Pediatric Lori L. Barr, MD #{149} Diane S. Babcock, MD William S. Ball, MD #{149} Em C. Prenger, DO

Color Doppler Neurosurgical

U

(2-4). has

adult

From the (K.R.C.,

cine,

Elland

and

Received February 24. Address reprint ‘ RSNA, 1991

Bethesda

12, 1991; requests

the

ious clinical conditions.

utility

AND

for either neurosurgery intervention. CDI

16 intraoperative dures

and

one

one mycotic malformation. defect

cysts, two ventricular

aneurysm, In three was

the transducer

not

large

needed

to admit

to perform

Cincinnati,

OH

CDI;

thus, those patients were omitted from further evaluation. Either an Ultramank (Advanced Technology Laboratories, Bothell, Wash) or Acuson 128 (Acuson, Mountain View, Calif) system was used

during system

the procedures. employed either

phased-array

size

The Ultramark a 5.0-MHz

scanning

of 10 mm

transducer

Acuson

head

with

or a 5.0-MHz with

a 40-mm

128 system

was

45229-2899.

March

From

15; revision

the

1990

received

RSNA

June

to the prepared

and

equipment before it

donning

a gown

and

gloves

technologist posiThe transducer

and cord were covered with a sterile sheath (Civco Surgi-cover 914623; Cone Instruments, Solon, Ohio) into which sten-

cases,

timal dunes.

Scanning

was per-

orthogonal planes through defect or open fontanelle. a separate

by the

burr

hole

neurosurgeon

US scanning during The neurosurgeon

drainage moistened

dura

or panenchyma

sterile

thus

providing

with

an

was

to allow

acoustic

saline,

interface

which The

op-

procethe for

was conradiologist

with on

a color

video

videotape and review.

printer

field

or ne-

for subsequent phoTotal US scanning

time was recorded. Scanning was penformed through the anterior fontanelle during the percutaneous embolization. In each case, patient age, preoperative imaging

results,

length

of intraoperative

scanning time, postoperative sults, and pathologic diagnosis viewed to assess whether the

imaging were addition

the color maps to the B-scan images 9

9

either structural identified with

as to whether of the

rereof

provided

or flow information not other imaging techniques. If provided additional flow data

the CDI maps or structural detail, course

an

US

on a an

operating room. The for surgery by scrub-

the sonography the US machine.

an assessment

information

the

was

altered

made

the

surgery.

aper-

linear-

RESULTS

aperture.

equipped

and Pediatrics (L.L.B., D.S.B., W.S.B., E.C.P.) and Medical Center, University of Cincinnati College

revision requested to L.L.B.

Chiari I the calva-

enough

while honed

corded tography

nonneoplastic obstruction,

and one patients,

bing

printed

pnoce-

intracranial percutaneous treatment of six ceremalformations, four tumors,

two arachnoid conditions,

rial

sched-

for the

brovascular

was moved radiologist

trans-

size of 28 mm transducer with

scanned while a sonography technologist operated the controls. The neunosurgeons observed or manipulated the surgical as necessary. The images were either

one

embolization

phased-array

the covered transducer, trolled by the radiologist.

or neunoradiused during

neurosurgical

a 3.0-MHz

aperture size of 38 mm. The was cleaned with disinfectant

In two

pathologic

was

either

ducen with an aperture 7.5-MHz linear-array

opened

in var-

performed children

with

formed in two the craniotomy

METHODS

was seen

MD

ite get was applied.

of

of CDI

and

A prospective study with 20 consecutively uted ologic

(CDI)

evaluation

situations

MATERIALS

and

malformations and neoplasms with children

experience

demonstrates

The

Ayes,

imaging

aneurysms,

Our

approach resection

in the

arteriovenous

(5,6).

181:567-571

Departments of Radiology T.S.B.), Children’s Hospital

Doppler

useful

(AVMs),

tune

I

Color been

surgical surgical

S. Berger,

Pediatric Procedures’

LTRASOUND

fies the optimal aids in complete

array

gery

#{149} Thomas

(US) imaging has become a common method for intraoperative evaluation of the central nervous system. Real-time monitoting aids in the ventricular insertion of catheters and drainage of cysts (i,2). Localization and characterization of intracerebral and intraspinal masses with US before resection clari-

Index terms: Brain, US studies, 10.12983, 10.12984, 10.12985, 10.12986 #{149} Brain neoplasms, US studies, 10.36, 10.12983, 10.12984, 10.12985, 10.12986 #{149} Cerebral blood vessels, 172.434, 172.75 #{149} Ultrasound (US), Doppler studies Ultrasound (US), in infants and children #{149} UItrasound (US) guidance

1991;

MD

US Imaging during and Neuroradiologic

Experience with color Doppler imaging (CDI) during 16 pediatric intraoperative and one percutaneous neuroradiologic procedures was reviewed to assess whether CDI increased the success rate or decreased the procedure time, thus contributing substantially to procedure performance. Intraoperative CDI was used to rapidly identify abnormal vessels or displacement of normal vessels and correlated with preoperative studies. In six cases (four vascular malformations, one mycotic aneurysm, and one hamartoma), surgical resection was altered on the basis of flow information obtained. In one case of percutaneous embolization of a Galenic malformation, CDI provided information contributing to the cessation of the procedure. In six cases (debulking of three gliomas, resection of one vascular malformation, and two biopsies of nonneoplastic conditions), information was added but did not alter the surgical approach. In the remaining four cases (three cerebrospinal fluid drainage procedures, one posterior fossa decompression), no additional information was obtained. Consultation among the ultrasound staff, neuroradiologists, and neurosurgeons before the operative procedure maximized the usefulness of CDI, thus aiding in the success of surgery.

Radiology

R. Crone,

#{149} Kerry

Radiology

Neurosurof Medi-

scientific

20; accepted

The results of the study are outlined in the Table. The surgical or interventional approach was altered on

assembly.

June Abbreviations: mation,

CDI

AVM = arteriovenous Doppler imaging.

malfor-

= color

567

Cases

in Which

Type or

CDI Was Performed

of Lesion

Patient

Procedure

Vascular

Age

lesions

Tumors

Nonvascular

masses

Drainage

procedures

12 3 3 14 6 17 1

y y y y d y y

3 5 7 2 4 13 9 7 7 7

y y y y y y y y y mo

Diagnosis

Site*

AVM AVM Cavernous angioma Cavernous angioma Galenic malformation Mycotic aneurysm Venous angioma, cavernous changes Glioma Glioma Glioma Extraaxial hamartoma Nonspecific inflammation Nonspecific inflammation Arachnoid cyst Arachnoid cyst Chiani I malformation Obstructed dysmorphic

Scanning Time (mm)

Procedure

R frontal R temporal L frontal L panietooccipital Vein of Galen R occipital L cerebellum

Resection Resection

< 10

Resection Resection Percutaneous Resection Resection

< 10

Brain stem Brain stem Brain stem R frontal Conus L frontal R temporal Quadrigeminal Craniocervical R lateral

Debulking Debulking Debulking Resection Biopsy Biopsy Shunt revision Laser fenestration Decompression Laser fenestration

plate

Usefulness of CDIt + + + + + + +

< 10

embolization

Preoperative Studies

< 10 < 10

4- +

< 10 > 30

+ + + +

< 10

+

< 10 < 10

+ +

10-20 < 10 20-30 20-30 > 30 < 10 > 30

+ +

MR, anglo CT, MR. anglo CT, MR Anglo, MR US, MR. anglo CT, anglo CT, MR CT, MR CT, CT, CT, CT, CT CT, MR CT,

+ +

0 0 0 0

MR MR MR MR. MR

US

MR MR.

US

ventricle L = left, R = right.

*

0 = no information anglo = angiography, AVM = arteriovenous

Figure 1. (a) Lateral phy of internal carotid

displacement the

mass

exists

added, + = information added CT = computed tomography,

along

of the

the

unaltered,

+ + = information

added

and

surgical

approach

altered.

malformation.

view during angiograartery demonstrates

of the sylvian effect

but surgical approach MR = MR imaging.

triangle

hematoma.

due A tiny

superopostenior

margin

to AVM of

the hematoma (arrows). (b) Anteropostenior view during angiography demonstrates the AVM along the superomedial margin of the hematoma (arrows). After removal of the bone flap, US was performed with a 7.5-MHz

linear

array

positioned

along

the night

tern-

poral dura (D). (c) Axial sonogram reveals the AVM nidus (A) 2-3 cm deep to the dural surface (D). The intraparenchymal hematoma (H) is hypoechoic. The position of the AVM along the superopostenior portion of the mass was confirmed (anterior is to the left). (d) Coronal sonogram correlates with the preoperative angiogram in b, demonstrating the AVM (A) superomedial to the hematoma (H).

a.

the basis of color seven of i7 cases. cases

were

and

the

Doppler findings Six of the seven

vascular

seventh

b.

in

abnormalities,

was

a nonvascular

mass lesion. A 3-year-old child had severe aches after minor head trauma.

headAt CT

examination, an acute intraparenchymat hemorrhage was seen in the right temporal lobe. Magnetic resonance (MR) imaging revealed prominent vessels gestive

medial of an

vealed medial

a small margin

ia, ib). After sonography

to the hematoma AVM. Angiography

sugre-

AVM draped along the of the hematoma (Fig removal of a bone flap, was performed through

the dura. Within a short period, the exact location of the small AVM was identified with CDI in all three planes 568

#{149} Radiology

c.

d. November

1991

thus was not resected. The location of the remaining aneurysm was determined at head CT, which was marred by artifact.

a.

b.

Figure

[repetition time msec/echo time msec]) MR imvein (arrowhead) and arteries supplying the AVM nidus (arrows). angiography of the night internal carotid artery after selective injection arteries (arrowheads) supplying the AVM. There is early filling of the draining veins (arrows). (c) Lateral view during angiography of the right internal carotid artery obtained with selective injection after embolization demonstrates decreased flow into the inferopostenior aspect of the malformation. (d) Sagittal CDI obtained through the surgical defect, with the dura intact, demonstrates residual draining veins (anrows) superficial to the echogenic mass. The hypoechoic focus (c) demonstrating posterior shadowing in the inferopostenior portion of the mass represents a coil used for embolization. age

2. (a) Coronal demonstrates flow

Ti-weighted (1,600/20 void in a large draining (b) Lateral view during demonstrates multiple

(Fig ic, id) and subsequently confirmed at intraoperative

was angiogra-

phy.

tion. At surgery, a venous angioma was found and hemostasis was difficult. At the end of the initial exploration, it was difficult to determine whether resection was complete. Intraoperative CDI demonstrated a fo-

Subtracted

angiography

performed in the anteroposterior and lateral planes demonstrated doubling of the size of the residual aneurysm, which was near the surgical clips (Fig 4a). Inability to locate the lesion previously in the operating room prompted the use of CDI before the dura was opened. The most direct course to the aneurysm was visualized, as well as a thrombosed portion of the aneurysm that was not apparent on the angiogram (Fig 4b). Resection was subsequently successful. In a 2-year-old child referred for repair of hypertelorism, a nasofrontal cephalocele was found at CT (Fig 5a) and MR imaging (Fig 5b). A normal gyral pattern in the right frontal lobe was not identified. Clear cleavage planes separated the mass from the remainder of the brain. Whether this represented a congenital brain anomaly or an extraaxial neoplasm was uncertain before surgery. After the dura was opened, the cleavage plane was not apparent to the neurosurgeon because of the size of the lesion. Both the morphologic appearance and the appearance of the mass at B-mode US demonstrated little difference between the mass and the adjacent, normat

left

frontal

lobe

of the

brain.

CDI

readily demonstrated feeding arteries to the hamartoma from the anterior and middle cerebral arteries, as well as displacement

of the

normal

ante-

nor cerebral arteries (Fig 5c, 5d). Since the middle cerebral artery and anterior cerebral arteries were known to lie within the cleavage planes on the basis

of preoperative

intraoperative vessels

imaging

identification

allowed

exact

studies,

of the

localization

of

cerebellar hemisphere was demonstrated at MR imaging (Fig 3a, 3b). Before surgery, it was uncertain whether the mass represented a cystic

had three known mycotic aneurysms, with previous surgical resection of two of the aneurysms through two separate craniotomies. The remaining aneurysm could not be localized dur-

the epicenter of the mass, which facilitated focused laser cavitation. CDI was also useful in percutaneous embolization performed in a 6-day-old neonate with severe congestive heart failure in whom a Galenic malformation was seen at US of the head. Arteriography was performed, and embolization followed. Both B-mode scanning and CDI were employed during the embolization. Videotaping of the real-time study demonstrated the coils and partial filling of the dilated vein of Galen by clotted blood and extruded embolic material. A substantial decrease in flow within the vein of Galen was confirmed at CDI, adding information that supported cessation of the procedure. In six cases (three tumors, two non-

neoplasm

ing

vascular

A 12-year-old

child

with

a right

frontal vascular malformation underwent preoperative embolization to decrease intraoperative blood loss (Fig 2a, 2b). Preoperative angiography performed after partial embolization demonstrated thrombosis of the middte and inferoposterior portions of the

mass

(Fig

2c).

Intraoperative

CDI

allowed exact localization of the nonthrombosed portion of the matformation before resection, thus expediting the procedure (Fig 2d). In a i-year-old infant with a head tilt, a large lesion intensity occupying

Volume

with

mixed signal most of the left

or a cavernous

181

#{149} Number

malforma2

cus

of flow

deep

in the

posterior

fossa

abutting the tentorium (Fig 3c), which at pathologic examination represented residual venous angioma. This was impossible to distinguish from residual

blood

and

absorbable

gelatin

sterile sponge on the B-mode images, and CDI allowed successful removal of the entire venous angioma with cavernous changes before A i7-year-old adolescent tory of subacute bacterial

the

first

surgical

closure. with a hisendocarditis

procedure

and

masses,

and

one

cavernous

Radiology

#{149} 569

a. Figure 3. cerebellar level

b.

(a) Axial

hemisphere.

as that

consistent

T2-weighted

The

in a demonstrates with

a vascular

malformation,

but

fect (SD) in the occiput allowed scanning ing blood and absorbable gelatin sterile deep in the posterior fossa superimposed

angioma),

CDI

provided

C.

(2,500/30) MR image demonstrates center of the largest cyst has low signal areas of high and low signal intensity a cystic

neoplasm

a mass with mixed signal intensity in the posterior fossa occupying the left intensity. (b) Axial Ti-weighted (619/20) MR image obtained at the same suggestive of blood in various stages of decomposition. The findings are

could

not

be

excluded.

(c)

Intraoperative

CDI

in the posterior fossa. Abutting the tentorium (arrowheads), sponge in the posterior fossa. CDI permits accurate identification with the color map (photographed in black and white).

there

has

been

inverted.

is echogenic of the residual

A surgical

material venous

de-

(B) representangioma (A)

information

but did not justify altering the procedune. In all of these cases, CDI rapidly allowed identification of the least vascular path for biopsy, debulking, or resection. as planned,

whether

All would however,

utilized. was added with CDI in the four drainage procedures. Laser fenestration was used to drain one arachnoid cyst and one obstructed dysmorphic ventricle. A second arachnoid cyst was drained with a ventriculoperitoneal shunt catheter. Posterior fossa decompression related to Amold-Chiari I malformation was also performed. Although CDI was perNo

formed

CDI

have proceeded regardless of

was

information

in each

case,

no information

added to the procedures, undertaken as planned.

which

was

were

a.

b.

Figure

4.

(a) Anteropostenior view during angiography obtained with injection into the left vertebral artery demonstrates the remaining aneurysm in the right occipital lobe (arrow) close to the surgical clips from previous aneurysm resections (arrowheads). The right calvanial flap was subsequently lifted for reoperation. (b) Axial CDI obtained before the dura was opened demonstrates flow to the bulbous mycotic aneurysm (M), which lies 1.3 cm deep to the dura. A crescentic hyperechoic area (I) superficial to the flow represents the thrombosed portion of

the aneurysm.

DISCUSSION Intraoperative evaluation of vascutar malformations of the brain (5-7) and spinal cord (4,8) includes localization of the malformation, determination of the extent and location of the feeding arteries and draining veins, and confirmation of the completeness of resection. The four vascular matformations

resected

the

addition US, although with spinal

in our

series

support

of CDI to intnaoperative we have no experience malformations. The three-

dimensional location of a small AVM may be rapidly confirmed in the oper-

ating

room.

570

#{149} Radiology

The

neurosungical

ap-

proach

to the

versus

formations operating sets

resection

of thrombosed

nonthrombosed

vascular

varies, since time is focused

surrounding

the

mat-

most of the on the yes-

nonthrombosed

portion of the malformation. The ability to determine the results of preoperative embolization with CDI and to identify the remaining arteriovenous shunting immediately before invasion allows alteration of the surgical approach when necessary.

mation

Assessment

may

of residual

be based

on visual

malfor-

in-

spection, intraoperative angiography, on intraoperative US. Intraoperative CDI allows rapid identification of malformations intraoperative

in three angiography

two. Differentiation residual malformation, and absorbable gelatin strates

within

since idly

the

wound

all three identifies

may

planes,

while demon-

among hematoma, sterile sponge be difficult,

are echogenic. blood

flow

CDI within

rapresid-

ual malformation, however, allowing complete resection in our case. As CDI becomes readily available, the November

1991

a.

b.

d.

C.

Figure

5. (a) Coronal CT scan demonstrates hypertelorism and a nasofrontal cephalocele. While no significant difference is noted in the attenuation of the frontal regions, a clear cleavage plane (arrow) is contiguous with the falx, although the falx is shifted to the left. (b) Right parasagittal Ti-weighted (600/15) MR image demonstrates a hypoplastic right frontal lobe abutting an extraaxial subfrontal mass. Flow void is identified in branches arising from the anterior cerebral artery (arrowheads), which supply the mass. At surgery, the surface of the subfrontal mass was similar to that of the remaining brain tissue except for a small area that had surface markings resembling the cerebellum. The echogenicity was similar to that of the remaining brain tissue at B-mode scanning. (c) Coronal CDI obtained through the anterior portion of the hamartoma (H) demonstrates deviation of the anterior cerebral arteries (arrows) lying

nal

within

CDI

the

falx

to

demonstrates

the

the

left

of the

abnormal

midline

vessels

near

the

mass

(arrowheads)

bral artery (arrow), thus identifying the cleavage mal brain tissue, which was not otherwise visible

plane

of mixed

arising

between

echogenicity.

from

the

(d)

right

Coro-

anterior cere(H) and nor-

the hamartoma

presently available and the sometimes small surgical opening in the calvaria have been reported (5,6). CDI was not used in three mtraoperative cases because of limitations imposed by transducer size. This limitation has largely been avoided by means of consultation among the US staff, neuroradiotogists, and neurosurgeons before the operative procedure. When neurosurgeons become familiar with the transducer sizes available, an adequate surgical window is usually possible. In conclusion, we have found intraoperative CDI useful in a number of different neurosurgical procedures involving the treatment of vascular malformations, aneurysms, congenital malformations, and tumors. Structural or flow information, identified in seven of 17 cases, was not apparent to the neurosurgeons from the preoperative studies. Most of these cases were vascular malformations about which CDI provided exclusive information for localization or residual detection. CDI added no information to that obtamed with real-time US scanning during percutaneous embolization of a vascular malformation or drainage of cerebrospinat fluid collections. With the development of smaller scan heads, intraoperative applications of CDI may expand. Preoperative consultation among the US staff, neuroradiologists, and neurosurgeons optimized US scanning conditions, which led to additional information foltowed by procedural alterations in seven of our i7 cases. #{149} References 1.

at surgery. 2.

need for intraoperative angiography and for repeated surgery because of incomplete resection may decrease. CDI may be used to determine whether the neck of a giant aneurysm has been adequately clipped by depicting the presence of residual blood flow (6). Since mycotic aneurysms occur in a variety of locations deep within the brain, they may be especially difficult to locate on the basis of preoperative imaging studies alone.

was

Mycotic

scanning in a plane perpendicular to the vessel to be visualized, since such scanning also limits detection of flow with both duplex US and CDI. In addition to the pitfalls in detecting flow, such as suboptimal scan angle, low-velocity flow, and small yes-

aneurysms

with extradural neurosurgeon the aneurysm mycotic

series, fication rysm;

not

be identified

CDI, allowing the to dissect directly to site. While only one

aneurysm

was

present

CDI allowed immediate of the location of the thus,

visible

Volume

can

181

the

in our

identianeu-

thrombosed

portion

on preoperative

studies

#{149} Number

2

identified,

resulting

ity

flow.

Care

sel diameters,

relatively

must

be taken

limitations

large

size

of the

sonography

in precise

resection. All of the neoplasms in our series appeared avascular with CDI, and this unanimity correlated with the preoperative imaging studies. Vascular masses may not be recognized as such with CDI, however, because of the inability to identify small vessels ( < 0.6 mm in diameter with present equipment) or vessels with low-veloc-

3.

5.

to the

transducers

6.

8.

brain

tumor

Radiology

1985;

localization

AJR 1982;

and

139:

Efficacy of intraintracranial 157:509-511.

Rubin JM, DePietro MA, Chandler WF, Venes JL. Spinal ultrasonography: ultraoperative and pediatric applications. Radiol Clin North Am 1988; 26:1-27. Rubin JM, Hatfield MK, Chandler WF, Black KL, DiPietro MA. Intracerebral arteriovenous malformations: intraoperative color Doppler flow imaging. Radiology 1989; 170: 219-222. Black KL, Rubin JM, Chandler WF, McGillicuddy

7.

for

ventricular shunt placement. 733-738. Rubin JM, Dohrmann GJ. operative US for evaluating masses.

4.

to avoid

due

Chandler WF, Knake JE, McGillicuddy JE, Lillehei KO, Silver TM. Intraoperative use of real-time ultrasonography in neurosurgery. J Neurosurg 1982; 57:157-163. Knake JE, Chandler WF, McGillicuddy JE, Silver TM, Gabrielsen TO. Intraoperative

JE.

Intraoperative

len imaging

of AVM’s

Neurosurg

1988;

color-flow

and

Dopp-

aneurysms.

68:635-639.

Nornes H, Grip A, Wikeby P. Intraoperative evaluation of cerebral hemodynamics using directional Doppler technique. J Neurosurg 1979; 50:145-151. Rubin JM, Knake JE. Intraoperative sonography of a spinal cord arteriovenous malformation. AJNR 1987; 8:730-731.

Radiology

#{149} 571