Masatoshi Masahiro

Kudo, Hirasa,

MD MD

Hepatic Specific US with

Shusuke Tomita, Akio Todo, MD

a a

MD

a

Hitoshi

Tochio,

RT

a

Hiroshi

Kashida,

W

the recent advances in imaging technology including the development of ultrasonography (US), computed tomography (CT), single photon emission computed tomography (SPECT), and magnetic resonance (MR) imaging, the mcidental discovery of asymptomatic small liver tumors has been increasing. In Asia, where hepatocellulan carcinoma (HCC) is endemic, differentiation of benign hepatic nodules from HCC is one of the most important considerations in the fields of gastroenterology and radiology. Focal nodular hyperplasia (FNH) is a rare benign hypervascular lesion (i) that is often found incidentally and is confirmed by the postoperative pathologic study of the mesected specimen. However, since it is a benign nodule, it would be preferable to distinguish it from a malignant vascular tumor before surgery, thus avoiding unnecessary surgical intervention. A large FNH has hemodynamic and histopatho!ogic characteristics that permit accurate diagnosis with imaging modalities (2-4). However, with a small FNH, it is difficult to recognize characteristic radiologic findings, such as the “spoke-wheel sign,” with angiography. Recently we have examined four patients with histologically proved small FNH less than 3 cm in diameten. The characteristic findings of FNH were observed with dynamic contrast material-enhanced US dunITH

ing

hepatic

PATIENTS

study

the

AT.)

761.3119

Liver

Liver,

#{149} Liver

focal

nodular

neoplasms,

neoplasms.

neop!asrns,

diagnosis,

761.3114,

neoplasms,

secondary,

761.33

plasms,

US

studies,

761.12988

14, 761.321

761.124

#{149} Liver

761.321 a

Liver

a

neo-

Liver

and

Section

Service (H.T.), Minatojima-Nakamachi,

hyperplasia.

761.31

angiography.

a

1991;

179:377-382

of Castroenterology, (M.K.,

ST.,

of Abdominal Kobe

City

De-

H.K.,

M.H.,

Ultrasound

Ceneral Chuo-ku,

Hospital, 4-6, Kobe 650, Ja-

of

METHODS

FNH

smaller

(three

underwent

went

percutaneous

ing

144

of

3 cm in diameter

biopsy). had

HCC

(n

a total

113),

=

nodule

(n

nodule (n = 5), hemangioma

tastasis granuloma

(n

=

pemplastic

(n

the

one

under-

The

remain-

of

167

hepatic

hy14),

(i

(n

me-

4), and

=

1).

US was dioxide

catheter

patients

adenomatous 6), regenerative

24), lymphoma

Dynamic ing carbon

artery branch)

148

resection,

patients

nodules:

Four

than

performed microbubbles

placed

in

the

by

injectthrough

proper

hepatic

(or in the right or left hepatic after conventional superse!ective

hepatic

angiogmaphy.

catheter

tip

for

The dynamic

position US

of the

was

the

same

as for conventional angiography. Injection of carbon dioxide microbubbles was performed slowly, at a rate of approximately

2 mL/sec,

evaluation tern within early the

arterial

phase.

patients

them

period,

to hold

accurate

dioxide dioxide, saline,

blood

given

early

US

to study

breath an

for a

accurate

pattern arterial

of the phase.

microbubbles

were

mixing

10 mL

vigorously

normal

their

vascular

at the

patat the

was

dynamic

permits

of the

nodule

carbon

own

the

which

observation

by

the

Oxygen

during

to enable long

permitting

of the dynamic vascular the nodule, especially

preof

10 mL of heparinized

and 5 mL of the patient’s with use of two syringes con-

nected to each cock. The total

other by a three-way volume of microbubbles

stop-

pan. Received October 3, 1990; revision requested October 31; revision received Decernber 12; accepted December 26. Supported in part by Medical

a grant Research

ta!. Address Radiology

Division

of Medicine

AND

population.

had

pared From

infusion

From March 1989 to August 1990, dynamic US was performed on 148 patients with histologically proved hepatic tumoms. These patients constituted our

Carbon

partment

arterial

carbon dioxide microbubbles. In this article, these findings will be described, and new diagnostic criteria for FNH will be discussed and cornpared with the findings observed in HCC and other hepatic tumors.

small

terms:

a

Focal Nodular Hyperplasia: Findings at Dynamic Contrast-enhanced Carbon Dioxide Microbubbles’

Dynamic contrast material-enhanced ultrasonography (US) with intraarterial infusion of carbon dioxide microbubbles was performed for four cases of histologically proved focal nodular hyperplasia (FNH) in four patients and for i67 cases of various hepatic nodules in 144 patients. No complications due to dynamic US were observed in any of the 148 patients. All FNH nodules were less than 3 cm in diameter. Consistent specific findings of FNH were not obtained with US, computed tomography, magnetic resonance imaging, radiocolloid scanning, or angiography in the four cases of FNH. In contrast, the characteristic vascular pattern (ie, early central hypervascular supply with centrifugal filling to the periphery at the arterial phase and a uniform or lobulated dense stain at the capillary phase) was observed in all four cases of FNH with dynamic US. This vascular pattern demonstrated in FNH with dynamic US was not seen in any of the i67 hepatic nodules, including 44 small hepatocellular carcinomas less than 3 cm in diameter. Therefore, the newly developed, dynamic contrastenhanced US technique seems to be extremely sensitive and specific for diagnosing FNH and is useful in the differentiation of FNH from other hepatic tumors, especially hepatocellular carcinoma.

Index

MD

C

RSNA,

from

reprint 1991

the Grant-in-Aid of Kobe City General

requests

to M.K.

1989 for Hospi-

Abbreviations: plasia,

SPECT

FNH

HCC =

=

single

focal

hepatocel!ular photon

nodular

hyper-

carcinoma, emission

computed

to-

mography.

p7,7

injected US

was

adequately

monitoring

determined

to avoid

attenuation

with

producing

of the deeper

ma. Dynamic images corded on a videotape

strong

liver

pamenchy-

of all cases were recorder during

me-

the period prior to the injection dioxide microbubbles until the

of carbon bubbles

were

paren-

washed

chyma.

out from

The

with three

vascular

dynamic patterns,

vascularity

the liver distributions

US were depending

compared

seen

classified into on the tumor

with

the surround-

ing liver vascular,

parenchyma: hypervascular, isoand hypovascular. The vascular patterns of the four cases of FNH and 44 cases of HCC smaller than 3 cm in diame-

ter (24 patients underwent

underwent

percutaneous

precisely

biopsy)

compared.

linear-array

or 5.0-MHz studies Most scanning

20 were

A convex-

and/or

electronic

scanner

with

transducer

was used

a 3.5-

for the

(SSD-650; Aloka, Tokyo). of the 148 patients underwent CT with a 9800 scanner (GE Medi-

cal Systems,

Milwaukee)

administration imaging

before

of contrast was

imager

netom;

Siemens,

weighted with

with

operating

1.5 T (Mag-

Germany).

images

a repetition

MR

a supercon-

at

Erlangen,

spin-echo

and after

material.

performed

ducting

tamed

resection,

were

time

Ti-

ob-

of 600 msec

and an echo time of 15 msec. T2-weighted spin-echo images were obtained with a repetition time of 2,000 msec and an echo time with

of 70 msec. SPECT was performed a model 400T or Starcam 400AC

tating gamma camera (GE Medical Systems). nous

injection

and minicomputer After an intrave-

of 185 MBq

technetium-99m

phytate,

liver scintigmaphy formed.

ro-

(5 mCi) conventional

and SPECT

were

of

per-

RESULTS FNH The nadiologic findings in the four cases of FNH are listed in Table 1 All four patients had normal liver function and had no symptoms related to FNH. US showed no specific finding except in patient 4 in whom the “central scar sign” was seen. One FNH nodule was not demonstrated with CT and MR imaging, probably because of its size (patient 2). CT and MR imaging findings were not specific for FNH except in patient 4, in whom a central fibrous scar was demonstrated with MR imaging. The spoke-wheel sign was demonstrated with angiography in only one case (patient 3), and in all four cases a dense capillary stain was seen. In two of four cases of FNH, decreased opacity centrally was demonstrated on .

the

venous

In the US

phase

four

showed

cases consistent

of angiography.

of FNH, findings

dynamic of a

hypenechoic spot at 1 second after jection of carbon dioxide, centrifugal 378

a

Radiology

in-

filling pattern to the periphery at 2-5 seconds after injection, and dense staining at the late phase after injection. In one case, a 61-year-old man (patient 1) was referred for US examination to screen for liver metastasis aften an operation for scrotal Paget disease, which is a carcinoma arising in the apocnine gland epithelium. The results of blood tests showed normal liver function. Abdominal US mevealed a homogeneous hypoechoic nodule 1.8 cm in diameter in the posterosupenion area of the liver. The nodule was demonstrated as a highattenuation mass on a CT scan obtamed with contrast material enhancement. The lesion was homogeneously hypointense on Ti-weighted sequences and homogeneously hypenintense on T2-weighted sequences at MR imaging. A SPECT madiocolloid liver scan showed normal distribution of radioactivity. Hepatic angiognaphy revealed the nodule to be hypervascular with a dense stain at the capillary phase, but the spokewheel sign was not demonstrated (Fig la, ib). Dynamic contrast-enhanced US cleanly revealed the antenial supply arising from the center of the nodule at 1-2 seconds after infusion of carbon dioxide (early arterial phase). The arterial supply gradually radiated to the periphery at 3-4 seconds after infusion of carbon dioxide (arterial phase), and a hypervascular stain with sharp margin was observed at 10 seconds after infusion (capillary phase), which is diagnostic of FNH (Fig ic). The resected nodule was a well-circumscribed, but nonencapsulated, nodular mass in an otherwise normal liver. Pathologic study of the resected nodule revealed a central, stellate fibrotic scar with connective tissue septa radiating to the periphery.

The

hepatocytes

were

C .

0.

C (05’

u

bc 0

u0.

++++

0

C IL.

.L Ic

++++

Cu’ 5’

u. -

(50 5)

I

I ++

50

I

1+1,

5)0

L)3 -C 0. (0 bO

0 50 C

5) 5)

-C .

0 0.

‘5

50

Iz

I ±

0

5’

50

C

-C 50

50

a)

s

0.

0.0.

‘

(5

0 0.

00 0.0.

IZI

I C

0 I-

5) 0

C

U

5) .C

5)

C

.p

. .

0.

.

.C

C

+

C

C

z

cy-

tologically normal, and abundant dilated arterial vessels and bile duct proliferation were seen in the central scar and in the radiating septa (Fig id). In another case, a 68-year-old woman (patient 2) was referred to our hospital for further evaluation of a hepatic nodule incidentally detected with abdominal US. The results of blood tests showed normal liver function. There was no history of the patient taking oral contraceptives. Abdominal US revealed a homogeneous hypoechoic nodule 1.0 cm in diameter with a clear margin in the anterosupenior area of the right lobe of the liver. The nodule was not dernonstrated with CT or MR imaging.

‘

(0

0

5)

00

5)

0.

0. 0

.

E

(5

C

(5(5

E

uu2 OO

0

0.

0.

0

zzzz

0

0

0 5) U 0 5)

C)

‘U

bO

0

C

111+

0

lv

z

0 ia.1

0000 0. 0.

-

0.

0.

.

..‘

C)

0

50 5) 0.

‘0 0

r’4 ,-

,-‘

0i

-

‘LI

II

U

00 e

.2

-

.

:

00

‘

0

c’

IL.

IL.

5)5)

z 0.

:

5;

z*

0.

May

1991

d.

b.

a.

Plain

isec

2sec

3sec

4sec

lOsec

C.

Figure 1. (a) Hepatic capillary phase shows lobe (arrow). al radiation the capillary stain; original

angiogram

obtained

a dense

stain

during

with

a clear

Dynamic contrast-enhanced to the periphery at 2-3 seconds, phase (10 seconds), diagnostic magnification, Xi.)

Table 2 Vascular

Patterns

at Dy namic

the

arterial

margin.

US scans filling of FNH

phase

(c) Plain

with carbon of the nodule (arrow). (d)

U S of Histologic

ally Proved

Adenomatous Hyperplastic

Regenerative

shows

US scan

hypervascular

(intercostal)

tumor.

shows

dioxide microbubbles show at 4 seconds, and a uniform Cross section of the resected

H epatic

Hepatic

angiogram

obtained

hypoechoic

tumor

during

central hyperechoic spot at 1 second with gradudense stain in the nodule with sharp margin at specimen shows typical finding of FNH. (Silver

Nodules

FNH

3

6*

2

0

3

3

1

0

Isovascular

8

0

3*

0

0

0

0

0

102* 0 0 0

0 0 0 0

0 0 0 0

0 14* 0 0

5 0 16* 0

1 0 0 0

0 0 0 0

0 0 0 4*

6

5

14

24

4

1

4

supply

Total *

These

113 nodules

showed

typical

vascular

Metastasis

Granu!oma

Hypovascular Hypervascular Uniformormosaic Spottypooling Peripheral Central arterial

Hemangioma

Lymphoma

Nodule

Pattern

pattern.

HCC

Radiocolloid liver scanning was not performed. Hepatic angiography showed the nodule as hypervascular with a dense stain at the capillary

to the

phase,

not demonstrated (Fig 2a, 2b). Dynamic contrast-enhanced US clearly mevealed the arterial supply arising from the center of the nodule at the early

ule, radiating to the periphery, and no atypia in the hepatocyte, consistent with typical FNH (Fig 2d). Histopathologic examination showed abundant dilated arteries in the central fibrous

arterial

scar

Volume

but

the

spoke-wheel

phase

179

then

a

sign

gradually

Number

2

the

in the right

HCC

Vascular

Nodule

(b)

a homogeneous

was

radiating

periphery,

and

a hypervascular

stain was noted at the capillary phase (Fig 2c). Pathologic study revealed fibrous

septa

and

in the

in the

center

radiating

of the

septa.

nod-

The characteristic observed in FNH strated

with

vascular were not

dynamic

patterns demon-

contrast-en-

hanced US in any of the 1 i3 histologically proved HCCs. One hundred two HCC nodules showed a hypervascular nodules

pattern, showed

while three a hypovascular

Radiology

HCC pat-

a

379

a.

b.

d.

C.

Figure

2. the

during poechoic

arterial

(a) Hepatic angiogram capillary phase shows tumor in the right lobe

vascular

stain Xl.)

(10

seconds).

pattern (d)

of FNH Cross

obtained during the arterial hypervascular tumor with of the liver (arrow). Dynamic

(arrow)

section

(ie, central

of the

resected

phase shows hypervasculam a dense stain (arrow). (c) Plain contrast-enhanced US scans

hypervascu!ar specimen

supply) shows

arterial

phase

compared

those

arterial

radiation

of FNH.

HCCs

in FNH

vascular

HCCs arose from infiltrated to the

seconds),

finding

in 44 small

with

3). The

(1-2

typical

tumor (arrow). (b) Hepatic angiogram obtained US scan (intercostal) shows a homogeneous hywith carbon dioxide microbubbles show typical

were

(Table

supply

in a!!

the periphery and center in a willow-

branch fashion in 37 of 44 HCCs (Fig 3). In seven of 44 HCCs, the filling

pattern of the nodule determined because filling

within

phase lary

stain

HCCs,

tern,

and

eight

showed

lam pattern (Table 2). The namic patterns, especially

vascular 380

a

supply Radiology

pattern,

an isovascuhemodythe arterial

at the

early

was

and

in four. lary

the

stain

Other

(Table

Hepatic

At

the

US, a dense observed

a faint

Two

could not be of simultaneous

nodule.

of dynamic

stain

HCCs

late

demonstrated

vascular

with

dynamic

with

cu!ar

at

seconds),

original

gradually over time very late dynamic

and

dense

magnification,

US

early

arterial

(ie,

hypovasphase

and

infiltrating to the center with spotty pooling at the phase) reflecting the hemocharacteristic of extremely

blood vascular

was

observed

tumors

no

lam

in

dynamic the

The

flow

were 67%

within

the

patterns

of

peripheral (16

of

24

tumor.

metastatic

hypervascutumors),

uniform

or mosaic hypervascular in 2i% of 24 tumors), and hypovascular 13% (three of 24 tumors). None

Tumors

were

tern

slow

capi!-

(2-3 stain;

generative nodules, two were demonstrated as hypovasculam, and three were isovascular. All i4 hemangiomas showed a typical vascular pat-

in 38 of 44

3).

nodules

trichrome

capil-

showed

A!! six adenomatous

to the periphery

(Masson

hepatic

hyperplastic as hypo-

US.

Of five

tumors,

demonstrated lam supply

me-

No

other

than

a central pattern

complications

FNH,

arterial

(Table

were

(five in of the vascu-

2).

observed

May

in

1991

Plain

isec

2sec

3sec

10 sec

4sec

a. Figure 3. right lobe lar supply ion,

(a) Plain US scan (intercostal) shows a homogeneous hypoechoic tumor in the of the liver (arrow). Dynamic contrast-enhanced US scans show the arterial vascuarising from the periphery and infiltrating to the center in a willow-branch fash-

a typical

finding

ab!ed confirmation

b.

the i48 patients namic US.

who

underwent

dy-

an

essential

for the

are

ing

The pathologic findings usually distinctive and

istic.

FNH

consists

of FNH character-

of hepatocytes;

Kupffer cells; bile duct elements; a central, stellate fibrotic scar with connective tissue septa; and a specific antenial vascular structure (5-7). The madiologic diagnosis of FNH, based on the results of radiocolloid liven scanning,

hepatobiliany

scanning,

US,

CT, MR imaging, and angiography, has focused on the identification this

pathologic

(1-8).

signature

of How-

even, as seen in our presented cases, although US, CT, and MR imaging are sensitive in the detection of FNH, they have been less specific in the agnosis of FNH than scmntigraphy angiography (1-4,6,8).

The

diagnosis

tamed

if the

of FNH

can

radiocolloid

liven

reveals normal or increased the lesion (1,8). However, graphic findings for small sions

are

poor

resolution

raphy

the

not

for

use

reliable

space-occupying 2 cm in diameter only ii%-45%

if normal Volume

because

the

179

of the

Even

ability

lesions smaller was reported (9,iO). Therefore,

radiocolloid Number

#{149}

scan

of scintig-

lesions.

of SPECT,

be ob-

uptake in scintiFNH le-

capability

small

dion

uptake 2

with

to detect than even

is ob-

Tumor

procedure

differential

FNH

but

not

diagnosis

also

for

only

includ-

treatment

plan-

ning. Moreover, histologic diagnosis of the relatively small specimen obtamed at percutaneous biopsy is not always pathognomonic for FNH since it is difficult to determine if the nodule consists of normal hepatocytes on if a sampling error has occumred when normal hepatocytes are obtained. Therefore, angiography memains agnosis

of considerable of FNH. The

value in the angiographic

di-

diagnosis is based on the demonstration of the nodule’s pathologic vasculam feature; namely, the arteries enter into the center of the FNH and then fill the nodule in a centrifugal dinection as demonstrated at the three-dimensional pathologic study (7). The characteristic angiographic appearance of FNH is, therefore, a hypervascular, septated lesion with a central vascular supply or central spokewheel arterial pattern (1,2,6) and a uniform dense blush gins at the capillary

with phase

sharp man(8). How-

ever, since this typical finding can be detected in only 57%-90% of cases of large FNH (1,3), it is more difficult to detect this vascular pattern with conventional angiography in small FNH nodules,

as

(arrow).

served in hepatic lesions smaller than 2 cm in diameter, such as in patients 1-3 in our series, a definitive diagnosis of FNH cannot be made. Although percutaneous biopsy with US guidance is a very useful procedure in the differential diagnosis of hepatic tumors, angiography is still

DISCUSSION

of HCC

especially

cm in diameter phy

in three

vessel

is evident.

(b)

Resected

specimen

en-

of HCC.

those

(1,3,8). of four

less

than

3

At angiognacases

of FNH

de-

scnibed

in this

article

the

nonspecific

finding of hypervasculan stain at the capillary phase was seen; this mimics the finding with HCC. Dynamic contrast-enhanced US with arterial infusion of carbon dioxide microbubbles was first applied to the diagnosis of hepatic tumors by Matsuda and Yabuuchi (ii). This technique is a combination of US and angiography and is considered extremely sensitive in the detection of tumor vascular pattern even if the tumom is very small. Matsuda and Yabuuchi found that findings on sonograms enhanced with carbon dioxide microbubbles and obtained by means of their technique are similar to the tumor stain seen in the venous phase of angiography. Moreover, since contnast-enhanced US is sensitive in the detection of arterial vasculanity because of its tomographic imaging capability, it is superior to angiography in most

cases.

We modified two ways. First, method

their technique we modified

of preparing

the

in their

carbon

di-

oxide microbubbles to simplify the technique of dynamic contrast-enhanced US. We made carbon dioxide microbubbles by agitation of a mixtune of carbon dioxide, hepaminized normal saline, and the patient’s own blood, instead of using a solution of 20%

glucose

and

Our modified carbon dioxide sidemed solution

25%

albumin

(ii).

method of preparing microbubbles is con-

much simpler since necessary, including

any

extra hepa-

rinized normal saline, is routinely used at the conventional angiographic procedure. It was proved with microscopic

measurement

that

the

Radiology

a

381

mean size of the microbubbles made with our method was 34 rn ± 18, similar to the size of those made by Matsuda and Yabuuchi (1 1), and moreover, the flow of microbubbles can be satisfactorily regarded as the same as blood flow. Second, we evaluated arterial flow by making a videotape recording from the start of the carbon dioxide infusion until the time when carbon dioxide was washed out of the liver parenchyma, instead of evaluating only at the late phase after injection of carbon dioxide as was done by Matsuda and Yabuuchi. By observing the vascular flow from the early phase to the late phase, more accurate evaluation of the vascular pattern of the hepatic nodule is possible. With our method, early, middle, and late phases at dynamic US are assumed to correspond to the arterial, capillary, and venous phases at conventional angiography, respectively. Our method of dynamic US, therefore, permits much more precise three-dimensional evaluation of anterial flow, not only in the capillary phase but also in the arterial phase, than does conventional angiography or the contrast-enhanced US technique used by Matsuda and Yabuuchi (11). In our four cases of FNH smaller than 3 cm in diameter, the typical vascular pattern (ie, the vascular supply to the tumors arising centrally and radiating peripherally, as well as a uniform or lobulated dense hypervascular

stain

at the

capillary

phase)

was clearly demonstrated with our technique. This vascular pattern was not observed with any of the other histologically proved hepatic nodules, suggesting that these findings are specific for FNH (Table 2). Since dynamic US is an extremely sensitive method for detecting arterial vascularity within the tumor, even faint arterial flow, which is not demonstrated at conventional angiography, is expected to be detected in the case of a hypovascular FNH (i). However, further study is necessary to clarify this point. CT arteniography (12,13) is a technique that combines CT and angiography and is considered highly sensitive in the detection of small hepatic

nodules. that with

Takayasu et al (14) reported CT arteriognaphy the dilat-

ed

was

sary.

382

a

Radiology

artery

demonstrated

in the

center of nodules in 100% of cases of large FNH, and they stated that this would be a specific finding for FNH. Although, to our knowledge, there has been no other case report of small FNHs that were diagnosed with CT arteniography, CT arteniography is potentially a sensitive diagnostic tool for detecting the arterial vascular pattern of FNH as well, when dynamic CT is performed. However, since the early arterial phase, which is the most important period in evaluating the vascular pattern of FNH, lasts only a few seconds after injection in small (less than 3 cm) nodules, CT arteniography seems of limited value in the confirmation of FNH due to the inherent limitation of the relatively long scanning time compared with the real-time scanning in dynamic contrast-enhanced US. In summary, dynamic contrast-enhanced US is a highly sensitive technique in the evaluation of tumor vascular patterns especially at the early arterial phase, and is useful in the diagnosis of FNH even if the lesion size is smaller than 3 cm in diameter. The specific arterial vascular pattern for FNH with dynamic contrast-enhanced US is early central hypervascular supply with centrifugal filling to the periphery and a uniform or lobuiated dense stain at the capillary phase, and this finding is extremely characteristic for FNH (Tables 2, 3). Moreover, since dynamic US is cornplementary to the conventional angiographic procedure, dynamic US is beneficial in the following ways: (a) it is easy to perform, (b) it produces no complications, (c) it takes negligible extra time to perform (less than 10 minutes), and (d) it provides specific findings in the diagnosis of FNH. Therefore, we emphasize that dynamic contrast-enhanced US should be included routinely with the angiognaphic procedure for a suspected FNH nodule found incidentally in a patient with a normal liver, when findings at conventional angiography are not specific. If the specific vascular pattern described above is seen with dynamic US, additional wonkup or surgery is unneces-

Acknowledgments: We thank Masatoshi Makuuchi, MD, Osamu Matsui, MD, and Kazutaka Yamamoto, MD, for their critical review of the manuscript.

Kozue

We

also

Mayurni

preparation

thank

Kristin

for their of the

Coulon

secretarial

and

support

in

manuscript.

References 1.

Rogers

JV, Mack

Hepatic

focal

ography,

CT,

phy. 2.

AJR

JC,

EK,

Sanders

RC.

in focal

ular

hyperplasia

of the

liver.

J Ultrasound

1982; D,

Pointreau mas and

CC, focal CT

Bruneton

CR,

N.

adenody-

1986;

GM,

160:53-

Quint

of hepatic

LE,

focal

characterization primary

J,

Drouil!ard

Hepatic hyperplasia:

Radiology

Glazer

imaging

malignant

et a!.

nodular

hy-

and

distinction

hepatic

tumors.

1987; 148:711-715.

Knowles

DM

hyperplasia study

Hum

Pathol

Fechner

and

Il, Wolff of the

logic

cal

JN,

Vasile nodular

studs’.

58. Mattison

AJR

nod-

1:275-278.

Mathieu

from

7.

137:983-990.

Fishman

sign”

perplasia:

6.

scintigra-

“scar

MR

5.

angi-

and

sonographic

namic 4.

PC, et a!.

hyperplasia:

The Med 3.

Freeny

sonography.

1981;

Scatarige

LA,

nodular

and RE,

M.

liver:

review

of the

1976;

7:533-545.

Roehrn

JOF

pathologic nodular

liver.

literature. Angiographic

C, Adams

of focal

of the

nodular

of hepatic foAm J Surg Pathol

hyperplasia.

pathogenesis

plasia

Jr.

correlations

1977; 1:217-224. Wanless IR, Mawdsley

the

Focal

a clinicopatho-

R.

nodular

On

hyper-

Hepatology

1985;

5:1194-1200. 8.

D’Souza VJ, Focal nodular aging diol

9.

by

differing

1983;

modalities.

K, Mukai usefulness for

ceedings

of

eration

III Congress

ic liver

disease:

diology

1986;

Matsuda

12.

US contrast crobubb!es. Moss AA,

14.

of the

World

and

Takakuwa

detection

Fed-

Biology

H,

et a!. in chron-

with

SPECT.

Ra-

159:697-703.

Y, Yabuuchi

I.

Hepatic

tumors:

enhancement with CO2 miRadiology 1986; 161:701-705. Dean PB, Axe! L, et a!. Dynam-

masses

N, Miura

angiography

computed Pro-

carcinomas

ic CT of hepatic

Nakao

Ra-

1, et a!.

Medicine M,

and intraarterial 1982; 138:847-852.

13.

T, Dodo

hepatocellular

11.

NE, et a!. liver: im-

Pediatr

of emission liver scintigraphy.

of Nuclear

1982; 2866-2869. Kudo M, Hirasa Small

Watson of the

13:77-81.

Yamamoto Clinical tomography

10.

Sumner TE, hyperplasia

in

with

contrast

K, Takayasu hepatocellular

intravenous

material.

AJR

Y, et at.

CT

carcinoma.

J Comput Assist Tomogr 1983; 7:780-787. Takayasu K, Muramatsu Y, Moriyama N, et a!. Computed tomography: hepatic neoplasms hepatocellular

showing similar carcinoma.

features to Diagn Imaging

Abdomen

1990;

[Japanesej

10:22-30.

U

May

1991