage.
Radiology 156:99-103, 1985
4. Chin D, Harper C: Angiographically occult cerebral vascular malformationswith abnormal computedtomography.
Surg Neurol 20:138-142, 1983
5. Clark JV: Familial occurrence of cavernous angiomata of
the brain. J Neurol Neurosurg Psychiatry 33:871-876,
1970
6. Cohen HCM, Tucker WS, Humphreys RP, et al: Anglographically cryptic histologicallyverified cerebrovascular
malformations. Neurosurgery 10:704-714, 1982
7. CrawfordJV, RussellDS: Crypticarteriovenousand venous
hamartomas of the brain. J Nearol Neurosurg Psychiatry
19:1-11, 1956
8. Curling OD Jr, Kelly DL Jr, Elster AD, et al: An analysis
of the natural history of cavernousangiomas. J Neurosurg
75:702-708, 1991
9. Davis DH, Kelly PJ: Stereotactic resection of occult vascular malformations. J Neurosurg 72:698-702, 1990
J. Neurosurg. / Volume 80 / March, 1994
Natural history of familial cavernous malformations
10. Dobyns WB, Michels VV, Groover RV, et al: Familial cavernous malformations of the central nervous system and
retina. Ann Neurol 21:578-583, 1987
11. Ebeling JD, Tranmer BI, Davis KA, et al: Thrombosed arteriovenous malformations: a type of occult vascular malformation. Magnetic resonance imaging and histopathological correlations. Neurosurgery 23:605-610, 1988
12. E1-Gohary EM, Tomita T, Gutierrez FA, et al: Angiographically occult vascular malformations in childhood. Neurosurgery 20:759-766, 1987
13. Fahlbusch R, Strauss C, Huk W, et al: Surgical removal of
pontomesencephalic cavernous hemangiomas. Neurosurgery 26:449-457, 1990
14. Farmer JP, Cosgrove GR, Villemure JG, et al: Intracerebral
cavernous angiomas. Neurology 38:1699-1704, 1988
15. Giombini S, Morello G: Cavernous angiomas of the brain.
Account of fourteen personal cases and review of the literature. Acta Neurochir 40:61-82, 1978
16. Gomori JM, Grossman RI, Goldberg HI, et al: Intracranial
hematomas: imaging by high-field MR. Radiology 157:
87-93, 1985
17. Gomori JM, Grossman RI, Goldberg HI, et al: Occult cerebral vascular malformations: high-field MR imaging. Ra.
diology 158:707-713, 1986
18. Hayman LA, Evans RA, Ferrell RE, et al: Familial cavernous angiomas: natural history and genetic study over a
5-year period. Am J Med Genet lh147-160, 1982
19. Kidd HA, Cumings JN: Cerebral angJomata in an Icelandic
family. Lancet 1:747-748, 1947
20. Kufs H: Uber heredofamili~re Angiomatosc des Gehirns
und der Retina, ihre Beziehungen zueinander und zur Angiomatose der Haut. Z Gesamte Neurol Psychiatr 113:
651-686, 1928
21. Lange SC, Neafsey El, Wyler AR: Neuronal activity in
chronic ferric chloride epileptic foci in cats and monkey.
Epilepsia 21:251-254, 1980
22. Lobato RD, Perez C, Rivas JJ, et al: Clinical, radiological,
and pathological spectrum of angiographically occult intracranial vascular malformations. Analysis of 21 cases
and review of the literature. J Neurosurg 68:518-531,
1988
23. Martin N, Vinters H: Pathology and grading of intracranial
malformations, in Awad IA, Barrow DL (eds): lntracranial
Vascular Malformations. Neurosugical Topics. Park
Ridge, Ill:American Association of Neurological Surgeons,
1990, pp 1-30
24. Martin NA, Wilson CB, Stein BM: Venous and cavernous
malformations, in Wilson CB, Stein BM (eds): lntracranial
Arteriovenous Malformations. Baltimore: Williams &
Wilkins, 1984, pp 234-245
25. Mason I, Aase JM, Orrison WW, et al: Familial cavernous
angiomas of the brain in an Hispanic family. Neurology
38:324-326, 1988
26. McCormick PC, Michelsen W J: Management of intracranial cavernous and venous malformations, inAwad IA, Barrow DL (eds): Intracranial Vascular Malformations.
Neurosurgieal Topics. Park Ridge, II1:American Association of Neurological Surgeons, 1990, pp 197-217
27. McCormick WF: Classification, pathology, and natural history of angiomas of the central nervous system. Weekly
Update: Neurol Neurosurg 1(14):1978
28. McCormick WF: The pathology of vascular ("arteriovenous") malformations. J Neurosurg 24:807-816, 1966
29. McCormick WF: Pathology of vascular malformations of
the brain, in Wilson CB, Stein BM (eds): Intraerauial
Arteriovenous Malformations. Baltimore: Williams &
J. Neurosurg. / Volume80 / March, 1994
Wilkins, 1984, pp 44-63
30. McCormick WF: Vascular disorders of nervous tissue:
anomalies, malformations, and aneurysms, in Bourne GH
(cd): Structure and Function of Nervous Tissue. New
York: Academic Press, 1969, pp 537-596
31. McCormick WE Hardman JM, Bouller TR: Vascular malformations Cangiomas'') of the brain, with special reference
to those occurring in the posterior fossa. J Neurosurg 28:
241-256, 1968
32. McCormick WE Nofzinger JD: "Cryptic" vascular malformations of the central nervous system. J Neurosurg 24:
865-875, 1966
33. Mizutani T, Goldberg HI, Kerson LA, et al: Cavernous hemangioma in the dieneephalon, Arch Neuro138:379-382,
1981
34. Pozzati E, Giuliani G, Nuzzo G, et al: The growth of cerebral cavernous angiomas. Neurosurgery 25:92-97, 1989
35. Reid SA, Sypert GW, Boggs WM, et al: Histopathology of
the ferric-induced chronic epileptic focus in cat: a Golgi
study. Exp Neurol 66:205-219, 1979
36. Rigamonti D: Natural history of cavernous malformations,
capillary malformations (telangiectases), and venous malformations, in Awad IA, Barrow DL (eds): Intracrauial
Vascular Malformations. Neurosurgical Topics. Park
Ridge, Ill: American Association of Neurological Surgeons,
1990, pp 45-51
37. Rigamonti D, Drayer BP, Johnson PC, et al: The MR[ appearance of cavernous malformations (angiomas). J Neurosurg 67:518-524, 1987
38. Rigamonti D, Hadley MN, Drayer BF, et al: Cerebral cavernous malformations. Incidence and familial occurrence.
N Engl J Med 319:343-347, 1988
39. Rigamonti D, Johnson PC, Spetzler RE et al: Cavernous
malformations and capillary telangiectasia: a spectrum
within a single pathological entity. Neurosurgery 28:
6ff454, 1991
40. Robinson JR, Awad IA, Little JR: Natural history of the
cavernous angioma. J Neurosorg 75:709-714, 1991
41. Robinson JR Jr, Little JR, Awad IA: The natural history of
cavernous angiomas. J Neurosurg 72:333A, 1990 (Abstract)
42. Runnels JB, Gifford DB, Forsberg PL, et al: Dense calcification in a large cavernous angioma. Case report. J Neurosurg 30:293-298, 1969
43. Russell DS, Rubinstein LJ: Pathology of Tumours of the
Nervous System, ed 5. Baltimore: Williams & Wilkins,
1989
44. Rutka JT, Brant-Zawadzki M, Wilson CB, et al: Familial
cavernous malformations. Diagnostic potential of magnetic
resonance imaging. Surg Neurol 29:467-474, 1988
45. Saito N, Yamakawa K, Sasaki T, et al: Intramedullary cavernous angioma with trigeminal neuralgia: a case report and
review of the litcrature. Neurosurgery 25:97-101, 1989
46. Shuey HM Jr, Day AL, Quisling RG, et al: Angiograpbically
cryptic cerebrovascular malformations. Neurosurgery
5:476-479, 1979
47. Simard JM, Garcia-Bengochea F, Ballinger WE Jr, et al:
Cavernous angioma: a review of 126 collected and 12 new
clinical cases. Neurosurgery 18:162-172, 1986
48. Steiger HJ, Tew JM Jr: Hemorrhage and epilepsy in cryptic
cerebrovascular malformations. Arch Neurol 14:722-724,
1984
49. Tagle P, Huete I, M~ndez J, et al: lntracranial cavernous
angioma: presentation and management. J Neurosurg 64:
720-723, 1986
50. Vaquero J, Leunda G, Martinez R, et al: Cavernomas of the
431
J. M. Zabramski, et al.
brain. Neurosurgery 12:208-210, 1983
51. Voigt K, Ya~argil MG: Cerebral cavernous haemangiomas
or cavernomas. Incidence, pathology, localization, diagnosis, clinical features and treatment. Review of the literature
and report of an unusual case. Neurochirurgia 19:59-68,
1976
52. Wakai S, Ueda Y, Inoh S, et al: Angiographically occult
angiomas: a report of thirteen cases with analysis of the
cases documented in the literature. Neurosurgery 17:
54%556, 1985
53. Willmore LJ, Sypert GW, Munson JB: Recurrent seizures
induced by cortical iron injection: a model of posttraumatic
epilepsy. Ann Neurol 4:32%336, 1978
54. Yoshimoto T, Suzuki J: Radical surgery on cavernous angioma of the brainstem. Surg Neurol 26:72-78, 1986
55. Zentner J, Hassler W, Gawehn J, et al: Intramedullary cav-
432
ernous angiomas. Surg Neurol 31:64-68, 1989
56. Zimmerman RS, Spetzler RE Lee KS, et al: Cavernous malformations of the brain stem. J Neurosurg 75:32-39, 1991
Manuscript received March 23, 1993.
Accepted in final form August 4, 1993.
This work was supported in part by Arizona Disease Control
Research Commission Contract 82-1703.
Address for Dr. Brown: Division of Neuroradiology, Kaiser
Permanente Hospital, Portland, Oregon.
Address for Dr. Rigamonti: Division of Neurological Surgery,
University of Maryland Hospital, Baltimore, Maryland.
Address reprint requests to: Joseph M. Zabramski, M.D.,
c/o Editorial Office, Barrow Neurological Institute, 350 West
Thomas Road, Phoenix, Arizona 85013.
J. Neurosurg. / Volume 80 / March, 1994