Charles
A. Jungreis,
MD
Skull-Base Tumors: Ethanol Embolizatlon ofthe Cavernous Carotid Artery’ Tumors of the skull base frequenfly have some blood supply from cavernous branches of the internal carotid artery (ICA). tion of such
Preoperative
completed
after
emboli-
zation. The other five patients had much drier surgical fields than expected. Index terms: Alcohol #{149} Arteries, therapeutic blockade, 172.1299 #{149} Carotid arteries, therapeutic blockade, 172.1299
Radiology
1991; 181:741-743
From the Departments of Radiology and Neurological Surgery, University of Pittsburgh School of Medicine, Presbyterian University Hospital, DeSoto at OHara Sts, Pittsburgh, PA 15213. Received June 10, 1991; revision requested July 3; revision received July 12; accepted July 22. Address reprint requests to the author. C RSNA, 1991 I
Data
Age/Sex
emboliza-
pedicles can expedite the subsequently performed surgery. Sometimes, however, the tumor yessels are multiple and very small so selective catheterizations are not possible, particularly when the tumor invades the cavernous sinus. In eight procedures in seven patients of this type, the tumor was embolized with 100% ethyl alcohol by temporarily occluding the ICA above the feeders while infusing ethanol with a microcatheter as close to the feeders as possible. At fluoroscopy, tumor blush was seen to have decreased markedly in all cases. In one patient,. no obvious benefit was gained at surgery. In another patient, a first surgery that was aborted due to blood loss was successfully
Patient
.
Diagnosis
28/M
Meningioma
31/F
Meningioma
32/M 37/F
Meningioma Meningioma
Neurobogic Increased nerve
weakness Vi already
S
in cranial paralyzed);
nerves HI and IV (cranial transient mild upper
extremity weakness on balloon deflation Transient right cranial nerve VI palsy (first procedure on right ICA); delayed onset of transient paralysis of cranial nerves ifi-VI (second procedure on left ICA) No change Transient
decreased
of ipsilaterab 42/F 45/F 60/F
Change
Meningioma Meningioma Meningioma
sensation
trigeminab
in face
nerve
(maxillary
division
) on balloon
deflation
No change No change No change
tumors sometimes receive a portion of their blood supply from cavernous branches of the internal carotid artery (ICA). For example, the meningohypophyseal trunk frequently feeds tumors of the cavernous sinus and/or the tentonium cenebelli. If catheters can be sebectively placed into a particular feeding vessel, then preoperative embobization can be performed in the usual fashion with relative safety. However, the blood supply to many tumors of this region is in the form of a myriad of small vessels that originate at the cavernous portion of the ICA. These small vessels are too small to be cathetenized individually, but nevertheless they provide a significant supply to the tumor and are a source of considKULL-BASE
enable bleeding at surgery. In an attempt to decrease the intraoperative blood boss, we treated seven patients preopenatively (one bilaterally) by infusing 100% ethyl alcohol through a microcatheter that was positioned as close as possible to the pathologic yessels while the distal ICA was being occluded by a temporarily placed endovascuban balloon.
tion
with
Blood
preservation
supply
stantial
of the
to each
ICA (Table).
tumor
contribution
had
from
the
a sub-
cavernous
portion of the ICA. Two patients had supply bilaterally from the cavernous portion of both ICAs. All patients had some supply from the ipsibateral external carotid artery. In most cases the external supply was embolized in either the same session or during a procedure performed with standard techniques on the previous day. All embobizations were performed in the angiography suite. Patients were awake but sedat-
ed with intravenously administered doses of fentanyl citrate (Elkins-Sinn, Cherry Hill, NJ) and Laboratories,
trated
midazobam Nutbey,
clinically.
Through
an
8-F
B Kit; femoral
catheter
(Introducer
ventional cisco,
sheath
(Radiofocus
Terumo,
troducer into the
Tokyo)
artery,
an
Cabif)
was
portion obtained
activated
tial “dead
8-F
coagulation
time
spaces”
passed
Franinto
of the ICA. A blood for measurement (1).
between
were perfused A nondetachable
was
Inter-
San
coaxiabby
with
All
the samof the
poten-
coaxial
cathe-
heparinized silicon balloon
catheter (model 1505 NDSB loon Catheter; Interventional tics)
introducer Set;
South
passed
In-
placed
Catheter
Therapeutics,
cervical ple was
ters line.
(Versed; Roche that were ti-
NJ)
through
the
sa-
Occlusion Therapeuintroducer
catheter until the balloon was just distal the vessels supplying the tumor. Five
MATERIALS Eight seven
AND
procedures patients
were
scheduled
Bal-
to
METHODS performed for
tumor
in resec-
Abbreviation:
ICA
=
internal
carotid
artery.
741
a.
c.
b.
d.
f.
C.
Images were dimeglumine;
obtained of the first Magnevist; Berlex
procedure Laboratories,
performed Wayne,
on the right ICA of the NJ) MR image (repetition
aged) demonstrates a large skull-base meningioma along the right petrous extends into the ICA and into both cavernous sinuses. (b) Lateral angiogram cavernous
portion
of the
ICA.
The tip of a microcatheter sion
of the
the distal (compare
distal
ICA
(c) Lateral
used with
fluoroscopic
for the ethanol
the
balloon
shows
ICA shows that while some tumor with d). (f) Lateral postembolization
view
infusion extensive
shows
(white tumor
vessels from angiogram
an
arrow) blush.
31-year-old patient. time = 520 msec,
(gadopentetate four signals
bone that compresses the brain stem and cerebellum. The of the right ICA shows multiple tumor vessels originating
endovascular
balloon
is also apparent.
(black
(d) Lateral
(e) Postembobization
the ICA opacified, of the ICA shows
(a) Axial contrast-enhanced echo time = 25 msec, with
arrow)
in position
angiogram
angiogram
of the
to occlude
the
of the ICA obtained ICA
obtained
the flow was almost static, and the tumor that the integrity of the ICA is maintained,
during
blush and
aver-
tumor in the distal
ICA.
during
occlu-
occlusion
of
was decreased the tumor blush
is
decreased.
thousand
units
of heparin
sodium
Pak Laboratories, Franklin administered intravenously placement
of the
crocatheter
(model
Therapeutics, passed
eter
catheter.
the
A mi-
was
then
introducer
cath-
as possible to the abnormal feeders just proximal to the balloon. Dehydrated ethanol (American Regent Laboratories, Shirley, NY) was opacified with metrizamide (Amipaque; Winthrop
the tip was
Target
Calif)
same
as close
Pharmaceuticals,
New
York)
to
a concentration of approximately 240 mg of iodine per milliliter. The balloon was inflated, and Neurobogic
the ICA testing
was occluded. ensued, including
tests of motor and sensory extremities, of visual acuity
function and
in all fields,
of
cranial nerve function, and of memory. If the patient remained stable, the ethanol mixture was infused slowly through the microcatheter with fluoroscopic control to minimize
exposure
nob. Aliquots of ethanol bowed point
742
of the
ICA
of approximately were
to wash occurred
#{149} Radiology
injected into when
tumor. the
etha-
0.1-0.4 slowly
the
to the
tumor
mL
and The
abend
vessels
no
longer
be
clearly
opacified,
and
the ICA could no longer be washed out. The total volume of ethanol injected varied from 1 to 12 mL, with 5-40 minutes required for infusion, excluding the time required to obtain serial angiograms and
Ill) were prior to
Tracker-18;
San Jose,
through
until
balloon
could
(Solo-
Park, just
to reposition
the
catheters.
Prior
to balloon
deflation, 20 mL of saline was infused into the microcatheter while a similar volume was withdrawn from the introducer catheter in an attempt to irrigate the ICA. Control angiography was performed. Infusion of more ethanol after reinflation of the balloon was performed as necessary. At the conclusion of the procedure, protamine sulfate (Eli Lilly, Indianapolis) was administered in a dose titrated according to results of a repeated activated coagubation time test (Figure).
RESULTS Tumor blush depicted in the ICA at angiography decreased on was ebiminated in all cases. Surgical findings were difficult to quantify precisely with respect to the benefit of preoper-
ative embolization. However, in one case a first surgical attempt had been aborted as a result of difficulty in controlling hemorrhage; after embolization, the resection was completed with minimal difficulty. In a second case,
embolization
with
alcohol
re-
subted in the tumor becoming a firm dry mass that remained nonhemonrhagic even when sectioned. Four additional cases seemed far less bloody than expected. In one case, no obvious benefit was gained. Pathologic findings in tumor specimens included scattered areas of necrosis that were possibly secondary to embobization but were not diagnostic for embobization. In no sected. Neurobogically, mained unchanged. ported a subjective tion in the maxillary ipsilaterab trigeminal bess than 1 minute.
case
was
four One
the
ICA
patients patient
ne-
nene-
decrease in division of nerve that One patient
sensathe lasted had a
December
1991
contralateral that basted
of these rally
less
upper than
transient
extremity 5 minutes.
deficits
paresis Both
were
tempo-
related to balloon deflation and presumably embolic in origin.
of considerable
morbidity accompanies surgery. Ideally, preoperative embolization should not have a risk of morbidity or mortality, but to be effective that may not be possible. How-
were One patient had worsening of panesis of the ocubomotor and trochbean nerves (the abducens nerve was already paralyzed by the tumor prior to
ever, any surgical advantage be gained with embolization well worth the risk. As was least
one
the
under-
able
on nearby
Mild
be converted
procedure).
went
bilateral
transient
nerve dune. days
One
patient
embolization.
paresis occurred
of the abducens after the first pnoce-
After
second
the
later,
occurred ethanol
complete
cab cases
procedure
2
ophthabmopbegia
several infusion
minutes after the but resolved in less
than 1 hour; however, the deficit neturned 6 hours later and required 3 months to resolve totally. All patients
experienced nob. The but was
some degree tolerable
travenously and fentanyb.
pain
from
of discomfort with the
administered No deaths
curred. All seven patients quently performed
with cranial embolization,
persisted
postoperatively.
etha-
varied use of in-
midazobam have oc-
survived surgery.
two cases following
the
nerve the
subseIn the
palsy deficit
DISCUSSION Management
of this
group
of pa-
tients is difficult. Extensive masses in the skull base may invade one on both cavernous sinuses, compress the brain stem, encase major vessels and cranial nerves, and be histologically benign. Surgical which multiple
resection of these tumors, must often be performed in stages, involves vascular,
nerve, tients
and
tissue
grafting.
typically present neurobogic impairment,
The with and
pa-
marked the risk
case
from
series,
inoperable
into after
our
that can may be true in at
manageable
tumors
mopermay
surgi-
embolization.
In a previously published article, performance of alcohol infusion after permanent occlusion of the distal ICA was described (2). In contrast, our senies ICA
consisted of patients in whom the was to be preserved, and therefore we only temporarily occluded the vessel during embolization. In addition, in our patients the infusions were performed with a microcatheter to limit the exposure of the ICA to ethanol as much as possible. Our choice of a liquid embobization agent also reflects the fact that at the conclusion of the procedure the balloon was deflated, and the blood flow in the ICA was reestablished. Particulate embolization has the additional theoretic risk that some particles might remain in the ICA despite thorough irrigation and might cause a cerebral infarction after balloon deflation. The two cases of transient neunobogic deficit in our series were rebated temporally to balloon deflation and were most likely embolic in origin, probabby secondary to the induction of small aggregates of blood products by the ethanol. The long-term effect of ethanol on the ICA is not clear. In studies of canine renal arteries, tissue necrosis and endothebiab damage with secondary vascular thrombosis were observed in areas exposed to concentrated ethanob but not in vessels exposed to di-
luted ethanol (3,4). No delayed neurologic deficits have occurred in our group of patients in the 9 months since the first embolization was penformed. Quantification of effect is difficult. Dramatic decreases in tumor blush have been demonstrated at fluonoscopy. One case in which a repeat angiognam was obtained 2 days after embobization showed no evidence of revasculanization. Performance of a controlled study to compare the results of surgery performed with and without embolization is conceivable but would be difficult. The impression of the surgeons has been favorable regarding preoperative embolization because of the achievement of a reduction in expected blood boss, a drier than expected surgical field, and a savings in time needed to perform surgery. In conclusion, we emphasize that appropriate patient selection is crucial since ethanol embolization of skullbase tumors involving the cavernous ICA has substantial risk. Nevertheless, significant benefit may be gained. U
References 1.
Scott JA, Berenstein A, Blumenthal D. Use of the activated coagulation time as a measure of anticoagubation during interventional procedures. Radiology 1986; 158:849850.
2.
Lasjaunias roangiography.
ment 3.
4.
P. Berenstein A. Surgical Vol 2. Endovascular
of craniofacial
lesions.
neutreat-
New York:
Springer-Verlag, 1987; 62-63. ElIman BA, Green CE, Eigenbrodt E, GarriottJC, Curry TS. Renal infarction with absolute ethanol. Invest Radiol 1980; 15:318322. Buchta K, Sands J, Rosenkrantz H, Roche WD. Early mechanism of action of arteri-
ally infused ization.
alcohol
Radiology
U.S.P. in renal 1982;
devital-
145:45-48.
Radioboev
#{149} 743